1 2013-10-23 Ian Lance Taylor <iant@google.com>
3 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
4 (C++ Dialect Options): Likewise.
5 (Optimize Options): Likewise.
7 2013-10-23 Tom de Vries <tom@codesourcery.com>
9 PR tree-optimization/58805
10 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
12 2013-10-23 Jakub Jelinek <jakub@redhat.com>
14 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
15 sequence based on get_range_info returned range.
17 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
19 * tree-ssa.h: Remove all #include's
20 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
21 * alias.c: Move required includes from tree-ssa.h.
23 * builtins.c: Likewise.
25 * cfgexpand.c: Likewise.
26 * cfghooks.c: Likewise.
27 * cfgloop.c: Likewise.
28 * cfgloopmanip.c: Likewise.
30 * cgraphbuild.c: Likewise.
31 * cgraphclones.c: Likewise.
32 * cgraphunit.c: Likewise.
37 * fold-const.c: Likewise.
38 * ggc-page.c: Likewise.
39 * gimple-builder.c: Likewise.
40 * gimple-fold.c: Likewise.
41 * gimple-iterator.c: Likewise.
42 * gimple-low.c: Likewise.
43 * gimple-pretty-print.c: Likewise.
44 * gimple-ssa-strength-reduction.c: Likewise.
45 * gimple-streamer-in.c: Likewise.
46 * gimple-streamer-out.c: Likewise.
47 * gimplify.c: Likewise.
48 * graphite-blocking.c: Likewise.
49 * graphite-clast-to-gimple.c: Likewise.
50 * graphite-dependences.c: Likewise.
51 * graphite-interchange.c: Likewise.
52 * graphite-optimize-isl.c: Likewise.
53 * graphite-poly.c: Likewise.
54 * graphite-scop-detection.c: Likewise.
55 * graphite-sese-to-poly.c: Likewise.
56 * graphite.c: Likewise.
58 * ipa-inline-analysis.c: Likewise.
59 * ipa-inline-transform.c: Likewise.
60 * ipa-inline.c: Likewise.
61 * ipa-prop.c: Likewise.
62 * ipa-pure-const.c: Likewise.
63 * ipa-reference.c: Likewise.
64 * ipa-split.c: Likewise.
65 * ipa-utils.c: Likewise.
66 * loop-init.c: Likewise.
67 * lto-cgraph.c: Likewise.
68 * lto-section-in.c: Likewise.
69 * lto-section-out.c: Likewise.
70 * lto-streamer-in.c: Likewise.
71 * lto-streamer-out.c: Likewise.
72 * lto-streamer.c: Likewise.
73 * omp-low.c: Likewise.
75 * predict.c: Likewise.
76 * print-tree.c: Likewise.
77 * profile.c: Likewise.
79 * targhooks.c: Likewise.
81 * trans-mem.c: Likewise.
82 * tree-call-cdce.c: Likewise.
83 * tree-cfg.c: Likewise.
84 * tree-cfgcleanup.c: Likewise.
85 * tree-chrec.c: Likewise.
86 * tree-complex.c: Likewise.
87 * tree-data-ref.c: Likewise.
88 * tree-dfa.c: Likewise.
89 * tree-eh.c: Likewise.
90 * tree-emutls.c: Likewise.
91 * tree-if-conv.c: Likewise.
92 * tree-inline.c: Likewise.
93 * tree-into-ssa.c: Likewise.
94 * tree-loop-distribution.c: Likewise.
95 * tree-mudflap.c: Likewise.
96 * tree-nested.c: Likewise.
97 * tree-nrv.c: Likewise.
98 * tree-object-size.c: Likewise.
99 * tree-outof-ssa.c: Likewise.
100 * tree-parloops.c: Likewise.
101 * tree-phinodes.c: Likewise.
102 * tree-predcom.c: Likewise.
103 * tree-pretty-print.c: Likewise.
104 * tree-profile.c: Likewise.
105 * tree-scalar-evolution.c: Likewise.
106 * tree-sra.c: Likewise.
107 * tree-ssa-address.c: Likewise.
108 * tree-ssa-alias.c: Likewise.
109 * tree-ssa-ccp.c: Likewise.
110 * tree-ssa-coalesce.c: Likewise.
111 * tree-ssa-copy.c: Likewise.
112 * tree-ssa-copyrename.c: Likewise.
113 * tree-ssa-dce.c: Likewise.
114 * tree-ssa-dom.c: Likewise.
115 * tree-ssa-dse.c: Likewise.
116 * tree-ssa-forwprop.c: Likewise.
117 * tree-ssa-ifcombine.c: Likewise.
118 * tree-ssa-live.c: Likewise.
119 * tree-ssa-loop-ch.c: Likewise.
120 * tree-ssa-loop-im.c: Likewise.
121 * tree-ssa-loop-ivcanon.c: Likewise.
122 * tree-ssa-loop-ivopts.c: Likewise.
123 * tree-ssa-loop-manip.c: Likewise.
124 * tree-ssa-loop-niter.c: Likewise.
125 * tree-ssa-loop-prefetch.c: Likewise.
126 * tree-ssa-loop-unswitch.c: Likewise.
127 * tree-ssa-loop.c: Likewise.
128 * tree-ssa-math-opts.c: Likewise.
129 * tree-ssa-operands.c: Likewise.
130 * tree-ssa-phiopt.c: Likewise.
131 * tree-ssa-phiprop.c: Likewise.
132 * tree-ssa-pre.c: Likewise.
133 * tree-ssa-propagate.c: Likewise.
134 * tree-ssa-reassoc.c: Likewise.
135 * tree-ssa-sccvn.c: Likewise.
136 * tree-ssa-sink.c: Likewise.
137 * tree-ssa-strlen.c: Likewise.
138 * tree-ssa-structalias.c: Likewise.
139 * tree-ssa-tail-merge.c: Likewise.
140 * tree-ssa-ter.c: Likewise.
141 * tree-ssa-threadedge.c: Likewise.
142 * tree-ssa-threadupdate.c: Likewise.
143 * tree-ssa-uncprop.c: Likewise.
144 * tree-ssa-uninit.c: Likewise.
145 * tree-ssa.c: Likewise.
146 * tree-ssanames.c: Likewise.
147 * tree-stdarg.c: Likewise.
148 * tree-streamer-in.c: Likewise.
149 * tree-switch-conversion.c: Likewise.
150 * tree-tailcall.c: Likewise.
151 * tree-vect-data-refs.c: Likewise.
152 * tree-vect-generic.c: Likewise.
153 * tree-vect-loop-manip.c: Likewise.
154 * tree-vect-loop.c: Likewise.
155 * tree-vect-patterns.c: Likewise.
156 * tree-vect-slp.c: Likewise.
157 * tree-vect-stmts.c: Likewise.
158 * tree-vectorizer.c: Likewise.
159 * tree-vrp.c: Likewise.
162 * value-prof.c: Likewise.
163 * var-tracking.c: Likewise.
164 * varpool.c: Likewise.
165 * vtable-verify.c: Likewise.
167 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
169 * config/tilegx/tilegx.c: Include "tree.h".
171 2013-10-23 Jakub Jelinek <jakub@redhat.com>
173 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
174 the info, not after it.
175 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
176 dump_ssaname_info call.
177 (pp_gimple_stmt_1): Adjust caller.
178 (dump_phi_nodes): Likewise. Don't print "# " here.
180 2013-10-22 Jan Hubicka <jh@suse.cz>
182 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
183 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
184 * i386.c (expand_small_movmem_or_setmem): New function.
185 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
186 (alg_usable_p): Add support for value ranges; cleanup.
187 (ix86_expand_set_or_movmem): Add support for misaligned moves.
189 2013-10-22 Sterling Augustine <saugustine@google.com>
191 * doc/invoke.texi: Document -ggnu-pubnames.
192 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
194 * dwarf2out.c: Include gdb/gdb-index.h.
195 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
196 debug_generate_pub_sections.
197 (is_java, output_pubtables, output_pubname): New functions.
198 (include_pubname_in_output): Handle debug_generate_pub_sections at
200 (size_of_pubnames): Use new local space_for_flags based on
201 debug_generate_pub_sections.
202 (output_pubnames): Unify pubnames and pubtypes output logic.
203 Genericize comments. Call output_pubname.
204 (dwarf2out_finish): Move logic to output_pubnames and call it.
206 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
209 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
210 Remove CCCmode handling.
211 <case LTU>: Return 'c' suffix for CCCmode.
212 <case GEU>: Return 'nc' suffix for CCCmode.
213 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
214 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
215 (*sub<mode>3_cc_overflow): Ditto.
216 (*subsi3_zext_cc_overflow): Ditto.
218 2013-10-22 Steve Ellcey <sellcey@mips.com>
220 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
221 (AND (NOT OP1) (NOT OP2)).
223 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
225 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
226 meaning of merge-high and merge-low masks for little endian; avoid
227 use of vector-pack masks for little endian for mismatched modes.
229 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
231 * config/tilepro/tilepro.c: Include "tree.h".
233 2013-10-22 Andreas Schwab <schwab@suse.de>
235 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
238 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
240 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
241 [Ed Smith-Rowland]): New entries.
242 ([Stephen M. Webb]): Update.
244 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
246 * tree-ssa-ter.h: Remove duplicate copy of file contents.
248 2013-10-21 Marek Polacek <polacek@redhat.com>
251 * fold-const.c (fold_range_test): Return 0 if the type is not
254 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
256 * system.h: Move hwint.h include further down.
257 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
258 gcc_checking_asserts.
259 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
261 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
263 Fix volatile issues in optimize_bit_field_compare.
264 * fold-const.c (optimize_bit_field_compare): Bail out if
265 lvolatilep or rvolatilep.
267 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
269 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
270 and get_inner_reference returning different pmode for non-volatile
271 bit-field members dependent on flag_strict_volatile_bitfields.
272 * stor-layout.c (layout_decl): Remove special handling of
273 flag_strict_volatile_bitfields.
274 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
275 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
277 2013-10-21 Paulo Matos <pmatos@broadcom.com>
279 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
282 2013-10-21 Jeff Law <law@redhat.com>
284 * tree-ssa-threadedge.c (thread_through_normal_block): New
285 argument VISITED. Remove VISISTED as a local variable. When we
286 have a threadable jump, verify the destination of the jump has not
288 (thread_across_edge): Allocate VISITED bitmap once at function
289 scope and use it throughout. Make sure to set appropriate bits in
290 VISITED for E (start of jump thread path).
291 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
292 through a joiner if any edge on the path has a recorded jump thread.
294 2013-10-21 Ian Lance Taylor <iant@google.com>
296 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
297 don't imply that attributes can solve all problems.
298 (Directory Options): Fix typo.
300 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
302 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
303 extend and extend_add.
305 2013-10-21 Richard Biener <rguenther@suse.de>
307 PR tree-optimization/58794
308 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
309 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
311 2013-10-21 Richard Biener <rguenther@suse.de>
314 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
315 to (T) X for sign-changing conversions (or no conversion).
317 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
319 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
321 2013-10-20 Jan Hubicka <jh@suse.cz>
323 * config/i386/i386-tune.def: Add comment; organize into categories
325 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
327 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
328 argument. Update function comment.
329 (expand_set_or_movmem_via_rep): New function combining
330 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
331 (expand_movmem_via_rep_mov): Remove.
332 expand_setmem_via_rep_stos): Remove.
333 (expand_movmem_epilogue): Update calls correspondingly.
334 (expand_setmem_epilogue_via_loop): Likewise.
336 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
337 (expand_set_or_movmem_prologue): New function combining
338 expand_movmem_prologue and expand_setmem_prologue.
339 (expand_movmem_prologue): Remove.
340 (expand_setmem_prologue): Remove.
341 (expand_set_or_movmem_constant_prologue): New function combining
342 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
343 (expand_constant_movmem_prologue): Remove.
344 (expand_constant_setmem_prologue): Remove.
345 (promote_duplicated_reg): Allow vector-const0 value.
346 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
347 and ix86_expand_setmem.
348 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
349 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
351 2013-10-21 Diego Novillo <dnovillo@google.com>
353 * asan.c: Include tree.h
354 * bb-reorder.c: Likewise.
355 * cfgcleanup.c: Likewise.
356 * cfgloopmanip.c: Likewise.
357 * data-streamer-in.c: Likewise.
358 * data-streamer-out.c: Likewise.
359 * data-streamer.c: Likewise.
360 * dwarf2cfi.c: Likewise.
361 * graphite-blocking.c: Likewise.
362 * graphite-clast-to-gimple.c: Likewise.
363 * graphite-dependences.c: Likewise.
364 * graphite-interchange.c: Likewise.
365 * graphite-optimize-isl.c: Likewise.
366 * graphite-poly.c: Likewise.
367 * graphite-scop-detection.c: Likewise.
368 * graphite-sese-to-poly.c: Likewise.
369 * graphite.c: Likewise.
370 * ipa-devirt.c: Likewise.
371 * ipa-profile.c: Likewise.
374 * loop-init.c: Likewise.
375 * loop-unroll.c: Likewise.
376 * lower-subreg.c: Likewise.
377 * lto/lto-object.c: Likewise.
379 * reginfo.c: Likewise.
380 * tree-loop-distribution.c: Likewise.
381 * tree-parloops.c: Likewise.
382 * tree-ssa-strlen.c: Likewise.
383 * tree-streamer.c: Likewise.
384 * value-prof.c: Likewise.
385 * target-globals.c: Likewise.
386 * expr.h: Include tree-core.h instead of tree.h.
387 * gimple.h: Likewise.
388 * ipa-prop.h: Likewise.
389 * ipa-utils.h: Likewise.
390 * lto-streamer.h: Likewise.
391 * streamer-hooks.h: Likewise.
392 * ipa-reference.h: Include cgraph.h instead of tree.h.
393 * cgraph.h: Include basic-block.h instead of tree.h.
394 * tree-streamer.h: Do not include tree.h.
395 * genattrtab.c (write_header): Generate inclusion of
397 * genautomata.c (main): Likewise.
398 * genemit.c: Likewise.
399 * genopinit.c: Likewise.
400 * genoutput.c (output_prologue): Likewise.
401 * genpeep.c: Likewise.
403 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
405 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
407 (vec_unpacku_hi_v8hi): Likewise.
408 (vec_unpacku_lo_v16qi): Likewise.
409 (vec_unpacku_lo_v8hi): Likewise.
411 2013-10-20 Jan Hubicka <jh@suse.cz>
413 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
414 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
415 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
417 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
419 * config/mips/mips.h (ISA_HAS_WSBH): Define.
420 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
422 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
424 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
427 * system.h: Undef m_slot.
429 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
431 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
432 all elements for both endian flavors.
434 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
437 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
438 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
439 and SI_REG for 64bit SYSV ABI targets.
441 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
443 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
444 to multi_reg_return. Clarify that we are skipping USEs of multiple
445 return registers. Use bool type where appropriate.
447 2013-10-18 Jan Hubicka <jh@suse.cz>
449 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
451 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
452 (X86_TUNE_PUSH_MEMORY): Likewise.
453 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
454 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
455 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
457 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
458 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
460 (ix86_option_override_internal): Update to use tune features instead
463 2013-10-18 Cong Hou <congh@google.com>
465 PR tree-optimization/58508
466 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
467 statement that contains data refs with zero-step.
469 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
471 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
472 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
473 * gimple-low.c (gimple_check_call_arg,
474 gimple_check_call_matching_types): Move to cgraph.c.
475 * gimple-low.h: Remove prototype.
476 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
477 Relocate from gimple-low.c.
478 * cgraph.h: Add prototype. Don't include basic-block.h.
479 * gimplify.c: Add gimple-low to include list.
480 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
481 * tree-eh.c: Add gimple-low to include list.
482 * tree-nested.c: Likewise.
483 * cfgexpand.c: Add tree-ssa-address.h to include list.
485 * gimple-fold.c: Likewise.
486 * gimple-ssa-strength-reduction.c: Likewise.
487 * trans-mem.c: Likewise.
488 * tree-mudflap.c: Likewise.
489 * tree-ssa-loop-ivopts.c: Likewise.
490 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
491 (degenerate_phi_result): Move to tree-phinodes.c.
492 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
493 * tree-ssa-threadedge.c: Likewise.
494 * tree-vrp.c: Likewise.
495 * tree-phinodes.c (degenerate_phi_result): Relocate here.
496 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
497 * tree-phinodes.h (degenerate_phi_result): Add prototype.
498 * tree-ssa-copy.c: Include tree-ssa-dom.h.
499 * tree-ssa-forwprop.c: Likewise.
500 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
501 pass_data_cleanup_cfg_post_optimizing,
502 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
503 * tree-optimize.c: Delete File.
504 * graphite.c: Include tree-cfgcleanup.h.
505 * passes.c: Likewise.
506 * tree-cfg.c: Likewise.
507 * tree-profile.c: Likewise.
508 * tree-ssa-dse.c: Likewise.
509 * tree-ssa-loop-ivcanon.c: Likewise.
510 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
511 * tree-outof-ssa.c: Include sbitmap.h.
512 * tree-ssa-live.c: Likewise.
513 * tree-ssa-propagate.c: Likewise.
514 * tree-ssa-structalias.c: Likewise.
515 * tree-stdarg.c: Likewise.
516 * Makefile.in (OBJS): Delete tree-optimize.o.
517 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
518 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
519 * varasm.c: Include basic-block.h.
520 * cfgloop.h: Include function.h instead of basic-block.h
521 (bb_loop_depth): Move to cfgloop.c.
522 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
524 2013-10-18 Teresa Johnson <tejohnson@google.com>
526 * predict.c (probably_never_executed): Compare frequency-based
527 count to number of training runs.
528 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
530 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
532 * config/arm/arm.c (cortexa9_extra_costs): New table.
533 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
535 2013-10-18 Jeff Law <law@redhat.com>
537 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
539 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
540 (register_jump_thread): Add "registered_jump_thread" debug
542 * dbgcnt.def (registered_jump_thread): New debug counter.
544 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
546 * config/rs6000/rs6000.c: Include cgraph.h.
548 2013-10-18 Teresa Johnson <tejohnson@google.com>
550 * tree-ssa-tail-merge.c (replace_block_by): Update edge
551 weights during merging.
553 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
555 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
556 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
557 * tree-ssa-operands.h: Remove prototype.
558 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
559 * gimple.c (virtual_operand_p): Relocate from gimple.c.
560 * gimple.h: Add prototype.
561 * gimple-ssa.h: Include tree-ssa-operands.h.
562 * tree-dump.c: Add tree-cfg.h to include list.
563 * tree-ssa-alias.c: Add ipa-reference.h to include list.
564 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
565 * config/i386/i386.c: Don't include tree-flow.h.
566 * config/rs6000/rs6000.c: Likewise.
568 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
570 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
572 2013-10-18 Richard Biener <rguenther@suse.de>
574 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
575 or TYPE_UNSIGNED of integral types.
576 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
577 NULL_TREE for zero-precision integral types.
579 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
581 * config/aarch64/arm_neon.h
582 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
584 2013-10-17 Sriraman Tallam <tmsriram@google.com>
586 * opth-gen.awk: Define target_flags_explicit.
588 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
590 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
591 fields to the reg_addr array that describes the valid addressing
592 mode for any register, general purpose registers, floating point
593 registers, and Altivec registers.
594 (FIRST_RELOAD_REG_CLASS): Likewise.
595 (LAST_RELOAD_REG_CLASS): Likewise.
596 (struct reload_reg_map_type): Likewise.
597 (reload_reg_map_type): Likewise.
598 (RELOAD_REG_VALID): Likewise.
599 (RELOAD_REG_MULTIPLE): Likewise.
600 (RELOAD_REG_INDEXED): Likewise.
601 (RELOAD_REG_OFFSET): Likewise.
602 (RELOAD_REG_PRE_INCDEC): Likewise.
603 (RELOAD_REG_PRE_MODIFY): Likewise.
604 (reg_addr): Likewise.
605 (mode_supports_pre_incdec_p): New helper functions to say whether
606 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
607 (mode_supports_pre_modify_p): Likewise.
608 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
609 print the valid address mode bits for each mode.
610 (rs6000_debug_print_mode): Likewise.
611 (rs6000_debug_reg_global): Likewise.
612 (rs6000_setup_reg_addr_masks): New function to set up the address
613 mask bits for each type.
614 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
615 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
616 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
617 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
618 PRE_MODIFY are supported.
619 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
620 registers, instead of {src,dest}_av_p.
621 (rs6000_print_options_internal): Tweak the debug output slightly.
623 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
625 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
628 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
630 * tree-flow.h (struct omp_region): Move to omp-low.c.
631 Remove omp_ prototypes and variables.
632 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
633 (copy_var_decl): Relocate prototype from tree-flow.h.
634 * gimple.c (copy_var_decl): Relocate from omp-low.c.
635 * tree.h: Move prototype to omp-low.h.
636 * omp-low.h: New File. Relocate prototypes here.
637 * omp-low.c (struct omp_region): Make local here.
638 (root_omp_region): Make static.
639 (copy_var_decl) Move to gimple.c.
640 (new_omp_region): Make static.
641 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
642 * tree-cfg.c: Include omp-low.h.
643 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
644 * gimplify.c: Include omp-low.h.
645 * tree-parloops.c: Likewise.
647 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
649 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
650 operands are in memory, prefer to force non-matched operand 1 to
653 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
656 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
657 restrict TImode addresses to single indirect registers if both
658 -mquad-memory and -mvsx-timode are used.
659 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
660 we should emit load/store quad. Remove using %y for quad memory
663 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
664 constraints to allow load/store quad on machines where TImode is
665 not allowed in VSX registers. Use 'n' instead of 'F' constraint
666 for TImode to load integer constants.
668 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
670 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
672 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
674 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
675 handling of STACK_REG.
677 2013-10-17 Richard Biener <rguenther@suse.de>
679 PR tree-optimization/58143
680 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
682 (rewrite_to_defined_overflow): Likewise.
683 (move_computations_dom_walker::before_dom): Rewrite stmts
684 with undefined signed overflow that are not always executed
685 into unsigned arithmetic.
687 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
690 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
691 explicit isa flag to be an options variable, instead of using
692 global_options_set. Remove define from rs6000.h.
693 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
695 * config/rs6000/rs6000.c (rs6000_option_override_internal):
696 Initialize rs6000_isa_flags_explicit.
697 (rs6000_function_specific_save): Add gcc_options* parameter, so
698 that the powerpc builds after the 2013-10-15 changes.
699 (rs6000_function_specific_restore): Likewise.
701 2013-10-16 DJ Delorie <dj@redhat.com>
703 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
704 op is a REG before checking REGNO.
705 (rl78_alloc_physical_registers): Verify pattern is a SET before
708 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
710 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
712 (vec_unpacks_lo_v4sf): Likewise.
713 (vec_unpacks_float_hi_v4si): Likewise.
714 (vec_unpacks_float_lo_v4si): Likewise.
715 (vec_unpacku_float_hi_v4si): Likewise.
716 (vec_unpacku_float_lo_v4si): Likewise.
718 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
720 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
721 (vsx_concat_v2sf): Likewise.
723 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
725 * config/aarch64/aarch64.md
726 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
728 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
730 PR tree-optimization/58697
731 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
732 estimated_loop_iterations_int.
733 (max_stmt_executions_int): Call get_max_loop_iterations_int.
734 (get_max_loop_iterations_int): New. HWINT version of
735 get_max_loop_iterations.
736 * cfgloop.h: Add prototypes.
737 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
738 * loop-unroll.c (decide_peel_once_rolling): Call
739 get_estimated_loop_iterations_int.
740 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
741 * tree-ssa-loop-niter.h: Tweak prototypes.
743 2013-10-16 David Malcolm <dmalcolm@redhat.com>
745 * gengtype-parse.c (struct_field_seq): Ignore access-control
746 keywords ("public:" etc).
748 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
750 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
751 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
753 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
755 * config/arm/arm.opt (mlra): New option.
756 * config/arm/arm.c (arm_lra_p): New function.
757 (TARGET_LRA_P): Define.
759 2013-10-16 Paulo Matos <pmatos@broadcom.com>
761 * tree-core.h (tree_code_name): Remove.
762 * tree.h (get_tree_code_name): New prototype.
763 * tree.c (tree_code_name): Make static.
764 (get_tree_code_name): New function.
765 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
766 tree_class_check_failed, tree_range_check_failed,
767 tree_not_class_check_failed, omp_clause_check_failed,
768 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
769 wrapper get_tree_code_name instead of calling tree_code_name directly.
770 * tree-vrp.c (dump_asserts_for): Likewise.
771 * tree-dump.c (dequeue_and_dump): Likewise.
772 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
773 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
774 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
775 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
776 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
777 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
778 dump_gimple_omp_for): Likewise.
779 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
780 * tree-ssa-pre.c (print_pre_expr): Likewise.
781 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
782 * print-tree.c (print_node_brief, print_node): Likewise.
783 * gimple.c (gimple_check_failed): Likewise.
784 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
785 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
786 * config/mep/mep.c (mep_validate_vliw): Likewise.
787 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
788 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
790 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
792 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
795 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
797 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
798 (MULTILIB_MATCHES): Add multilib for -march=v8.
800 2013-10-15 Sriraman Tallam <tmsriram@google.com>
803 * optc-save-gen.awk: Add extra parameter to the save and restore
805 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
806 * tree.c (build_optimization_node): New parameter. Add extra parameter
807 to call to cl_optimization_save.
808 (build_target_option_node): New parameter. Add extra parameter
809 to call to cl_target_option_save.
810 * tree.h (build_optimization_node): New parameter.
811 (build_target_option_node): New parameter.
812 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
813 build_optimization_node and build_target_option_node.
814 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
815 (handle_pragma_push_options): Ditto.
816 * toplev.c (process_options): Ditto.
817 * opts.c (init_options_struct): Check for opts_set non-null.
818 * target.def (target_option.save): New parameter.
819 (target_option.restore): New parameter.
821 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
822 (ix86_pragma_target_parse): Ditto.
823 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
825 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
826 to build_optimization_node and build_target_option_node.
827 (rs6000_valid_attribute_p): Ditto.
828 (rs6000_pragma_target_parse): Ditto.
829 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
831 * config/i386/i386.h:
832 TARGET_64BIT_P: New Macro
833 TARGET_MMX_P: New Macro.
834 TARGET_3DNOW_P: New Macro.
835 TARGET_3DNOW_A_P: New Macro.
836 TARGET_SSE_P: New Macro.
837 TARGET_SSE2_P: New Macro.
838 TARGET_SSE3_P: New Macro.
839 TARGET_SSSE3_P: New Macro.
840 TARGET_SSE4_1_P: New Macro.
841 TARGET_SSE4_2_P: New Macro.
842 TARGET_AVX_P: New Macro.
843 TARGET_AVX2_P: New Macro.
844 TARGET_AVX512F_P: New Macro.
845 TARGET_AVX512PF_P: New Macro.
846 TARGET_AVX512ER_P: New Macro.
847 TARGET_AVX512CD_P: New Macro.
848 TARGET_FMA_P: New Macro.
849 TARGET_SSE4A_P: New Macro.
850 TARGET_FMA4_P: New Macro.
851 TARGET_XOP_P: New Macro.
852 TARGET_LWP_P: New Macro.
853 TARGET_ABM_P: New Macro.
854 TARGET_BMI_P: New Macro.
855 TARGET_BMI2_P: New Macro.
856 TARGET_LZCNT_P: New Macro.
857 TARGET_TBM_P: New Macro.
858 TARGET_POPCNT_P: New Macro.
859 TARGET_SAHF_P: New Macro.
860 TARGET_MOVBE_P: New Macro.
861 TARGET_CRC32_P: New Macro.
862 TARGET_AES_P: New Macro.
863 TARGET_PCLMUL_P: New Macro.
864 TARGET_CMPXCHG16B_P: New Macro.
865 TARGET_FSGSBASE_P: New Macro.
866 TARGET_RDRND_P: New Macro.
867 TARGET_F16C_P: New Macro.
868 TARGET_RTM_P: New Macro.
869 TARGET_HLE_P: New Macro.
870 TARGET_RDSEED_P: New Macro.
871 TARGET_PRFCHW_P: New Macro.
872 TARGET_ADX_P: New Macro.
873 TARGET_FXSR_P: New Macro.
874 TARGET_XSAVE_P: New Macro.
875 TARGET_XSAVEOPT_P: New Macro.
876 TARGET_LP64_P: New Macro.
877 TARGET_X32_P: New Macro.
878 TARGET_FPMATH_DEFAULT_P: New Macro.
879 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
880 * config/i386/i386.c (ix86_option_override_internal): New parameters.
882 Change ix86_tune_string to access opts->x_ix86_tune_string.
883 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
884 Change ix86_arch_string to access opts->x_ix86_arch_string.
885 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
886 Change ix86_pmode to access opts->x_ix86_pmode.
887 Change ix86_abi to access opts->x_ix86_abi.
888 Change ix86_cmodel to access opts->x_ix86_cmodel.
889 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
890 Change ix86_isa_flags_explicit to access
891 opts->x_ix86_isa_flags_explicit.
892 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
893 Change ix86_regparm to access opts->x_ix86_regparm.
894 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
895 Change ix86_preferred_stack_boundary_arg to access
896 opts->x_ix86_preferred_stack_boundary_arg.
897 Change ix86_force_align_arg_pointer to access
898 opts->x_ix86_force_align_arg_pointer.
899 Change ix86_incoming_stack_boundar_arg to access
900 opts->x_ix86_incoming_stack_boundar_arg.
901 Change ix86_fpmath to access opts->x_ix86_fpmath.
902 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
903 Change ix86_recip_name to access opts->x_ix86_recip_name.
904 Change ix86_stack_protector_guard to access
905 opts->x_ix86_stack_protector_guard.
906 Change ix86_tune_memcpy_strategy to access
907 opts->x_ix86_tune_memcpy_strategy.
908 Change ix86_tune_memset_strategy to access
909 opts->x_ix86_tune_memset_strategy.
910 Change global_options to access opts.
911 Change global_options_set to access opts_set.
912 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
913 Change TARGET_MMX to TARGET_MMX_P (opts->...).
914 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
915 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
916 Change TARGET_SSE to TARGET_SSE_P (opts->...).
917 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
918 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
919 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
920 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
921 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
922 Change TARGET_AVX to TARGET_AVX_P (opts->...).
923 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
924 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
925 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
926 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
927 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
928 Change TARGET_FMA to TARGET_FMA_P (opts->...).
929 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
930 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
931 Change TARGET_XOP to TARGET_XOP_P (opts->...).
932 Change TARGET_LWP to TARGET_LWP_P (opts->...).
933 Change TARGET_ABM to TARGET_ABM_P (opts->...).
934 Change TARGET_BMI to TARGET_BMI_P (opts->...).
935 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
936 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
937 Change TARGET_TBM to TARGET_TBM_P (opts->...).
938 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
939 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
940 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
941 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
942 Change TARGET_AES to TARGET_AES_P (opts->...).
943 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
944 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
945 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
946 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
947 Change TARGET_F16C to TARGET_F16C_P (opts->...).
948 Change TARGET_RTM to TARGET_RTM_P (opts->...).
949 Change TARGET_HLE to TARGET_HLE_P (opts->...).
950 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
951 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
952 Change TARGET_ADX to TARGET_ADX_P (opts->...).
953 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
954 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
955 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
956 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
957 Change TARGET_X32 to TARGET_X32_P (opts->...).
958 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
959 Change TARGET_FLOAT_RETURNS_IN_80387 to
960 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
961 (ix86_function_specific_save): New parameter. Use opts-> fields
962 to replace global fields.
963 (ix86_function_specific_restore): Ditto.
964 (ix86_valid_target_attribute_inner_p): New parameters.
966 Fix call to ix86_handle_option and set_option.
967 (ix86_valid_target_attribute_tree): New parameters.
968 Change global_options to access opts.
969 Change global_options_set to access opts_set.
970 Fix call to ix86_valid_target_attribute_inner_p.
971 Change ix86_tune_string to access opts->x_ix86_tune_string.
972 Change ix86_arch_string to access opts->x_ix86_arch_string.
973 Change ix86_fpmath to access opts->x_ix86_fpmath
974 Fix call to ix86_option_override_internal.
975 Fix call to ix86_add_new_builtins.
976 Fix calls to build_optimization_node and build_target_option_node.
977 (ix86_valid_target_attribute_p): Remove access to global_options.
978 Use new gcc_options structure func_options.
979 Fix call to ix86_valid_target_attribute_tree.
980 Fix call to build_optimization_node.
981 (get_builtin_code_for_version): Fix call to
982 ix86_valid_target_attribute_tree.
984 2013-10-15 David Malcolm <dmalcolm@redhat.com>
986 * Makefile.in (PICFLAG): New.
987 (enable_host_shared): New.
988 (INTERNAL_CFLAGS): Use PICFLAG.
989 (LIBIBERTY): Use pic build of libiberty.a if configured with
990 --enable-host-shared.
991 * configure.ac: Add --enable-host-shared, setting up new
993 * configure: Regenerate.
994 * doc/install.texi (--enable-shared): Add note contrasting it with ...
995 (--enable-host-shared): New option.
997 2013-10-15 Richard Biener <rguenther@suse.de>
999 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1000 for built-in functions.
1002 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1004 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1005 (optimize_range_tests_1): New function,
1006 extracted from optimize_range_tests.
1007 (optimize_range_tests_xor): Similarly.
1008 (optimize_range_tests_diff): New function.
1009 (optimize_range_tests): Use optimize_range_tests_1.
1011 2013-10-15 Cong Hou <congh@google.com>
1013 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1014 requirement of the reduction pattern so that one operand of the
1015 reduction operation can come from outside of the loop.
1017 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1019 * config/arm/neon-schedgen.ml: Remove.
1020 * config/arm/cortex-a9-neon.md: Remove comment regarding
1023 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1025 * config/arm/types: Remove old neon types.
1027 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1029 * config/arm/cortex-a7.md
1030 (cortex_a7_neon_type): New.
1031 (cortex_a7_neon_mul): Update for new types.
1032 (cortex_a7_neon_mla): Likewise.
1033 (cortex_a7_neon): Likewise.
1035 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1037 * config/arm/cortex-a15-neon.md
1038 (cortex_a15_neon_type): New,
1040 (cortex_a15_neon_int_1): Remove.
1041 (cortex_a15_neon_int_2): Likewise.
1042 (cortex_a15_neon_int_3): Likewise.
1043 (cortex_a15_neon_int_4): Likewise.
1044 (cortex_a15_neon_int_5): Likewise.
1045 (cortex_a15_neon_vqneg_vqabs): Likewise.
1046 (cortex_a15_neon_vmov): Likewise.
1047 (cortex_a15_neon_vaba): Likewise.
1048 (cortex_a15_neon_vaba_qqq): Likewise.
1049 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1050 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1051 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1053 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1054 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1055 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1056 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1057 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1058 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1059 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1060 (cortex_a15_neon_shift_1): Likewise.
1061 (cortex_a15_neon_shift_2): Likewise.
1062 (cortex_a15_neon_shift_3): Likewise.
1063 (cortex_a15_neon_vshl_ddd): Likewise.
1064 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1065 (cortex_a15_neon_vsra_vrsra): Likewise.
1066 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1067 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1068 (cortex_a15_neon_bp_3cycle): Likewise.
1069 (cortex_a15_neon_ldm_2): Likewise.
1070 (cortex_a15_neon_stm_2): Likewise.
1071 (cortex_a15_neon_mcr): Likewise.
1072 (cortex_a15_neon_mrc): Likewise.
1073 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1074 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1075 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1076 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1077 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1078 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1079 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1080 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1081 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1082 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1083 (cortex_a15_neon_bp_simple): Likewise.
1084 (cortex_a15_neon_bp_2cycle): Likewise.
1085 (cortex_a15_neon_bp_3cycle): Likewise.
1086 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1087 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1088 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1089 (cortex_a15_neon_vld2_4_regs): Likewise.
1090 (cortex_a15_neon_vld3_vld4): Likewise.
1091 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1092 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1093 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1094 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1095 (cortex_a15_neon_vst3_vst4): Rename to...
1096 (cortex_a15_neon_vst4): This, update for new attributes.
1097 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1098 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1099 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1100 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1101 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1102 (cortex_a15_neon_ldm_2): Likewise.
1103 (cortex_a15_neon_stm_2): Likewise.
1104 (cortex_a15_neon_mcr): Likewise.
1105 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1106 (cortex_a15_neon_mrc): Likewise.
1107 (cortex_a15_neon_mrrc): Likewise.
1109 (cortex_a15_neon_abd): New.
1110 (cortex_a15_neon_abd_q): Likewise.
1111 (cortex_a15_neon_aba): Likewise.
1112 (cortex_a15_neon_aba_q): Likewise.
1113 (cortex_a15_neon_acc): Likewise.
1114 (cortex_a15_neon_acc_q): Likewise.
1115 (cortex_a15_neon_arith_basic): Likewise.
1116 (cortex_a15_neon_arith_complex): Likewise.
1117 (cortex_a15_neon_multiply): Likewise.
1118 (cortex_a15_neon_multiply_q): Likewise.
1119 (cortex_a15_neon_mla): Likewise.
1120 (cortex_a15_neon_mla_q): Likewise.
1121 (cortex_a15_neon_sat_mla_long): Likewise.
1122 (cortex_a15_neon_shift_acc): Likewise.
1123 (cortex_a15_neon_shift_imm_basic): Likewise.
1124 (cortex_a15_neon_shift_imm_complex): Likewise.
1125 (cortex_a15_neon_shift_reg_basic): Likewise.
1126 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1127 (cortex_a15_neon_shift_reg_complex): Likewise.
1128 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1129 (cortex_a15_neon_fp_negabs): Likewise
1130 (cortex_a15_neon_fp_arith): Likewise
1131 (cortex_a15_neon_fp_arith_q): Likewise
1132 (cortex_a15_neon_fp_cvt_int): Likewise
1133 (cortex_a15_neon_fp_cvt_int_q): Likewise
1134 (cortex_a15_neon_fp_cvt_16): Likewise
1135 (cortex_a15_neon_fp_mul): Likewise
1136 (cortex_a15_neon_fp_mul_q): Likewise
1137 (cortex_a15_neon_fp_mla): Likewise
1138 (cortex_a15_neon_fp_mla_q): Likewise
1139 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1140 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1141 (cortex_a15_neon_bitops): Likewise.
1142 (cortex_a15_neon_bitops_q): Likewise.
1143 (cortex_a15_neon_from_gp): Likewise.
1144 (cortex_a15_neon_from_gp_q): Likewise.
1145 (cortex_a15_neon_tbl3_tbl4): Likewise.
1146 (cortex_a15_neon_zip_q): Likewise.
1147 (cortex_a15_neon_to_gp): Likewise.
1148 (cortex_a15_neon_load_a): Likewise.
1149 (cortex_a15_neon_load_b): Likewise.
1150 (cortex_a15_neon_load_c): Likewise.
1151 (cortex_a15_neon_load_d): Likewise.
1152 (cortex_a15_neon_load_e): Likewise.
1153 (cortex_a15_neon_load_f): Likewise.
1154 (cortex_a15_neon_store_a): Likewise.
1155 (cortex_a15_neon_store_b): Likewise.
1156 (cortex_a15_neon_store_c): Likewise.
1157 (cortex_a15_neon_store_d): Likewise.
1158 (cortex_a15_neon_store_e): Likewise.
1159 (cortex_a15_neon_store_f): Likewise.
1160 (cortex_a15_neon_store_g): Likewise.
1161 (cortex_a15_neon_store_h): Likewise.
1162 (cortex_a15_vfp_to_from_gp): Likewise.
1164 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1166 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1168 (cortex_a9_neon_vshl_ddd): Remove.
1169 (cortex_a9_neon_vst3_vst4): Likewise.
1170 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1172 (cortex_a9_neon_bit_ops_q): New.
1174 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1175 (cortex_a9_neon_int_2): Likewise.
1176 (cortex_a9_neon_int_3): Likewise.
1177 (cortex_a9_neon_int_4): Likewise.
1178 (cortex_a9_neon_int_5): Likewise.
1179 (cortex_a9_neon_vqneg_vqabs): Likewise.
1180 (cortex_a9_neon_vmov): Likewise.
1181 (cortex_a9_neon_vaba): Likewise.
1182 (cortex_a9_neon_vaba_qqq): Likewise.
1183 (cortex_a9_neon_shift_1): Likewise.
1184 (cortex_a9_neon_shift_2): Likewise.
1185 (cortex_a9_neon_shift_3): Likewise.
1186 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1187 (cortex_a9_neon_vsra_vrsra): Likewise.
1188 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1189 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1190 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1192 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1193 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1194 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1196 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1197 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1198 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1199 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1200 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1201 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1202 (cortex_a9_neon_fp_vsum): Likewise.
1203 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1204 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1205 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1206 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1207 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1208 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1209 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1210 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1211 (cortex_a9_neon_bp_simple): Likewise.
1212 (cortex_a9_neon_bp_2cycle): Likewise.
1213 (cortex_a9_neon_bp_3cycle): Likewise.
1214 (cortex_a9_neon_ldr): Likewise.
1215 (cortex_a9_neon_str): Likewise.
1216 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1217 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1218 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1219 (cortex_a9_neon_vld2_4_regs): Likewise.
1220 (cortex_a9_neon_vld3_vld4): Likewise.
1221 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1222 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1223 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1224 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1225 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1226 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1227 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1228 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1229 (cortex_a9_neon_mcr): Likewise.
1230 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1231 (cortex_a9_neon_mrc): Likewise.
1232 (cortex_a9_neon_mrrc): Likewise.
1234 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1236 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1238 (cortex_a8_neon_vshl_ddd): Remove.
1239 (cortex_a8_neon_vst3_vst4): Likewise.
1240 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1242 (cortex_a8_neon_bit_ops_q): New.
1244 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1245 (cortex_a8_neon_int_2): Likewise..
1246 (cortex_a8_neon_int_3): Likewise.
1247 (cortex_a8_neon_int_5): Likewise.
1248 (cortex_a8_neon_vqneg_vqabs): Likewise.
1249 (cortex_a8_neon_int_4): Likewise.
1250 (cortex_a8_neon_vaba): Likewise.
1251 (cortex_a8_neon_vaba_qqq): Likewise.
1252 (cortex_a8_neon_shift_1): Likewise.
1253 (cortex_a8_neon_shift_2): Likewise.
1254 (cortex_a8_neon_shift_3): Likewise.
1255 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1256 (cortex_a8_neon_vsra_vrsra): Likewise.
1257 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1258 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1259 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1261 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1262 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1263 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1265 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1266 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1267 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1268 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1269 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1270 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1271 (cortex_a8_neon_fp_vsum): Likewise.
1272 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1273 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1274 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1275 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1276 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1277 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1278 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1279 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1280 (cortex_a8_neon_bp_simple): Likewise.
1281 (cortex_a8_neon_bp_2cycle): Likewise.
1282 (cortex_a8_neon_bp_3cycle): Likewise.
1283 (cortex_a8_neon_ldr): Likewise.
1284 (cortex_a8_neon_str): Likewise.
1285 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1286 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1287 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1288 (cortex_a8_neon_vld2_4_regs): Likewise.
1289 (cortex_a8_neon_vld3_vld4): Likewise.
1290 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1291 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1292 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1293 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1294 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1295 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1296 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1297 (cortex_a8_neon_mcr): Likewise.
1298 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1299 (cortex_a8_neon_mrc): Likewise.
1300 (cortex_a8_neon_mrrc): Likewise.
1302 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1304 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1306 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1307 (aarch64_simd_dup<mode>): Add "type" attribute.
1308 (aarch64_dup_lane<mode>): Likewise.
1309 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1310 (*aarch64_simd_mov<mode>): Likewise.
1311 (aarch64_simd_mov_from_<mode>low): Likewise.
1312 (aarch64_simd_mov_from_<mode>high): Likewise.
1313 (orn<mode>3): Likewise.
1314 (bic<mode>3): Likewise.
1315 (add<mode>3): Likewise.
1316 (sub<mode>3): Likewise.
1317 (mul<mode>3): Likewise.
1318 (*aarch64_mul3_elt<mode>): Likewise.
1319 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1320 (*aarch64_mul3_elt_to_128df): Likewise.
1321 (*aarch64_mul3_elt_to_64v2df): Likewise.
1322 (neg<mode>2): Likewise.
1323 (abs<mode>2): Likewise.
1324 (abd<mode>_3): Likewise.
1325 (aba<mode>_3): Likewise.
1326 (fabd<mode>_3): Likewise.
1327 (*fabd_scalar<mode>3): Likewise.
1328 (and<mode>3): Likewise.
1329 (ior<mode>3): Likewise.
1330 (xor<mode>3): Likewise.
1331 (one_cmpl<mode>2): Likewise.
1332 (aarch64_simd_vec_set<mode>): Likewise.
1333 (aarch64_simd_lshr<mode>): Likewise.
1334 (aarch64_simd_ashr<mode>): Likewise.
1335 (aarch64_simd_imm_shl<mode>): Likewise.
1336 (aarch64_simd_reg_sshl<mode): Likewise.
1337 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1338 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1339 (aarch64_simd_vec_setv2di): Likewise.
1340 (aarch64_simd_vec_set<mode>): Likewise.
1341 (aarch64_mla<mode>): Likewise.
1342 (*aarch64_mla_elt<mode>): Likewise.
1343 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1344 (aarch64_mls<mode>): Likewise.
1345 (*aarch64_mls_elt<mode>): Likewise.
1346 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1347 (<su><maxmin><mode>3): Likewise.
1348 (move_lo_quad_<mode>): Likewise.
1349 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1350 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1351 (vec_pack_trunc_<mode>): Likewise.
1352 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1353 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1354 (*aarch64_<su>mlal_lo<mode>): Likewise.
1355 (*aarch64_<su>mlal_hi<mode>): Likewise.
1356 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1357 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1358 (*aarch64_<su>mlal<mode>): Likewise.
1359 (*aarch64_<su>mlsl<mode>): Likewise.
1360 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1361 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1362 (add<mode>3): Likewise.
1363 (sub<mode>3): Likewise.
1364 (mul<mode>3): Likewise.
1365 (div<mode>3): Likewise.
1366 (neg<mode>2): Likewise.
1367 (abs<mode>2): Likewise.
1368 (fma<mode>4): Likewise.
1369 (*aarch64_fma4_elt<mode>): Likewise.
1370 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1371 (*aarch64_fma4_elt_to_128df): Likewise.
1372 (*aarch64_fma4_elt_to_64v2df): Likewise.
1373 (fnma<mode>4): Likewise.
1374 (*aarch64_fnma4_elt<mode>): Likewise.
1375 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1376 (*aarch64_fnma4_elt_to_128df): Likewise.
1377 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1378 (<frint_pattern><mode>2): Likewise.
1379 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1380 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1381 (vec_unpacks_lo_v4sf): Likewise.
1382 (aarch64_float_extend_lo_v2df): Likewise.
1383 (vec_unpacks_hi_v4sf): Likewise.
1384 (aarch64_float_truncate_lo_v2sf): Likewise.
1385 (aarch64_float_truncate_hi_v4sf): Likewise.
1386 (aarch64_vmls<mode>): Likewise.
1387 (<su><maxmin><mode>3): Likewise.
1388 (<maxmin_uns><mode>3): Likewise.
1389 (reduc_<sur>plus_<mode>): Likewise.
1390 (reduc_<sur>plus_v2di): Likewise.
1391 (reduc_<sur>plus_v2si): Likewise.
1392 (reduc_<sur>plus_<mode>): Likewise.
1393 (aarch64_addpv4sf): Likewise.
1394 (clz<mode>2): Likewise.
1395 (reduc_<maxmin_uns>_<mode>): Likewise.
1396 (reduc_<maxmin_uns>_v2di): Likewise.
1397 (reduc_<maxmin_uns>_v2si): Likewise.
1398 (reduc_<maxmin_uns>_<mode>): Likewise.
1399 (reduc_<maxmin_uns>_v4sf): Likewise.
1400 (aarch64_simd_bsl<mode>_internal): Likewise.
1401 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1402 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1403 (aarch64_get_lane<mode>): Likewise.
1404 (*aarch64_combinez<mode>): Likewise.
1405 (aarch64_combine<mode>): Likewise.
1406 (aarch64_simd_combine<mode>): Likewise.
1407 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1408 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1409 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1410 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1411 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1412 (aarch64_<sur>h<addsub><mode>): Likewise.
1413 (aarch64_<sur><addsub>hn<mode>): Likewise.
1414 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1415 (aarch64_pmul<mode>): Likewise.
1416 (aarch64_<su_optab><optab><mode>): Likewise.
1417 (aarch64_<sur>qadd<mode>): Likewise.
1418 (aarch64_sqmovun<mode>): Likewise.
1419 (aarch64_<sur>qmovn<mode>): Likewise.
1420 (aarch64_s<optab><mode>): Likewise.
1421 (aarch64_sq<r>dmulh<mode>): Likewise.
1422 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1423 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1424 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1425 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1426 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1427 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1428 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1429 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1430 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1431 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1432 (aarch64_sqdmull<mode>): Likewise.
1433 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1434 (aarch64_sqdmull_n<mode>): Likewise.
1435 (aarch64_sqdmull2<mode>_internal): Likewise.
1436 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1437 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1438 (aarch64_<sur>shl<mode>): Likewise.
1439 (aarch64_<sur>q<r>shl<mode>
1440 (aarch64_<sur>shll_n<mode>): Likewise.
1441 (aarch64_<sur>shll2_n<mode>): Likewise.
1442 (aarch64_<sur>shr_n<mode>): Likewise.
1443 (aarch64_<sur>sra_n<mode>): Likewise.
1444 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1445 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1446 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1447 (aarch64_cm<optab><mode>): Likewise.
1448 (aarch64_cm<optab>di): Likewise.
1449 (aarch64_cm<optab><mode>): Likewise.
1450 (aarch64_cm<optab>di): Likewise.
1451 (aarch64_cmtst<mode>): Likewise.
1452 (aarch64_cmtstdi): Likewise.
1453 (aarch64_cm<optab><mode>): Likewise.
1454 (*aarch64_fac<optab><mode>): Likewise.
1455 (aarch64_addp<mode>): Likewise.
1456 (aarch64_addpdi): Likewise.
1457 (sqrt<mode>2): Likewise.
1458 (vec_load_lanesoi<mode>): Likewise.
1459 (vec_store_lanesoi<mode>): Likewise.
1460 (vec_load_lanesci<mode>): Likewise.
1461 (vec_store_lanesci<mode>): Likewise.
1462 (vec_load_lanesxi<mode>): Likewise.
1463 (vec_store_lanesxi<mode>): Likewise.
1464 (*aarch64_mov<mode>): Likewise.
1465 (aarch64_ld2<mode>_dreg): Likewise.
1466 (aarch64_ld2<mode>_dreg): Likewise.
1467 (aarch64_ld3<mode>_dreg): Likewise.
1468 (aarch64_ld3<mode>_dreg): Likewise.
1469 (aarch64_ld4<mode>_dreg): Likewise.
1470 (aarch64_ld4<mode>_dreg): Likewise.
1471 (aarch64_tbl1<mode>): Likewise.
1472 (aarch64_tbl2v16qi): Likewise.
1473 (aarch64_combinev16qi): Likewise.
1474 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1475 (aarch64_st2<mode>_dreg): Likewise.
1476 (aarch64_st2<mode>_dreg): Likewise.
1477 (aarch64_st3<mode>_dreg): Likewise.
1478 (aarch64_st3<mode>_dreg): Likewise.
1479 (aarch64_st4<mode>_dreg): Likewise.
1480 (aarch64_st4<mode>_dreg): Likewise.
1481 (*aarch64_simd_ld1r<mode>): Likewise.
1482 (aarch64_frecpe<mode>): Likewise.
1483 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1484 (aarch64_frecps<mode>): Likewise.
1486 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1488 * config/arm/iterators.md (V_elem_ch): New.
1490 (VQH_type): Likewise.
1491 * config/arm/arm.md (is_neon_type): New.
1492 (conds): Use is_neon_type.
1493 (anddi3_insn): Update type attribute.
1494 (xordi3_insn): Likewise.
1495 (one_cmpldi2): Likewise.
1496 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1497 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1498 (*movmisalign<mode>_neon_store): Likewise.
1499 (*movmisalign<mode>_neon_load): Likewise.
1500 (vec_set<mode>_internal): Likewise.
1501 (vec_set<mode>_internal): Likewise.
1502 (vec_setv2di_internal): Likewise.
1503 (vec_extract<mode>): Likewise.
1504 (vec_extract<mode>): Likewise.
1505 (vec_extractv2di): Likewise.
1506 (*add<mode>3_neon): Likewise.
1507 (adddi3_neon): Likewise.
1508 (*sub<mode>3_neon): Likewise.
1509 (subdi3_neon): Likewise.
1510 (fma<VCVTF:mode>4): Likewise.
1511 (fma<VCVTF:mode>4_intrinsic): Likewise.
1512 (*fmsub<VCVTF:mode>4): Likewise.
1513 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1514 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1515 (ior<mode>3): Likewise.
1516 (and<mode>3): Likewise.
1517 (orn<mode>3_neon): Likewise.
1518 (orndi3_neon): Likewise.
1519 (bic<mode>3_neon): Likewise.
1520 (bicdi3_neon): Likewise.
1521 (xor<mode>3): Likewise.
1522 (one_cmpl<mode>2): Likewise.
1523 (abs<mode>2): Likewise.
1524 (neg<mode>2): Likewise.
1525 (negdi2_neon): Likewise.
1526 (*umin<mode>3_neon): Likewise.
1527 (*umax<mode>3_neon): Likewise.
1528 (*smin<mode>3_neon): Likewise.
1529 (*smax<mode>3_neon): Likewise.
1530 (vashl<mode>3): Likewise.
1531 (vashr<mode>3_imm): Likewise.
1532 (vlshr<mode>3_imm): Likewise.
1533 (ashl<mode>3_signed): Likewise.
1534 (ashl<mode>3_unsigned): Likewise.
1535 (neon_load_count): Likewise.
1536 (ashldi3_neon_noclobber): Likewise.
1537 (ashldi3_neon): Likewise.
1538 (signed_shift_di3_neon): Likewise.
1539 (unsigned_shift_di3_neon): Likewise.
1540 (ashrdi3_neon_imm_noclobber): Likewise.
1541 (lshrdi3_neon_imm_noclobber): Likewise.
1542 (<shift>di3_neon): Likewise.
1543 (widen_ssum<mode>3): Likewise.
1544 (widen_usum<mode>3): Likewise.
1545 (quad_halves_<code>v4si): Likewise.
1546 (quad_halves_<code>v4sf): Likewise.
1547 (quad_halves_<code>v8hi): Likewise.
1548 (quad_halves_<code>v16qi): Likewise.
1549 (reduc_splus_v2di): Likewise.
1550 (neon_vpadd_internal<mode>): Likewise.
1551 (neon_vpsmin<mode>): Likewise.
1552 (neon_vpsmax<mode>): Likewise.
1553 (neon_vpumin<mode>): Likewise.
1554 (neon_vpumax<mode>): Likewise.
1555 (*ss_add<mode>_neon): Likewise.
1556 (*us_add<mode>_neon): Likewise.
1557 (*ss_sub<mode>_neon): Likewise.
1558 (*us_sub<mode>_neon): Likewise.
1559 (neon_vadd<mode>_unspec): Likewise.
1560 (neon_vaddl<mode>): Likewise.
1561 (neon_vaddw<mode>): Likewise.
1562 (neon_vhadd<mode>): Likewise.
1563 (neon_vqadd<mode>): Likewise.
1564 (neon_vaddhn<mode>): Likewise.
1565 (neon_vmul<mode>): Likewise.
1566 (neon_vfms<VCVTF:mode>): Likewise.
1567 (neon_vmlal<mode>): Likewise.
1568 (neon_vmls<mode>): Likewise.
1569 (neon_vmlsl<mode>): Likewise.
1570 (neon_vqdmulh<mode>): Likewise.
1571 (neon_vqdmlal<mode>): Likewise.
1572 (neon_vqdmlsl<mode>): Likewise.
1573 (neon_vmull<mode>): Likewise.
1574 (neon_vqdmull<mode>): Likewise.
1575 (neon_vsub<mode>_unspec): Likewise.
1576 (neon_vsubl<mode>): Likewise.
1577 (neon_vsubw<mode>): Likewise.
1578 (neon_vqsub<mode>): Likewise.
1579 (neon_vhsub<mode>): Likewise.
1580 (neon_vsubhn<mode>): Likewise.
1581 (neon_vceq<mode>): Likewise.
1582 (neon_vcge<mode>): Likewise.
1583 (neon_vcgeu<mode>): Likewise.
1584 (neon_vcgt<mode>): Likewise.
1585 (neon_vcgtu<mode>): Likewise.
1586 (neon_vcle<mode>): Likewise.
1587 (neon_vclt<mode>): Likewise.
1588 (neon_vcage<mode>): Likewise.
1589 (neon_vcagt<mode>): Likewise.
1590 (neon_vtst<mode>): Likewise.
1591 (neon_vabd<mode>): Likewise.
1592 (neon_vabdl<mode>): Likewise.
1593 (neon_vaba<mode>): Likewise.
1594 (neon_vabal<mode>): Likewise.
1595 (neon_vmax<mode>): Likewise.
1596 (neon_vmin<mode>): Likewise.
1597 (neon_vpaddl<mode>): Likewise.
1598 (neon_vpadal<mode>): Likewise.
1599 (neon_vpmax<mode>): Likewise.
1600 (neon_vpmin<mode>): Likewise.
1601 (neon_vrecps<mode>): Likewise.
1602 (neon_vrsqrts<mode>): Likewise.
1603 (neon_vqabs<mode>): Likewise.
1604 (neon_vqneg<mode>): Likewise.
1605 (neon_vcls<mode>): Likewise.
1606 (clz<mode>2): Likewise.
1607 (popcount<mode>2): Likewise.
1608 (neon_vrecpe<mode>): Likewise.
1609 (neon_vrsqrte<mode>): Likewise.
1610 (neon_vget_lane<mode>_sext_internal): Likewise.
1611 (neon_vget_lane<mode>_zext_internal): Likewise.
1612 (neon_vdup_n<mode>): Likewise.
1613 (neon_vdup_n<mode>): Likewise.
1614 (neon_vdup_nv2di): Likewise.
1615 (neon_vdup_lane<mode>_interal): Likewise.
1616 (*neon_vswp<mode>): Likewise.
1617 (neon_vcombine<mode>): Likewise.
1618 (float<mode><V_cvtto>2): Likewise.
1619 (floatuns<mode><V_cvtto>2): Likewise.
1620 (fix_trunc<mode><V_cvtto>2): Likewise.
1621 (fixuns_trunc<mode><V_cvtto>2
1622 (neon_vcvt<mode>): Likewise.
1623 (neon_vcvt<mode>): Likewise.
1624 (neon_vcvtv4sfv4hf): Likewise.
1625 (neon_vcvtv4hfv4sf): Likewise.
1626 (neon_vcvt_n<mode>): Likewise.
1627 (neon_vcvt_n<mode>): Likewise.
1628 (neon_vmovn<mode>): Likewise.
1629 (neon_vqmovn<mode>): Likewise.
1630 (neon_vqmovun<mode>): Likewise.
1631 (neon_vmovl<mode>): Likewise.
1632 (neon_vmul_lane<mode>): Likewise.
1633 (neon_vmul_lane<mode>): Likewise.
1634 (neon_vmull_lane<mode>): Likewise.
1635 (neon_vqdmull_lane<mode>): Likewise.
1636 (neon_vqdmulh_lane<mode>): Likewise.
1637 (neon_vqdmulh_lane<mode>): Likewise.
1638 (neon_vmla_lane<mode>): Likewise.
1639 (neon_vmla_lane<mode>): Likewise.
1640 (neon_vmlal_lane<mode>): Likewise.
1641 (neon_vqdmlal_lane<mode>): Likewise.
1642 (neon_vmls_lane<mode>): Likewise.
1643 (neon_vmls_lane<mode>): Likewise.
1644 (neon_vmlsl_lane<mode>): Likewise.
1645 (neon_vqdmlsl_lane<mode>): Likewise.
1646 (neon_vext<mode>): Likewise.
1647 (neon_vrev64<mode>): Likewise.
1648 (neon_vrev32<mode>): Likewise.
1649 (neon_vrev16<mode>): Likewise.
1650 (neon_vbsl<mode>_internal): Likewise.
1651 (neon_vshl<mode>): Likewise.
1652 (neon_vqshl<mode>): Likewise.
1653 (neon_vshr_n<mode>): Likewise.
1654 (neon_vshrn_n<mode>): Likewise.
1655 (neon_vqshrn_n<mode>): Likewise.
1656 (neon_vqshrun_n<mode>): Likewise.
1657 (neon_vshl_n<mode>): Likewise.
1658 (neon_vqshl_n<mode>): Likewise.
1659 (neon_vqshlu_n<mode>): Likewise.
1660 (neon_vshll_n<mode>): Likewise.
1661 (neon_vsra_n<mode>): Likewise.
1662 (neon_vsri_n<mode>): Likewise.
1663 (neon_vsli_n<mode>): Likewise.
1664 (neon_vtbl1v8qi): Likewise.
1665 (neon_vtbl2v8qi): Likewise.
1666 (neon_vtbl3v8qi): Likewise.
1667 (neon_vtbl4v8qi): Likewise.
1668 (neon_vtbl1v16qi): Likewise.
1669 (neon_vtbl2v16qi): Likewise.
1670 (neon_vcombinev16qi): Likewise.
1671 (neon_vtbx1v8qi): Likewise.
1672 (neon_vtbx2v8qi): Likewise.
1673 (neon_vtbx3v8qi): Likewise.
1674 (neon_vtbx4v8qi): Likewise.
1675 (*neon_vtrn<mode>_insn): Likewise.
1676 (*neon_vzip<mode>_insn): Likewise.
1677 (*neon_vuzp<mode>_insn): Likewise.
1678 (neon_vld1<mode>): Likewise.
1679 (neon_vld1_lane<mode>): Likewise.
1680 (neon_vld1_lane<mode>): Likewise.
1681 (neon_vld1_dup<mode>): Likewise.
1682 (neon_vld1_dup<mode>): Likewise.
1683 (neon_vld1_dupv2di): Likewise.
1684 (neon_vst1<mode>): Likewise.
1685 (neon_vst1_lane<mode>): Likewise.
1686 (neon_vst1_lane<mode>): Likewise.
1687 (neon_vld2<mode>): Likewise.
1688 (neon_vld2<mode>): Likewise.
1689 (neon_vld2_lane<mode>): Likewise.
1690 (neon_vld2_lane<mode>): Likewise.
1691 (neon_vld2_dup<mode>): Likewise.
1692 (neon_vst2<mode>): Likewise.
1693 (neon_vst2<mode>): Likewise.
1694 (neon_vst2_lane<mode>): Likewise.
1695 (neon_vst2_lane<mode>): Likewise.
1696 (neon_vld3<mode>): Likewise.
1697 (neon_vld3qa<mode>): Likewise.
1698 (neon_vld3qb<mode>): Likewise.
1699 (neon_vld3_lane<mode>): Likewise.
1700 (neon_vld3_lane<mode>): Likewise.
1701 (neon_vld3_dup<mode>): Likewise.
1702 (neon_vst3<mode>): Likewise.
1703 (neon_vst3qa<mode>): Likewise.
1704 (neon_vst3qb<mode>): Likewise.
1705 (neon_vst3_lane<mode>): Likewise.
1706 (neon_vst3_lane<mode>): Likewise.
1707 (neon_vld4<mode>): Likewise.
1708 (neon_vld4qa<mode>): Likewise.
1709 (neon_vld4qb<mode>): Likewise.
1710 (neon_vld4_lane<mode>): Likewise.
1711 (neon_vld4_lane<mode>): Likewise.
1712 (neon_vld4_dup<mode>): Likewise.
1713 (neon_vst4<mode>): Likewise.
1714 (neon_vst4qa<mode>): Likewise.
1715 (neon_vst4qb<mode>): Likewise.
1716 (neon_vst4_lane<mode>): Likewise.
1717 (neon_vst4_lane<mode>): Likewise.
1718 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1719 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1720 (neon_vec_<US>mult_lo_<mode>): Likewise.
1721 (neon_vec_<US>mult_hi_<mode>): Likewise.
1722 (neon_vec_<US>shiftl_<mode>): Likewise.
1723 (neon_unpack<US>_<mode>): Likewise.
1724 (neon_vec_<US>mult_<mode>): Likewise.
1725 (vec_pack_trunc_<mode>): Likewise.
1726 (neon_vec_pack_trunc_<mode>): Likewise.
1727 (neon_vabd<mode>_2): Likewise.
1728 (neon_vabd<mode>_3): Likewise.
1730 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1732 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1733 (load_pair): Update type attribute.
1734 (store_pair): Update type attribute.
1735 * config/aarch64/iterators.md (q): New.
1737 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1739 * config/arm/types.md: Add new types for Neon insns.
1741 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1742 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1743 Sergey Lega <sergey.s.lega@intel.com>
1744 Anna Tikhonova <anna.tikhonova@intel.com>
1745 Ilya Tocar <ilya.tocar@intel.com>
1746 Andrey Turetskiy <andrey.turetskiy@intel.com>
1747 Ilya Verbin <ilya.verbin@intel.com>
1748 Kirill Yukhin <kirill.yukhin@intel.com>
1749 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1751 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1752 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1753 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1755 (srcp14<mode>): Ditto.
1756 (rsqrt14<mode>): Ditto.
1757 (rsqrt14<mode>): Ditto.
1758 (avx512f_vmscalef<mode>): Ditto.
1759 (avx512f_scalef<mode>): Ditto.
1760 (avx512f_getexp<mode>): Ditto.
1761 (avx512f_sgetexp<mode>): Ditto.
1762 (avx512f_fixupimm<mode>): Ditto.
1763 (avx512f_sfixupimm<mode>): Ditto.
1764 (avx512f_rndscale<mode>): Ditto.
1765 (*avx512er_exp2<mode>): Ditto.
1766 (*avx512er_rcp28<mode>): Ditto.
1767 (avx512er_rsqrt28<mode>): Ditto.
1768 (avx512f_getmant<mode>): Ditto.
1769 (avx512f_getmant<mode>): Ditto.
1770 (avx512f_rndscale<mode>): Fix formatting.
1772 2013-10-15 Martin Jambor <mjambor@suse.cz>
1774 * ipa-utils.h (ipa_edge_within_scc): Declare.
1775 * ipa-cp.c (edge_within_scc): Moved...
1776 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1778 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1779 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1780 Sergey Lega <sergey.s.lega@intel.com>
1781 Anna Tikhonova <anna.tikhonova@intel.com>
1782 Ilya Tocar <ilya.tocar@intel.com>
1783 Andrey Turetskiy <andrey.turetskiy@intel.com>
1784 Ilya Verbin <ilya.verbin@intel.com>
1785 Kirill Yukhin <kirill.yukhin@intel.com>
1786 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1788 * config/i386/predicates.md (const_8_to_15_operand): New.
1789 (const_16_to_31_operand): Ditto.
1790 * config/i386/sse.md (V8FI): New.
1792 (reduc_splus_v8df): Ditto.
1793 (reduc_splus_v16sf): Ditto.
1794 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1795 (vec_extract_hi_<mode>): Ditto.
1796 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1797 (vec_set_lo_<mode>): Ditto.
1798 (vec_set_hi_<mode>): Ditto.
1799 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1800 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1801 (avx512f_pshufd_1): Ditto.
1802 (avx512f_broadcast<mode>): Ditto.
1803 (avx512f_broadcast<mode>): Ditto.
1804 (define_split): Split vec_extract_lo into move.
1805 (ssequartermode): Ditto.
1806 (ssedoublemode): Extened with wider modes.
1807 (vec_extract_lo_<mode>): Ditto.
1809 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1810 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1811 Sergey Lega <sergey.s.lega@intel.com>
1812 Anna Tikhonova <anna.tikhonova@intel.com>
1813 Ilya Tocar <ilya.tocar@intel.com>
1814 Andrey Turetskiy <andrey.turetskiy@intel.com>
1815 Ilya Verbin <ilya.verbin@intel.com>
1816 Kirill Yukhin <kirill.yukhin@intel.com>
1817 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1819 * config/i386/predicates.md (register_or_constm1_operand): New.
1820 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1821 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1822 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1823 UNSPEC_SCATTER_PREFETCH
1825 (avx512f_ucmp<mode>3): Ditto.
1826 (avx512f_vternlog<mode>): Ditto.
1827 (avx512f_align<mode>): Ditto.
1828 (<shift_insn><mode>3): Ditto.
1829 (avx512f_<rotate>v<mode>): Ditto.
1830 (avx512f_<rotate><mode>): Ditto.
1831 (avx512f_eq<mode>3): Ditto.
1832 (avx512f_eq<mode>3_1): Ditto.
1833 (avx512f_gt<mode>3): Ditto.
1834 (avx512f_testm<mode>3): Ditto.
1835 (avx512f_testnm<mode>3): Ditto.
1836 (avx512pf_gatherpf<mode>): Ditto.
1837 (*avx512pf_gatherpf<mode>_mask): Ditto.
1838 (*avx512pf_gatherpf<mode>): Ditto.
1839 (avx512pf_scatterpf<mode>): Ditto.
1840 (*avx512pf_scatterpf<mode>_mask): Ditto.
1841 (*avx512pf_scatterpf<mode>): Ditto.
1842 (avx512f_vec_dup_gpr<mode>): Ditto.
1843 (clz<mode>2): Ditto.
1844 (conflict<mode>): Ditto.
1845 (REDUC_SMINMAX_MODE): Extened with wider modes.
1846 (reduc_<code>_<mode>): Ditto.
1847 (vlshr<mode>3): Ditto.
1848 (vashl<mode>3): Ditto.
1850 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1851 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1852 Sergey Lega <sergey.s.lega@intel.com>
1853 Anna Tikhonova <anna.tikhonova@intel.com>
1854 Ilya Tocar <ilya.tocar@intel.com>
1855 Andrey Turetskiy <andrey.turetskiy@intel.com>
1856 Ilya Verbin <ilya.verbin@intel.com>
1857 Kirill Yukhin <kirill.yukhin@intel.com>
1858 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1860 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1863 (avx512fmaskmode): Ditto.
1864 (bcstscalarsuff): Ditto.
1865 (avx512f_blendm<mode>): Ditto.
1866 (cmp_imm_predicate): Ditto.
1867 (avx512f_cmp<mode>3): Ditto.
1868 (avx512f_vec_dup<mode>): Ditto.
1869 (avx512f_vec_dup_mem<mode>): Ditto.
1870 (avx512f_vpermi2var<mode>3): Ditto.
1871 (avx512f_vpermt2var<mode>3): Ditto.
1872 (vec_init<mode>): Ditto.
1873 (avx512f_gathersi<mode>): Ditto.
1874 (*avx512f_gathersi<mode>): Ditto.
1875 (*avx512f_gathersi<mode>_2): Ditto.
1876 (avx512f_gatherdi<mode>): Ditto.
1877 (*avx512f_gatherdi<mode>): Ditto.
1878 (*avx512f_gatherdi<mode>_2): Ditto.
1879 (avx512f_scattersi<mode>): Ditto.
1880 (*avx512f_scattersi<mode>): Ditto.
1881 (avx512f_scatterdi<mode>): Ditto.
1882 (*avx512f_scatterdi<mode>): Ditto.
1883 (sseintprefix): Extened with wider modes.
1884 (VEC_GATHER_IDXSI): Ditto.
1885 (VEC_GATHER_IDXDI): Ditto.
1886 (VEC_GATHER_SRCDI): Ditto.
1888 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1889 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1891 * config/arm/t-aprofile: New file.
1892 * config.gcc: Handle --with-multilib-list option.
1894 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
1896 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
1899 2013-10-15 Richard Biener <rguenther@suse.de>
1901 * tree-loop-distribution.c (build_empty_rdg): Inline into
1903 (rdg_flag_vertex): Inline into single user.
1904 (rdg_flag_vertex_and_dependent): Likewise.
1905 (build_rdg_partition_for_vertex): Remove processed bitmap.
1906 (rdg_build_partitions): Simplify.
1908 2013-10-15 Richard Biener <rguenther@suse.de>
1910 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
1911 Restructure forwarding through conversions and copies to
1912 avoid performing copy-propagation the wrong way. Adjust
1913 recursion invocations.
1914 (forward_propagate_addr_expr): Add argument stating if we
1915 are recursing from a single-use.
1916 (ssa_forward_propagate_and_combine): Adjust.
1918 2013-10-14 David Malcolm <dmalcolm@redhat.com>
1920 * dumpfile.h (gcc::dump_manager): New class, to hold state
1921 relating to dumpfile management.
1922 (get_dump_file_name): Remove in favor of method of dump_manager.
1923 (dump_initialized_p): Likewise.
1924 (dump_start): Likewise.
1925 (dump_finish): Likewise.
1926 (dump_switch_p): Likewise.
1927 (dump_register): Likewise.
1928 (get_dump_file_info): Likewise.
1929 * context.c (gcc::context::context): Construct the dump_manager
1931 * context.h (gcc::context::get_dumps): New.
1932 (gcc::context::m_dumps): New.
1933 * coverage.c (coverage_init): Port to dump_manager API.
1934 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
1935 (extra_dump_files_in_use): Likewise.
1936 (extra_dump_files_alloced): Likewise.
1937 (gcc::dump_manager::dump_manager): New.
1938 (dump_register): Convert to...
1939 (gcc::dump_manager::dump_register): ...method, replacing
1940 function-static next_dump with m_next_dump field.
1941 (get_dump_file_info): Convert to...
1942 (gcc::dump_manager::get_dump_file_info): ...method.
1943 (get_dump_file_name): Convert to...
1944 (gcc::dump_manager::get_dump_file_name): ...method.
1945 (dump_start): Convert to...
1946 (gcc::dump_manager::dump_start): ...method.
1947 (dump_finish): Convert to...
1948 (gcc::dump_manager::dump_finish): ...method.
1949 (dump_begin): Replace body with...
1950 (gcc::dump_manager::dump_begin): ...new method.
1951 (dump_phase_enabled_p): Convert to...
1952 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1953 (dump_phase_enabled_p): Convert to...
1954 (gcc::dump_manager::dump_phase_enabled_p): ...method.
1955 (dump_initialized_p): Convert to...
1956 (gcc::dump_manager::dump_initialized_p): ...method.
1957 (dump_flag_name): Replace body with...
1958 (gcc::dump_manager::dump_flag_name): ...new method.
1959 (dump_enable_all): Convert to...
1960 (gcc::dump_manager::dump_enable_all): ...new method.
1961 (opt_info_enable_passes): Convert to...
1962 (gcc::dump_manager::opt_info_enable_passes): ...new method.
1963 (dump_switch_p_1): Convert to...
1964 (gcc::dump_manager::dump_switch_p_1): ...new method.
1965 (dump_switch_p): Convert to...
1966 (gcc::dump_manager::dump_switch_p): ...new method.
1967 (opt_info_switch_p): Port to dump_manager API.
1968 (enable_rtl_dump_file): Likewise.
1969 * opts-global.c (handle_common_deferred_options): Port to new
1971 * passes.c (pass_manager::finish_optimization_passes): Likewise.
1972 (pass_manager::register_one_dump_file): Likewise.
1973 (pass_manager::register_pass): Likewise.
1974 (pass_init_dump_file): Likewise.
1975 (pass_fini_dump_file): Likewise.
1976 * statistics.c (statistics_early_init): Likewise.
1978 2013-10-14 Richard Biener <rguenther@suse.de>
1980 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1981 iterative_hash_canonical_type, gimple_canonical_type_hash,
1982 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1983 gimple_register_canonical_type, print_gimple_types_stats,
1984 free_gimple_type_tables): Move to lto/lto.c
1985 (gt-gimple.h): Do not include.
1986 * gimple.h (gimple_register_canonical_type,
1987 print_gimple_types_stats, free_gimple_type_tables): Remove.
1988 * Makefile.in (GTFILES): Remove gimple.c.
1990 2013-10-14 Travis Snoozy <quandary@remstate.com>
1993 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1994 scanning for msp430x targets.
1996 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1999 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2000 (registers_ok_for_ldd_peep): Move around.
2001 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2002 * config/sparc/sparc.md (widening peepholes): Use it.
2004 2013-10-14 Richard Biener <rguenther@suse.de>
2008 * gimple.c (iterative_hash_canonical_type): Make sure to
2009 record the hash into the correct hashtable slot.
2011 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2013 PR rtl-optimization/58662
2014 * combine.c (try_combine): Take into account death nodes on I2 when
2015 splitting a PARALLEL of two independent SETs. Fix dump message.
2017 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2020 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2021 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2022 * config/sh/t-sh (sh_treg_combine.o): New entry.
2023 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2024 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2025 (register_sh_passes): New function. Register sh_treg_combine pass.
2026 (sh_option_override): Invoke it.
2027 (sh_canonicalize_comparison): Handle op0_preserve_value.
2028 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2029 opportunities. Canonicalize branch condition.
2030 (nott): Allow only if pseudos can be created for non-SH2A.
2032 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2035 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2036 (ix86_expand_movmem): Replace copy_addr_to_reg with
2037 ix86_copy_addr_to_reg.
2038 (ix86_expand_setmem): Likewise.
2040 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2042 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2043 provided by ix86_fp_compare_mode instead of CCFPUmode.
2045 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2047 * config/aarch64/arm_neon.h
2048 (vtbx<1,3>_<psu>8): Fix register constriants.
2050 2013-10-11 Jeff Law <law@redhat.com>
2052 PR tree-optimization/58640
2053 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2054 threading paths that cross over two loop entry points.
2056 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2058 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2059 handle vector float as well.
2060 (*vsx_le_perm_load_v4si): Likewise.
2061 (*vsx_le_perm_store_v2di): Likewise.
2062 (*vsx_le_perm_store_v4si): Likewise.
2064 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2066 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2067 directly to circumvent subtract from splat{31} workaround.
2068 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2070 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2071 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2072 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2073 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2074 separate big and little endian logic.
2075 (*altivec_vperm_<mode>_internal): New define_insn.
2076 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2077 separate big and little endian logic.
2078 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2079 (vec_permv16qi): Add little endian logic.
2081 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2083 * doc/extend.texi (returns_nonnull): Remove arguments.
2085 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2086 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2087 Sergey Lega <sergey.s.lega@intel.com>
2088 Anna Tikhonova <anna.tikhonova@intel.com>
2089 Ilya Tocar <ilya.tocar@intel.com>
2090 Andrey Turetskiy <andrey.turetskiy@intel.com>
2091 Ilya Verbin <ilya.verbin@intel.com>
2092 Kirill Yukhin <kirill.yukhin@intel.com>
2093 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2095 * config/i386/sse.md (VI48F_256_512): New.
2096 (avx2_permvar<mode>): Change to ...
2097 (<avx2_avx512f>_permvar<mode>): This.
2099 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2100 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2101 Sergey Lega <sergey.s.lega@intel.com>
2102 Anna Tikhonova <anna.tikhonova@intel.com>
2103 Ilya Tocar <ilya.tocar@intel.com>
2104 Andrey Turetskiy <andrey.turetskiy@intel.com>
2105 Ilya Verbin <ilya.verbin@intel.com>
2106 Kirill Yukhin <kirill.yukhin@intel.com>
2107 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2109 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2110 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2111 * config/i386/sse.md (VI4_AVX): New.
2112 (sf2simodelower): Ditto.
2113 (sse2_cvtps2dq): Change to ...
2114 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2116 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2117 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2118 Sergey Lega <sergey.s.lega@intel.com>
2119 Anna Tikhonova <anna.tikhonova@intel.com>
2120 Ilya Tocar <ilya.tocar@intel.com>
2121 Andrey Turetskiy <andrey.turetskiy@intel.com>
2122 Ilya Verbin <ilya.verbin@intel.com>
2123 Kirill Yukhin <kirill.yukhin@intel.com>
2124 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2126 * config/i386/sse.md (V_512): New.
2128 (vcond<V_512:mode><VF_512:mode>): Ditto.
2129 (vcond<V_512:mode><VI_512:mode>): Ditto.
2130 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2132 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2133 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2134 Sergey Lega <sergey.s.lega@intel.com>
2135 Anna Tikhonova <anna.tikhonova@intel.com>
2136 Ilya Tocar <ilya.tocar@intel.com>
2137 Andrey Turetskiy <andrey.turetskiy@intel.com>
2138 Ilya Verbin <ilya.verbin@intel.com>
2139 Kirill Yukhin <kirill.yukhin@intel.com>
2140 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2142 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2143 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2145 (fma<mode>4): Removed condition.
2146 (fms<mode>4): Ditto.
2147 (fnma<mode>4): Ditto.
2148 (fnms<mode>4): Ditto.
2149 (fma4i_fmadd_<mode>): Ditto.
2150 (*fma_fmadd_<mode>): Ditto.
2151 (*fma_fmsub_<mode>): Ditto.
2152 (*fma_fnmadd_<mode>): Ditto.
2153 (*fma_fnmsub_<mode>): Ditto.
2154 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2155 (*fma_fmaddsub_<mode>): Ditto.
2156 (*fma_fmsubadd_<mode>): Ditto.
2157 (*fmai_fmadd_<mode>): Ditto.
2158 (*fmai_fmsub_<mode>): Ditto.
2159 (*fmai_fnmadd_<mode>): Ditto.
2160 (*fmai_fnmsub_<mode>): Ditto.
2162 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2163 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2164 Sergey Lega <sergey.s.lega@intel.com>
2165 Anna Tikhonova <anna.tikhonova@intel.com>
2166 Ilya Tocar <ilya.tocar@intel.com>
2167 Andrey Turetskiy <andrey.turetskiy@intel.com>
2168 Ilya Verbin <ilya.verbin@intel.com>
2169 Kirill Yukhin <kirill.yukhin@intel.com>
2170 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2172 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2173 (VI124_256): Changed to ...
2174 (VI124_256_48_512): This.
2175 (ssepackmode): Extended with wider modes.
2176 (<code><mode>3): Changed iterator.
2177 (*avx2_<code><mode>3): Ditto.
2178 (vec_pack_trunc_<mode>): Ditto.
2180 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2181 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2182 Sergey Lega <sergey.s.lega@intel.com>
2183 Anna Tikhonova <anna.tikhonova@intel.com>
2184 Ilya Tocar <ilya.tocar@intel.com>
2185 Andrey Turetskiy <andrey.turetskiy@intel.com>
2186 Ilya Verbin <ilya.verbin@intel.com>
2187 Kirill Yukhin <kirill.yukhin@intel.com>
2188 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2190 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2191 (VI8F_256_512): Ditto.
2192 (abs<mode>2): Changed iterator.
2193 (avx2_perm<mode>): Changed to ...
2194 (<avx2_avx512f>_perm<mode>): This.
2195 (avx2_perm<mode>_1): Changed to ...
2196 (<avx2_avx512f>_perm<mode>_1): This.
2198 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2199 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2200 Sergey Lega <sergey.s.lega@intel.com>
2201 Anna Tikhonova <anna.tikhonova@intel.com>
2202 Ilya Tocar <ilya.tocar@intel.com>
2203 Andrey Turetskiy <andrey.turetskiy@intel.com>
2204 Ilya Verbin <ilya.verbin@intel.com>
2205 Kirill Yukhin <kirill.yukhin@intel.com>
2206 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2208 * config/i386/sse.md (VI48_AVX512F): New.
2209 (VI48_AVX2): Changed to ...
2210 (VI48_AVX2_48_AVX512F): This.
2211 (avx2_ashrv<mode>): Changed to ...
2212 (<avx2_avx512f>_ashrv<mode>): This.
2213 (avx2_<shift_insn>v<mode>): Changed to ...
2214 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2216 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2217 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2218 Sergey Lega <sergey.s.lega@intel.com>
2219 Anna Tikhonova <anna.tikhonova@intel.com>
2220 Ilya Tocar <ilya.tocar@intel.com>
2221 Andrey Turetskiy <andrey.turetskiy@intel.com>
2222 Ilya Verbin <ilya.verbin@intel.com>
2223 Kirill Yukhin <kirill.yukhin@intel.com>
2224 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2226 * config/i386/sse.md (VI4_AVX512F): New.
2227 (VI8_AVX2_AVX512F): Ditto.
2228 (mul<mode>3): Extended with wider modes.
2229 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2230 (mul<mode>3): Ditto.
2231 (vec_widen_<s>mult_odd_<mode>): Ditto.
2233 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2234 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2235 Sergey Lega <sergey.s.lega@intel.com>
2236 Anna Tikhonova <anna.tikhonova@intel.com>
2237 Ilya Tocar <ilya.tocar@intel.com>
2238 Andrey Turetskiy <andrey.turetskiy@intel.com>
2239 Ilya Verbin <ilya.verbin@intel.com>
2240 Kirill Yukhin <kirill.yukhin@intel.com>
2241 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2243 * config/i386/sse.md (VI2_AVX512F): New.
2244 (VI124_AVX512F): Ditto.
2245 (sseunpackmode): Extended with wider modes.
2246 (sseunpackfltmode): Ditto.
2247 (vec_unpacks_float_hi_<mode>): Ditto.
2248 (vec_unpacks_float_lo_<mode>): Ditto.
2249 (vec_unpacku_float_hi_<mode>): Ditto.
2250 (vec_unpacku_float_lo_<mode>): Ditto.
2251 (vec_unpacks_lo_<mode>): Ditto.
2252 (vec_unpacks_hi_<mode>): Ditto.
2253 (vec_unpacku_lo_<mode>): Ditto.
2254 (vec_unpacku_hi_<mode>): Ditto.
2256 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2257 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2258 Sergey Lega <sergey.s.lega@intel.com>
2259 Anna Tikhonova <anna.tikhonova@intel.com>
2260 Ilya Tocar <ilya.tocar@intel.com>
2261 Andrey Turetskiy <andrey.turetskiy@intel.com>
2262 Ilya Verbin <ilya.verbin@intel.com>
2263 Kirill Yukhin <kirill.yukhin@intel.com>
2264 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2266 * config/i386/i386.md (multdiv): New.
2267 (multdiv_mnemonic): Ditto.
2268 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2269 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2270 (<sse>_vmdiv<mode>3): Removed.
2272 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2273 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2274 Sergey Lega <sergey.s.lega@intel.com>
2275 Anna Tikhonova <anna.tikhonova@intel.com>
2276 Ilya Tocar <ilya.tocar@intel.com>
2277 Andrey Turetskiy <andrey.turetskiy@intel.com>
2278 Ilya Verbin <ilya.verbin@intel.com>
2279 Kirill Yukhin <kirill.yukhin@intel.com>
2280 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2282 * config/i386/sse.md (V): Extended with wider modes.
2284 (ssehalfvecmode): Ditto.
2286 (ssepackfltmode): Ditto.
2287 (avx_vec_concat<mode>): Ditto.
2288 (V_256_512): New iterator.
2289 (VF2_512_256): Ditto.
2290 (si2dfmode): New attribute.
2291 (si2dfmodelower): Ditto.
2293 (concat_tg_mode): Ditto.
2294 (floatv4siv4df2): Changed to ...
2295 (float<si2dfmodelower><mode>2): This.
2296 (avx_cvtps2pd256): Changed to ...
2297 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2298 (vec_pack_trunc_v4df): Changed to ...
2299 (vec_pack_trunc_<mode>): This.
2300 (avx_vpermil<mode>): Changed to ...
2301 (<sse2_avx_avx512f>_vpermil<mode>): This.
2302 (<fixsuffix>fix_truncv8dfv8si2): New.
2303 (vec_pack_sfix_trunc_v8df): Ditto.
2304 (avx512f_rndscale<mode>): Ditto.
2305 (avx512f_roundpd512): Ditto.
2306 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2308 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2309 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2310 Sergey Lega <sergey.s.lega@intel.com>
2311 Anna Tikhonova <anna.tikhonova@intel.com>
2312 Ilya Tocar <ilya.tocar@intel.com>
2313 Andrey Turetskiy <andrey.turetskiy@intel.com>
2314 Ilya Verbin <ilya.verbin@intel.com>
2315 Kirill Yukhin <kirill.yukhin@intel.com>
2316 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2318 * config/i386/i386.md (any_fix): New iterator.
2319 (fixsuffix): New attribute.
2320 * config/i386/sse.md (VF1): Extened with wider modes.
2324 (sseintvecmodelower): Ditto.
2325 (ssescalarmode): Ditto.
2326 (ssescalarnum): Ditto.
2328 (ssexmmmode): Ditto.
2329 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2330 (<sse>_rcp<mode>2): Change iterator.
2331 (rsqrt<mode>2): Ditto.
2332 (<sse>_rsqrt<mode>2): Ditto.
2333 (avx2_vec_dup<mode>): Ditto.
2334 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2335 (round<mode>2_sfix): Ditto.
2336 (avx2_pbroadcast<mode>): Ditto.
2337 (*andnot<mode>3): Handle XI mode.
2338 (*<code><mode>3): Ditto.
2339 (AVXTOSSEMODE): Removed.
2340 (avx_vpermil<mode>): Changed to ...
2341 (<sse2_avx_avx512f>_vpermil<mode>): This.
2343 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2344 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2345 Sergey Lega <sergey.s.lega@intel.com>
2346 Anna Tikhonova <anna.tikhonova@intel.com>
2347 Ilya Tocar <ilya.tocar@intel.com>
2348 Andrey Turetskiy <andrey.turetskiy@intel.com>
2349 Ilya Verbin <ilya.verbin@intel.com>
2350 Kirill Yukhin <kirill.yukhin@intel.com>
2351 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2353 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2354 (<sse>_comi): Ditto.
2355 (<sse>_ucomi): Ditto.
2356 (sse_cvtss2siq_2): Ditto.
2357 (sse2_cvtsd2si): Ditto.
2358 (sse2_cvtsd2siq): Ditto.
2359 (sse2_cvttsd2si): Ditto.
2360 (sse2_cvttsd2siq): Ditto.
2361 (<shift_insn><mode>3): Ditto.
2362 (sse2_cvtsi2sdq): Update constraint and prefix.
2363 (sse_cvtsi2ss): Update prefix.
2364 (sse_cvtsi2ssq): Ditto.
2366 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2368 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2369 ignore internal calls.
2371 2013-10-11 Richard Biener <rguenther@suse.de>
2373 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2374 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2375 TREE_OVERFLOW is set.
2377 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2379 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2381 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2382 * gimple.h: Likewise.
2384 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2386 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2389 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2391 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2392 * configure: Regenerate.
2394 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2396 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2397 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2398 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2399 similarly pp_right_paren.
2400 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2401 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2402 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2403 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2405 (omp_declare_simd_clauses_equal,
2406 omp_remove_redundant_declare_simd_attrs): New functions.
2407 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2408 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2409 * tree.h (OMP_LOOP_CHECK): Define.
2410 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2411 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2412 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2413 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2414 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2415 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2416 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2417 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2418 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2419 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2420 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2421 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2422 (omp_remove_redundant_declare_simd_attrs): New prototype.
2423 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2424 GIMPLE_OMP_TEAMS): New codes.
2425 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2426 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2428 (target_nesting_level): New variable.
2429 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2430 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2431 for collapse > 1 static schedule unless ordered.
2432 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2433 (determine_parallel_type): Adjust get_ws_args_for caller.
2434 (install_var_field): Handle mask & 4 for double indirection.
2435 (scan_sharing_clauses): Ignore shared clause on teams construct.
2436 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2437 (create_omp_child_function): If inside target or declare target
2438 constructs, set "omp declare target" attribute on the child function.
2439 (find_combined_for): New function.
2440 (scan_omp_parallel): Handle combined loops.
2441 (scan_omp_target, scan_omp_teams): New functions.
2442 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2443 restrictions and set ctx->cancellable for cancellable constructs.
2444 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2445 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2446 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2447 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2448 (omp_clause_aligned_alignment): New function.
2449 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2450 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2451 on teams constructs. Handle user defined reductions and new
2453 (lower_reduction_clauses): Don't set placeholder to address of ref
2454 if it has already the right type.
2455 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2456 (expand_parallel_call): Use the new non-_start suffixed builtins,
2457 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2458 and GOMP_parallel_end after the call.
2459 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2460 (expand_omp_for_init_counts): Handle combined loops.
2461 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2463 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2464 end of cancellable loops.
2465 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2466 Likewise. Handle collapse > 1 loops.
2467 (expand_omp_simd): Handle combined loops.
2468 (expand_omp_for): Add inner_stmt argument, adjust callers of
2469 expand_omp_for* functions, use expand_omp_for_static*chunk even
2470 for collapse > 1 unless ordered.
2471 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2472 of cancellable sections.
2473 (expand_omp_single): Remove need_barrier variable, just rely on
2474 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2475 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2476 (expand_omp_atomic_load, expand_omp_atomic_store,
2477 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2478 (expand_omp_target): New function.
2479 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2480 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2481 (build_omp_regions_1): Immediately close region for
2482 GF_OMP_TARGET_KIND_UPDATE.
2483 (maybe_add_implicit_barrier_cancel): New function.
2484 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2486 (lower_omp_single): Likewise. Add clobber after the barrier.
2487 (lower_omp_taskgroup): New function.
2488 (lower_omp_for): Handle combined loops. Adjust
2489 lower_rec_input_clauses caller. Handle cancellation.
2490 (lower_depend_clauses): New function.
2491 (lower_omp_taskreg): Lower depend clauses. Adjust
2492 lower_rec_input_clauses caller. Add clobber after the call. Handle
2494 (lower_omp_target, lower_omp_teams): New functions.
2495 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2496 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2497 and GOMP_cancellation_point calls.
2498 (lower_omp): Fold stmts inside of target region.
2499 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2500 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2501 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2502 (BT_FN_VOID_OMPFN_PTR_UINT,
2503 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2504 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2505 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2506 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2507 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2508 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2509 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2510 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2511 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2512 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2513 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2514 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2515 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2516 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2517 BUILT_IN_GOMP_PARALLEL_END.
2518 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2519 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2520 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2521 GF_OMP_FOR_KIND_DISTRIBUTE.
2522 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2523 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2524 (dump_gimple_omp_return): Print lhs if it has any.
2525 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2526 gimple_omp_atomic_seq_cst_p.
2527 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2528 and GIMPLE_OMP_TEAMS.
2529 * langhooks.c (lhd_omp_mappable_type): New function.
2530 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2531 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2533 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2534 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2535 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2536 (struct gimplify_omp_ctx): Add combined_loop field.
2537 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2538 on stmts inside of target region.
2539 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2540 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2541 ORT_TARGET and ORT_TARGET_DATA.
2542 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2544 (omp_notice_threadprivate_variable): Complain about threadprivate
2545 variables in target region.
2546 (omp_notice_variable): Complain about vars with non-mappable type
2547 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2548 (omp_check_private): Ignore ORT_TARGET* regions.
2549 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2550 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2551 (find_combined_omp_for): New function.
2552 (gimplify_omp_for): Handle gimplification of combined loops.
2553 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2555 (gimplify_omp_target_update): New function.
2556 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2557 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2558 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2559 (gimplify_body): If fndecl has "omp declare target" attribute, add
2560 implicit ORT_TARGET context around it.
2561 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2562 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2563 * tree-nested.c (convert_nonlocal_reference_stmt,
2564 convert_local_reference_stmt, convert_gimple_call): Handle
2565 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2566 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2567 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2568 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2569 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2570 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2571 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2572 BUILT_IN_GOMP_LOOP_END_CANCEL,
2573 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2574 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2575 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2576 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2577 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2578 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2579 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2580 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2581 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2582 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2583 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2584 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2585 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2586 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2587 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2588 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2589 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2590 gimple_build_omp_teams): New functions.
2591 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2592 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2593 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2594 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2595 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2596 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2597 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2598 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2599 GF_OMP_ATOMIC_SEQ_CST): New.
2600 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2601 gimple_build_omp_teams): New prototypes.
2602 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2603 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2604 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2605 GIMPLE_OMP_SINGLE as end of range.
2606 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2607 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2608 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2609 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2610 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2611 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2612 gimple_omp_target_kind, gimple_omp_target_set_kind,
2613 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2614 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2615 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2616 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2617 gimple_omp_teams_set_clauses): New inlines.
2618 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2619 and GIMPLE_OMP_TASKGROUP.
2620 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2621 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2622 enum omp_clause_proc_bind_kind): New.
2623 (union omp_clause_subcode): Add depend_kind, map_kind and
2624 proc_bind_kind fields.
2625 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2626 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2627 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2628 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2629 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2631 2013-10-10 Teresa Johnson <tejohnson@google.com>
2633 * predict.c (tree_estimate_probability): Add new parameter
2634 for estimate_bb_frequencies.
2635 (estimate_bb_frequencies): Add new parameter to force estimation.
2636 (rebuild_frequencies): When max frequency in function is small,
2637 recompute counts from frequencies.
2638 * predict.h (estimate_bb_frequencies): New parameter.
2640 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2642 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2643 optimizations are disabled.
2645 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2647 * coverage.c (coverage_finish): Fix leak of da_file_name.
2649 2013-10-10 Jan Hubicka <jh@suse.cz>
2651 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2652 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2655 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2658 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2659 if any labels are in FALLTHRU_BB, use a special label emitted
2660 immediately after the asm goto insn rather than label_rtx
2662 (expand_asm_stmt): Adjust caller.
2663 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2664 edge if the last insn in predecessor is a jump with single successor,
2665 but it isn't simplejump_p.
2667 2013-10-10 Richard Biener <rguenther@suse.de>
2669 PR tree-optimization/58656
2670 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2672 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2674 * gimplify.c: Include expr.h and tm_p.h for targets with special
2675 va-arg padding requirements.
2677 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2679 * tree-flow.h: Move some prototypes to gimple.h.
2680 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2681 * gimple.h: Relocate some prototypes from tree-flow.h
2682 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2684 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2685 (build_va_arg_indirect_ref): Relocate and make static.
2686 (std_gimplify_va_arg_expr): Relocate here.
2687 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2688 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2690 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2692 * doc/md.texi: Document the mnemonic attribute.
2694 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2697 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2698 (set_attr_alternative x ...) when searching for user defined
2701 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2703 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2705 2013-10-09 Easwaran Raman <eraman@google.com>
2707 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2708 * cfgexpand.c (defer_stack_allocation): ...use here
2709 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2711 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2713 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2714 (operand_equal_for_value_replacement): New function, extracted from
2715 value_replacement and enhanced to catch more cases.
2716 (value_replacement): Use operand_equal_for_value_replacement.
2718 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2720 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2721 get_max_loop_iterations.
2723 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2725 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2726 Place comment above function.
2728 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2730 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2731 are not related to tree-cfg.c.
2732 * tree-ssa-address.h: New file. Relocate prototypes.
2733 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2734 (addr_for_mem_ref): New. Combine call to get_address_description and
2735 return addr_for_mem_ref.
2736 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2737 * tree-ssa-live.h: Adjust prototypes.
2738 * passes.c: Include tree-ssa-live.h.
2739 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2740 * graphite.c (graphite_transform_loops): Make static.
2741 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2742 make_pass_graphite, pass_data_graphite_transforms,
2743 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2744 * ipa-pure-const.c (warn_function_noreturn): Make static.
2745 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2746 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2747 Relocate from tree-cfg.c
2748 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2750 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2751 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2752 Move to ipa-pure-const.c.
2753 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2754 Relocate from tree-optimize.c.
2755 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2756 make_pass_fixup_cfg): Move to tree-cfg.c.
2757 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2758 Relocate some prototypes.
2759 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2760 * tree-dump.c (dump_function_to_file): Remove prototype.
2761 Add tree-flow.h to the include file.
2762 * tree-dump.h: Remove prototype.
2763 * tree-parloops.h: New File. Add prototypes.
2764 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2765 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2766 from tree-ssa-loop.c.
2767 * tree-predcom.c (run_tree_predictive_commoning,
2768 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2769 Relocate here from tree-ssa-loop.c.
2770 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2771 ssa_name_values.release ().
2772 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2773 (ssa_name_values): Relocate from tree-flow.h.
2774 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2775 * tree-ssa-loop.c (run_tree_predictive_commoning,
2776 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2777 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2778 make_pass_graphite, pass_data_graphite_transforms,
2779 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2780 tree_parallelize_loops, pass_data_parallelize_loops,
2781 make_pass_parallelize_loops): Move to other files.
2782 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2784 * tree.h: Remove prototypes from tree-address.c.
2786 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2788 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2789 (struct int_tree_map): Move to tree-hasher.h
2790 (SCALE, LABEL, PERCENT): Move to gimple.h
2791 * tree-flow-inline.h: Delete. Move functions to other files.
2792 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2793 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2794 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2795 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2796 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2797 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2798 * tree-hasher.h: Don't include tree-flow.h.
2799 (struct int_tree_map): Relocate from tree-flow.h.
2800 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2801 tree-flow-inline.h and make static.
2802 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2804 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2805 tree-flow-inline.h and make static.
2806 * tree.h (is_global_var, may_be_aliased): Relocate from
2808 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2809 * value-prof.c: No longer include tree-flow-inline.h.
2810 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2812 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2814 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2815 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2816 (enum move_pos): Move to tree-ssa-loop-im.h
2817 * cfgloop.h: Move some prototypes.
2818 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2819 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2820 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2821 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2822 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2823 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2824 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2825 (enum move_pos): Relocate here.
2826 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2828 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2829 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2830 make_pass_lim): Relocate here from tree-ssa-loop.c.
2831 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2833 (loop_edge_to_cancel, unloop_loops): Make static.
2834 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2835 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2836 (tree_complete_unroll, gate_tree_complete_unroll,
2837 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2838 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2839 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2840 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2841 (stmt_invariant_in_loop_p): Remove unused function.
2842 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2843 * tree-ssa-loop-manip.h: New file. Add prototypes.
2844 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2845 (gcov_type_to_double_int): Move to cfgloop.h.
2846 (double_int_cmp, bound_index,
2847 estimate_numbers_of_iterations_loop): Make static.
2848 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2849 (max_loop_iterations): Factor out get_max_loop_iterations.
2850 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2852 * tree-ssa-loop-niter.h: New file. Add prototypes.
2853 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2854 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2855 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2856 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2857 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2858 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2859 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2860 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2861 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2862 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2863 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2864 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2865 pass_data_complete_unroll, make_pass_complete_unroll,
2866 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2867 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2868 tree-ssa-loop-ivcanon.c.
2869 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2870 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2871 tree-ssa-loop-prefetch.c.
2872 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2874 (get_lsm_tmp_name): Relocate and add suffix parameter.
2875 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2876 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2877 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2878 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2879 (get_estimated_loop_iterations): Factor out accessor from
2880 estimated_loop_iterations in tree-ssa-loop-niter.c.
2881 (get_max_loop_iterations): Factor out accessor from
2882 _max_loop_iterations in tree-ssa-niter.c.
2883 * loop-unroll.c (decide_unroll_constant_iterations,
2884 decide_unroll_runtime_iterations, decide_peel_simple,
2885 decide_unroll_stupid): Use new get_* accessors.
2887 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2889 PR tree-optimization/20318
2890 * doc/extend.texi (returns_nonnull): New function attribute.
2891 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
2893 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
2894 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
2896 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
2899 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
2900 false if both components are bitfields.
2902 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2904 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
2905 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
2906 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
2907 * config/aarch64/aarch64.h
2908 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
2909 * config/aarch64/aarch64-simd-builtins.def
2910 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
2912 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2914 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
2916 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2918 * config/aarch64/arm_neon.h (vdiv_f64): Added.
2920 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
2922 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
2923 (vneg_f64): New intrinsic.
2924 (vneg_s8): Asm replaced with C.
2925 (vneg_s16): Likewise.
2926 (vneg_s32): Likewise.
2927 (vneg_s64): New intrinsic.
2928 (vnegq_f32): Asm replaced with C.
2929 (vnegq_f64): Likewise.
2930 (vnegq_s8): Likewise.
2931 (vnegq_s16): Likewise.
2932 (vnegq_s32): Likewise.
2933 (vnegq_s64): Likewise.
2935 2013-10-09 Renlin Li <Renlin.Li@arm.com>
2937 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
2939 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2941 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
2942 packed stack special handling.
2943 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
2944 back to fixed stack slots for FPRs saved due to stdarg.
2946 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2948 * config/s390/s390.c (s390_frame_info): Restructure function.
2950 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2952 * config/s390/s390.c (struct s390_frame_layout): New field
2954 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
2955 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
2956 regs_ever_clobbered to char*.
2957 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
2958 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
2959 (s390_register_info): Enable FPR save slots. Move/Copy some
2960 functionality into ...
2961 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
2962 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
2964 (s390_frame_info): Do gpr slot allocation here now. stdarg does
2965 not imply a stack frame.
2966 (s390_init_frame_layout): Remove variable clobbered_regs.
2967 (s390_update_register_info): Remove function.
2968 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
2969 cannot be used for register renaming.
2970 (s390_hard_regno_scratch_ok): New function.
2971 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
2972 (s390_initial_elimination_offset): Change offset calculation of
2973 the return address pointer.
2974 (save_gprs): Deal with only r6 being saved from the call-saved regs.
2975 (restore_gprs): Set frame related flag.
2976 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
2977 (s390_emit_prologue): Call s390_register_info instead of
2978 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2979 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2980 (s390_optimize_prologue): Call s390_optimize_register_info.
2981 Try to remove also FPR slot save/restore INSNs. Remove frame
2982 related flags from restore INSNs.
2984 2013-10-08 DJ Delorie <dj@redhat.com>
2986 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2989 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2990 avoid conflict with the MI use of %c.
2991 * config/rl78/rl78-real.md: change %c to %C throughout.
2992 * config/rl78/rl78-virt.md: Likewise.
2994 2013-10-08 Jan Hubicka <jh@suse.cz>
2996 * config/i386/i386.c (ix86_option_override_internal): Switch
2997 to SSE math for -ffast-math when target ISA supports SSE2.
2999 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3001 * tree-flow.h: Remove some prototypes.
3002 * tree.h: Remove some protypes, add a couple.
3003 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3004 using_eh_for_cleanups_p): Add interface routines for front ends.
3005 * tree-eh.h: New file. Add protoptyes.
3006 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3007 (add_stmt_to_eh_lp_fn): Make static.
3008 (lower_try_finally): Use new using_eh_for_cleanups_p.
3009 * emit-rtl.c: Include tree-eh.h.
3010 * gimple.h: Include tree-eh.h.
3012 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3014 PR tree-optimization/58480
3015 * tree-vrp.c (infer_nonnull_range): New function.
3016 (infer_value_range): Call infer_nonnull_range.
3018 2013-10-08 Dehao Chen <dehao@google.com>
3020 PR tree-optimization/58619
3021 * tree-inline.c (copy_phis_for_bb): Combine location data
3024 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3027 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3028 RTX_FRAME_RELATED_P on INSN.
3030 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3032 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3033 (altivec_expand_vec_perm_const): Call it.
3035 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3037 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3038 sequences for LE VSX loads and stores at expand time.
3039 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3041 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3042 (rs6000_gen_le_vsx_permute): New.
3043 (rs6000_gen_le_vsx_load): New.
3044 (rs6000_gen_le_vsx_store): New.
3045 (rs6000_gen_le_vsx_move): New.
3046 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3047 (*vsx_le_perm_load_v4si): New.
3048 (*vsx_le_perm_load_v8hi): New.
3049 (*vsx_le_perm_load_v16qi): New.
3050 (*vsx_le_perm_store_v2di): New.
3051 (*vsx_le_perm_store_v4si): New.
3052 (*vsx_le_perm_store_v8hi): New.
3053 (*vsx_le_perm_store_v16qi): New.
3054 (*vsx_xxpermdi2_le_<mode>): New.
3055 (*vsx_xxpermdi4_le_<mode>): New.
3056 (*vsx_xxpermdi8_le_V8HI): New.
3057 (*vsx_xxpermdi16_le_V16QI): New.
3058 (*vsx_lxvd2x2_le_<mode>): New.
3059 (*vsx_lxvd2x4_le_<mode>): New.
3060 (*vsx_lxvd2x8_le_V8HI): New.
3061 (*vsx_lxvd2x16_le_V16QI): New.
3062 (*vsx_stxvd2x2_le_<mode>): New.
3063 (*vsx_stxvd2x4_le_<mode>): New.
3064 (*vsx_stxvd2x8_le_V8HI): New.
3065 (*vsx_stxvd2x16_le_V16QI): New.
3067 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3069 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3071 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3073 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3074 loop to work also for 31bit ABI.
3075 Save the stack pointer for frame_size > 0.
3077 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3079 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3080 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3081 constraint letters from expanders.
3082 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3083 retry count to general_operand.
3084 ("tabort"): Give operand 0 a mode.
3085 ("tabort_1"): Add mode and constraint letter for operand 0.
3086 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3088 2013-10-04 Jeff Law <law@redhat.com>
3090 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3092 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3094 (thread_across_edge): Here. Call it.
3096 2013-10-04 Cary Coutant <ccoutant@google.com>
3098 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3099 discarding a location list expression (or a piece of one).
3101 2013-10-03 Jan Hubicka <jh@suse.cz>
3103 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3104 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3105 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3107 2013-10-03 Jan Hubicka <jh@suse.cz>
3109 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3110 accumulate-outgoing-args when producing unwind info.
3112 2013-10-03 Wei Mi <wmi@google.com>
3114 * lra-constraints.c (insert_move_for_subreg): New function
3115 extracted from simplify_operand_subreg.
3116 (simplify_operand_subreg): Add reload for paradoxical subreg.
3118 2013-10-03 Rong Xu <xur@google.com>
3120 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3121 the candidate of builtin_expect such that we should fix the
3122 size/time estimation.
3123 (estimate_function_body_sizes): Do the acutally size/time fix-up
3126 2013-10-03 Rong Xu <xur@google.com>
3128 * predict.c (tree_predict_by_opcode): Get the probability
3129 for builtin_expect from param builtin_expect_probability.
3130 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3131 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3132 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3134 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3137 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3138 * calls.c (alloca_call_p): Use get_callee_fndecl.
3139 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3140 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3142 (vrp_visit_stmt): Remove duplicated code.
3144 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3146 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3147 ceildf2, btruncdf2, instead of vsx_* name.
3149 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3150 iterators to only do V2DF and V4SF here. Move the DF code to
3151 rs6000.md where it is combined with SF mode. Replace <VSv> with
3152 just 'v' since only vector operations are handled with these insns
3153 after moving the DF support to rs6000.md.
3154 (vsx_sub<mode>3): Likewise.
3155 (vsx_mul<mode>3): Likewise.
3156 (vsx_div<mode>3): Likewise.
3157 (vsx_fre<mode>2): Likewise.
3158 (vsx_neg<mode>2): Likewise.
3159 (vsx_abs<mode>2): Likewise.
3160 (vsx_nabs<mode>2): Likewise.
3161 (vsx_smax<mode>3): Likewise.
3162 (vsx_smin<mode>3): Likewise.
3163 (vsx_sqrt<mode>2): Likewise.
3164 (vsx_rsqrte<mode>2): Likewise.
3165 (vsx_fms<mode>4): Likewise.
3166 (vsx_nfma<mode>4): Likewise.
3167 (vsx_copysign<mode>3): Likewise.
3168 (vsx_btrunc<mode>2): Likewise.
3169 (vsx_floor<mode>2): Likewise.
3170 (vsx_ceil<mode>2): Likewise.
3171 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3172 (vsx_sminsf3): Likewise.
3173 (vsx_fmadf4): Likewise.
3174 (vsx_fmsdf4): Likewise.
3175 (vsx_nfmadf4): Likewise.
3176 (vsx_nfmsdf4): Likewise.
3177 (vsx_cmpdf_internal1): Likewise.
3179 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3180 simpler to select whether a target has SPE or traditional floating
3181 point support in iterators.
3182 (TARGET_DF_SPE): Likewise.
3183 (TARGET_SF_FPR): Likewise.
3184 (TARGET_DF_FPR): Likewise.
3185 (TARGET_SF_INSN): Macros to say whether floating point support
3186 exists for a given operation for expanders.
3187 (TARGET_DF_INSN): Likewise.
3189 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3190 combining of SF/DF mode operations, using both traditional and VSX
3198 (abs<mode>2): Combine SF/DF modes using traditional floating point
3199 instructions. Add support for using the upper DF registers with
3200 VSX support, and SF registers with power8-vector support. Update
3201 expanders for operations supported by both the SPE and traditional
3202 floating point units.
3203 (abs<mode>2_fpr): Likewise.
3204 (nabs<mode>2): Likewise.
3205 (nabs<mode>2_fpr): Likewise.
3206 (neg<mode>2): Likewise.
3207 (neg<mode>2_fpr): Likewise.
3208 (add<mode>3): Likewise.
3209 (add<mode>3_fpr): Likewise.
3210 (sub<mode>3): Likewise.
3211 (sub<mode>3_fpr): Likewise.
3212 (mul<mode>3): Likewise.
3213 (mul<mode>3_fpr): Likewise.
3214 (div<mode>3): Likewise.
3215 (div<mode>3_fpr): Likewise.
3216 (sqrt<mode>3): Likewise.
3217 (sqrt<mode>3_fpr): Likewise.
3218 (fre<Fs>): Likewise.
3219 (rsqrt<mode>2): Likewise.
3220 (cmp<mode>_fpr): Likewise.
3221 (smax<mode>3): Likewise.
3222 (smin<mode>3): Likewise.
3223 (smax<mode>3_vsx): Likewise.
3224 (smin<mode>3_vsx): Likewise.
3225 (negsf2): Delete SF operations that are merged with DF.
3232 (fmasf4_fpr): Likewise.
3233 (fmssf4_fpr): Likewise.
3234 (nfmasf4_fpr): Likewise.
3235 (nfmssf4_fpr): Likewise.
3236 (sqrtsf2): Likewise.
3237 (rsqrtsf_internal1): Likewise.
3238 (smaxsf3): Likewise.
3239 (sminsf3): Likewise.
3240 (cmpsf_internal1): Likewise.
3241 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3242 (negdf2): Delete DF operations that are merged with SF.
3244 (nabsdf2): Likewise.
3250 (rsqrtdf_internal1): Likewise.
3251 (fmadf4_fpr): Likewise.
3252 (fmsdf4_fpr): Likewise.
3253 (nfmadf4_fpr): Likewise.
3254 (nfmsdf4_fpr): Likewise.
3255 (sqrtdf2): Likewise.
3256 (smaxdf3): Likewise.
3257 (smindf3): Likewise.
3258 (cmpdf_internal1): Likewise.
3259 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3260 (btrunc<mode>2): Delete separate expander, and combine with the
3261 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3262 (btrunc<mode>2_fpr): Likewise.
3263 (ceil<mode>2): Likewise.
3264 (ceil<mode>2_fpr): Likewise.
3265 (floor<mode>2): Likewise.
3266 (floor<mode>2_fpr): Likewise.
3267 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3268 Add support for using the upper registers with VSX and
3269 power8-vector. Move insns to be closer to the define_expands. On
3270 VSX systems, prefer the traditional form of FMA over the VSX
3271 version, since the traditional form allows the target not to
3272 overlap with the inputs.
3273 (fms<mode>4_fpr): Likewise.
3274 (nfma<mode>4_fpr): Likewise.
3275 (nfms<mode>4_fpr): Likewise.
3277 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3278 Richard Earnshaw <richard.earnshaw@arm.com>
3280 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3281 (struct mult_cost_table): Likewise.
3282 (struct mem_cost_table): Likewise.
3283 (struct fp_cost_table): Likewise.
3284 (struct vector_cost_table): Likewise.
3285 (cpu_cost_table): Likewise.
3286 * config/arm/arm.opt (mold-rts-costs): New option.
3287 (mnew-generic-costs): Likewise.
3288 * config/arm/arm.c (generic_extra_costs): New table.
3289 (cortexa15_extra_costs): Likewise.
3290 (arm_slowmul_tune): Use NULL as new costs.
3291 (arm_fastmul_tune): Likewise.
3292 (arm_strongarm_tune): Likewise.
3293 (arm_xscale_tune): Likewise.
3294 (arm_9e_tune): Likewise.
3295 (arm_v6t2_tune): Likewise.
3296 (arm_cortex_a5_tune): Likewise.
3297 (arm_cortex_a9_tune): Likewise.
3298 (arm_v6m_tune): Likewise.
3299 (arm_fa726te_tune): Likewise.
3300 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3301 (arm_cortex_tune): Use generict_extra_costs.
3302 (shifter_op_p): New function.
3303 (arm_unspec_cost): Likewise.
3304 (LIBCALL_COST): Define.
3305 (arm_new_rtx_costs): New function.
3306 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3307 table is available. Use old costs otherwise unless mnew-generic-costs
3309 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3310 (cpu_cost_table): Declare.
3312 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3315 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3316 (*subs_mul_imm_<mode>)
3317 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3318 (*sub_<shift>_<mode>)
3319 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3320 Remove k constraint.
3322 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3324 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3326 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3328 2013-10-02 Teresa Johnson <tejohnson@google.com>
3330 * predict.c (probably_never_executed): New function.
3331 (probably_never_executed_bb_p): Invoke probably_never_executed.
3332 (probably_never_executed_edge_p): Ditto.
3333 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3334 Treat profile insanities conservatively.
3336 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3338 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3340 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3342 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3343 use it. Use smaller increase for scratch. Don't increase reject
3344 for early clobber scratch.
3345 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3346 setting eliminated regs except setting fp from hfp.
3347 (lra_eliminate): Check lra_insn_recog_data on NULL.
3349 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3352 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3353 setting -mvsx-timode by default until the underlying problem is fixed.
3354 (RS6000_CPU, power7 defaults): Likewise.
3356 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3358 * config/x-linux (host-linux.o): Remove header dependencies.
3359 Use $(COMPILE) and $(POSTCOMPILE).
3360 * config/t-linux-android (linux-android.o): Ditto.
3362 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3364 * Makefile.in (expmed.o-warn): Remove.
3366 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3368 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3369 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3371 2013-10-02 Teresa Johnson <tejohnson@google.com>
3373 * dojump.c (do_jump_1): Divide probability between
3374 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3376 2013-10-02 Tom Tromey <tromey@redhat.com>
3378 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3380 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3382 * tree-flow.h: Remove some prototypes.
3383 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3384 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3385 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3386 mark_virtual_phi_result_for_renaming): Relocate here.
3387 * tree-into-ssa.h: Add prototypes.
3388 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3389 single_pred_before_succ_order.
3390 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3391 (nonfreeing_call_p) Move to gimple.c.
3392 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3394 * basic-block.h (single_pred_before_succ_order): Add prototype.
3395 * gimple.c (nonfreeing_call_p): Relocate here.
3396 * gimple.h: Add prototype.
3397 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3398 * tree-ssa-dom.h: New file. Relocate prototypes here.
3399 * tree-ssa.h: Include tree-ssa-dom.h.
3401 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3403 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3404 Use $(COMPILE) and $(POSTCOMPILE).
3406 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3408 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3410 * tree-flow.h: Remove some prototypes.
3411 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3412 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3413 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3414 propagate_tree_value*): Move from here to...
3415 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3416 propagate_tree_value*): Relocate here.
3417 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3418 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3419 * gimple-fold.c: Remove gimple-fold.h from include list.
3420 * tree-vrp.c: Remove gimple-fold.h from include list.
3421 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3422 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3423 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3424 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3425 * sese.c: Add tree-ssa-propagate.h to include list.
3427 2013-10-02 Richard Biener <rguenther@suse.de>
3429 * tree-loop-distribution.c: Include tree-vectorizer.h for
3431 (enum partition_kind): Remove PKIND_REDUCTION.
3432 (struct partition_s): Remove has_writes member, add reduction_p member.
3433 (partition_alloc): Adjust.
3434 (partition_builtin_p): Likewise.
3435 (partition_has_writes): Remove.
3436 (partition_reduction_p): New function.
3437 (partition_merge_into): Likewise.
3438 (generate_code_for_partition): Commonize builtin partition
3440 (rdg_cannot_recompute_vertex_p): Remove.
3441 (already_processed_vertex_p): Likewise.
3442 (rdg_flag_vertex): Do not set has_writes.
3443 (classify_partition): Adjust.
3444 (rdg_build_partitions): Do not set has_writes, treat all
3445 partitions as useful.
3446 (distribute_loop): Record number of library calls generated. Adjust.
3447 (tree_loop_distribution): Report number of loops and library
3448 calls generated as opt-info.
3450 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3452 * tree-flow.h: Include new .h files. Move prototypes.
3453 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3454 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3455 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3456 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3457 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3458 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3459 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3460 * tree.h (get_ref_base_and_extent): Move prototype out.
3461 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3463 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3464 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3465 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3466 * tree-scalar-evolution.c: Include tree.h.
3467 * sese.c: Include tree.h.
3468 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3469 * dwarf2out.c: Include tree-dfa.h.
3470 * tree-chrec.c: Include tree.h.
3471 * tree-data-ref.c: Include tree.h.
3473 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3475 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3478 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3480 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3481 Use $(COMPILE) and $(POSTCOMPILE).
3482 (sol2-cxx.o): Likewise.
3483 (sol2-stubs.o): Likewise.
3485 * config/x-solaris (host-solaris.o): Likewise.
3487 * config/sparc/t-sparc (sparc.o): Remove.
3488 (sparc-c.o): Remove header dependencies.
3489 Use $(COMPILE) and $(POSTCOMPILE).
3490 * config/sparc/x-sparc: Likewise.
3492 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3494 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3495 * config/arc/arc700.md: Likewise.
3496 * config/arc/arc-modes.def: Likewise.
3497 * config/arc/arc-simd.h: Likewise.
3498 * config/arc/t-arc-uClibc: Likewise.
3499 * config/arc/t-arc-newlib: Likewise.
3501 2013-10-02 Renlin Li <renlin.li@arm.com>
3503 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3505 (aarch64_expand_epilogue): Likewise.
3507 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3508 Yufeng Zhang <yufeng.zhang@arm.com>
3510 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3512 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3513 'base_in' represent a conversion and legal_cast_p_1 holds; set
3514 'base_in' with the returned value from get_unwidened.
3516 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3518 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3519 plus_constant is not used.
3521 2013-10-01 Wei Mi <wmi@google.com>
3523 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3524 * config/i386/i386.md: Add define_peephole2 to
3525 break partial reg stall for cvtss2sd/cvtsd2ss.
3527 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3529 * config/arc/arc.c (pass_arc_ifcvt::clone):
3530 Update for ctxt_ -> m_ctxt change.
3532 2013-10-01 Jeff Law <law@redhat.com>
3534 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3535 outgoing_edge and intermediate_edge fields. Instead store the path.
3536 (redirection_data::hash): Hash on the last edge's destination index.
3537 (redirection_data::equal): Check the entire thread path.
3538 (lookup_redirectio_data): Corresponding changes.
3539 (create_edge_and_update_destination_phis): Likewise.
3540 (thread_single_edge): Likewise.
3542 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3543 Diego Novillo <dnovillo@google.com>
3545 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3546 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3547 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3548 versions of these patterns.
3550 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3551 (__builtin_arc_aligned): Likewise.
3553 * config/arc/arc.md: Expand adc_0 comment stating the intended
3554 purpose and why it isn't ready.
3555 Replace commented out call_value_via_label_mixed with a
3556 plain comment about bl_s.
3558 * config/arc/arc.c (stdio.h): Don't include directly.
3559 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3560 Remove if (1) condition.
3561 (arc_encode_section_info): Fix comment.
3563 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3565 * config/arc/arc.c (arc_conditional_register_usage):
3566 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3567 Also set reg_alloc_order for DMA config regs.
3569 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3570 Jeremy Bennett <jeremy.bennett@embecosm.com>
3572 * doc/install.texi (--with-cpu): Mention ARC.
3573 (arc-*-elf32): New paragraph.
3574 (arc-linux-uclibc): Likewise.
3575 * doc/md.texi (Machine Constraints): Add ARC part.
3576 * doc/invoke.texi: (menu): Add ARC Options.
3577 (Machine Dependent Options) <ARC Options>: Add synopsis.
3578 (node ARC Options): Add.
3579 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3580 (ARC Built-in Functions): New section defining
3581 generic ARC built-in functions.
3582 (ARC SIMD Built-in Functions): New section defining SIMD specific
3584 (Declaring Attributes of Functions): Extended
3585 description of short_call and long_call attributes for ARC and
3586 added index entries.
3588 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3589 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3590 Joern Rennecke <joern.rennecke@embecosm.com>
3591 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3592 Brendan Kehoe <brendan@zen.org>
3593 Michael Eager <eager@eagercon.com>
3594 Simon Cook <simon.cook@embecosm.com>
3595 Jeremy Bennett <jeremy.bennett@embecosm.com>
3597 * config/arc, common/config/arc: New directories.
3599 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3600 Brendan Kehoe <brendan@zen.org>
3601 Simon Cook <simon.cook@embecosm.com>
3603 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3605 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3607 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3608 * tree-ssa-coalesce.h: New. Move prototype to here.
3609 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3610 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3611 (gimple_can_coalesce_p): Move to...
3612 * gimple.c (gimple_can_coalesce_p): Here.
3614 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3616 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3617 (dump_decl_set): Move to gimple.c.
3618 * gimple.h: Don't include tree-ssa-operands.h.
3619 (dump_decl_set): Add prototype.
3620 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3621 Move to gimple-ssa.h.
3622 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3623 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3624 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3625 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3626 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3628 (dump_decl_set): Relocate here.
3629 * gimple-ssa.h: New file.
3630 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3631 Relocate from gimple.h.
3632 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3633 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3634 swap_tree_operands and remove non-ssa path.
3635 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3636 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3638 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3639 vect_is_simple_reduction_1): Use swap_ssa_operands.
3640 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3641 (enum need_phi_state): Move to tree-into-ssa.c.
3642 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3643 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3644 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3645 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3646 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3647 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3648 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3649 num_imm_uses): Move to ssa-iterators.h.
3650 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3651 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3653 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3654 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3655 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3656 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3657 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3658 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3659 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3660 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3661 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3662 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3663 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3664 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3665 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3666 (set_phi_nodes): Move to tree-phinodes.h.
3667 * tree-ssa-operands.h (enum ssa_op_iter_type,
3668 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3669 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3670 (dump_decl_set): Remove prototype.
3671 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3672 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3673 (set_phi_nodes): Relocate from tree-flow-inline.h.
3674 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3676 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3677 include list. Temporarily add gimple.h to include list.
3678 * ssa-iterators.h: New file.
3679 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3680 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3681 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3682 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3683 Relocate from tree-ssa-operands.h.
3684 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3685 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3686 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3687 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3688 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3689 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3690 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3691 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3692 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3693 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3694 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3695 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3696 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3697 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3698 Relocate from tree-flow-inline.h.
3699 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3701 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3704 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3705 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3706 Insert '\t' to output template.
3707 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3708 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3709 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3710 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3712 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3714 * doc/install.texi (Host/target specific installation notes for GCC):
3715 Put @anchor before @heading.
3716 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3717 Use @email for email addresses.
3719 2013-10-01 Jeff Law <law@redhat.com>
3721 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3722 a vec. Only delete the path if we create one without successfully
3723 registering a jump thread.
3724 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3726 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3727 (paths): New vector of jump threading paths.
3728 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3729 (THREAD_PATH): New accessor macro for the entire thread path.
3730 (lookup_redirection_data): Get intermediate and final outgoing edge
3731 from the thread path.
3732 (create_edge_and_update_destination_phis): Copy the threading path.
3733 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3734 jump threading path.
3735 (ssa_redirect_edges): Get edges and block types from the jump threading
3736 path. Free the path vector.
3737 (thread_block): Get edges from the jump threading path. Look at the
3738 entire path to see if we thread to a loop exit. If we cancel a jump
3739 thread request, then free the path vector.
3740 (thread_single_edge): Get edges and block types from the jump threading
3741 path. Free the path vector.
3742 (thread_through_loop_header): Get edges and block types from the jump
3743 threading path. Free the path vector.
3744 (mark_threaded_blocks): Iterate over the vector of paths and store
3745 the path on the appropriate edge. Get edges and block types from the
3746 jump threading path.
3747 (mark_threaded_blocks): Get edges and block types from the jump
3748 threading path. Free the path vector.
3749 (thread_through_all_blocks): Use the vector of paths rather than
3750 a vector of 3-edge sets.
3751 (register_jump_thread): Accept pointer to a path vector rather
3752 than the path vector itself. Store the path vector for later use.
3755 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3756 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3759 * config/s390/s390.c (s390_split_branches): Modify check for table
3761 (s390_chunkify_start): Rearrange table jump insn check in order to
3762 deal with compare and branch insns correctly.
3764 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3768 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3769 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3770 define_insn_and_split.
3771 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3772 (shiftsi3_compare): New pattern.
3774 * config/arm/unspecs.md (UNSPEC_RRX): New.
3776 2013-10-01 Alan Modra <amodra@gmail.com>
3778 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3779 casing inout operands.
3781 2013-10-01 Richard Biener <rguenther@suse.de>
3783 PR tree-optimization/58553
3784 * tree-loop-distribution.c (struct partition_s): Add niter member.
3785 (classify_partition): Populate niter member for the partition
3786 and properly identify whether the relevant store happens before
3787 or after the loop exit.
3788 (generate_memset_builtin): Use niter member from the partition.
3789 (generate_memcpy_builtin): Likewise.
3791 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3793 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3794 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3796 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3798 * basic-block.h (edge_list): Prefix member names with "m_".
3799 * context.h (context): Likewise.
3800 * domwalk.h (dom_walker): Likewise.
3801 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3802 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3803 * hash-table.h (hash_table): Likewise.
3804 * machmode.h (bit_field_mode_iterator): Likewise.
3805 * pass_manager.h (pass_list): Likewise.
3806 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3807 * tree-pass.h (pass_data): Likewise.
3808 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3809 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3810 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3811 * asan.c (pass_data_asan): Update accordingly.
3812 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3813 (control_dependences::control_dependences): Likewise.
3814 (control_dependences::~control_dependences): Likewise.
3815 (control_dependences::~control_dependences): Likewise.
3816 (control_dependences::get_edges_dependent_on): Likewise.
3817 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3818 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3819 * context.c (gcc::context::context): Likewise.
3820 * cprop.c (pass_rtl_cprop::clone): Likewise.
3821 * domwalk.c (dom_walker::walk): Likewise.
3822 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3823 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3824 * mode-switching.c (pass_mode_switching::clone): Likewise.
3825 * passes.c (opt_pass::opt_pass): Likewise.
3826 (pass_manager::pass_manager): Likewise.
3827 * predict.c (pass_strip_predict_hints::clone): Likewise.
3828 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3829 (pass_split_all_insns::clone): Likewise.
3830 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3832 (bit_field_mode_iterator::next_mode): Likewise.
3833 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3834 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3835 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3836 * tree-complex.c (pass_lower_complex::clone): Likewise.
3837 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3838 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3839 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3840 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3841 (pass_fold_builtins::clone): Likewise.
3842 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3843 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3844 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3845 (pass_cd_dce::clone): Likewise.
3846 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3847 (pass_phi_only_cprop::clone): Likewise.
3848 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3849 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3850 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3851 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3852 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3853 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3854 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3855 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3856 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3857 * tree-vrp.c (pass_vrp::clone): Likewise.
3858 * tsan.c (pass_tsan::clone): Likewise.
3860 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3863 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3864 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3867 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3868 optimization, punt if sign_bit_p looked through any zero extension.
3870 2013-09-30 Teresa Johnson <tejohnson@google.com>
3872 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3873 Update redirected out edge count in joiner case.
3874 (ssa_redirect_edges): Common the joiner and non-joiner cases
3875 so that joiner case gets profile updates.
3877 2013-09-30 Richard Biener <rguenther@suse.de>
3879 PR tree-optimization/58554
3880 * tree-loop-distribution.c (classify_partition): Require
3881 unconditionally executed stores for memcpy and memset recognition.
3882 (tree_loop_distribution): Calculate dominance info.
3884 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3886 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3887 (NO_PROFILE_COUNTERS): Likewise.
3888 (PROFILE_HOOK): Likewise.
3889 (FUNCTION_PROFILER): Likewise.
3890 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3892 2013-09-30 Iain Sandoe <iain@codesourcery.com>
3894 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
3895 calls and defines in TARGET_MACHO conditional.
3896 (load_macho_picbase_di): Likewise.
3897 (reload_macho_picbase): Likewise.
3898 (reload_macho_picbase_si): Likewise.
3899 (reload_macho_picbase_di): Likewise.
3900 (nonlocal_goto_receiver): Likewise.
3902 2013-09-30 Nick Clifton <nickc@redhat.com>
3904 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
3905 that use the MSP430X ISA.
3906 (msp430_option_override): Scan -mmcu command line option for any
3907 MCU name that supports the MSP430X ISA.
3908 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
3909 -mmcu options which enable the MSP430X ISA.
3911 2013-09-30 Richard Biener <rguenther@suse.de>
3914 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
3915 before looking for setjmp-like calls.
3917 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3920 * config/darwin-protos.h (machopic_get_function_picbase): New.
3921 * config/darwin.c (machopic_get_function_picbase): New.
3922 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
3923 label for a new func. (load_macho_picbase_di): Likewise.
3924 (reload_macho_picbase): New expand.
3925 (reload_macho_picbase_si): New insn.
3926 (reload_macho_picbase_di): New insn.
3927 (nonlocal_goto_receiver): New define and split.
3928 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
3929 (unspecv enum): Add UNSPECV_NLGR.
3931 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3933 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
3934 that altivec registers are correctly sized on Darwin.
3936 2013-09-29 Iain Sandoe <iain@codesourcery.com>
3938 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
3939 darwin-driver.o): Use COMPILE and POSTCOMPILE.
3940 * config/x-darwin (host-darwin.o): Likewise.
3941 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
3942 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
3943 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
3945 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
3947 * doc/invoke.texi: Fix usage of @tie{} command.
3949 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
3951 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
3953 2013-09-28 Jan Hubicka <jh@suse.cz>
3955 * config/i386/x86-tune.def: Add documentation for each of the options;
3958 2013-09-28 Jan Hubicka <jh@suse.cz>
3960 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
3962 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
3963 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
3964 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
3965 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
3966 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
3968 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
3970 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
3971 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
3972 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
3973 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
3974 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
3975 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
3976 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
3977 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3978 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3979 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3980 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3981 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3982 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3983 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3984 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3985 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3986 gimple.h, godump.c, graphite-clast-to-gimple.c,
3987 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3988 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3989 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3990 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3991 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3992 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3993 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3994 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3995 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3996 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3997 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3998 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3999 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4000 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4001 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4002 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4003 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4004 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4005 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4006 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4007 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4008 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4009 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4010 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4011 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4012 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4013 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4014 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4015 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4016 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4017 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4018 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4019 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4020 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4021 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4022 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4023 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4024 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4025 whitespace before "(".
4027 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4029 * expr.h (extract_bit_field): Remove packedp parameter.
4030 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4031 from forward declaration.
4032 (store_split_bit_field): Remove packedp arg from calls to
4033 extract_fixed_bit_field.
4034 (extract_bit_field_1): Remove packedp parameter and packedp
4035 argument from recursive calls and calls to extract_fixed_bit_field.
4036 (extract_bit_field): Remove packedp parameter and corresponding
4037 arg to extract_bit_field_1.
4038 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4040 (extract_split_bit_field): Remove packedp arg from call to
4041 extract_fixed_bit_field.
4042 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4043 (copy_blkmode_from_reg): Likewise.
4044 (copy_blkmode_to_reg): Likewise.
4045 (read_complex_part): Likewise.
4046 (store_field): Likewise.
4047 (expand_expr_real_1): Likewise.
4048 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4049 to extract_bit_field.
4050 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4051 call to extract_bit_field.
4052 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4053 call to extract_bit_field.
4054 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4055 and special packedp behavior from -fstrict-volatile-bitfields
4058 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4060 * lra-eliminations.c (init_elim_table): Guard value_p.
4062 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4064 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4065 DFmode, DImode, and SFmode in the upper VSX registers based on the
4066 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4067 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4068 of the VSX register handling.
4070 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4072 (zero_extendsidi2_lfiwzx): Add support for loading into the
4073 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4074 only do VSX memory options on Altivec registers.
4075 (extendsidi2_lfiwax): Likewise.
4076 (extendsfdf2_fpr): Likewise.
4077 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4078 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4079 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4080 (movdi_internal64): Likewise.
4082 2013-09-27 Xinliang David Li <davidxl@google.com>
4084 * opts.c (finish_options): Adjust parameters
4085 according to vect cost model.
4086 (common_handle_option): Set dynamic vect cost
4088 targhooks.c (default_add_stmt_cost): Compute stmt cost
4090 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4091 Use helper function.
4092 * tree-vectorizer.h (unlimited_cost_model): New function.
4093 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4094 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4096 (vect_enhance_data_refs_alignment): Ditto.
4097 * flag-types.h: New enum.
4098 * common/config/i386/i386-common.c (ix86_option_init_struct):
4099 No need to initialize vect_cost_model flag.
4100 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4103 2013-09-27 Diego Novillo <dnovillo@google.com>
4105 * gimple.h (enum ssa_mode): Remove.
4107 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4109 * cfgloop.h (number_of_loops): Fix typo in check for null.
4111 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4114 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4115 are children of outermost saved_cfun's loop, and set it up to
4116 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4117 if loop != loop0 and SESE region contains bbs that belong to loop0.
4119 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4121 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4122 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4123 (set_address_segment, set_address_base, set_address_index)
4124 (set_address_disp): Accept the argument unconditionally.
4125 (baseness): Remove must_be_base_p and must_be_index_p checks.
4126 (decompose_normal_address): Classify as much as possible in the
4129 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4131 * cse.c (count_reg_usage): Handle INT_LIST.
4132 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4133 * reginfo.c (reg_scan_mark_refs): Likewise.
4134 * reload1.c (eliminate_regs_1): Likewise.
4136 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4139 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4142 2013-09-27 Richard Biener <rguenther@suse.de>
4144 PR tree-optimization/58459
4145 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4146 restriction not propagating into loops.
4148 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4150 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4151 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4152 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4154 2013-09-26 Richard Biener <rguenther@suse.de>
4156 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4158 2013-09-26 Richard Biener <rguenther@suse.de>
4160 * alias.h (component_uses_parent_alias_set): Rename to ...
4161 (component_uses_parent_alias_set_from): ... this.
4162 * alias.c (component_uses_parent_alias_set): Rename to ...
4163 (component_uses_parent_alias_set_from): ... this and return
4165 (reference_alias_ptr_type_1): Use the result from
4166 component_uses_parent_alias_set_from instead of stripping
4167 components one at a time.
4168 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4170 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4172 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4173 Move prototypes to...
4174 * tree-ssa-ter.h: New File. Move prototypes here.
4175 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4176 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4177 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4178 from is_replaceable_p.
4179 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4180 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4181 (process_replaceable): Use ter_is_replaceable_p.
4182 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4183 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4184 newly refactored ssa_is_replaceable_p.
4185 * cfgexpand.c: Include tree-outof-ssa.h.
4186 * ssaexpand.h: Delete.
4188 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4190 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4191 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4193 (create_gimple_tmp): Delete.
4194 (get_expr_type, build_assign, build_type_cast): Move to...
4195 * gimple-builder.c: New File.
4196 (get_expr_type): Relocate from gimple.c.
4197 (build_assign, build_type_cast): Change to only create ssanames.
4198 * gimple.h: Move prototypes to...
4199 * gimple-builder.h: New File. Here.
4200 * tree-ssa.h: And here.
4201 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4202 count_uses_and_derefs): Relocate from gimple.c.
4203 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4204 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4205 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4206 gimple_replace_ssa_lhs.
4207 * asan.c: Include gimple-builder.h.
4208 * Makefile.in: Add gimple-builder.o.
4210 2013-09-26 Richard Biener <rguenther@suse.de>
4212 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4213 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4214 (loe_visit_block): Use gcc_checking_assert.
4215 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4216 gimple_assign_ssa_name_copy_p.
4217 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4220 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4222 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4223 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4225 2013-09-26 Richard Biener <rguenther@suse.de>
4227 PR tree-optimization/58539
4228 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4229 the fact that debug statements are not taking part in loop-closed
4232 2013-09-26 Nick Clifton <nickc@redhat.com>
4234 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4235 time warning message.
4236 (msp430_print_operand_raw): Delete unused letter parameter.
4237 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4238 (msp430_print_operand_address): New function.
4239 (msp430_print_operand): Move address printing code from here to
4241 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4243 (zero_extendpsisi2): Likewise.
4244 (extendpsisi2): New pattern.
4245 (andneghi3): New pattern.
4247 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4249 * config/aarch64/aarch64.opt (mlra): New option.
4250 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4251 (TARGET_LRA_P): Define.
4253 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4255 * expr.c (expand_assignment): Remove obsolete comment.
4257 2013-09-25 Jeff Law <law@redhat.com>
4259 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4260 tree-ssa-threadupdate.h.
4261 (register_jump_thread): Similarly.
4262 * tree-ssa-threadupdate.h: New header file.
4263 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4264 * tree-vrp.c: Likewise.
4265 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4266 (thread_around_empty_blocks): Change type of path vector argument to
4267 an edge,type pair from just an edge. Initialize both elements when
4268 appending to a jump threading path. Tweak references to elements
4270 (thread_across_edge): Similarly. Release memory for the elements
4272 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4273 (dump_jump_thread_path): New function broken out from
4274 register_jump_thread.
4275 (register_jump_thread): Use dump_jump_thread_path. Change type of
4276 path vector entries. Search the path for NULL edges and dump
4277 the path if one is found. Tweak the conversion of path to 3-edge
4278 form to use the block copy type information embedded in the path.
4280 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4282 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4284 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4285 Vladimir Makarov <vmakarov@redhat.com>
4287 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4288 from the least significant bit.
4289 (strip_address_mutations): Add bitfield operations handling.
4290 (must_be_index_p): Add shifting and rotate operations handling.
4291 (set_address_base): Use must_be_base_p predicate.
4292 (set_address_index): Use must_be_index_p predicate.
4294 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4295 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4296 Sergey Lega <sergey.s.lega@intel.com>
4297 Anna Tikhonova <anna.tikhonova@intel.com>
4298 Ilya Tocar <ilya.tocar@intel.com>
4299 Andrey Turetskiy <andrey.turetskiy@intel.com>
4300 Ilya Verbin <ilya.verbin@intel.com>
4301 Kirill Yukhin <kirill.yukhin@intel.com>
4302 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4304 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4306 (ix86_expand_vector_move_misalign): Support new unaligned load and
4307 stores and use new names.
4308 (CODE_FOR_sse2_storedqu): Rename to ...
4309 (CODE_FOR_sse2_storedquv16qi): ... this.
4310 (CODE_FOR_sse2_loaddqu): Rename to ...
4311 (CODE_FOR_sse2_loaddquv16qi): ... this.
4312 (CODE_FOR_avx_loaddqu256): Rename to ...
4313 (CODE_FOR_avx_loaddquv32qi): ... this.
4314 (CODE_FOR_avx_storedqu256): Rename to ...
4315 (CODE_FOR_avx_storedquv32qi): ... this.
4316 * config/i386/i386.md (fpint_logic): New.
4317 * config/i386/sse.md (VMOVE): Extend for AVX512.
4321 (VI_UNALIGNED_LOADSTORE): Ditto.
4322 (sse2_avx_avx512f): Ditto.
4323 (sse2_avx2): Extend for AVX512.
4324 (sse4_1_avx2): Ditto.
4325 (avx2_avx512f): New.
4326 (sse): Extend for AVX512.
4329 (avxsizesuffix): Ditto.
4330 (sseintvecmode): Ditto.
4332 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4333 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4334 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4335 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4336 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4337 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4338 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4339 (STORENT_MODE): Extend for AVX512.
4340 (*absneg<mode>2): Replace constraint "x" with "v".
4341 (*mul<mode>3): Ditto.
4342 (*ieee_smin<mode>3): Ditto.
4343 (*ieee_smax<mode>3): Ditto.
4344 (avx_cmp<mode>3): Replace VF with VF_128_256.
4345 (*<sse>_maskcmp<mode>3_comm): Ditto.
4346 (<sse>_maskcmp<mode>3): Ditto.
4347 (<sse>_andnot<mode>3): Extend for AVX512.
4348 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4349 (<code><mode>3, fpint_logic): New.
4350 (*<code><mode>3): Extend for AVX512.
4351 (avx512flogicsuff): New.
4352 (avx512f_<logic><mode>): Ditto.
4353 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4355 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4356 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4357 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4358 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4359 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4360 (xop_vpermil2<mode>3): Ditto.
4361 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4362 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4363 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4364 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4366 2013-09-25 Tom Tromey <tromey@redhat.com>
4368 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4369 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4370 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4371 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4372 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4373 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4374 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4375 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4376 (GRAPHITE_HTAB_H): Remove.
4378 2013-09-25 Tom Tromey <tromey@redhat.com>
4380 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4382 2013-09-25 Tom Tromey <tromey@redhat.com>
4384 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4386 2013-09-25 Tom Tromey <tromey@redhat.com>
4388 * config/i386/t-i386 (i386.o): Remove.
4389 (i386-c.o): Use COMPILE and POSTCOMPILE.
4391 2013-09-25 Tom Tromey <tromey@redhat.com>
4393 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4395 2013-09-25 Tom Tromey <tromey@redhat.com>
4397 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4398 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4399 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4400 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4401 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4402 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4403 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4404 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4405 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4406 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4407 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4408 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4409 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4410 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4411 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4412 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4413 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4414 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4415 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4416 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4417 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4418 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4419 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4420 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4421 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4422 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4423 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4424 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4425 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4426 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4427 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4428 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4429 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4430 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4431 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4432 (graphite-clast-to-gimple.o, graphite-dependences.o)
4433 (graphite-interchange.o, graphite-poly.o)
4434 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4435 (graphite-optimize-isl.o, tree-vect-loop.o)
4436 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4437 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4438 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4439 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4440 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4441 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4442 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4443 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4444 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4445 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4446 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4447 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4448 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4449 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4450 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4451 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4452 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4453 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4454 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4455 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4456 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4457 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4458 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4459 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4460 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4461 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4462 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4463 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4464 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4465 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4466 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4467 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4468 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4469 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4470 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4471 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4472 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4473 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4474 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4475 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4476 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4477 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4478 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4479 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4480 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4481 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4482 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4483 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4484 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4485 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4486 (gcov-dump.o): Remove.
4487 (default-c.o): Use COMPILE and POSTCOMPILE.
4488 (CFLAGS-gcc.o): New variable.
4489 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4491 2013-09-25 Tom Tromey <tromey@redhat.com>
4493 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4494 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4495 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4496 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4497 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4498 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4499 (c-family/array-notation-common.o, c-family/stub-objc.o)
4500 (c-family/c-ubsan.o): Remove.
4502 2013-09-25 Tom Tromey <tromey@redhat.com>
4504 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4506 2013-09-25 Tom Tromey <tromey@redhat.com>
4508 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4509 to add -DENABLE_SHARED_LIBGCC.
4510 (gcc.o): Don't use subshell.
4512 2013-09-25 Tom Tromey <tromey@redhat.com>
4514 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4515 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4516 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4517 * configure, config.in: Rebuild.
4519 2013-09-25 Tom Tromey <tromey@redhat.com>
4521 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4522 (COMPILE, POSTCOMPILE): New variables.
4523 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4524 (DEPFILES): New variable.
4525 Include ".Po" files.
4526 * configure.ac: Add checks for dependency checking.
4527 * configure, aclocal.m4: Regenerate.
4529 2013-09-25 Tom Tromey <tromey@redhat.com>
4531 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4532 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4534 2013-09-25 Tom Tromey <tromey@redhat.com>
4536 * Makefile.in (generated_files): Add options.h,
4537 target-hooks-def.h, insn-opinit.h,
4538 common/common-target-hooks-def.h, pass-instances.def,
4539 c-family/c-target-hooks-def.h.
4541 2013-09-25 Jeff Law <law@redhat.com>
4543 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4544 than foo[foo.length () - 1] to access last member in a vec.
4545 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4547 2013-09-25 Richard Biener <rguenther@suse.de>
4550 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4552 2013-09-25 Jan Hubicka <jh@suse.cz>
4554 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4557 2013-09-25 Marek Polacek <polacek@redhat.com>
4560 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4561 when determining the type name.
4563 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4565 * config/sh/sh.md: Fix formatting.
4567 2013-09-24 Xinliang David Li <davidxl@google.com>
4569 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4570 max peel iterations parameter.
4571 * param.def: New parameter.
4572 * doc/invoke.texi: Document New parameter.
4574 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4576 * gimple-pretty-print.c: Various whitespace tweaks.
4577 * tree-core.h: Likewise.
4578 * tree-pretty-print.c: Likewise.
4579 * tree-ssa-alias.c: Likewise.
4580 * tree-ssa-copy.c: Likewise.
4581 * tree-ssanames.c: Likewise.
4582 * tree-ssanames.h: Likewise.
4583 * tree-vrp.c: Likewise.
4585 2013-09-24 Alan Modra <amodra@gmail.com>
4589 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4590 for output operands that disallow regs. Don't use EXPAND_WRITE on
4593 2013-09-24 Richard Biener <rguenther@suse.de>
4596 * tree.c (reference_alias_ptr_type): Move ...
4597 * alias.c (reference_alias_ptr_type): ... here and implement
4598 in terms of the new reference_alias_ptr_type_1.
4599 (ref_all_alias_ptr_type_p): New helper.
4600 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4601 use ref_all_alias_ptr_type_p.
4602 (get_deref_alias_set): Add flag_strict_aliasing check here.
4603 (reference_alias_ptr_type_1): New function, split out from ...
4604 (get_alias_set): ... here.
4605 (alias_ptr_types_compatible_p): New function.
4606 * alias.h (reference_alias_ptr_type): Declare.
4607 (alias_ptr_types_compatible_p): Likewise.
4608 * tree.h (reference_alias_ptr_type): Remove.
4609 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4610 to compare MEM_REF alias types.
4612 2013-09-24 Richard Biener <rguenther@suse.de>
4614 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4616 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4618 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4619 reload helper function arrays into a single array reg_addr.
4620 (reload_fpr_gpr): Likewise.
4621 (reload_gpr_vsx): Likewise.
4622 (reload_vsx_gpr): Likewise.
4623 (struct rs6000_reg_addr): Likewise.
4624 (reg_addr): Likewise.
4625 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4626 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4627 (rs6000_init_hard_regno_mode_ok): Likewise.
4628 (rs6000_secondary_reload_direct_move): Likewise.
4629 (rs6000_secondary_reload): Likewise.
4631 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4632 constraints: wu, ww, and wy. Repurpose wv constraint added during
4633 power8 changes. Put wg constraint in alphabetical order.
4635 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4636 for future work to add ISA 2.07 VSX single precision support.
4637 (-mvsx-scalar-double): Change default from -1 to 1, update
4638 documentation comment.
4639 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4640 (-mupper-regs-df): New debug switch to control whether DF values
4641 can go in the traditional Altivec registers.
4642 (-mupper-regs-sf): New debug switch to control whether SF values
4643 can go in the traditional Altivec registers.
4645 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4647 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4648 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4649 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4650 (rs6000_debug_legitimate_address_p): Print if we are running
4651 before, during, or after reload.
4652 (rs6000_secondary_reload): Add a comment.
4653 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4655 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4656 constraints. Update documentation string.
4657 (wd constraint): Likewise.
4658 (wf constraint): Likewise.
4659 (wg constraint): Likewise.
4660 (wn constraint): Likewise.
4661 (ws constraint): Likewise.
4662 (wt constraint): Likewise.
4663 (wx constraint): Likewise.
4664 (wz constraint): Likewise.
4665 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4667 (ww constraint): Likewise.
4668 (wy constraint): Likewise.
4669 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4670 in the previous submissions.
4671 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4673 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4675 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4676 (REG_BR_PROB): Say that the probability is stored in an int_list.
4677 * reg-notes.def: Update commentary to mention INT_LIST.
4678 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4679 (INT_LIST): New rtx.
4680 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4681 * rtlanal.c (int_reg_note_p): New function.
4682 (alloc_reg_note): Assert that the note does not have an int argument.
4683 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4684 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4685 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4686 rather than an INSN_LIST. Handle INT_LIST.
4687 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4688 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4689 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4690 Manipulate them as ints rather than rtxes.
4691 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4692 * regmove.c (copy_src_to_dest): Likewise.
4693 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4695 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4696 into the cases that need it.
4697 * config/arm/arm.c (arm_unwind_emit): Likewise.
4699 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4700 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4701 * loop-doloop.c (add_test, doloop_modify): Likewise.
4702 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4703 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4704 * predict.c (combine_predictions_for_insn): Likewise.
4705 * print-rtl.c (print_rtx): Handle INT_LIST.
4706 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4707 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4708 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4709 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4710 (ix86_split_fp_branch, predict_jump): Likewise.
4711 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4712 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4713 * config/spu/spu.c (ea_load_store_inline): Likewise.
4715 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4716 value of a REG_BR_PROB note.
4717 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4718 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4719 * emit-rtl.c (try_split): Likewise.
4720 * predict.c (br_prob_note_reliable_p): Likewise.
4721 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4722 * reorg.c (mostly_true_jump): Likewise.
4723 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4724 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4725 * config/i386/i386.c (ix86_print_operand): Likewise.
4726 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4727 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4728 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4729 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4730 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4731 * config/sparc/sparc.c (output_cbranch): Likewise.
4732 * config/spu/spu.c (get_branch_target): Likewise.
4733 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4734 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4736 2013-09-23 Jan Hubicka <jh@suse.cz>
4738 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4740 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4741 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4743 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4744 (gimple_fold_call): Dump inconsistent devirtualizations; add
4745 sanity check for type based devirtualizations.
4746 * ipa-prop.c: Include ipa-utils.h
4747 (ipa_intraprocedural_devirtualization): Add sanity check.
4748 (try_make_edge_direct_virtual_call): Likewise.
4750 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4752 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4753 assignment statements.
4755 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4757 * gimple-pretty-print.c (dump_ssaname_info): New function.
4758 (dump_gimple_phi): Call it.
4759 (pp_gimple_stmt_1): Likewise.
4760 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4761 (range_info_def): Declare.
4762 * tree-pretty-print.c (pp_double_int): New function.
4763 (dump_generic_node): Call it.
4764 * tree-pretty-print.h (pp_double_int): Declare.
4765 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4766 * tree-ssanames.h (range_info_def): New structure.
4767 (value_range_type): Move definition here.
4768 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4770 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4772 (set_range_info): New function.
4773 (get_range_info): Likewise.
4774 (duplicate_ssa_name_range_info): Likewise.
4775 (duplicate_ssa_name_fn): Check pointer type and call
4776 duplicate_ssa_name_range_info.
4777 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4778 * tree-vrp.c (value_range_type): Remove definition, now in
4780 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4781 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4782 (SSA_NAME_RANGE_INFO): New macro.
4784 2013-09-23 Richard Biener <rguenther@suse.de>
4786 PR tree-optimization/58464
4787 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4788 (phi_trans_add): Change to add conditionally on being not present.
4789 (phi_translate_1): Remove recursion detection here.
4790 (phi_translate): Pre-seed the cache with NULL to catch
4791 recursion here in a more generic way.
4792 (bitmap_find_leader): Adjust comment.
4793 (get_representative_for): Dump value-numbers.
4794 (create_expression_by_pieces): Likewise.
4795 (insert_into_preds_of_block): Likewise.
4797 2013-09-23 Christian Bruel <christian.bruel@st.com>
4800 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4801 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4803 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4806 2013-09-20 Renlin Li <renlin.li@arm.com>
4808 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4810 (aarch64_expand_epilogue): Likewise.
4811 (aarch64_legitimize_reload_address): Likewise.
4813 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4815 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4816 NULL_TREE before pushing them onto the vector. Likewise for labels.
4818 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4820 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4821 comparison operators when -fno-trapping-math is in effect.
4822 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4823 comparison operators in TFmode and assert that unsupported operators
4825 (ia64_print_operand): Likewise.
4827 2013-09-21 Jan Hubicka <jh@suse.cz>
4829 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4830 (sse_typeless_stores): Enable for core
4831 (sse_load0_by_pxor): Likewise.
4832 (four_jump_limit): Disable for core.
4833 (pad_returns): Likewise.
4834 (avoid_vector_decode): Likewise.
4835 (fuse_cmp_and_branch): Enable for cores.
4836 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4838 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4841 * config/pa/pa.c (pa_option_override): Disable auto increment and
4842 decrement instructions until reload is completed.
4844 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4845 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4846 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4848 2013-09-20 DJ Delorie <dj@redhat.com>
4849 Nick Clifton <nickc@redhat.com>
4851 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4852 (need_to_save): Save bank 0 on interrupts.
4853 (characterize_address): Strip far address wrappers.
4854 (rl78_as_legitimate_address): Likewise.
4855 (transcode_memory_rtx): Likewise.
4856 (rl78_peep_movhi_p): Disable this peephole after devirt.
4857 (rl78_propogate_register_origins): Forget all origins when a
4859 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4860 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4862 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4863 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4864 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4866 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4868 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4870 * config/pa/pa.md: In "scc" insn patterns, change output template to
4871 handle const0_rtx in reg_or_0_operand operands.
4873 2013-09-20 Martin Husemann <martin@NetBSD.org>
4876 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4877 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4879 2013-09-20 Richard Biener <rguenther@suse.de>
4882 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4883 remembering SSA name version and block index.
4884 (new_scev_info_str): Adjust.
4885 (hash_scev_info): Likewise. Also hash the block index.
4886 (eq_scev_info): Adjust.
4887 (find_var_scev_info): Likewise.
4888 (struct instantiate_cache_entry): Remove.
4889 (struct instantiate_cache_type): Use a htab to map name, block
4891 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4892 (get_instantiated_value_entry): Likewise.
4893 (hash_idx_scev_info, eq_idx_scev_info): New functions.
4894 (instantiate_scev_name): Adjust.
4896 2013-09-20 Jeff Law <law@redhat.com>
4898 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
4900 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
4902 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
4903 Call aarch64_simd_expand_args to update op[argc].
4905 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4907 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
4910 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
4912 * gengtype.c (file_rules): Added rule for *.cc files.
4913 (get_output_file_with_visibility): Give fatal message when no
4916 2013-09-20 Renlin Li <renlin.li@arm.com>
4918 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
4919 (aarch64_expand_epilogue): Likewise.
4920 (aarch64_legitimize_reload_address): Likewise.
4922 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
4925 * expr.c (expand_assignment): Remove misalignp code path.
4927 2013-09-20 Marek Polacek <polacek@redhat.com>
4930 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
4931 TYPE_PRECISION. Add asserts.
4933 2013-09-20 Richard Biener <rguenther@suse.de>
4935 PR tree-optimization/58453
4936 * tree-loop-distribution.c (distribute_loop): Apply the cost
4937 model for -ftree-loop-distribute-patterns, too.
4939 2013-09-20 Richard Biener <rguenther@suse.de>
4942 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
4943 make type comparison less strict.
4945 2013-09-20 Alan Modra <amodra@gmail.com>
4947 * configure: Regenerate.
4948 * aclocal.m4: Regenerate.
4950 2013-09-20 Marek Polacek <polacek@redhat.com>
4953 * doc/extend.texi: Document that attribute used is meant to be used
4954 on variables with static storage duration.
4956 2013-09-19 Jakub Jelinek <jakub@redhat.com>
4958 PR tree-optimization/58472
4959 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
4960 simd_lane_access set inv_p = false.
4961 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
4962 the simduid magic VAR_DECL.
4964 2013-09-19 Jan Hubicka <jh@suse.cz>
4966 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
4968 2013-09-17 Jeff Law <law@redhat.com>
4970 * tree-ssa-dom.c (record_temporary_equivalences): New function
4971 split out of dom_opt_dom_walker::after_dom_children.
4972 (dom_opt_dom_walker::thread_across_edge): Move common code
4973 in here from dom_opt_dom_walker::after_dom_children.
4974 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4976 2013-09-19 Jan Hubicka <jh@suse.cz>
4978 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4979 (TARGET_GENERIC): Use PROCESOR_GENERIC
4980 (enum processor_type): Unify generic32 and 64.
4981 * i386.md (cpu): Likewise.
4982 * x86-tune.def (use_leave): Enable for generic32.
4983 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4984 * athlon.md: Change generic64 to generic in all occurences.
4985 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4986 (ix86_target_macros_internal): Likewise.
4987 * driver-i386.c (host_detect_local_cpu): Likewise.
4988 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4990 (generic_memcpy, generic_memset, generic_cost): This one.
4991 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4992 (m_GENERIC32, m_GENERIC64): Remove.
4993 (m_GENERIC): Turn into one flag.
4994 (processor_target): Unify generic tunnings.
4995 (ix86_option_override_internal): Replace generic32/64 by generic.
4996 (ix86_issue_rate): Likewise.
4997 (ix86_adjust_cost): Likewise.
4999 2013-09-19 Jan Hubicka <jh@suse.cz>
5001 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5002 of speculative flag.
5004 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5006 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5007 GOMP_sections_start, even if !exit_reachable.
5009 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5011 * lra-constraints.c (need_for_all_save_p): Use macro
5012 HARD_REGNO_CALL_PART_CLOBBERED.
5013 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5014 set up pseudo conflict hard regs.
5016 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5019 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5022 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5024 PR rtl-optimization/58438
5025 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5026 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5027 from previous subpasses.
5029 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5031 * arm.c (arm_get_frame_offsets): Validate architecture supports
5032 LDRD/STRD before accepting the tuning preference.
5033 (arm_expand_prologue): Likewise.
5034 (arm_expand_epilogue): Likewise.
5036 2013-09-18 Richard Biener <rguenther@suse.de>
5038 PR tree-optimization/58417
5039 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5040 have chrecs with symbols defined in the loop as operands.
5041 (chrec_fold_multiply): Likewise.
5042 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5043 parameters before folding binary operations.
5044 (struct instantiate_cache_entry_hasher): Remove.
5045 (struct instantiate_cache_type): Use a pointer-map.
5046 (instantiate_cache_type::instantiate_cache_type): New function.
5047 (instantiate_cache_type::get): Likewise.
5048 (instantiate_cache_type::set): Likewise.
5049 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5050 (get_instantiated_value_entry): Likewise.
5051 (global_cache): New global.
5052 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5053 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5054 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5055 (instantiate_scev_name): Adjust.
5056 (instantiate_scev): Construct global instead of local cache.
5057 (resolve_mixers): Likewise.
5059 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5060 Paolo Carlini <paolo.carlini@oracle.com>
5063 * doc/implement-cxx.texi: Fix references to the C++ standards.
5065 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5067 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5068 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5069 simd_lane_access drs, update also DR_ALIGNED_TO.
5071 2013-09-18 Marek Polacek <polacek@redhat.com>
5074 * doc/extend.texi: Document no_sanitize_undefined attribute.
5075 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5076 no_sanitize_undefined attribute.
5078 2013-09-18 Nick Clifton <nickc@redhat.com>
5080 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5081 (ASM_DECLARE_FUNCTION_NAME): Define.
5083 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5085 * compare-elim.c (find_comparison_dom_walker): New class
5086 (find_comparisons_in_bb): Rename to
5087 find_comparison_dom_walker::before_dom_children
5088 (find_comparisons): Adjust
5089 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5091 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5092 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5093 (init_walk_dominator_tree): Remove declaration.
5094 (fini_walk_dominator_tree): Remove declaration.
5095 * fwprop.c (single_def_use_dom_walker): New class
5096 (single_def_use_enter_block): Convert to
5097 single_def_use_dom_walker::before_dom_children.
5098 (single_def_use_leave_block): Convert to
5099 single_def_use_dom_walker::after_dom_children.
5100 (build_single_def_use_links): Adjust.
5101 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5103 (find_candidates_in_block): Convert to
5104 find_candidates_dom_walker::before_dom_children.
5105 (execute_strength_reduction): Adjust.
5106 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5107 (sese_dom_walker): New class.
5108 (sese_dom_walker::sese_dom_walker): New constructor.
5109 (sese_dom_walker::~sese_dom_walker): New destructor.
5110 (build_sese_conditions_before): Convert to
5111 sese_dom_walker::before_dom_children.
5112 (build_sese_conditions_after): Convert to
5113 sese_dom_walker::after_dom_children.
5114 (build_poly_scop): Adjust
5115 * tree-into-ssa.c (rewrite_dom_walker): New class
5116 (rewrite_enter_block): Convert to
5117 rewrite_dom_walker::before_dom_children.
5118 (rewrite_leave_block): Convert to
5119 rewrite_dom_walker::after_dom_children.
5120 (rewrite_update_dom_walker): New class.
5121 (rewrite_update_enter_block): Convert to
5122 rewrite_update_dom_walker::before_dom_children.
5123 (rewrite_update_leave_block): Convert to
5124 rewrite_update_dom_walker::after_dom_children.
5125 (rewrite_blocks, rewrite_into_ssa): Adjust.
5126 (mark_def_dom_walker): New class.
5127 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5128 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5129 (mark_def_sites_blocks): Convert to
5130 mark_def_dom_walker::before_dom_children.
5131 (mark_def_site_blocks): Remove.
5132 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5133 (tree_ssa_dominator_optimize): Adjust.
5134 (dom_thread_across_edge): Convert to method
5135 dom_opt_dom_walker::thread_across_edge.
5136 (dom_opt_enter_block): Convert to member function
5137 dom_opt_dom_walker::before_dom_children.
5138 (dom_opt_leave_block): Convert to member function
5139 dom_opt_dom_walker::after_dom_children.
5140 * tree-ssa-dse.c (dse_dom_walker): New class.
5141 (dse_enter_block): Convert to member function
5142 dse_dom_walker::before_dom_children.
5143 (tree_ssa_dse): Adjust.
5144 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5145 (determine_invariantness_stmt): Convert to method
5146 invariantness_dom_walker::before_dom_children.
5147 (determine_invariantness): Remove
5148 (move_computations_dom_walker): New class.
5149 (move_computations_stmt): Convert to method
5150 move_computations_dom_walker::before_dom_children.
5151 (move_computations, tree_ssa_lim): Adjust.
5152 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
5153 (nt_init_block): Make method
5154 notrappping_dom_walker::before_dom_children.
5155 (nt_fini_block): Make
5156 method nontrapping_dom_walker::after_dom_children.
5157 (get_non_trapping): Adjust.
5158 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5159 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
5160 (eliminate_leave_block): Make method.
5161 eliminate_dom_walker::after_dom_children.
5163 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5164 (strlen_enter_block): Make method
5165 strlen_dom_walker::before_dom_children.
5166 (strlen_leave_block): Make
5167 method strlen_dom_walker::after_dom_children.
5168 (tree_ssa_strlen): Adjust.
5169 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5170 (tree_ssa_uncprop): Adjust.
5171 (uncprop_leave_block): Make method
5172 uncprop_dom_walker::after_dom_children.
5173 (uncprop_leave_block): Make method
5174 uncprop_dom_walker::before_dom_children.
5176 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5178 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5179 branch in same basic block. Check both src and dest of the move insn.
5181 2013-09-17 Nick Clifton <nickc@redhat.com>
5183 * config/rl78/rl78-real.md (bf): New pattern.
5185 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5186 (rl78_print_operand): Do not put a # before a %B.
5187 * config/rl78/rl78.opt: Tweak doc strings.
5189 2013-09-17 DJ Delorie <dj@redhat.com>
5191 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5192 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5193 (rl78_option_override): Likewise, if -mallregs.
5194 (is_virtual_register): Likewise.
5195 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5196 (REGNO_OK_FOR_BASE_P): Likewise.
5197 * config/rl78/rl78.opt (-mallregs): New.
5199 2013-09-17 Nick Clifton <nickc@redhat.com>
5201 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5202 For interrupt functions: save all call clobbered registers if the
5203 interrupt handler is not a leaf function.
5204 (rl78_expand_prologue): Always recompute the frame information.
5205 For interrupt functions: only select bank 0 if one of the bank 0
5206 registers is going to be psuhed.
5208 2013-09-17 DJ Delorie <dj@redhat.com>
5210 * config/rl78/constraints.md: For each W* constraint, rename to C*
5211 and create a W* constraint that checks for an optional ES: prefix
5213 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5214 (es_addr): New. Used to wrap far addresses.
5215 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5216 (rl78_es_base): New.
5217 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5218 wrapped far addresses.
5219 (rl78_print_operand_1): Unwrap far addresses before processing.
5220 (rl78_lo16): Wrap far addresses in unspecs.
5221 (rl78_es_addr): New.
5222 (rl78_es_base): New.
5223 (insn_ok_now): Check for not-yet-wrapped far addresses.
5224 (transcode_memory_rtx): Properly re-wrap far addresses.
5226 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5228 * config/sparc/t-rtems: Add leon3 multilibs.
5230 2013-09-17 Cong Hou <congh@google.com>
5232 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5233 when checking the dot production pattern. The type of rhs operand
5234 of multiply is now checked correctly.
5236 2013-09-17 Jeff Law <law@redhat.com>
5238 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5239 edge implied equivalences into successor phis.
5240 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5241 here from tree-ssa-threadedge.c.
5242 (mark_threaded_blocks): When threading through a joiner, if both
5243 successors of the joiner's clone reach the same block, verify the
5244 PHI arguments are equal. If not, cancel the jump threading request.
5245 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5246 tree-ssa-threadupdate.c
5247 (thread_across_edge): Don't check PHI argument equality when
5248 threading through joiner block here.
5250 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5252 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5253 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5254 (warn_uninit, warn_uninitialized_vars,
5255 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5256 Move to tree-ssa-uninit.c.
5257 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5258 (has_undefined_value_p): New. Pass dependant parts of
5259 ssa_undefined_value_p.
5260 (uninit_undefined_value_p): Use has_undefined_value_p.
5261 (warn_uninit, warn_uninitialized_vars,
5262 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5263 Move from tree-ssa.c.
5264 * tree-ssa.h: Adjust prototypes.
5266 2013-09-17 Jan Hubicka <jh@suse.cz>
5269 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5270 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5271 FUNCTION_NOT_OPTIMIZED.
5272 * ipa-inline-analysis.c (compute_inline_parameters): Function
5273 not optimized is not inlinable unless it is alwaysinline.
5274 (inline_analyze_function): Force calls in not optimized
5275 function not inlinable.
5277 2013-09-17 Jan Hubicka <jh@suse.cz>
5280 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5282 * ipa.c (function_and_variable_visibility): Likewise.
5283 * ipa-profile.c (ipa_profile): Likewise.
5285 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5288 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5289 attribute, and don't inline the resolver in this case.
5291 2013-09-17 Teresa Johnson <tejohnson@google.com>
5293 * coverage.c (get_coverage_counts): Add missing newline.
5295 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5297 PR tree-optimization/58088
5298 * fold-const.c (mask_with_trailing_zeros): New function.
5299 (fold_binary_loc): Make sure we don't recurse infinitely
5300 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5301 Use mask_with_trailing_zeros where appropriate.
5303 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5305 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5306 of 'prev' var to get better distance estimation.
5308 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5310 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5311 * tree-inline.c (is_parameter_of): New predicate.
5312 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5313 a parameter has been remapped.
5314 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5315 (optimize_inline_calls): Initialize transform_parameter.
5316 (copy_gimple_seq_and_replace_locals): Likewise.
5317 (tree_function_versioning): Likewise.
5318 (maybe_inline_call_in_expr): Likewise.
5320 2013-09-17 Nick Clifton <nickc@redhat.com>
5322 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5323 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5325 (is_attr_func): New function.
5326 (msp430_is_interrupt_func): New function.
5327 (is_naked_func): New function.
5328 (is_reentrant_func): New function.
5329 (is_critical_func): New function.
5330 (msp430_start_function): Add annotations for function attributes.
5331 (msp430_attr): New function.
5332 (msp430_attribute_table): New.
5333 (msp430_function_section): New function.
5334 (TARGET_ASM_FUNCTION_SECTION): Define.
5335 (msp430_builtin): New enum.
5336 (msp430_init_builtins): New function.
5337 (msp430_builtin_devl): New function.
5338 (msp430_expand_builtin): New function.
5339 (TARGET_INIT_BUILTINS): Define.
5340 (TARGET_EXPAND_BUILTINS): Define.
5341 (TARGET_BUILTIN_DECL): Define.
5342 (msp430_expand_prologue): Add support for naked, interrupt,
5343 critical and reentrant functions.
5344 (msp430_expand_epilogue): Likewise.
5345 (msp430_print_operand): Handle 'O' character.
5346 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5348 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5349 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5350 (pushm): Use a 'n' rather than an 'i' constraint.
5351 (msp_return): Add generation of the interrupt return instruction.
5352 (disable_interrupts): New pattern.
5353 (enable_interrupts): New pattern.
5354 (push_intr_state): New pattern.
5355 (pop_intr_state): New pattern.
5356 (bic_SR): New pattern.
5357 (bis_SR): New pattern.
5358 * doc/extend.texi: Document MSP430 function attributes and builtin
5361 2013-09-17 Richard Biener <rguenther@suse.de>
5363 PR tree-optimization/58432
5364 * tree-loop-distribution.c (tree_loop_distribution): Also
5365 scan PHIs for outside loop uses and seed a partition from them.
5367 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5369 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5370 (restructure_reference): Call backtrace_base_for_ref.
5372 2013-09-17 Alan Modra <amodra@gmail.com>
5375 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5378 2013-09-16 DJ Delorie <dj@redhat.com>
5380 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5381 vregs location for RL78/G10.
5382 (rl78_expand_prologue): Avoid SEL on G10.
5383 (rl78_expand_epilogue): Likewise.
5384 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5385 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5386 __RL78_G10__ when appropriate.
5387 (ASM_SPEC): Pass -mg10 along to the assembler.
5388 * config/rl78/rl78.md (sel_rb): Disable for G10.
5389 * config/rl78/rl78.opt: Add -mg10 option.
5390 * config/rl78/t-rl78: Add -mg10 multilib.
5392 2013-09-16 Xinliang David Li <davidxl@google.com>
5394 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5395 * omp-low.c (omp_max_vf): Ditto.
5396 (expand_omp_simd): Ditto.
5397 * tree-vectorizer.c (vectorize_loops): Ditto.
5398 (gate_vect_slp): Ditto.
5399 (gate_increase_alignment): Ditto.
5400 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5401 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5402 (gate_tree_vectorize): Name change.
5403 (tree_vectorize): Ditto.
5404 (pass_vectorize::gate): Call new function.
5405 (pass_vectorize::execute): Ditto.
5406 * opts.c: O3 default setting change.
5407 (finish_options): Check new flag.
5408 * doc/invoke.texi: Document new flags.
5409 * common.opt: New flags.
5411 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5413 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5414 (Condition Code): Don't reference it.
5415 * doc/tm.texi: Regenerate.
5417 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5420 * lra-constraints.c (undo_optional_reloads): Consider all optional
5421 reload even if it did not get a hard reg.
5423 2013-09-16 Teresa Johnson <tejohnson@google.com>
5425 * dumpfile.c (dump_loc): Remove newline emission.
5426 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5427 emission to dump_printf_loc calls where missing.
5428 (vect_mark_for_runtime_alias_test): Ditto.
5429 (vect_analyze_data_ref_dependence): Ditto.
5430 (vect_analyze_data_ref_dependences): Ditto.
5431 (vect_slp_analyze_data_ref_dependence): Ditto.
5432 (vect_slp_analyze_data_ref_dependences): Ditto.
5433 (vect_compute_data_ref_alignment): Ditto.
5434 (vect_update_misalignment_for_peel): Ditto.
5435 (vect_verify_datarefs_alignment): Ditto.
5436 (vector_alignment_reachable_p): Ditto.
5437 (vect_get_data_access_cost): Ditto.
5438 (vect_enhance_data_refs_alignment): Ditto.
5439 (vect_find_same_alignment_drs): Ditto.
5440 (vect_analyze_data_refs_alignment): Ditto.
5441 (vect_analyze_group_access): Ditto.
5442 (vect_analyze_data_ref_access): Ditto.
5443 (vect_analyze_data_ref_accesses): Ditto.
5444 (vect_prune_runtime_alias_test_list): Ditto.
5445 (vect_analyze_data_refs): Ditto.
5446 (vect_create_addr_base_for_vector_ref): Ditto.
5447 (vect_create_data_ref_ptr): Ditto.
5448 (vect_grouped_store_supported): Ditto.
5449 (vect_grouped_load_supported): Ditto.
5450 * value-prof.c (check_counter): Ditto.
5451 (check_ic_target): Ditto.
5452 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5453 (vect_recog_widen_mult_pattern): Ditto.
5454 (vect_recog_widen_sum_pattern): Ditto.
5455 (vect_recog_over_widening_pattern): Ditto.
5456 (vect_recog_widen_shift_pattern): Ditto.
5457 (vect_recog_rotate_pattern): Ditto.
5458 (vect_recog_vector_vector_shift_pattern): Ditto.
5459 (vect_recog_divmod_pattern): Ditto.
5460 (vect_recog_mixed_size_cond_pattern): Ditto.
5461 (vect_recog_bool_pattern): Ditto.
5462 (vect_pattern_recog_1): Ditto.
5463 (vect_pattern_recog): Ditto.
5464 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5465 (vect_is_simple_iv_evolution): Ditto.
5466 (vect_analyze_scalar_cycles_1): Ditto.
5467 (vect_get_loop_niters): Ditto.
5468 (vect_analyze_loop_1): Ditto.
5469 (vect_analyze_loop_form): Ditto.
5470 (vect_analyze_loop_operations): Ditto.
5471 (vect_analyze_loop_2): Ditto.
5472 (vect_analyze_loop): Ditto.
5473 (report_vect_op): Ditto.
5474 (vect_is_slp_reduction): Ditto.
5475 (vect_is_simple_reduction_1): Ditto.
5476 (vect_get_known_peeling_cost): Ditto.
5477 (vect_estimate_min_profitable_iters): Ditto.
5478 (vect_model_reduction_cost): Ditto.
5479 (vect_model_induction_cost): Ditto.
5480 (get_initial_def_for_induction): Ditto.
5481 (vect_create_epilog_for_reduction): Ditto.
5482 (vectorizable_reduction): Ditto.
5483 (vectorizable_induction): Ditto.
5484 (vectorizable_live_operation): Ditto.
5485 (vect_loop_kill_debug_uses): Ditto.
5486 (vect_transform_loop): Ditto.
5487 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5488 (vect_stmt_relevant_p): Ditto.
5489 (process_use): Ditto.
5490 (vect_mark_stmts_to_be_vectorized): Ditto.
5491 (vect_model_simple_cost): Ditto.
5492 (vect_model_promotion_demotion_cost): Ditto.
5493 (vect_model_store_cost): Ditto.
5494 (vect_get_store_cost): Ditto.
5495 (vect_model_load_cost): Ditto.
5496 (vect_get_load_cost): Ditto.
5497 (vect_init_vector_1): Ditto.
5498 (vect_get_vec_def_for_operand): Ditto.
5499 (vect_finish_stmt_generation): Ditto.
5500 (vectorizable_call): Ditto.
5501 (vectorizable_conversion): Ditto.
5502 (vectorizable_assignment): Ditto.
5503 (vectorizable_shift): Ditto.
5504 (vectorizable_operation): Ditto.
5505 (vectorizable_store): Ditto.
5506 (vectorizable_load): Ditto.
5507 (vectorizable_condition): Ditto.
5508 (vect_analyze_stmt): Ditto.
5509 (vect_transform_stmt): Ditto.
5510 (vect_is_simple_use): Ditto.
5511 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5512 (vect_can_advance_ivs_p): Ditto.
5513 (vect_update_ivs_after_vectorizer): Ditto.
5514 (vect_do_peeling_for_loop_bound): Ditto.
5515 (vect_gen_niters_for_prolog_loop): Ditto.
5516 (vect_update_inits_of_drs): Ditto.
5517 (vect_create_cond_for_alias_checks): Ditto.
5518 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5519 (vect_build_slp_tree_1): Ditto.
5520 (vect_supported_load_permutation_p): Ditto.
5521 (vect_analyze_slp_instance): Ditto.
5522 (vect_analyze_slp): Ditto.
5523 (vect_make_slp_decision): Ditto.
5524 (vect_detect_hybrid_slp): Ditto.
5525 (vect_bb_vectorization_profitable_p): Ditto.
5526 (vect_slp_analyze_bb_1): Ditto.
5527 (vect_update_slp_costs_according_to_vf): Ditto.
5528 (vect_get_mask_element): Ditto.
5529 (vect_transform_slp_perm_load): Ditto.
5530 (vect_schedule_slp_instance): Ditto.
5531 (vect_schedule_slp): Ditto.
5532 (vect_slp_transform_bb): Ditto.
5533 * profile.c (read_profile_edge_counts): Ditto.
5534 (compute_branch_probabilities): Ditto.
5535 * coverage.c (get_coverage_counts): Ditto.
5537 2013-09-16 Diego Novillo <dnovillo@google.com>
5539 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5541 2013-09-16 Jan Hubicka <jh@suse.cz>
5543 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5544 abstract functions; for static functions check the presence of body.
5546 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5548 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5549 * config/aarch64/aarch64-simd.md
5550 (aarch64_mla_elt<mode>): New.
5551 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5552 (aarch64_mls_elt<mode>): Likewise.
5553 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5554 (aarch64_fma4_elt<mode>): Likewise.
5555 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5556 (aarch64_fma4_elt_to_128v2df): Likewise.
5557 (aarch64_fma4_elt_to_64df): Likewise.
5558 (fnma<mode>4): Likewise.
5559 (aarch64_fnma4_elt<mode>): Likewise.
5560 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5561 (aarch64_fnma4_elt_to_128v2df): Likewise.
5562 (aarch64_fnma4_elt_to_64df): Likewise.
5563 * config/aarch64/iterators.md (VDQSF): New.
5564 * config/aarch64/arm_neon.h
5565 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5566 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5568 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5570 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5571 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5572 (aarch64_mul3_elt_to_128df): Likewise.
5573 (aarch64_mul3_elt_to_64v2df): Likewise.
5574 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5576 (VMUL_CHANGE_NLANES) Likewise.
5579 * config/aarch64/arm_neon.h
5580 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5582 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5584 * config/aarch64/arm_neon.h
5585 (vcvtx_high_f32_f64): Fix parameters.
5587 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5588 Uros Bizjak <ubizjak@gmail.com>
5590 * config/alpha.c: Include tree-ssanames.h.
5592 2013-09-16 Richard Biener <rguenther@suse.de>
5594 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5595 (dot_rdg_1): Handle control_dd.
5596 (create_edge_for_control_dependence): New function.
5597 (create_rdg_edges): Add control dependences if asked for.
5598 (build_rdg): Likewise.
5599 (generate_loops_for_partition): If there are not necessary
5600 control stmts remove all their dependencies.
5601 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5602 (distribute_loop): Pass on control dependences.
5603 (tree_loop_distribution): Compute control dependences and remove
5604 restriction on number of loop nodes.
5606 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5608 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5611 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5613 * cse.c (try_const_anchors): Punt on CC modes.
5615 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5617 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5619 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5622 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5625 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5628 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5630 (darwin_objc1_section): Likewise.
5631 (darwin_file_end): Emit Image Info section when required.
5633 2013-09-14 Jan Hubicka <jh@suse.cz>
5635 * tree-into-ssa.c (gate_into_ssa): New.
5636 (pass_data_build_ssa): Use it.
5637 * cgraph.h (expand_thunk): Update prototype.
5638 * cgraphunit.c (analyze_function): Expand thunks early.
5639 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5640 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5641 set lowered flag; do not add new function.
5642 (assemble_thunks_and_aliases): Update.
5643 * tree-ssa.c (gate_init_datastructures): New gate.
5644 (pass_data_init_datastructures): Use it.
5646 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5649 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5650 xmm register set described in the psABI.
5652 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5654 * dwarf2out.c (should_emit_struct_debug): Add check
5655 for type_decl variable is not NULL.
5657 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5659 * config.gcc: Use new winnt-c.c target hooks
5660 * config/t-winnt: New file
5661 * config/winnt-c.c: New file
5662 * doc/tm.texi.in: Document new hook
5663 * doc/tm.texi: Regenerated
5665 2013-09-13 Jan Hubicka <jh@suse.cz>
5668 * ipa-inline.c (check_callers): New function.
5669 (check_caller_edge): Remove.
5670 (want_inline_function_to_all_callers_p): Also permit alises that are
5672 (inline_to_all_callers): Terminate the walk when devirtualization
5673 introduce new calls.
5675 2013-09-13 Jan Hubicka <jh@suse.cz>
5677 * ipa-inline-analysis.c (struct growth_data): Add node.
5678 (do_estimate_growth_1): Fix detection of recursion.
5680 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5682 PR tree-optimization/58392
5683 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5684 to avoid shadowing of outer loop variable. If
5685 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5686 replace_by_duplicate_decl simduid of loops that have it set and
5687 set dest_cfun->has_simduid_loops and/or
5688 dest_cfun->has_force_vect_loops.
5689 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5690 instead of maybe_lookup_decl.
5691 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5692 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5693 call. Copy over force_vect and copy and remap simduid. Set
5694 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5695 (copy_cfg_body): Remove blocks_to_copy argument. Use
5696 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5697 caller. Don't set cfun->has_simduid_loops and/or
5698 cfun->has_force_vect_loops here.
5699 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5701 (expand_call_inline, tree_function_versioning): Adjust copy_body
5704 2013-09-13 Martin Jambor <mjambor@suse.cz>
5707 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5708 the assert if the edge was a speculative one.
5710 2013-09-13 Richard Biener <rguenther@suse.de>
5712 * tree-data-ref.h (known_dependences_p): Move here ...
5713 * tree-loop-distribution.c (known_dependences_p): ... from here.
5714 (dump_rdg_component, debug_rdg_component): Remove.
5716 (generate_loops_for_partition): Use gimple_uid instead of
5717 relying on matching stmt visit order.
5718 (rdg_build_partitions): Take starting stmt vector.
5719 (ldist_gen): Merge into ...
5720 (distribute_loop): ... this function. Do not compute starting
5722 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5724 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5726 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5727 Set type attribute correctly. Set predicable_short_it attribute.
5728 (cmpsi_shiftsi): Remove %? from output template.
5730 2013-09-13 Richard Biener <rguenther@suse.de>
5732 * tree-loop-distribution.c (struct rdg_component,
5733 rdg_defs_used_in_other_loops_p, free_rdg_components,
5734 rdg_build_components): Remove.
5735 (stmts_from_loop): Do not record virtual PHIs.
5736 (generate_loops_for_partition): Skip virtual PHIs.
5737 (build_rdg_partition_for_component): Rename to ...
5738 (build_rdg_partition_for_vertex): ... this and adjust.
5739 (rdg_build_partitions): Take a vector of starting vertices
5740 instead of components. Remove unnecessary leftover handling.
5741 (ldist_gen): Do not build components or record other stores.
5742 (distribute_loop): Do not distribute loops containing stmts
5745 2013-09-13 Christian Bruel <christian.bruel@st.com>
5748 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5750 2013-09-13 Kai Tietz <ktietz@redhat.com>
5752 * config.gcc: Separate cases for mingw and cygwin targets,
5753 and add 64-bit cygwin target case.
5755 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5756 dll-export inline-functions.
5757 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5760 2013-09-13 Jeff Law <law@redhat.com>
5764 2013-09-06 Jeff Law <law@redhat.com>
5766 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5767 edge implied equivalences into successor phis.
5769 2013-09-12 DJ Delorie <dj@redhat.com>
5771 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5774 2013-09-12 Brooks Moses <bmoses@google.com>
5777 * Makefile.in: Do not install driver binaries in $(target)/bin.
5779 2013-09-12 DJ Delorie <dj@redhat.com>
5781 * config/rl78/rl78.opt (mrelax): New.
5782 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5783 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5785 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5787 (rl78_expand_epilogue): Likewise.
5789 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5792 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5793 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5794 pointer elimination with using remove_reg_equal_offset_note.
5796 2013-09-12 DJ Delorie <dj@redhat.com>
5798 * config/msp430/: New port.
5799 * config.gcc (msp430): Added.
5800 * doc/invoke.texi: Document MSP430 options.
5801 * doc/install.texi: Document msp430-elf
5802 * doc/md.texi: Document msp430-elf
5803 * doc/contrib.texi: Document msp430-elf
5805 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5808 2013-09-12 Martin Jambor <mjambor@suse.cz>
5811 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5812 reference descriptor is NULL.
5813 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5816 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5818 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5819 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5820 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5821 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5822 make_temp_ssa_name): move to tree-ssanames.h
5823 * tree-ssa-alias.h: Move prototype.
5824 * tree-ssa.h: Include tree-ssanames.h.
5825 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5826 * tree-ssanames.h: New. Move items from tree-flow*.h
5827 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5829 2013-09-12 Richard Biener <rguenther@suse.de>
5831 PR tree-optimization/58404
5832 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5833 propagate non-invariant addresses into dereferences wrapped
5834 in component references.
5836 2013-09-12 Richard Biener <rguenther@suse.de>
5838 PR tree-optimization/58402
5839 * passes.def: Move pass_late_warn_uninitialized later.
5841 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5843 * tree-ssa.h: New. Move content from tree-flow.h and
5845 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5846 Move prototypes belonging to tree-ssa.c.
5847 * tree-flow-inline.h (redirect_edge_var_map_def,
5848 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5850 * gimple.h: Adjust prototypes.
5851 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5853 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5854 * tree.h: Move prototype to tree-ssa.h.
5855 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5856 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5857 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5858 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5859 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5860 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5861 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5862 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5863 graphite-blocking.c, graphite-clast-to-gimple.c,
5864 graphite-dependences.c, graphite-interchange.c,
5865 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5866 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5867 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5868 ipa-reference.c, ipa-split.c, ipa-utils.c,
5869 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5870 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5871 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5872 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5873 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5874 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5875 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5876 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5877 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5878 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5879 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5880 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5881 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5882 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5883 value-prof.c, var-tracking.c,
5884 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5886 2013-09-12 Richard Biener <rguenther@suse.de>
5888 PR tree-optimization/58396
5889 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5890 (build_rdg): Take a loop-nest parameter, fix memleaks.
5891 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
5893 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
5895 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
5898 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
5900 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
5901 The FMA instruction names should have a 'v' prefix.
5903 2013-09-12 Richard Biener <rguenther@suse.de>
5905 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
5906 (dot_rdg): Use popen instead of system in optional code.
5907 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
5908 (already_processed_vertex_p): Adjust.
5909 (has_anti_or_output_dependence, predecessor_has_mem_write,
5910 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
5911 rdg_flag_uses): Remove.
5912 (rdg_flag_vertex): Simplify.
5913 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
5915 (build_rdg_partition_for_component): Process the first vertex
5916 of a component only.
5917 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
5919 2013-09-12 Alan Modra <amodra@gmail.com>
5921 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
5923 2013-09-11 DJ Delorie <dj@redhat.com>
5924 Nick Clifton <nickc@redhat.com>
5926 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
5927 (rl78_stack_based_mem): New.
5928 * config/rl78/constraints.md (Iv08): New.
5937 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
5939 (movsi): Change from expand to insn-and-split.
5940 (ashrsi3): Clobber AX.
5944 * config/rl78/rl78.md (CC_REG): Fix.
5945 (addsi3): Allow memory and immediate operands.
5946 (addsi3_internal): Split into...
5947 (addsi3_internal_virt): ...new, and ...
5948 (addsi3_internal_real): ...new.
5950 (subsi3_internal_virt): New.
5951 (subsi3_internal_real): New.
5952 (mulsi3): Add memory operand.
5953 (mulsi3_rl78): Likewise.
5954 (mulsi3_g13): Likewise.
5955 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
5956 (cbranchqi4_real): Add more constraint options.
5957 (cbranchhi4_real): Expand pattern.
5958 (cbranchhi4_real_signed): New.
5959 (cbranchhi4_real_inverted): New.
5960 (cbranchsi4_real_lt): New.
5961 (cbranchsi4_real_ge): New.
5962 (cbranchsi4_real_signed): New.
5963 (cbranchsi4_real): New.
5965 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
5967 (lshrsi3_virt): Likewise.
5968 (ashlsi3_virt): Likewise.
5969 (cbranchqi4_virt_signed): New.
5970 (cbranchhi4_virt_signed): New.
5971 (cbranchsi4_virt): New.
5972 * config/rl78/rl78.c: Whitespace fixes throughout.
5973 (move_elim_pass): New.
5974 (pass_data_rl78_move_elim): New.
5975 (pass_rl78_move_elim): New.
5976 (make_pass_rl78_move_elim): New.
5977 (rl78_devirt_info): Run devirt earlier.
5978 (rl78_move_elim_info): New.
5979 (rl78_asm_file_start): Register it.
5980 (rl78_split_movsi): New.
5981 (rl78_as_legitimate_address): Allow virtual base registers when
5983 (rl78_addr_space_convert): Remove spurious debug stuff.
5984 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5985 (rl78_print_operand): More cases for not printing '#'.
5986 (rl78_expand_compare): Remove most of the logic.
5987 (content_memory): New.
5988 (clear_content_memory): New.
5989 (get_content_index): New.
5990 (get_content_name): New.
5991 (display_content_memory): New.
5992 (update_content): New.
5993 (record_content): New.
5994 (already_contains): New.
5995 (insn_ok_now): Re-recog insns with virtual registers.
5996 (add_postponed_content_update): New.
5997 (process_postponed_content_update): New.
5998 (gen_and_emit_move): New.
5999 (transcode_memory_rtx): Record new location content.
6000 Use gen_and_emit_move.
6001 (force_into_acc): New.
6002 (move_to_acc): Use gen_and_emit_move.
6003 (move_from_acc): Likewise.
6004 (move_acc_to_reg): Likewise.
6005 (move_to_x): Likewise.
6006 (move_to_hl): Likewise.
6007 (move_to_de): Likewise.
6008 (rl78_alloc_physical_registers_op1): Record location content.
6009 (has_constraint): New.
6010 (rl78_alloc_physical_registers_op2): Record location content.
6012 (rl78_alloc_physical_registers_ro1): Likewise.
6013 (rl78_alloc_physical_registers_cmp): Likewise.
6014 (rl78_alloc_physical_registers_umul): Likewise.
6015 (rl78_alloc_address_registers_macax): New.
6016 (rl78_alloc_physical_registers): Initialize and set location
6017 content memory as needed.
6018 (rl78_reorg): Make sure split2 is called.
6019 (rl78_rtx_costs): New.
6021 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6023 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6024 for (not (neg ...)) and (neg (not ...)) cases.
6026 2013-09-11 Richard Biener <rguenther@suse.de>
6029 * passes.def: Split critical edges before late uninit warning passes.
6030 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6032 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6034 PR tree-optimization/58385
6035 * fold-const.c (build_range_check): If both low and high are NULL,
6036 use omit_one_operand_loc to preserve exp side-effects.
6038 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6040 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6042 2013-09-11 Richard Biener <rguenther@suse.de>
6044 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6045 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6046 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6047 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6048 stmts_from_loop, known_dependences_p, build_empty_rdg,
6049 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6050 * tree-loop-distribution.c: ... here.
6051 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6052 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6053 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6054 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6055 * tree-loop-distribution.c: ... here.
6056 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6057 (struct partition_s): Add loops member.
6058 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6059 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6060 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6062 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6063 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6064 Sergey Lega <sergey.s.lega@intel.com>
6065 Anna Tikhonova <anna.tikhonova@intel.com>
6066 Ilya Tocar <ilya.tocar@intel.com>
6067 Andrey Turetskiy <andrey.turetskiy@intel.com>
6068 Ilya Verbin <ilya.verbin@intel.com>
6069 Kirill Yukhin <kirill.yukhin@intel.com>
6070 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6072 * config/i386/constraints.md (k): New.
6074 * config/i386/i386.c (const regclass_map): Add new mask registers.
6075 (dbx_register_map): Ditto.
6076 (dbx64_register_map): Ditto.
6077 (svr4_dbx_register_map): Ditto.
6078 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6080 (ix86_preferred_reload_class): Disable constants for mask registers.
6081 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6082 (ix86_hard_regno_mode_ok): Support new mask registers.
6083 (x86_order_regs_for_local_alloc): Ditto.
6084 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6085 (FIXED_REGISTERS): Add new mask registers.
6086 (CALL_USED_REGISTERS): Ditto.
6087 (REG_ALLOC_ORDER): Ditto.
6088 (VALID_MASK_REG_MODE): New.
6089 (FIRST_MASK_REG): Ditto.
6090 (LAST_MASK_REG): Ditto.
6091 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6092 (MAYBE_MASK_CLASS_P): New.
6093 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6094 (REG_CLASS_CONTENTS): Ditto.
6095 (MASK_REGNO_P): New.
6096 (ANY_MASK_REG_P): Ditto.
6097 (HI_REGISTER_NAMES): Add new mask registers.
6098 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6099 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6101 (attribute "type"): Add mskmov, msklog.
6102 (attribute "length_immediate"): Support them.
6103 (attribute "memory"): Ditto.
6104 (attribute "prefix_0f"): Ditto.
6105 (*movhi_internal): Support new mask registers.
6106 (*movqi_internal): Ditto.
6107 (define_split): Split out clobber pattern is a logic
6108 insn on mask registers.
6109 (*k<logic><mode>): New.
6110 (*andhi_1): Extend to support mask regs.
6111 (*andqi_1): Extend to support mask regs.
6113 (define_split): Split and-not to and and not if operands
6115 (*<code><mode>_1): Separate HI mode to new pattern...
6116 (*<code>hi_1): This.
6117 (*<code>qi_1): Extend to support mask regs.
6119 (kortestzhi): Ditto.
6120 (kortestchi): Ditto.
6122 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6123 (*one_cmplhi2_1): ...Here, now with mask registers support.
6124 (*one_cmplqi2_1): Support new mask registers.
6125 (HI/QImode arithmetics splitter): Don't split if mask registers
6127 (HI/QImode not splitter): Ditto.
6128 * config/i386/predicated.md (mask_reg_operand): New.
6129 (general_reg_operand): Ditto.
6131 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6133 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6134 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6136 2013-09-10 Jeff Law <law@redhat.com>
6138 PR tree-optimization/58380
6139 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6140 of threading through a buried loop header.
6142 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6143 return value for single successor case.
6145 2013-09-10 Jan Hubicka <jh@suse.cz>
6147 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6149 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6152 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6153 support conditional execution.
6154 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6156 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6158 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6160 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6161 Check destination too to check move insn.
6162 (undo_optional_reloads): Add check that the original peudo did not
6163 changed its allocation and the optional reload was inherited on last
6164 inheritance pass. Break loop after deciding to keep optional reload.
6165 (lra_undo_inheritance): Add check that inherited pseudo still in
6168 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6170 * config/aarch64/aarch64.md (generic_sched): New.
6171 * config/aarch64/aarch64-generic.md (load): Make conditional
6172 on generic_sched attribute.
6173 (nonload): Likewise.
6175 2013-09-10 Jan Hubicka <jh@suse.cz>
6177 * lto-cgraph.c: Include ipa-utils.h.
6178 (compute_ltrans_boundary): Also add possible targets into the boundary.
6180 2013-09-10 Jan Hubicka <jh@suse.cz>
6182 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6183 VAR_DECL of vtable rather than full expression.
6185 2013-09-10 Jan Hubicka <jh@suse.cz>
6186 Paolo Carlini <paolo.carlini@oracle.com>
6188 * cgraphunit.c (analyze_functions): Save input_location, set it
6189 to UNKNOWN_LOCATION and restore it at the end.
6191 2013-09-10 Martin Jambor <mjambor@suse.cz>
6193 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6194 represented by a thunk.
6196 2013-09-10 Jeff Law <law@redhat.com>
6198 PR tree-optimization/58343
6199 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6200 jump threading requests through loop headers buried in the middle
6201 of a jump threading path.
6203 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6204 in return value/type.
6206 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6208 PR rtl-optimization/58365
6209 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6210 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6213 2013-09-10 Richard Biener <rguenther@suse.de>
6215 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6216 * tree-data-ref.c (create_rdg_vertices): Collect all data
6217 references, signal failure to the caller, use data-ref API.
6218 (build_rdg): Compute data references only once. Maintain lifetime
6219 of data references and data dependences from within RDG.
6220 (free_rdg): Free dependence relations.
6221 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6222 inventing extra dependences.
6223 (distribute_loop): Update for RDG API changes.
6225 2013-09-10 Kai Tietz <ktietz@redhat.com>
6227 * doc/invoke.texi (fms-extensions): Document changed
6228 behavior for ms-abi targets.
6229 * config/i386/i386.c (ix86_option_override_internal):
6230 Set default value of option -fms-extension for ms-abi targets.
6232 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6234 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6236 2013-09-10 Alan Modra <amodra@gmail.com>
6239 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6241 2013-09-10 Alan Modra <amodra@gmail.com>
6243 * config/rs6000/predicates.md (add_cint_operand): New.
6244 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6245 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6246 using add_cint_operand.
6247 (largetoc_high_plus_aix): Likewise.
6249 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6251 PR tree-optimization/58364
6252 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6253 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6254 the current range can't be an unconditional true or false.
6256 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6258 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6260 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6262 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6264 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6267 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6269 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6271 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6272 NO_REGS for immediate that can't be moved directly into FP_REGS.
6274 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6276 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6277 comparison with negated operand.
6278 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6281 2013-09-09 Richard Biener <rguenther@suse.de>
6284 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6285 of a subloop record all its block as affecting loop-closed SSA form.
6287 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6289 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6290 of an rtx/bitpos pair.
6291 (store_fixed_bit_field): Update accordingly.
6293 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6295 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6297 * builtins.c (expand_errno_check): Likewise.
6298 * dwarf2cfi.c (init_return_column_size): Likewise.
6299 * except.c (sjlj_mark_call_sites): Likewise.
6300 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6301 * lra-constraints.c (emit_inc): Likewise.
6302 * ree.c (combine_set_extension): Likewise.
6303 * regmove.c (fixup_match_2): Likewise.
6304 * reload1.c (inc_for_reload): Likewise.
6306 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6308 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6309 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6310 Use gen_int_mode with the mode of the associated simplify_* call.
6311 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6312 * expmed.c (expand_shift_1): Likewise.
6313 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6314 * loop-iv.c (iv_number_of_iterations): Likewise.
6315 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6316 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6318 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6320 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6321 of the associated expand_* call.
6322 (asan_emit_stack_protection): Likewise.
6323 * builtins.c (round_trampoline_addr): Likewise.
6324 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6325 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6326 (emit_store_flag): Likewise.
6327 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6328 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6330 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6331 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6332 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6333 * modulo-sched.c (generate_prolog_epilog): Likewise.
6334 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6335 (expand_ctz, expand_ffs, expand_unop): Likewise.
6337 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6339 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6340 of the associated gen_rtx_* call.
6341 * caller-save.c (init_caller_save): Likewise.
6342 * combine.c (find_split_point, make_extraction): Likewise.
6343 (make_compound_operation): Likewise.
6344 * dwarf2out.c (mem_loc_descriptor): Likewise.
6345 * explow.c (plus_constant, probe_stack_range): Likewise.
6346 * expmed.c (expand_mult_const): Likewise.
6347 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6348 * reload1.c (init_reload): Likewise.
6349 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6350 * var-tracking.c (adjust_mems): Likewise.
6351 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6352 rather than gen_rtx_fmt_ee.
6354 2013-09-09 Jan Hubicka <jh@suse.cz>
6357 * value-prof.c (gimple_ic): Copy also abnormal edges.
6359 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6361 * asan.c (asan_shadow_cst): Use gen_int_mode.
6363 2013-09-08 Jan Hubicka <jh@suse.cz>
6365 * ipa-profile.c: Add toplevel comment.
6366 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6367 (contains_hot_call_p): New function.
6368 (ipa_propagate_frequency): Set frequencies based on counts when
6370 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6371 profile; do not tamper with profile after inlining if it is read.
6373 2013-09-08 Jan Hubicka <jh@suse.cz>
6375 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6378 2013-09-08 Jan Hubicka <jh@suse.cz>
6380 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6383 2013-09-08 Jeff Law <law@redhat.com>
6386 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6389 2013-09-08 Andi Kleen <ak@linux.intel.com>
6391 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6393 2013-09-08 Jan Hubicka <jh@suse.cz>
6395 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6397 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6399 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6400 for non-debug insns.
6401 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6402 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6403 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6406 2013-09-08 Jan Hubicka <jh@suse.cz>
6408 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6409 targets and devirtualize to BUILT_IN_UNREACHABLE.
6410 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6411 * ipa.c (walk_polymorphic_call_targets): New function.
6412 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6413 functions; use the new timevar.
6414 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6415 was removed from the program.
6416 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6417 not consider it in the walk when its vtable is dead.
6418 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6420 (devirt_variable_node_removal_hook): New function.
6421 (possible_polymorphic_call_targets): Also register
6422 devirt_variable_node_removal_hook.
6423 (ipa_devirt): Do not do non-speculative devirtualization.
6424 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6426 2013-09-08 Jan Hubicka <jh@suse.cz>
6428 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6429 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6430 varpool_remove_variable_insertion_hook): Declare.
6431 * varpool.c (varpool_node_hook_list): New structure.
6432 (first_varpool_node_removal_hook,
6433 first_varpool_variable_insertion_hook): New variables.
6434 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6435 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6436 varpool_remove_variable_insertion_hook,
6437 varpool_call_variable_insertion_hooks): New functions.
6438 (varpool_remove_node): Use it.
6440 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6443 * diagnostic.c (diagnostic_build_prefix): When s.file is
6444 "<built-in>" don't output line and column numbers.
6446 2013-09-06 Jan Hubicka <jh@suse.cz>
6448 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6449 * lto-streamer-out.c: Stream thunks, too.
6450 * lto-streamer-in.c (input_function): Pop cfun here
6451 (lto_read_body): Instead of here.
6453 2013-09-06 Caroline Tice <cmtice@google.com>
6455 * doc/install.texi: Add documentation for the --enable-vtable-verify
6456 and the --disable-libvtv configure options.
6458 2013-09-06 Jeff Law <law@redhat.com>
6460 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6461 edge implied equivalences into successor phis.
6463 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6465 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6467 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6469 * resource.c (mark_target_live_regs): Compute resources taking
6470 into account if a call is predicated or not.
6472 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6474 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6475 the compiler in the function names.
6477 2013-09-06 Jan Hubicka <jh@suse.cz>
6480 * ipa-inline.c (has_caller_p): New function.
6481 (want_inline_function_to_all_callers_p): Use it.
6482 (sum_callers, inline_to_all_callers): Break out from ...
6483 (ipa_inline): ... here.
6485 2013-09-06 Jan Hubicka <jh@suse.cz>
6487 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6488 only when AVX is enabled.
6490 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6492 * config/aarch64/aarch64.md
6493 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6494 is fpsimd_<load/store>2.
6495 (load_pair<mode>): Likewise.
6496 (store_pair<mode>): Likewise.
6498 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6500 * config/arm/types.md (type): Add "mrs" type.
6501 * config/aarch64/aarch64.md
6502 (aarch64_load_tp_hard): Make type "mrs".
6504 (load_tp_hard): Make type "mrs".
6505 * config/arm/cortex-a15.md: Update with new attributes.
6506 * config/arm/cortex-a5.md: Update with new attributes.
6507 * config/arm/cortex-a53.md: Update with new attributes.
6508 * config/arm/cortex-a7.md: Update with new attributes.
6509 * config/arm/cortex-a8.md: Update with new attributes.
6510 * config/arm/cortex-a9.md: Update with new attributes.
6511 * config/arm/cortex-m4.md: Update with new attributes.
6512 * config/arm/cortex-r4.md: Update with new attributes.
6513 * config/arm/fa526.md: Update with new attributes.
6514 * config/arm/fa606te.md: Update with new attributes.
6515 * config/arm/fa626te.md: Update with new attributes.
6516 * config/arm/fa726te.md: Update with new attributes.
6518 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6520 * config/aarch64/aarch64.md
6521 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6522 (*movtf_aarch64): Likewise.
6524 (thumb1_movdi_insn): Use "multiple" for type where more than one
6525 instruction is used for a move.
6526 (*arm32_movhf): Likewise.
6527 (*thumb_movdf_insn): Likewise.
6529 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6531 * config/arm/types.md (type): Rename fcpys to fmov.
6533 (*arm_movsi_vfp): Rename type fcpys as fmov.
6534 (*thumb2_movsi_vfp): Likewise
6535 (*movhf_vfp_neon): Likewise
6536 (*movhf_vfp): Likewise
6537 (*movsf_vfp): Likewise
6538 (*thumb2_movsf_vfp): Likewise
6539 (*movsfcc_vfp): Likewise
6540 (*thumb2_movsfcc_vfp): Likewise
6541 * config/aarch64/aarch64-simd.md
6542 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6543 * config/aarch64/aarch64.md
6544 (*movsi_aarch64): Replace type mov_reg with fmovs.
6545 (*movdi_aarch64): Likewise
6546 (*movsf_aarch64): Likewise
6547 (*movdf_aarch64): Likewise
6549 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6550 * config/arm/iwmmxt.md
6551 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6552 * config/arm/arm1020e.md: Update with new attributes.
6553 * config/arm/cortex-a15-neon.md: Update with new attributes.
6554 * config/arm/cortex-a5.md: Update with new attributes.
6555 * config/arm/cortex-a53.md: Update with new attributes.
6556 * config/arm/cortex-a7.md: Update with new attributes.
6557 * config/arm/cortex-a8-neon.md: Update with new attributes.
6558 * config/arm/cortex-a9.md: Update with new attributes.
6559 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6560 * config/arm/cortex-r4f.md: Update with new attributes.
6561 * config/arm/marvell-pj4.md: Update with new attributes.
6562 * config/arm/vfp11.md: Update with new attributes.
6564 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6566 * config/aarch64/aarch64.md
6567 (*madd<mode>): Fix type attribute.
6568 (*maddsi_uxtw): Likewise.
6569 (*msub<mode>): Likewise.
6570 (*msubsi_uxtw): Likewise.
6571 (<su_optab>maddsidi4): Likewise.
6572 (<su_optab>msubsidi4): Likewise.
6574 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6576 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6577 * config/arm/arm.md (core_cycles): Remove fdiv.
6578 * config/arm/vfp.md:
6579 (*sqrtsf2_vfp): Update for attribute changes.
6580 (*sqrtdf2_vfp): Likewise.
6581 * config/aarch64/aarch64.md:
6582 (sqrt<mode>2): Update for attribute changes.
6583 * config/arm/arm1020e.md: Update with new attributes.
6584 * config/arm/cortex-a15-neon.md: Update with new attributes.
6585 * config/arm/cortex-a5.md: Update with new attributes.
6586 * config/arm/cortex-a53.md: Update with new attributes.
6587 * config/arm/cortex-a7.md: Update with new attributes.
6588 * config/arm/cortex-a8-neon.md: Update with new attributes.
6589 * config/arm/cortex-a9.md: Update with new attributes.
6590 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6591 * config/arm/cortex-r4f.md: Update with new attributes.
6592 * config/arm/marvell-pj4.md: Update with new attributes.
6593 * config/arm/vfp11.md: Update with new attributes.
6595 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6597 * config/arm/types.md
6598 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6599 * config/aarch64/aarch64.md
6600 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6602 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6603 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6604 (float<GPI:mode><GPF:mode>2): Likewise.
6606 (*truncsisf2_vfp): Update with new attributes.
6607 (*truncsidf2_vfp): Likewise.
6608 (fixuns_truncsfsi2): Likewise.
6609 (fixuns_truncdfsi2): Likewise.
6610 (*floatsisf2_vfp): Likewise.
6611 (*floatsidf2_vfp): Likewise.
6612 (floatunssisf2): Likewise.
6613 (floatunssidf2): Likewise.
6614 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6615 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6616 * config/arm/arm1020e.md: Update with new attributes.
6617 * config/arm/cortex-a15-neon.md: Update with new attributes.
6618 * config/arm/cortex-a5.md: Update with new attributes.
6619 * config/arm/cortex-a53.md: Update with new attributes.
6620 * config/arm/cortex-a7.md: Update with new attributes.
6621 * config/arm/cortex-a8-neon.md: Update with new attributes.
6622 * config/arm/cortex-a9.md: Update with new attributes.
6623 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6624 * config/arm/cortex-r4f.md: Update with new attributes.
6625 * config/arm/marvell-pj4.md: Update with new attributes.
6626 * config/arm/vfp11.md: Update with new attributes.
6628 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6630 * config/aarch64/arm_neon.h
6631 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6632 (vqtbx<1,2,3,4><q>_s8): Likewise.
6634 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6636 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6637 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6638 (add<mode>3): Add type attribute.
6639 (add<mode>3): Likewise.
6640 (usadd<mode>3): Likewise.
6641 (ssadd<mode>3): Likewise.
6642 (sub<mode>3): Likewise.
6643 (sub<mode>3): Likewise.
6644 (ussub<mode>3): Likewise.
6645 (sssub<mode>3): Likewise.
6646 (ssmulsa3): Likewise.
6647 (usmulusa3): Likewise.
6648 (arm_usatsihi): Likewise.
6650 (*movdi_vfp): Add types for all instructions.
6651 (*movdi_vfp_cortexa8): Likewise.
6652 (*movhf_vfp_neon): Likewise.
6653 (*movhf_vfp): Likewise.
6654 (*movdf_vfp): Likewise.
6655 (*thumb2_movdf_vfp): Likewise.
6656 (*thumb2_movdfcc_vfp): Likewise.
6657 * config/arm/arm.md: Add type attribute to all insn patterns.
6658 (*thumb1_adddi3): Add type attribute.
6659 (*arm_adddi3): Likewise.
6660 (*adddi_sesidi_di): Likewise.
6661 (*adddi_zesidi_di): Likewise.
6662 (*thumb1_addsi3): Likewise.
6663 (addsi3_compare0): Likewise.
6664 (*addsi3_compare0_scratch): Likewise.
6665 (*compare_negsi_si): Likewise.
6666 (cmpsi2_addneg): Likewise.
6667 (*addsi3_carryin_<optab>): Likewise.
6668 (*addsi3_carryin_alt2_<optab>): Likewise.
6669 (*addsi3_carryin_clobercc_<optab>): Likewise.
6670 (*subsi3_carryin): Likewise.
6671 (*subsi3_carryin_const): Likewise.
6672 (*subsi3_carryin_compare): Likewise.
6673 (*subsi3_carryin_compare_const): Likewise.
6674 (*arm_subdi3): Likewise.
6675 (*thumb_subdi3): Likewise.
6676 (*subdi_di_zesidi): Likewise.
6677 (*subdi_di_sesidi): Likewise.
6678 (*subdi_zesidi_di): Likewise.
6679 (*subdi_sesidi_di): Likewise.
6680 (*subdi_zesidi_ze): Likewise.
6681 (thumb1_subsi3_insn): Likewise.
6682 (*arm_subsi3_insn): Likewise.
6683 (*anddi3_insn): Likewise.
6684 (*anddi_zesidi_di): Likewise.
6685 (*anddi_sesdi_di): Likewise.
6686 (*ne_zeroextracts): Likewise.
6687 (*ne_zeroextracts): Likewise.
6688 (*ite_ne_zeroextr): Likewise.
6689 (*ite_ne_zeroextr): Likewise.
6690 (*anddi_notdi_di): Likewise.
6691 (*anddi_notzesidi): Likewise.
6692 (*anddi_notsesidi): Likewise.
6693 (andsi_notsi_si): Likewise.
6694 (thumb1_bicsi3): Likewise.
6695 (*iordi3_insn): Likewise.
6696 (*iordi_zesidi_di): Likewise.
6697 (*iordi_sesidi_di): Likewise.
6698 (*thumb1_iorsi3_insn): Likewise.
6699 (*xordi3_insn): Likewise.
6700 (*xordi_zesidi_di): Likewise.
6701 (*xordi_sesidi_di): Likewise.
6702 (*arm_xorsi3): Likewise.
6703 (*andsi_iorsi3_no): Likewise.
6704 (*smax_0): Likewise.
6705 (*smax_m1): Likewise.
6706 (*arm_smax_insn): Likewise.
6707 (*smin_0): Likewise.
6708 (*arm_smin_insn): Likewise.
6709 (*arm_umaxsi3): Likewise.
6710 (*arm_uminsi3): Likewise.
6711 (*minmax_arithsi): Likewise.
6712 (*minmax_arithsi_): Likewise.
6713 (*satsi_<SAT:code>): Likewise.
6714 (arm_ashldi3_1bit): Likewise.
6715 (arm_ashrdi3_1bit): Likewise.
6716 (arm_lshrdi3_1bit): Likewise.
6717 (*arm_negdi2): Likewise.
6718 (*thumb1_negdi2): Likewise.
6719 (*arm_negsi2): Likewise.
6720 (*thumb1_negsi2): Likewise.
6721 (*negdi_extendsid): Likewise.
6722 (*negdi_zero_extend): Likewise.
6723 (*arm_abssi2): Likewise.
6724 (*thumb1_abssi2): Likewise.
6725 (*arm_neg_abssi2): Likewise.
6726 (*thumb1_neg_abss): Likewise.
6727 (one_cmpldi2): Likewise.
6728 (extend<mode>di2): Likewise.
6729 (*compareqi_eq0): Likewise.
6730 (*arm_extendhisi2addsi): Likewise.
6731 (*arm_movdi): Likewise.
6732 (*thumb1_movdi_insn): Likewise.
6733 (*arm_movt): Likewise.
6734 (*thumb1_movsi_insn): Likewise.
6735 (pic_add_dot_plus_four): Likewise.
6736 (pic_add_dot_plus_eight): Likewise.
6737 (tls_load_dot_plus_eight): Likewise.
6738 (*thumb1_movhi_insn): Likewise.
6739 (*thumb1_movsf_insn): Likewise.
6740 (*movdf_soft_insn): Likewise.
6741 (*thumb_movdf_insn): Likewise.
6742 (cbranchsi4_insn): Likewise.
6743 (cbranchsi4_scratch): Likewise.
6744 (*negated_cbranchsi4): Likewise.
6745 (*tbit_cbranch): Likewise.
6746 (*tlobits_cbranch): Likewise.
6747 (*tstsi3_cbranch): Likewise.
6748 (*cbranchne_decr1): Likewise.
6749 (*addsi3_cbranch): Likewise.
6750 (*addsi3_cbranch_scratch): Likewise.
6751 (*arm_cmpdi_insn): Likewise.
6752 (*arm_cmpdi_unsig): Likewise.
6753 (*arm_cmpdi_zero): Likewise.
6754 (*thumb_cmpdi_zero): Likewise.
6755 (*deleted_compare): Likewise.
6756 (*mov_scc): Likewise.
6757 (*mov_negscc): Likewise.
6758 (*mov_notscc): Likewise.
6759 (*cstoresi_eq0_thumb1_insn): Likewise.
6760 (cstoresi_nltu_thumb1): Likewise.
6761 (cstoresi_ltu_thu): Likewise.
6762 (thumb1_addsi3_addgeu): Likewise.
6763 (*arm_jump): Likewise.
6764 (*thumb_jump): Likewise.
6765 (*check_arch2): Likewise.
6766 (arm_casesi_internal): Likewise.
6767 (thumb1_casesi_dispatch): Likewise.
6768 (*arm_indirect_jump): Likewise.
6769 (*thumb1_indirect_jump): Likewise.
6771 (*and_scc): Likewise.
6772 (*ior_scc): Likewise.
6773 (*compare_scc): Likewise.
6774 (*cond_move): Likewise.
6775 (*cond_arith): Likewise.
6776 (*cond_sub): Likewise.
6777 (*cmp_ite0): Likewise.
6778 (*cmp_ite1): Likewise.
6779 (*cmp_and): Likewise.
6780 (*cmp_ior): Likewise.
6781 (*ior_scc_scc): Likewise.
6782 (*ior_scc_scc_cmp): Likewise.
6783 (*and_scc_scc): Likewise.
6784 (*and_scc_scc_cmp): Likewise.
6785 (*and_scc_scc_nod): Likewise.
6786 (*negscc): Likewise.
6787 (movcond_addsi): Likewise.
6788 (movcond): Likewise.
6789 (*ifcompare_plus_move): Likewise.
6790 (*if_plus_move): Likewise.
6791 (*ifcompare_move_plus): Likewise.
6792 (*if_move_plus): Likewise.
6793 (*ifcompare_arith_arith): Likewise.
6794 (*if_arith_arith): Likewise.
6795 (*ifcompare_arith_move): Likewise.
6796 (*if_arith_move): Likewise.
6797 (*ifcompare_move_arith): Likewise.
6798 (*if_move_arith): Likewise.
6799 (*ifcompare_move_not): Likewise.
6800 (*if_move_not): Likewise.
6801 (*ifcompare_not_move): Likewise.
6802 (*if_not_move): Likewise.
6803 (*ifcompare_shift_move): Likewise.
6804 (*if_shift_move): Likewise.
6805 (*ifcompare_move_shift): Likewise.
6806 (*if_move_shift): Likewise.
6807 (*ifcompare_shift_shift): Likewise.
6808 (*ifcompare_not_arith): Likewise.
6809 (*ifcompare_arith_not): Likewise.
6810 (*if_arith_not): Likewise.
6811 (*ifcompare_neg_move): Likewise.
6812 (*if_neg_move): Likewise.
6813 (*ifcompare_move_neg): Likewise.
6814 (*if_move_neg): Likewise.
6815 (prologue_thumb1_interwork): Likewise.
6816 (*cond_move_not): Likewise.
6817 (*sign_extract_onebit): Likewise.
6818 (*not_signextract_onebit): Likewise.
6819 (stack_tie): Likewise.
6820 (align_4): Likewise.
6821 (align_8): Likewise.
6822 (consttable_end): Likewise.
6823 (consttable_1): Likewise.
6824 (consttable_2): Likewise.
6825 (consttable_4): Likewise.
6826 (consttable_8): Likewise.
6827 (consttable_16): Likewise.
6828 (*thumb1_tablejump): Likewise.
6829 (prefetch): Likewise.
6830 (force_register_use): Likewise.
6831 (thumb_eh_return): Likewise.
6832 (load_tp_hard): Likewise.
6833 (load_tp_soft): Likewise.
6834 (tlscall): Likewise.
6835 (*arm_movtas_ze): Likewise.
6836 (*arm_rev): Likewise.
6837 (*arm_revsh): Likewise.
6838 (*arm_rev16): Likewise.
6839 * config/arm/thumb2.md
6840 (*thumb2_smaxsi3): Likewise.
6841 (*thumb2_sminsi3): Likewise.
6842 (*thumb32_umaxsi3): Likewise.
6843 (*thumb2_uminsi3): Likewise.
6844 (*thumb2_negdi2): Likewise.
6845 (*thumb2_abssi2): Likewise.
6846 (*thumb2_neg_abss): Likewise.
6847 (*thumb2_movsi_insn): Likewise.
6848 (tls_load_dot_plus_four): Likewise.
6849 (*thumb2_movhi_insn): Likewise.
6850 (*thumb2_mov_scc): Likewise.
6851 (*thumb2_mov_negs): Likewise.
6852 (*thumb2_mov_negs): Likewise.
6853 (*thumb2_mov_nots): Likewise.
6854 (*thumb2_mov_nots): Likewise.
6855 (*thumb2_movsicc_): Likewise.
6856 (*thumb2_movsfcc_soft_insn): Likewise.
6857 (*thumb2_indirect_jump): Likewise.
6858 (*thumb2_and_scc): Likewise.
6859 (*thumb2_ior_scc): Likewise.
6860 (*thumb2_ior_scc_strict_it): Likewise.
6861 (*thumb2_cond_move): Likewise.
6862 (*thumb2_cond_arith): Likewise.
6863 (*thumb2_cond_ari): Likewise.
6864 (*thumb2_cond_sub): Likewise.
6865 (*thumb2_negscc): Likewise.
6866 (*thumb2_movcond): Likewise.
6867 (thumb2_casesi_internal): Likewise.
6868 (thumb2_casesi_internal_pic): Likewise.
6869 (*thumb2_alusi3_short): Likewise.
6870 (*thumb2_mov<mode>_shortim): Likewise.
6871 (*thumb2_addsi_short): Likewise.
6872 (*thumb2_subsi_short): Likewise.
6873 (thumb2_addsi3_compare0): Likewise.
6874 (*thumb2_cbz): Likewise.
6875 (*thumb2_cbnz): Likewise.
6876 (*thumb2_one_cmplsi2_short): Likewise.
6877 (*thumb2_negsi2_short): Likewise.
6878 (*orsi_notsi_si): Likewise.
6879 * config/arm/arm1020e.md: Update with new attributes.
6880 * config/arm/arm1026ejs.md: Update with new attributes.
6881 * config/arm/arm1136jfs.md: Update with new attributes.
6882 * config/arm/arm926ejs.md: Update with new attributes.
6883 * config/arm/cortex-a15.md: Update with new attributes.
6884 * config/arm/cortex-a5.md: Update with new attributes.
6885 * config/arm/cortex-a53.md: Update with new attributes.
6886 * config/arm/cortex-a7.md: Update with new attributes.
6887 * config/arm/cortex-a8.md: Update with new attributes.
6888 * config/arm/cortex-a9.md: Update with new attributes.
6889 * config/arm/cortex-m4.md: Update with new attributes.
6890 * config/arm/cortex-r4.md: Update with new attributes.
6891 * config/arm/fa526.md: Update with new attributes.
6892 * config/arm/fa606te.md: Update with new attributes.
6893 * config/arm/fa626te.md: Update with new attributes.
6894 * config/arm/fa726te.md: Update with new attributes.
6896 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6898 * config/aarch64/aarch64-simd.md
6899 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
6900 <vwx> iterator to ensure correct register choice.
6901 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
6902 (aarch64_sqdmull_n<mode>): Likewise.
6903 (aarch64_sqdmull2_n<mode>_internal): Likewise.
6904 * config/aarch64/arm_neon.h
6905 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
6906 (vml<as><q>_n_<su>16): Likewise.
6907 (vml<as>l_high_lane<q>_<su>16): Likewise.
6908 (vml<as>l_high_n_<su>16): Likewise.
6909 (vml<as>l_lane<q>_<su>16): Likewise.
6910 (vml<as>l_n_<su>16): Likewise.
6911 (vmul<q>_lane<q>_<su>16): Likewise.
6912 (vmul<q>_n_<su>16): Likewise.
6913 (vmull_lane<q>_<su>16): Likewise.
6914 (vmull_n_<su>16): Likewise.
6915 (vmull_high_lane<q>_<su>16): Likewise.
6916 (vmull_high_n_<su>16): Likewise.
6917 (vqrdmulh<q>_n_s16): Likewise.
6919 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
6921 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
6922 have the correct lane parameter.
6924 2013-09-06 Richard Biener <rguenther@suse.de>
6926 * cfganal.c (control_dependences::~control_dependences):
6927 Properly free all of the vector.
6929 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
6932 * config/i386/i386.c (ix86_conditional_register_usage):
6933 Proper initialize extended SSE registers.
6935 2013-09-06 Jan Hubicka <jh@suse.cz>
6937 PR tree-optimization/58311
6938 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
6940 2013-09-06 Jan Hubicka <jh@suse.cz>
6942 * Makefile.in (tree-sra.o): Update dependencies.
6943 * tree-sra.c: Include ipa-utils.h
6944 (scan_function): Use recursive_call_p.
6945 (has_caller_p): New function.
6946 (cgraph_for_node_and_aliases): Count also callers of aliases.
6948 2013-09-06 Jan Hubicka <jh@suse.cz>
6951 * cgraph.h (symtab_semantically_equivalent_p): Declare.
6952 * tree-tailcall.c: Include ipa-utils.h.
6953 (find_tail_calls): Use it.
6954 * ipa-pure-const.c (check_call): Likewise.
6955 * ipa-utils.c (recursive_call_p): New function.
6956 * ipa-utils.h (recursive_call_p): Dclare.
6957 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
6958 (symtab_semantically_equivalent_p): New function.
6959 * Makefile.in (tree-tailcall.o): Update dependencies.
6961 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6963 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
6966 2013-09-06 Richard Biener <rguenther@suse.de>
6968 * lto-streamer.h (lto_global_var_decls): Remove.
6969 * Makefile.in (OBJS): Remove lto-symtab.o.
6970 (lto-symtab.o): Remove.
6971 (GTFILES): Remove lto-symtab.c
6972 * lto-symtab.c: Move to lto/
6974 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6976 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6977 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6978 (UNSPEC_FPINT_RINT): New constant definitions.
6979 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6980 definition with 2 attributes.
6981 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6982 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6985 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6987 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
6988 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6989 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6990 attribute to "z196_alone".
6991 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6994 2013-09-06 Richard Biener <rguenther@suse.de>
6996 * basic-block.h (class control_dependences): New.
6997 * tree-ssa-dce.c (control_dependence_map): Remove.
6999 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7000 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7001 find_pdom, find_control_dependence, find_all_control_dependences):
7003 (mark_control_dependent_edges_necessary,
7004 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7005 tree_dce_done, perform_tree_ssa_dce): Adjust.
7006 * cfganal.c (set_control_dependence_map_bit,
7007 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7008 find_all_control_dependences): Move from tree-ssa-dce.c and
7009 implement as methods of control_dependences class.
7010 (control_dependences::control_dependences): New.
7011 (control_dependences::~control_dependences): Likewise.
7012 (control_dependences::get_edges_dependent_on): Likewise.
7013 (control_dependences::get_edge): Likewise.
7015 2013-09-04 Jan Hubicka <jh@suse.cz>
7017 * tree.c (types_same_for_odr): Drop overactive check.
7018 * ipa-devirt.c (hash_type_name): Likewise.
7020 2013-09-04 Jan Hubicka <jh@suse.cz>
7022 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7023 (analyze_functions): ... here.
7025 2013-09-04 Jan Hubicka <jh@suse.cz>
7028 * cgraphunit.c (analyze_functions): Clear AUX fields
7029 after processing; initialize assembler name has.
7031 2013-09-05 Jeff Law <law@redhat.com>
7033 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7034 from thread_around_empty_block. Record threading path into PATH.
7035 Recurse if threading through the initial block is successful.
7036 (thread_across_edge): Corresponding changes to slightly simplify.
7038 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7040 * config/aarch64/aarch64.md
7041 (type): Remove frecpe, frecps, frecpx.
7042 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7043 fix to be a TARGET_SIMD instruction.
7044 (aarch64_frecps): Remove.
7045 * config/aarch64/aarch64-simd.md
7046 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7047 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7049 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7050 Sofiane Naci <sofiane.naci@arm.com>
7052 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7053 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7054 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7055 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7056 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7057 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7058 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7059 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7060 into "clz, "rbit". Rename "shift" to "shift_imm".
7061 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7062 changes. Update for attribute changes all occurrences of arlo_* and
7064 * config/arm/arm-fixed.md: Update for attribute changes
7065 all occurrences of arlo_* types.
7066 * config/arm/thumb2.md: Update for attribute changes all occurrences
7068 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7069 (cortexa7_older_only): Likewise.
7070 (cortexa7_younger): Likewise.
7071 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7072 (1020alu_shift_op): Likewise.
7073 (1020alu_shift_reg_op): Likewise.
7074 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7075 (alu_shift_op): Likewise.
7076 (alu_shift_reg_op): Likewise.
7077 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7078 (11_alu_shift_op): Likewise.
7079 (11_alu_shift_reg_op): Likewise.
7080 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7081 (9_alu_shift_reg_op): Likewise.
7082 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7084 (cortex_a15_alu_shift): Likewise.
7085 (cortex_a15_alu_shift_reg): Likewise.
7086 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7088 (cortex_a5_alu_shift): Likewise.
7089 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7091 (cortex_a53_alu_shift): Likewise.
7092 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7094 (cortex_a7_alu_reg): Likewise.
7095 (cortex_a7_alu_shift): Likewise.
7096 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7098 (cortex_a8_alu_shift): Likewise.
7099 (cortex_a8_alu_shift_reg): Likewise.
7100 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7101 (cortex_a9_dp_shift): Likewise.
7102 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7104 * config/arm/cortex-r4.md
7105 (cortex_r4_alu): Update for attribute changes.
7106 (cortex_r4_mov): Likewise.
7107 (cortex_r4_alu_shift_reg): Likewise.
7108 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7109 (526_alu_shift_op): Likewise.
7110 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7111 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7112 (626te_alu_shift_op): Likewise.
7113 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7114 (726te_alu_shift_op): Likewise.
7115 (726te_alu_shift_reg_op): Likewise.
7116 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7117 (mp626_alu_shift_op): Likewise.
7118 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7119 (pj4_alu_conds): Likewise.
7120 (pj4_shift): Likewise.
7121 (pj4_shift_conds): Likewise.
7122 (pj4_alu_shift): Likewise.
7123 (pj4_alu_shift_conds): Likewise.
7124 * config/aarch64/aarch64.md: Update for attribute change
7125 all occurrences of arlo_* and shift* types.
7127 2013-09-05 Mike Stump <mikestump@comcast.net>
7129 * tree.h: Move documentation for tree_function_decl to tree-core.h
7130 with the declaration.
7132 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7135 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7136 looking for widest mode.
7138 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7140 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7142 2013-09-05 Richard Biener <rguenther@suse.de>
7144 PR tree-optimization/58137
7145 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7146 Do not create vectors of pointers.
7147 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7148 types for the components of the vector initializer.
7149 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7150 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7152 2013-09-05 Martin Jambor <mjambor@suse.cz>
7154 * ipa-prop.c (remove_described_reference): Accept missing references,
7155 return false if that hppens, otherwise return true.
7156 (cgraph_node_for_jfunc): New function.
7157 (try_decrement_rdesc_refcount): Likewise.
7158 (try_make_edge_direct_simple_call): Use them.
7159 (ipa_edge_removal_hook): Remove references from rdescs.
7160 (ipa_edge_duplication_hook): Clone rdescs and their references
7161 when the new edge has the same caller as the old one.
7162 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7163 reference before removing any edges.
7165 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7167 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7169 * thumb2.md (thumb2_storewb_parisi): New pattern.
7171 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7173 * config/aarch64/aarch64-option-extensions.def: Add
7174 AARCH64_OPT_EXTENSION of 'crc'.
7175 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7176 (AARCH64_ISA_CRC): Ditto.
7177 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7178 description of the CRC extension.
7180 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7182 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7183 * config/rs6000/linux.h: Ditto.
7184 * alpha/linux.h: Ditto.
7185 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7186 no_c99_libc_has_function.
7187 * config/c6x/uclinux-elf.h: Ditto.
7188 * config/lm32/uclinux-elf.h: Ditto.
7189 * config/m68k/uclinux.h: Ditto.
7190 * config/moxie/uclinux.h: Ditto.
7191 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7192 (crisv32-*-linux*, cris-*-linux*): Ditto.
7193 * config/bfin/bfin.c: Include "tm_p.h".
7195 2013-09-05 Richard Biener <rguenther@suse.de>
7197 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7198 check for a definition without a basic-block.
7200 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7201 Sofiane Naci <sofiane.naci@arm.com>
7203 * config/aarch64/aarch64.md
7204 (*movti_aarch64): Rename r_2_f and f_2_r.
7205 (*movsf_aarch64): Likewise.
7206 (*movdf_aarch64): Likewise.
7207 (*movtf_aarch64): Likewise.
7208 (aarch64_movdi_<mode>low): Likewise.
7209 (aarch64_movdi_<mode>high): Likewise.
7210 (aarch64_mov<mode>high_di): Likewise.
7211 (aarch64_mov<mode>low_di): Likewise.
7212 (aarch64_movtilow_tilow): Likewise.
7213 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7214 values to config/arm/types.md
7215 (attribute "conds"): Update for attribute change.
7216 (anddi3_insn): Likewise.
7217 (iordi3_insn): Likewise.
7218 (xordi3_insn): Likewise.
7219 (one_cmpldi2): Likewise.
7220 * config/arm/types.md (type): Add Neon types.
7221 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7222 use "type" attribute.
7223 (movmisalign<mode>_neon_store): Likewise.
7224 (movmisalign<mode>_neon_load): Likewise.
7225 (vec_set<mode>_internal): Likewise.
7226 (vec_setv2di_internal): Likewise.
7227 (vec_extract<mode>): Likewise.
7228 (vec_extractv2di): Likewise.
7229 (add<mode>3_neon): Likewise.
7230 (adddi3_neon): Likewise.
7231 (sub<mode>3_neon): Likewise.
7232 (subdi3_neon): Likewise.
7233 (mul<mode>3_neon): Likewise.
7234 (mul<mode>3add<mode>_neon): Likewise.
7235 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7236 (fma<VCVTF:mode>4)): Likewise.
7237 (fma<VCVTF:mode>4_intrinsic): Likewise.
7238 (fmsub<VCVTF:mode>4)): Likewise.
7239 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7240 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7241 (ior<mode>3): Likewise.
7242 (and<mode>3): Likewise.
7243 (anddi3_neon): Likewise.
7244 (orn<mode>3_neon): Likewise.
7245 (orndi3_neon): Likewise.
7246 (bic<mode>3_neon): Likewise.
7247 (bicdi3_neon): Likewise.
7248 (xor<mode>3): Likewise.
7249 (one_cmpl<mode>2): Likewise.
7250 (abs<mode>2): Likewise.
7251 (neg<mode>2): Likewise.
7252 (umin<mode>3_neon): Likewise.
7253 (umax<mode>3_neon): Likewise.
7254 (smin<mode>3_neon): Likewise.
7255 (smax<mode>3_neon): Likewise.
7256 (vashl<mode>3): Likewise.
7257 (vashr<mode>3_imm): Likewise.
7258 (vlshr<mode>3_imm): Likewise.
7259 (ashl<mode>3_signed): Likewise.
7260 (ashl<mode>3_unsigned): Likewise.
7261 (neon_load_count): Likewise.
7262 (ashldi3_neon_noclobber): Likewise.
7263 (signed_shift_di3_neon): Likewise.
7264 (unsigned_shift_di3_neon): Likewise.
7265 (ashrdi3_neon_imm_noclobber): Likewise.
7266 (lshrdi3_neon_imm_noclobber): Likewise.
7267 (widen_ssum<mode>3): Likewise.
7268 (widen_usum<mode>3): Likewise.
7269 (quad_halves_<code>v4si): Likewise.
7270 (quad_halves_<code>v4sf): Likewise.
7271 (quad_halves_<code>v8hi): Likewise.
7272 (quad_halves_<code>v16qi): Likewise.
7273 (reduc_splus_v2di): Likewise.
7274 (neon_vpadd_internal<mode>): Likewise.
7275 (neon_vpsmin<mode>): Likewise.
7276 (neon_vpsmax<mode>): Likewise.
7277 (neon_vpumin<mode>): Likewise.
7278 (neon_vpumax<mode>): Likewise.
7279 (ss_add<mode>_neon): Likewise.
7280 (us_add<mode>_neon): Likewise.
7281 (ss_sub<mode>_neon): Likewise.
7282 (us_sub<mode>_neon): Likewise.
7283 (neon_vadd<mode>_unspec): Likewise.
7284 (neon_vaddl<mode>): Likewise.
7285 (neon_vaddw<mode>): Likewise.
7286 (neon_vhadd<mode>): Likewise.
7287 (neon_vqadd<mode>): Likewise.
7288 (neon_vaddhn<mode>): Likewise.
7289 (neon_vmul<mode>): Likewise.
7290 (neon_vmla<mode>): Likewise.
7291 (neon_vmlal<mode>): Likewise.
7292 (neon_vmls<mode>): Likewise.
7293 (neon_vmlsl<mode>): Likewise.
7294 (neon_vqdmulh<mode>): Likewise.
7295 (neon_vqdmlal<mode>): Likewise.
7296 (neon_vqdmlsl<mode>): Likewise.
7297 (neon_vmull<mode>): Likewise.
7298 (neon_vqdmull<mode>): Likewise.
7299 (neon_vsub<mode>_unspec): Likewise.
7300 (neon_vsubl<mode>): Likewise.
7301 (neon_vsubw<mode>): Likewise.
7302 (neon_vqsub<mode>): Likewise.
7303 (neon_vhsub<mode>): Likewise.
7304 (neon_vsubhn<mode>): Likewise.
7305 (neon_vceq<mode>): Likewise.
7306 (neon_vcge<mode>): Likewise.
7307 (neon_vcgeu<mode>): Likewise.
7308 (neon_vcgt<mode>): Likewise.
7309 (neon_vcgtu<mode>): Likewise.
7310 (neon_vcle<mode>): Likewise.
7311 (neon_vclt<mode>): Likewise.
7312 (neon_vcage<mode>): Likewise.
7313 (neon_vcagt<mode>): Likewise.
7314 (neon_vtst<mode>): Likewise.
7315 (neon_vabd<mode>): Likewise.
7316 (neon_vabdl<mode>): Likewise.
7317 (neon_vaba<mode>): Likewise.
7318 (neon_vabal<mode>): Likewise.
7319 (neon_vmax<mode>): Likewise.
7320 (neon_vmin<mode>): Likewise.
7321 (neon_vpaddl<mode>): Likewise.
7322 (neon_vpadal<mode>): Likewise.
7323 (neon_vpmax<mode>): Likewise.
7324 (neon_vpmin<mode>): Likewise.
7325 (neon_vrecps<mode>): Likewise.
7326 (neon_vrsqrts<mode>): Likewise.
7327 (neon_vqabs<mode>): Likewise.
7328 (neon_vqneg<mode>): Likewise.
7329 (neon_vcls<mode>): Likewise.
7330 (clz<mode>2): Likewise.
7331 (popcount<mode>2): Likewise.
7332 (neon_vrecpe): Likewise.
7333 (neon_vrsqrte): Likewise.
7334 (neon_vget_lane<mode>_sext_internal): Likewise.
7335 (neon_vget_lane<mode>_zext_internal): Likewise.
7336 (neon_vdup_n<mode>): Likewise.
7337 (neon_vdup_nv2di): Likewise.
7338 (neon_vdpu_lane<mode>_internal): Likewise.
7339 (neon_vswp<mode>): Likewise.
7340 (float<mode><V_cvtto>2): Likewise.
7341 (floatuns<mode><V_cvtto>2): Likewise.
7342 (fix_trunc<mode><V_cvtto>)2): Likewise
7343 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7344 (neon_vcvt<mode>): Likewise.
7345 (neon_vcvtv4sfv4hf): Likewise.
7346 (neon_vcvtv4hfv4sf): Likewise.
7347 (neon_vcvt_n<mode>): Likewise.
7348 (neon_vmovn<mode>): Likewise.
7349 (neon_vqmovn<mode>): Likewise.
7350 (neon_vqmovun<mode>): Likewise.
7351 (neon_vmovl<mode>): Likewise.
7352 (neon_vmul_lane<mode>): Likewise.
7353 (neon_vmull_lane<mode>): Likewise.
7354 (neon_vqdmull_lane<mode>): Likewise.
7355 (neon_vqdmulh_lane<mode>): Likewise.
7356 (neon_vmla_lane<mode>): Likewise.
7357 (neon_vmlal_lane<mode>): Likewise.
7358 (neon_vqdmlal_lane<mode>): Likewise.
7359 (neon_vmls_lane<mode>): Likewise.
7360 (neon_vmlsl_lane<mode>): Likewise.
7361 (neon_vqdmlsl_lane<mode>): Likewise.
7362 (neon_vext<mode>): Likewise.
7363 (neon_vrev64<mode>): Likewise.
7364 (neon_vrev32<mode>): Likewise.
7365 (neon_vrev16<mode>): Likewise.
7366 (neon_vbsl<mode>_internal): Likewise.
7367 (neon_vshl<mode>): Likewise.
7368 (neon_vqshl<mode>): Likewise.
7369 (neon_vshr_n<mode>): Likewise.
7370 (neon_vshrn_n<mode>): Likewise.
7371 (neon_vqshrn_n<mode>): Likewise.
7372 (neon_vqshrun_n<mode>): Likewise.
7373 (neon_vshl_n<mode>): Likewise.
7374 (neon_vqshl_n<mode>): Likewise.
7375 (neon_vqshlu_n<mode>): Likewise.
7376 (neon_vshll_n<mode>): Likewise.
7377 (neon_vsra_n<mode>): Likewise.
7378 (neon_vsri_n<mode>): Likewise.
7379 (neon_vsli_n<mode>): Likewise.
7380 (neon_vtbl1v8qi): Likewise.
7381 (neon_vtbl2v8qi): Likewise.
7382 (neon_vtbl3v8qi): Likewise.
7383 (neon_vtbl4v8qi): Likewise.
7384 (neon_vtbx1v8qi): Likewise.
7385 (neon_vtbx2v8qi): Likewise.
7386 (neon_vtbx3v8qi): Likewise.
7387 (neon_vtbx4v8qi): Likewise.
7388 (neon_vtrn<mode>_internal): Likewise.
7389 (neon_vzip<mode>_internal): Likewise.
7390 (neon_vuzp<mode>_internal): Likewise.
7391 (neon_vld1<mode>): Likewise.
7392 (neon_vld1_lane<mode>): Likewise.
7393 (neon_vld1_dup<mode>): Likewise.
7394 (neon_vld1_dupv2di): Likewise.
7395 (neon_vst1<mode>): Likewise.
7396 (neon_vst1_lane<mode>): Likewise.
7397 (neon_vld2<mode>): Likewise.
7398 (neon_vld2_lane<mode>): Likewise.
7399 (neon_vld2_dup<mode>): Likewise.
7400 (neon_vst2<mode>): Likewise.
7401 (neon_vst2_lane<mode>): Likewise.
7402 (neon_vld3<mode>): Likewise.
7403 (neon_vld3qa<mode>): Likewise.
7404 (neon_vld3qb<mode>): Likewise.
7405 (neon_vld3_lane<mode>): Likewise.
7406 (neon_vld3_dup<mode>): Likewise.
7407 (neon_vst3<mode>): Likewise.
7408 (neon_vst3qa<mode>): Likewise.
7409 (neon_vst3qb<mode>): Likewise.
7410 (neon_vst3_lane<mode>): Likewise.
7411 (neon_vld4<mode>): Likewise.
7412 (neon_vld4qa<mode>): Likewise.
7413 (neon_vld4qb<mode>): Likewise.
7414 (neon_vld4_lane<mode>): Likewise.
7415 (neon_vld4_dup<mode>): Likewise.
7416 (neon_vst4<mode>): Likewise.
7417 (neon_vst4qa<mode>): Likewise.
7418 (neon_vst4qb<mode>): Likewise.
7419 (neon_vst4_lane<mode>): Likewise.
7420 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7421 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7422 (neon_vec_<US>mult_lo_<mode>): Likewise.
7423 (neon_vec_<US>mult_hi_<mode>): Likewise.
7424 (neon_vec_<US>shiftl_<mode>): Likewise.
7425 (neon_unpack<US>_<mode>): Likewise.
7426 (neon_vec_<US>mult_<mode>): Likewise.
7427 (vec_pack_trunc_<mode>): Likewise.
7428 (neon_vec_pack_trunk_<mode>): Likewise.
7429 (neon_vabd<mode>_2): Likewise.
7430 (neon_vabd<mode>_3): Likewise.
7431 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7432 (thumb2_movsi_vfp): Likewise.
7433 (movdi_vfp): Likewise.
7434 (movdi_vfp_cortexa8): Likewise.
7435 (movhf_vfp_neon): Likewise.
7436 (movhf_vfp): Likewiwse.
7437 (movsf_vfp): Likewiwse.
7438 (thumb2_movsf_vfp): Likewiwse.
7439 (movdf_vfp): Likewise.
7440 (thumb2_movdf_vfp): Likewise.
7441 (movsfcc_vfp): Likewise.
7442 (thumb2_movsfcc_vfp): Likewise.
7443 (movdfcc_vfp): Likewise.
7444 (thumb2_movdfcc_vfp): Likewise.
7445 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7446 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7447 (v10_v2c): Likewise.
7448 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7450 (cortex_a15_neon_int_2): Likewise.
7451 (cortex_a15_neon_int_3): Likewise.
7452 (cortex_a15_neon_int_4): Likewise.
7453 (cortex_a15_neon_int_5): Likewise.
7454 (cortex_a15_neon_vqneg_vqabs): Likewise.
7455 (cortex_a15_neon_vmov): Likewise.
7456 (cortex_a15_neon_vaba): Likewise.
7457 (cortex_a15_neon_vaba_qqq): Likewise.
7458 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7459 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7460 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7461 scalar_64_32_long_scalar): Likewise.
7462 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7463 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7464 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7465 lotype_qdd_64_32_long): Likewise.
7466 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7467 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7468 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7469 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7470 (cortex_a15_neon_shift_1): Likewise.
7471 (cortex_a15_neon_shift_2): Likewise.
7472 (cortex_a15_neon_shift_3): Likewise.
7473 (cortex_a15_neon_vshl_ddd): Likewise.
7474 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7475 (cortex_a15_neon_vsra_vrsra): Likewise.
7476 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7477 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7478 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7479 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7480 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7481 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7482 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7483 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7484 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7485 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7486 (cortex_a15_neon_bp_simple): Likewise.
7487 (cortex_a15_neon_bp_2cycle): Likewise.
7488 (cortex_a15_neon_bp_3cycle): Likewise.
7489 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7490 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7491 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7492 (cortex_a15_neon_vld2_4_regs): Likewise.
7493 (cortex_a15_neon_vld3_vld4): Likewise.
7494 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7495 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7496 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7497 (cortex_a15_neon_vst3_vst4): Likewise.
7498 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7499 (cortex_a15_neon_vld3_vld4_lane" 10
7500 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7501 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7502 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7503 (cortex_a15_neon_ldm_2): Likewise.0
7504 (cortex_a15_neon_stm_2): Likewise.
7505 (cortex_a15_neon_mcr): Likewise.
7506 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7507 (cortex_a15_neon_mrc): Likewise.
7508 (cortex_a15_neon_mrrc): Likewise.
7509 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7511 (cortex_a15_alu_shift): Likewise.
7512 (cortex_a15_alu_shift_reg): Likewise.
7513 (cortex_a15_mult32): Likewise.
7514 (cortex_a15_mult64): Likewise.
7515 (cortex_a15_block): Likewise.
7516 (cortex_a15_branch): Likewise.
7517 (cortex_a15_load1): Likewise.
7518 (cortex_a15_load3): Likewise.
7519 (cortex_a15_store1): Likewise.
7520 (cortex_a15_store3): Likewise.
7521 (cortex_a15_call): Likewise.
7522 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7523 (cortex_a5_f2r): Likewise.
7524 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7526 (cortex_a53_f2r): Likewise.
7527 * config/arm/cortex-a7.md
7528 (cortex_a7_branch): Update for attribute change.
7529 (cortex_a7_call): Likewise.
7530 (cortex_a7_alu_imm): Likewise.
7531 (cortex_a7_alu_reg): Likewise.
7532 (cortex_a7_alu_shift): Likewise.
7533 (cortex_a7_mul): Likewise.
7534 (cortex_a7_load1): Likewise.
7535 (cortex_a7_store1): Likewise.
7536 (cortex_a7_load2): Likewise.
7537 (cortex_a7_store2): Likewise.
7538 (cortex_a7_load3): Likewise.
7539 (cortex_a7_store3): Likewise.
7540 (cortex_a7_load4): Likewise.
7541 (cortex_a7_store4): Likewise.
7542 (cortex_a7_fpalu): Likewise.
7543 (cortex_a7_fconst): Likewise.
7544 (cortex_a7_fpmuls): Likewise.
7545 (cortex_a7_neon_mul): Likewise.
7546 (cortex_a7_fpmacs): Likewise.
7547 (cortex_a7_neon_mla: Likewise.
7548 (cortex_a7_fpmuld: Likewise.
7549 (cortex_a7_fpmacd: Likewise.
7550 (cortex_a7_fpfmad: Likewise.
7551 (cortex_a7_fdivs: Likewise.
7552 (cortex_a7_fdivd: Likewise.
7553 (cortex_a7_r2f: Likewise.
7554 (cortex_a7_f2r: Likewise.
7555 (cortex_a7_f_flags: Likewise.
7556 (cortex_a7_f_loads: Likewise.
7557 (cortex_a7_f_loadd: Likewise.
7558 (cortex_a7_f_stores: Likewise.
7559 (cortex_a7_f_stored: Likewise.
7560 (cortex_a7_neon): Likewise.
7561 * config/arm/cortex-a8-neon.md
7562 (cortex_a8_neon_mrc): Update for attribute change.
7563 (cortex_a8_neon_mrrc): Likewise.
7564 (cortex_a8_neon_int_1): Likewise.
7565 (cortex_a8_neon_int_2): Likewise.
7566 (cortex_a8_neon_int_3): Likewise.
7567 (cortex_a8_neon_int_4): Likewise.
7568 (cortex_a8_neon_int_5): Likewise.
7569 (cortex_a8_neon_vqneg_vqabs): Likewise.
7570 (cortex_a8_neon_vmov): Likewise.
7571 (cortex_a8_neon_vaba): Likewise.
7572 (cortex_a8_neon_vaba_qqq): Likewise.
7573 (cortex_a8_neon_vsma): Likewise.
7574 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7575 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7576 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7577 long_scalar): Likewise.
7578 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7579 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7580 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7581 long_scalar_qdd_64_32_long): Likewise.
7582 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7583 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7584 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7585 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7586 (cortex_a8_neon_shift_1): Likewise.
7587 (cortex_a8_neon_shift_2): Likewise.
7588 (cortex_a8_neon_shift_3): Likewise.
7589 (cortex_a8_neon_vshl_ddd): Likewise.
7590 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7591 (cortex_a8_neon_vsra_vrsra): Likewise.
7592 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7593 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7594 (cortex_a8_neon_fp_vsum): Likewise.
7595 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7596 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7597 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7598 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7599 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7600 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7601 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7602 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7603 (cortex_a8_neon_bp_simple): Likewise.
7604 (cortex_a8_neon_bp_2cycle): Likewise.
7605 (cortex_a8_neon_bp_3cycle): Likewise.
7606 (cortex_a8_neon_ldr): Likewise.
7607 (cortex_a8_neon_str): Likewise.
7608 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7609 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7610 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7611 (cortex_a8_neon_vld2_4_regs): Likewise.
7612 (cortex_a8_neon_vld3_vld4): Likewise.
7613 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7614 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7615 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7616 (cortex_a8_neon_vst3_vst4): Likewise.
7617 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7618 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7619 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7620 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7621 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7622 (cortex_a8_neon_mcr): Likewise.
7623 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7624 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7625 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7627 (ca9_neon_mrrc): Likewise.
7628 (cortex_a9_neon_int_1): Likewise.
7629 (cortex_a9_neon_int_2): Likewise.
7630 (cortex_a9_neon_int_3): Likewise.
7631 (cortex_a9_neon_int_4): Likewise.
7632 (cortex_a9_neon_int_5): Likewise.
7633 (cortex_a9_neon_vqneg_vqabs): Likewise.
7634 (cortex_a9_neon_vmov): Likewise.
7635 (cortex_a9_neon_vaba): Likewise.
7636 (cortex_a9_neon_vaba_qqq): Likewise.
7637 (cortex_a9_neon_vsma): Likewise.
7638 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7639 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7640 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7641 long_scalar): Likewise.
7642 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7643 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7644 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7645 long_scalar_qdd_64_32_long): Likewise.
7646 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7647 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7648 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7649 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7650 (cortex_a9_neon_shift_1): Likewise.
7651 (cortex_a9_neon_shift_2): Likewise.
7652 (cortex_a9_neon_shift_3): Likewise.
7653 (cortex_a9_neon_vshl_ddd): Likewise.
7654 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7655 (cortex_a9_neon_vsra_vrsra): Likewise.
7656 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7657 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7658 (cortex_a9_neon_fp_vsum): Likewise.
7659 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7660 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7661 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7662 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7663 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7664 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7665 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7666 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7667 (cortex_a9_neon_bp_simple): Likewise.
7668 (cortex_a9_neon_bp_2cycle): Likewise.
7669 (cortex_a9_neon_bp_3cycle): Likewise.
7670 (cortex_a9_neon_ldr): Likewise.
7671 (cortex_a9_neon_str): Likewise.
7672 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7673 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7674 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7675 (cortex_a9_neon_vld2_4_regs): Likewise.
7676 (cortex_a9_neon_vld3_vld4): Likewise.
7677 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7678 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7679 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7680 (cortex_a9_neon_vst3_vst4): Likewise.
7681 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7682 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7683 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7684 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7685 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7686 (cortex_a9_neon_mcr): Likewise.
7687 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7688 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7689 (cortex_a9_fps): Likewise.
7690 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7692 (cortex_m4_fmuls): Likewise.
7693 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7695 (cortex_r4_mrc): Likewise.
7696 * config/arm/iterators.md: Update comment referring to neon_type.
7697 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7698 (iwmmxt_movsi_insn): Likewise.
7699 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7701 (pj4_core_to_vfp): Likewise.
7702 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7704 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7705 (vfp_fstore): Likewise.
7706 * doc/md.texi: Change references to neon_type to refer to type.
7708 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7710 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7712 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7714 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7717 2013-09-04 Easwaran Raman <eraman@google.com>
7720 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7721 (build_and_add_sum): Use it.
7722 (appears_later_in_bb): Simplify code.
7724 2013-09-04 Teresa Johnson <tejohnson@google.com>
7726 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7728 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7730 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7732 2013-09-04 Jan Hubicka <jh@suse.cz>
7734 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7735 * ipa-devirt.c: Include diganostic.h
7736 (odr_type_d): Add types and types_set.
7737 (hash_type_name): Work for types with vtables during LTO.
7738 (odr_hasher::remove): Fix comment; destroy types_set.
7739 (add_type_duplicate): New function,
7740 (get_odr_type): Use it.
7741 (dump_type_inheritance_graph): Dump type duplicates.
7742 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7744 * tree.c (types_same_for_odr): Give exact answers on types with
7747 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7749 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7750 explaining their differences.
7752 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7754 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7756 2013-09-03 Jeff Law <law@redhat.com>
7758 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7759 when not threading through a joiner block. Pass joiner/no joiner
7760 state to register_jump_thread.
7761 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7762 state from argument rather than implying on path length.
7763 Dump the entire jump thread path into debugging dump.
7764 * tree-flow.h (register_jump_thread): Update prototype.
7766 2013-08-29 Xinliang David Li <davidxl@google.com>
7768 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7769 Remove a trivial gcc_assert.
7771 2013-08-29 Xinliang David Li <davidxl@google.com>
7773 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7774 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7775 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7776 Delay base decl alignment adjustment.
7777 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7778 * tree-vectorizer.h: New data structure.
7779 (set_dr_misalignment): New function.
7780 (dr_misalignment): Ditto.
7781 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7782 (vectorizable_load): Ditto.
7783 (ensure_base_align): New function.
7784 (vectorize_loops): Add dbg_cnt support.
7785 (execute_vect_slp): Ditto.
7786 * dbgcnt.def: New debug counter.
7787 * Makefile: New dependency.
7789 2013-09-03 Meador Inge <meadori@codesourcery.com>
7793 2013-08-30 Meador Inge <meadori@codesourcery.com>
7795 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7797 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7799 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7800 function descriptor.
7802 2013-09-03 Richard Biener <rguenther@suse.de>
7804 * tree-affine.c (add_elt_to_tree): Fix association issue,
7805 avoid useless converts and make sure to always return a
7806 properly typed result.
7808 2013-09-03 Richard Biener <rguenther@suse.de>
7811 * fold-const.c (negate_expr_p): Fix division case.
7812 (negate_expr): Likewise.
7814 2013-09-03 Richard Biener <rguenther@suse.de>
7817 * tree-streamer-out.c: Include tm.h.
7818 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7820 2013-09-03 Jan Hubicka <jh@suse.cz>
7822 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7824 2013-09-03 Alan Modra <amodra@gmail.com>
7826 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7827 multilibs to big-endian target and vice versa.
7828 * config/rs6000/t-linux64: Use := assignment on all vars.
7829 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7830 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7831 * config/rs6000/t-linux64le: New file.
7832 * config/rs6000/t-linux64bele: New file.
7833 * config/rs6000/t-linux64lebe: New file.
7835 2013-09-02 Jan Hubicka <jh@suse.cz>
7837 * ipa-inline-transform.c (inline_transform): Do not
7838 optimize_inline_calls when not optimizing.
7840 2013-09-02 Jan Hubicka <jh@suse.cz>
7842 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7843 duplicated nodes for assembler names.
7844 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7845 hard registers from assembler name hash.
7847 2013-09-02 Jan Hubicka <jh@suse.cz>
7849 * ipa-split.c (execute_split_functions): Split externally visible
7850 functions called once.
7852 2013-09-02 Martin Jambor <mjambor@suse.cz>
7855 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7856 linked list. When finding the correct duplicate, also consider also
7857 the caller in additon to its inlined_to node.
7859 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7861 * config/aarch64/aarch64-simd-builtins.def
7862 (dup_lane_scalar): Remove.
7863 * config/aarch64/aarch64-simd.md
7864 (aarch64_simd_dup): Add 'w->w' alternative.
7865 (aarch64_dup_lane<mode>): Allow for VALL.
7866 (aarch64_dup_lane_scalar<mode>): Remove.
7867 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7868 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7869 (aarch64_get_lane_unsigned<mode>): Likewise.
7870 (aarch64_get_lane<mode>): Likewise.
7871 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7872 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7873 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7874 (VCON): Change container of V2SF.
7875 (vswap_width_name): Likewise.
7876 * config/aarch64/arm_neon.h
7877 (__aarch64_vdup_lane_any): New.
7878 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7879 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7880 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7882 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7885 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7886 the source or the destination is a single hard register.
7888 2013-09-02 Richard Biener <rguenther@suse.de>
7891 * tree-scalar-evolution.c (instantiate_scev_name): Allow
7894 2013-09-02 Richard Biener <rguenther@suse.de>
7896 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
7897 arithmetic to sizetype.
7899 2013-09-02 Bin Cheng <bin.cheng@arm.com>
7901 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
7902 Find auto-increment use both before and after candidate.
7904 2013-09-02 Marek Polacek <polacek@redhat.com>
7906 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
7908 2013-09-01 Jan Hubicka <jh@suse.cz>
7910 * Makefile.in: Add ipa-profile.o
7911 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
7912 * cgraph.c (struct cgraph_propagate_frequency_data,
7913 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
7914 ipa-profile.c; replace cgraph_ by ipa_ prefix.
7915 * cgraph.h (cgraph_propagate_frequency): Remove.
7916 * ipa-inline-analysis.c: Include ipa-utils.h;
7917 drop duplicated cfgloop.h.
7918 (inline_update_callee_summaries): Update.
7919 * ipa-profile.c: New file.
7920 * ipa-utils.h (ipa_propagate_frequency): Declare.
7921 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
7922 data-streamer.h, value-prof.h.
7923 (symtab_remove_unreachable_nodes): Update profile.
7924 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
7925 account_time_size, cmp_counts, dump_histogram,
7926 ipa_profile_generate_summary, ipa_profile_write_summary,
7927 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
7928 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
7929 Move to ipa-profile.c.
7931 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
7933 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
7935 2013-09-01 Jan Hubicka <jh@suse.cz>
7937 * common.opt (fdevirtualize-speculatively): New function.
7938 * invoke.texi (fdevirtualize-speculatively): Document.
7939 * ipa-devirt.c: Include ipa-inline.h
7940 (likely_target_p): New function.
7941 (ipa_devirt): New function.
7942 (gate_ipa_devirt): New function.
7943 (pass_data_ipa_devirt): New static var.
7944 (pass_ipa_devirt): Likewise.
7945 (make_pass_ipa_devirt): New function.
7946 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
7947 (common_handle_option): Disable devirtualization when
7948 value range profiling is available.
7949 * passes.def (pass_ipa_devirt): Add.
7950 * timever.def (TV_IPA_DEVIRT): New timevar.
7951 * tree-pass.h (make_pass_ipa_devirt):
7953 2013-09-01 Iain Sandoe <iain@codesourcery.com>
7955 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
7956 include sanitize(undefined).
7958 2013-08-31 Diego Novillo <dnovillo@google.com>
7960 * Makefile.in (TREE_CORE_H): Define.
7962 (GTFILES): Add tree-core.h.
7963 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
7965 * tree-core.h: New file.
7966 Move all data structures, enum, typedefs, global
7967 declarations and constants from ...
7970 2013-08-31 Jan Hubicka <jh@suse.cz>
7972 * bulitins.c (expand_builtin): Do not early exit for gcov
7973 instrumented functions.
7975 2013-08-31 Marek Polacek <polacek@redhat.com>
7977 * ubsan.c: Include tm_p.h.
7979 2013-08-31 Jan Hubicka <jh@suse.cz>
7981 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7984 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7985 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
7986 * tree-cfg.c (verify_gimple_label): ... here.
7987 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7988 (ipa_merge_profiles): New function.
7989 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7990 (lto_input_function_body): Likewise.
7991 * ipa-utils.h (ipa_merge_profiles): Declare.
7992 * lto-streamer.h (lto_input_function_body): Update prototype.
7993 (emit_label_in_global_context_p): Remove.
7994 * lto-symtab.c: Include ipa-utils.h
7995 (lto_cgraph_replace_node): Use ipa_merge_profiles.
7997 2013-08-31 Jan Hubicka <jh@suse.cz>
7999 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8001 2013-08-31 Jan Hubicka <jh@suse.cz>
8003 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8005 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8007 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8008 "cmp" RTX before signed_comparison_operator check to account
8011 2013-08-30 Jan Hubicka <jh@suse.cz>
8013 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8014 (detect_type_change_1): Rename to ...
8015 (detect_type_change): ... this one; early return on non-polymorphic
8017 (detect_type_change_ssa): Add comp_type parameter; update
8018 use of detect_type_change.
8019 (compute_complex_assign_jump_func): Add param_type parameter;
8020 update use of detect_type_change_ssa.
8021 (compute_complex_ancestor_jump_func): Likewise.
8022 (ipa_get_callee_param_type): New function.
8023 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8024 update calls to the jump function computation functions.
8026 2013-08-30 Teresa Johnson <tejohnson@google.com>
8027 Steven Bosscher <steven@gcc.gnu.org>
8029 * cfgrtl.c (fixup_new_cold_bb): New routine.
8030 (commit_edge_insertions): Invoke fixup_partitions.
8031 (find_partition_fixes): New routine.
8032 (fixup_partitions): Ditto.
8033 (verify_hot_cold_block_grouping): Update comments.
8034 (rtl_verify_edges): Invoke find_partition_fixes.
8035 (rtl_verify_bb_pointers): Update comments.
8036 (rtl_verify_bb_layout): Ditto.
8037 * basic-block.h (probably_never_executed_edge_p): Declare.
8038 (fixup_partitions): Ditto.
8039 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8040 * bb-reorder.c (sanitize_hot_paths): New function.
8041 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8043 * predict.c (probably_never_executed_edge_p): New routine.
8044 * cfg.c (check_bb_profile): Add partition insanity warnings.
8046 2013-08-30 Meador Inge <meadori@codesourcery.com>
8048 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8050 2013-08-30 Marek Polacek <polacek@redhat.com>
8052 * Makefile.in (ubsan.o): Add.
8053 (c-family/c-ubsan.o): Add.
8054 (builtins.o): Add ubsan.h dependency.
8055 * ubsan.h: New file.
8056 * ubsan.c: New file.
8057 * common.opt: Add -fsanitize=undefined option.
8058 (flag_sanitize): Add variable.
8059 (fsanitize=): Add option. Add Driver.
8060 (fsanitize=thread): Remove option.
8061 (fsanitize=address): Likewise.
8062 (static-libubsan): New option.
8063 * doc/invoke.texi: Document the new flag and -static-libubsan.
8064 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8065 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8066 * builtin-attrs.def (ATTR_COLD): Define.
8067 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8068 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8069 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8070 * flag-types.h (sanitize_code): New enum.
8071 * opts.c (common_handle_option): Parse command line arguments
8072 of -fsanitize=. Add -fsanitize=unreachable option.
8073 * varasm.c (get_variable_section): Adjust.
8074 (assemble_noswitch_variable): Likewise.
8075 (assemble_variable): Likewise.
8076 (output_constant_def_contents): Likewise.
8077 (categorize_decl_for_section): Likewise.
8078 (place_block_symbol): Likewise.
8079 (output_object_block): Likewise.
8080 * builtins.def: Likewise.
8081 * toplev.c (compile_file): Likewise.
8082 (process_options): Likewise.
8083 * cppbuiltin.c: Likewise.
8084 * tsan.c (tsan_pass): Likewise.
8085 (tsan_gate): Likewise.
8086 (tsan_gate_O0): Likewise.
8087 * cfgexpand.c (partition_stack_vars): Likewise.
8088 (expand_stack_vars): Likewise.
8089 (defer_stack_allocation): Likewise.
8090 (expand_used_vars): Likewise.
8091 * cfgcleanup.c (old_insns_match_p): Likewise.
8092 * asan.c (asan_finish_file): Likewise.
8093 (asan_instrument): Likewise.
8094 (gate_asan): Likewise.
8095 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8096 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8097 (asan_global_struct): Use pointer_sized_int_node instead
8098 calling build_nonstandard_integer_type.
8099 (initialize_sanitizer_builtins): Likewise.
8100 (asan_finish_file): Likewise.
8101 * gcc.c: Document %{%:function(args):X}.
8102 (static_spec_functions): Add sanitize.
8103 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8104 store funcval != NULL there.
8105 (do_spec_1): Adjust handle_spec_function caller.
8106 (handle_braces): Allow %:function(args) as condition.
8107 (sanitize_spec_function): New function.
8108 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8109 (LIBUBSAN_SPEC): Likewise.
8110 (LIBUBSAN_EARLY_SPEC): Likewise.
8111 (SANITIZER_SPEC): Handle libubsan.
8112 (SANITIZER_EARLY_SPEC): Likewise.
8113 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8114 instead of fsanitize=address.
8115 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8116 instead of fsanitize=address*.
8117 * builtins.c: Include ubsan.h.
8118 (fold_builtin_0): Instrument __builtin_unreachable.
8119 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8120 instead of flag_asan.
8121 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8122 (pointer_sized_int_node): Define.
8123 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8125 2013-08-30 Mike Stump <mikestump@comcast.net>
8127 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8130 2013-08-29 Jan Hubicka <jh@suse.cz>
8132 * cgraph.c (cgraph_function_body_availability): Handle weakref
8134 * passes.def: Remove pass_fixup_cfg.
8135 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8136 track when we need to remove functions.
8137 (gate_ipa_inline): Execute inlining always; add comment why.
8138 (pass_data_ipa_inline): Remove TODO_remove_functions.
8139 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8140 do not produce summaries.
8141 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8142 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8143 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8144 constant pool and vtable.
8146 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8148 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8149 New arm_neon.h's internal macros to specify 64-bit constants.
8150 Avoid using stdint.h's macros.
8152 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8154 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8155 before applying REG_P.
8157 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8159 PR tree-optimization/58277
8160 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8161 after seeing too many stmts with vdef in between dombb and current
8162 bb, invalidate everything.
8164 2013-08-30 Richard Biener <rguenther@suse.de>
8166 * fold-const.c (fold_single_bit_test): Fix overflow test.
8168 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8170 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8171 and which can live in a register, always retrieve the value on entry.
8172 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8173 passed by invisible reference specially.
8174 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8175 (vt_add_function_parameter): Correctly deal with a parameter passed by
8176 invisible reference.
8178 2013-08-30 Jan Hubicka <jh@suse.cz>
8180 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8182 2013-08-30 Richard Biener <rguenther@suse.de>
8184 PR tree-optimization/58228
8185 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8186 allow invariant loads in nested loop vectorization.
8188 2013-08-30 Richard Biener <rguenther@suse.de>
8190 PR tree-optimization/58223
8191 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8192 (has_anti_or_output_dependence): ... this and adjust to also
8193 look for output dependences.
8194 (mark_nodes_having_upstream_mem_writes): Adjust.
8195 (rdg_flag_uses): Likewise.
8197 2013-08-30 Richard Biener <rguenther@suse.de>
8199 PR tree-optimization/58010
8200 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8201 assert that we have a loop-closed PHI.
8203 2013-08-29 Jan Hubicka <jh@suse.cz>
8205 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8206 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8207 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8208 (lto_free_function_in_decl_state_for_node): New function.
8210 2013-08-29 Xinliang David Li <davidxl@google.com>
8212 * loop-unroll.c (report_unroll_peel): Minor message change.
8213 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8214 Emit alignment peeling message with default -fopt-info.
8215 (vect_loop_versioning): Emit loop version info message.
8216 * tree-vectorizer.c (vectorize_loops): Minor message change.
8217 (execute_vect_slp): Ditto.
8219 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8221 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8222 of the clone from the DECL_NAME of the original function.
8224 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8226 * passes.c (register_pass): Add overload.
8227 * tree-pass.h (register_pass): Forward declare it. Add comment.
8229 2013-08-29 Jan Hubicka <jh@suse.cz>
8231 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8232 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8233 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8234 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8235 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8236 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8237 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8238 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8240 2013-08-29 Teresa Johnson <tejohnson@google.com>
8242 * dumpfile.c (dump_loc): Output column number.
8243 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8244 * doc/invoke.texi: Document optall -fopt-info flag.
8245 * profile.c (read_profile_edge_counts): Use new dump framework.
8246 (compute_branch_probabilities): Ditto.
8247 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8248 when pass not in any opt group.
8249 * pass_manager.h (pass_manager::get_pass_profile): New method.
8250 * value-prof.c (check_counter): Use new dump framework.
8251 (check_ic_target): Ditto.
8252 * coverage.c (get_coverage_counts): Ditto.
8253 (coverage_init): Setup new dump framework.
8255 2013-08-29 Richard Biener <rguenther@suse.de>
8257 PR tree-optimization/58246
8258 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8259 handle the dominance check inside a basic-block.
8261 2013-08-29 Richard Biener <rguenther@suse.de>
8264 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8265 of default defs that appear in abnormal PHI nodes.
8267 2013-08-29 Richard Biener <rguenther@suse.de>
8269 PR tree-optimization/57685
8270 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8271 single-use operands to avoid exponential complexity.
8273 2013-08-28 Dehao Chen <dehao@google.com>
8275 * ipa-inline.c (edge_badness): Fix integer underflow.
8277 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8279 * gtm-builtins.def (_ITM_free): Declare leaf.
8281 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8284 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8285 (*tls_local_dynamic_base_64_largepic): Likewise.
8286 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8287 Remove predicate from call operand.
8288 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8289 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8291 2013-08-28 Jeff Law <law@redhat.com>
8293 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8294 checks for the number of predecessors and successors allowed.
8295 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8296 which require copying a joiner block if there is a request which
8297 is a subpath that requires no joiner block copying.
8299 2013-08-28 Jan Hubicka <jh@suse.cz>
8301 * lto-streamer-out.c (DFS_write_tree_body): Drop
8302 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8303 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8304 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8305 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8306 Do not read DECL_ERROR_ISSUED.
8307 (unpack_ts_decl_with_vis_value_fields): Do not read
8309 (lto_input_ts_binfo_tree_pointers): Do not read
8310 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8311 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8312 write DECL_ERROR_ISSUED..
8313 (pack_ts_decl_with_vis_value_fields): Do not write
8315 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8316 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8317 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8318 * tree.h (tree_decl_common): Update comment.
8319 (DECL_ERROR_ISSUED): Remove.
8321 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8324 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8326 2013-08-28 Jan Hubicka <jh@suse.cz>
8328 * builtins.def (free): Declare leaf.
8330 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8333 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8334 * configure: Regenerate.
8336 2013-08-27 Martin Jambor <mjambor@suse.cz>
8338 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8339 (ipa_ancestor_jf_data): Likewise.
8340 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8341 (ipa_get_jf_pass_through_type_preserved): New function.
8342 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8343 (ipa_get_jf_ancestor_type_preserved): New function.
8344 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8345 (ipa_get_jf_ancestor_result): Likewise.
8346 (propagate_vals_accross_pass_through): Use
8347 ipa_get_jf_pass_through_result to do all the value mappings.
8348 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8349 type_preserved flag.
8350 (ipa_set_jf_cst_copy): New function.
8351 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8352 (ipa_set_jf_arith_pass_through): Likewise.
8353 (ipa_set_ancestor_jf): Likewise.
8354 (compute_complex_assign_jump_func): Set type_preserved instead of
8356 (ipa_compute_jump_functions_for_edge): Likewise.
8357 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8358 (update_jump_functions_after_inlining): Update type_preserved.
8359 Explicitely create jump functions when combining one with pass_through.
8360 (ipa_write_jump_function): Stream the type_preserved flags.
8361 (ipa_read_jump_function): Likewise.
8363 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8364 Aldy Hernandez <aldyh@redhat.com>
8366 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8367 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8368 * function.h (struct function): Add has_force_vect_loops and
8370 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8371 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8373 * gimple.def: Update CLAUSES comments.
8374 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8375 (gimple_build_omp_for): Add argument to prototype.
8376 (gimple_omp_for_kind): New.
8377 (gimple_omp_for_set_kind): New.
8378 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8379 GOVD_DATA_SHARE_CLASS.
8380 (enum omp_region_type): Add ORT_SIMD.
8381 (gimple_add_tmp_var): Handle ORT_SIMD.
8382 (gimplify_var_or_parm_decl): Same.
8383 (is_gimple_stmt): Same.
8384 (omp_firstprivatize_variable): Same.
8385 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8386 (omp_notice_variable): Handle ORT_SIMD.
8387 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8388 (omp_check_private): Handle ORT_SIMD.
8389 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8391 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8392 Handle OMP_CLAUSE_LASTPRIVATE.
8393 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8395 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8396 (gimplify_expr): Handle OMP_SIMD.
8397 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8398 (expand_GOMP_SIMD_VF): New.
8399 (expand_GOMP_SIMD_LAST_LANE): New.
8400 * internal-fn.def (GOMP_SIMD_LANE): New.
8401 (GOMP_SIMD_VF): New.
8402 (GOMP_SIMD_LAST_LANE): New.
8403 * omp-low.c: Include target.h.
8404 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8406 (check_omp_nesting_restrictions): Same.
8408 (lower_rec_simd_input_clauses): New.
8409 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8411 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8412 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8413 (expand_omp_build_assign): New.
8414 (expand_omp_for_init_counts): New.
8415 (expand_omp_for_init_vars): New.
8416 (extract_omp_for_update_vars): New.
8417 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8418 and rewrite accordingly.
8419 (expand_omp_simd): New.
8420 (expand_omp_for): Use expand_omp_simd.
8421 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8422 (lower_omp_for): Do not lower the body.
8423 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8425 * tree-flow.h (find_omp_clause): Remove prototype.
8426 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8427 forcing vectorization of the loop, or if flag_tree_vectorize.
8428 (gate_tree_if_conversion): Similarly.
8429 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8430 gimple_build_omp_for.
8431 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8432 * tree-parloops (create_parallel_loop): Pass kind argument to
8433 gimple_build_omp_for.
8434 * tree-pretty-print.c (dump_omp_clause): Add cases for
8435 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8436 OMP_CLAUSE__SIMDUID_.
8437 (dump_generic_node): Handle OMP_SIMD.
8438 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8439 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8440 unroll OMP_SIMD loops here.
8441 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8442 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8444 (vect_analyze_data_refs): Handle simd loops.
8445 * tree-vect-loop.c (vectorizable_live_operation): Handle
8447 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8448 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8449 (vectorizable_load): Same.
8450 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8451 (struct simduid_to_vf): New.
8452 (simduid_to_vf::hash): New.
8453 (simduid_to-vf::equal): New.
8454 (struct simd_array_to_simduid): New.
8455 (simd_array_to_simduid::hash): New.
8456 (simd_array_to_simduid::equal): New.
8457 (adjust_simduid_builtins): New.
8458 (struct note_simd_array_uses_struct): New.
8459 (note_simd_array_uses_cb): New.
8460 (note_simd_array_uses): New.
8461 (vectorize_loops): Handle simd hints and adjust simd builtins
8463 * tree-vectorizer.h (struct _stmt_vec_info): Add
8464 simd_lane_access_p field.
8465 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8466 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8467 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8468 (omp_clause_code_name): Same.
8469 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8470 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8471 * tree.def (OMP_SIMD): New entry.
8472 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8473 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8474 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8475 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8476 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8477 (OMP_CLAUSE_LINEAR_STEP): New.
8478 (OMP_CLAUSE_SAFELEN_EXPR): New.
8479 (OMP_CLAUSE__SIMDUID__DECL): New.
8480 (find_omp_clause): New prototype.
8482 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8484 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8485 Haswell processor detection.
8487 2013-08-27 Christian Widmer <shadow@umbrox.de>
8490 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8491 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8492 AVX2 capable processors.
8494 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8496 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8497 of vget_low_* with implementations in terms of other intrinsics.
8499 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8502 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8503 values to -1, 0 and 1.
8505 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8507 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8508 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8509 (<optab><mode>3_insn): Remove.
8510 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8511 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8512 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8513 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8514 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8515 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8516 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8517 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8518 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8519 (ror<mode>3_insn): Likewise.
8520 * config/aarch64/predicates.md (aarch64_simd_register): New.
8522 2013-08-27 Richard Biener <rguenther@suse.de>
8524 PR tree-optimization/57521
8525 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8526 one edge is non-critical.
8527 (find_phi_replacement_condition): Make sure to use a non-critical
8528 edge. Cleanup and remove old bug workarounds.
8529 (bb_postdominates_preds): Remove.
8530 (if_convertible_loop_p_1): Do not compute post-dominators.
8531 (combine_blocks): Do not free post-dominators.
8532 (main_tree_if_conversion): Likewise.
8533 (pass_data_if_conversion): Add TODO_verify_ssa.
8535 2013-08-27 DJ Delorie <dj@redhat.com>
8537 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8539 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8541 * function.c (assign_parm_find_data_types): Set passed_mode and
8542 nominal_mode to the TYPE_MODE of nominal_type for the built
8543 pointer type in case of the struct-pass-by-reference.
8545 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8547 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8548 (UINT16_TYPE): Change default to "unsigned int".
8550 * config/avr/avr.opt (mfract-convert-truncate): New option.
8551 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8552 is set, round negative fractional integers according to n1169
8553 when converting to integer types.
8555 2013-08-26 Jan Hubicka <jh@suse.cz>
8557 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8558 methods can not be called indirectly when their address is not taken.
8560 2013-08-26 Jan Hubicka <jh@suse.cz>
8562 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8565 2013-08-26 Jan Hubicka <jh@suse.cz>
8567 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8570 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8572 * reload.c (find_valid_class): Allow classes that do not include
8573 FIRST_PSEUDO_REGISTER - 1.
8575 2013-08-26 Jan Hubicka <jh@suse.cz>
8577 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8578 fix edge count/frequency when speculation failed; fix type check
8579 for the direct call.
8581 2013-08-26 Jan Hubicka <jh@suse.cz>
8583 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8585 2013-08-26 Jan Hubicka <jh@suse.cz>
8587 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8588 to be changed by edge redirection.
8590 2013-08-26 Jan Hubicka <jh@suse.cz>
8592 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8593 formating; add sanity check.
8594 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8595 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8596 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8597 (resolve_noninline_speculation): Update callee keys, too.
8599 2013-08-26 Jan Hubicka <jh@suse.cz>
8601 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8602 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8604 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8606 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8609 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8611 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8612 (x86_64_elf_select_section): Ditto.
8613 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8615 (ix86_pass_by_reference): Ditto.
8616 (output_set_got): Ditto.
8617 (ix86_unary_operator_ok): Ditto.
8618 (ix86_expand_builtin): Ditto.
8620 2013-08-23 Jan Hubicka <jh@suse.cz>
8622 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8624 2013-08-23 Jan Hubicka <jh@suse.cz>
8626 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8627 (tree_decl_with_vis): Add FINAL field.
8629 2013-08-23 Jeff Law <law@redhat.com>
8631 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8632 the debugging dump when the expression is fully redundant.
8634 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8636 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8637 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8638 * pretty-print.c (pp_formatted_text_data): Likewise.
8639 (pp_write_text_to_stream): Likewise.
8640 (pp_write_text_as_dot_label_to_stream): Likewise.
8641 (pp_append_r): Likewise.
8642 (pp_format): Likewise.
8643 (pp_flush): Likewise.
8644 (pp_clear_output_area): Likewise.
8645 (pp_append_text): Likewise.
8646 (pp_formatted_text): Likewise.
8647 (pp_remaining_character_count_for_line): Likewise.
8648 (pp_newline): Likewise.
8649 (pp_character): Likewise.
8650 (output_buffer::~output_buffer): Define.
8651 (pretty_printer::~pretty_printer): Destruct output buffer.
8652 * pretty-print.h (output_buffer::~output_buffer): Declare.
8653 (pretty_printer::~pretty_printer): Declare virtual.
8655 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8658 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8659 HOST_WIDE_INT_M1U): New macros.
8660 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8661 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8662 unsigned -1 for lshift.
8663 * cse.c (cse_insn): Likewise.
8664 * double-int.c (rshift_double, lshift_double): Likewise.
8665 * builtins.c (fold_builtin_bitop): Likewise.
8666 * combine.c (force_to_mode): Likewise.
8667 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8668 * simplify-rtx.c (simplify_const_unary_operation,
8669 simplify_const_binary_operation): Likewise.
8670 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8671 check_va_list_escapes): Likewise.
8672 * rtlanal.c (nonzero_bits1): Likewise.
8673 * expmed.c (expand_smod_pow2): Likewise.
8674 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8676 2013-08-23 Jan Hubicka <jh@suse.cz>
8678 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8679 as having address taken.
8681 2013-08-23 Jan Hubicka <jh@suse.cz>
8683 * ipa-utils.h (method_class_type): Declare.
8684 * ipa-devirt.c (method_class_type): Export.
8686 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8687 do not walk base classes of anonymous types.
8689 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8691 PR rtl-optimization/58220
8693 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8694 handle SEQUENCE insns properly.
8696 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8698 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8700 (pp_newline_and_indent): Likewise.
8701 (pp_separate_with): Likewise.
8702 * pretty-print.c (pp_newline_and_flush): Define.
8703 (pp_newline_and_indent): Likewise.
8704 (pp_separate_with): Likewise.
8706 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8709 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8710 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8712 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8714 * config/i386/predicates.md (ext_sse_reg_operand): New.
8715 * config/i386/i386.md (*movti_internal): Use
8716 predicate to determine if EVEX is needed.
8717 (*movsi_internal): Ditto.
8718 (*movdf_internal): Ditto.
8719 (*movsf_internal): Ditto.
8720 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8722 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8724 PR tree-optimization/58209
8725 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8726 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8727 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8728 emit POINTER_PLUS_EXPR.
8729 (create_tailcall_accumulator): For pointer result type accumulate in
8732 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8734 * configure.ac: Add backslashes missing from the last change.
8735 * configure: Regenerate.
8737 2013-08-22 Jan Hubicka <jh@susue.cz>
8739 * ipa.c (function_and_variable_visibility): First remember function
8740 was global and then make it local.
8742 2013-08-22 Julian Brown <julian@codesourcery.com>
8744 * configure.ac: Add aarch64 to list of arches which use "nop" in
8746 * configure: Regenerate.
8748 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8750 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8751 gnu_libc_has_function.
8752 * config/s390/tpf.h: Likewise.
8754 2013-08-22 Jan Hubicka <jh@susue.cz>
8756 * timevar.c (validate_phases): Add cast.
8758 2013-08-22 Jan Hubicka <jh@susue.cz>
8760 * timevar.c (validate_phases): Use size_t for memory.
8761 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8763 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8765 * pretty-print.h (output_buffer::output_buffer): Declare.
8766 (pretty_printer::pretty_printer): Likewise.
8767 (pp_construct): Remove.
8768 * pretty-print.c (output_buffer::output_buffer): Define.
8769 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8770 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8771 (print_gimple_expr): Likewise.
8772 (print_gimple_seq): Likewise.
8773 (gimple_dump_bb): Likewise.
8774 * sched-vis.c (dump_value_slim): Likewise.
8775 (dump_insn_slim): Likewise.
8776 (dump_rtl_slim): Likewise.
8777 (str_pattern_slim): Likewise.
8778 * tree-mudflap.c (mf_varname_tree): Likewise.
8779 * graph.c (print_graph_cfg): Likewise.
8780 (start_graph_dump): Likewise.
8781 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8783 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8784 pretty printer initialization.
8785 * coretypes.h (diagnostic_context): Remove superflous type alias
8787 (pretty_printer): Likewise. Declare directly as a class.
8788 (pretty_print_info): Remove declaration as class.
8789 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8790 and pp_clear_output_area.
8791 (asan_add_global): Likewise.
8793 2013-08-22 Jan Hubicka <jh@suse.cz>
8795 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8796 * ipa-utils.h (update_type_inheritance_graph): Declare.
8797 (possible_polymorphic_call_target_p): Declare.
8798 (possible_polymorphic_call_target_p): New.
8799 * ipa-devirt.c: Update toplevel comments.
8800 (cached_polymorphic_call_targets): Move up.
8801 (odr_type_d): Move ID down.
8802 (polymorphic_type_binfo_p): Update comment.
8803 (odr_hasher::remove): Likewise;
8804 (get_odr_type): Set anonymous_namespace.
8805 (dump_odr_type): Dump it.
8806 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8807 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8808 (record_binfo): Add comment.
8809 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8811 (devirt_node_removal_hook): Do not iCE when cache is freed.
8812 (possible_polymorphic_call_target_p): New predicate.
8813 (update_type_inheritance_graph): New function.
8815 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8816 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8817 Sergey Lega <sergey.s.lega@intel.com>
8818 Anna Tikhonova <anna.tikhonova@intel.com>
8819 Ilya Tocar <ilya.tocar@intel.com>
8820 Andrey Turetskiy <andrey.turetskiy@intel.com>
8821 Ilya Verbin <ilya.verbin@intel.com>
8822 Kirill Yukhin <kirill.yukhin@intel.com>
8823 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8825 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8826 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8827 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8828 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8829 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8830 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8831 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8832 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8833 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8834 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8835 OPT_mavx512pf, OPT_mavx512er cases.
8836 * config/i386/constraints.md (v): New constraint.
8837 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8838 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8839 (bit_AVX512CD): New.
8840 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8841 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8842 * config/i386/i386-c.c (ix86_target_macros_internal):
8843 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8845 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8846 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8847 * config/i386/i386.c (regclass_map, dbx_register_map)
8848 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8849 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8850 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8852 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8853 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8854 -mavx512pf options. Fix formatting.
8855 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8856 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8857 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8858 -mavx512cd, -mavx512pf options.
8859 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8860 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8861 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8863 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8864 (ix86_set_reg_reg_cost): Ditto.
8865 (x86_order_regs_for_local_alloc): Ditto.
8866 (MAX_VECT_LEN): Extend to 64-byte.
8867 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8868 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8869 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8870 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8871 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8872 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8873 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8874 (SSE_REG_MODE_P): Support new modes.
8875 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8876 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8877 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8878 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8879 (REG_CLASS_CONTENTS): Add new registers.
8880 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8881 (EXT_REX_SSE_REGNO_P): New.
8882 (HI_REGISTER_NAMES): Add new registers.
8883 * config/i386/i386.md: Define constants for new registers.
8884 (mode): Add new 512-bit modes.
8885 (prefix): Support evex prefix.
8886 (isa): Support avx512f, noavx512f, fma_avx512f.
8887 (ssemodesuffix): Add new 512-bit modes.
8889 (*movxi_internal_avx512f): Ditto.
8890 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8892 (*movsi_internal): Likewise.
8893 (*movdf_internal): Likewise.
8894 (*movsf_internal): Likewise.
8895 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
8896 (<code><mode>3): Likewise.
8897 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
8899 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
8900 with the new constraint "v".
8901 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
8903 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
8904 with the new constraint "v".
8905 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
8906 (<sse2>_storedqu<avxsizesuffix>): Likewise.
8907 (*<plusminus_insn><mode>3): Likewise.
8908 (<sse>_vm<plusminus_insn><mode>3): Likewise.
8909 (*mul<mode>3): Likewise.
8910 (<sse>_vmmul<mode>3): Likewise.
8911 (<sse>_div<mode>3): Likewise.
8912 (<sse>_vmdiv<mode>3): Likewise.
8913 (<sse>_sqrt<mode>2): Likewise.
8914 (<sse>_vmsqrt<mode>2): Likewise.
8915 (*<code><mode>3_finite): Likewise.
8916 (*<code><mode>3) <smaxmin>: Likewise.
8917 (<sse>_vm<code><mode>3): Likewise.
8918 (*<code><mode>3) <any_logic>: Likewise.
8919 (*fma_fmadd_<mode>): Likewise.
8920 (*fma_fmsub_<mode>): Likewise.
8921 (*fma_fnmadd_<mode>): Likewise.
8922 (*fma_fnmsub_<mode>): Likewise.
8923 (*fma_fmaddsub_<mode>): Likewise.
8924 (*fma_fmsubadd_<mode>): Likewise.
8925 (*fmai_fmadd_<mode>): Likewise.
8926 (*fmai_fmsub_<mode>): Likewise.
8927 (*fmai_fnmadd_<mode>): Likewise.
8928 (*fmai_fnmsub_<mode>): Likewise.
8929 (sse_cvtsi2ss): Likewise.
8930 (sse_cvtsi2ssq): Likewise.
8931 (sse_cvtss2si): Likewise.
8932 (sse_cvtss2si_2): Likewise.
8933 (sse_cvtss2siq): Likewise.
8934 (sse_cvtss2siq_2): Likewise.
8935 (sse_cvttss2si): Likewise.
8936 (sse_cvtss2siq_2): Likewise.
8937 (float<sseintvecmodelower><mode>2): Likewise.
8938 (sse2_cvtsd2si_2): Likewise.
8939 (sse2_cvtsd2siq_2): Likewise.
8940 (*<plusminus_insn><mode>3): Likewise.
8941 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
8942 (*<sse4_1_avx2>_mul<mode>3): Likewise.
8943 (ashr<mode>3): Likewise.
8944 (<shift_insn><mode>3): Likewise.
8945 (avx2_<code><mode>3): Likewise.
8946 (*avx2_<code><mode>3): Likewise.
8947 (*andnot<mode>3): Likewise.
8948 (*<code><mode>3) <any_logic>: Likewise.
8949 (abs<mode>2): Likewise.
8950 (avx2_permvar<mode>): Likewise.
8951 (avx2_perm<mode>_1): Likewise.
8952 (*avx_vpermilp<mode>): Likewise.
8953 (avx_vpermilvar<mode>3): Likewise.
8954 (avx2_ashrv<mode>): Likewise.
8955 (avx2_<shift_insn>v<mode>): Likewise.
8956 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
8958 * doc/rtl.texi: Document XImode.
8960 2013-08-21 Jeff Law <law@redhat.com>
8962 * tree-flow.h (register_jump_thread): Pass vector of edges
8963 instead of each important edge.
8964 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
8965 thread path into a vector and pass that to register_jump_thread.
8966 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
8967 passed in edge vector to the current 3-edge form.
8970 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8972 * dce.c (fini_dce): Call df_analyze again just in case
8973 delete_unmarked_insns removed anything.
8975 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8977 * reload.h (struct reg_equivs): Rename to ..
8978 (struct reg_equivs_s): .. this.
8980 2013-08-20 Martin Liska <marxin.liska@gmail.com>
8982 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8984 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8986 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
8988 2013-08-21 Jeff Law <law@redhat.com>
8990 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8991 simplify assignments too. If the RHS collapses to a singleton
8992 range, then return the value for the range.
8994 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8996 * config/i386/sse.md (V16): Rename to...
8998 (mov<mode>): Update iterator name.
8999 (*mov<mode>_internal): Ditto.
9000 (push<mode>1): Ditto.
9001 (movmisalign<mode>): Ditto.
9003 2013-08-20 Jan Hubicka <jh@suse.cz>
9006 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9007 entry for direct edges.
9008 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9010 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9012 Revert my last two changes, r201865 and r201864:
9015 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9017 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9018 instances can own GC refs.
9020 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9021 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9022 (gcc::context::gt_pch_nx): Likewise.
9023 (gcc::context::gt_pch_nx): Likewise.
9024 * ggc.h (gt_ggc_mx <T>): New.
9025 (gt_pch_nx_with_op <T>): New.
9026 (gt_pch_nx <T>): New.
9027 * passes.c (opt_pass::gt_ggc_mx): New.
9028 (opt_pass::gt_pch_nx): New.
9029 (opt_pass::gt_pch_nx_with_op): New.
9030 (pass_manager::gt_ggc_mx): New.
9031 (pass_manager::gt_pch_nx): New.
9032 (pass_manager::gt_pch_nx_with_op): New.
9033 (pass_manager::operator new): Use
9034 ggc_internal_cleared_alloc_stat rather than xcalloc.
9035 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9036 (pass_manager::gt_ggc_mx): New.
9037 (pass_manager::gt_pch_nx): New.
9038 (pass_manager::gt_pch_nx_with_op): New.
9039 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9040 (opt_pass::operator new): New.
9041 (opt_pass::gt_ggc_mx): New.
9042 (opt_pass::gt_pch_nx): New.
9043 (opt_pass::gt_pch_nx_with_op): New.
9046 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9048 * Makefile.in (GTFILES): Add context.h.
9049 * context.c (gcc::context::operator new): New.
9050 (gcc::context::gt_ggc_mx): New.
9051 (gcc::context::gt_pch_nx): New.
9052 (gcc::context::gt_pch_nx): New.
9053 * context.h (gcc::context): Add GTY((user)) marking.
9054 (gcc::context::operator new): New.
9055 (gcc::context::gt_ggc_mx): New.
9056 (gcc::context::gt_pch_nx): New.
9057 (gcc::context::gt_pch_nx): New.
9058 (g): Add GTY marking.
9059 (gt_ggc_mx (gcc::context *)): New.
9060 (gt_pch_nx (gcc::context *)): New.
9061 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9062 void *cookie)): New.
9063 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9065 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9067 * dce.c (fini_dce): Call df_analyze again just in case
9068 delete_unmarked_insns removed anything.
9070 2013-08-20 Teresa Johnson <tejohnson@google.com>
9072 PR rtl-optimizations/57451
9073 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9074 from crossing split section boundaries.
9076 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9078 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9079 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9080 with MULTLIB_DEFAULTS.
9082 2013-08-20 Nick Clifton <nickc@redhat.com>
9084 * target.def (narrow_volatile_bitfield): Note that the default
9085 value is false, not !TARGET_STRICT_ALIGN.
9086 * doc/tm.texi: Regenerate.
9088 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9090 Fix LIB_SPEC for systems without libpthread.
9092 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9093 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9095 * config/i386/linux-common.h: Likewise.
9096 * config/mips/linux-common.h: Likewise.
9098 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9100 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9103 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9105 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9106 instances can own GC refs.
9108 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9109 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9110 (gcc::context::gt_pch_nx): Likewise.
9111 (gcc::context::gt_pch_nx): Likewise.
9112 * ggc.h (gt_ggc_mx <T>): New.
9113 (gt_pch_nx_with_op <T>): New.
9114 (gt_pch_nx <T>): New.
9115 * passes.c (opt_pass::gt_ggc_mx): New.
9116 (opt_pass::gt_pch_nx): New.
9117 (opt_pass::gt_pch_nx_with_op): New.
9118 (pass_manager::gt_ggc_mx): New.
9119 (pass_manager::gt_pch_nx): New.
9120 (pass_manager::gt_pch_nx_with_op): New.
9121 (pass_manager::operator new): Use
9122 ggc_internal_cleared_alloc_stat rather than xcalloc.
9123 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9124 (pass_manager::gt_ggc_mx): New.
9125 (pass_manager::gt_pch_nx): New.
9126 (pass_manager::gt_pch_nx_with_op): New.
9127 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9128 (opt_pass::operator new): New.
9129 (opt_pass::gt_ggc_mx): New.
9130 (opt_pass::gt_pch_nx): New.
9131 (opt_pass::gt_pch_nx_with_op): New.
9133 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9135 * Makefile.in (GTFILES): Add context.h.
9136 * context.c (gcc::context::operator new): New.
9137 (gcc::context::gt_ggc_mx): New.
9138 (gcc::context::gt_pch_nx): New.
9139 (gcc::context::gt_pch_nx): New.
9140 * context.h (gcc::context): Add GTY((user)) marking.
9141 (gcc::context::operator new): New.
9142 (gcc::context::gt_ggc_mx): New.
9143 (gcc::context::gt_pch_nx): New.
9144 (gcc::context::gt_pch_nx): New.
9145 (g): Add GTY marking.
9146 (gt_ggc_mx (gcc::context *)): New.
9147 (gt_pch_nx (gcc::context *)): New.
9148 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9149 void *cookie)): New.
9150 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9152 2013-08-20 Alan Modra <amodra@gmail.com>
9155 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9156 (rs6000_emit_epilogue): Likewise.
9158 2013-08-19 Dehao Chen <dehao@google.com>
9160 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9162 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9163 Jakub Jelinek <jakub@redhat.com>
9165 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9166 (BUILT_IN_FABSD64): Likewise.
9167 (BUILT_IN_FABSD128): Likewise.
9168 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9169 (fold_builtin_1): Likewise.
9170 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9171 destination and source operands.
9172 (*abstd2_fpr): Likewise.
9173 (*nabstd2_fpr): Likewise.
9175 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9177 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9180 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9182 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9185 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9187 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9188 * builtins.c (default_libc_has_function): New.
9189 (gnu_libc_has_function): Ditto.
9190 (no_c99_libc_has_function): Ditto.
9191 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9192 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9193 (fold_builtin_sincos): Likewise.
9194 (fold_builtin_cexp): Likewise.
9195 * builtins.def (DEF_C94_BUILTIN): Likewise.
9196 (DEF_C99_BUILTIN): Likewise.
9197 (DEF_C99_C90RES_BUILTIN): Likewise.
9198 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9199 definitions to using this define.
9200 * config/darwin-protos.h (darwin_libc_has_function): New.
9201 * config/darwin.c (darwin_libc_has_function): Ditto.
9202 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9203 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9204 * config/darwin.h: Ditto.
9205 * config/elfos.h: Ditto.
9206 * config/freebsd.h: Ditto.
9207 * config/i386/cygming.h: Ditto.
9208 * config/i386/djgpp.h: Ditto.
9209 * config/i386/i386-interix.h: Ditto.
9210 * config/microblaze/microblaze.h: Ditto.
9211 * config/mmix/mmix.h: Ditto.
9212 * config/gnu-user.h: Ditto.
9213 * config/ia64/hpux.h: Ditto.
9214 * config/pa/pa-hpux.h: Ditto.
9215 * config/pdp11/pdp11.h: Ditto.
9216 * config/picochip/picochip.h: Ditto.
9217 * config/linux.h: Ditto.
9218 * config/netbsd.h: Ditto.
9219 * config/openbsd.h: Ditto.
9220 * config/rs6000/aix43.h: Ditto.
9221 * config/rs6000/aix51.h: Ditto.
9222 * config/rs6000/aix52.h: Ditto.
9223 * config/rs6000/aix53.h: Ditto.
9224 * config/rs6000/aix61.h: Ditto.
9225 * config/rs6000/darwin.h: Ditto.
9226 * config/rs6000/linux.h: Ditto.
9227 * config/rs6000/linux64.h: Ditto.
9228 * config/s390/tpf.h: Ditto.
9229 * config/sol2-10.h: Ditto.
9230 * config/sol2.h: Ditto.
9231 * config/vms/vms.h: Ditto.
9232 * config/vxworks.h: Ditto.
9233 * config/linux-android.c (linux_android_libc_has_function):
9234 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9235 * config/linux-protos.h (linux_android_libc_has_function):
9237 * config/i386/i386.c (ix86_libc_has_function): New.
9238 * config/i386/i386-protos.h
9239 (ix86_libc_has_function): New declaration.
9240 * config/i386/i386.md
9241 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9242 ("isinf<mode>2): Likewise.
9243 * convert.c (convert_to_integer): Using new target hook
9244 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9245 TARGET_C99_FUNCTIONS.
9246 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9247 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9248 * coretypes.h (function_class): New enum for different
9249 classes of functions.
9250 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9251 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9252 (TARGET_HAS_SINCOS): Likewise.
9253 (TARGET_LIBC_HAS_FUNCTION): New.
9254 * doc/tm.texi: Regenerated.
9255 * targhooks.h (default_libc_has_function): New declaration.
9256 (no_c99_libc_has_function): Ditto.
9257 (gnu_libc_has_function): Ditto.
9258 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9259 and TARGET_HAS_SINCOS.
9261 2013-08-18 Jan Hubicka <jh@suse.cz>
9263 * Makeifle-in (ipa-devirt.o): New.
9264 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9265 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9266 (analyze_functions): Look into possible targets of polymorphic call.
9267 * dumpfile.c (dump_files): Add type-inheritance dump.
9268 * dumpfile.h (TDI_inheritance): New.
9269 * ipa-devirt.c: New file.
9270 * ipa-utils.h (odr_type_d): Forward declare.
9271 (odr_type): New type.
9272 (build_type_inheritance_graph): Declare.
9273 (possible_polymorphic_call_targets): Declare and introduce inline
9274 variant when only edge is pased.
9275 (dump_possible_polymorphic_call_targets): Likewise.
9276 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9277 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9278 (types_same_for_odr): ... here.
9279 * tree.h (type_in_anonymous_namespace_p): Declare.
9281 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9283 PR tree-optimization/58006
9284 * tree-parloops.c (take_address_of): Don't ICE if get_name
9286 (eliminate_local_variables_stmt): Remove clobber stmts.
9288 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9290 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9291 error message is issued for an alias to undefined symbol.
9293 2013-08-18 Jan Hubicka <jh@suse.cz>
9295 * cgraph.c (cgraph_create_indirect_edge): Discover
9296 polymorphic calls and record basic info into indirect_info.
9297 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9298 devirtualization, ignore objc function calls.
9299 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9300 call with no parm index info.
9301 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9302 * tree.c (virtual_method_call_p): New function.
9303 * tree.h (virtual_method_call_p): Declare.
9305 2013-08-16 Jan Hubicka <jh@suse.cz>
9308 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9310 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9312 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9315 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9317 * gengtype.c (type_for_name): Add special-case support for
9318 locating types within the "gcc::" namespace.
9319 (open_base_files): Emit a "using namespace gcc" directive.
9321 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9324 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9325 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9327 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9328 array instead of each individual operand as a separate argument.
9329 (emit_fusion_gpr_load): Likewise.
9330 (expand_fusion_gpr_load): Add new function declaration.
9332 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9333 signature to have the operands passed as an array, instead of as
9334 separate arguments. Allow ZERO_EXTEND to be in the memory
9335 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9336 depend on the register live/dead flags when peepholes are run.
9337 (expand_fusion_gpr_load): New function to be called from the
9338 peephole2 pass, to change the register that addis sets to be the
9340 (emit_fusion_gpr_load): Change the calling signature to have the
9341 operands passed as an array, instead of as separate arguments.
9342 Allow ZERO_EXTEND to be in the memory address, and also
9343 SIGN_EXTEND if -mpower8-fusion-sign.
9345 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9347 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9348 adjust the register addis loads up in the peephole2 pass. Do not
9349 depend on the register live/dead state when the peephole pass is done.
9351 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9353 * gengtype.c (create_user_defined_type): Ensure that the kind
9354 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9355 declaration is seen before the GTY((user)) marking.
9357 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9360 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9362 2013-08-16 Jan Hubicka <jh@suse.cz>
9364 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9365 arugment expected_type.
9366 (gimple_fold_call): Use it.
9367 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9368 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9369 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9370 (try_make_edge_direct_virtual_call): Likewise.
9371 * tree.c (obj_type_ref_class): New.
9372 * tree.h (obj_type_ref_class): Use it.
9374 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9376 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9377 (rtl_slim_pp): Likewise.
9378 (init_rtl_slim_pretty_print): Likewise.
9379 (dump_value_slim): Don't call it. Use local pretty printer.
9380 (dump_insn_slim): Likewise.
9381 (dump_rtl_slim): Likewise.
9382 (str_pattern_slim): Likewise.
9383 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9386 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9388 PR tree-optimization/58164
9389 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9390 walk gimple_goto_dest of GIMPLE_GOTO.
9392 PR tree-optimization/58165
9393 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9394 bi_call must be the last stmt in a bb, don't split_block, instead
9395 use fallthru edge from it and give up if there is none.
9396 Release conds vector when returning early.
9398 2013-08-14 Xinliang David Li <davidxl@google.com>
9400 * config/i386/i386.c (ix86_option_override_internal):
9401 Remove unused variable and field.
9403 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9406 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9407 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9408 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9409 and Linux, correct BLKmode alignment when 128-bit alignment is
9410 required and compatibility flag is not set.
9411 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9412 for zero-size arguments when compatibility flag is not set.
9414 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9416 PR tree-optimization/58145
9417 * tree-sra.c (build_ref_for_offset): If prev_base has
9418 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9420 2013-08-14 Xinliang David Li <davidxl@google.com>
9422 * config/i386/i386.c (ix86_option_override_internal):
9423 Fix uninitialized variable error.
9425 2013-08-14 Xinliang David Li <davidxl@google.com>
9427 * config/i386/i386.opt: Define two new options.
9428 * config/i386/x86-tune.def: Add arch selector field in macros.
9429 * config/i386/i386.h: Adjust macro definition.
9430 * config/i386/i386.c (ix86_option_override_internal):
9432 (parse_mtune_ctrl_str): New function.
9433 (set_ix86_tune_features): New function.
9434 (ix86_function_specific_restore): Call the new helper function.
9436 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9438 PR rtl-optimization/57662
9439 * sel-sched.c (code_motion_process_successors): When the current insn
9440 is removed after the recursive traversal, break from the loop.
9441 Add comments and debug printouts.
9443 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9444 Alexandre Oliva <aoliva@redhat.com>
9447 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9448 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9449 there also UNSPEC_PLTOFF.
9451 2013-08-14 Marek Polacek <polacek@redhat.com>
9453 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9454 NUM_CONDITIONS bit positions.
9456 2013-08-13 Cary Coutant <ccoutant@google.com>
9458 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9459 (attr_checksum): Hash vector contents instead of pointer.
9460 (attr_checksum_ordered): Likewise.
9462 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9464 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9465 when Pmode != word_mode. Add length_address attribute.
9466 (sse3_monitor_<mode>): Merge from sse3_monitor and
9467 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9468 Pmode != word_mode. Update insn length attribute.
9469 * config/i386/i386.c (ix86_option_override_internal): Update
9470 ix86_gen_monitor selection for merged sse3_monitor insn.
9472 2013-08-13 Julian Brown <julian@codesourcery.com>
9474 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9475 perform invalid legitimization on greater-than-word-size modes for
9478 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9480 * ira.c (setup_class_translate_array): Use aclass instead of cl
9481 for classes not fully covered by allocno classes.
9483 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9485 PR tree-optimization/57661
9486 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9487 * tree-inline.c (tree_function_versioning): Initialize it.
9488 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9489 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9490 that is not being copied.
9493 * asan.c (instrument_strlen_call): Fix typo in comment.
9494 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9496 2013-08-13 Steve Ellcey <sellcey@mips.com>
9498 * config/mips/mips.md (prefetch): Use lw instead of ld on
9499 loongson in 32bit mode.
9501 2013-08-13 Nick Clifton <nickc@redhat.com>
9503 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9505 2013-08-13 Jan Hubicka <jh@suse.cz>
9507 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9508 introduced edge; fix typo in sanity check.
9509 (cgraph_resolve_speculation): Export; improve diagnostic.
9510 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9511 speculation at type mismatch.
9512 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9513 (cgraph_resolve_speculation): Declare.
9514 (symtab_can_be_discarded): New function.
9515 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9516 * ipa-inline-transform.c (speculation_removed): New global var.
9517 (clone_inlined_nodes): See if speculation can be removed.
9518 (inline_call): If speculations was removed, we growths may not match.
9519 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9520 (speculation_useful_p): New function.
9521 (resolve_noninline_speculation): New function.
9522 (inline_small_functions): Resolve useless speculations.
9523 * ipa-inline.h (speculation_useful_p): Declare
9524 * ipa.c (can_replace_by_local_alias): Simplify.
9525 (ipa_profile): Produce speculative calls in non-lto, too;
9526 add simple cost model; produce local aliases.
9528 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9530 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9533 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9535 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9537 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9539 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9540 iterator instead of X87MODEF.
9542 2013-08-12 Perez Read <netfirewall@gmail.com>
9545 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9546 operand 0 for intel asm alternative.
9547 (*movabs<mode>_2): Ditto for operand 1.
9549 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9551 * config/aarch64/arm_none.h
9552 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9554 2013-08-12 Nick Clifton <nickc@redhat.com>
9556 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9558 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9560 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9561 expand for QI/HImode operand to produce more effictive code for
9562 unsigned char(short) --> float(double) conversion.
9564 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9566 * doc/invoke.texi: Mention that -ftls-model does not force the final
9569 2013-08-12 Marek Polacek <polacek@redhat.com>
9570 Marc Glisse <marc.glisse@inria.fr>
9572 PR tree-optimization/57980
9573 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9574 when creating -1 constant.
9576 2013-08-10 Jan Hubicka <jh@suse.cz>
9578 Workaround binutils PR14342.
9579 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9580 (gimple_init_edge_profiler): Likewise.
9581 (gimple_gen_ic_func_profiler): Likewise.
9583 2013-08-09 Jan Hubicka <jh@suse.cz>
9585 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9587 2013-08-09 Xinliang David Li <davidxl@google.com>
9589 * config/i386/stringop.def: New file.
9590 * config/i386/stringop.opt: New file.
9591 * config/i386/i386-opts.h: Include stringopt.def.
9592 * config/i386/i386.opt: Include stringopt.opt.
9593 * config/i386/i386.c (ix86_option_override_internal):
9594 Override default size based stringop inline strategies with options.
9595 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9598 2013-08-09 Jan Hubicka <jh@suse.cz>
9600 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9602 2013-08-09 Jan Hubicka <jh@suse.cz>
9604 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9606 (dump_cgraph_node): Dump profile-id.
9607 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9608 and common_target_probability.
9609 * lto-cgraph.c (lto_output_edge): Stream common targets.
9610 (lto_output_node): Stream profile ids.
9611 (input_node): Stream profile ids.
9612 (input_edge): Stream common targets.
9613 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9614 * ipa.c: Include value-prof.h
9615 (ipa_profile_generate_summary): Turn indirect call statement histograms
9616 into common targets.
9617 (ipa_profile): Turn common targets into speculative edges.
9619 2013-08-09 Jan Hubicka <jh@suse.cz>
9621 * cgraph.h (cgraph_node): Add profile_id.
9622 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9623 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9624 (del_node_map): Update.
9625 (find_func_by_funcdef_no): Replace by ...
9626 (find_func_by_profile_id): ... this one.
9627 (gimple_ic_transform): Do not remove useful histograms when
9628 speculation is not done; dump info when indirect call removal
9630 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9631 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9632 (__gcov_indirect_call_profiler_v2): .. this one.
9633 * profile.h (init_node_map): Update.
9634 * coverage.c (coverage_compute_profile_id): New function.
9635 * coverage.h (coverage_compute_profile_id): Declare.
9636 * tree-profile.c (init_ic_make_global_vars): Make
9637 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9638 (gimple_init_edge_profiler): Update prototype of
9639 __gcov_indirect_call_profiler.
9640 (gimple_gen_ic_func_profiler): Simplify.
9641 (tree_profiling): Use init_node_map
9643 2013-08-09 Jan Hubicka <jh@suse.cz>
9645 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9646 non-speculative refs.
9647 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9648 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9649 (cgraph_set_call_stmt): Likewise.
9650 (cgraph_create_edge_1): Fix release checking compilatoin;
9652 (cgraph_free_edge): Free indirect info.
9653 (cgraph_turn_edge_to_speculative): New function.
9654 (cgraph_speculative_call_info): New function.
9655 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9656 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9657 (dump_cgraph_node): Dump speculation.
9658 (verify_edge_count_and_frequency): Accept speculative edges.
9659 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9660 (verify_cgraph_node): Handle speculation.
9661 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9662 (cgraph_set_call_stmt): Update prototype.
9663 (cgraph_make_edge_direct): Update prototype.
9664 (cgraph_speculative_call_info): Declare.
9665 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9666 to change; update call of ipa_find_references.
9667 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9668 lto_stmt_uid and speculative flags.
9669 (ipa_dump_references): Dump speculation.
9670 (ipa_clone_references): Clone speculative flag.
9671 (ipa_clone_referring): Likewise.
9672 (ipa_clone_ref): New function.
9673 (ipa_find_reference): Look into lto_stmt_uids
9674 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9675 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9676 (ipa_find_reference): Update declaration.
9677 (ipa_clone_ref): Declare.
9678 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9679 stream speculative flag.
9680 (lto_output_ref): Stream statements uids and speculation.
9681 (input_ref): Likewise.
9682 (input_edge): Stream speuclation.
9683 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9684 (cgraph_set_call_stmt_including_clones): Handle speculation.
9685 * ipa-inline.c (heap_edge_removal_hook): New function.
9686 (inline_small_functions): Register it.
9687 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9688 also initialize refs.
9689 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9691 (try_make_edge_direct_simple_call): Likewise.
9692 (try_make_edge_direct_simple_call): Likewise.
9693 (update_indirect_edges_after_inlining): Likewise.
9694 (remove_described_reference): Look proper lto_stmt_uid.
9695 (propagate_controlled_uses): Likewise.
9696 (propagate_controlled_uses): Liekwise.
9697 * tree-inline.c (copy_bb): Copy speculative edges.
9698 (redirect_all_calls): New function.
9699 (copy_cfg_body): Do redirection after loop info is updated.
9700 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9702 2013-08-09 Jan Hubicka <jh@suse.cz>
9704 * lto-streamer-out.c (output_function): Renumber PHIs.
9705 * lto-streamer-in.c (input_function): Likewise.
9707 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9709 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9710 (get_lane_unsigned): Likewise.
9711 (dup_lane_scalar): Likewise.
9712 (get_lane): enable for VALL.
9713 * config/aarch64/aarch64-simd.md
9714 (aarch64_dup_lane_scalar<mode>): Remove.
9715 (aarch64_get_lane_signed<mode>): Likewise.
9716 (aarch64_get_lane_unsigned<mode>): Likewise.
9717 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9718 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9719 (aarch64_get_lane<mode>): Enable for all vector modes.
9720 (aarch64_get_lanedi): Remove misleading constraints.
9721 * config/aarch64/arm_neon.h
9722 (__aarch64_vget_lane_any): Define.
9723 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9724 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9725 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9726 * config/aarch64/iterators.md (VDQQH): New.
9730 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9732 * configure.ac: Add GAS check for LEON instructions on SPARC.
9733 * configure: Regenerate.
9734 * config.in: Likewise.
9735 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9737 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9738 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9740 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9741 (AS_LEON_FLAG): New macro.
9742 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9743 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9744 Deal with LEON and LEON3 for the memory model.
9745 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9746 (atomic_compare_and_swap<mode>_1): Likewise.
9747 (*atomic_compare_and_swap<mode>_1): Likewise.
9749 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9751 * config/arm/neon.md (vcond): Fix floating-point vector
9752 comparisons against 0.
9754 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9756 * lra-constraints.c (emit_spill_move): Remove assert.
9757 (process_alt_operands): Add more debugging
9758 output. Increase reject for spilling into memory. Decrease
9759 reject for reloading scratch.
9760 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9762 2013-08-08 Steve Ellcey <sellcey@mips.com>
9764 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9765 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9766 micromips incompatible. Add nan2008.
9767 (MULTILIB_DIRNAMES): Add nan2008.
9768 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9769 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9770 and micromips incompatible. Add nan2008.
9771 (MULTILIB_DIRNAMES): Add nan2008.
9772 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9774 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9776 PR rtl-optimization/58079
9777 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9778 simplify_comparison has widened a comparison with an integer.
9780 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9782 * config/arm/neon.md (movmisalign<mode>): Disable when we
9783 don't allow unaligned accesses.
9784 (*movmisalign<mode>_neon_store): Likewise.
9785 (*movmisalign<mode>_neon_load): Likewise.
9786 (*movmisalign<mode>_neon_store): Likewise.
9787 (*movmisalign<mode>_neon_load): Likewise.
9789 2013-08-08 Jan Hubicka <jh@suse.cz>
9791 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9792 (make_pass_rebuild_cgraph_edges): Also clear references.
9793 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9794 * ipa-inline-transform.c (inline_transform): Remove all references
9796 * cgraphunit.c (expand_function): Remove all references after
9798 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9799 (ipa_find_reference): Rewrite to iterator.
9800 (remove_stmt_references): Likewise.
9801 (ipa_clear_stmts_in_references): New function.
9802 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9803 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9805 * ipa-split.c (split_function): Remove references in split function.
9807 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9810 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9811 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9813 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9816 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9817 suggested mode for the assignment isn't compatible with the
9820 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9823 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9825 2013-08-07 Xinliang David Li <davidxl@google.com>
9827 * config/i386/i386.opt: New option -mtune-ctrl=.
9828 * config/i386/x86-tune.def: New file.
9829 * config/i386/i386.h: include x86-tune.def.
9830 * config/i386/i386.c (ix86_option_override_internal):
9831 Parsing -mtune-ctrl= option and set tune features.
9833 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9836 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9837 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9840 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9842 * rtl.h (update_alignments): Declare.
9843 * final.c (grow_label_align): New function extracted from...
9844 (shorten_branches): ...here. Call it.
9845 (update_alignments): New function.
9846 * reorg.c (sibling_labels): New variable.
9847 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9848 the new label along with it onto the sibling_labels vector.
9849 (fill_simple_delay_slots): Adjust call to get_label_before.
9850 (fill_slots_from_thread): Likewise.
9851 (relax_delay_slots): Likewise.
9852 (make_return_insns): Likewise.
9853 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9855 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9857 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9858 document its semantics.
9859 (diagnostic_report_diagnostic): Adjust accordingly.
9861 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9863 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9864 (sparc_option_override): ...and port to new C++ pass API.
9865 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9867 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9869 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9871 2013-08-06 Caroline Tice <cmtice@google.com>
9873 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9874 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9875 * tree-pass.h: Add pass_vtable_verify.
9876 * varasm.c (assemble_variable): Add code to properly set the comdat
9877 section and name for the .vtable_map_vars section.
9878 (assemble_vtyv_preinit_initializer): New function.
9879 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9881 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9882 * vtable-verify.c: New file.
9883 * vtable-verify.h: New file.
9884 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9885 initialiation levels.
9886 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9887 * passes.def: Insert pass_vtable_verify.
9888 * aclocal.m4: Reorder includes.
9889 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9890 -fvtv-counts options.
9891 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9892 as appropriate, if -fvtable-verify=... is used.
9893 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
9894 -fvtable-verify=... is used.
9895 * Makefile.in (OBJS): Add vtable-verify.o to list.
9896 (vtable-verify.o): Add new build rule.
9897 (GTFILES): Add vtable-verify.c to list.
9898 * common.opt (fvtable-verify=): New flag.
9899 (vtv_priority): Values for fvtable-verify= flag.
9900 (fvtv-counts): New flag.
9901 (fvtv-debug): New flag.
9902 * tree.h (save_vtable_map_decl): New extern function decl.
9904 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9906 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
9907 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
9908 (pass_data_rl78_devirt): ...new pass_data instance and...
9909 (make_pass_rl78_devirt): ...new function.
9910 (rl78_asm_file_start): Port pass registration to new C++ API.
9912 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9914 * coretypes.h (rtl_opt_pass): Add.
9915 (gcc::context): Add.
9916 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
9917 (epiphany_init): Port to new C++ pass API.
9918 (epiphany_optimize_mode_switching): Likewise.
9919 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
9920 (pass_manager::get_pass_mode_switching): New.
9921 (pass_manager::get_pass_peephole2): New.
9922 * mode-switching.c (pass_mode_switching): Add clone method.
9923 * recog.c (pass_peephole2): Add clone method.
9924 (pass_split_all_insns): Add clone method.
9926 2013-08-06 David Malcolm <dmalcolm@redhat.com>
9928 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
9929 (mips_option_override): ...here, porting to new C++ API for passes.
9931 2013-08-06 Jan Hubicka <jh@suse.cz>
9933 * cgraph.c (cgraph_get_body): New function based on lto.c
9935 * cgraph.h (cgraph_get_body): Declare.
9936 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
9938 * cgraphunit.c (expand_function): Get body prior expanding.
9939 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
9940 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
9942 * passes.c (do_per_function_toporder): Get body.
9943 * tree-inline.c (expand_call_inline): Get body prior inlining it.
9944 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
9946 2013-08-06 Martin Jambor <mjambor@suse.cz>
9949 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
9950 is not re-finalized. Rename second parameter to no_collect.
9952 2013-08-06 Martin Jambor <mjambor@suse.cz>
9955 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
9956 MEM_REF has proper alignment information.
9958 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
9961 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
9963 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
9964 rtx (*) (rtx, ...) with insn_gen_fn.
9965 * genoutput.c (output_insn_data): Cast gen_? function pointers to
9966 insn_gen_fn::stored_funcptr. Add initializer braces.
9968 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9970 Rewrite how instances of passes are cloned to remove assumptions
9971 about their sizes (thus allowing pass subclasses to have
9972 additional data fields, albeit non-GC-managed ones at this point).
9974 * passes.c (make_pass_instance): Now that passes have clone
9975 methods, rewrite this function to eliminate XNEW and memcpy
9976 calls that used hardcoded sizes. Since this function no longer
9977 creates pass instances, rename it to...
9978 (add_pass_instance): ...this. Document the old way that passes were
9979 numbered and flagged, and rework this function to continue using it.
9980 (next_pass_1): Add an initial_pass argument for use by
9982 (position_pass): When adding multiple instances of a pass, use
9983 the pass's clone method, rather than relying on the XNEW/memcpy
9984 within the former make_pass_instance (now add_pass_instance).
9985 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9986 the initial instance of the current pass within the pass manager.
9988 2013-08-05 David Malcolm <dmalcolm@redhat.com>
9990 This is the automated part of the conversion of passes from C
9991 structs to C++ classes.
9993 Patch autogenerated by refactor_passes.py from
9994 https://github.com/davidmalcolm/gcc-refactoring-scripts
9995 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9997 * asan.c (pass_asan): Convert from a global struct to a subclass of
9998 gimple_opt_pass along with...
9999 (pass_data_asan): ...new pass_data instance and...
10000 (make_pass_asan): ...new function.
10001 (pass_asan_O0): Convert from a global struct to a subclass of
10002 gimple_opt_pass along with...
10003 (pass_data_asan_O0): ...new pass_data instance and...
10004 (make_pass_asan_O0): ...new function.
10005 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10006 subclass of rtl_opt_pass along with...
10007 (pass_data_inc_dec): ...new pass_data instance and...
10008 (make_pass_inc_dec): ...new function.
10009 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10010 a subclass of rtl_opt_pass along with...
10011 (pass_data_reorder_blocks): ...new pass_data instance and...
10012 (make_pass_reorder_blocks): ...new function.
10013 (pass_duplicate_computed_gotos): Convert from a global struct to a
10014 subclass of rtl_opt_pass along with...
10015 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10016 (make_pass_duplicate_computed_gotos): ...new function.
10017 (pass_partition_blocks): Convert from a global struct to a subclass of
10018 rtl_opt_pass along with...
10019 (pass_data_partition_blocks): ...new pass_data instance and...
10020 (make_pass_partition_blocks): ...new function.
10021 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10022 struct to a subclass of rtl_opt_pass along with...
10023 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10025 (make_pass_branch_target_load_optimize1): ...new function.
10026 (pass_branch_target_load_optimize2): Convert from a global struct to a
10027 subclass of rtl_opt_pass along with...
10028 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10030 (make_pass_branch_target_load_optimize2): ...new function.
10031 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10032 of rtl_opt_pass along with...
10033 (pass_data_jump): ...new pass_data instance and...
10034 (make_pass_jump): ...new function.
10035 (pass_jump2): Convert from a global struct to a subclass of
10036 rtl_opt_pass along with...
10037 (pass_data_jump2): ...new pass_data instance and...
10038 (make_pass_jump2): ...new function.
10039 * cfgexpand.c (pass_expand): Convert from a global struct to a
10040 subclass of rtl_opt_pass along with...
10041 (pass_data_expand): ...new pass_data instance and...
10042 (make_pass_expand): ...new function.
10043 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10044 of rtl_opt_pass along with...
10045 (pass_data_free_cfg): ...new pass_data instance and...
10046 (make_pass_free_cfg): ...new function.
10047 (pass_into_cfg_layout_mode): Convert from a global struct to a
10048 subclass of rtl_opt_pass along with...
10049 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10050 (make_pass_into_cfg_layout_mode): ...new function.
10051 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10052 subclass of rtl_opt_pass along with...
10053 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10054 (make_pass_outof_cfg_layout_mode): ...new function.
10055 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10056 struct to a subclass of gimple_opt_pass along with...
10057 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10058 (make_pass_build_cgraph_edges): ...new function.
10059 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10060 subclass of gimple_opt_pass along with...
10061 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10062 (make_pass_rebuild_cgraph_edges): ...new function.
10063 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10064 subclass of gimple_opt_pass along with...
10065 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10067 (make_pass_remove_cgraph_callee_edges): ...new function.
10068 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10069 struct to a subclass of rtl_opt_pass along with...
10070 (pass_data_stack_adjustments): ...new pass_data instance and...
10071 (make_pass_stack_adjustments): ...new function.
10072 * combine.c (pass_combine): Convert from a global struct to a subclass
10073 of rtl_opt_pass along with...
10074 (pass_data_combine): ...new pass_data instance and...
10075 (make_pass_combine): ...new function.
10076 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10077 global struct to a subclass of rtl_opt_pass along with...
10078 (pass_data_compare_elim_after_reload): ...new pass_data instance
10080 (make_pass_compare_elim_after_reload): ...new function.
10081 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10082 of rtl_opt_pass along with...
10083 (pass_data_rtl_cprop): ...new pass_data instance and...
10084 (make_pass_rtl_cprop): ...new function.
10085 * cse.c (pass_cse): Convert from a global struct to a subclass of
10086 rtl_opt_pass along with...
10087 (pass_data_cse): ...new pass_data instance and...
10088 (make_pass_cse): ...new function.
10089 (pass_cse2): Convert from a global struct to a subclass of
10090 rtl_opt_pass along with...
10091 (pass_data_cse2): ...new pass_data instance and...
10092 (make_pass_cse2): ...new function.
10093 (pass_cse_after_global_opts): Convert from a global struct to a
10094 subclass of rtl_opt_pass along with...
10095 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10096 (make_pass_cse_after_global_opts): ...new function.
10097 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10098 of rtl_opt_pass along with...
10099 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10100 (make_pass_ud_rtl_dce): ...new function.
10101 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10102 rtl_opt_pass along with...
10103 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10104 (make_pass_fast_rtl_dce): ...new function.
10105 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10106 a subclass of rtl_opt_pass along with...
10107 (pass_data_df_initialize_opt): ...new pass_data instance and...
10108 (make_pass_df_initialize_opt): ...new function.
10109 (pass_df_initialize_no_opt): Convert from a global struct to a
10110 subclass of rtl_opt_pass along with...
10111 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10112 (make_pass_df_initialize_no_opt): ...new function.
10113 (pass_df_finish): Convert from a global struct to a subclass of
10114 rtl_opt_pass along with...
10115 (pass_data_df_finish): ...new pass_data instance and...
10116 (make_pass_df_finish): ...new function.
10117 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10118 rtl_opt_pass along with...
10119 (pass_data_rtl_dse1): ...new pass_data instance and...
10120 (make_pass_rtl_dse1): ...new function.
10121 (pass_rtl_dse2): Convert from a global struct to a subclass of
10122 rtl_opt_pass along with...
10123 (pass_data_rtl_dse2): ...new pass_data instance and...
10124 (make_pass_rtl_dse2): ...new function.
10125 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10126 subclass of rtl_opt_pass along with...
10127 (pass_data_dwarf2_frame): ...new pass_data instance and...
10128 (make_pass_dwarf2_frame): ...new function.
10129 * except.c (pass_set_nothrow_function_flags): Convert from a global
10130 struct to a subclass of rtl_opt_pass along with...
10131 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10133 (make_pass_set_nothrow_function_flags): ...new function.
10134 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10135 subclass of rtl_opt_pass along with...
10136 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10138 (make_pass_convert_to_eh_region_ranges): ...new function.
10139 * final.c (pass_compute_alignments): Convert from a global struct to a
10140 subclass of rtl_opt_pass along with...
10141 (pass_data_compute_alignments): ...new pass_data instance and...
10142 (make_pass_compute_alignments): ...new function.
10143 (pass_final): Convert from a global struct to a subclass of
10144 rtl_opt_pass along with...
10145 (pass_data_final): ...new pass_data instance and...
10146 (make_pass_final): ...new function.
10147 (pass_shorten_branches): Convert from a global struct to a subclass of
10148 rtl_opt_pass along with...
10149 (pass_data_shorten_branches): ...new pass_data instance and...
10150 (make_pass_shorten_branches): ...new function.
10151 (pass_clean_state): Convert from a global struct to a subclass of
10152 rtl_opt_pass along with...
10153 (pass_data_clean_state): ...new pass_data instance and...
10154 (make_pass_clean_state): ...new function.
10155 * function.c (pass_instantiate_virtual_regs): Convert from a global
10156 struct to a subclass of rtl_opt_pass along with...
10157 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10158 (make_pass_instantiate_virtual_regs): ...new function.
10159 (pass_leaf_regs): Convert from a global struct to a subclass of
10160 rtl_opt_pass along with...
10161 (pass_data_leaf_regs): ...new pass_data instance and...
10162 (make_pass_leaf_regs): ...new function.
10163 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10164 subclass of rtl_opt_pass along with...
10165 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10167 (make_pass_thread_prologue_and_epilogue): ...new function.
10168 (pass_match_asm_constraints): Convert from a global struct to a
10169 subclass of rtl_opt_pass along with...
10170 (pass_data_match_asm_constraints): ...new pass_data instance and...
10171 (make_pass_match_asm_constraints): ...new function.
10172 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10173 subclass of rtl_opt_pass along with...
10174 (pass_data_rtl_fwprop): ...new pass_data instance and...
10175 (make_pass_rtl_fwprop): ...new function.
10176 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10177 rtl_opt_pass along with...
10178 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10179 (make_pass_rtl_fwprop_addr): ...new function.
10180 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10181 rtl_opt_pass along with...
10182 (pass_data_rtl_pre): ...new pass_data instance and...
10183 (make_pass_rtl_pre): ...new function.
10184 (pass_rtl_hoist): Convert from a global struct to a subclass of
10185 rtl_opt_pass along with...
10186 (pass_data_rtl_hoist): ...new pass_data instance and...
10187 (make_pass_rtl_hoist): ...new function.
10188 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10189 subclass of gimple_opt_pass along with...
10190 (pass_data_lower_cf): ...new pass_data instance and...
10191 (make_pass_lower_cf): ...new function.
10192 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10193 from a global struct to a subclass of gimple_opt_pass along with...
10194 (pass_data_strength_reduction): ...new pass_data instance and...
10195 (make_pass_strength_reduction): ...new function.
10196 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10197 of rtl_opt_pass along with...
10198 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10199 (make_pass_rtl_ifcvt): ...new function.
10200 (pass_if_after_combine): Convert from a global struct to a subclass of
10201 rtl_opt_pass along with...
10202 (pass_data_if_after_combine): ...new pass_data instance and...
10203 (make_pass_if_after_combine): ...new function.
10204 (pass_if_after_reload): Convert from a global struct to a subclass of
10205 rtl_opt_pass along with...
10206 (pass_data_if_after_reload): ...new pass_data instance and...
10207 (make_pass_if_after_reload): ...new function.
10208 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10209 a subclass of rtl_opt_pass along with...
10210 (pass_data_initialize_regs): ...new pass_data instance and...
10211 (make_pass_initialize_regs): ...new function.
10212 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10213 of ipa_opt_pass_d along with...
10214 (pass_data_ipa_cp): ...new pass_data instance and...
10215 (make_pass_ipa_cp): ...new function.
10216 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10217 global struct to a subclass of gimple_opt_pass along with...
10218 (pass_data_inline_parameters): ...new pass_data instance and...
10219 (make_pass_inline_parameters): ...new function.
10220 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10221 subclass of gimple_opt_pass along with...
10222 (pass_data_early_inline): ...new pass_data instance and...
10223 (make_pass_early_inline): ...new function.
10224 (pass_ipa_inline): Convert from a global struct to a subclass of
10225 ipa_opt_pass_d along with...
10226 (pass_data_ipa_inline): ...new pass_data instance and...
10227 (make_pass_ipa_inline): ...new function.
10228 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10229 struct to a subclass of gimple_opt_pass along with...
10230 (pass_data_local_pure_const): ...new pass_data instance and...
10231 (make_pass_local_pure_const): ...new function.
10232 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10233 ipa_opt_pass_d along with...
10234 (pass_data_ipa_pure_const): ...new pass_data instance and...
10235 (make_pass_ipa_pure_const): ...new function.
10236 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10237 to a subclass of ipa_opt_pass_d along with...
10238 (pass_data_ipa_reference): ...new pass_data instance and...
10239 (make_pass_ipa_reference): ...new function.
10240 * ipa-split.c (pass_split_functions): Convert from a global struct to
10241 a subclass of gimple_opt_pass along with...
10242 (pass_data_split_functions): ...new pass_data instance and...
10243 (make_pass_split_functions): ...new function.
10244 (pass_feedback_split_functions): Convert from a global struct to a
10245 subclass of gimple_opt_pass along with...
10246 (pass_data_feedback_split_functions): ...new pass_data instance and...
10247 (make_pass_feedback_split_functions): ...new function.
10248 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10249 global struct to a subclass of simple_ipa_opt_pass along with...
10250 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10252 (make_pass_ipa_function_and_variable_visibility): ...new function.
10253 (pass_ipa_free_inline_summary): Convert from a global struct to a
10254 subclass of simple_ipa_opt_pass along with...
10255 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10256 (make_pass_ipa_free_inline_summary): ...new function.
10257 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10258 subclass of ipa_opt_pass_d along with...
10259 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10261 (make_pass_ipa_whole_program_visibility): ...new function.
10262 (pass_ipa_profile): Convert from a global struct to a subclass of
10263 ipa_opt_pass_d along with...
10264 (pass_data_ipa_profile): ...new pass_data instance and...
10265 (make_pass_ipa_profile): ...new function.
10266 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10267 ipa_opt_pass_d along with...
10268 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10269 (make_pass_ipa_cdtor_merge): ...new function.
10270 * ira.c (pass_ira): Convert from a global struct to a subclass of
10271 rtl_opt_pass along with...
10272 (pass_data_ira): ...new pass_data instance and...
10273 (make_pass_ira): ...new function.
10274 (pass_reload): Convert from a global struct to a subclass of
10275 rtl_opt_pass along with...
10276 (pass_data_reload): ...new pass_data instance and...
10277 (make_pass_reload): ...new function.
10278 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10279 subclass of rtl_opt_pass along with...
10280 (pass_data_cleanup_barriers): ...new pass_data instance and...
10281 (make_pass_cleanup_barriers): ...new function.
10282 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10283 of rtl_opt_pass along with...
10284 (pass_data_loop2): ...new pass_data instance and...
10285 (make_pass_loop2): ...new function.
10286 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10287 rtl_opt_pass along with...
10288 (pass_data_rtl_loop_init): ...new pass_data instance and...
10289 (make_pass_rtl_loop_init): ...new function.
10290 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10291 rtl_opt_pass along with...
10292 (pass_data_rtl_loop_done): ...new pass_data instance and...
10293 (make_pass_rtl_loop_done): ...new function.
10294 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10295 subclass of rtl_opt_pass along with...
10296 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10297 (make_pass_rtl_move_loop_invariants): ...new function.
10298 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10299 rtl_opt_pass along with...
10300 (pass_data_rtl_unswitch): ...new pass_data instance and...
10301 (make_pass_rtl_unswitch): ...new function.
10302 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10303 subclass of rtl_opt_pass along with...
10304 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10306 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10307 (pass_rtl_doloop): Convert from a global struct to a subclass of
10308 rtl_opt_pass along with...
10309 (pass_data_rtl_doloop): ...new pass_data instance and...
10310 (make_pass_rtl_doloop): ...new function.
10311 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10312 a subclass of rtl_opt_pass along with...
10313 (pass_data_lower_subreg): ...new pass_data instance and...
10314 (make_pass_lower_subreg): ...new function.
10315 (pass_lower_subreg2): Convert from a global struct to a subclass of
10316 rtl_opt_pass along with...
10317 (pass_data_lower_subreg2): ...new pass_data instance and...
10318 (make_pass_lower_subreg2): ...new function.
10319 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10320 struct to a subclass of ipa_opt_pass_d along with...
10321 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10322 (make_pass_ipa_lto_gimple_out): ...new function.
10323 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10324 of ipa_opt_pass_d along with...
10325 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10326 (make_pass_ipa_lto_finish_out): ...new function.
10327 * mode-switching.c (pass_mode_switching): Convert from a global struct
10328 to a subclass of rtl_opt_pass along with...
10329 (pass_data_mode_switching): ...new pass_data instance and...
10330 (make_pass_mode_switching): ...new function.
10331 * modulo-sched.c (pass_sms): Convert from a global struct to a
10332 subclass of rtl_opt_pass along with...
10333 (pass_data_sms): ...new pass_data instance and...
10334 (make_pass_sms): ...new function.
10335 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10336 subclass of gimple_opt_pass along with...
10337 (pass_data_expand_omp): ...new pass_data instance and...
10338 (make_pass_expand_omp): ...new function.
10339 (pass_lower_omp): Convert from a global struct to a subclass of
10340 gimple_opt_pass along with...
10341 (pass_data_lower_omp): ...new pass_data instance and...
10342 (make_pass_lower_omp): ...new function.
10343 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10344 of gimple_opt_pass along with...
10345 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10346 (make_pass_diagnose_omp_blocks): ...new function.
10347 * passes.c (pass_early_local_passes): Convert from a global struct to
10348 a subclass of simple_ipa_opt_pass along with...
10349 (pass_data_early_local_passes): ...new pass_data instance and...
10350 (make_pass_early_local_passes): ...new function.
10351 (pass_all_early_optimizations): Convert from a global struct to a
10352 subclass of gimple_opt_pass along with...
10353 (pass_data_all_early_optimizations): ...new pass_data instance and...
10354 (make_pass_all_early_optimizations): ...new function.
10355 (pass_all_optimizations): Convert from a global struct to a subclass
10356 of gimple_opt_pass along with...
10357 (pass_data_all_optimizations): ...new pass_data instance and...
10358 (make_pass_all_optimizations): ...new function.
10359 (pass_all_optimizations_g): Convert from a global struct to a subclass
10360 of gimple_opt_pass along with...
10361 (pass_data_all_optimizations_g): ...new pass_data instance and...
10362 (make_pass_all_optimizations_g): ...new function.
10363 (pass_rest_of_compilation): Convert from a global struct to a subclass
10364 of rtl_opt_pass along with...
10365 (pass_data_rest_of_compilation): ...new pass_data instance and...
10366 (make_pass_rest_of_compilation): ...new function.
10367 (pass_postreload): Convert from a global struct to a subclass of
10368 rtl_opt_pass along with...
10369 (pass_data_postreload): ...new pass_data instance and...
10370 (make_pass_postreload): ...new function.
10371 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10372 subclass of rtl_opt_pass along with...
10373 (pass_data_gcse2): ...new pass_data instance and...
10374 (make_pass_gcse2): ...new function.
10375 * postreload.c (pass_postreload_cse): Convert from a global struct to
10376 a subclass of rtl_opt_pass along with...
10377 (pass_data_postreload_cse): ...new pass_data instance and...
10378 (make_pass_postreload_cse): ...new function.
10379 * predict.c (pass_profile): Convert from a global struct to a subclass
10380 of gimple_opt_pass along with...
10381 (pass_data_profile): ...new pass_data instance and...
10382 (make_pass_profile): ...new function.
10383 (pass_strip_predict_hints): Convert from a global struct to a subclass
10384 of gimple_opt_pass along with...
10385 (pass_data_strip_predict_hints): ...new pass_data instance and...
10386 (make_pass_strip_predict_hints): ...new function.
10387 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10388 of rtl_opt_pass along with...
10389 (pass_data_peephole2): ...new pass_data instance and...
10390 (make_pass_peephole2): ...new function.
10391 (pass_split_all_insns): Convert from a global struct to a subclass of
10392 rtl_opt_pass along with...
10393 (pass_data_split_all_insns): ...new pass_data instance and...
10394 (make_pass_split_all_insns): ...new function.
10395 (pass_split_after_reload): Convert from a global struct to a subclass
10396 of rtl_opt_pass along with...
10397 (pass_data_split_after_reload): ...new pass_data instance and...
10398 (make_pass_split_after_reload): ...new function.
10399 (pass_split_before_regstack): Convert from a global struct to a
10400 subclass of rtl_opt_pass along with...
10401 (pass_data_split_before_regstack): ...new pass_data instance and...
10402 (make_pass_split_before_regstack): ...new function.
10403 (pass_split_before_sched2): Convert from a global struct to a subclass
10404 of rtl_opt_pass along with...
10405 (pass_data_split_before_sched2): ...new pass_data instance and...
10406 (make_pass_split_before_sched2): ...new function.
10407 (pass_split_for_shorten_branches): Convert from a global struct to a
10408 subclass of rtl_opt_pass along with...
10409 (pass_data_split_for_shorten_branches): ...new pass_data instance
10411 (make_pass_split_for_shorten_branches): ...new function.
10412 * ree.c (pass_ree): Convert from a global struct to a subclass of
10413 rtl_opt_pass along with...
10414 (pass_data_ree): ...new pass_data instance and...
10415 (make_pass_ree): ...new function.
10416 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10417 subclass of rtl_opt_pass along with...
10418 (pass_data_stack_regs): ...new pass_data instance and...
10419 (make_pass_stack_regs): ...new function.
10420 (pass_stack_regs_run): Convert from a global struct to a subclass of
10421 rtl_opt_pass along with...
10422 (pass_data_stack_regs_run): ...new pass_data instance and...
10423 (make_pass_stack_regs_run): ...new function.
10424 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10425 subclass of rtl_opt_pass along with...
10426 (pass_data_cprop_hardreg): ...new pass_data instance and...
10427 (make_pass_cprop_hardreg): ...new function.
10428 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10429 subclass of rtl_opt_pass along with...
10430 (pass_data_reginfo_init): ...new pass_data instance and...
10431 (make_pass_reginfo_init): ...new function.
10432 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10433 of rtl_opt_pass along with...
10434 (pass_data_regmove): ...new pass_data instance and...
10435 (make_pass_regmove): ...new function.
10436 * regrename.c (pass_regrename): Convert from a global struct to a
10437 subclass of rtl_opt_pass along with...
10438 (pass_data_regrename): ...new pass_data instance and...
10439 (make_pass_regrename): ...new function.
10440 * reorg.c (pass_delay_slots): Convert from a global struct to a
10441 subclass of rtl_opt_pass along with...
10442 (pass_data_delay_slots): ...new pass_data instance and...
10443 (make_pass_delay_slots): ...new function.
10444 (pass_machine_reorg): Convert from a global struct to a subclass of
10445 rtl_opt_pass along with...
10446 (pass_data_machine_reorg): ...new pass_data instance and...
10447 (make_pass_machine_reorg): ...new function.
10448 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10449 of rtl_opt_pass along with...
10450 (pass_data_sched): ...new pass_data instance and...
10451 (make_pass_sched): ...new function.
10452 (pass_sched2): Convert from a global struct to a subclass of
10453 rtl_opt_pass along with...
10454 (pass_data_sched2): ...new pass_data instance and...
10455 (make_pass_sched2): ...new function.
10456 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10457 to a subclass of rtl_opt_pass along with...
10458 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10459 (make_pass_stack_ptr_mod): ...new function.
10460 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10461 to a subclass of rtl_opt_pass along with...
10462 (pass_data_rtl_store_motion): ...new pass_data instance and...
10463 (make_pass_rtl_store_motion): ...new function.
10464 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10465 of gimple_opt_pass along with...
10466 (pass_data_tracer): ...new pass_data instance and...
10467 (make_pass_tracer): ...new function.
10468 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10469 to a subclass of gimple_opt_pass along with...
10470 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10471 (make_pass_diagnose_tm_blocks): ...new function.
10472 (pass_lower_tm): Convert from a global struct to a subclass of
10473 gimple_opt_pass along with...
10474 (pass_data_lower_tm): ...new pass_data instance and...
10475 (make_pass_lower_tm): ...new function.
10476 (pass_tm_init): Convert from a global struct to a subclass of
10477 gimple_opt_pass along with...
10478 (pass_data_tm_init): ...new pass_data instance and...
10479 (make_pass_tm_init): ...new function.
10480 (pass_tm_mark): Convert from a global struct to a subclass of
10481 gimple_opt_pass along with...
10482 (pass_data_tm_mark): ...new pass_data instance and...
10483 (make_pass_tm_mark): ...new function.
10484 (pass_tm_edges): Convert from a global struct to a subclass of
10485 gimple_opt_pass along with...
10486 (pass_data_tm_edges): ...new pass_data instance and...
10487 (make_pass_tm_edges): ...new function.
10488 (pass_tm_memopt): Convert from a global struct to a subclass of
10489 gimple_opt_pass along with...
10490 (pass_data_tm_memopt): ...new pass_data instance and...
10491 (make_pass_tm_memopt): ...new function.
10492 (pass_ipa_tm): Convert from a global struct to a subclass of
10493 simple_ipa_opt_pass along with...
10494 (pass_data_ipa_tm): ...new pass_data instance and...
10495 (make_pass_ipa_tm): ...new function.
10496 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10497 subclass of gimple_opt_pass along with...
10498 (pass_data_call_cdce): ...new pass_data instance and...
10499 (make_pass_call_cdce): ...new function.
10500 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10501 subclass of gimple_opt_pass along with...
10502 (pass_data_build_cfg): ...new pass_data instance and...
10503 (make_pass_build_cfg): ...new function.
10504 (pass_split_crit_edges): Convert from a global struct to a subclass of
10505 gimple_opt_pass along with...
10506 (pass_data_split_crit_edges): ...new pass_data instance and...
10507 (make_pass_split_crit_edges): ...new function.
10508 (pass_warn_function_return): Convert from a global struct to a
10509 subclass of gimple_opt_pass along with...
10510 (pass_data_warn_function_return): ...new pass_data instance and...
10511 (make_pass_warn_function_return): ...new function.
10512 (pass_warn_function_noreturn): Convert from a global struct to a
10513 subclass of gimple_opt_pass along with...
10514 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10515 (make_pass_warn_function_noreturn): ...new function.
10516 (pass_warn_unused_result): Convert from a global struct to a subclass
10517 of gimple_opt_pass along with...
10518 (pass_data_warn_unused_result): ...new pass_data instance and...
10519 (make_pass_warn_unused_result): ...new function.
10520 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10521 a subclass of gimple_opt_pass along with...
10522 (pass_data_merge_phi): ...new pass_data instance and...
10523 (make_pass_merge_phi): ...new function.
10524 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10525 a subclass of gimple_opt_pass along with...
10526 (pass_data_lower_complex): ...new pass_data instance and...
10527 (make_pass_lower_complex): ...new function.
10528 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10529 gimple_opt_pass along with...
10530 (pass_data_lower_complex_O0): ...new pass_data instance and...
10531 (make_pass_lower_complex_O0): ...new function.
10532 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10533 subclass of gimple_opt_pass along with...
10534 (pass_data_lower_eh): ...new pass_data instance and...
10535 (make_pass_lower_eh): ...new function.
10536 (pass_refactor_eh): Convert from a global struct to a subclass of
10537 gimple_opt_pass along with...
10538 (pass_data_refactor_eh): ...new pass_data instance and...
10539 (make_pass_refactor_eh): ...new function.
10540 (pass_lower_resx): Convert from a global struct to a subclass of
10541 gimple_opt_pass along with...
10542 (pass_data_lower_resx): ...new pass_data instance and...
10543 (make_pass_lower_resx): ...new function.
10544 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10545 of gimple_opt_pass along with...
10546 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10547 (make_pass_lower_eh_dispatch): ...new function.
10548 (pass_cleanup_eh): Convert from a global struct to a subclass of
10549 gimple_opt_pass along with...
10550 (pass_data_cleanup_eh): ...new pass_data instance and...
10551 (make_pass_cleanup_eh): ...new function.
10552 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10553 to a subclass of simple_ipa_opt_pass along with...
10554 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10555 (make_pass_ipa_lower_emutls): ...new function.
10556 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10557 a subclass of gimple_opt_pass along with...
10558 (pass_data_if_conversion): ...new pass_data instance and...
10559 (make_pass_if_conversion): ...new function.
10560 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10561 subclass of gimple_opt_pass along with...
10562 (pass_data_build_ssa): ...new pass_data instance and...
10563 (make_pass_build_ssa): ...new function.
10564 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10565 global struct to a subclass of gimple_opt_pass along with...
10566 (pass_data_loop_distribution): ...new pass_data instance and...
10567 (make_pass_loop_distribution): ...new function.
10568 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10569 subclass of gimple_opt_pass along with...
10570 (pass_data_mudflap_1): ...new pass_data instance and...
10571 (make_pass_mudflap_1): ...new function.
10572 (pass_mudflap_2): Convert from a global struct to a subclass of
10573 gimple_opt_pass along with...
10574 (pass_data_mudflap_2): ...new pass_data instance and...
10575 (make_pass_mudflap_2): ...new function.
10576 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10577 subclass of gimple_opt_pass along with...
10578 (pass_data_mudflap_1): ...new pass_data instance and...
10579 (make_pass_mudflap_1): ...new function.
10580 (pass_mudflap_2): Convert from a global struct to a subclass of
10581 gimple_opt_pass along with...
10582 (pass_data_mudflap_2): ...new pass_data instance and...
10583 (make_pass_mudflap_2): ...new function.
10584 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10585 gimple_opt_pass along with...
10586 (pass_data_nrv): ...new pass_data instance and...
10587 (make_pass_nrv): ...new function.
10588 (pass_return_slot): Convert from a global struct to a subclass of
10589 gimple_opt_pass along with...
10590 (pass_data_return_slot): ...new pass_data instance and...
10591 (make_pass_return_slot): ...new function.
10592 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10593 to a subclass of gimple_opt_pass along with...
10594 (pass_data_object_sizes): ...new pass_data instance and...
10595 (make_pass_object_sizes): ...new function.
10596 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10597 global struct to a subclass of gimple_opt_pass along with...
10598 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10600 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10601 (pass_fixup_cfg): Convert from a global struct to a subclass of
10602 gimple_opt_pass along with...
10603 (pass_data_fixup_cfg): ...new pass_data instance and...
10604 (make_pass_fixup_cfg): ...new function.
10605 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10606 (make_pass_mudflap_1): ...new function.
10607 (pass_mudflap_2): Replace declaration with that of...
10608 (make_pass_mudflap_2): ...new function.
10609 (pass_asan): Replace declaration with that of...
10610 (make_pass_asan): ...new function.
10611 (pass_asan_O0): Replace declaration with that of...
10612 (make_pass_asan_O0): ...new function.
10613 (pass_tsan): Replace declaration with that of...
10614 (make_pass_tsan): ...new function.
10615 (pass_tsan_O0): Replace declaration with that of...
10616 (make_pass_tsan_O0): ...new function.
10617 (pass_lower_cf): Replace declaration with that of...
10618 (make_pass_lower_cf): ...new function.
10619 (pass_refactor_eh): Replace declaration with that of...
10620 (make_pass_refactor_eh): ...new function.
10621 (pass_lower_eh): Replace declaration with that of...
10622 (make_pass_lower_eh): ...new function.
10623 (pass_lower_eh_dispatch): Replace declaration with that of...
10624 (make_pass_lower_eh_dispatch): ...new function.
10625 (pass_lower_resx): Replace declaration with that of...
10626 (make_pass_lower_resx): ...new function.
10627 (pass_build_cfg): Replace declaration with that of...
10628 (make_pass_build_cfg): ...new function.
10629 (pass_early_tree_profile): Replace declaration with that of...
10630 (make_pass_early_tree_profile): ...new function.
10631 (pass_cleanup_eh): Replace declaration with that of...
10632 (make_pass_cleanup_eh): ...new function.
10633 (pass_sra): Replace declaration with that of...
10634 (make_pass_sra): ...new function.
10635 (pass_sra_early): Replace declaration with that of...
10636 (make_pass_sra_early): ...new function.
10637 (pass_early_ipa_sra): Replace declaration with that of...
10638 (make_pass_early_ipa_sra): ...new function.
10639 (pass_tail_recursion): Replace declaration with that of...
10640 (make_pass_tail_recursion): ...new function.
10641 (pass_tail_calls): Replace declaration with that of...
10642 (make_pass_tail_calls): ...new function.
10643 (pass_tree_loop): Replace declaration with that of...
10644 (make_pass_tree_loop): ...new function.
10645 (pass_tree_loop_init): Replace declaration with that of...
10646 (make_pass_tree_loop_init): ...new function.
10647 (pass_lim): Replace declaration with that of...
10648 (make_pass_lim): ...new function.
10649 (pass_tree_unswitch): Replace declaration with that of...
10650 (make_pass_tree_unswitch): ...new function.
10651 (pass_predcom): Replace declaration with that of...
10652 (make_pass_predcom): ...new function.
10653 (pass_iv_canon): Replace declaration with that of...
10654 (make_pass_iv_canon): ...new function.
10655 (pass_scev_cprop): Replace declaration with that of...
10656 (make_pass_scev_cprop): ...new function.
10657 (pass_empty_loop): Replace declaration with that of...
10658 (make_pass_empty_loop): ...new function.
10659 (pass_record_bounds): Replace declaration with that of...
10660 (make_pass_record_bounds): ...new function.
10661 (pass_graphite): Replace declaration with that of...
10662 (make_pass_graphite): ...new function.
10663 (pass_graphite_transforms): Replace declaration with that of...
10664 (make_pass_graphite_transforms): ...new function.
10665 (pass_if_conversion): Replace declaration with that of...
10666 (make_pass_if_conversion): ...new function.
10667 (pass_loop_distribution): Replace declaration with that of...
10668 (make_pass_loop_distribution): ...new function.
10669 (pass_vectorize): Replace declaration with that of...
10670 (make_pass_vectorize): ...new function.
10671 (pass_slp_vectorize): Replace declaration with that of...
10672 (make_pass_slp_vectorize): ...new function.
10673 (pass_complete_unroll): Replace declaration with that of...
10674 (make_pass_complete_unroll): ...new function.
10675 (pass_complete_unrolli): Replace declaration with that of...
10676 (make_pass_complete_unrolli): ...new function.
10677 (pass_parallelize_loops): Replace declaration with that of...
10678 (make_pass_parallelize_loops): ...new function.
10679 (pass_loop_prefetch): Replace declaration with that of...
10680 (make_pass_loop_prefetch): ...new function.
10681 (pass_iv_optimize): Replace declaration with that of...
10682 (make_pass_iv_optimize): ...new function.
10683 (pass_tree_loop_done): Replace declaration with that of...
10684 (make_pass_tree_loop_done): ...new function.
10685 (pass_ch): Replace declaration with that of...
10686 (make_pass_ch): ...new function.
10687 (pass_ccp): Replace declaration with that of...
10688 (make_pass_ccp): ...new function.
10689 (pass_phi_only_cprop): Replace declaration with that of...
10690 (make_pass_phi_only_cprop): ...new function.
10691 (pass_build_ssa): Replace declaration with that of...
10692 (make_pass_build_ssa): ...new function.
10693 (pass_build_alias): Replace declaration with that of...
10694 (make_pass_build_alias): ...new function.
10695 (pass_build_ealias): Replace declaration with that of...
10696 (make_pass_build_ealias): ...new function.
10697 (pass_dominator): Replace declaration with that of...
10698 (make_pass_dominator): ...new function.
10699 (pass_dce): Replace declaration with that of...
10700 (make_pass_dce): ...new function.
10701 (pass_dce_loop): Replace declaration with that of...
10702 (make_pass_dce_loop): ...new function.
10703 (pass_cd_dce): Replace declaration with that of...
10704 (make_pass_cd_dce): ...new function.
10705 (pass_call_cdce): Replace declaration with that of...
10706 (make_pass_call_cdce): ...new function.
10707 (pass_merge_phi): Replace declaration with that of...
10708 (make_pass_merge_phi): ...new function.
10709 (pass_split_crit_edges): Replace declaration with that of...
10710 (make_pass_split_crit_edges): ...new function.
10711 (pass_pre): Replace declaration with that of...
10712 (make_pass_pre): ...new function.
10713 (pass_profile): Replace declaration with that of...
10714 (make_pass_profile): ...new function.
10715 (pass_strip_predict_hints): Replace declaration with that of...
10716 (make_pass_strip_predict_hints): ...new function.
10717 (pass_lower_complex_O0): Replace declaration with that of...
10718 (make_pass_lower_complex_O0): ...new function.
10719 (pass_lower_complex): Replace declaration with that of...
10720 (make_pass_lower_complex): ...new function.
10721 (pass_lower_vector): Replace declaration with that of...
10722 (make_pass_lower_vector): ...new function.
10723 (pass_lower_vector_ssa): Replace declaration with that of...
10724 (make_pass_lower_vector_ssa): ...new function.
10725 (pass_lower_omp): Replace declaration with that of...
10726 (make_pass_lower_omp): ...new function.
10727 (pass_diagnose_omp_blocks): Replace declaration with that of...
10728 (make_pass_diagnose_omp_blocks): ...new function.
10729 (pass_expand_omp): Replace declaration with that of...
10730 (make_pass_expand_omp): ...new function.
10731 (pass_expand_omp_ssa): Replace declaration with that of...
10732 (make_pass_expand_omp_ssa): ...new function.
10733 (pass_object_sizes): Replace declaration with that of...
10734 (make_pass_object_sizes): ...new function.
10735 (pass_strlen): Replace declaration with that of...
10736 (make_pass_strlen): ...new function.
10737 (pass_fold_builtins): Replace declaration with that of...
10738 (make_pass_fold_builtins): ...new function.
10739 (pass_stdarg): Replace declaration with that of...
10740 (make_pass_stdarg): ...new function.
10741 (pass_early_warn_uninitialized): Replace declaration with that of...
10742 (make_pass_early_warn_uninitialized): ...new function.
10743 (pass_late_warn_uninitialized): Replace declaration with that of...
10744 (make_pass_late_warn_uninitialized): ...new function.
10745 (pass_cse_reciprocals): Replace declaration with that of...
10746 (make_pass_cse_reciprocals): ...new function.
10747 (pass_cse_sincos): Replace declaration with that of...
10748 (make_pass_cse_sincos): ...new function.
10749 (pass_optimize_bswap): Replace declaration with that of...
10750 (make_pass_optimize_bswap): ...new function.
10751 (pass_optimize_widening_mul): Replace declaration with that of...
10752 (make_pass_optimize_widening_mul): ...new function.
10753 (pass_warn_function_return): Replace declaration with that of...
10754 (make_pass_warn_function_return): ...new function.
10755 (pass_warn_function_noreturn): Replace declaration with that of...
10756 (make_pass_warn_function_noreturn): ...new function.
10757 (pass_cselim): Replace declaration with that of...
10758 (make_pass_cselim): ...new function.
10759 (pass_phiopt): Replace declaration with that of...
10760 (make_pass_phiopt): ...new function.
10761 (pass_forwprop): Replace declaration with that of...
10762 (make_pass_forwprop): ...new function.
10763 (pass_phiprop): Replace declaration with that of...
10764 (make_pass_phiprop): ...new function.
10765 (pass_tree_ifcombine): Replace declaration with that of...
10766 (make_pass_tree_ifcombine): ...new function.
10767 (pass_dse): Replace declaration with that of...
10768 (make_pass_dse): ...new function.
10769 (pass_nrv): Replace declaration with that of...
10770 (make_pass_nrv): ...new function.
10771 (pass_rename_ssa_copies): Replace declaration with that of...
10772 (make_pass_rename_ssa_copies): ...new function.
10773 (pass_sink_code): Replace declaration with that of...
10774 (make_pass_sink_code): ...new function.
10775 (pass_fre): Replace declaration with that of...
10776 (make_pass_fre): ...new function.
10777 (pass_check_data_deps): Replace declaration with that of...
10778 (make_pass_check_data_deps): ...new function.
10779 (pass_copy_prop): Replace declaration with that of...
10780 (make_pass_copy_prop): ...new function.
10781 (pass_vrp): Replace declaration with that of...
10782 (make_pass_vrp): ...new function.
10783 (pass_uncprop): Replace declaration with that of...
10784 (make_pass_uncprop): ...new function.
10785 (pass_return_slot): Replace declaration with that of...
10786 (make_pass_return_slot): ...new function.
10787 (pass_reassoc): Replace declaration with that of...
10788 (make_pass_reassoc): ...new function.
10789 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10790 (make_pass_rebuild_cgraph_edges): ...new function.
10791 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10792 (make_pass_remove_cgraph_callee_edges): ...new function.
10793 (pass_build_cgraph_edges): Replace declaration with that of...
10794 (make_pass_build_cgraph_edges): ...new function.
10795 (pass_local_pure_const): Replace declaration with that of...
10796 (make_pass_local_pure_const): ...new function.
10797 (pass_tracer): Replace declaration with that of...
10798 (make_pass_tracer): ...new function.
10799 (pass_warn_unused_result): Replace declaration with that of...
10800 (make_pass_warn_unused_result): ...new function.
10801 (pass_diagnose_tm_blocks): Replace declaration with that of...
10802 (make_pass_diagnose_tm_blocks): ...new function.
10803 (pass_lower_tm): Replace declaration with that of...
10804 (make_pass_lower_tm): ...new function.
10805 (pass_tm_init): Replace declaration with that of...
10806 (make_pass_tm_init): ...new function.
10807 (pass_tm_mark): Replace declaration with that of...
10808 (make_pass_tm_mark): ...new function.
10809 (pass_tm_memopt): Replace declaration with that of...
10810 (make_pass_tm_memopt): ...new function.
10811 (pass_tm_edges): Replace declaration with that of...
10812 (make_pass_tm_edges): ...new function.
10813 (pass_split_functions): Replace declaration with that of...
10814 (make_pass_split_functions): ...new function.
10815 (pass_feedback_split_functions): Replace declaration with that of...
10816 (make_pass_feedback_split_functions): ...new function.
10817 (pass_strength_reduction): Replace declaration with that of...
10818 (make_pass_strength_reduction): ...new function.
10819 (pass_ipa_lower_emutls): Replace declaration with that of...
10820 (make_pass_ipa_lower_emutls): ...new function.
10821 (pass_ipa_function_and_variable_visibility): Replace declaration with
10823 (make_pass_ipa_function_and_variable_visibility): ...new function.
10824 (pass_ipa_tree_profile): Replace declaration with that of...
10825 (make_pass_ipa_tree_profile): ...new function.
10826 (pass_early_local_passes): Replace declaration with that of...
10827 (make_pass_early_local_passes): ...new function.
10828 (pass_ipa_whole_program_visibility): Replace declaration with that
10830 (make_pass_ipa_whole_program_visibility): ...new function.
10831 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10832 (make_pass_ipa_lto_gimple_out): ...new function.
10833 (pass_ipa_increase_alignment): Replace declaration with that of...
10834 (make_pass_ipa_increase_alignment): ...new function.
10835 (pass_ipa_inline): Replace declaration with that of...
10836 (make_pass_ipa_inline): ...new function.
10837 (pass_ipa_free_lang_data): Replace declaration with that of...
10838 (make_pass_ipa_free_lang_data): ...new function.
10839 (pass_ipa_free_inline_summary): Replace declaration with that of...
10840 (make_pass_ipa_free_inline_summary): ...new function.
10841 (pass_ipa_cp): Replace declaration with that of...
10842 (make_pass_ipa_cp): ...new function.
10843 (pass_ipa_reference): Replace declaration with that of...
10844 (make_pass_ipa_reference): ...new function.
10845 (pass_ipa_pure_const): Replace declaration with that of...
10846 (make_pass_ipa_pure_const): ...new function.
10847 (pass_ipa_pta): Replace declaration with that of...
10848 (make_pass_ipa_pta): ...new function.
10849 (pass_ipa_lto_finish_out): Replace declaration with that of...
10850 (make_pass_ipa_lto_finish_out): ...new function.
10851 (pass_ipa_tm): Replace declaration with that of...
10852 (make_pass_ipa_tm): ...new function.
10853 (pass_ipa_profile): Replace declaration with that of...
10854 (make_pass_ipa_profile): ...new function.
10855 (pass_ipa_cdtor_merge): Replace declaration with that of...
10856 (make_pass_ipa_cdtor_merge): ...new function.
10857 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10859 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10860 (pass_init_datastructures): Replace declaration with that of...
10861 (make_pass_init_datastructures): ...new function.
10862 (pass_fixup_cfg): Replace declaration with that of...
10863 (make_pass_fixup_cfg): ...new function.
10864 (pass_expand): Replace declaration with that of...
10865 (make_pass_expand): ...new function.
10866 (pass_instantiate_virtual_regs): Replace declaration with that of...
10867 (make_pass_instantiate_virtual_regs): ...new function.
10868 (pass_rtl_fwprop): Replace declaration with that of...
10869 (make_pass_rtl_fwprop): ...new function.
10870 (pass_rtl_fwprop_addr): Replace declaration with that of...
10871 (make_pass_rtl_fwprop_addr): ...new function.
10872 (pass_jump): Replace declaration with that of...
10873 (make_pass_jump): ...new function.
10874 (pass_jump2): Replace declaration with that of...
10875 (make_pass_jump2): ...new function.
10876 (pass_lower_subreg): Replace declaration with that of...
10877 (make_pass_lower_subreg): ...new function.
10878 (pass_cse): Replace declaration with that of...
10879 (make_pass_cse): ...new function.
10880 (pass_fast_rtl_dce): Replace declaration with that of...
10881 (make_pass_fast_rtl_dce): ...new function.
10882 (pass_ud_rtl_dce): Replace declaration with that of...
10883 (make_pass_ud_rtl_dce): ...new function.
10884 (pass_rtl_dce): Replace declaration with that of...
10885 (make_pass_rtl_dce): ...new function.
10886 (pass_rtl_dse1): Replace declaration with that of...
10887 (make_pass_rtl_dse1): ...new function.
10888 (pass_rtl_dse2): Replace declaration with that of...
10889 (make_pass_rtl_dse2): ...new function.
10890 (pass_rtl_dse3): Replace declaration with that of...
10891 (make_pass_rtl_dse3): ...new function.
10892 (pass_rtl_cprop): Replace declaration with that of...
10893 (make_pass_rtl_cprop): ...new function.
10894 (pass_rtl_pre): Replace declaration with that of...
10895 (make_pass_rtl_pre): ...new function.
10896 (pass_rtl_hoist): Replace declaration with that of...
10897 (make_pass_rtl_hoist): ...new function.
10898 (pass_rtl_store_motion): Replace declaration with that of...
10899 (make_pass_rtl_store_motion): ...new function.
10900 (pass_cse_after_global_opts): Replace declaration with that of...
10901 (make_pass_cse_after_global_opts): ...new function.
10902 (pass_rtl_ifcvt): Replace declaration with that of...
10903 (make_pass_rtl_ifcvt): ...new function.
10904 (pass_into_cfg_layout_mode): Replace declaration with that of...
10905 (make_pass_into_cfg_layout_mode): ...new function.
10906 (pass_outof_cfg_layout_mode): Replace declaration with that of...
10907 (make_pass_outof_cfg_layout_mode): ...new function.
10908 (pass_loop2): Replace declaration with that of...
10909 (make_pass_loop2): ...new function.
10910 (pass_rtl_loop_init): Replace declaration with that of...
10911 (make_pass_rtl_loop_init): ...new function.
10912 (pass_rtl_move_loop_invariants): Replace declaration with that of...
10913 (make_pass_rtl_move_loop_invariants): ...new function.
10914 (pass_rtl_unswitch): Replace declaration with that of...
10915 (make_pass_rtl_unswitch): ...new function.
10916 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
10917 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10918 (pass_rtl_doloop): Replace declaration with that of...
10919 (make_pass_rtl_doloop): ...new function.
10920 (pass_rtl_loop_done): Replace declaration with that of...
10921 (make_pass_rtl_loop_done): ...new function.
10922 (pass_web): Replace declaration with that of...
10923 (make_pass_web): ...new function.
10924 (pass_cse2): Replace declaration with that of...
10925 (make_pass_cse2): ...new function.
10926 (pass_df_initialize_opt): Replace declaration with that of...
10927 (make_pass_df_initialize_opt): ...new function.
10928 (pass_df_initialize_no_opt): Replace declaration with that of...
10929 (make_pass_df_initialize_no_opt): ...new function.
10930 (pass_reginfo_init): Replace declaration with that of...
10931 (make_pass_reginfo_init): ...new function.
10932 (pass_inc_dec): Replace declaration with that of...
10933 (make_pass_inc_dec): ...new function.
10934 (pass_stack_ptr_mod): Replace declaration with that of...
10935 (make_pass_stack_ptr_mod): ...new function.
10936 (pass_initialize_regs): Replace declaration with that of...
10937 (make_pass_initialize_regs): ...new function.
10938 (pass_combine): Replace declaration with that of...
10939 (make_pass_combine): ...new function.
10940 (pass_if_after_combine): Replace declaration with that of...
10941 (make_pass_if_after_combine): ...new function.
10942 (pass_ree): Replace declaration with that of...
10943 (make_pass_ree): ...new function.
10944 (pass_partition_blocks): Replace declaration with that of...
10945 (make_pass_partition_blocks): ...new function.
10946 (pass_match_asm_constraints): Replace declaration with that of...
10947 (make_pass_match_asm_constraints): ...new function.
10948 (pass_regmove): Replace declaration with that of...
10949 (make_pass_regmove): ...new function.
10950 (pass_split_all_insns): Replace declaration with that of...
10951 (make_pass_split_all_insns): ...new function.
10952 (pass_fast_rtl_byte_dce): Replace declaration with that of...
10953 (make_pass_fast_rtl_byte_dce): ...new function.
10954 (pass_lower_subreg2): Replace declaration with that of...
10955 (make_pass_lower_subreg2): ...new function.
10956 (pass_mode_switching): Replace declaration with that of...
10957 (make_pass_mode_switching): ...new function.
10958 (pass_sms): Replace declaration with that of...
10959 (make_pass_sms): ...new function.
10960 (pass_sched): Replace declaration with that of...
10961 (make_pass_sched): ...new function.
10962 (pass_ira): Replace declaration with that of...
10963 (make_pass_ira): ...new function.
10964 (pass_reload): Replace declaration with that of...
10965 (make_pass_reload): ...new function.
10966 (pass_clean_state): Replace declaration with that of...
10967 (make_pass_clean_state): ...new function.
10968 (pass_branch_prob): Replace declaration with that of...
10969 (make_pass_branch_prob): ...new function.
10970 (pass_value_profile_transformations): Replace declaration with that
10972 (make_pass_value_profile_transformations): ...new function.
10973 (pass_postreload_cse): Replace declaration with that of...
10974 (make_pass_postreload_cse): ...new function.
10975 (pass_gcse2): Replace declaration with that of...
10976 (make_pass_gcse2): ...new function.
10977 (pass_split_after_reload): Replace declaration with that of...
10978 (make_pass_split_after_reload): ...new function.
10979 (pass_branch_target_load_optimize1): Replace declaration with that
10981 (make_pass_branch_target_load_optimize1): ...new function.
10982 (pass_thread_prologue_and_epilogue): Replace declaration with that
10984 (make_pass_thread_prologue_and_epilogue): ...new function.
10985 (pass_stack_adjustments): Replace declaration with that of...
10986 (make_pass_stack_adjustments): ...new function.
10987 (pass_peephole2): Replace declaration with that of...
10988 (make_pass_peephole2): ...new function.
10989 (pass_if_after_reload): Replace declaration with that of...
10990 (make_pass_if_after_reload): ...new function.
10991 (pass_regrename): Replace declaration with that of...
10992 (make_pass_regrename): ...new function.
10993 (pass_cprop_hardreg): Replace declaration with that of...
10994 (make_pass_cprop_hardreg): ...new function.
10995 (pass_reorder_blocks): Replace declaration with that of...
10996 (make_pass_reorder_blocks): ...new function.
10997 (pass_branch_target_load_optimize2): Replace declaration with that
10999 (make_pass_branch_target_load_optimize2): ...new function.
11000 (pass_leaf_regs): Replace declaration with that of...
11001 (make_pass_leaf_regs): ...new function.
11002 (pass_split_before_sched2): Replace declaration with that of...
11003 (make_pass_split_before_sched2): ...new function.
11004 (pass_compare_elim_after_reload): Replace declaration with that of...
11005 (make_pass_compare_elim_after_reload): ...new function.
11006 (pass_sched2): Replace declaration with that of...
11007 (make_pass_sched2): ...new function.
11008 (pass_stack_regs): Replace declaration with that of...
11009 (make_pass_stack_regs): ...new function.
11010 (pass_stack_regs_run): Replace declaration with that of...
11011 (make_pass_stack_regs_run): ...new function.
11012 (pass_df_finish): Replace declaration with that of...
11013 (make_pass_df_finish): ...new function.
11014 (pass_compute_alignments): Replace declaration with that of...
11015 (make_pass_compute_alignments): ...new function.
11016 (pass_duplicate_computed_gotos): Replace declaration with that of...
11017 (make_pass_duplicate_computed_gotos): ...new function.
11018 (pass_variable_tracking): Replace declaration with that of...
11019 (make_pass_variable_tracking): ...new function.
11020 (pass_free_cfg): Replace declaration with that of...
11021 (make_pass_free_cfg): ...new function.
11022 (pass_machine_reorg): Replace declaration with that of...
11023 (make_pass_machine_reorg): ...new function.
11024 (pass_cleanup_barriers): Replace declaration with that of...
11025 (make_pass_cleanup_barriers): ...new function.
11026 (pass_delay_slots): Replace declaration with that of...
11027 (make_pass_delay_slots): ...new function.
11028 (pass_split_for_shorten_branches): Replace declaration with that of...
11029 (make_pass_split_for_shorten_branches): ...new function.
11030 (pass_split_before_regstack): Replace declaration with that of...
11031 (make_pass_split_before_regstack): ...new function.
11032 (pass_convert_to_eh_region_ranges): Replace declaration with that
11034 (make_pass_convert_to_eh_region_ranges): ...new function.
11035 (pass_shorten_branches): Replace declaration with that of...
11036 (make_pass_shorten_branches): ...new function.
11037 (pass_set_nothrow_function_flags): Replace declaration with that of...
11038 (make_pass_set_nothrow_function_flags): ...new function.
11039 (pass_dwarf2_frame): Replace declaration with that of...
11040 (make_pass_dwarf2_frame): ...new function.
11041 (pass_final): Replace declaration with that of...
11042 (make_pass_final): ...new function.
11043 (pass_rtl_seqabstr): Replace declaration with that of...
11044 (make_pass_rtl_seqabstr): ...new function.
11045 (pass_release_ssa_names): Replace declaration with that of...
11046 (make_pass_release_ssa_names): ...new function.
11047 (pass_early_inline): Replace declaration with that of...
11048 (make_pass_early_inline): ...new function.
11049 (pass_inline_parameters): Replace declaration with that of...
11050 (make_pass_inline_parameters): ...new function.
11051 (pass_update_address_taken): Replace declaration with that of...
11052 (make_pass_update_address_taken): ...new function.
11053 (pass_convert_switch): Replace declaration with that of...
11054 (make_pass_convert_switch): ...new function.
11055 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11056 to a subclass of simple_ipa_opt_pass along with...
11057 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11058 (make_pass_ipa_tree_profile): ...new function.
11059 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11060 subclass of gimple_opt_pass along with...
11061 (pass_data_sra_early): ...new pass_data instance and...
11062 (make_pass_sra_early): ...new function.
11063 (pass_sra): Convert from a global struct to a subclass of
11064 gimple_opt_pass along with...
11065 (pass_data_sra): ...new pass_data instance and...
11066 (make_pass_sra): ...new function.
11067 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11068 gimple_opt_pass along with...
11069 (pass_data_early_ipa_sra): ...new pass_data instance and...
11070 (make_pass_early_ipa_sra): ...new function.
11071 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11072 subclass of gimple_opt_pass along with...
11073 (pass_data_ccp): ...new pass_data instance and...
11074 (make_pass_ccp): ...new function.
11075 (pass_fold_builtins): Convert from a global struct to a subclass of
11076 gimple_opt_pass along with...
11077 (pass_data_fold_builtins): ...new pass_data instance and...
11078 (make_pass_fold_builtins): ...new function.
11079 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11080 subclass of gimple_opt_pass along with...
11081 (pass_data_copy_prop): ...new pass_data instance and...
11082 (make_pass_copy_prop): ...new function.
11083 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11084 global struct to a subclass of gimple_opt_pass along with...
11085 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11086 (make_pass_rename_ssa_copies): ...new function.
11087 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11088 subclass of gimple_opt_pass along with...
11089 (pass_data_dce): ...new pass_data instance and...
11090 (make_pass_dce): ...new function.
11091 (pass_dce_loop): Convert from a global struct to a subclass of
11092 gimple_opt_pass along with...
11093 (pass_data_dce_loop): ...new pass_data instance and...
11094 (make_pass_dce_loop): ...new function.
11095 (pass_cd_dce): Convert from a global struct to a subclass of
11096 gimple_opt_pass along with...
11097 (pass_data_cd_dce): ...new pass_data instance and...
11098 (make_pass_cd_dce): ...new function.
11099 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11100 subclass of gimple_opt_pass along with...
11101 (pass_data_dominator): ...new pass_data instance and...
11102 (make_pass_dominator): ...new function.
11103 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11104 gimple_opt_pass along with...
11105 (pass_data_phi_only_cprop): ...new pass_data instance and...
11106 (make_pass_phi_only_cprop): ...new function.
11107 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11108 subclass of gimple_opt_pass along with...
11109 (pass_data_dse): ...new pass_data instance and...
11110 (make_pass_dse): ...new function.
11111 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11112 a subclass of gimple_opt_pass along with...
11113 (pass_data_forwprop): ...new pass_data instance and...
11114 (make_pass_forwprop): ...new function.
11115 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11116 struct to a subclass of gimple_opt_pass along with...
11117 (pass_data_tree_ifcombine): ...new pass_data instance and...
11118 (make_pass_tree_ifcombine): ...new function.
11119 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11120 subclass of gimple_opt_pass along with...
11121 (pass_data_ch): ...new pass_data instance and...
11122 (make_pass_ch): ...new function.
11123 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11124 subclass of gimple_opt_pass along with...
11125 (pass_data_tree_loop): ...new pass_data instance and...
11126 (make_pass_tree_loop): ...new function.
11127 (pass_tree_loop_init): Convert from a global struct to a subclass of
11128 gimple_opt_pass along with...
11129 (pass_data_tree_loop_init): ...new pass_data instance and...
11130 (make_pass_tree_loop_init): ...new function.
11131 (pass_lim): Convert from a global struct to a subclass of
11132 gimple_opt_pass along with...
11133 (pass_data_lim): ...new pass_data instance and...
11134 (make_pass_lim): ...new function.
11135 (pass_tree_unswitch): Convert from a global struct to a subclass of
11136 gimple_opt_pass along with...
11137 (pass_data_tree_unswitch): ...new pass_data instance and...
11138 (make_pass_tree_unswitch): ...new function.
11139 (pass_predcom): Convert from a global struct to a subclass of
11140 gimple_opt_pass along with...
11141 (pass_data_predcom): ...new pass_data instance and...
11142 (make_pass_predcom): ...new function.
11143 (pass_vectorize): Convert from a global struct to a subclass of
11144 gimple_opt_pass along with...
11145 (pass_data_vectorize): ...new pass_data instance and...
11146 (make_pass_vectorize): ...new function.
11147 (pass_graphite): Convert from a global struct to a subclass of
11148 gimple_opt_pass along with...
11149 (pass_data_graphite): ...new pass_data instance and...
11150 (make_pass_graphite): ...new function.
11151 (pass_graphite_transforms): Convert from a global struct to a subclass
11152 of gimple_opt_pass along with...
11153 (pass_data_graphite_transforms): ...new pass_data instance and...
11154 (make_pass_graphite_transforms): ...new function.
11155 (pass_check_data_deps): Convert from a global struct to a subclass of
11156 gimple_opt_pass along with...
11157 (pass_data_check_data_deps): ...new pass_data instance and...
11158 (make_pass_check_data_deps): ...new function.
11159 (pass_iv_canon): Convert from a global struct to a subclass of
11160 gimple_opt_pass along with...
11161 (pass_data_iv_canon): ...new pass_data instance and...
11162 (make_pass_iv_canon): ...new function.
11163 (pass_scev_cprop): Convert from a global struct to a subclass of
11164 gimple_opt_pass along with...
11165 (pass_data_scev_cprop): ...new pass_data instance and...
11166 (make_pass_scev_cprop): ...new function.
11167 (pass_record_bounds): Convert from a global struct to a subclass of
11168 gimple_opt_pass along with...
11169 (pass_data_record_bounds): ...new pass_data instance and...
11170 (make_pass_record_bounds): ...new function.
11171 (pass_complete_unroll): Convert from a global struct to a subclass of
11172 gimple_opt_pass along with...
11173 (pass_data_complete_unroll): ...new pass_data instance and...
11174 (make_pass_complete_unroll): ...new function.
11175 (pass_complete_unrolli): Convert from a global struct to a subclass of
11176 gimple_opt_pass along with...
11177 (pass_data_complete_unrolli): ...new pass_data instance and...
11178 (make_pass_complete_unrolli): ...new function.
11179 (pass_parallelize_loops): Convert from a global struct to a subclass
11180 of gimple_opt_pass along with...
11181 (pass_data_parallelize_loops): ...new pass_data instance and...
11182 (make_pass_parallelize_loops): ...new function.
11183 (pass_loop_prefetch): Convert from a global struct to a subclass of
11184 gimple_opt_pass along with...
11185 (pass_data_loop_prefetch): ...new pass_data instance and...
11186 (make_pass_loop_prefetch): ...new function.
11187 (pass_iv_optimize): Convert from a global struct to a subclass of
11188 gimple_opt_pass along with...
11189 (pass_data_iv_optimize): ...new pass_data instance and...
11190 (make_pass_iv_optimize): ...new function.
11191 (pass_tree_loop_done): Convert from a global struct to a subclass of
11192 gimple_opt_pass along with...
11193 (pass_data_tree_loop_done): ...new pass_data instance and...
11194 (make_pass_tree_loop_done): ...new function.
11195 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11196 struct to a subclass of gimple_opt_pass along with...
11197 (pass_data_cse_reciprocals): ...new pass_data instance and...
11198 (make_pass_cse_reciprocals): ...new function.
11199 (pass_cse_sincos): Convert from a global struct to a subclass of
11200 gimple_opt_pass along with...
11201 (pass_data_cse_sincos): ...new pass_data instance and...
11202 (make_pass_cse_sincos): ...new function.
11203 (pass_optimize_bswap): Convert from a global struct to a subclass of
11204 gimple_opt_pass along with...
11205 (pass_data_optimize_bswap): ...new pass_data instance and...
11206 (make_pass_optimize_bswap): ...new function.
11207 (pass_optimize_widening_mul): Convert from a global struct to a
11208 subclass of gimple_opt_pass along with...
11209 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11210 (make_pass_optimize_widening_mul): ...new function.
11211 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11212 subclass of gimple_opt_pass along with...
11213 (pass_data_phiopt): ...new pass_data instance and...
11214 (make_pass_phiopt): ...new function.
11215 (pass_cselim): Convert from a global struct to a subclass of
11216 gimple_opt_pass along with...
11217 (pass_data_cselim): ...new pass_data instance and...
11218 (make_pass_cselim): ...new function.
11219 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11220 subclass of gimple_opt_pass along with...
11221 (pass_data_phiprop): ...new pass_data instance and...
11222 (make_pass_phiprop): ...new function.
11223 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11224 subclass of gimple_opt_pass along with...
11225 (pass_data_pre): ...new pass_data instance and...
11226 (make_pass_pre): ...new function.
11227 (pass_fre): Convert from a global struct to a subclass of
11228 gimple_opt_pass along with...
11229 (pass_data_fre): ...new pass_data instance and...
11230 (make_pass_fre): ...new function.
11231 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11232 subclass of gimple_opt_pass along with...
11233 (pass_data_reassoc): ...new pass_data instance and...
11234 (make_pass_reassoc): ...new function.
11235 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11236 subclass of gimple_opt_pass along with...
11237 (pass_data_sink_code): ...new pass_data instance and...
11238 (make_pass_sink_code): ...new function.
11239 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11240 subclass of gimple_opt_pass along with...
11241 (pass_data_strlen): ...new pass_data instance and...
11242 (make_pass_strlen): ...new function.
11243 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11244 struct to a subclass of gimple_opt_pass along with...
11245 (pass_data_build_alias): ...new pass_data instance and...
11246 (make_pass_build_alias): ...new function.
11247 (pass_build_ealias): Convert from a global struct to a subclass of
11248 gimple_opt_pass along with...
11249 (pass_data_build_ealias): ...new pass_data instance and...
11250 (make_pass_build_ealias): ...new function.
11251 (pass_ipa_pta): Convert from a global struct to a subclass of
11252 simple_ipa_opt_pass along with...
11253 (pass_data_ipa_pta): ...new pass_data instance and...
11254 (make_pass_ipa_pta): ...new function.
11255 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11256 subclass of gimple_opt_pass along with...
11257 (pass_data_uncprop): ...new pass_data instance and...
11258 (make_pass_uncprop): ...new function.
11259 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11260 global struct to a subclass of gimple_opt_pass along with...
11261 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11262 (make_pass_late_warn_uninitialized): ...new function.
11263 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11264 to a subclass of gimple_opt_pass along with...
11265 (pass_data_init_datastructures): ...new pass_data instance and...
11266 (make_pass_init_datastructures): ...new function.
11267 (pass_early_warn_uninitialized): Convert from a global struct to a
11268 subclass of gimple_opt_pass along with...
11269 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11270 (make_pass_early_warn_uninitialized): ...new function.
11271 (pass_update_address_taken): Convert from a global struct to a
11272 subclass of gimple_opt_pass along with...
11273 (pass_data_update_address_taken): ...new pass_data instance and...
11274 (make_pass_update_address_taken): ...new function.
11275 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11276 struct to a subclass of gimple_opt_pass along with...
11277 (pass_data_release_ssa_names): ...new pass_data instance and...
11278 (make_pass_release_ssa_names): ...new function.
11279 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11280 subclass of gimple_opt_pass along with...
11281 (pass_data_stdarg): ...new pass_data instance and...
11282 (make_pass_stdarg): ...new function.
11283 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11284 global struct to a subclass of gimple_opt_pass along with...
11285 (pass_data_convert_switch): ...new pass_data instance and...
11286 (make_pass_convert_switch): ...new function.
11287 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11288 to a subclass of gimple_opt_pass along with...
11289 (pass_data_tail_recursion): ...new pass_data instance and...
11290 (make_pass_tail_recursion): ...new function.
11291 (pass_tail_calls): Convert from a global struct to a subclass of
11292 gimple_opt_pass along with...
11293 (pass_data_tail_calls): ...new pass_data instance and...
11294 (make_pass_tail_calls): ...new function.
11295 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11296 struct to a subclass of gimple_opt_pass along with...
11297 (pass_data_lower_vector): ...new pass_data instance and...
11298 (make_pass_lower_vector): ...new function.
11299 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11300 gimple_opt_pass along with...
11301 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11302 (make_pass_lower_vector_ssa): ...new function.
11303 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11304 to a subclass of gimple_opt_pass along with...
11305 (pass_data_slp_vectorize): ...new pass_data instance and...
11306 (make_pass_slp_vectorize): ...new function.
11307 (pass_ipa_increase_alignment): Convert from a global struct to a
11308 subclass of simple_ipa_opt_pass along with...
11309 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11310 (make_pass_ipa_increase_alignment): ...new function.
11311 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11312 gimple_opt_pass along with...
11313 (pass_data_vrp): ...new pass_data instance and...
11314 (make_pass_vrp): ...new function.
11315 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11316 subclass of simple_ipa_opt_pass along with...
11317 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11318 (make_pass_ipa_free_lang_data): ...new function.
11319 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11320 gimple_opt_pass along with...
11321 (pass_data_tsan): ...new pass_data instance and...
11322 (make_pass_tsan): ...new function.
11323 (pass_tsan_O0): Convert from a global struct to a subclass of
11324 gimple_opt_pass along with...
11325 (pass_data_tsan_O0): ...new pass_data instance and...
11326 (make_pass_tsan_O0): ...new function.
11327 * var-tracking.c (pass_variable_tracking): Convert from a global
11328 struct to a subclass of rtl_opt_pass along with...
11329 (pass_data_variable_tracking): ...new pass_data instance and...
11330 (make_pass_variable_tracking): ...new function.
11331 * web.c (pass_web): Convert from a global struct to a subclass of
11332 rtl_opt_pass along with...
11333 (pass_data_web): ...new pass_data instance and...
11334 (make_pass_web): ...new function.
11335 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11336 declaration with that of...
11337 (make_pass_mode_switch_use): ...new function.
11338 (pass_resolve_sw_modes): Replace declaration with that of...
11339 (make_pass_resolve_sw_modes): ...new function.
11340 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11341 from a global struct to a subclass of rtl_opt_pass along with...
11342 (pass_data_mode_switch_use): ...new pass_data instance and...
11343 (make_pass_mode_switch_use): ...new function.
11344 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11345 from a global struct to a subclass of rtl_opt_pass along with...
11346 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11347 (make_pass_resolve_sw_modes): ...new function.
11348 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11349 struct to a subclass of rtl_opt_pass along with...
11350 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11351 (make_pass_insert_vzeroupper): ...new function.
11352 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11353 global struct to a subclass of rtl_opt_pass along with...
11354 (pass_data_work_around_errata): ...new pass_data instance and...
11355 (make_pass_work_around_errata): ...new function.
11356 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11357 struct to a subclass of rtl_opt_pass along with...
11358 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11359 (make_pass_mips_machine_reorg2): ...new function.
11361 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11363 * passes.c (pass_manager::operator new): New.
11365 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11367 Handwritten part of conversion of passes to C++ classes.
11369 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11370 (toplev.o): Add dep on PASS_MANAGER_H.
11371 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11372 of early local pases to reflect this moving from a global to a
11373 member of gcc::pass_manager.
11374 (cgraph_add_new_function): Likewise.
11375 * lto-cgraph.c (lto_output_node): Update for conversion of
11376 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11377 * passes.c (opt_pass::clone): New.
11378 (opt_pass::gate): New.
11379 (opt_pass::execute): New.
11380 (opt_pass::opt_pass): New.
11381 (pass_manager::execute_early_local_passes): New.
11382 (pass_manager::execute_pass_mode_switching): new.
11383 (finish_optimization_passes): Convert to...
11384 (pass_manager::finish_optimization_passes): ...this.
11385 (finish_optimization_passes): Update for conversion of passes to
11387 (register_dump_files_1): Use has_gate since we cannot portably
11388 check a vtable entry against NULL.
11389 (dump_one_pass): Likewise.
11390 (ipa_write_summaries_2): Likewise.
11391 (ipa_write_optimization_summaries_1): Likewise.
11392 (ipa_read_summaries_1): Likewise.
11393 (ipa_read_optimization_summaries_1): Likewise.
11394 (execute_ipa_stmt_fixups): Likewise.
11395 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11396 pass-creation functions rather than wiring up globals, and
11397 storing the results in fields of pass_manager generated using
11398 pass-instances.def.
11399 (pass_manager::dump_profile_report): Update for conversion of
11400 passes to C++ classes.
11401 (pass_manager::execute_ipa_summary_passes): Likewise.
11402 (execute_one_ipa_transform_pass): Likewise.
11403 (execute_one_pass): Use has_gate and has_execute since we cannot
11404 portably check a vtable entry against NULL.
11405 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11406 (pass_manager): Use pass-instances.def to add fields for the
11407 various pass instances.
11408 * toplev.c (finalize): Update for move of
11409 finish_optimization_passes to a method of gcc::pass_manager.
11410 * toplev.h (finish_optimization_passes): Move to method of class
11412 * tree-pass.h (struct pass_data): New.
11413 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11414 (opt_pass::gate): Convert to virtual function.
11415 (opt_pass::~opt_pass): New.
11416 (opt_pass::clone): New.
11417 (opt_pass::execute): Convert to virtual function.
11418 (opt_pass::opt_pass): New.
11419 (opt_pass::ctxt_): new.
11420 (gimple_opt_pass): Convert to subclass of opt_pass.
11421 (gimple_opt_pass::gimple_opt_pass): New.
11422 (rtl_opt_pass): Convert to subclass of opt_pass.
11423 (rtl_opt_pass::rtl_opt_pass): New.
11424 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11425 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11426 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11427 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11428 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11429 invocation of pass_mode_switching to reflect this moving from a
11430 global to a member of gcc::pass_manager.
11431 (ix86_option_override): Rework how pass_insert_vzeroupper is
11432 added to the pass_manager to reflect autogenerated changes.
11433 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11435 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11437 PR rtl-optimization/57708
11438 * recog.c (peep2_find_free_register): Validate all regs in a
11441 2013-08-05 Jan Hubicka <jh@suse.cz>
11444 * cgraph.c (verify_cgraph_node): Accept local flags from other
11446 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11447 (function_and_variable_visibility): Likewise.
11448 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11450 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11452 * graph.c (init_graph_slim_pretty_print): Remove.
11453 (print_graph_cfg): Do not call it. Use local pretty printer.
11454 (start_graph_dump): Likewise.
11456 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11458 * gimple-pretty-print.c (buffer): Remove.
11459 (initialized): Likewise.
11460 (maybe_init_pretty_print): Likewise.
11461 (print_gimple_stmt): Do not call it. Use non-static local
11462 pretty_printer variable.
11463 (print_gimple_expr): Likewise.
11464 (print_gimple_seq): Likewise.
11465 (gimple_dump_bb): Likewise.
11467 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11469 * asan.c (asan_pp): Remove.
11470 (asan_pp_initialized): Likewise.
11471 (asan_pp_initialize): Likewise.
11472 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11473 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11474 (asan_add_global): Likewise.
11476 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11478 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11479 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11480 * pretty-print.c (pp_formatted_text_data): Likewise.
11481 (pp_indent): Rename from pp_base_indent.
11482 (pp_format): Rename from pp_base_format.
11483 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11484 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11485 (pp_flush): Rename from pp_base_flush.
11486 (pp_set_line_maximum_length): Rename from
11487 pp_base_set_line_maximum_length.
11488 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11489 (pp_set_prefix): Rename from pp_base_set_prefix.
11490 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11491 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11492 (pp_append_text): Rename from pp_base_append_text.
11493 (pp_formatted_text): Rename from pp_base_formatted_text.
11494 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11495 (pp_remaining_character_count_for_line): Rename from
11496 pp_base_remaining_character_count_for_line.
11497 (pp_newline): Rename from pp_base_newline.
11498 (pp_character): Rename from pp_base_character.
11499 (pp_string): Rename from pp_base_string.
11500 (pp_maybe_space): Rename from pp_base_maybe_space.
11501 * asan.c (asan_pp_string): Adjust.
11502 (asan_emit_stack_protection): Likewise.
11503 (asan_add_global): Likewise.
11504 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11505 * tree-mudflap.c (mf_varname_tree): Likewise.
11506 * tree-pretty-print.c (pp_tree_identifier): Rename from
11507 pp_base_tree_identifier.
11508 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11509 Declare as function.
11511 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11513 * pretty-print.h (pp_bar_bar): New.
11514 (pp_ampersand_ampersand): Likewise.
11515 (pp_less_equal): Likewise.
11516 (pp_greater_equal): Likewise.
11517 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11518 printer functions instead of pp_string or operators and punctuators.
11519 (dump_gimple_call): Likewise.
11520 (dump_gimple_omp_for): Likewise.
11521 (dump_gimple_transaction): Likewise.
11522 (dump_gimple_phi): Likewise.
11523 (pp_gimple_stmt_1): Likewise.
11524 * sched-vis.c (print_insn): Likewise.
11525 * tree-mudflap.c (mf_varname_tree): Likewise.
11526 * tree-pretty-print.c (dump_block_node): Likewise.
11527 (dump_generic_node): Likewise.
11529 2013-08-02 Jan Hubicka <jh@suse.cz>
11531 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11533 * lto-streamer-out.c (tree_is_indexable): Results decls and
11534 parm decls are not indexable.
11535 (DFS_write_tree_body): Do not follow args and results.
11536 (hash_tree): Likewise.
11537 (output_functions): Rearrange so struct function is needed
11538 only when real body is output; be able to also ouptut abstract
11539 functions; output DECL_ARGUMENTS and DECL_RESULT.
11540 (lto_output): When not in WPA, ale store abstract functions.
11541 (write_symbol): Do not care about RESULT_DECL.
11542 (output_symbol_p): Handle correctly sbtract decls.
11543 * lto-streamer-in.c (input_function): Rearrange so struct
11544 function can be NULL at entry; allow streaming of
11545 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11546 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11547 sanity check during LTO.
11548 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11549 RESULT_DECl and DECL_ARGUMENTS.
11550 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11553 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11555 * pretty-print.h (pp_underscore): New.
11557 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11558 printer functions instead of pp_character.
11559 (dump_binary_rhs): Likewise.
11560 (dump_ternary_rhs): Likewise.
11561 (dump_gimple_call_args): Likewise.
11562 (pp_points_to_solution): Likewise.
11563 (dump_gimple_call): Likewise.
11564 (dump_gimple_switch): Likewise.
11565 (dump_gimple_cond): Likewise.
11566 (dump_gimple_bind): Likewise.
11567 (dump_gimple_try): Likewise.
11568 (dump_gimple_omp_for): Likewise.
11569 (dump_gimple_omp_continue): Likewise.
11570 (dump_gimple_omp_single): Likewise.
11571 (dump_gimple_omp_sections): Likewise.
11572 (dump_gimple_omp_block): Likewise.
11573 (dump_gimple_omp_critical): Likewise.
11574 (dump_gimple_transaction): Likewise.
11575 (dump_gimple_asm): Likewise.
11576 (dump_gimple_phi): Likewise.
11577 (dump_gimple_omp_parallel): Likewise.
11578 (dump_gimple_omp_task): Likewise.
11579 (dump_gimple_omp_atomic_load): Likewise.
11580 (dump_gimple_omp_atomic_store): Likewise.
11581 (dump_gimple_mem_ops): Likewise.
11582 (pp_gimple_stmt_1): Likewise.
11583 (pp_cfg_jump): Likewise.
11584 (dump_implicit_edges): Likewise.
11585 (gimple_dump_bb_for_graph): Likewise.
11586 * graph.c (draw_cfg_node): Likewise.
11587 * langhooks.c (lhd_print_error_function): Likewise.
11588 * sched-vis.c (print_exp): Likewise.
11589 (print_value): Likewise.
11590 (print_pattern): Likewise.
11591 (print_insn): Likewise.
11592 (rtl_dump_bb_for_graph): Likewise.
11593 * tree-pretty-print.c (dump_function_declaration): Likewise.
11594 (dump_array_domain): Likewise.
11595 (dump_omp_clause): Likewise.
11596 (dump_location): Likewise.
11597 (dump_generic_node): Likewise.
11598 (print_struct_decl): Likewise.
11599 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11601 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11603 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11604 candidate table when replacing a candidate statement.
11605 (replace_rhs_if_not_dup): Likewise.
11606 (replace_one_candidate): Likewise.
11608 2013-08-02 Jan Hubicka <jh@suse.cz>
11609 Martin Liska <marxin.liska@gmail.com>
11611 * cgraphunit.c (add_new_function): Fix logic when adding from
11613 (assemble_thunk): Rename to ...
11614 (expand_thunk); .. this one; export; get it working with
11615 general functions; make produced gimple valid.
11616 * cgraph.h (expand_thunk): Declare.
11618 2013-08-02 Jan Hubicka <jh@suse.cz>
11620 * ipa-cp.c (gather_context_independent_values): Use
11621 ipa_get_param_move_cost.
11622 (get_replacement_map): Remove PARAM; move parameter folding
11624 (create_specialized_node): Update.
11625 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11626 assert that we have gimple body; update move_cost.
11627 (count_formal_params): Assert that we have gimple body.
11628 (ipa_dump_param): New function.
11629 (ipa_alloc_node_params): Break out from ...
11630 (ipa_initialize_node_params): ... here.
11631 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11632 (ipa_write_node_info): Stream move costs.
11633 (ipa_read_node_info): Read move costs.
11634 (ipa_update_after_lto_read): Do not recompute node params.
11635 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11636 (ipa_get_param): Check we are not in WPA.
11637 (ipa_get_param_move_cost): New.
11638 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11639 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11640 parm numbers to be present.
11642 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11644 PR rtl-optimization/58048
11645 * lra-constraints.c (process_alt_operands): Don't check asm
11646 operand on register.
11648 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11650 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11651 the implied StoreLoad barrier for atomic operations if before.
11653 2013-08-02 Jan Hubicka <jh@suse.cz>
11654 Martin Liska <marxin.liska@gmail.com>
11656 * cgraph.c (cgraph_function_body_availability): Do not check
11658 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11659 symtab_node_availability): Declare.
11660 * ipa.c (can_replace_by_local_alias): New.
11661 (function_and_variable_visibility): Use it.
11662 * symtab.c (symtab_for_node_and_aliases,
11663 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11665 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11667 PR rtl-optimization/57963
11668 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11669 (lra_constraints): Use them.
11671 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11673 * config/arm/types.md (define_attr "type"): Add "load_acq"
11675 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11677 (cortex_a53_store1): Likewise.
11679 2013-08-01 Jan Hubicka <jh@suse.cz>
11681 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11682 partitions are not needed.
11684 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11686 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11687 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11688 MAYBE_NON_Q_CLASS_P where appropriate.
11690 2013-08-01 Jan Hubicka <jh@suse.cz>
11692 * cgraph.h (release_function_body): Declare.
11693 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11694 of unused delcarations.
11696 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11698 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11699 RTL form when subtracting a constant.
11701 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11703 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11704 Generate canonical plus rtx with negated immediate instead of minus
11706 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11708 2013-08-01 Jan Hubicka <jh@suse.cz>
11710 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11711 (cgraph_release_function_body): Likewise.
11712 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11713 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11714 to used_as_abstract_origin.
11715 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11716 symbols used as abstract origins.
11717 * cgraphunit.c (analyze_functions): Update.
11718 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11719 used_as_abstract_origin.
11720 * tree-inline.c (tree_function_versioning): Update
11721 used_as_abstract_origin; be ready for DECL_RESULT and
11722 DECL_ARGUMENTS to be NULL.
11724 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11725 for abstract functions.
11726 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11729 2013-08-01 Jan Hubicka <jh@suse.cz>
11731 * profile.c (compute_value_histograms): Fix thinko.
11733 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11735 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11736 aarch-common-protos.h to extra_headers.
11737 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11738 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11739 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11741 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11743 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11744 Include "../arm/types.md". Define "type" attribute for all patterns.
11745 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11748 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11750 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11751 to support power8 load fusion.
11752 (fusion_gpr_mem_load): Likewise.
11754 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11756 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11757 declarations for power8 load fusion.
11758 (emit_fusion_gpr_load): Likewise.
11760 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11761 tuning for power8, turn on fusion mode by default. Turn on sign
11762 extending fusion mode if normal fusion mode is on, and we are at
11764 (fusion_gpr_load_p): New function, return true if we can fuse an
11765 addis instruction with a dependent load to a GPR.
11766 (emit_fusion_gpr_load): Emit the instructions for power8 load
11769 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11770 (VSX load fusion peepholes): New peepholes to fuse together an
11771 addi instruction with a VSX load instruction.
11773 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11774 peepholes to fuse an addis instruction with a load to a GPR base
11775 register. If we are supporting sign extending fusions, convert
11776 sign extending loads to zero extending loads and add an explicit
11779 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11781 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11782 aarch-common-protos.h to extra_headers.
11783 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11784 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11785 (arm_early_store_addr_dep): Likewise.
11786 (arm_no_early_alu_shift_dep): Likewise.
11787 (arm_no_early_alu_shift_value_dep): Likewise.
11788 (arm_no_early_mul_dep): Likewise.
11789 (arm_no_early_store_addr_dep): Likewise.
11790 (arm_mac_accumulator_is_mul_result): Likewise.
11791 (arm_mac_accumulator_is_result): Likewise.
11792 * config/arm/aarch-common.c: ... here. New file.
11793 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11795 (arm_early_store_addr_dep): Likewise.
11796 (arm_no_early_alu_shift_dep): Likewise.
11797 (arm_no_early_alu_shift_value_dep): Likewise.
11798 (arm_no_early_mul_dep): Likewise.
11799 (arm_no_early_store_addr_dep): Likewise.
11800 (arm_mac_accumulator_is_mul_result): Likewise.
11801 (arm_mac_accumulator_is_result): Likewise.
11802 * config/arm/aarch-common-protos.h: ... here. New file.
11803 * config/arm/t-arm (aarch-common.o): Define.
11805 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11807 * config/arm/arm.md: Include new file "types.md".
11808 (define_attr "type"): Move from here to ...
11809 (define_attr "mul32"): Likewise.
11810 (define_attr "mul64"): Likewise.
11811 * config/arm/types.md: ... here. New file.
11813 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11815 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11816 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11818 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11820 * gen-pass-instances.awk: Fix offset of substr().
11822 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11824 * Makefile.in (pass-instances.def): New.
11825 (passes.o): Replace dependency on passes.def with one on
11828 * gen-pass-instances.awk: New.
11830 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11831 rather than passes.def, updating local definition of NEXT_PASS
11832 macro to add an extra NUM parameter (currently unused).
11834 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11836 * Makefile.in (PASS_MANAGER_H): New.
11837 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11838 (passes.o): Likewise.
11839 (statistics.o): Likewise.
11840 (cgraphunit.o): Likewise.
11841 (context.o): Depend on PASS_MANAGER_H.
11843 * pass_manager.h: New.
11845 * cgraphunit.c (cgraph_add_new_function): Update for moves
11846 of globals to fields of pass_manager.
11847 (analyze_function): Likewise.
11848 (expand_function): Likewise.
11849 (ipa_passes): Likewise.
11850 (compile): Likewise.
11852 * context.c (context::context): New.
11853 * context.h (context::context): New.
11854 (context::get_passes): New.
11855 (context::passes_): New.
11857 * lto-cgraph.c (input_node): Update for moves of globals to
11858 fields of pass_manager.
11860 * passes.c (all_passes): Remove, in favor of a field of the
11861 same name within the new class pass_manager.
11862 (all_small_ipa_passes): Likewise.
11863 (all_lowering_passes): Likewise.
11864 (all_regular_ipa_passes): Likewise.
11865 (all_late_ipa_passes): Likewise.
11866 (all_lto_gen_passes): Likewise.
11867 (passes_by_id): Likewise.
11868 (passes_by_id_size): Likewise.
11869 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11870 the new class pass_manager.
11871 (set_pass_for_id): Convert to...
11872 (pass_manager::set_pass_for_id): ...method.
11873 (get_pass_for_id): Convert to...
11874 (pass_manager::get_pass_for_id): ...method.
11875 (register_one_dump_file): Move body of implementation into...
11876 (pass_manager::register_one_dump_file): ...here.
11877 (register_dump_files_1): Convert to...
11878 (pass_manager::register_dump_files_1): ...method.
11879 (register_dump_files): Convert to...
11880 (pass_manager::register_dump_files): ...method.
11881 (create_pass_tab): Update for moves of globals to fields of
11883 (dump_passes): Move body of implementation into...
11884 (pass_manager::dump_passes): ...here.
11885 (register_pass): Move body of implementation into...
11886 (pass_manager::register_pass): ...here.
11887 (init_optimization_passes): Convert into...
11888 (pass_manager::pass_manager): ...constructor for new
11889 pass_manager class, and initialize the pass_lists array.
11890 (check_profile_consistency): Update for moves of globals to
11891 fields of pass_manager.
11892 (dump_profile_report): Move body of implementation into...
11893 (pass_manager::dump_profile_report): ...here.
11894 (ipa_write_summaries_1): Update for moves of pass lists from
11895 being globals to fields of pass_manager.
11896 (ipa_write_optimization_summaries): Likewise.
11897 (ipa_read_summaries): Likewise.
11898 (ipa_read_optimization_summaries): Likewise.
11899 (execute_all_ipa_stmt_fixups): Likewise.
11901 * statistics.c (statistics_fini): Update for moves of globals to
11902 fields of pass_manager.
11904 * toplev.c (general_init): Replace call to
11905 init_optimization_passes with construction of the pass_manager
11908 * tree-pass.h (all_passes): Remove, in favor of a field of the
11909 same name within the new class pass_manager.
11910 (all_small_ipa_passes): Likewise.
11911 (all_lowering_passes): Likewise.
11912 (all_regular_ipa_passes): Likewise.
11913 (all_lto_gen_passes): Likewise.
11914 (all_late_ipa_passes): Likewise.
11915 (passes_by_id): Likewise.
11916 (passes_by_id_size): Likewise.
11917 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11918 the new class pass_manager.
11919 (get_pass_for_id): Remove.
11921 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11923 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
11926 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
11928 * arm.md (mulhi3): New expand pattern.
11930 2013-07-30 Jan Hubicka <jh@suse.cz>
11931 Martin Liska <marxin.liska@gmail.com>
11933 * profile.c (compute_value_histograms): Do not ICE when
11934 there is mismatch only on some counters.
11936 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11938 PR rtl-optimization/57637
11939 * function.c (move_insn_for_shrink_wrap): Also check the
11940 GEN set of the LIVE problem for the liveness analysis
11941 if it exists, otherwise give up.
11943 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11945 PR tree-optimization/57993
11946 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
11947 replaced statement in the candidate table.
11948 (phi_add_costs): Return infinite cost when the hidden basis does
11949 not dominate all phis on which the candidate is dependent.
11950 (replace_one_candidate): Record replaced statement in the
11953 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
11955 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
11956 (ashlv2si3): New expander.
11957 (*ashlv2si3_i): New define_insn_and_split.
11958 * predicates.md (float_operation): Allow patterns with three
11959 basic sub-patterns.
11961 PR rtl-optimization/58021
11962 * mode-switching.c (create_pre_exit): Always split off preceding
11963 insns if we are not at the basic block head.
11965 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
11967 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
11968 (UCLIBC_DYNAMIC_LINKER): New macro.
11969 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
11971 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
11972 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11974 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11975 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11976 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11977 for SF and DF modes. Use ieee_quad_format for TF mode.
11978 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11979 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11980 (mips_option_override): Handle `-mnan=legacy'.
11981 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11982 `-mabs=2008' and `-mnan=2008'.
11983 (OPTION_DEFAULT_SPECS): Add "nan" default.
11984 (ASM_SPEC): Handle `-mnan='.
11985 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11986 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11987 comment accordingly.
11988 (neg<mode>2): Likewise.
11989 * config/mips/mips.opt (mabs, mnan): New options.
11990 * doc/install.texi (Configuration): Document `--with-nan=' option.
11991 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11993 (MIPS Options): Document them.
11994 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11995 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11996 * configure: Regenerate.
11997 * config.in: Regenerate.
11999 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12001 * config/i386/i386.md (float post-reload splitters): Do not check
12002 for subregs of SSE registers.
12004 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12005 H.J. Lu <hongjiu.lu@intel.com>
12009 * config/i386/i386.md (post-reload splitter
12010 to avoid partial SSE reg dependency stalls): New pattern.
12012 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12014 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12016 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12018 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12019 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12021 2013-07-26 Cary Coutant <ccoutant@google.com>
12023 * dwarf2out.c (die_checksum_ordered): Don't include template
12024 instantiations in signature.
12025 (is_template_parameter): New function.
12026 (is_template_instantiation): New function.
12027 (generate_skeleton_bottom_up): Don't include template instantiations
12029 (generate_skeleton): Likewise.
12030 (break_out_comdat_types): Move recursive call to break out nested
12032 (prune_unused_types_mark_generic_parms_dies): Call
12033 is_template_parameter.
12035 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12037 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12038 uses vector registers.
12039 * config/aarch64/iterators.md: Add attributes rtn and vas.
12041 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12042 Richard Earnshaw <richard.earnshaw@arm.com>
12044 * combine.c (simplify_comparison): Re-canonicalize operands
12046 * config/arm/arm.md (movcond_addsi): New splitter.
12048 2013-07-25 Sterling Augustine <saugustine@google.com>
12050 * dwarf2out.c (size_of_pubnames): Move code to...
12051 (include_pubname_in_output): ...here. New.
12052 (want_pubnames): Rearrange.
12053 (output_pubnames): Call include_pubname_in_output. Move assertion.
12055 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12057 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12059 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12062 * doc/extend.texi: Remove obsolete builtins. Fix
12063 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12065 2013-07-25 Jan Hubicka <jh@suse.cz>
12067 * cgraph.c (release_function_body): Break out from ...
12068 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12069 and DECL_ARGUMENTS.
12070 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12071 old_tree in the map.
12072 (create_specialized_node): Update.
12073 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12075 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12076 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12077 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12078 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12081 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12083 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12084 addsi3_carryin_alt2_<optab>): Correct output template.
12086 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12088 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12089 Adjust for arm_restrict_it.
12090 Remove trailing whitespace.
12092 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12094 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12095 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12097 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12099 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12101 PR rtl-optimization/57960
12102 * lra-constraints.c (process_alt_operands): Use the right mode
12103 when checking strict_low.
12105 2013-07-25 Jan Hubicka <jh@suse.cz>
12107 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12108 * cgraph.c (cgraph_remove_node): Do not release function body
12109 when in cgraph streaming.
12110 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12111 in other partitions are not considered reachable; fix handling of
12114 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12116 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12118 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12123 * config/arm/arm.md ("*sibcall_insn): Replace use of
12124 Ss with US. Adjust output for v5 and v4t.
12125 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12127 * config/arm/constraints.md ("Ss"): Rename to US.
12129 2013-07-25 Terry Guo <terry.guo@arm.com>
12131 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12132 shift_add/shift_sub0/shift_sub1 RTXs.
12134 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12135 Anton Blanchard <anton@au1.ibm.com>
12137 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12138 (altivec_vpks<VI_char>ss): Likewise.
12139 (altivec_vpks<VI_char>us): Likewise.
12140 (altivec_vpku<VI_char>us): Likewise.
12141 (altivec_vpku<VI_char>um): Likewise.
12143 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12145 Introduce context class.
12147 * Makefile.in (CONTEXT_H): New.
12148 (OBJS): Add context.o.
12149 (toplev.o): Add CONTEXT_H to dependencies.
12152 * toplev.c (general_init): Create the singleton gcc::context instance.
12158 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12160 PR rtl-optimization/57968
12161 * mode-switching.c (create_pre_exit): Allow instructions that
12162 don't set a return register to need a non-exit mode.
12164 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12165 Anton Blanchard <anton@au1.ibm.com>
12167 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12168 operands to vperm for little endian.
12169 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12170 of lvsl to create the control mask for a vperm for little endian.
12172 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12173 Anton Blanchard <anton@au1.ibm.com>
12175 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12176 two operands for little-endian.
12178 2013-07-23 Steve Ellcey <sellcey@mips.com>
12180 * config/mips/mips.c (mips_case_values_threshold): New.
12181 (TARGET_CASE_VALUES_THRESHOLD): Define.
12183 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12184 Anton Blanchard <anton@au1.ibm.com>
12186 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12187 selection of field for vector splat in little endian mode.
12189 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12191 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12192 expanders to rs6000.md.
12193 (ior<mode>3): Likewise.
12194 (and<mode>3): Likewise.
12195 (one_cmpl<mode>2): Likewise.
12196 (nor<mode>3): Likewise.
12197 (andc<mode>3): Likewise.
12198 (eqv<mode>3): Likewise.
12199 (nand<mode>3): Likewise.
12200 (orc<mode>3): Likewise.
12202 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12205 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12206 to split multi-word logical operations.
12207 (rs6000_split_logical_di): Likewise.
12208 (rs6000_split_logical): Likewise.
12210 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12211 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12212 and allow TImode operations in 32-bit.
12213 (vsx_and<mode>3_64bit): Likewise.
12214 (vsx_ior<mode>3_32bit): Likewise.
12215 (vsx_ior<mode>3_64bit): Likewise.
12216 (vsx_xor<mode>3_32bit): Likewise.
12217 (vsx_xor<mode>3_64bit): Likewise.
12218 (vsx_one_cmpl<mode>2_32bit): Likewise.
12219 (vsx_one_cmpl<mode>2_64bit): Likewise.
12220 (vsx_nor<mode>3_32bit): Likewise.
12221 (vsx_nor<mode>3_64bit): Likewise.
12222 (vsx_andc<mode>3_32bit): Likewise.
12223 (vsx_andc<mode>3_64bit): Likewise.
12224 (vsx_eqv<mode>3_32bit): Likewise.
12225 (vsx_eqv<mode>3_64bit): Likewise.
12226 (vsx_nand<mode>3_32bit): Likewise.
12227 (vsx_nand<mode>3_64bit): Likewise.
12228 (vsx_orc<mode>3_32bit): Likewise.
12229 (vsx_orc<mode>3_64bit): Likewise.
12231 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12232 logical types in GPRs.
12234 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12235 logical insns to rs6000.md, and allow TImode operations in
12237 (altivec_ior<mode>3): Likewise.
12238 (altivec_xor<mode>3): Likewise.
12239 (altivec_one_cmpl<mode>2): Likewise.
12240 (altivec_nor<mode>3): Likewise.
12241 (altivec_andc<mode>3): Likewise.
12243 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12244 attributes for moving the 128-bit logical operations into
12246 (BOOL_REGS_OUTPUT): Likewise.
12247 (BOOL_REGS_OP1): Likewise.
12248 (BOOL_REGS_OP2): Likewise.
12249 (BOOL_REGS_UNARY): Likewise.
12250 (BOOL_REGS_AND_CR0): Likewise.
12251 (one_cmpl<mode>2): Add support for DI logical operations on
12252 32-bit, splitting the operations to 32-bit.
12253 (anddi3): Likewise.
12254 (iordi3): Likewise.
12255 (xordi3): Likewise.
12256 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12257 changes to combine the 32/64-bit code, allow logical operations on
12258 TI mode in 32-bit, and to use similar match_operator patterns like
12259 scalar mode uses. Combine the Altivec and VSX code for logical
12260 operations, and move it here.
12261 (ior<mode>3, 128-bit types): Likewise.
12262 (xor<mode>3, 128-bit types): Likewise.
12263 (one_cmpl<mode>3, 128-bit types): Likewise.
12264 (nor<mode>3, 128-bit types): Likewise.
12265 (andc<mode>3, 128-bit types): Likewise.
12266 (eqv<mode>3, 128-bit types): Likewise.
12267 (nand<mode>3, 128-bit types): Likewise.
12268 (orc<mode>3, 128-bit types): Likewise.
12269 (and<mode>3_internal): Likewise.
12270 (bool<mode>3_internal): Likewise.
12271 (boolc<mode>3_internal1): Likewise.
12272 (boolc<mode>3_internal2): Likewise.
12273 (boolcc<mode>3_internal1): Likewise.
12274 (boolcc<mode>3_internal2): Likewise.
12275 (eqv<mode>3_internal1): Likewise.
12276 (eqv<mode>3_internal2): Likewise.
12277 (one_cmpl1<mode>3_internal): Likewise.
12279 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12281 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12282 Rename flag_stack_usage to flag_stack_usage_info.
12284 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12286 * config/microblaze/sync.md: New file.
12287 * config/microblaze/microblaze.md: Include sync.md
12288 * config/microblaze/microblaze.c: Add print_operand 'y'.
12289 * config/microblaze/constraints.md: Add memory_contraint
12290 'Q' which is a single register.
12292 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12294 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12296 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12298 * reload.c (find_reloads): Exit loop once we find this operand
12299 cannot be reloaded somehow for this alternative.
12301 * reload.c (find_reloads): Exit loop once we find a hard register.
12303 * rtlanal.c (computed_jump_p): Exit loop once we find label
12306 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12308 * cfgloopmanip.c (remove_path): Exit loop after setting
12311 * gensupport.c (subst_dup): Avoid loop if code is not
12312 MATCH_DUP nor MATCH_OP_DUP.
12314 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12316 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12318 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12320 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12321 true for SP_REGNUM if mode == ptr_mode.
12322 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12323 with value R0_REGNUM + 31.
12325 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12327 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12328 pad pointer-typed argument downward.
12330 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12332 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12333 and __ILP32__ when the ILP32 model is in use.
12335 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12337 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12338 (aarch64_load_symref_appropriately): In the case of
12339 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12340 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12341 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12342 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12343 if 'mode' doesn't equal to 'ptr_mode'.
12344 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12345 'vcall_offset'; change to call aarch64_emit_move differently depending
12346 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12347 to calculate the upper bound of 'vcall_offset'.
12348 (aarch64_cannot_force_const_mem): Change to also return true if
12350 (aarch64_legitimize_reload_address): In the case of large
12351 displacements, add new local variable 'xmode' and an assertion
12352 based on it; change to use 'xmode' to generate the new rtx and
12354 (aarch64_asm_trampoline_template): Change to generate the template
12355 differently depending on TARGET_ILP32 or not; change to use
12356 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12357 (aarch64_trampoline_size): Removed.
12358 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12359 and replace immediate literals with it. Change to use 'ptr_mode'
12360 instead of 'DImode' and call convert_memory_address if the mode
12361 of 'fnaddr' doesn't equal to 'ptr_mode'.
12362 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12364 (aarch64_elf_asm_destructor): Likewise.
12365 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12366 on TARGET_ILP32 instead of aarch64_trampoline_size.
12367 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12368 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12369 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12370 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12371 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12372 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12373 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12374 depending on the value of 'mode'.
12375 (add_losym_<mode>): New.
12376 (ldr_got_small_<mode>): New, based on ldr_got_small.
12377 (ldr_got_small): Remove.
12378 (ldr_got_small_sidi): New.
12379 * config/aarch64/iterators.md (P): New.
12380 (PTR): Change to 'ptr_mode' in the condition.
12382 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12384 * config.gcc (aarch64*-*-*): Support --with-abi.
12385 (aarch64*-*-elf): Support --with-multilib-list.
12386 (aarch64*-*-linux*): Likewise.
12387 (supported_defaults): Add abi to aarch64*-*-*.
12388 * configure.ac: Mention AArch64 for --with-multilib-list.
12389 * configure: Re-generated.
12390 * config/aarch64/biarchilp32.h: New file.
12391 * config/aarch64/biarchlp64.h: New file.
12392 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12394 (MULTILIB_DEFAULTS): Ditto.
12395 (DRIVER_SELF_SPECS): Ditto.
12396 (ASM_SPEC): Update to also substitute -mabi.
12397 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12398 file whose name depends on -mabi= and -mbig-endian.
12399 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12401 (POINTER_SIZE): New define.
12402 (POINTERS_EXTEND_UNSIGNED): Ditto.
12403 (enum aarch64_abi_type): New enumeration tag.
12404 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12405 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12406 (TARGET_ILP32): New define.
12407 * config/aarch64/aarch64.opt (mabi): New.
12408 (aarch64_abi): New.
12409 (ilp32, lp64): New values for -mabi.
12410 * config/aarch64/t-aarch64 (comma): New define.
12411 (MULTILIB_OPTIONS): Ditto.
12412 (MULTILIB_DIRNAMES): Ditto.
12413 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12414 * doc/invoke.texi: Document -mabi for AArch64.
12416 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12418 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12420 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12421 Anton Blanchard <anton@au1.ibm.com>
12423 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12424 endianness when selecting field to splat.
12426 2013-07-22 Eric Christopher <echristo@gmail.com>
12428 * dwarf2out.c (die_odr_checksum): New function to use
12429 CHECKSUM_ macros and ULEB128 for DIE tag.
12430 (generate_type_signature): Use.
12432 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12434 * config.gcc (sparc*-*-*): Accept leon3 processor.
12435 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12436 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12437 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12438 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12439 (mfix-ut699): Adjust comment.
12440 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12441 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12442 (CPP_CPU_SPEC): Likewise.
12443 (ASM_CPU_SPEC): Likewise.
12444 * config/sparc/sparc.c (leon3_cost): New constant.
12445 (sparc_option_override): Add leon3 support.
12446 (mem_ref): New function.
12447 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12448 (sparc_do_work_around_errata): Look into the instruction in the delay
12449 slot and adjust accordingly. Add fix for the data cache nullify issues
12450 of the UT699. Change insertion position for the NOP.
12451 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12452 (leon3_load): New reservation.
12453 (leon_store): Bump latency to 2.
12454 (grfpu): New automaton.
12455 (grfpu_alu): New unit.
12456 (grfpu_ds): Likewise.
12457 (leon_fp_alu): Adjust.
12458 (leon_fp_mult): Delete.
12459 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12460 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12461 * config/sparc/sparc.md (cpu): Add leon3.
12462 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12463 (swapsi): Likewise.
12464 (atomic_test_and_set): Likewise.
12465 (ldstub): Likewise.
12467 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12469 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12470 default for R5900 targets.
12471 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12472 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12473 * config/mips/mips.c (mips_option_override): Report an error for
12474 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12475 for -march=r5900 -mhard-float.
12477 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12479 * df-problems.c (can_move_insns_across): Exit loop once we
12480 find a non-fixed, non-global register.
12482 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12485 * omega.c (omega_eliminate_red): Break after setting red_found.
12486 (omega_problem_has_red_equations): Similarly after setting found.
12487 (omega_query_variable): Similarly after setting coupled.
12489 2013-07-22 Marek Polacek <polacek@redhat.com>
12491 * gimplify.c: Don't include gimple.h twice.
12493 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12495 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12496 instead of TARGET_THUMB1.
12497 (Pz): New constraint.
12498 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12500 (compare_negsi_si): Likewise.
12501 (compare_addsi2_op0): Likewise.
12502 (compare_addsi2_op1): Likewise.
12503 (addsi3_carryin_<optab>): Likewise.
12504 (addsi3_carryin_alt2_<optab>): Likewise.
12505 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12506 for arm_restrict_it.
12507 (subsi3_carryin): Likewise.
12508 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12509 (minmax_arithsi): Disable for arm_restrict_it.
12510 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12511 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12512 (satsi_<SAT:code>_shift): Likewise.
12513 (arm_shiftsi3): Add alternative for 16-bit encoding.
12514 (arm32_movhf): Disable for arm_restrict_it.
12515 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12516 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12518 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12520 * config/arm/arm.md (attribute "insn"): Delete.
12521 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12522 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12523 (not_shiftsi): Update for attribute change.
12524 (not_shiftsi_compare0): Likewise.
12525 (not_shiftsi_compare0_scratch): Likewise.
12526 (arm_one_cmplsi2): Likewise.
12527 (thumb1_one_cmplsi2): Likewise.
12528 (notsi_compare0): Likewise.
12529 (notsi_compare0_scratch): Likewise.
12530 (thumb1_movdi_insn): Likewise.
12531 (arm_movsi_insn): Likewise.
12532 (movhi_insn_arch4): Likewise.
12533 (movhi_bytes): Likewise.
12534 (arm_movqi_insn): Likewise.
12535 (thumb1_movqi_insn): Likewise.
12536 (arm32_movhf): Likewise.
12537 (thumb1_movhf): Likewise.
12538 (arm_movsf_soft_insn): Likewise.
12539 (thumb1_movsf_insn): Likewise.
12540 (thumb_movdf_insn): Likewise.
12541 (movsicc_insn): Likewise.
12542 (movsfcc_soft_insn): Likewise.
12543 (and_scc): Likewise.
12544 (cond_move): Likewise.
12545 (if_move_not): Likewise.
12546 (if_not_move): Likewise.
12547 (if_shift_move): Likewise.
12548 (if_move_shift): Likewise.
12549 (if_shift_shift): Likewise.
12550 (if_not_arith): Likewise.
12551 (if_arith_not): Likewise.
12552 (cond_move_not): Likewise.
12553 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12554 (neon_mov<mode>): Likewise.
12555 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12556 (thumb2_movsi_vfp): Likewise.
12557 (movsf_vfp): Likewise.
12558 (thumb2_movsf_vfp): Likewise.
12559 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12561 (cortexa7_older_only): Likewise.
12562 (cortexa7_younger): Likewise.
12563 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12564 (1020alu_shift_op): Likewise.
12565 (1020alu_shift_reg_op): Likewise.
12566 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12567 (alu_shift_op): Likewise.
12568 (alu_shift_reg_op): Likewise.
12569 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12570 (11_alu_shift_op): Likewise.
12571 (11_alu_shift_reg_op): Likewise.
12572 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12573 (9_alu_shift_reg_op): Likewise.
12574 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12576 (cortex_a15_alu_shift): Likewise.
12577 (cortex_a15_alu_shift_reg): Likewise.
12578 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12579 (cortex_a5_alu_shift): Likewise.
12580 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12582 (cortex_a53_alu_shift): Likewise.
12583 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12585 (cortex_a7_alu_reg): Likewise.
12586 (cortex_a7_alu_shift): Likewise.
12587 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12588 (cortex_a8_alu_shift): Likewise.
12589 (cortex_a8_alu_shift_reg): Likewise.
12590 (cortex_a8_mov): Likewise.
12591 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12592 (cortex_a9_dp_shift): Likewise.
12593 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12594 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12595 (cortex_r4_mov): Likewise.
12596 (cortex_r4_alu_shift): Likewise.
12597 (cortex_r4_alu_shift_reg): Likewise.
12598 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12599 (526_alu_shift_op): Likewise.
12600 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12601 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12602 (626te_alu_shift_op): Likewise.
12603 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12604 (726te_alu_op): Likewise.
12605 (726te_alu_shift_op): Likewise.
12606 (726te_alu_shift_reg_op): Likewise.
12607 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12608 (mp626_alu_shift_op): Likewise.
12609 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12610 (pj4_alu_e1_conds): Likewise.
12611 (pj4_alu): Likewise.
12612 (pj4_alu_conds): Likewise.
12613 (pj4_shift): Likewise.
12614 (pj4_shift_conds): Likewise.
12615 (pj4_alu_shift): Likewise.
12616 (pj4_alu_shift_conds): Likewise.
12618 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12620 * config/arm/predicates.md (shiftable_operator_strict_it):
12622 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12623 Disable cond_exec version for arm_restrict_it.
12624 (thumb2_smaxsi3): Convert to generate cond_exec.
12625 (thumb2_sminsi3): Likewise.
12626 (thumb32_umaxsi3): Likewise.
12627 (thumb2_uminsi3): Likewise.
12628 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12629 (thumb2_neg_abssi2): Likewise.
12630 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12631 (thumb2_movsicc_insn): Adjust alternatives.
12632 (thumb2_mov_negscc): Disable for arm_restrict_it.
12633 (thumb2_mov_negscc_strict_it): New pattern.
12634 (thumb2_mov_notscc_strict_it): New pattern.
12635 (thumb2_mov_notscc): Disable for arm_restrict_it.
12636 (thumb2_ior_scc): Likewise.
12637 (thumb2_ior_scc_strict_it): New pattern.
12638 (thumb2_cond_move): Adjust for arm_restrict_it.
12639 (thumb2_cond_arith): Disable for arm_restrict_it.
12640 (thumb2_cond_arith_strict_it): New pattern.
12641 (thumb2_cond_sub): Adjust for arm_restrict_it.
12642 (thumb2_movcond): Likewise.
12643 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12644 (thumb2_zero_extendhisi2_v6): Likewise.
12645 (thumb2_zero_extendqisi2_v6): Likewise.
12646 (orsi_notsi_si): Likewise.
12647 (orsi_not_shiftsi_si): Likewise.
12649 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12651 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12652 instruction sequence is 1 byte shorter.
12654 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12656 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12657 it is not needed after split.
12659 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12662 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12663 second label for nonlocal goto receivers. Don't output pic base labels
12664 unless we're producing PIC; mark that action unreachable().
12665 (ix86_save_reg): If the function contains a nonlocal label, save the
12667 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12668 * config/darwin.c (emitted_pic_label_num): New GTY.
12669 (update_pic_label_number_if_needed): New.
12670 (machopic_output_function_base_name): Adjust for nonlocal receiver
12672 (machopic_should_output_picbase_label): New.
12673 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12674 (nonlocal_goto_receiver): New insn and split.
12676 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12678 * config/aarch64/aarch64-builtins.c
12679 (aarch64_fold_builtin): Fold abs in all modes.
12680 * config/aarch64/aarch64-simd-builtins.def
12681 (abs): Enable for all modes.
12682 * config/aarch64/arm_neon.h
12683 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12684 (vabs_f64): Add missing intrinsic.
12686 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12688 * config/aarch64/arm_neon.h (vabs_s64): New function
12690 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12693 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12694 * config/avr/avr.md (adjust_len): Add `round'.
12695 * config/avr/avr-protos.h (avr_out_round): New prototype.
12696 (avr_out_plus): Add `out_label' argument.
12697 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12698 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12699 Handle the case where `insn' is just a pattern.
12700 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12701 (avr_out_round): New function.
12702 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12704 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12706 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12707 Add check for flag_stack_usage to handle -fstack-usage support
12709 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12711 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12712 interaction for new Power8 flags and VSX.
12714 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12716 PR middle-end/57698
12717 * tree-inline.c (expand_call_inline): Emit errors during
12718 early_inlining only if optimization is not turned on.
12720 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12724 * passes.c (init_optimization_passes): Move the construction of
12725 the pass hierarchy into a new passes.def file.
12727 * Makefile.in (passes.o): Add dependency on passes.def.
12729 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12731 * passes.c (init_optimization_passes): Introduce macros for
12732 constructing the tree of passes (INSERT_PASSES_AFTER,
12733 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12734 TERMINATE_PASS_LIST).
12736 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12737 Wei Mi <wmi@google.com>
12739 PR rtl-optimization/57878
12740 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12742 (reload_pseudo_compare_func): Check nregs first for reload
12745 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12747 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12749 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12751 * read-rtl.c (validate_const_int): Once an invalid character is
12752 seen, quit the loop.
12754 * gengtype.c (write_roots): Similarly once we find the "deletable"
12755 or "if_marked" option.
12757 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12759 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12760 "xtab" and "sat". Move value "clz" from here to ...
12761 (attriubte "type"): ... here.
12762 (satsi_<SAT:code>): Delete "insn" attribute.
12763 (satsi_<SAT:code>_shift): Likewise.
12764 (arm_zero_extendqisi2addsi): Likewise.
12765 (arm_extendqisi2addsi): Likewise.
12766 (clzsi2): Update for attribute changes.
12767 (rbitsi2): Likewise.
12768 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12770 (arm_usatsihi): Likewise.
12771 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12773 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12775 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12776 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12777 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12778 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12779 in alphabetical order.
12780 (attribute "core_cycles"): Update for attribute changes.
12781 (arm_addsi3): Likewise.
12782 (addsi3_compare0): Likewise.
12783 (addsi3_compare0_scratch): Likewise.
12784 (addsi3_compare_op1): Likewise.
12785 (addsi3_compare_op2): Likewise.
12786 (compare_addsi2_op0): Likewise.
12787 (compare_addsi2_op1): Likewise.
12788 (addsi3_carryin_shift_<optab>): Likewise.
12789 (subsi3_carryin_shift): Likewise.
12790 (rsbsi3_carryin_shift): Likewise.
12791 (arm_subsi3_insn): Likewise.
12792 (subsi3_compare0): Likewise.
12793 (subsi3_compare): Likewise.
12794 (arm_andsi3_insn): Likewise.
12795 (thumb1_andsi3_insn): Likewise.
12796 (andsi3_compare0): Likewise.
12797 (andsi3_compare0_scratch): Likewise.
12798 (zeroextractsi_compare0_scratch
12799 (andsi_not_shiftsi_si): Likewise.
12800 (iorsi3_insn): Likewise.
12801 (iorsi3_compare0): Likewise.
12802 (iorsi3_compare0_scratch): Likewise.
12803 (arm_xorsi3): Likewise.
12804 (thumb1_xorsi3_insn): Likewise.
12805 (xorsi3_compare0): Likewise.
12806 (xorsi3_compare0_scratch): Likewise.
12807 (satsi_<SAT:code>_shift): Likewise.
12809 (arm_shiftsi3): Likewise.
12810 (shiftsi3_compare0): Likewise.
12811 (not_shiftsi): Likewise.
12812 (not_shiftsi_compare0): Likewise.
12813 (not_shiftsi_compare0_scratch): Likewise.
12814 (arm_one_cmplsi2): Likewise.
12815 (thumb_one_complsi2): Likewise.
12816 (notsi_compare0): Likewise.
12817 (notsi_compare0_scratch): Likewise.
12818 (thumb1_zero_extendhisi2): Likewise.
12819 (arm_zero_extendhisi2): Likewise.
12820 (arm_zero_extendhisi2_v6): Likewise.
12821 (arm_zero_extendhisi2addsi): Likewise.
12822 (thumb1_zero_extendqisi2): Likewise.
12823 (thumb1_zero_extendqisi2_v6): Likewise.
12824 (arm_zero_extendqisi2): Likewise.
12825 (arm_zero_extendqisi2_v6): Likewise.
12826 (arm_zero_extendqisi2addsi): Likewise.
12827 (thumb1_extendhisi2): Likewise.
12828 (arm_extendhisi2): Likewise.
12829 (arm_extendhisi2_v6): Likewise.
12830 (arm_extendqisi): Likewise.
12831 (arm_extendqisi_v6): Likewise.
12832 (arm_extendqisi2addsi): Likewise.
12833 (thumb1_extendqisi2): Likewise.
12834 (thumb1_movdi_insn): Likewise.
12835 (arm_movsi_insn): Likewise.
12836 (movsi_compare0): Likewise.
12837 (movhi_insn_arch4): Likewise.
12838 (movhi_bytes): Likewise.
12839 (arm_movqi_insn): Likewise.
12840 (thumb1_movqi_insn): Likewise.
12841 (arm32_movhf): Likewise.
12842 (thumb1_movhf): Likewise.
12843 (arm_movsf_soft_insn): Likewise.
12844 (thumb1_movsf_insn): Likewise.
12845 (movdf_soft_insn): Likewise.
12846 (thumb_movdf_insn): Likewise.
12847 (arm_cmpsi_insn): Likewise.
12848 (cmpsi_shiftsi): Likewise.
12849 (cmpsi_shiftsi_swp): Likewise.
12850 (arm_cmpsi_negshiftsi_si): Likewise.
12851 (movsicc_insn): Likewise.
12852 (movsfcc_soft_insn): Likewise.
12853 (arith_shiftsi): Likewise.
12854 (arith_shiftsi_compare0
12855 (arith_shiftsi_compare0_scratch
12856 (sub_shiftsi): Likewise.
12857 (sub_shiftsi_compare0
12858 (sub_shiftsi_compare0_scratch
12859 (and_scc): Likewise.
12860 (cond_move): Likewise.
12861 (if_plus_move): Likewise.
12862 (if_move_plus): Likewise.
12863 (if_move_not): Likewise.
12864 (if_not_move): Likewise.
12865 (if_shift_move): Likewise.
12866 (if_move_shift): Likewise.
12867 (if_shift_shift): Likewise.
12868 (if_not_arith): Likewise.
12869 (if_arith_not): Likewise.
12870 (cond_move_not): Likewise.
12871 (thumb1_ashlsi3): Set type attribute.
12872 (thumb1_ashrsi3): Likewise.
12873 (thumb1_lshrsi3): Likewise.
12874 (thumb1_rotrsi3): Likewise.
12875 (shiftsi3_compare0_scratch): Likewise.
12876 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12877 (neon_mov<mode>): Likewise.
12878 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12880 (thumb2_movsi_insn): Likewise.
12881 (thumb2_cmpsi_neg_shiftsi): Likewise.
12882 (thumb2_extendqisi_v6): Likewise.
12883 (thumb2_zero_extendhisi2_v6): Likewise.
12884 (thumb2_zero_extendqisi2_v6): Likewise.
12885 (thumb2_shiftsi3_short): Likewise.
12886 (thumb2_addsi3_compare0_scratch): Likewise.
12887 (orsi_not_shiftsi_si): Likewise.
12888 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12889 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12891 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12892 (1020alu_shift_op): Likewise.
12893 (1020alu_shift_reg_op): Likewise.
12894 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
12895 (alu_shift_op): Likewise.
12896 (alu_shift_reg_op): Likewise.
12897 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
12898 (11_alu_shift_op): Likewise.
12899 (11_alu_shift_reg_op): Likewise.
12900 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
12901 (9_alu_shift_reg_op): Likewise.
12902 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12904 (cortex_a15_alu_shift): Likewise.
12905 (cortex_a15_alu_shift_reg): Likewise.
12906 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
12908 (cortex_a5_alu_shift): Likewise.
12909 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
12911 (cortex_a53_alu_shift): Likewise.
12912 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12914 (cortex_a7_alu_reg): Likewise.
12915 (cortex_a7_alu_shift): Likewise.
12916 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
12918 (cortex_a8_alu_shift): Likewise.
12919 (cortex_a8_alu_shift_reg): Likewise.
12920 (cortex_a8_mov): Likewise.
12921 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
12922 (cortex_a9_dp_shift): Likewise.
12923 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
12925 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
12927 (cortex_r4_mov): Likewise.
12928 (cortex_r4_alu_shift): Likewise.
12929 (cortex_r4_alu_shift_reg): Likewise.
12930 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
12931 (526_alu_shift_op): Likewise.
12932 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
12933 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
12934 (626te_alu_shift_op): Likewise.
12935 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
12936 (726te_alu_op): Likewise.
12937 (726te_alu_shift_op): Likewise.
12938 (726te_alu_shift_reg_op): Likewise.
12939 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
12940 (mp626_alu_shift_op): Likewise.
12941 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
12942 (pj4_alu_e1_conds): Likewise.
12943 (pj4_alu): Likewise.
12944 (pj4_alu_conds): Likewise.
12945 (pj4_shift): Likewise.
12946 (pj4_shift_conds): Likewise.
12947 (pj4_alu_shift): Likewise.
12948 (pj4_alu_shift_conds): Likewise.
12949 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12951 (cortexa7_older_only): Likewise.
12952 (cortexa7_younger): Likewise.
12954 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12956 * ipa-pure-const.c (generate_summary): Rename to...
12957 (pure_const_generate_summary): ... this.
12959 2013-07-17 Iain Sandoe <iain@codesourcery.com>
12961 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
12963 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
12966 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
12969 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12971 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
12972 enabled without -march=zEC12.
12973 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12976 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12978 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12979 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12980 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12981 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12982 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12983 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12984 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12986 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12987 and ISA_HAS_NMADD3_NMSUB3.
12988 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12989 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12990 (nmsub4<mode>, nmsub3<mode>): Likewise.
12991 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12993 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12995 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12996 TARGET_MIPS5400 checking.
12998 2013-07-16 Jakub Jelinek <jakub@redhat.com>
12999 Peter Bergner <bergner@vnet.ibm.com>
13001 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13002 registers in the comment.
13003 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13004 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13005 rather than FIRST_PSEUDO_REGISTERS.
13007 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13009 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13010 enable extra ISA flags with TARGET_HTM.
13012 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13014 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13017 2013-07-15 Cong Hou <congh@google.com>
13019 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13020 in compare function for sorting.
13022 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13024 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13025 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13026 * config/rs6000/rs6000.opt: Add -mhtm option.
13027 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13028 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13029 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13030 __HTM__ if the HTM instructions are available.
13031 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13032 htm_spr_reg_operand): New define_predicates.
13033 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13034 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13036 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13037 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13038 HTM builtin functions.
13039 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13040 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13041 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13042 (rs6000_builtin_mask_calculate): Likewise.
13043 (rs6000_option_override_internal): Likewise.
13044 (bdesc_htm): Add new HTM builtin support.
13045 (htm_spr_num): New function.
13046 (htm_spr_regno): Likewise.
13047 (rs6000_htm_spr_icode): Likewise.
13048 (htm_expand_builtin): Likewise.
13049 (htm_init_builtins): Likewise.
13050 (rs6000_expand_builtin): Add support for HTM builtin functions.
13051 (rs6000_init_builtins): Likewise.
13052 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13054 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13055 (TARGET_HTM, MASK_HTM): Define macros.
13056 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13057 (FIXED_REGISTERS): Likewise.
13058 (CALL_USED_REGISTERS): Likewise.
13059 (CALL_REALLY_USED_REGISTERS): Likewise.
13060 (REG_ALLOC_ORDER): Likewise.
13061 (enum reg_class): Likewise.
13062 (REG_CLASS_NAMES): Likewise.
13063 (REG_CLASS_CONTENTS): Likewise.
13064 (REGISTER_NAMES): Likewise.
13065 (ADDITIONAL_REGISTER_NAMES): Likewise.
13066 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13067 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13068 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13069 * config/rs6000/htm.md: New file.
13070 * config/rs6000/htmintrin.h: New file.
13071 * config/rs6000/htmxlintrin.h: New file.
13073 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13075 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13076 Define SYMBOL_TINY_GOT, update comment.
13077 * config/aarch64/aarch64.c
13078 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13079 (aarch64_expand_mov_immediate): Likewise.
13080 (aarch64_print_operand): Likewise.
13081 (aarch64_classify_symbol): Likewise.
13082 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13083 (ldr_got_tiny): Define.
13085 2013-07-13 Tobias Grosser <tobias@grosser.es>
13087 PR tree-optimization/54094
13088 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13089 scheduling dimension for the parallelism check from the polyhedral
13090 information in the AST.
13091 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13094 2013-07-13 Jason Merrill <jason@redhat.com>
13096 * print-tree.c (debug_vec_tree): Use debug_raw.
13097 (debug_raw (vec<tree, va_gc> &)): New.
13098 (debug_raw (vec<tree, va_gc> *)): New.
13099 * tree.h: Declare them.
13101 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13103 * ifcvt.c (ifcvt_after_combine): New static variable.
13104 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13106 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13107 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13108 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13110 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13112 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13114 2013-07-12 Michael Matz <matz@suse.de>
13116 PR middle-end/55771
13117 * convert.c (convert_to_real): Reject non-float inner types.
13119 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13121 * config/aarch64/aarch64-protos.h
13122 (aarch64_simd_immediate_valid_for_move): Remove.
13123 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13124 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13126 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13128 2013-07-11 Steve Ellcey <sellcey@mips.com>
13130 * config/mips/mips.c (mips_conditional_register_usage): Do not
13131 use t[0-7] registers in MIPS16 mode when optimizing for size.
13133 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13135 * config/i386/i386.c (dispatch_function_versions): Fix array
13136 indexing of function_version_info to match actual_versions.
13138 2013-07-11 Teresa Johnson <tejohnson@google.com>
13140 * vec.h (struct va_gc): Move release out-of-line.
13141 (va_gc::release): Call ggc_free on released vec.
13143 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13145 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13146 Require GOT register as additional operand in UNSPEC.
13147 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13148 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13149 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13150 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13151 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13152 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13153 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13155 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13158 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13159 name seen by assembler/linker rather if available.
13161 2013-07-11 Andreas Schwab <schwab@suse.de>
13163 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13165 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13167 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13169 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13171 * doc/tm.texi.in: Move hook documentation to ....
13172 * target.def: ... here.
13174 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13175 text on @hook line.
13176 * doc/tm.texi: Regenerate.
13178 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13181 * doc/invoke.texi: Document -Wconditionally-supported.
13183 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13186 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13189 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13192 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13193 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13194 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13195 Remove duplicate devices.
13196 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13197 * config/avr/t-multilib: Regenerate.
13198 * config/avr/avr-tables.opt: Regenerate.
13199 * doc/avr-mmcu.texi: Regenerate.
13201 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13204 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13206 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13208 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13209 the cost of MULT when optimizing for size.
13211 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13213 * config/cr16/cr16-protos.h: Don't include target.h.
13215 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13217 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13218 adjust register size for TDmode and TFmode for VSX registers.
13220 2013-07-08 Kai Tietz <ktietz@redhat.com>
13223 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13224 hook_bool_const_tree_true.
13226 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13228 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13229 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13230 * config/s390/s390.md: Define FPR*_REGNUM constants.
13231 Fix FPR2_REGNUM constant (18 -> 17).
13232 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13233 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13234 ("trunc<BFP:mode><DFP_ALL:mode>2")
13235 ("trunc<DFP_ALL:mode><BFP:mode>2")
13236 ("*extend<BFP:mode><DFP_ALL:mode>2")
13237 ("*extend<DFP_ALL:mode><BFP:mode>2")
13238 ("extend<BFP:mode><DFP_ALL:mode>2")
13239 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13242 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13244 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13246 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13248 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13249 and cfun_fpr_bit_p to cfun_fpr_save_p.
13250 (s390_frame_area, s390_register_info, s390_frame_info)
13251 (s390_emit_prologue, s390_emit_epilogue)
13252 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13254 * config/s390/s390.h: Define *_REGNUM macros for floating point
13257 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13259 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13261 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13263 PR rtl-optimization/57786
13264 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13265 and break out of the loop when it is set to false.
13267 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13270 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13271 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13272 (const_int 63)) 0)).
13273 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13274 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13275 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13277 PR rtl-optimization/57829
13278 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13279 mask bits outside of mode are just sign-extension from mode to HWI.
13281 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13283 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13284 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13285 adjust_address instead of change_address to keep info about alignment.
13286 (emit_strmov): Remove.
13287 (emit_memmov): New function.
13288 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13289 (expand_movmem_epilogue): Likewise and return updated rtx for
13291 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13292 destination and source.
13293 (decide_alignment): Refactor, handle vector_loop.
13294 (ix86_expand_movmem): Likewise.
13295 (ix86_expand_setmem): Likewise.
13296 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13298 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13300 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13301 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13303 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13305 * config/i386/sse.md (sse_movlhps): Change alternative 3
13306 of operand 2 to "m".
13308 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13311 * config/i386/sse.md (iptr): New mode attribute.
13312 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13313 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13314 (<sse>_vmmul<mode>3): Ditto.
13315 (<sse>_vmdiv<mode>3): Ditto.
13316 (sse_vmrcpv4sf2): Ditto.
13317 (<sse>_vmsqrt<mode>2): Ditto.
13318 (sse_vmrsqrtv4sf2): Ditto.
13319 (<sse>_vm<code><mode>3): Ditto.
13320 (avx_vmcmp<mode>3): Ditto.
13321 (<sse>_vmmaskcmp<mode>3): Ditto.
13322 (<sse>_comi): Ditto.
13323 (<sse>_ucomi): Ditto.
13324 (*xop_vmfrcz_<mode>): Ditto.
13325 (*fmai_fmadd_<mode>): Ditto.
13326 (*fmai_fmsub_<mode>): Ditto.
13327 (*fmai_fnmadd_<mode>): Ditto.
13328 (*fmai_fnmsub_<mode>): Ditto.
13329 (*fma4i_vmfmadd_<mode>): Ditto.
13330 (*fma4i_vmfmsub_<mode>): Ditto.
13331 (*fma4i_vmfnmadd_<mode>): Ditto.
13332 (*fma4i_vmfnmsub_<mode>): Ditto.
13333 (*xop_vmfrcz_<mode>): Ditto.
13334 (sse_cvtps2pi): Ditto.
13335 (sse_cvttps2pi): Ditto.
13336 (sse_cvtss2si): Ditto.
13337 (sse_cvtss2si_2): Ditto.
13338 (sse_cvtss2siq_2): Ditto.
13339 (sse_cvttss2si): Ditto.
13340 (sse_cvttss2siq): Ditto.
13341 (sse_cvtsd2si): Ditto.
13342 (sse_cvtsd2si_2): Ditto.
13343 (sse_cvtsd2siq_2): Ditto.
13344 (sse_cvttsd2si): Ditto.
13345 (sse_cvttsd2siq): Ditto.
13346 (sse_cvtsd2ss): Ditto.
13347 (sse_cvtss2sd): Ditto.
13348 (avx2_pbroadcast<mode>): Ditto.
13349 (avx2_pbroadcast<mode>_1): Ditto.
13350 (*avx_vperm_broadcast_v4sf): Ditto.
13352 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13353 (sse_movlhps): Ditto.
13354 (sse_storehps): Ditto.
13355 (sse_loadhps): Ditto.
13356 (sse_storelps): Ditto.
13357 (sse_loadlps): Ditto.
13358 (*vec_concatv4sf): Ditto.
13359 (*vec_interleave_highv2df): Ditto.
13360 (*vec_interleave_lowv2df): Ditto.
13361 (*vec_extractv2df_1_sse): Ditto.
13362 (*vec_extractv2df_0_sse): Ditto.
13363 (sse2_storelpd): Ditto.
13364 (sse2_loadlpd): Ditto.
13365 (sse2_movsd): Ditto.
13366 (*vec_concatv4si): Ditto.
13367 (vec_concatv2di): Ditto.
13369 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13370 for Intel asm dialect.
13371 (mmx_punpcklwd): Ditto.
13372 (mmx_punpckldq): Ditto.
13374 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13375 for intel assembler dialect.
13377 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13380 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13381 for BUILT_IN_C{LZ,LRSB}*.
13382 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13383 * tree-vrp.c (extract_range_basic): Handle
13384 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13385 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13386 fall thru to code calling set_value*.
13387 * builtins.c (expand_builtin): Remove *IMAX cases.
13388 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13389 if width is bigger than 2*HWI.
13391 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13393 PR rtl-optimization/55342
13394 * lra-int.h (lra_subreg_reload_pseudos): New.
13395 * lra.c: Add undoing optional reloads to the block diagram.
13396 (lra_subreg_reload_pseudos): New.
13397 (lra_optional_reload_pseudos): Change comments.
13398 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13399 lra_optional_reload_pseudos after undo transformations.
13400 * lra-assigns.c (pseudo_prefix_title): New.
13401 (lra_setup_reg_renumber): Use it.
13402 (spill_for): Ditto. Check subreg reload pseudos too.
13403 (assign_by_spills): Consider subreg reload pseudos too.
13404 * lra-constraints.c (simplify_operand_subreg): Use
13405 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13406 (curr_insn_transform): Recognize and do optional reloads.
13407 (undo_optional_reloads): New.
13408 (lra_undo_inheritance): Call undo_optional_reloads.
13410 2013-07-05 Thomas Quinot <quinot@adacore.com>
13412 * tree-complex.c (expand_complex_operations_1): Fix typo.
13414 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13416 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13417 (tune_params): New member 'const vec_costs'.
13418 * config/aarch64/aarch64.c (generic_vector_cost): New.
13419 (generic_tunings): New member 'generic_vector_cost'.
13420 (aarch64_builtin_vectorization_cost): New.
13421 (aarch64_add_stmt_cost): New.
13422 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13423 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13425 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13428 * config/i386/predicates.md (vsib_address_operand): Disallow
13429 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13431 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13433 PR middle-end/55030
13434 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13435 expand_builtin_setjmp_receiver.
13436 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13437 with NULL for the label parameter.
13438 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13439 the frame-pointer. Adjust comments.
13440 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13441 only if LABEL is non-NULL.
13443 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13445 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13446 (ARM_ABI_AAPCS64): Ditto.
13448 (ARM_DEFAULT_ABI): Ditto.
13450 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13452 * config/aarch64/aarch64-builtins.c
13453 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13454 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13456 * config/aarch64/aarch64-simd.md
13457 (aarch64_ld1<VALL:mode>): New.
13458 (aarch64_st1<VALL:mode>): Likewise.
13459 * config/aarch64/arm_neon.h
13460 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13462 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13464 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13466 (ix86_option_override_internal):Turn on all -mavx target flags by
13467 default as they are dependent on AVX anyway.
13469 2013-07-02 Cary Coutant <ccoutant@google.com>
13471 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13472 deterministic hash.
13473 (loc_checksum_ordered): Likewise.
13474 (hash_loc_operands): Remove inline keyword.
13476 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13478 PR tree-optimization/57741
13479 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13480 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13481 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13482 Allow REAL_CST step_exprs if flag_associative_math.
13483 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13485 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13487 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13489 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13491 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13493 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13495 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13497 (iorsi3_insn): Likewise.
13498 (arm_xorsi3): Likewise.
13500 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13502 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13504 (attribute "type"): ... here, and prefix with "wmmx_".
13505 (attribute "core_cycles"): Update for attribute changes.
13506 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13507 (tbcstv4hi): Likewise.
13508 (tbcstv2si): Likewise.
13509 (iwmmxt_iordi3): Likewise.
13510 (iwmmxt_xordi3): Likewise.
13511 (iwmmxt_anddi3): Likewise.
13512 (iwmmxt_nanddi3): Likewise.
13513 (iwmmxt_arm_movdi): Likewise.
13514 (iwmmxt_movsi_insn): Likewise.
13515 (mov<mode>_internal): Likewise.
13516 (and<mode>3_iwmmxt): Likewise.
13517 (ior<mode>3_iwmmxt): Likewise.
13518 (xor<mode>3_iwmmxt): Likewise.
13519 (add<mode>3_iwmmxt): Likewise.
13520 (ssaddv8qi3): Likewise.
13521 (ssaddv4hi3): Likewise.
13522 (ssaddv2si3): Likewise.
13523 (usaddv8qi3): Likewise.
13524 (usaddv4hi3): Likewise.
13525 (usaddv2si3): Likewise.
13526 (sub<mode>3_iwmmxt): Likewise.
13527 (sssubv8qi3): Likewise.
13528 (sssubv4hi3): Likewise.
13529 (sssubv2si3): Likewise.
13530 (ussubv8qi3): Likewise.
13531 (ussubv4hi3): Likewise.
13532 (ussubv2si3): Likewise.
13533 (mulv4hi3_iwmmxt): Likewise.
13534 (smulv4hi3_highpart): Likewise.
13535 (umulv4hi3_highpart): Likewise.
13536 (iwmmxt_wmacs): Likewise.
13537 (iwmmxt_wmacsz): Likewise.
13538 (iwmmxt_wmacu): Likewise.
13539 (iwmmxt_wmacuz): Likewise.
13540 (iwmmxt_clrdi): Likewise.
13541 (iwmmxt_clrv8qi): Likewise.
13542 (iwmmxt_clr4hi): Likewise.
13543 (iwmmxt_clr2si): Likewise.
13544 (iwmmxt_uavgrndv8qi3): Likewise.
13545 (iwmmxt_uavgrndv4hi3): Likewise.
13546 (iwmmxt_uavgv8qi3): Likewise.
13547 (iwmmxt_uavgv4hi3): Likewise.
13548 (iwmmxt_tinsrb): Likewise.
13549 (iwmmxt_tinsrh): Likewise.
13550 (iwmmxt_tinsrw): Likewise.
13551 (iwmmxt_textrmub): Likewise.
13552 (iwmmxt_textrmsb): Likewise.
13553 (iwmmxt_textrmuh): Likewise.
13554 (iwmmxt_textrmsh): Likewise.
13555 (iwmmxt_textrmw): Likewise.
13556 (iwmxxt_wshufh): Likewise.
13557 (eqv8qi3): Likewise.
13558 (eqv4hi3): Likewise.
13559 (eqv2si3): Likewise.
13560 (gtuv8qi3): Likewise.
13561 (gtuv4hi3): Likewise.
13562 (gtuv2si3): Likewise.
13563 (gtv8qi3): Likewise.
13564 (gtv4hi3): Likewise.
13565 (gtv2si3): Likewise.
13566 (smax<mode>3_iwmmxt): Likewise.
13567 (umax<mode>3_iwmmxt): Likewise.
13568 (smin<mode>3_iwmmxt): Likewise.
13569 (umin<mode>3_iwmmxt): Likewise.
13570 (iwmmxt_wpackhss): Likewise.
13571 (iwmmxt_wpackwss): Likewise.
13572 (iwmmxt_wpackdss): Likewise.
13573 (iwmmxt_wpackhus): Likewise.
13574 (iwmmxt_wpackwus): Likewise.
13575 (iwmmxt_wpackdus): Likewise.
13576 (iwmmxt_wunpckihb): Likewise.
13577 (iwmmxt_wunpckihh): Likewise.
13578 (iwmmxt_wunpckihw): Likewise.
13579 (iwmmxt_wunpckilb): Likewise.
13580 (iwmmxt_wunpckilh): Likewise.
13581 (iwmmxt_wunpckilw): Likewise.
13582 (iwmmxt_wunpckehub): Likewise.
13583 (iwmmxt_wunpckehuh): Likewise.
13584 (iwmmxt_wunpckehuw): Likewise.
13585 (iwmmxt_wunpckehsb): Likewise.
13586 (iwmmxt_wunpckehsh): Likewise.
13587 (iwmmxt_wunpckehsw): Likewise.
13588 (iwmmxt_wunpckelub): Likewise.
13589 (iwmmxt_wunpckeluh): Likewise.
13590 (iwmmxt_wunpckeluw): Likewise.
13591 (iwmmxt_wunpckelsb): Likewise.
13592 (iwmmxt_wunpckelsh): Likewise.
13593 (iwmmxt_wunpckelsw): Likewise.
13594 (ror<mode>3): Likewise.
13595 (ashr<mode>3_iwmmxt): Likewise.
13596 (lshr<mode>3_iwmmxt): Likewise.
13597 (ashl<mode>3_iwmmxt): Likewise.
13598 (ror<mode>3_di): Likewise.
13599 (ashr<mode>3_di): Likewise.
13600 (lshr<mode>3_di): Likewise.
13601 (ashl<mode>3_di): Likewise.
13602 (iwmmxt_wmadds): Likewise.
13603 (iwmmxt_wmaddu): Likewise.
13604 (iwmmxt_tmia): Likewise.
13605 (iwmmxt_tmiaph): Likewise.
13606 (iwmmxt_tmiabb): Likewise.
13607 (iwmmxt_tmiatb): Likewise.
13608 (iwmmxt_tmiabt): Likewise.
13609 (iwmmxt_tmiatt): Likewise.
13610 (iwmmxt_tmovmskb): Likewise.
13611 (iwmmxt_tmovmskh): Likewise.
13612 (iwmmxt_tmovmskw): Likewise.
13613 (iwmmxt_waccb): Likewise.
13614 (iwmmxt_wacch): Likewise.
13615 (iwmmxt_waccw): Likewise.
13616 (iwmmxt_waligni): Likewise.
13617 (iwmmxt_walignr): Likewise.
13618 (iwmmxt_walignr0): Likewise.
13619 (iwmmxt_walignr1): Likewise.
13620 (iwmmxt_walignr2): Likewise.
13621 (iwmmxt_walignr3): Likewise.
13622 (iwmmxt_wsadb): Likewise.
13623 (iwmmxt_wsadh): Likewise.
13624 (iwmmxt_wsadbz): Likewise.
13625 (iwmmxt_wsadhz): Likewise.
13626 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13627 (iwmmxt_wabsdiffb): Likewise.
13628 (iwmmxt_wabsdiffh): Likewise.
13629 (iwmmxt_wabsdiffw): Likewise.
13630 (iwmmxt_waddsubhx): Likewise
13631 (iwmmxt_wsubaddhx): Likewise.
13632 (addc<mode>3): Likewise.
13633 (iwmmxt_avg4): Likewise.
13634 (iwmmxt_avg4r): Likewise.
13635 (iwmmxt_wmaddsx): Likewise.
13636 (iwmmxt_wmaddux): Likewise.
13637 (iwmmxt_wmaddsn): Likewise.
13638 (iwmmxt_wmaddun): Likewise.
13639 (iwmmxt_wmulwsm): Likewise.
13640 (iwmmxt_wmulwum): Likewise.
13641 (iwmmxt_wmulsmr): Likewise.
13642 (iwmmxt_wmulumr): Likewise.
13643 (iwmmxt_wmulwsmr): Likewise.
13644 (iwmmxt_wmulwumr): Likewise.
13645 (iwmmxt_wmulwl): Likewise.
13646 (iwmmxt_wqmulm): Likewise.
13647 (iwmmxt_wqmulwm): Likewise.
13648 (iwmmxt_wqmulmr): Likewise.
13649 (iwmmxt_wqmulwmr): Likewise.
13650 (iwmmxt_waddbhusm): Likewise.
13651 (iwmmxt_waddbhusl): Likewise.
13652 (iwmmxt_wqmiabb): Likewise.
13653 (iwmmxt_wqmiabt): Likewise.
13654 (iwmmxt_wqmiatb): Likewise.
13655 (iwmmxt_wqmiatt): Likewise.
13656 (iwmmxt_wqmiabbn): Likewise.
13657 (iwmmxt_wqmiabtn): Likewise.
13658 (iwmmxt_wqmiatbn): Likewise.
13659 (iwmmxt_wqmiattn): Likewise.
13660 (iwmmxt_wmiabb): Likewise.
13661 (iwmmxt_wmiabt): Likewise.
13662 (iwmmxt_wmiatb): Likewise.
13663 (iwmmxt_wmiatt): Likewise.
13664 (iwmmxt_wmiabbn): Likewise.
13665 (iwmmxt_wmiabtn): Likewise.
13666 (iwmmxt_wmiatbn): Likewise.
13667 (iwmmxt_wmiattn): Likewise.
13668 (iwmmxt_wmiawbb): Likewise.
13669 (iwmmxt_wmiawbt): Likewise.
13670 (iwmmxt_wmiawtb): Likewise.
13671 (iwmmxt_wmiawtt): Likewise.
13672 (iwmmxt_wmiawbbn): Likewise.
13673 (iwmmxt_wmiawbtn): Likewise.
13674 (iwmmxt_wmiawtbn): Likewise.
13675 (iwmmxt_wmiawttn): Likewise.
13676 (iwmmxt_wmerge): Likewise.
13677 (iwmmxt_tandc<mode>3): Likewise.
13678 (iwmmxt_torc<mode>3): Likewise.
13679 (iwmmxt_torvsc<mode>3): Likewise.
13680 (iwmmxt_textrc<mode>3): Likewise.
13681 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13682 (wmmxt_pack): Likewise.
13683 (wmmxt_mult_c1): Likewise.
13684 (wmmxt_mult_c2): Likewise.
13685 (wmmxt_alu_c1): Likewise.
13686 (wmmxt_alu_c2): Likewise.
13687 (wmmxt_alu_c3): Likewise.
13688 (wmmxt_transfer_c1): Likewise.
13689 (wmmxt_transfer_c2): Likewise.
13690 (wmmxt_transfer_c3): Likewise.
13691 (marvell_f_iwmmxt_wstr): Likewise.
13692 (marvell_f_iwmmxt_wldr): Likewise.
13694 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13696 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13698 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13701 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13702 * lra-constraints.c (need_for_split_p): Check call used hard regs
13703 living through calls.
13705 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13706 call used regs for call insn.
13708 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13711 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13712 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13714 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13716 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13718 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13720 * lra-constraints.c (need_for_split_p): Check call used hard regs
13721 living through calls.
13723 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13726 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13727 to tie with any other modes. Eliminate Altivec vector mode tests,
13728 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13729 code, to return 0 if testing MODE2 for a condition, if we've
13730 already tested MODE1 for the same condition.
13732 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13734 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13737 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13739 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13740 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13742 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13744 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13746 (aarch64_symbolic_constant_p): Remove.
13747 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13748 static. Fix line length and white space.
13749 (aarch64_symbolic_constant_p): Remove.
13750 * config/aarch64/predicates.md (aarch64_valid_symref):
13751 Use aarch64_classify_symbol_expression.
13753 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13755 * config/arm/constraints.md (Ts): New constraint.
13756 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13758 (compare_scc): Use "Ts" constraint for operand 0.
13759 (ior_scc_scc): Likewise.
13760 (and_scc_scc): Likewise.
13761 (and_scc_scc_nodom): Likewise.
13762 (ior_scc_scc_cmp): Likewise for operand 7.
13763 (and_scc_scc_cmp): Likewise.
13764 * config/arm/thumb2.md (thumb2_movsi_insn):
13765 Add alternatives for 16-bit encodings.
13766 (thumb2_movhi_insn): Likewise.
13767 (thumb2_movsicc_insn): Likewise.
13768 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13769 (thumb2_negscc): Use "Ts" constraint.
13770 Move mvn instruction outside cond_exec block.
13771 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13772 for 16-bit encodings.
13774 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13776 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13778 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13779 (mulsi3subsi): Likewise.
13780 (mulsidi3adddi): Likewise.
13781 (mulsidi3_v6): Likewise.
13782 (umulsidi3_v6): Likewise.
13783 (umulsidi3adddi_v6): Likewise.
13784 (smulsi3_highpart_v6): Likewise.
13785 (umulsi3_highpart_v6): Likewise.
13786 (mulhisi3tb): Likewise.
13787 (mulhisi3bt): Likewise.
13788 (mulhisi3tt): Likewise.
13789 (maddhisi4): Likewise.
13790 (maddhisi4tb): Likewise.
13791 (maddhisi4tt): Likewise.
13792 (maddhidi4): Likewise.
13793 (maddhidi4tb): Likewise.
13794 (maddhidi4tt): Likewise.
13795 (zeroextractsi_compare0_scratch): Likewise.
13796 (insv_zero): Likewise.
13797 (insv_t2): Likewise.
13798 (anddi_notzesidi_di): Likewise.
13799 (anddi_notsesidi_di): Likewise.
13800 (andsi_notsi_si): Likewise.
13801 (iordi_zesidi_di): Likewise.
13802 (xordi_zesidi_di): Likewise.
13803 (andsi_iorsi3_notsi): Likewise.
13804 (smax_0): Likewise.
13805 (smax_m1): Likewise.
13806 (smin_0): Likewise.
13807 (not_shiftsi): Likewise.
13808 (unaligned_loadsi): Likewise.
13809 (unaligned_loadhis): Likewise.
13810 (unaligned_loadhiu): Likewise.
13811 (unaligned_storesi): Likewise.
13812 (unaligned_storehi): Likewise.
13813 (extv_reg): Likewise.
13814 (extzv_t2): Likewise.
13815 (divsi3): Likewise.
13816 (udivsi3): Likewise.
13817 (arm_zero_extendhisi2addsi): Likewise.
13818 (arm_zero_extendqisi2addsi): Likewise.
13819 (compareqi_eq0): Likewise.
13820 (arm_extendhisi2_v6): Likewise.
13821 (arm_extendqisi2addsi): Likewise.
13822 (arm_movt): Likewise.
13823 (thumb2_ldrd): Likewise.
13824 (thumb2_ldrd_base): Likewise.
13825 (thumb2_ldrd_base_neg): Likewise.
13826 (thumb2_strd): Likewise.
13827 (thumb2_strd_base): Likewise.
13828 (thumb2_strd_base_neg): Likewise.
13829 (arm_negsi2): Add alternative for 16-bit encoding.
13830 (arm_one_cmplsi2): Likewise.
13832 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13834 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13835 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13836 (movdfcc): Likewise.
13837 * config/arm/vfp.md (*thumb2_movsf_vfp):
13838 Disable predication for arm_restrict_it.
13839 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13840 (*thumb2_movdfcc_vfp): Likewise.
13841 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13842 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13843 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13844 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13845 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13846 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13847 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13848 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13849 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13850 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13851 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13852 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13853 Disable predication for arm_restrict_it.
13855 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13857 * config/i386/bmiintrin.h (_bextr_u32): New.
13858 (_bextr_u64): Ditto.
13860 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13862 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13863 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13864 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13865 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13866 * config/mips/n32-elf.h: ...this new file.
13868 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13871 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13872 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13874 2013-06-27 Catherine Moore <clm@codesourcery.com>
13876 * config/mips/mips-tables.opt: Regenerate.
13877 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13878 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13879 * doc/invoke.texi: Add -m14kc.
13881 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13884 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13885 constraints of operand 1 and 2.
13888 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13889 to match RTL canonicalization. Swap predicates and
13890 constraints of operand 1 and 2.
13892 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
13894 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
13895 Process OP_INOUT regs for splitting too.
13897 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13899 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
13900 decl before the loop, initialize to NULL.
13901 (vectorizable_load): Initialize ptr_incr to NULL.
13903 2013-06-27 Martin Jambor <mjambor@suse.cz>
13906 * ipa-ref.h (ipa_maybe_record_reference): Declare.
13907 * ipa-ref.c (ipa_maybe_record_reference): New function.
13908 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
13909 * ipa-cp.c (create_specialized_node): Record potential references from
13911 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
13913 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13915 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
13916 parameter 'mode' of type 'enum machine_mode mode'; change to pass
13917 'mode' to force_reg.
13918 (aarch64_add_offset): Update calls to aarch64_force_temporary.
13919 (aarch64_expand_mov_immediate): Likewise.
13921 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
13923 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
13924 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
13926 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13928 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
13929 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13930 (struct machine_function): Add tbegin_p.
13931 (s390_canonicalize_comparison): Fold CC mode compares to
13932 conditional jump if possible.
13933 (s390_emit_jump): Return the emitted jump.
13934 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
13935 Handle CCRAWmode compares.
13936 (s390_option_override): Default to -mhtm if available.
13937 (s390_reg_clobbered_rtx): Handle floating point regs as well.
13938 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
13939 FPRs instead of df_regs_ever_live_p.
13940 (s390_optimize_nonescaping_tx): New function.
13941 (s390_init_frame_layout): Extend clobbered_regs array to cover
13943 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
13944 (s390_expand_tbegin): New function.
13945 (enum s390_builtin): New enum definition.
13946 (code_for_builtin): New array definition.
13947 (s390_init_builtins): New function.
13948 (s390_expand_builtin): New function.
13949 (TARGET_INIT_BUILTINS): Define.
13950 (TARGET_EXPAND_BUILTIN): Define.
13951 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
13952 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
13953 (s390_alc_comparison): Likewise.
13954 * config/s390/s390-modes.def: Add CCRAWmode.
13955 * config/s390/s390.h (processor_flags): Add PF_TX.
13956 (TARGET_CPU_HTM): Define macro.
13957 (TARGET_HTM): Define macro.
13958 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
13959 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
13960 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
13961 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
13962 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
13963 (TBEGIN_MASK, TBEGINC_MASK): New constants.
13964 ("*cc_to_int"): Move up.
13965 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
13966 constants other than 0.
13967 ("*ccraw_to_int"): New insn and splitter definition.
13968 ("tbegin", "tbegin_nofloat", "tbegin_retry")
13969 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
13970 ("tx_assist"): New expander.
13971 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
13972 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13973 * config/s390/s390.opt: Add -mhtm option.
13974 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13975 * config/s390/htmxlintrin.h: New file.
13976 * config/s390/htmintrin.h: New file.
13977 * config/s390/s390intrin.h: New file.
13978 * doc/extend.texi: Document htm builtins.
13979 * config.gcc: Add the new header files to extra_headers.
13981 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13983 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
13984 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
13986 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
13987 Pat Haugen <pthaugen@us.ibm.com>
13988 Peter Bergner <bergner@vnet.ibm.com>
13990 * config/rs6000/power8.md: New.
13991 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13992 setting for power8 entry.
13993 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13994 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13995 test for Power4/Power5 only.
13996 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13998 (force_new_group): Adjust comment.
13999 * config/rs6000/rs6000.md: Include power8.md.
14001 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14003 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14004 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14006 (tune_params): Update comment.
14007 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14008 (arm_max_conditional_execute): New function.
14009 (thumb2_final_prescan_insn): Use max_insn_skipped and
14010 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14012 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14014 PR tree-optimization/57705
14015 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14016 SSA_NAME step, provided that it is not defined inside the loop.
14017 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14018 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14020 2013-06-25 Martin Jambor <mjambor@suse.cz>
14022 PR middle-end/57670
14023 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14024 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14026 (ipa_note_param_call): Initialize member_ptr flag.
14027 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14028 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14029 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14030 (ipa_read_indirect_edge_info): Likewise.
14032 2013-06-25 Richard Biener <rguenther@suse.de>
14034 PR middle-end/56977
14035 * passes.c (init_optimization_passes): Move pass_fold_builtins
14036 and pass_dce earlier with -Og.
14038 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14040 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14041 <BIT_FIELD_REF>: Remove trailing TAB.
14042 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14045 2013-06-24 Martin Jambor <mjambor@suse.cz>
14047 PR tree-optimization/57358
14048 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14049 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14050 (ipa_analyze_params_uses): Generate pessimistic info when true.
14052 2013-06-24 Martin Jambor <mjambor@suse.cz>
14054 PR tree-optimization/57539
14055 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14056 global.inlined_to of the new node to it. All callers changed.
14057 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14058 inlining_into, pass it to cgraph_clone_node.
14059 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14060 ipa_free_edge_args_substructures.
14061 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14062 rdesc linked list. Do not assert rdesc edges have inlined caller.
14063 Assert we have found an rdesc in the rdesc list.
14065 2013-06-24 Richard Biener <rguenther@suse.de>
14067 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14068 (pointer_set_lookup): Declare.
14069 (class pointer_map): New template class implementing a
14070 generic pointer to T map.
14071 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14072 pointer_map<T>::contains, pointer_map<T>::insert,
14073 pointer_map<T>::traverse): New functions.
14074 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14075 (pointer_set_lookup): New function.
14076 (pointer_set_contains): Use pointer_set_lookup.
14077 (pointer_set_insert): Likewise.
14078 (insert_aux): Remove.
14079 (struct pointer_map_t): Embed a pointer_set_t.
14080 (pointer_map_create): Adjust.
14081 (pointer_map_destroy): Likewise.
14082 (pointer_map_contains): Likewise.
14083 (pointer_map_insert): Likewise.
14084 (pointer_map_traverse): Likewise.
14085 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14086 pointer_map<unsigned> instead of a pointer_map_t.
14087 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14088 (streamer_tree_cache_lookup): Likewise.
14089 (streamer_tree_cache_create): Likewise.
14090 (streamer_tree_cache_delete): Likewise.
14091 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14092 pointer_map<unsigned> instead of a pointer_map_t.
14093 (lto_init_tree_ref_encoder): Adjust.
14094 (lto_destroy_tree_ref_encoder): Likewise.
14095 * lto-section-out.c (lto_output_decl_index): Likewise.
14096 (lto_record_function_out_decl_state): Likewise.
14097 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14099 2013-06-24 Richard Biener <rguenther@suse.de>
14101 PR tree-optimization/57488
14102 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14104 2013-06-24 Alan Modra <amodra@gmail.com>
14106 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14107 (gen_easy_altivec_constant): Likewise.
14108 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14109 easy_vector_constant_msb): Likewise.
14111 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14114 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14115 add missing return true.
14117 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14120 * config/sh/predicates.md (general_extend_operand): Invoke
14121 general_movsrc_operand for memory operands.
14122 (general_movsrc_operand): Allow reg+reg addressing, do not use
14123 general_operand for memory operands.
14125 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14127 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14128 when current target options does not apply.
14129 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14130 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14131 * config/i386/bmiintrin.h: Pass appropriate target
14132 attributes to header.
14133 * config/i386/mmintrin.h: Ditto.
14134 * config/i386/nmmintrin.h: Ditto.
14135 * config/i386/avx2intrin.h: Ditto.
14136 * config/i386/fxsrintrin.h: Ditto.
14137 * config/i386/tbmintrin.h: Ditto.
14138 * config/i386/xsaveintrin.h: Ditto.
14139 * config/i386/f16cintrin.h: Ditto.
14140 * config/i386/xtestintrin.h: Ditto.
14141 * config/i386/xsaveoptintrin.h: Ditto.
14142 * config/i386/bmi2intrin.h: Ditto.
14143 * config/i386/lzcntintrin.h: Ditto.
14144 * config/i386/smmintrin.h: Ditto.
14145 * config/i386/wmmintrin.h: Ditto.
14146 * config/i386/x86intrin.h: Remove all header include guards.
14147 * config/i386/prfchwintrin.h: Ditto.
14148 * config/i386/pmmintrin.h: Ditto.
14149 * config/i386/tmmintrin.h: Ditto.
14150 * config/i386/xmmintrin.h: Ditto.
14151 * config/i386/popcntintrin.h: Ditto.
14152 * config/i386/rdseedintrin.h: Ditto.
14153 * config/i386/ammintrin.h: Ditto.
14154 * config/i386/emmintrin.h: Ditto.
14155 * config/i386/immintrin.h: Remove all header include guards.
14156 * config/i386/fma4intrin.h: Ditto.
14157 * config/i386/lwpintrin.h: Ditto.
14158 * config/i386/xopintrin.h: Ditto.
14159 * config/i386/ia32intrin.h: Ditto.
14160 * config/i386/avxintrin.h: Ditto.
14161 * config/i386/rtmintrin.h: Ditto.
14162 * config/i386/fmaintrin.h: Ditto.
14163 * config/i386/mm3dnow.h: Ditto.
14165 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14167 * common/config/i386/i386-common.c: Handle LZCNT.
14169 2013-06-22 Andi Kleen <ak@linux.intel.com>
14171 * doc/extend.texi: Use __atomic_store_n instead of
14172 __atomic_store in HLE example.
14174 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14176 * config/sh/sh.c: Remove <cstdlib> workaround.
14178 2013-06-21 Andi Kleen <ak@linux.intel.com>
14180 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14182 2013-06-21 Andi Kleen <ak@linux.intel.com>
14184 * doc/extend.texi: Document that __atomic_clear and
14185 __atomic_test_and_set should only be used with bool.
14187 2013-06-20 Jan Hubicka <jh@suse.cz>
14189 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14190 types_same_for_odr.
14191 * tree.c (decls_same_for_odr): New function.
14192 (same_for_edr): New function.
14193 (types_same_for_odr): New function.
14194 (get_binfo_at_offset): Use it.
14195 * tree.h (types_same_for_odr): Declare.
14197 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14198 Jason Merrill <jason@redhat.com>
14200 * system.h: Include <cstdlib> as well as <stdlib.h>.
14202 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14205 * config/i386/i386.c (construct_container): Report error if
14206 long double is used with disabled x87 float returns.
14208 2013-06-20 Jan Hubicka <jh@suse.cz>
14210 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14212 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14214 PR rtl-optimization/57425
14215 PR rtl-optimization/57569
14216 * alias.c (write_dependence_p): Remove parameters mem_mode and
14217 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14218 Changed all callers.
14219 (canon_anti_dependence): Get comments and semantics in sync.
14220 Add parameter mem_canonicalized. Changed all callers.
14221 * rtl.h (canon_anti_dependence): Update prototype.
14223 2013-06-20 Richard Biener <rguenther@suse.de>
14225 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14226 byte case, inline streamer_read_uchar and defer section
14229 2013-06-20 Richard Biener <rguenther@suse.de>
14231 PR tree-optimization/57584
14232 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14233 SSA names into the expanded expression that take part in
14234 abnormal coalescing.
14236 2013-06-19 Sharad Singhai <singhai@google.com>
14238 * gcov.c (print_usage): Handle new option.
14239 (process_args): Ditto.
14240 (get_gcov_intermediate_filename): New function.
14241 (output_intermediate_file): New function.
14242 (output_gcov_file): New function
14243 (generate_results): Handle new option.
14244 (release_function): Relase demangled name.
14245 (read_graph_file): Handle demangled name.
14246 (output_lines): Ditto.
14247 * doc/gcov.texi: Document gcov intermediate format.
14249 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14252 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14253 (lra_emit_add): Use the functions. Add comment about Y as an
14256 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14259 * collect2.c (collect_atexit): New.
14260 (collect_exit): Delete.
14261 (main): Register collect_atexit with atexit.
14262 (collect_wait): Change collect_exit to exit.
14264 * collect2.h (collect_exit): Delete.
14265 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14267 2013-06-19 Wei Mi <wmi@google.com>
14269 PR rtl-optimization/57518
14270 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14271 if regno is used in paradoxical subreg.
14272 (update_equiv_regs): Check pdx_subregs[regno] before
14273 set a reg to be equivalent with a mem.
14275 2013-06-19 Matthias Klose <doko@ubuntu.com>
14278 * file-find.h (find_a_file): Add a mode parameter.
14279 * file-find.c (find_a_file): Likewise.
14280 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14281 with X_OK for the executables.
14282 * collect2.c (main): Call find_a_file with X_OK.
14284 2013-06-19 Steve Ellcey <sellcey@mips.com>
14287 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14288 Use NEXT_INSN instead of next_real_insn.
14290 2013-06-19 Jan Hubicka <jh@suse.cz>
14292 * cgraph.h (const_value_known_p): Replace by ...
14293 (ctor_for_folding): .. this one.
14294 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14295 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14296 * expr.c (expand_expr_real_1): Likewise.
14297 (string_constant): Likewise.
14298 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14299 * ipa.c (process_references): Likewise.
14300 (symtab_remove_unreachable_nodes): Likewise.
14301 * ipa-inline-analysis.c (param_change_prob): Likewise.
14302 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14303 (get_base_constructor): Likwise.
14304 * varpool.c (varpool_remove_node): Likewise.
14305 (varpool_remove_initializer): LIkewise.
14306 (dump_varpool_node): LIkwise.
14307 (const_value_known_p): Rewrite to ...
14308 (ctor_for_folding): ... this one.
14310 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14313 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14314 PERSONALITY in $PATH derived prefixes.
14316 2013-06-19 Jeff Law <law@redhat.com>
14318 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14321 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14322 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14325 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14327 * config/arm/vfp.md: Move VFP instruction classification documentation
14329 * config/arm/arm.md: ... here. Update instruction classification
14332 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14334 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14335 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14336 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14338 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14341 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14342 with NEXT_INSN. Use tablejump_p to check for jump table data
14345 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14348 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14349 that now in C++ the value is correct per the C++ standards.
14351 2013-06-19 Richard Biener <rguenther@suse.de>
14353 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14354 for global context.
14356 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14359 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14362 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14363 with next_active_insn.
14365 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14367 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14368 functions are inlined during failures to flag an error.
14369 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14370 in early inline pass.
14372 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14374 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14377 2013-06-18 Julian Brown <julian@codesourcery.com>
14379 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14380 Permit virtual register pre-reload if !strict.
14381 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14383 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14385 * config/arm/neon.md (movmisalign<mode>): Use
14386 neon_perm_struct_or_reg_operand instead of
14387 neon_struct_or_register_operand.
14388 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14389 neon_permissive_struct_operand instead of neon_struct_operand.
14390 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14391 neon_vector_mem_operand.
14392 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14393 neon_vector_mem_operand.
14394 (neon_permissive_struct_operand): New.
14395 (neon_struct_or_register_operand): Rename to...
14396 (neon_perm_struct_or_reg_operand): This. Adjust call to
14397 neon_vector_mem_operand.
14399 2013-06-18 Richard Biener <rguenther@suse.de>
14401 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14402 * lto-streamer.h: Include pointer-set.h.
14403 (struct lto_decl_slot): Remove.
14404 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14405 Remove next_index entry.
14406 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14407 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14408 (lto_init_tree_ref_encoder): Adjust.
14409 (lto_destroy_tree_ref_encoder): Likewise.
14410 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14411 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14412 (lto_output_decl_index): Adjust.
14413 (lto_new_out_decl_state): Likewise.
14414 (lto_record_function_out_decl_state): Likewise.
14415 * lto-streamer-out.c (copy_function): Likewise.
14417 2013-06-18 Richard Biener <rguenther@suse.de>
14419 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14420 * cgraphunit.c: Include cfgloop.h.
14421 (init_lowered_empty_function): Initialize the loop tree.
14422 (assemble_thunk): Insert new BBs into loops.
14424 2013-06-18 Richard Biener <rguenther@suse.de>
14426 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14427 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14428 the map from cache entry to cache index optional.
14429 (streamer_tree_cache_replace_tree): Adjust accordingly.
14430 (streamer_tree_cache_append): Likewise.
14431 (streamer_tree_cache_delete): Likewise.
14432 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14433 streamer cache map from cache entry to cache index.
14434 * lto-streamer-out.c (create_output_block): Adjust.
14436 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14438 * config/arm/arm.md (attribute "insn"): Move multiplication and
14439 division attributes to...
14440 (attribute "type"): ... here. Remove mult.
14441 (attribute "mul32"): New attribute.
14442 (attribute "mul64"): Add umaal.
14443 (*arm_mulsi3): Update attributes.
14444 (*arm_mulsi3_v6): Likewise.
14445 (*thumb_mulsi3): Likewise.
14446 (*thumb_mulsi3_v6): Likewise.
14447 (*mulsi3_compare0): Likewise.
14448 (*mulsi3_compare0_v6): Likewise.
14449 (*mulsi_compare0_scratch): Likewise.
14450 (*mulsi_compare0_scratch_v6): Likewise.
14451 (*mulsi3addsi): Likewise.
14452 (*mulsi3addsi_v6): Likewise.
14453 (*mulsi3addsi_compare0): Likewise.
14454 (*mulsi3addsi_compare0_v6): Likewise.
14455 (*mulsi3addsi_compare0_scratch): Likewise.
14456 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14457 (*mulsi3subsi): Likewise.
14458 (*mulsidi3adddi): Likewise.
14459 (*mulsi3addsi_v6): Likewise.
14460 (*mulsidi3adddi_v6): Likewise.
14461 (*mulsidi3_nov6): Likewise.
14462 (*mulsidi3_v6): Likewise.
14463 (*umulsidi3_nov6): Likewise.
14464 (*umulsidi3_v6): Likewise.
14465 (*umulsidi3adddi): Likewise.
14466 (*umulsidi3adddi_v6): Likewise.
14467 (*smulsi3_highpart_nov6): Likewise.
14468 (*smulsi3_highpart_v6): Likewise.
14469 (*umulsi3_highpart_nov6): Likewise.
14470 (*umulsi3_highpart_v6): Likewise.
14471 (mulhisi3): Likewise.
14472 (*mulhisi3tb): Likewise.
14473 (*mulhisi3bt): Likewise.
14474 (*mulhisi3tt): Likewise.
14475 (maddhisi4): Likewise.
14476 (*maddhisi4tb): Likewise.
14477 (*maddhisi4tt): Likewise.
14478 (maddhidi4): Likewise.
14479 (*maddhidi4tb): Likewise.
14480 (*maddhidi4tt): Likewise.
14481 (divsi3): Likewise.
14482 (udivsi3): Likewise.
14483 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14484 (thumb2_mulsi_short_compare0): Likewise.
14485 (thumb2_mulsi_short_compare0_scratch): Likewise.
14486 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14487 (1020mult2): Likewise.
14488 (1020mult3): Likewise.
14489 (1020mult4): Likewise.
14490 (1020mult5): Likewise.
14491 (1020mult6): Likewise.
14492 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14494 (cortex_a15_mult64): Likewise.
14495 (cortex_a15_sdiv): Likewise.
14496 (cortex_a15_udiv): Likewise.
14497 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14503 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14504 (pj4_ir_div): Likewise.
14505 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14506 (11_mult2): Likewise.
14507 (11_mult3): Likewise.
14508 (11_mult4): Likewise.
14509 (11_mult5): Likewise.
14510 (11_mult6): Likewise.
14511 (11_mult7): Likewise.
14512 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14513 (cortex_a8_mla): Likewise.
14514 (cortex_a8_mull): Likewise.
14515 (cortex_a8_smulwy): Likewise.
14516 (cortex_a8_smlald): Likewise.
14517 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14518 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14519 (cortex_r4_mul_3): Likewise.
14520 (cortex_r4_mla_4): Likewise.
14521 (cortex_r4_mla_3): Likewise.
14522 (cortex_r4_smlald): Likewise.
14523 (cortex_r4_mull): Likewise.
14524 (cortex_r4_sdiv): Likewise.
14525 (cortex_r4_udiv): Likewise.
14526 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14527 (cortex_a7_idiv): Likewise.
14528 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14529 (9_mult2): Likewise.
14530 (9_mult3): Likewise.
14531 (9_mult4): Likewise.
14532 (9_mult5): Likewise.
14533 (9_mult6): Likewise.
14534 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14535 (cortex_a53_sdiv): Likewise.
14536 (cortex_a53_udiv): Likewise.
14537 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14538 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14539 (mp626_mult2): Likewise.
14540 (mp626_mult3): Likewise.
14541 (mp626_mult4): Likewise.
14542 * config/arm/fa526.md (526_mult1): Update attribute change.
14543 (526_mult2): Likewise.
14544 * config/arm/arm-generic.md (mult): Update attribute change.
14545 (mult_ldsched_strongarm): Likewise.
14546 (mult_ldsched): Likewise.
14547 (multi_cycle): Likewise.
14548 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14549 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14550 (606te_mult2): Likewise.
14551 (606te_mult3): Likewise.
14552 (606te_mult4): Likewise.
14553 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14554 (cortex_a9_mac16): Likewise.
14555 (cortex_a9_multiply): Likewise.
14556 (cortex_a9_mac): Likewise.
14557 (cortex_a9_multiply_long): Likewise.
14558 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14559 (626te_mult2): Likewise.
14560 (626te_mult3): Likewise.
14561 (626te_mult4): Likewise.
14563 2013-06-18 Richard Biener <rguenther@suse.de>
14566 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14568 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14571 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14572 with next_active_insn.
14574 2013-06-18 Alan Modra <amodra@gmail.com>
14576 * config/rs6000/rs6000.h (enum data_align): New.
14577 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14578 (DATA_ABI_ALIGNMENT): Define.
14579 (CONSTANT_ALIGNMENT): Correct comment.
14580 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14581 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14583 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14585 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14586 ATTRIBUTE_UNUSED marking.
14588 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14590 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14591 alternative and update.
14592 (aarch64_dup_lanedi): Delete.
14593 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14594 * config/aarch64/aarch64-simd-builtins.def: Update.
14596 2013-06-17 Richard Biener <rguenther@suse.de>
14598 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14599 (lto_input_scc): Declare.
14600 (lto_input_tree_1): Likewise.
14601 (struct lto_stats_d): Add num_tree_bodies_output and
14602 num_pickle_refs_output.
14603 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14604 (lto_read_tree_1): Split out from ...
14605 (lto_read_tree): ... this.
14606 (lto_input_scc): New function.
14607 (lto_input_tree_1): Split out from ...
14608 (lto_input_tree): ... this. Handle LTO_tree_scc.
14609 (lto_data_in_create): Create the streamer cache without hashes.
14610 * lto-streamer-out.c (create_output_block): Create the streamer
14611 cache with hashes when not doing WPA.
14612 (lto_write_tree_1): Split out from ...
14613 (lto_write_tree): ... this.
14614 (get_symbol_initial_value): New function.
14615 (lto_output_tree_1): Split out from ...
14616 (lto_output_tree): ... this. Write trees as series of SCCs
14617 using a DFS walk via DFS_write_tree.
14618 (struct sccs, struct scc_entry): New types.
14619 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14620 (DFS_write_tree_body): New function.
14621 (DFS_write_tree): Likewise.
14622 (hash_tree): Likewise.
14623 (scc_entry_compare): Likewise.
14624 (hash_scc): Likewise.
14625 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14626 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14627 TREE_CHAIN as regular reference.
14628 (streamer_read_integer_cst): Remove.
14629 (streamer_get_pickled_tree): Adjust.
14630 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14631 of DECL_EXTERNALs in BLOCK_VARS for now.
14632 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14634 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14635 Add hash value argument and record that if hashes are recorded
14637 (streamer_tree_cache_insert_1): Adjust.
14638 (streamer_tree_cache_insert): Likewise.
14639 (streamer_tree_cache_insert_at): Rename to ...
14640 (streamer_tree_cache_replace_tree): ... this and adjust.
14641 (streamer_tree_cache_append): Adjust.
14642 (record_common_node): Likewise.
14643 (streamer_tree_cache_create): Add argument whether to
14644 record hash values together with trees.
14645 (streamer_tree_cache_delete): Adjust.
14646 * tree-streamer.h (struct streamer_tree_cache_d): Add
14648 (streamer_read_integer_cst): Remove.
14649 (streamer_tree_cache_insert): Adjust.
14650 (streamer_tree_cache_append): Likewise.
14651 (streamer_tree_cache_insert_at): Rename to ...
14652 (streamer_tree_cache_replace_tree): ... this and adjust.
14653 (streamer_tree_cache_create): Add argument whether to record hashes.
14654 (streamer_tree_cache_get): Rename to ...
14655 (streamer_tree_cache_get_tree): ... this.
14656 (streamer_tree_cache_get_hash): New function.
14657 * tree.c (cache_integer_cst): New function.
14658 * tree.h (cache_integer_cst): Declare.
14659 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14660 * lto-symtab.c (lto_varpool_replace_node): Only release
14661 DECL_INITIAL of non-prevailing decls.
14662 * varpool.c (varpool_remove_initializer): Do not release
14663 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14665 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14667 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14668 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14669 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14670 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14671 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14672 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14673 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14674 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14675 instead of TARGET_64BIT.
14676 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14677 Require ISA_HAS_<D>DIV.
14679 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14681 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14682 (mips*-*-linux*): Move default with_llsc setting to where other
14684 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14686 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14687 Likewise. Remove default with_tune setting. Move default float
14688 setting to its own block. Handle with_llsc in the same block as above.
14690 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14692 PR rtl-optimization/57425
14693 PR rtl-optimization/57569
14694 * alias.c (write_dependence_p): Add new parameters mem_mode,
14695 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14696 Changed all callers.
14697 (canon_anti_dependence): New function.
14698 * cse.c (check_dependence): Use canon_anti_dependence.
14699 * cselib.c (cselib_invalidate_mem): Likewise.
14700 * rtl.h (canon_anti_dependence): Declare.
14702 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14704 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14705 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14706 ".set mips3" for 64-bit targets.
14708 2013-06-15 Dehao Chen <dehao@google.com>
14710 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14711 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14712 (gimple_check_call_args): Likewise.
14713 * value-prof.c (check_ic_target): Likewise.
14714 * ipa-inline.c (early_inliner): Likewise.
14715 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14716 * cgraph.c (cgraph_create_edge_1): Likewise.
14717 (cgraph_make_edge_direct): Likewise.
14719 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14722 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14723 rs6000_output_move_128bit to handle emitting quad memory
14724 operations. Set attribute length to 8 bytes.
14726 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14728 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14730 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14731 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14732 (aarch64_<su>mlsl<mode>): Likewise.
14734 2013-06-14 Mike Stump <mikestump@comcast.net>
14736 * Makefile.in (TARGET_H): Add insn-codes.h.
14738 2013-06-14 Alan Modra <amodra@gmail.com>
14740 PR middle-end/57134
14741 PR middle-end/57586
14742 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14743 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14744 bitfield expansion when EXPAND_MEMORY.
14745 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14747 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14749 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14750 test for clearing quad memory on 32-bit later.
14752 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14754 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14755 (fold_negate_expr): Likewise.
14756 (fold_real_zero_addition_p): Handle vectors.
14757 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14759 2013-06-14 Alan Modra <amodra@gmail.com>
14761 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14763 2013-06-13 Jan Hubicka <jh@suse.cz>
14765 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14766 Local comdats are not externally visible.
14767 * symtab.c (dump_symtab_base): Dump externally visible.
14768 (verify_symtab_base): Verify back links in the symtab hash.
14770 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14772 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14773 CONVERT_EXPR as equal nodes.
14775 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14777 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14779 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14781 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14782 Generalize to complex and vector.
14783 * tree.c (build_all_ones_cst): New function.
14784 * tree.h (build_all_ones_cst): Declare it.
14786 2013-06-13 Alan Modra <amodra@gmail.com>
14788 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14789 * config/rs6000/rs6000.md (signbittf2): New insn.
14790 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14791 (abstf2_internal, cmptf_internal2): Likewise.
14792 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14794 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14795 Pat Haugen <pthaugen@us.ibm.com>
14796 Peter Bergner <bergner@vnet.ibm.com>
14798 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14799 power8 byte, half-word, and quad-word atomic instructions.
14800 (emit_store_conditional): Likewise.
14801 (rs6000_expand_atomic_compare_and_swap): Likewise.
14802 (rs6000_expand_atomic_op): Likewise.
14804 * config/rs6000/sync.md (larx): Add new modes for power8.
14806 (AINT): New mode iterator to include TImode as well as normal
14807 integer modes on power8.
14808 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14809 that VSX registers are not considered. Use AINT mode iterator
14810 instead of INT1 to allow inclusion of quad word atomic operations
14812 (load_locked<mode>): Likewise.
14813 (store_conditional<mode>): Likewise.
14814 (atomic_compare_and_swap<mode>): Likewise.
14815 (atomic_exchange<mode>): Likewise.
14816 (atomic_nand<mode>): Likewise.
14817 (atomic_fetch_<fetchop_name><mode>): Likewise.
14818 (atomic_nand_fetch<mode>): Likewise.
14819 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14821 (ATOMIC): On power8, add QImode, HImode modes.
14822 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14823 modes that promote to SImode.
14824 (load_lockedti): Convert TImode arguments to PTImode, so that we
14825 get a guaranteed even/odd register pair.
14826 (load_lockedpti): Likewise.
14827 (store_conditionalti): Likewise.
14828 (store_conditionalpti): Likewise.
14830 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14831 atomic load/store instructions.
14834 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14836 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14838 (insn_count): New attribute, with most cases extracted from...
14839 (length): ...here. Redefine most cases in terms of insn_count.
14840 (single_insn): Delete.
14841 (can_delay): Use insn_count to check for single instructions.
14842 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14843 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14844 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14845 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14846 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14847 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14848 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14849 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14850 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14851 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14852 rather than "length".
14853 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14854 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14855 Use "insn_count" rather than "length".
14856 * config/mips/mips-dsp.md
14857 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14858 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14861 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14863 PR tree-optimization/57361
14864 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14866 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14868 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14870 (aarch64_simd_combine<mode>): New instruction expansion.
14871 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14872 function prototype.
14873 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14874 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14876 2013-06-12 Jan Hubicka <jh@suse.cz>
14878 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14879 decl has when in streaming stage.
14880 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14881 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14883 2013-06-12 Roland Stigge <stigge@antcom.de>
14886 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14888 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14890 PR tree-optimization/57537
14891 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14892 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
14894 2013-06-12 Richard Biener <rguenther@suse.de>
14896 * data-streamer.h (streamer_write_char_stream): CSE
14897 obs->current_pointer.
14898 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
14899 streamer_write_char_stream manually and optimize the resulting loop.
14900 (streamer_write_hwi_stream): Likewise.
14902 2013-06-12 Jan Hubicka <jh@suse.cz>
14904 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
14905 * cgraph.h (varpool_create_empty_node): Declare.
14906 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
14908 * symtab.c (symtab_unregister_node): Be lax about missin entries
14910 (symtab_get_node): Update comment.
14911 * varpool.c (varpool_create_empty_node): Break out from ...
14912 (varpool_node_for_decl): ... here.
14913 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
14915 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
14917 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
14918 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
14919 part. Use straight-line flow at the end.
14920 <COMPONENT_REF>: Remove superfluous else.
14921 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
14923 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14926 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
14927 target hook even for !TREE_PUBLIC decls. If no resolution info
14928 is available, return false for common and external decls.
14930 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14932 * config/rl78/constraints.md (U): New constraint.
14933 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
14936 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
14939 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
14940 to allow returning address to AT_PLATFORM name.
14942 2013-06-11 Jan Hubicka <jh@suse.cz>
14944 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
14945 * cgraph.h (symtab_node_base): Add weakref flag.
14946 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
14947 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
14948 (output_weakrefs): Use weakref flag.
14949 * fold-const.c (simple_operand_p): Handle WEAK.
14950 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
14951 * ipa.c (varpool_externally_visible_p): Drop weakref.
14952 (function_and_variable_visibility): Update comment; fix weakref
14953 sanity checks; do not clear DECL_WEAK on them.
14954 * lto-cgraph.c (lto_output_node): update.
14955 (lto_output_varpool_node): Update.
14956 (input_overwrite_node): Update.
14957 (input_node): Update.
14958 (input_varpool_node): Update.
14959 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
14960 (lto_symtab_merge_symbols): Add sanity check.
14961 (lto_symtab_prevailing_decl): Do not special case weakrefs.
14962 * passes.c (rest_of_decl_compilation): Set static flag, too.
14963 * symtab.c (dump_symtab_base): Dump weakref.
14964 (verify_symtab_base): Sanity check weakrefs.
14965 (symtab_make_decl_local): Remove duplicated code.
14966 (symtab_alias_ultimate_target): Simplify.
14967 * varpool.c (varpool_create_variable_alias): Set weakref flag.
14969 2013-06-11 Tom de Vries <tom@codesourcery.com>
14971 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
14972 sequence_vect == NULL.
14974 2013-06-11 DJ Delorie <dj@redhat.com>
14976 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14977 (rl78_unwind_word_mode): New.
14979 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14981 * final.c (debug_prefix_maps): Make static.
14983 2013-06-11 David Malcolm <dmalcolm@redhat.com>
14985 * function.c (initial_trampoline): Remove stray copy.
14987 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14989 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14991 2013-06-11 Martin Jambor <mjambor@suse.cz>
14993 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14994 within bounds at the beginning of the function.
14996 2013-06-11 Alan Modra <amodra@gmail.com>
14998 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15000 (get_named_section): Don't NULL !DECL_P decl.
15002 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15004 * doc/invoke.texi (core-avx2): Document.
15006 (atom): Updated with MOVBE.
15008 2013-06-11 Richard Biener <rguenther@suse.de>
15010 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15012 2013-06-11 Anton Blanchard <anton@samba.org>
15014 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15015 correct shift value in little-endian mode.
15017 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15020 * varasm.c (get_variable_align): Move #endif to the right place.
15022 2013-06-10 Cary Coutant <ccoutant@google.com>
15024 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15025 for hash so that hash table traversal order is deterministic.
15027 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15028 Pat Haugen <pthaugen@us.ibm.com>
15029 Peter Bergner <bergner@vnet.ibm.com>
15031 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15032 of vectors in GPRS if they are direct moves or quad word load or
15035 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15037 (direct_move_p): Likewise.
15038 (quad_load_store_p): Likewise.
15040 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15041 classes into bins based on the physical register type.
15042 (reg_class_to_reg_type): Likewise.
15043 (IS_STD_REG_TYPE): Likewise.
15044 (IS_FP_VECT_REG_TYPE): Likewise.
15045 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15046 use direct move instructions.
15047 (reload_gpr_vsx): Likewise.
15048 (reload_vsx_gpr): Likewise.
15049 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15050 information that is a simplification of register classes. Also
15051 precalculate direct move reload helpers.
15052 (direct_move_p): New function to return true if the operation can
15053 be done as a direct move instruciton.
15054 (quad_load_store_p): New function to return true if the operation
15055 is a quad memory operation.
15056 (rs6000_legitimize_address): If quad memory, only allow register
15057 indirect for TImode addresses.
15058 (rs6000_legitimate_address_p): Likewise.
15059 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15060 (rs6000_reload_register_type): Likewise.
15061 (register_to_reg_type): Return register type.
15062 (rs6000_secondary_reload_simple_move): New helper function for
15063 secondary reload and secondary memory needed to identify anything
15064 that is a simple move, and does not need reloading.
15065 (rs6000_secondary_reload_direct_move): New helper function for
15066 secondary reload to identify cases that can be done with several
15067 instructions via the direct move instructions.
15068 (rs6000_secondary_reload_move): New helper function for secondary
15069 reload to identify moves between register types that can be done.
15070 (rs6000_secondary_reload): Add support for quad memory operations
15071 and for direct move.
15072 (rs6000_secondary_memory_needed): Likewise.
15073 (rs6000_debug_secondary_memory_needed): Change argument names.
15074 (rs6000_output_move_128bit): New function to return the move to
15075 use for 128-bit moves, including knowing about the various
15076 limitations of quad memory operations.
15078 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15079 memory operations. call rs6000_output_move_128bit for the actual
15080 instruciton(s) to generate.
15081 (vsx_movti_64bit): Likewise.
15083 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15084 (UNSPEC_P8V_MTVSRWZ): Likewise.
15085 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15086 (UNSPEC_P8V_MTVSRD): Likewise.
15087 (UNSPEC_P8V_XXPERMDI): Likewise.
15088 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15089 (UNSPEC_FUSION_GPR): Likewise.
15090 (FMOVE128_GPR): New iterator for direct move.
15091 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15092 (f32_sv): Likewise.
15093 (f32_dm): Likewise.
15094 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15095 loads and direct move instructions.
15096 (zero_extendsidi2_lfiwzx): Likewise.
15097 (extendsidi2_lfiwax): Likewise.
15098 (extendsidi2_nocell): Likewise.
15099 (floatsi<mode>2_lfiwax): Likewise.
15100 (lfiwax): Likewise.
15101 (floatunssi<mode>2_lfiwzx): Likewise.
15102 (lfiwzx): Likewise.
15103 (fix_trunc<mode>_stfiwx): Likewise.
15104 (fixuns_trunc<mode>_stfiwx): Likewise.
15105 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15106 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15107 (parity<mode>2_cmpb): Set length/type attr.
15108 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15109 for 'mr.' to fast_compare.
15110 (bpermd_<mode>): Change type attr to popcnt.
15111 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15112 (p8_mtvsrwz_1): Likewise.
15113 (p8_mtvsrwz_2): Likewise.
15114 (reload_fpr_from_gpr<mode>): Likewise.
15115 (p8_mtvsrd_1): Likewise.
15116 (p8_mtvsrd_2): Likewise.
15117 (p8_xxpermdi_<mode>): Likewise.
15118 (reload_vsx_from_gpr<mode>): Likewise.
15119 (reload_vsx_from_gprsf): Likewise.
15120 (p8_mfvsrd_3_<mode>): LIkewise.
15121 (reload_gpr_from_vsx<mode>): Likewise.
15122 (reload_gpr_from_vsxsf): Likewise.
15123 (p8_mfvsrd_4_disf): Likewise.
15124 (multi-word GPR splits): Do not split direct moves or quad memory
15127 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15129 * tree-into-ssa.c (interesting_blocks): Make static.
15131 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15134 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15135 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15136 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15137 (get_variable_align): New function.
15138 (get_variable_section, emit_bss, emit_common,
15139 assemble_variable_contents, place_block_symbol): Use
15140 get_variable_align instead of DECL_ALIGN.
15141 (assemble_noswitch_variable): Add align argument, use it
15142 instead of DECL_ALIGN.
15143 (assemble_variable): Adjust caller. Use get_variable_align
15144 instead of DECL_ALIGN.
15145 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15147 (DATA_ABI_ALIGNMENT): Define.
15148 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15149 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15150 opt is false, only return the psABI mandated alignment increase.
15151 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15152 (DATA_ABI_ALIGNMENT): ... this.
15153 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15154 (DATA_ABI_ALIGNMENT): ... this.
15155 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15156 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15157 (DATA_ABI_ALIGNMENT): ... this.
15158 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15159 * doc/tm.texi: Regenerated.
15161 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15163 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15164 cmp_code to construct REG_EQUAL note.
15166 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15169 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15170 that operands[2] doesn't overlap with operands[0].
15172 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15173 Jan Hubicka <jh@suse.cz>
15175 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15176 hack to mark symbols as used.
15178 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15180 PR rtl-optimization/57559
15181 * lra-constraints.c (process_alt_operands): Don't discourage
15182 memory with known offset for offsetable memory constraint.
15183 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15185 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15187 * varasm.c (struct oc_local_state): Reorder fields.
15188 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15189 and adjust accordingly.
15190 (output_constructor): Reorder initialization code and adjust call to
15191 output_constructor_bitfield.
15193 2013-06-07 Jan Hubicka <jh@suse.cz>
15195 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15197 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15199 * tree-object-size.c (unknown): Make const.
15201 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15203 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15204 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15205 for last alternative in the cpu_facility attribute.
15207 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15210 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15211 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15212 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15213 * config/arm/constraints.md (Dg): New constraint.
15214 * config/arm/neon.md (xordi3_neon): Remove.
15215 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15216 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15218 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15220 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15221 Clean up alternatives.
15223 2013-06-07 Alan Modra <amodra@gmail.com>
15225 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15228 2013-06-07 Alan Modra <amodra@gmail.com>
15230 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15232 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15234 * config/arm/constraints.md (Df): New constraint.
15235 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15236 Correct length attribute for last two alternatives.
15238 2013-06-07 Alan Modra <amodra@gmail.com>
15240 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15241 override user -mfp-in-toc.
15242 (offsettable_ok_by_alignment): Consider just the current access
15243 rather than the whole object, unless BLKmode. Handle
15244 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15245 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15246 for -mcmodel=medium.
15247 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15248 override user -mfp-in-toc or -msum-in-toc. Default to
15249 -mno-fp-in-toc for -mcmodel=medium.
15251 2013-06-06 DJ Delorie <dj@redhat.com>
15253 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15254 TARGET_VALID_POINTER_MODE.
15256 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15257 Pat Haugen <pthaugen@us.ibm.com>
15258 Peter Bergner <bergner@vnet.ibm.com>
15260 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15261 Document new power8 builtins.
15263 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15264 condition code register, to allow 128-bit logical operations to be
15265 done in the VSX or GPR registers.
15266 (nor<mode>3): Use the canonical form for nor.
15267 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15268 vclz*, and vpopcnt* vector instructions.
15269 (nand<mode>3): Likewise.
15270 (orc<mode>3): Likewise.
15271 (clz<mode>2): LIkewise.
15272 (popcount<mode>2): Likewise.
15274 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15275 that only the GPRs are recognized.
15277 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15278 support for new power8 builtins.
15280 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15282 (xscvdpspn): Likewise.
15288 (vpopcnt): Likewise.
15289 (vpopcntb): Likewise.
15290 (vpopcnth): Likewise.
15291 (vpopcntw): Likewise.
15292 (vpopcntd): Likewise.
15294 (vmrgew): Likewise.
15295 (vmrgow): Likewise.
15297 (eqv_v16qi3): Likewise.
15298 (eqv_v8hi3): Likewise.
15299 (eqv_v4si3): Likewise.
15300 (eqv_v2di3): Likewise.
15301 (eqv_v4sf3): Likewise.
15302 (eqv_v2df3): Likewise.
15304 (nand_v16qi3): Likewise.
15305 (nand_v8hi3): Likewise.
15306 (nand_v4si3): Likewise.
15307 (nand_v2di3): Likewise.
15308 (nand_v4sf3): Likewise.
15309 (nand_v2df3): Likewise.
15311 (orc_v16qi3): Likewise.
15312 (orc_v8hi3): Likewise.
15313 (orc_v4si3): Likewise.
15314 (orc_v2di3): Likewise.
15315 (orc_v4sf3): Likewise.
15316 (orc_v2df3): Likewise.
15318 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15319 allow power8 quad mode in 64-bit.
15320 (rs6000_builtin_vectorized_function): Add support to vectorize
15321 ISA 2.07 count leading zeros, population count builtins.
15322 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15323 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15324 (builtin_function_type): Add vgbbd builtin function which takes an
15326 (altivec_expand_vec_perm_const): Add support for new power8 merge
15329 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15330 that does not include TImdoe for use with 32-bit.
15331 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15333 (UNSPEC_VSX_CVDPSPN): Likewise.
15334 (vsx_xscvdpspn): Likewise.
15335 (vsx_xscvspdpn): Likewise.
15336 (vsx_xscvdpspn_scalar): Likewise.
15337 (vsx_xscvspdpn_directmove): Likewise.
15338 (vsx_and<mode>3): Split logical operations into 32-bit and
15339 64-bit. Add support to do logical operations on TImode as well as
15340 VSX vector types. Allow logical operations to be done in either
15341 VSX registers or in general purpose registers in 64-bit mode. Add
15342 splitters if GPRs were used. For AND, add clobber of CCmode to
15343 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15345 (vsx_and<mode>3_32bit): Likewise.
15346 (vsx_and<mode>3_64bit): Likewise.
15347 (vsx_ior<mode>3): Likewise.
15348 (vsx_ior<mode>3_32bit): Likewise.
15349 (vsx_ior<mode>3_64bit): Likewise.
15350 (vsx_xor<mode>3): Likewise.
15351 (vsx_xor<mode>3_32bit): Likewise.
15352 (vsx_xor<mode>3_64bit): Likewise.
15353 (vsx_one_cmpl<mode>2): Likewise.
15354 (vsx_one_cmpl<mode>2_32bit): Likewise.
15355 (vsx_one_cmpl<mode>2_64bit): Likewise.
15356 (vsx_nor<mode>3): Likewise.
15357 (vsx_nor<mode>3_32bit): Likewise.
15358 (vsx_nor<mode>3_64bit): Likewise.
15359 (vsx_andc<mode>3): Likewise.
15360 (vsx_andc<mode>3_32bit): Likewise.
15361 (vsx_andc<mode>3_64bit): Likewise.
15362 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15363 and xxlorc instructions.
15364 (vsx_eqv<mode>3_64bit): Likewise.
15365 (vsx_nand<mode>3_32bit): Likewise.
15366 (vsx_nand<mode>3_64bit): Likewise.
15367 (vsx_orc<mode>3_32bit): Likewise.
15368 (vsx_orc<mode>3_64bit): Likewise.
15370 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15372 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15374 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15375 (p8_vmrgow): Likewise.
15376 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15377 GPRs to be split under VSX.
15378 (p8v_clz<mode>2): Add power8 count leading zero support.
15379 (p8v_popcount<mode>2): Add power8 population count support.
15380 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15383 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15386 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15388 (vec_nand): Likewise.
15389 (vec_vclz): Likewise.
15390 (vec_vclzb): Likewise.
15391 (vec_vclzd): Likewise.
15392 (vec_vclzh): Likewise.
15393 (vec_vclzw): Likewise.
15394 (vec_vgbbd): Likewise.
15395 (vec_vmrgew): Likewise.
15396 (vec_vmrgow): Likewise.
15397 (vec_vpopcnt): Likewise.
15398 (vec_vpopcntb): Likewise.
15399 (vec_vpopcntd): Likewise.
15400 (vec_vpopcnth): Likewise.
15401 (vec_vpopcntw): Likewise.
15403 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15405 PR rtl-optimization/57468
15406 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15409 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15411 PR rtl-optimization/57459
15412 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15413 type when setting live regs.
15415 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15417 * config/s390/s390.opt (mlra): New option.
15418 * config/s390/s390.c (s390_decompose_address): Check displacement
15419 for all registers for LRA.
15420 (s390_secondary_reload): Don't used secondary reloads for LRA.
15421 (s390_lra_p): New function.
15422 (TARGET_LRA_P): Define.
15423 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15424 of attribute cpu_facility to zarch for the last alternative.
15425 (*cmpmem_short): Ditto.
15427 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15429 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15430 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15431 (arm_expand_prologue): Likewise.
15433 2013-06-06 Teresa Johnson <tejohnson@google.com>
15436 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15437 as this is now done by redirect_edge_and_branch_force.
15438 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15439 barriers, and fix interaction with splitting.
15440 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15441 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15442 reflect changes made in the routine.
15443 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15444 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15445 since this is called in cfglayout mode, and replace partition fixup
15446 with assert as that is now done by force_nonfallthru_and_redirect.
15447 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15448 already be marked with region crossing note.
15449 (insert_section_boundary_note): Make non-static, gate on flag
15450 has_bb_partition, rewrite to also check for multiple partitions.
15451 (rest_of_handle_reorder_blocks): Remove call to
15452 insert_section_boundary_note, now done later during free_cfg.
15453 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15454 * bb-reorder.h (insert_section_boundary_note): Declare.
15455 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15456 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15457 invoke insert_section_boundary_note.
15458 (try_redirect_by_replacing_jump): Remove unnecessary
15459 check for region crossing note.
15460 (fixup_partition_crossing): New function.
15461 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15462 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15463 in non-cfglayout mode.
15464 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15465 remove old code that tried to do this. Emit barrier correctly
15466 when we are in cfglayout mode.
15467 (last_bb_in_partition): New function.
15468 (rtl_split_edge): Correctly fixup partition boundaries.
15469 (commit_one_edge_insertion): Remove old code that tried to
15470 fixup region crossing edge since this is now handled in
15471 split_block, and set up insertion point correctly since
15472 block may now end in a jump.
15473 (verify_hot_cold_block_grouping): Guard against checking when not in
15474 linearized RTL mode.
15475 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15477 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15478 rtl_verify_flow_info, so not called in cfglayout mode.
15479 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15480 (fixup_reorder_chain): Remove old code that attempted to fixup region
15481 crossing note as this is now handled in force_nonfallthru_and_redirect.
15482 (duplicate_insn_chain): Don't duplicate switch section notes.
15483 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15485 * basic-block.h (emit_barrier_after_bb): Declare.
15487 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15489 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15490 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15491 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15493 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15495 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15497 * config/arm/ldmstm.md: Regenerate.
15499 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15501 * config/arm/sync.md (atomic_loaddi_1):
15502 Disable predication for arm_restrict_it.
15503 (arm_load_exclusive<mode>): Likewise.
15504 (arm_load_exclusivesi): Likewise.
15505 (arm_load_exclusivedi): Likewise.
15506 (arm_load_acquire_exclusive<mode>): Likewise.
15507 (arm_load_acquire_exclusivesi): Likewise.
15508 (arm_load_acquire_exclusivedi): Likewise.
15509 (arm_store_exclusive<mode>): Likewise.
15510 (arm_store_exclusive<mode>): Likewise.
15511 (arm_store_release_exclusivedi): Likewise.
15512 (arm_store_release_exclusive<mode>): Likewise.
15514 2013-06-06 Richard Biener <rguenther@suse.de>
15516 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15518 (lto_tag_is_tree_code_p): Adjust.
15519 (lto_tag_is_gimple_code_p): Likewise.
15520 (lto_gimple_code_to_tag): Likewise.
15521 (lto_tag_to_gimple_code): Likewise.
15522 (lto_tree_code_to_tag): Likewise.
15523 (lto_tag_to_tree_code): Likewise.
15524 * data-streamer.h (streamer_write_hwi_in_range): Use
15525 uhwi streaming to stream the normalized range.
15526 (streamer_read_hwi_in_range): Likewise.
15528 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15530 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15531 (predicable_short_it): Likewise.
15532 (predicated): Likewise.
15533 (enabled): Handle above.
15534 (define_cond_exec): Set predicated attribute to yes.
15536 2013-06-05 Mike Stump <mikestump@comcast.net>
15538 * gdbinit.in (__FUNCTION__): Add.
15540 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15542 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15543 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15545 2013-06-05 Jan Hubicka <jh@suse.cz>
15547 * varasm.c (mark_decl_referenced): Revert the removal until targets
15550 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15552 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15553 instead of mark_decl_referenced.
15555 2013-06-05 Jan Hubicka <jh@suse.cz>
15557 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15558 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15559 and symtab_used_from_object_file_p.
15560 (cgraph_make_node_local_1): Clear forced_by_abi.
15561 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15562 * cgraph.h (symtab_node_base): Add forced_by_abi.
15563 (decide_is_variable_needed): Remove.
15564 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15565 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15566 (decide_is_symbol_needed): ... this one; handle symbols in general;
15567 always analyze virtuals; honnor forced_by_abi.
15568 (cgraph_finalize_function): Update.
15569 (varpool_finalize_decl): Update.
15570 (symbol_defined_and_needed): Remove.
15571 (analyze_functions): Update.
15572 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15573 output_refs, input_overwrite_node): Handle forced_by_abi.
15574 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15575 (address_taken_from_non_vtable_p): ... this one.
15576 (comdat_can_be_unshared_p_1): New function.
15577 (cgraph_comdat_can_be_unshared_p): Rename to ...
15578 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15579 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15580 (function_and_variable_visibility): Clear forced_by_abi as needed.
15581 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15582 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15583 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15584 * varpool.c (decide_is_variable_needed): Remove.
15586 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15588 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15589 (arm_option_override): Override arm_restrict_it where appropriate.
15590 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15591 * config/arm/arm.opt (mrestrict-it): New command-line option.
15592 * doc/invoke.texi: Document -mrestrict-it.
15594 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15596 * tsan.c (tsan_atomic_table): Make const.
15598 2013-06-05 Richard Biener <rguenther@suse.de>
15600 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15601 index associated with the tree we are supposed to replace.
15602 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15603 TREE_ASM_WRITTEN as zero for everything but SSA names.
15605 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15607 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15609 2013-06-04 Jan Hubicka <jh@suse.cz>
15611 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15612 (input_node, input_varpool_node): Handle correctly external same
15614 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15615 nodes at ltrans stage.
15617 2013-06-04 Jan Hubicka <jh@suse.cz>
15619 * ipa-inline.c (update_caller_keys): Fix availability test.
15620 (update_callee_keys): Likewise.
15621 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15622 to follow ELF standard.
15624 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15626 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15627 (mips64r5900el-*-elf*): New configurations.
15628 * config/mips/mips-cpus.def (r5900): New processor.
15629 * config/mips/mips-tables.opt: Regenerate.
15630 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15631 (mips_issue_rate): Handle PROCESSOR_R5900.
15632 (mips_reorg_process_insns): Force reorder mode for the R5900.
15633 * config/mips/mips.h (TARGET_MIPS5900): Define.
15634 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15636 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15638 * config/mips/mips.md (processor): Add r5900.
15639 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15641 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15643 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15644 into function to generate MOVI instruction.
15645 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15646 (aarch64_preferred_simd_mode): Turn into wrapper.
15647 (aarch64_output_scalar_simd_mov_immediate): New function.
15648 * config/aarch64/aarch64-protos.h: Add prototype for above.
15650 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15652 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15653 element_char member.
15654 (sizetochar): Return signed char.
15655 (aarch64_simd_valid_immediate): Remove elchar and other
15656 unnecessary variables.
15657 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15658 Calculate element_char as required.
15659 * config/aarch64/aarch64-protos.h: Update and move prototype
15660 for aarch64_output_simd_mov_immediate.
15661 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15664 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15666 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15667 information completed by aarch64_simd_valid_immediate.
15668 (aarch64_legitimate_constant_p): Update arguments.
15669 (aarch64_simd_valid_immediate): Work with struct rather than many
15671 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15672 (aarch64_simd_make_constant): Update arguments.
15673 (aarch64_output_simd_mov_immediate): Work with struct rather than
15674 many pointers. Output immediate directly rather than as operand.
15675 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15677 * config/aarch64/constraints.md (Dn): Update arguments.
15679 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15681 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15683 (aarch64_simd_immediate_valid_for_move): Remove.
15684 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15685 (aarch64_simd_make_constant): Update call.
15686 (aarch64_output_simd_mov_immediate): Update call.
15687 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15689 * config/aarch64/constraints.md (Dn): Update call.
15691 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15693 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15694 return type to bool for prototype.
15695 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15696 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15697 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15699 2013-06-04 Catherine Moore <clm@codesourcery.com>
15701 * config/mips/mips.opt (meva): New.
15702 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15703 (ASM_SPEC): Handle -meva.
15704 * doc/invoke.texi (meva): Document.
15706 2013-06-04 Alan Modra <amodra@gmail.com>
15708 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15711 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15713 * rtl.def: Add extra fourth optional field to define_cond_exec.
15714 * gensupport.c (process_one_cond_exec): Process attributes from
15716 * doc/md.texi: Document fourth field in define_cond_exec.
15718 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15720 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15721 out the processing order as in store_bit_field_1.
15723 2013-06-04 Jan Hubicka <jh@suse.cz>
15725 PR middle-end/57500
15726 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15727 non-VAR_DECL node if it does not exist yet.
15729 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15731 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15732 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15733 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15734 target_cpu_default setting.
15736 2013-06-03 Teresa Johnson <tejohnson@google.com>
15738 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15739 default to -fopt-info=optimized instead of all.
15740 * doc/invoke.texi: Ditto.
15741 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15742 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15743 (execute_vect_slp): Emit BB vectorization success under
15744 MSG_OPTIMIZED_LOCATIONS.
15745 * tree-vect-slp.c (vect_slp_transform_bb): Change
15746 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15747 * tree-vect-loop.c (vect_transform_loop): Ditto.
15749 2013-06-03 Jason Merrill <jason@redhat.com>
15752 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15753 Use TARGET_EXPR for C++.
15755 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15757 PR rtl-optimization/57268
15758 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15759 if DEBUG_INSN_P (insn).
15762 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15764 PR rtl-optimization/57268
15765 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15766 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15768 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15770 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15771 (ix86_avoid_lea_for_addr): Likewise.
15772 (exact_dependency_1): Likewise.
15773 (ix86_adjust_cost): Likewise.
15774 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15776 (do_reorder_for_imul): Fix typo, formatting and
15777 !reload_completed check removed.
15778 (ix86_sched_reorder): Fix typo and formatting.
15779 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15782 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15784 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15786 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15788 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15789 <STRING_CST>: Likewise.
15790 <VECTOR_CST>: Likewise.
15792 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15793 Mikael Morin <mikael@gcc.gnu.org>
15795 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15796 * config.in: Regenerated.
15797 * configure: Regenerated.
15799 2013-06-01 Jan Hubicka <jh@suse.cz>
15801 PR middle-end/57366
15802 * cgraphunit.c (compile): When weakref is not supported,
15803 set up transparent aliases before final output pass.
15804 * varasm.c (assemble_alias): Do not try to do it here.
15806 2013-06-01 Jan Hubicka <jh@suse.cz>
15808 PR middle-end/57467
15809 * passes.c (for_per_function): Skip unanalyzed functions.
15811 2013-06-01 Jan Hubicka <jh@suse.cz>
15813 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15814 (lto_symtab_merge_symbols_1): ... this one.
15815 (lto_symtab_merge_cgraph_nodes): Rename to ...
15816 (lto_symtab_merge_symbols): ... this one; simplify.
15817 * cgraph.c (same_body_aliases_done): Rename to ...
15818 (cpp_implicit_aliases_done): ... this one.
15819 (cgraph_create_function_alias): Update.
15820 (cgraph_same_body_alias): Update.
15821 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15822 (verify_edge_corresponds_to_fndecl): Simplify.
15823 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15824 (cgraph_node): Remove same_body_alias.
15825 (varpool_node): Remove alias_of and extra_name_alias.
15826 (same_body_aliases_done): Rename to ..
15827 (cpp_implicit_aliases_done): ... this one.
15828 (symtab_alias_ultimate_target): Add default parameter.
15829 (symtab_resolve_alias): New function.
15830 (fixup_same_cpp_alias_visibility): Declare.
15831 (cgraph_function_node): Add default parameter.
15832 (cgraph_node_asm_name): Likewise.
15833 (cgraph_function_or_thunk_node): Add default parameter; do
15834 not ICE when it is NULL.
15835 (varpool_variable_node): Likewise.
15836 * tree-emutls.c (create_emultls_var): Update.
15837 (ipa_lower_emutls): Update.
15838 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15839 (cgraph_reset_node): Reset alias info.
15840 (cgraph_finalize_function): Update.
15841 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15842 (analyze_function): Simplify.
15843 (cgraph_process_same_body_aliases): Simplify.
15844 (analyze_functions): Fixup same body aliases.
15845 (handle_alias_pairs): Simplify.
15846 (assemble_thunk): Update.
15847 (assemble_thunks_and_aliases): Update.
15848 (output_weakrefs): Rewrite.
15849 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15850 (lto_output_varpool_node): Likewise.
15851 (compute_ltrans_boundary): Remve assert.
15852 (get_alias_symbol): New functoin.
15853 (input_node): Rewrite alias handling.
15854 (input_varpool_node): Likewise.
15855 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15856 * ipa.c (process_references): Handle weakrefs correctly.
15857 (symtab_remove_unreachable_nodes): Likewise.
15858 * trans-mem.c (get_cg_data): Update.
15859 (ipa_tm_create_version_alias): Update.
15860 (ipa_tm_execute): Update.
15861 * symtab.c (dump_symtab_base): Dump aliases.
15862 (verify_symtab_base): Verify aliases.
15863 (symtab_node_availability): New function.
15864 (symtab_alias_ultimate_target): Simplify.
15865 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15866 handle all the fixup cases.
15867 (symtab_resolve_alias): New function.
15868 * passes.c (ipa_write_summaries): Handle weakrefs.
15869 * varpool.c (varpool_analyze_node): Simplify.
15870 (assemble_aliases): Update.
15871 (varpool_create_variable_alias): Simplify.
15872 (varpool_extra_name_alias): Simplify.
15873 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15874 (lto_symtab_merge_symbols): ... this one.
15876 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15879 PR rtl-optimization/57268
15880 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15881 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15883 2013-06-01 Tobias Burnus <burnus@net-b.de>
15885 Partially reverted:
15886 2013-05-31 Tobias Burnus <burnus@net-b.de>
15888 PR middle-end/57073
15889 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15892 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15894 PR rtl-optimization/57268
15895 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15896 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15898 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15900 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
15901 unordered comparison operators when -fno-trapping-math is in effect
15903 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
15904 and implement unordered comparison operators properly on the e500.
15906 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
15908 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
15909 for constant scalar integers.
15910 (simplify_relational_operation_1): Likewise.
15912 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
15914 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
15915 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
15918 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15919 Igor Zamyatin <igor.zamyatin@intel.com>
15921 Silvermont (SLM) architecture performance tuning.
15922 * config/i386/i386.h (enum ix86_tune_indices): Add
15923 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
15924 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
15926 * config/i386/i386.c (initial_ix86_tune_features)
15927 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
15928 (ix86_lea_outperforms): Handle Silvermont tuning.
15929 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
15931 (ix86_use_lea_for_mov): Likewise.
15932 (ix86_avoid_lea_for_addr): Likewise.
15933 (ix86_lea_for_add_ok): Likewise.
15934 (exact_dependency_1): New function.
15935 (exact_store_load_dependency): Likewise.
15936 (ix86_adjust_cost): Handle Silvermont tuning.
15937 (do_reoder_for_imul): Likewise.
15938 (swap_top_of_ready_list): New function.
15939 (ix86_sched_reorder): Changed to handle Silvermont tuning.
15941 * config/i386/i386.md (peepholes that split memory operand in fp
15944 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15946 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
15947 Remove un-necessary braces.
15949 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
15951 * config/aarch64/aarch64.c (aarch64_classify_symbol):
15952 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
15954 2013-05-31 Tobias Burnus <burnus@net-b.de>
15956 PR middle-end/57073
15957 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
15959 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15962 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
15963 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
15964 * config/arm/neon.md (iordi3_neon): Remove.
15965 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
15966 * config/arm/predicates.md (imm_for_neon_logic_operand):
15967 Move to earlier in the file.
15968 (neon_logic_op2): Likewise.
15969 (arm_iordi_operand_neon): New predicate.
15971 2013-05-31 Richard Biener <rguenther@suse.de>
15973 PR tree-optimization/57478
15974 PR tree-optimization/57453
15975 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15978 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15980 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15981 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15983 2013-05-30 Tobias Burnus <burnus@net-b.de>
15984 Thomas Koenig <tkoenig@gcc.gnu.org>
15986 PR middle-end/57073
15987 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15988 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15990 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15992 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
15994 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15996 * target.def (register_usage_leveling_p): New hook.
15997 * targhooks.c (default_register_usage_leveling_p): New.
15998 * targhooks.h (default_register_usage_leveling_p): New prototype.
15999 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16000 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16001 * doc/tm.texi: Update.
16002 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16004 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16006 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16007 (*insv_reg<mode>): New define_insn.
16009 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16011 PR rtl-optimization/57439
16012 * postreload.c (move2add_valid_value_p): Check that we have
16013 a zero subreg_regno_offset when accessing the register in
16014 the requested mode.
16016 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16017 Igor Zamyatin <igor.zamyatin@intel.com>
16019 Silvermont (SLM) architecture pipeline model, tuning and
16021 * config.gcc: Add slm config options and target.
16023 * config/i386/slm.md: New.
16025 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16027 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16029 (ix86_target_macros_internal): Likewise.
16031 * config/i386/i386.c (slm_cost): New cost.
16032 (m_SLM): New macro flag.
16033 (initial_ix86_tune_features): Set m_SLM.
16034 (x86_accumulate_outgoing_args): Likewise.
16035 (x86_arch_always_fancy_math_387): Likewise.
16036 (processor_target_table): Add slm cost.
16037 (cpu_names): Add slm cpu name.
16038 (x86_option_override_internal): Set SLM ISA.
16039 (ix86_issue_rate): New case PROCESSOR_SLM.
16040 (ia32_multipass_dfa_lookahead): Likewise.
16041 (fold_builtin_cpu): Add slm.
16043 * config/i386/i386.h (TARGET_SLM): New target macro.
16044 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16045 (processor_type): Add PROCESSOR_SLM.
16047 * config/i386/i386.md (cpu): Add new value "slm".
16048 (slm.md): Include slm.md.
16050 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16051 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16053 * config/arm/arm-protos.h: Add and update function protos.
16054 * config/arm/arm.c (use_simple_return_p): New added.
16055 (thumb2_expand_return): Check simple_return flag.
16056 * config/arm/arm.md: Add simple_return and conditional simple_return.
16057 * config/arm/iterators.md: Add iterator for return and simple_return.
16059 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16061 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16062 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16063 (arm_emit_vfp_multi_reg_pop): Likewise.
16064 (thumb2_emit_ldrd_pop): Likewise.
16065 (arm_expand_epilogue): Add misc REG_CFA notes.
16066 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16068 2013-05-29 Lawrence Crowl <crowl@google.com>
16070 * config/arm/t-arm: Update for below.
16072 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16073 Change type to hash_table. Update dependent calls and types.
16075 * config/i386/t-cygming: Update for below.
16077 * config/i386/t-interix: Update for below.
16079 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16080 Change type to hash_table. Update dependent calls and types.
16081 (i386_find_on_wrapper_list::wrappers): Likewise.
16083 * config/ia64/t-ia64: Update for below.
16085 * config/ia64/ia64.c (bundle_state_table):
16086 Change type to hash_table. Update dependent calls and types.
16088 * config/mips/mips.c (mips_reorg_process_insns::htab):
16089 Change type to hash_table. Update dependent calls and types.
16091 * config/sol2.c (solaris_comdat_htab):
16092 Change type to hash_table. Update dependent calls and types.
16094 * config/t-sol2: Update for above.
16096 2013-05-29 Teresa Johnson <tejohnson@google.com>
16098 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16099 functions are not yet marked as defined.
16101 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16102 Pat Haugen <pthaugen@us.ibm.com>
16103 Peter Bergner <bergner@vnet.ibm.com>
16105 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16109 (vrotl<mode>3): Likewise.
16110 (vashl<mode>3): Likewise.
16111 (vlshr<mode>3): Likewise.
16112 (vashr<mode>3): Likewise.
16114 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16115 support for power8 V2DI builtins.
16117 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16118 power8 V2DI builtins.
16119 (vupkhsw): Likewise.
16120 (vupklsw): Likewise.
16121 (vaddudm): Likewise.
16122 (vminsd): Likewise.
16123 (vmaxsd): Likewise.
16124 (vminud): Likewise.
16125 (vmaxud): Likewise.
16126 (vpkudum): Likewise.
16127 (vpksdss): Likewise.
16128 (vpkudus): Likewise.
16129 (vpksdus): Likewise.
16134 (vsubudm): Likewise.
16135 (vcmpequd): Likewise.
16136 (vcmpgtsd): Likewise.
16137 (vcmpgtud): Likewise.
16138 (vcmpequd_p): Likewise.
16139 (vcmpgtsd_p): Likewise.
16140 (vcmpgtud_p): Likewise.
16141 (vupkhsw): Likewise.
16142 (vupklsw): Likewise.
16143 (vaddudm): Likewise.
16144 (vmaxsd): Likewise.
16145 (vmaxud): Likewise.
16146 (vminsd): Likewise.
16147 (vminud): Likewise.
16148 (vpksdss): Likewise.
16149 (vpksdus): Likewise.
16150 (vpkudum): Likewise.
16151 (vpkudus): Likewise.
16156 (vsubudm): Likewise.
16158 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16159 support for power8 V2DI instructions.
16161 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16162 power8 V2DI instructions. Combine pack and unpack insns to use an
16163 iterator for each mode. Check whether a particular mode supports
16164 Altivec instructions instead of just checking TARGET_ALTIVEC.
16165 (UNSPEC_VPKUWUM): Likewise.
16166 (UNSPEC_VPKSHSS): Likewise.
16167 (UNSPEC_VPKSWSS): Likewise.
16168 (UNSPEC_VPKUHUS): Likewise.
16169 (UNSPEC_VPKSHUS): Likewise.
16170 (UNSPEC_VPKUWUS): Likewise.
16171 (UNSPEC_VPKSWUS): Likewise.
16172 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16173 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16174 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16175 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16176 (UNSPEC_VUPKHSB): Likewise.
16177 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16178 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16179 (UNSPEC_VUPKHSH): Likewise.
16180 (UNSPEC_VUPKLSB): Likewise.
16181 (UNSPEC_VUPKLSH): Likewise.
16183 (VI_char): Likewise.
16184 (VI_scalar): Likewise.
16185 (VI_unit): Likewise.
16187 (VP_small): Likewise.
16188 (VP_small_lc): Likewise.
16189 (VU_char): Likewise.
16190 (add<mode>3): Likewise.
16191 (altivec_vaddcuw): Likewise.
16192 (altivec_vaddu<VI_char>s): Likewise.
16193 (altivec_vadds<VI_char>s): Likewise.
16194 (sub<mode>3): Likewise.
16195 (altivec_vsubcuw): Likewise.
16196 (altivec_vsubu<VI_char>s): Likewise.
16197 (altivec_vsubs<VI_char>s): Likewise.
16198 (altivec_vavgs<VI_char>): Likewise.
16199 (altivec_vcmpbfp): Likewise.
16200 (altivec_eq<mode>): Likewise.
16201 (altivec_gt<mode>): Likewise.
16202 (altivec_gtu<mode>): Likewise.
16203 (umax<mode>3): Likewise.
16204 (smax<mode>3): Likewise.
16205 (umin<mode>3): Likewise.
16206 (smin<mode>3): Likewise.
16207 (altivec_vpkuhum): Likewise.
16208 (altivec_vpkuwum): Likewise.
16209 (altivec_vpkshss): Likewise.
16210 (altivec_vpkswss): Likewise.
16211 (altivec_vpkuhus): Likewise.
16212 (altivec_vpkshus): Likewise.
16213 (altivec_vpkuwus): Likewise.
16214 (altivec_vpkswus): Likewise.
16215 (altivec_vpks<VI_char>ss): Likewise.
16216 (altivec_vpks<VI_char>us): Likewise.
16217 (altivec_vpku<VI_char>us): Likewise.
16218 (altivec_vpku<VI_char>um): Likewise.
16219 (altivec_vrl<VI_char>): Likewise.
16220 (altivec_vsl<VI_char>): Likewise.
16221 (altivec_vsr<VI_char>): Likewise.
16222 (altivec_vsra<VI_char>): Likewise.
16223 (altivec_vsldoi_<mode>): Likewise.
16224 (altivec_vupkhsb): Likewise.
16225 (altivec_vupkhs<VU_char>): Likewise.
16226 (altivec_vupkls<VU_char>): Likewise.
16227 (altivec_vupkhsh): Likewise.
16228 (altivec_vupklsb): Likewise.
16229 (altivec_vupklsh): Likewise.
16230 (altivec_vcmpequ<VI_char>_p): Likewise.
16231 (altivec_vcmpgts<VI_char>_p): Likewise.
16232 (altivec_vcmpgtu<VI_char>_p): Likewise.
16233 (abs<mode>2): Likewise.
16234 (vec_unpacks_hi_v16qi): Likewise.
16235 (vec_unpacks_hi_v8hi): Likewise.
16236 (vec_unpacks_lo_v16qi): Likewise.
16237 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16238 (vec_unpacks_lo_v8hi): Likewise.
16239 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16240 (vec_pack_trunc_v8h): Likewise.
16241 (vec_pack_trunc_v4si): Likewise.
16242 (vec_pack_trunc_<mode>): Likewise.
16244 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16246 (vec_vmaxsd): Likewise.
16247 (vec_vmaxud): Likewise.
16248 (vec_vminsd): Likewise.
16249 (vec_vminud): Likewise.
16250 (vec_vpksdss): Likewise.
16251 (vec_vpksdus): Likewise.
16252 (vec_vpkudum): Likewise.
16253 (vec_vpkudus): Likewise.
16254 (vec_vrld): Likewise.
16255 (vec_vsld): Likewise.
16256 (vec_vsrad): Likewise.
16257 (vec_vsrd): Likewise.
16258 (vec_vsubudm): Likewise.
16259 (vec_vupkhsw): Likewise.
16260 (vec_vupklsw): Likewise.
16262 2013-05-29 Jan Hubicka <jh@suse.cz>
16264 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16265 flags; reorder rest of fields in more consistent way.
16266 (varpool_node): Remove analyzed, finalized and alias.
16267 (cgraph_ndoe): Likewise.
16268 (symtab_alias_ultimate_target): New function.
16269 (cgraph_function_node): Move offline.
16270 (cgraph_reset_node): Declare.
16271 (cgraph_comdat_can_be_unshared_p): Remove.
16272 (varpool_remove_initializer): Declare.
16273 (varpool_first_defined_variable, varpool_next_defined_variable
16274 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16275 (cgraph_function_with_gimple_body_p): Update.
16276 (varpool_all_refs_explicit_p): Update.
16277 (symtab_alias_target): New function.
16278 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16279 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16280 (cgraph_function_or_thunk_node): Simplify using
16281 symtab_alias_ultimate_target.
16282 (varpool_variable_node): Likewise.
16283 * cgraph.c (cgraph_create_function_alias): Update.
16284 (cgraph_add_thunk): Update.
16285 (cgraph_remove_node): Update.
16286 (dump_cgraph_node): Do not dump removed flags.
16287 (cgraph_function_body_availability): Update.
16288 (cgraph_propagate_frequency): Update.
16289 (verify_cgraph_node): Check sanity of local flag.
16290 (cgraph_function_node): Move here from cgraph.h; revamp for
16291 cgraph_function_or_thunk_node.
16292 * lto-symtab.c (lto_varpool_replace_node): Update.
16293 (lto_symtab_resolve_can_prevail_p): Update.
16294 (lto_symtab_merge_cgraph_nodes): Update.
16295 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16296 propagate_constants_accross_call, devirtualization_time_bonus,
16297 ipcp_propagate_stage): Update.
16298 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16299 * ipa-inline-transform.c (clone_inlined_nodes,
16300 preserve_function_body_p): Update.
16301 * ipa-reference.c (propagate): Update.
16302 (write_node_summary_p): Update.
16303 * toplev.c (wrapup_global_declaration_2): Update.
16304 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16305 (analyze_function) ... this one.
16306 (cgraph_process_new_functions): Update.
16307 (cgraph_reset_node): Export.
16308 (cgraph_finalize_function): Update.
16309 (cgraph_add_new_function): Update.
16310 (process_function_and_variable_attributes): Update.
16311 (varpool_finalize_decl): Update.
16312 (symbol_finalized): Remove.
16313 (symbol_finalized_and_needed): Rename to ...
16314 (symbol_defined_and_needed): ... update.
16315 (cgraph_analyze_functions): Update.
16316 (handle_alias_pairs): Update.
16317 (mark_functions_to_output): Update.
16318 (assemble_thunk): Update.
16319 (output_in_order): Update.
16320 (output_weakrefs): Update.
16321 (finalize_compilation_unit): Update.
16322 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16323 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16324 input_node, input_varpool_node): Update.
16325 * dbxout.c (dbxout_expand_expr): Update.
16326 * cgraphclones.c (cgraph_clone_node): Update.
16327 (cgraph_copy_node_for_versioning): Update.
16328 (cgraph_materialize_clone): Update.
16329 (cgraph_materialize_all_clones): Update.
16330 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16331 propagate_pure_const, propagate_nothrow): Update.
16332 * lto-streamer-out.c (lto_output, write_symbol): Update.
16333 * ipa-utils.c (ipa_reverse_postorder): Update.
16334 * ipa-inline.c (can_inline_edge_p): Update.
16335 (update_caller_keys, ipa_inline): Update.
16336 * dwarf2out.c (reference_to_unused,
16337 premark_types_used_by_global_vars_helper): Update.
16338 * tree-eh.c (tree_could_trap_p): Update.
16339 * ipa-split.c (consider_split, execute_split_functions): Update.
16340 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16341 has_addr_references_p): Update; move ahead in file for better
16343 (process_references): Simplify.
16344 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16345 bodies are removed.
16346 (cgraph_comdat_can_be_unshared_p): Make static.
16347 (cgraph_externally_visible_p): Update.
16348 (varpool_externally_visible_p): Update.
16349 (function_and_variable_visibility): Update.
16350 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16351 ipa_tm_mark_force_output_node): Update.
16352 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16353 estimate_edge_devirt_benefit, inline_generate_summary,
16354 inline_write_summary): Update.
16355 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16356 * ipa-prop.c (ipa_compute_jump_functions): Update.
16357 (ipa_print_node_params, ipa_prop_read_section,
16358 ipa_update_after_lto_read, read_replacements_section): Update.
16359 * varasm.c (mark_decl_referenced): Update.
16360 (assemble_alias, dump_tm_clone_pairs): Update.
16361 * tree-inline.c (copy_bb): Update.
16362 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16364 * symtab.c (dump_symtab_base): Print new flags.
16365 (verify_symtab_base): Verify new flags.
16366 (symtab_alias_ultimate_target): New function.
16367 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16368 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16370 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16372 * i386.c (ix86_get_function_versions_dispatcher,
16373 ix86_generate_version_dispatcher_body): Update.
16374 (fold_builtin_cpu): Use varpool_add_new_variable.
16375 * varpool.c (varpool_remove_initializer): Break out from ...
16376 (varpool_remove_node): ... this one.
16377 (dump_varpool_node, varpool_node_for_asm,
16378 cgraph_variable_initializer_availability, varpool_analyze_node,
16379 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16380 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16383 2013-05-29 Jan Hubicka <jh@suse.cz>
16385 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16387 2013-05-29 Easwaran Raman <eraman@google.com>
16389 PR tree-optimization/57442
16390 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16391 when control exits the main loop.
16393 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16395 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16397 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16398 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16399 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16402 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16404 PR tree-optimization/57441
16405 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16406 Don't limit size of incr_vec to number of candidates.
16408 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16410 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16411 and mips16 directories.
16412 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16413 (MULTILIB_DIRNAMES): Ditto.
16414 (MULTILIB_EXCEPTIONS): Add new exceptions.
16415 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16416 (MULTILIB_DIRNAMES): Ditto.
16417 (MULTILIB_EXCEPTIONS): Add new exceptions.
16419 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16420 Marcus Shawcroft <marcus.shawcroft@arm.com>
16422 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16423 SYMBOL_TINY_ABSOLUTE.
16424 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16425 SYMBOL_TINY_ABSOLUTE.
16426 (aarch64_expand_mov_immediate): Likewise.
16427 (aarch64_classify_symbol): Likewise.
16428 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16429 Permit SYMBOL_TINY_ABSOLUTE.
16430 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16432 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16433 Marcus Shawcroft <marcus.shawcroft@arm.com>
16435 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16436 Refactor if/switch. Replace gcc_assert with if.
16438 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16440 * config/i386/i386.c (initial_ix86_tune_features): Enable
16441 FP Reassociation for AMD bdver1 and bdver2.
16443 2013-05-29 Martin Jambor <mjambor@suse.cz>
16445 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16446 and IMAGPART_EXPR do not occur within other handled_components.
16448 2013-05-29 Richard Biener <rguenther@suse.de>
16450 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16451 access on whether the use is in the BB we currently try to
16453 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16454 vectorize to vect_bb_slp_scalar_cost.
16456 2013-05-29 Richard Biener <rguenther@suse.de>
16458 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16459 computing scalar cost offsetted by stmts that are kept live
16461 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16462 for computation of scalar cost.
16464 2013-05-28 Steve Ellcey <sellcey@mips.com>
16466 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16468 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16470 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16471 array notation built-in reduction functions.
16472 * doc/passes.texi (Passes): Added documentation about changes done
16474 * doc/invoke.texi (C Dialect Options): Added documentation about
16475 the -fcilkplus flag.
16476 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16477 (BUILTINS_DEF): Depend on cilkplus.def.
16478 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16479 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16480 * cilkplus.def: New file.
16482 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16484 PR rtl-optimization/57439
16485 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16487 2013-05-28 Easwaran Raman <eraman@google.com>
16489 PR tree-optimization/57337
16490 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16491 (find_insert_point): Correctly identify the insertion point
16492 when two statements with the same UID is compared.
16494 2013-05-28 Richard Biener <rguenther@suse.de>
16496 PR tree-optimization/56787
16497 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16498 from the list of data references.
16499 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16501 (vect_analyze_loop_operations): Likewise.
16502 (vect_transform_loop): Remove clobbers.
16504 2013-05-28 Martin Jambor <mjambor@suse.cz>
16506 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16507 and REALPART_EXPRs have scalar type.
16509 2013-05-28 Richard Biener <rguenther@suse.de>
16511 PR tree-optimization/57411
16512 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16514 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16515 virtual operand propagation.
16517 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16519 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16520 destination register for bmasksi_vis.
16521 (vector_init_bshuffle): Likewise.
16522 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16524 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16526 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16527 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16528 mode if the instruction isn't available in the original mode.
16529 * config/sparc/sparc.opt (mfix-ut699): New option.
16530 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16531 (divdf3): Turn into expander.
16532 (divdf3_nofix): New insn.
16533 (divdf3_fix): Likewise.
16534 (divsf3): Disable if -mfix-ut699.
16535 (sqrtdf2): Turn into expander.
16536 (sqrtdf2_nofix): New insn.
16537 (sqrtdf2_fix): Likewise.
16538 (sqrtsf2): Disable if -mfix-ut699.
16540 2013-05-27 Richard Biener <rguenther@suse.de>
16542 PR middle-end/57412
16543 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16544 block for the new loop.
16546 2013-05-27 Richard Biener <rguenther@suse.de>
16548 PR tree-optimization/57343
16549 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16550 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16551 (number_of_iterations_cond): Do not build the folded tree.
16553 2013-05-27 Richard Biener <rguenther@suse.de>
16556 PR middle-end/57381
16557 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16558 OEP_CONSTANT_ADDRESS_OF retained.
16560 PR tree-optimization/57417
16561 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16562 for unchanged base.
16563 (set_ssa_val_to): Compare addresses using
16564 get_addr_base_and_unit_offset.
16566 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16568 PR rtl-optimization/56833
16569 * postreload.c (move2add_record_mode): New function.
16570 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16571 (move2add_use_add2_insn): Use move2add_record_sym_value.
16572 (move2add_use_add3_insn): Likewise.
16573 (reload_cse_move2add): Use move2add_valid_value_p and
16574 move2add_record_mode. Invalidate call-clobbered and REG_INC
16575 affected regs by setting reg_mode to VOIDmode.
16576 (move2add_note_store): Don't pretend the inside of a SUBREG is
16577 the actual destination. Invalidate single/leading registers by
16578 setting reg_mode to VOIDmode.
16579 Use move2add_record_sym_value, move2add_valid_value_p and
16580 move2add_record_mode.
16582 2013-05-27 Richard Biener <rguenther@suse.de>
16584 PR tree-optimization/57396
16585 * tree-affine.c (double_int_constant_multiple_p): Properly
16586 return false for val == 0 and div != 0.
16588 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16590 * config/mips/mips.h: Use #elif in preprocessor conditions.
16592 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16595 * config/mips/constraints.md (kl): New constraint.
16596 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16597 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16598 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16599 from LO for MIPS16.
16600 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16602 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16605 * config/mips/mips.c (mips_can_inline_p): New function.
16606 (TARGET_CAN_INLINE_P): Define.
16608 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16610 * sched-int.h (ds_t, dw_t): Make unsigned int.
16611 Fix documentation that describes how all the ds_t bits are used.
16612 Reserve the last bit for delayed-branch scheduling.
16613 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16614 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16615 (gen_dep_weak_1): Remove prototype.
16616 * sched-deps.c (get_dep_weak_1): Make static.
16617 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16618 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16619 * doc/tm.texi: Regenerate.
16620 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16622 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16625 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16627 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16628 Sandra Loosemore <sandra@codesourcery.com>
16630 * config.gcc (powerpc-*): Allow native for with-cpu.
16632 2013-05-24 Jeff Law <law@redhat.com>
16634 PR tree-optimization/57124
16635 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16636 conversion feeding a condition if the range has an overflow
16637 if -fstrict-overflow. Add warnings for when we do make the
16640 2013-05-24 Dehao Chen <dehao@google.com>
16642 * tree-cfg.c (locus_discrim_map): Fix the typo.
16643 (locus_discrim_hasher): Likewise.
16644 (locus_discrim_hasher::hash): Likewise.
16645 (locus_discrim_hasher::equal): Likewise.
16647 2013-05-24 Martin Jambor <mjambor@suse.cz>
16649 PR tree-optimization/57294
16650 * cgraph.h (ipa_record_stmt_references): Declare.
16651 * cgraphbuild.c (ipa_record_stmt_references): New function.
16652 (build_cgraph_edges): Use ipa_record_stmt_references.
16653 (rebuild_cgraph_edges): Likewise.
16654 (cgraph_rebuild_references): Likewise.
16655 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16656 associated with the old statement and build references from the
16657 newly built statements.
16658 * ipa-ref.c (ipa_remove_stmt_references): New function.
16659 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16661 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16663 * lra-constraints.c (emit_spill_move): Use smaller mode for
16665 (check_and_process_move): Consider mem-reg moves for secondary
16667 (curr_insn_transform): Don't lose insns emitted before for
16668 secondary memory moves.
16669 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16670 reg set up in the current insn.
16672 2013-05-24 Dehao Chen <dehao@google.com>
16674 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16676 (locus_descrim_hasher::equal): Likewise.
16677 (build_gimple_cfg): New discriminator assignment algorithm.
16678 (make_edges): Likewise.
16679 (next_discriminator_for_locus): Likewise.
16680 (same_line_p): Likewise.
16681 (assign_discriminators): Likewise.
16682 (make_cond_expr_edges): Likewise.
16683 (make_gimple_switch_edges): Likewise.
16684 (make_goto_expr_edges): Likewise.
16685 (make_gimple_asm_edges): Likewise.
16687 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16689 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16690 X format specifier to only display bottom 16 bits.
16691 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16692 immediate to match for operand 2, since it will be masked.
16694 2013-05-24 Richard Biener <rguenther@suse.de>
16696 PR tree-optimization/57287
16697 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16698 all SSA names that occur in abnormal PHIs.
16700 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16703 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16704 that index is not negative.
16706 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16708 PR rtl-optimization/55177
16709 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16710 (simplify_byte_swapping_operation): New.
16711 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16712 (simplify_relational_operation_1): Deal with BSWAP.
16714 2013-05-23 Richard Henderson <rth@redhat.com>
16717 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16718 (ix86_reorg): Call it.
16720 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16723 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16724 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16725 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16727 2013-05-23 Christian Bruel <christian.bruel@st.com>
16730 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16732 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16733 Marcus Shawcroft <marcus.shawcroft@arm.com>
16735 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16736 * config/aarch64/constraints.md (Usa): Remove.
16737 * doc/md.texi (AArch64 Usa): Remove.
16739 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16740 Marcus Shawcroft <marcus.shawcroft@arm.com>
16742 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16743 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16744 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16745 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16747 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16749 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16750 instruction (AdvSIMD).
16751 * config/aarch64/aarch64-builtins.c
16752 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16753 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16755 2013-05-23 Martin Jambor <mjambor@suse.cz>
16757 PR middle-end/57347
16758 * tree.h (contains_bitfld_component_ref_p): Declare.
16759 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16760 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16762 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16763 not access a bit-field. Assert all final offsets are byte-aligned.
16765 2013-05-23 Richard Biener <rguenther@suse.de>
16767 PR tree-optimization/57380
16768 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16769 least one invariant or re-used load.
16770 * passes.c (init_optimization_passes): Move pass_phiprop before
16773 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16775 * config/aarch64/aarch64-simd.md
16776 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16778 2013-05-23 Richard Biener <rguenther@suse.de>
16780 PR middle-end/57381
16781 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16782 OEP_CONSTANT_ADDRESS_OF retained.
16784 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16786 PR middle-end/57344
16787 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16788 don't lower unit. Handle unit not being always BITS_PER_WORD.
16790 2013-05-23 Richard Biener <rguenther@suse.de>
16792 PR rtl-optimization/57341
16793 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16794 instead of true_dependence.
16796 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16798 * bb-reorder.c (branch_threshold): Make const.
16799 (exec_threshold): Ditto.
16801 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16802 Pat Haugen <pthaugen@us.ibm.com>
16803 Peter Bergner <bergner@vnet.ibm.com>
16805 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16806 documentation for the power8 crypto builtins.
16808 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16810 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16811 macros for defining power8 builtin functions.
16812 (BU_P8V_AV_2): Likewise.
16813 (BU_P8V_AV_P): Likewise.
16814 (BU_P8V_VSX_1): Likewise.
16815 (BU_P8V_OVERLOAD_1): Likewise.
16816 (BU_P8V_OVERLOAD_2): Likewise.
16817 (BU_CRYPTO_1): Likewise.
16818 (BU_CRYPTO_2): Likewise.
16819 (BU_CRYPTO_3): Likewise.
16820 (BU_CRYPTO_OVERLOAD_1): Likewise.
16821 (BU_CRYPTO_OVERLOAD_2): Likewise.
16822 (XSCVSPDP): Fix typo, point to the correct instruction.
16823 (VCIPHER): Add power8 crypto builtins.
16824 (VCIPHERLAST): Likewise.
16825 (VNCIPHER): Likewise.
16826 (VNCIPHERLAST): Likewise.
16827 (VPMSUMB): Likewise.
16828 (VPMSUMH): Likewise.
16829 (VPMSUMW): Likewise.
16830 (VPERMXOR_V2DI): Likewise.
16831 (VPERMXOR_V4SI: Likewise.
16832 (VPERMXOR_V8HI: Likewise.
16833 (VPERMXOR_V16QI: Likewise.
16834 (VSHASIGMAW): Likewise.
16835 (VSHASIGMAD): Likewise.
16836 (VPMSUM): Likewise.
16837 (VPERMXOR): Likewise.
16838 (VSHASIGMA): Likewise.
16840 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16841 __CRYPTO__ if the crypto instructions are available.
16842 (altivec_overloaded_builtins): Add support for overloaded power8
16845 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16846 support for power8 crypto builtins.
16847 (builtin_function_type): Likewise.
16848 (altivec_init_builtins): Add support for builtins that take vector
16849 long long (V2DI) arguments.
16851 * config/rs6000/crypto.md: New file, define power8 crypto
16854 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16855 Pat Haugen <pthaugen@us.ibm.com>
16856 Peter Bergner <bergner@vnet.ibm.com>
16858 * doc/invoke.texi (Option Summary): Add power8 options.
16859 (RS/6000 and PowerPC Options): Likewise.
16861 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16862 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16863 wm, wn, wr documentation.
16865 * config/rs6000/constraints.md (wm): New constraint for VSX
16866 registers if direct move instructions are enabled.
16867 (wn): New constraint for no registers.
16868 (wq): New constraint for quad word even GPR registers.
16869 (wr): New constraint if 64-bit instructions are enabled.
16870 (wv): New constraint if power8 vector instructions are enabled.
16871 (wQ): New constraint for quad word memory locations.
16873 * config/rs6000/predicates.md (const_0_to_15_operand): New
16874 constraint for 0..15 for crypto instructions.
16875 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16876 as GPR and floating point registers.
16877 (int_reg_operand): New predicate to match only GPR registers.
16878 (base_reg_operand): New predicate to match base registers.
16879 (quad_int_reg_operand): New predicate to match even GPR registers
16880 for quad memory operations.
16881 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16882 operations in both GPR and VSX registers.
16883 (quad_memory_operand): New predicate for quad memory operations.
16884 (reg_or_indexed_operand): New predicate for direct move support.
16886 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16887 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16888 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16889 (POWERPC_MASKS): Add power8 options.
16890 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
16893 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
16894 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
16896 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
16897 (-mpower8-fusion): New power8 options.
16898 (-mpower8-fusion-sign): Likewise.
16899 (-mpower8-vector): Likewise.
16900 (-mcrypto): Likewise.
16901 (-mdirect-move): Likewise.
16902 (-mquad-memory): Likewise.
16904 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
16905 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
16907 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
16908 (rs6000_debug_vector_unit): Add p8_vector.
16909 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
16910 definitions. Also print fusion state.
16911 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
16912 (rs6000_builtin_mask_calculate): Add power8 builtin support.
16913 (rs6000_option_override_internal): Add support for power8.
16914 (rs6000_common_init_builtins): Add debugging for skipped builtins
16915 if -mdebug=builtin.
16916 (rs6000_adjust_cost): Add power8 support.
16917 (rs6000_issue_rate): Likewise.
16918 (insn_must_be_first_in_group): Likewise.
16919 (insn_must_be_last_in_group): Likewise.
16920 (force_new_group): Likewise.
16921 (rs6000_register_move_cost): Likewise.
16922 (rs6000_opt_masks): Likewise.
16924 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
16925 power8 capable assembler, default to power7 options.
16926 (TARGET_DIRECT_MOVE): Likewise.
16927 (TARGET_CRYPTO): Likewise.
16928 (TARGET_P8_VECTOR): Likewise.
16929 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
16930 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
16931 (VECTOR_MEM_P8_VECTOR_P): Likewise.
16932 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
16933 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
16934 (TARGET_XSCVDPSPN): Likewise.
16935 (TARGET_XSCVSPDPN): Likewsie.
16936 (TARGET_SYNC_HI_QI): Likewise.
16937 (TARGET_SYNC_TI): Likewise.
16938 (MASK_CRYPTO): Likewise.
16939 (MASK_DIRECT_MOVE): Likewise.
16940 (MASK_P8_FUSION): Likewise.
16941 (MASK_P8_VECTOR): Likewise.
16942 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
16943 temporary used by some of the direct move instructions to get two FP
16944 temporary registers does not force creation of a stack frame.
16945 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
16946 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
16947 that any VSX registers are tieable, even if they are also an
16948 Altivec vector mode.
16949 (r6000_reg_class_enum): Add wm, wr, wv constraints.
16950 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
16951 (RS6000_BTM_CRYPTO): Likewise.
16952 (RS6000_BTM_COMMON): Likewise.
16954 * config/rs6000/rs6000.md (cpu attribute): Add power8.
16955 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
16956 (enum rs6000_vector): Add power8 vector support.
16958 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16962 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
16963 (any_sibcall_could_use_r3): this and handle indirect calls.
16964 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
16966 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16968 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
16970 2013-05-22 Richard Biener <rguenther@suse.de>
16972 PR middle-end/57349
16973 * profile.c (branch_prob): Do not split blocks that are
16974 abnormally receiving from ECF_RETURNS_TWICE functions.
16976 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16978 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16979 address mode. Move pointer mode initialization to the same place.
16981 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16983 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16984 while it has any effect.
16986 2013-05-21 Easwaran Raman <eraman@google.com>
16988 PR tree-optimization/57322
16989 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16990 UID of the statement added to the BB to be 1.
16992 2013-05-21 Jakub Jelinek <jakub@redhat.com>
16994 PR tree-optimization/57331
16995 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16996 of conversion from pointer type to integral type with integer.
16998 2013-05-21 Martin Jambor <mjambor@suse.cz>
17001 * ipa-prop.c (ipa_read_node_info): Process param_used and
17002 controlled_uses in the same order as when writing.
17004 2013-05-21 Magnus Granberg <baldrick@free.fr>
17007 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17009 2013-05-21 Richard Biener <rguenther@suse.de>
17011 PR tree-optimization/57318
17012 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17013 estimate stmts with side-effects as likely eliminated.
17015 2013-05-21 Richard Biener <rguenther@suse.de>
17017 PR tree-optimization/57330
17018 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17019 preserve the call stmts fntype.
17021 2013-05-21 Richard Biener <rguenther@suse.de>
17023 PR tree-optimization/57303
17024 * tree-ssa-sink.c (statement_sink_location): Improve killing
17025 stmt detection and properly handle self-assignments.
17027 2013-05-21 Christian Bruel <christian.bruel@st.com>
17029 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17030 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17031 registers. Set register size out of the PARALLEL loop.
17033 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17036 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17037 (*fmasf4, *fmasf4_media): New insns.
17039 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17041 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17042 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17043 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17044 (mips_idiv_insns): Update the comments to say that the returned
17045 instruction counts are in units of BASE_INSN_LENGTH.
17046 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17047 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17048 using 2 rather than 4 as the length of indirect MIPS16 and
17049 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17050 length of a NOP. Don't divide MIPS16 lengths by 2.
17051 (mips16_split_long_branches): Assume a branch is long if the
17052 length is greater than 4 rather than 8.
17053 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17054 rather than multiplying them by 2. Multiply instruction counts
17055 by BASE_INSN_LENGTH rather than 4.
17056 (*jump_mips16, tls_get_tp_mips16_<mode>)
17057 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17059 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17061 * config/mips/mips.md (extended_mips16): Remove branch case.
17062 (length): Remove duplicated extended_mips16 test.
17064 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17066 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17068 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17070 * recog.h (Recog_data): Rename to...
17071 (recog_data_d): ...this.
17072 (recog_data): Update accordingly.
17073 * recog.c (recog_data): Likewise.
17074 * reload.c (save_recog_data): Likewise.
17075 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17076 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17078 2013-05-17 Julian Brown <julian@codesourcery.com>
17080 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17081 found in a REG_EQUAL note, invalidate it.
17083 2013-05-17 Easwaran Raman <eraman@google.com>
17085 * tree-ssa-reassoc.c (find_insert_point): New function.
17086 (insert_stmt_after): Likewise.
17087 (get_def_stmt): Likewise.
17088 (ensure_ops_are_available): Likewise.
17089 (not_dominated_by): Likewise.
17090 (rewrite_expr_tree): Do not move statements beyond what is
17091 necessary. Remove call to swap_ops_for_binary_stmt...
17092 (reassociate_bb): ... and move it here.
17093 (build_and_add_sum): Assign UIDs for new statements.
17094 (linearize_expr): Likewise.
17095 (do_reassoc): Renumber gimple statement UIDs.
17097 2013-05-17 Jan Hubicka <jh@suse.cz>
17099 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17101 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17102 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17104 (output_weakrefs): Update.
17106 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17107 Martin Jambor <mjambor@suse.cz>
17109 PR middle-end/57276
17110 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17111 value that corresponds to the given aggval is found in values vector.
17113 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17115 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17116 sse, sse2, sse3, ssse3 and sse4a flags to options.
17118 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17120 * gengtype-state.c: (s_expr_writer): New class, to handle
17121 prettifying of output layout of s-expressions.
17122 (state_writer): New class, to write out gtype.state.
17123 (state_written_type_count): Move this variable into member data of
17125 (s_expr_writer::s_expr_writer): New code: constructor for new class
17126 (state_writer::state_writer(): ditto
17127 (s_expr_writer::write_new_line): New function
17128 (s_expr_writer::write_any_indent): ditto
17129 (s_expr_writer::begin_s_expr): ditto
17130 (s_expr_writer::end_s_expr): ditto
17131 (write_state_fileloc): convert to method of state_writer...
17132 (state_writer:: write_state_fileloc): ...and use methods of
17133 s_expr_writer to write indentation into the gtype.state output file
17134 to visually represent the hierarchical structure of the list
17136 (write_state_fields): ditto, renaming to...
17137 (state_writer::write_state_fields)
17138 (write_state_a_string): ditto, renaming to...
17139 (state_writer::write_state_a_string)
17140 (write_state_string_option): ditto, renaming to...
17141 (state_writer::write_state_string_option)
17142 (write_state_type_option): ditto, renaming to...
17143 (state_writer::write_state_type_option)
17144 (write_state_nested_option): ditto, renaming to...
17145 (state_writer::write_state_nested_option)
17146 (write_state_option): ditto, renaming to...
17147 (state_writer::write_state_option)
17148 (write_state_options): ditto, renaming to...
17149 (state_writer::write_state_options)
17150 (write_state_lang_bitmap): ditto, renaming to...
17151 (state_writer::write_state_lang_bitmap)
17152 (write_state_version): ditto, renaming to...
17153 (state_writer::write_state_version)
17154 (write_state_scalar_type): ditto, renaming to...
17155 (state_writer::write_state_scalar_type)
17156 (write_state_string_type): ditto, renaming to...
17157 (state_writer::write_state_string_type)
17158 (write_state_undefined_type): ditto, renaming to...
17159 (state_writer::write_state_undefined_type)
17160 (write_state_struct_union_type): ditto, renaming to...
17161 (state_writer::write_state_struct_union_type)
17162 (write_state_struct_type): ditto, renaming to...
17163 (state_writer::write_state_struct_type)
17164 (write_state_user_struct_type): ditto, renaming to...
17165 (state_writer::write_state_user_struct_type)
17166 (write_state_lang_struct_type): ditto, renaming to...
17167 (state_writer::write_state_lang_struct_type)
17168 (write_state_param_struct_type): ditto, renaming to...
17169 (state_writer::write_state_param_struct_type)
17170 (write_state_pointer_type): ditto, renaming to...
17171 (state_writer::write_state_pointer_type)
17172 (write_state_array_type): ditto, renaming to...
17173 (state_writer::write_state_array_type)
17174 (write_state_gc_used): ditto, renaming to...
17175 (state_writer::write_state_gc_used)
17176 (write_state_common_type_content): ditto, renaming to...
17177 (state_writer::write_state_common_type_content)
17178 (write_state_type): ditto, renaming to...
17179 (state_writer::write_state_type)
17180 (write_state_pair_list): ditto, renaming to...
17181 (state_writer::write_state_pair_list)
17182 (write_state_pair): ditto, renaming to...
17183 (state_writer::write_state_pair)
17184 (write_state_typedefs): ditto, renaming to...
17185 (state_writer::write_state_typedefs)
17186 (write_state_structures): ditto, renaming to...
17187 (state_writer::write_state_structures)
17188 (write_state_param_structs): ditto, renaming to...
17189 (state_writer::write_state_param_structs)
17190 (write_state_variables): ditto, renaming to...
17191 (state_writer::write_state_variables)
17192 (write_state_srcdir): ditto, renaming to...
17193 (state_writer::write_state_srcdir)
17194 (write_state_files_list): ditto, renaming to...
17195 (state_writer::write_state_files_list)
17196 (write_state_languages): ditto, renaming to...
17197 (state_writer::write_state_languages)
17198 (write_state): create a state_writer instance and use it when
17199 writing out the state file
17201 2013-05-17 Mike Stump <mikestump@comcast.net>
17203 PR rtl-optimization/57304
17204 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17205 accessing DF_REF_REAL_LOC.
17207 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17209 PR rtl-optimization/57281
17210 PR rtl-optimization/57300
17211 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17212 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17213 what the other splitter did if the registers are dead.
17215 2013-05-17 Richard Biener <rguenther@suse.de>
17217 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17220 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17222 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17225 2013-05-17 Marek Polacek <polacek@redhat.com>
17227 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17228 length when doing non-zero store of storing '\0' to '\0'.
17230 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17232 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17233 vect_external_def oprnd1 with loop_vinfo, try to emit
17234 optional cast, negation and and stmts on the loop preheader
17235 edge instead of into the pattern def seq.
17237 PR tree-optimization/57051
17238 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17239 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17241 2013-05-16 Nick Clifton <nickc@redhat.com>
17243 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17244 (rl78_is_naked_func): New function.
17245 (rl78_expand_prologue): Skip prologue generation for naked functions.
17246 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17247 * doc/extend.texi (naked): Add RL78 to the list of processors
17248 that supports this attribute.
17250 2013-05-16 Jeff Law <law@redhat.com>
17252 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17254 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17256 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17257 cache parameters using detect_caches_amd also for CYRIX,
17258 NSC and TM2 signatures.
17260 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17261 Dzianis Kahanovich <mahatma@eu.by>
17265 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17266 VIA/Centaur processors and determine their cache parameters
17267 using detect_caches_amd.
17269 2013-05-16 Teresa Johnson <tejohnson@google.com>
17271 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17272 (rtl_verify_edges): New function.
17273 (rtl_verify_bb_insns): Ditto.
17274 (rtl_verify_bb_pointers): Ditto.
17275 (rtl_verify_bb_insn_chain): Ditto.
17276 (rtl_verify_fallthru): Ditto.
17277 (rtl_verify_bb_layout): Ditto.
17278 (rtl_verify_flow_info_1): Outline checks into new functions.
17279 (rtl_verify_flow_info): Ditto.
17281 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17283 * cfghooks.c (copy_bbs): Add update_dominance argument.
17284 * cfghooks.h (copy_bbs): Update prototype.
17285 * tree-cfg.c (gimple_duplicate_sese_region):
17286 Add update_dominance argument.
17287 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17288 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17289 gimple_duplicate_sese_region call.
17290 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17291 Update copy_bbs call.
17292 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17293 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17295 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17297 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17298 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17299 vect_recog_rotate_pattern.
17300 (vect_recog_rotate_pattern): New function.
17302 2013-05-16 Jason Merrill <jason@redhat.com>
17304 * Makefile.in (LLINKER): New variable.
17305 (mostlyclean): Remove link mutex.
17306 * configure.ac: Handle --enable-link-mutex.
17307 * lock-and-run.sh: New script.
17309 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17312 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17315 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17317 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17319 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17321 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17322 * config/arm/arm.c (next_consecutive_mem): New function.
17323 (gen_movmem_ldrd_strd): Likewise.
17324 * config/arm/arm.md (movmemqi): Update condition and code.
17325 (unaligned_loaddi, unaligned_storedi): New patterns.
17327 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17329 * config.gcc: Obsolete *-*-solaris2.9*.
17330 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17332 2013-05-16 Richard Biener <rguenther@suse.de>
17334 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17335 earlier, after GRAPHITE transforms and IV canonicalization.
17337 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17339 * omp-low.c (extract_omp_for_data): For collapsed loops,
17340 if at least one of the loops is known at compile time to
17341 iterate zero times, set count to 0.
17342 (expand_omp_regimplify_p): New function.
17343 (expand_omp_for_generic): For collapsed loops, if at least
17344 one of the loops isn't known to iterate at least once,
17345 add runtime check with setting count to 0.
17346 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17347 For unsigned types if it isn't known at compile time that
17348 the loop will iterate at least once, add runtime check to bypass
17349 the whole loop if initial condition isn't true.
17351 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17353 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17355 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17357 PR middle-end/57286
17358 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17359 transformations to avoid an infinite loop.
17361 2013-05-16 Marek Polacek <polacek@redhat.com>
17363 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17365 2013-05-15 Leif Ekblad <leif@rdos.net>
17367 * config/i386/i386.c (ix86_decompose_address): Use
17368 DEFAULT_TLS_SEG_REG to access TLS segment register.
17369 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17370 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17371 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17373 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17376 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17377 sibling calls to functions that would normally be lazily bound,
17378 unless $gp is call-clobbered.
17380 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17382 * config/i386/i386.c (ix86_option_override_internal): Update
17383 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17384 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17385 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17386 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17388 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17390 2013-05-15 Andreas Schwab <schwab@suse.de>
17392 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17393 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17396 2013-05-15 Teresa Johnson <tejohnson@google.com>
17398 * loop-unroll.c (report_unroll_peel): Check decision before
17399 emitting unroll/peel message.
17401 2013-05-15 Teresa Johnson <tejohnson@google.com>
17403 * function.h (has_bb_partition): New rtl_data flag.
17404 (bb_reorder_complete): Ditto.
17405 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17406 instead of flag_reorder_blocks_and_partition.
17407 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17408 with some enhancements.
17409 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17410 * bb-reorder.c (connect_traces): Check for has_bb_partition
17411 instead of flag_reorder_blocks_and_partition.
17412 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17413 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17414 verify_hot_cold_block_grouping.
17415 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17417 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17420 * config/arm/predicates.md (call_insn_operand): New predicate.
17421 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17422 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17423 if insn is not a tail call.
17424 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17426 * config/arm/arm.h (enum reg_class): New caller save register class.
17427 (REG_CLASS_NAMES): Likewise.
17428 (REG_CLASS_CONTENTS): Likewise.
17429 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17432 2013-05-15 Richard Biener <rguenther@suse.de>
17434 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17435 of MSG_OPTIMIZED_LOCATIONS.
17436 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17437 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17439 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17440 of MSG_OPTIMIZED_LOCATIONS.
17441 (execute_vect_slp): Likewise.
17442 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17443 (vect_create_cond_for_alias_checks): Likewise.
17444 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17445 (vect_recog_widen_mult_pattern): Likewise.
17446 (vect_recog_widen_sum_pattern): Likewise.
17447 (vect_recog_over_widening_pattern): Likewise.
17448 (vect_recog_widen_shift_pattern): Likewise.
17449 (vect_recog_vector_vector_shift_pattern): Likewise.
17450 (vect_recog_divmod_pattern): Likewise.
17451 (vect_recog_mixed_size_cond_pattern): Likewise.
17452 (vect_recog_bool_pattern): Likewise.
17453 (vect_pattern_recog_1): Likewise.
17455 2013-05-15 Martin Jambor <mjambor@suse.cz>
17457 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17458 non-functions to builtin_unreachable.
17459 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17460 correct when new direct edges were discovered.
17462 2013-05-15 Martin Jambor <mjambor@suse.cz>
17464 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17465 header, print symbol order instead of node uid, print more information
17466 about indirect edge targets.
17467 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17469 (ipa_make_edge_direct_to_target): Likewise.
17470 (remove_described_reference): Likewise.
17471 (propagate_controlled_uses): Likewise.
17472 (ipa_print_node_params): Also print symbol order.
17473 (ipcp_transform_function): Print symbol order instead of node uids.
17474 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17475 (cgraph_get_create_real_symbol_node): Likewise.
17476 * ipa-cp.c (print_lattice): Likewise.
17477 (print_all_lattices): Likewise.
17478 (determine_versionability): Likewise.
17479 (initialize_node_lattices): Likewise.
17480 (estimate_local_effects): Likewise.
17481 (update_profiling_info): Likewise.
17482 (create_specialized_node): Likewise.
17483 (perhaps_add_new_callers): Likewise.
17484 (decide_about_value): Likewise.
17485 (decide_whether_version_node): Likewise.
17486 (identify_dead_nodes): Likewise.
17487 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17488 (dump_inline_summary): Likewise.
17489 (estimate_node_size_and_time): Likewise.
17490 (inline_analyze_function): Likewise.
17491 * ipa-inline.c (report_inline_failed_reason): Likewise.
17492 (want_early_inline_function_p): Likewise.
17493 (edge_badness): Likewise.
17494 (update_edge_key): Likewise.
17495 (inline_small_functions): Likewise. Add dumping of order to two other
17497 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17498 instead of node uids.
17499 (propagate_pure_const): Likewise.
17500 (propagate_pure_const): Likewise.
17501 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17502 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17503 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17505 * tree-pretty-print.c (dump_function_header): Likewise.
17506 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17507 Print symbol order instead of node uids.
17509 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17511 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17512 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17514 2013-05-15 Richard Biener <rguenther@suse.de>
17516 PR tree-optimization/57275
17517 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17518 return value for fail to do runtime alias checks for gather loads.
17520 2013-05-15 Jan Hubicka <jh@suse.cz>
17524 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17525 weakrefs are not external.
17526 (lto_symtab_merge_decls): Fix thinko when dealing with
17527 non-lto_symtab decls.
17528 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17529 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17530 * varpool.c (dump_varpool_node): Dump more flags.
17532 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17534 * config/i386/i386.c (processor_alias_table): Add instruction
17535 FSGSBASE for AMD bdver3 architecture.
17537 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17539 * tree.c (warn_deprecated_use): Print file:line using locus color.
17540 * diagnostic.c (diagnostic_report_current_module): Print file:line
17541 and file:line:column using locus color.
17543 2013-05-14 Mike Stump <mikestump@comcast.net>
17545 * gdbinit.in: Add __null.
17547 2013-05-14 Mike Stump <mikestump@comcast.net>
17549 * recog.h: Rename struct recog_data to Recog_data.
17550 * recog.c: Likewise.
17551 * reload.c (can_reload_into): Likewise.
17552 * config/picochip/picochip.c: Likewise.
17554 2013-05-14 Mike Stump <mikestump@comcast.net>
17556 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17558 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17560 * resource.h (struct resources): Remove unch_memory member.
17561 (CLEAR_RESOURCE): Don't clear unch_memory.
17562 * resource.c (mark_referenced_resources): Don't set it.
17563 (mark_set_resources): Likewise.
17564 (mark_target_live_regs): Don't clear it.
17565 (init_resource_info): Likewise.
17566 * reorg.c (resource_conflicts_p): Don't compare it.
17567 (redundant_insn): Don't set it.
17569 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17571 * emit-rtl.c (next_label): Remove unused function.
17572 (skip_consecutive_labels, link_cc0_insns): Move to ...
17573 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17574 only place where these functions are used, and make them static.
17576 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17578 * fold-const.c (fold_negate_expr): Handle vectors.
17579 (fold_truth_not_expr): Make it static.
17580 (fold_invert_truthvalue): New static function.
17581 (invert_truthvalue_loc): Handle vectors. Do not call
17582 fold_truth_not_expr directly.
17583 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17584 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17585 (fold_comparison): Handle vector constants.
17586 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17587 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17588 * tree.h (fold_truth_not_expr): Remove declaration.
17590 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17592 * config/aarch64/aarch64-simd.md
17593 (aarch64_vcond_internal<mode>): Rename to...
17594 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17595 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17596 float modes. Clarify all iterator modes.
17597 (vcond<mode><mode>): Use new name for vcond expanders.
17598 (vcond<v_cmp_result><mode>): Likewise.
17599 (vcondu<mode><mode>: Likewise.
17600 * config/aarch64/iterators.md (VDQF_COND): New.
17602 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17605 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17606 variable for the shift amount. Check that we shift by non-negative
17609 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17612 * config/arm/arm.h (EPILOGUE_USES): Only return true
17613 for LR_REGNUM after epilogue_completed.
17615 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17617 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17618 is error_mark_node.
17620 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17623 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17624 and Solaris 11+/x86 with gld.
17625 * configure: Regenerate.
17627 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17629 * expmed.c (expand_shift_1): Canonicalize rotates by
17630 constant bitsize / 2 to bitsize - 1.
17631 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17632 case ROTATERT>: Likewise.
17635 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17637 * config/i386/i386.md (rotateinv): New code attr.
17638 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17639 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17640 roll $31, %eax, etc.
17642 2013-05-14 Richard Biener <rguenther@suse.de>
17644 PR middle-end/57235
17645 * tree-eh.c (sink_clobbers): Give up for successors with
17646 multiple predecessors and no virtual uses.
17648 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17650 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17651 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17653 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17655 PR middle-end/57251
17656 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17657 the case when both op0 and op1 have VOIDmode.
17659 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17661 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17662 in multiply-accumulate mode.
17664 2013-05-13 Guozhi Wei <carrot@google.com>
17666 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17668 2013-05-13 Kai Tietz <ktietz@redhat.com>
17671 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17672 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17673 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17674 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17675 (ix86_expand_prologue): Likewise.
17676 (ix86_expand_split_stack_prologue): Likewise.
17677 (legitimate_pic_address_disp_p): Likewise.
17678 (legitimize_pic_address): Likewise.
17679 (legitimize_tls_address): Likewise.
17680 (legitimize_pe_coff_symbol): Likewise.
17681 (output_pic_addr_const): Likewise.
17682 (construct_plt_address): Likewise.
17683 (ix86_expand_call): Likewise.
17684 (x86_output_mi_thunk): Likewise.
17685 (x86_function_profiler): Likewise.
17687 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17689 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17690 similar switch cases.
17691 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17692 (aarch64_simd_mov_to_<mode>low): Delete.
17693 (aarch64_simd_mov_to_<mode>high): Delete.
17694 (move_lo_quad_<mode>): Add w<-r alternative.
17695 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17696 (aarch64_simd_mov_from_*): Update type attribute.
17697 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17700 2013-05-13 Jan Hubicka <jh@suse.cz>
17702 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17703 * config/i386/i386.c (ix86_compute_frame_layout,
17704 ix86_expand_epilogue, emit_i387_cw_initialization,
17705 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17706 ix86_local_alignment): Fix use of size/speed predicates.
17708 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17710 PR tree-optimization/45216
17711 PR tree-optimization/57157
17712 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17713 the (-Y) & (B - 1) variant if OP is |.
17714 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17715 return shifted. Use (-op1) & (prec - 1) as other_amount
17716 instead of prec - op1.
17718 2013-05-13 Martin Jambor <mjambor@suse.cz>
17720 PR middle-end/42371
17721 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17722 (ipa_constant_data): New type.
17723 (ipa_jump_func): Use ipa_constant_data to hold information about
17724 constant jump functions.
17725 (ipa_get_jf_constant): Adjust to jump function type changes.
17726 (ipa_get_jf_constant_rdesc): New function.
17727 (ipa_param_descriptor): New field controlled_uses.
17728 (ipa_get_controlled_uses): New function.
17729 (ipa_set_controlled_uses): Likewise.
17730 * ipa-ref.h (ipa_find_reference): Declare.
17731 * ipa-prop.c (ipa_cst_ref_desc): New type.
17732 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17734 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17735 New parameter cs. Adjust all callers.
17736 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17737 (remove_described_reference): New function.
17738 (jfunc_rdesc_usable): Likewise.
17739 (try_make_edge_direct_simple_call): Decrement controlled use count,
17740 attempt to remove reference if it hits zero.
17741 (combine_controlled_uses_counters): New function.
17742 (propagate_controlled_uses): Likewise.
17743 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17744 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17745 (ipa_print_node_params): Print described use counter.
17746 (ipa_write_jump_function): Adjust to jump function type changes.
17747 (ipa_read_jump_function): New parameter CS, pass it to
17748 ipa_set_jf_constant. Adjust caller.
17749 (ipa_write_node_info): Stream controlled use count
17750 (ipa_read_node_info): Likewise.
17751 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17753 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17754 count. Remove cloning-added reference if it reaches zero.
17755 * ipa-ref.c (ipa_find_reference): New function.
17757 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17759 * config/i386/i386.c (processor_target_table): Modified default
17760 alignment values for AMD BD and BT architectures.
17762 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17764 * tree-vect-generic.c (uniform_vector_p): Move ...
17765 * tree.c (uniform_vector_p): ... here.
17766 * tree.h (uniform_vector_p): Declare it.
17767 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17770 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17772 PR tree-optimization/57230
17773 * tree-ssa-strlen.c (handle_char_store): Record length for
17774 array store from STRING_CST.
17776 PR tree-optimization/57230
17777 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17780 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17782 * config/epiphany/epiphany.c (epiphany_init): Check size of
17783 NUM_MODES_FOR_MODE_SWITCHING.
17784 (epiphany_expand_prologue):
17785 Remove CONFIG_REGNUM initial value handling code.
17786 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17787 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17788 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17789 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17790 Don't return 1 for FP_MODE_NONE.
17791 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17792 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17793 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17794 * config/epiphany/epiphany.md (save_config): New pattern.
17796 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17798 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17800 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17802 * config/i386/i386.md (memory): Handle sseishft1.
17803 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17804 (*vec_extractv2di_1): Ditto.
17806 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17808 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17811 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17813 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17814 Add mthumb/march=armv7-a multilib.
17815 Add mthumb/march=armv7-r multilib.
17816 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17818 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17820 * config/v850/t-rtems: Add more multilibs.
17822 2013-05-10 Richard Biener <rguenther@suse.de>
17824 PR tree-optimization/57214
17825 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17826 not propagate from SSA names that occur in abnormal PHI nodes.
17828 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17830 * stor-layout.c (element_precision): New function.
17831 * machmode.h (element_precision): Declare it.
17832 * tree.c (build_minus_one_cst): New function.
17833 (element_precision): Likewise.
17834 * tree.h (build_minus_one_cst): Declare new function.
17835 (element_precision): Likewise.
17836 * fold-const.c (operand_equal_p): Use element_precision.
17837 (fold_binary_loc): Handle vector types.
17838 * convert.c (convert_to_integer): Use element_precision.
17839 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17842 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17844 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17845 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17846 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17847 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17848 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17849 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17850 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17851 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17852 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17853 (Uuw8): New constraints.
17854 (Usb4): Move into alphabetical order.
17855 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17856 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17857 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17858 previously unnamed patterns.
17859 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17860 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17861 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17862 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17863 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17864 of specific lengths.
17866 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17868 * config/i386/i386.md (rotateinv): New code attr.
17869 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17870 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17871 roll $31, %eax, etc.
17873 PR tree-optimization/45216
17874 PR tree-optimization/57157
17875 * tree-ssa-forwprop.c (simplify_rotate): New function.
17876 (ssa_forward_propagate_and_combine): Call it.
17878 2013-05-10 Richard Biener <rguenther@suse.de>
17880 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17881 disable peeling when we version for aliasing.
17882 (vector_alignment_reachable_p): Honor explicit user alignment.
17883 (vect_supportable_dr_alignment): Likewise.
17884 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17885 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17886 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17887 then peeling to arrange for the cost-model check to come first.
17889 2013-05-10 Alan Modra <amodra@gmail.com>
17891 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17892 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
17893 * configure: Regenerate.
17895 2013-05-10 Alan Modra <amodra@gmail.com>
17898 * varasm.c (default_elf_select_section): Move !DECL_P check..
17899 (get_named_section): ..to here before calling get_section_name.
17901 (default_section_type_flags): Add DECL_P check.
17902 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
17903 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
17905 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
17907 * config/epiphany/epiphany.c (epiphany_expand_prologue):
17908 When using gen_stack_adjust_str with a register offset, add a
17909 REG_FRAME_RELATED_EXPR note.
17911 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
17913 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
17914 (*vec_extractv4si_zext_mem): Ditto.
17915 (*vec_extractv2di): Add 0->x and x->x alternatives.
17916 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
17917 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
17919 2013-05-09 Jason Merrill <jason@redhat.com>
17921 N3639 C++1y VLA support
17922 * gimplify.c (gimplify_vla_decl): Don't touch an existing
17925 * tree.c (build_constructor_va): New.
17926 * tree.h: Declare it.
17928 2013-05-09 Martin Jambor <mjambor@suse.cz>
17931 * gimple-fold.c (canonicalize_constructor_val): Call
17932 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
17934 2013-05-09 Jan Hubicka <jh@suse.cz>
17935 Richard Biener <rguenther@suse.de>
17938 * symtab.c (symtab_make_decl_local): Do not add private names.
17940 2013-05-09 Jan Hubicka <jh@suse.cz>
17943 * symtab.c (insert_to_assembler_name_hash): Handle clones.
17944 (unlink_from_assembler_name_hash): Likewise.
17945 (symtab_prevail_in_asm_name_hash, symtab_register_node,
17946 symtab_unregister_node, symtab_initialize_asm_name_hash,
17947 change_decl_assembler_name): Update.
17949 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
17951 * config/aarch64/aarch64.md: New movtf split.
17952 (*movtf_aarch64): Update.
17953 (aarch64_movdi_tilow): Handle TF modes and rename to
17954 aarch64_movdi_<mode>low.
17955 (aarch64_movdi_tihigh): Handle TF modes and rename to
17956 aarch64_movdi_<mode>high
17957 (aarch64_movtihigh_di): Handle TF modes and rename to
17958 aarch64_mov<mode>high_di
17959 (aarch64_movtilow_di): Handle TF modes and rename to
17960 aarch64_mov<mode>low_di
17961 (aarch64_movtilow_tilow): Remove spurious whitespace.
17962 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
17964 (aarch64_print_operand): Update.
17966 2013-05-09 Alan Modra <amodra@gmail.com>
17968 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
17970 * configure: Regenerate.
17972 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17974 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17975 splitter preparation statements.
17976 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17977 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17980 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17982 * gimple-ssa-strength-reduction.c (count_candidates): Change
17983 return value to int.
17984 (analyze_candidates_and_replace): Change type of length to int.
17986 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
17988 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17989 (*vec_extract<mode>): Use VI12_128 mode iterator.
17990 (*vec_extract<mode>_mem): Ditto.
17991 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17994 2013-05-08 Diego Novillo <dnovillo@google.com>
17999 2012-08-17 Diego Novillo <dnovillo@google.com>
18002 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18003 * config.in: Regenerate.
18004 * configure: Regenerate.
18005 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18007 2013-05-08 Jan Hubicka <jh@suse.cz>
18010 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18011 * cgraph.h (symtab_node_base): Add unique_name.
18012 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18013 input_overwrite_node, input_varpool_node): Stream unique_name.
18014 * cgraphclones.c (cgraph_create_virtual_clone,
18015 cgraph_function_versioning): Set unique_name.
18016 * ipa.c (function_and_variable_visibility): Set unique_name.
18018 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18020 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18021 (alloc_cand_and_find_basis): Restrict conditional candidate
18022 processing to CAND_MULTs.
18024 2013-05-08 Jan Hubicka <jh@suse.cz>
18027 lto-symtab.c (lto_symtab_symbol_p): New function.
18028 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18029 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18030 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18031 Skip static symbols.
18033 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18035 PR tree-optimization/57200
18036 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18037 Only call inform if the preceding warning_at returns true.
18039 2013-05-07 Han Shen <shenhan@google.com>
18041 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18042 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18043 * common.opt (fstack-protector-strong): New option.
18044 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18045 * doc/invoke.texi (Optimization Options): Document
18046 "-fstack-protector-strong".
18047 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18049 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18051 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18053 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18055 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18056 it is not lhs of the insn.
18057 (setup_reg_equiv): Remove insn having equiv memory which it is not
18059 * lra-constraints.c (process_address): Try to improve generation
18060 code for address base + disp.
18061 (lra_constraints): Make correct the code for checking insn setting
18062 up backward equivalence. Remove insn only if it is in the init
18064 * lra-eliminations.c (update_reg_eliminate): Change return value.
18065 (lra_eliminate): Use the result.
18067 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18069 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18070 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18071 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18072 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18073 register target operands.
18074 (*vec_extractv8hi_sse2): New pattern.
18075 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18076 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18077 (*vec_extract<mode>_mem): New insn and split pattern.
18079 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18081 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18082 (TARGET_ASAN_SHADOW_OFFSET): Define.
18083 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18084 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18086 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18088 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18089 (incr_vec_index): Return -1 if increment not found.
18090 (create_add_on_incoming_edge): Assert if increment not found.
18091 (record_increment): Limit number of increments recorded.
18092 (all_phi_incrs_profitable): Return false if an increment not found.
18093 (replace_profitable_candidates): Don't process increments that were
18095 (analyze_candidates_and_replace): Limit size of incr_vec.
18097 2013-05-07 Richard Biener <rguenther@suse.de>
18099 * calls.c (special_function_p): setjmp-like functions are leaf.
18100 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18101 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18103 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18105 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18106 (aarch64_simd_mov<mode>): New expander.
18107 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18108 (aarch64_simd_mov_to_<mode>high): Likewise.
18109 (aarch64_simd_mov_from_<mode>low): Likewise.
18110 (aarch64_simd_mov_from_<mode>high): Likewise.
18111 (aarch64_dup_lane<mode>): Update.
18112 (aarch64_dup_lanedi): New instruction pattern.
18113 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18114 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18116 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18118 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18119 (replace_mult_candidate): Remove unnecessary argument; remove
18120 unnecessary parameter from call to introduce_cast_before_cand.
18121 (replace_unconditional_candidate): Remove unnecessary parameter
18122 from call to replace_mult_candidate.
18123 (replace_conditional_candidate): Likewise.
18124 (insert_initializers): Use make_temp_ssa_name.
18125 (introduce_cast_before_cand): Remove unnecessary argument; use
18126 make_temp_ssa_name.
18127 (replace_one_candidate): Remove unnecessary argument; remove
18128 unnecessary parameter from calls to introduce_cast_before_cand.
18129 (replace_profitable_candidates): Remove unnecessary parameters
18130 from calls to replace_one_candidate.
18132 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18134 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18135 phi def as possibly hiding a basis for a CAND_ADD whose operands
18136 have been commuted in the analysis.
18137 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18139 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18141 * config/aarch64/aarch64.md
18142 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18143 shift value between 0-4.
18145 2013-05-07 Richard Biener <rguenther@suse.de>
18147 * double-int.h (rshift): New overload.
18148 * double-int.c (rshift): New function.
18149 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18150 (create_reference_ops_from_ref): Remove.
18151 (vn_reference_insert): Use shared ops for constructing the
18152 reference and copy it.
18154 2013-05-07 Richard Biener <rguenther@suse.de>
18156 PR middle-end/57190
18157 * tree-eh.c (sink_clobbers): Properly propagate
18158 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18160 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18162 PR tree-optimization/57149
18163 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18164 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18165 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18166 uninit_undefined_value_p instead of ssa_undefined_value_p.
18169 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18170 for modifier == EXPAND_INITIALIZER.
18172 2013-05-07 Anton Blanchard <anton@samba.org>
18174 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18175 for powerpc64 little endian.
18176 * configure: Regenerate.
18178 2013-05-06 Graham Stott <grahams@btinternet.com>
18180 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18181 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18182 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18183 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18185 2013-05-06 Graham Stott <grahams@btinternet.com>
18187 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18188 codes which allow non-lvalues.
18190 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18192 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18193 components are all 1s.
18194 (integer_minus_onep): New function.
18195 * tree.h (integer_minus_onep): Declare it.
18196 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18197 integer_minus_onep instead of integer_all_onesp.
18199 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18202 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18203 variations of these patterns.
18205 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18207 * config/i386/i386.md (isa): Add x64_sse4 member.
18208 (enabled): Handle x64_sse4.
18209 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18210 instruction for 64bit SSE4_1 targets. Update insn attributes.
18211 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18212 instruction for SSE4_1 targets. Update insn attributes.
18213 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18214 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18215 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18217 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18218 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18219 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18221 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18224 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18226 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18228 * final.c (do_assembler_dialects): Don't handle curly braces and
18229 vertical bar escaped by % as dialect delimiters.
18230 (output_asm_insn): Print curly braces and vertical bar if escaped
18231 by % and ASSEMBLER_DIALECT defined.
18232 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18233 * doc/tm.texi: Regenerated.
18235 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18237 * config/mips/mips.c: Include tree-pass.h.
18238 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18239 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18240 (pass_mips_machine_reorg2): New machine specific pass.
18241 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18242 (mips_option_override): Register the new pass.
18243 * rtl.h (cleanup_barriers): Remove prototype.
18244 (dbr_schedule): Likewise.
18245 * jump.c (cleanup_barriers): Make static.
18246 * reorg.c (dbr_schedule): Likewise.
18248 2013-05-06 Richard Biener <rguenther@suse.de>
18250 PR tree-optimization/57185
18251 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18252 SSA names properly.
18254 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18257 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18258 between "first_arg" and "insn", not "last" and "insn".
18260 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18262 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18263 (find_candidates_in_block): Re-enable slsr_process_phi.
18264 (create_phi_basis): Fix double counting of candidate adjustment.
18266 2013-05-06 Richard Biener <rguenther@suse.de>
18268 PR middle-end/57147
18269 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18270 the edge is also fallthru, preserve it and just clear the
18272 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18273 also complex, preserve that and just clear the fallthru flag.
18274 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18275 update virtual operands.
18277 2013-05-06 Alan Modra <amodra@gmail.com>
18279 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18280 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18281 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18282 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18283 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18284 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18286 2013-05-06 Alan Modra <amodra@gmail.com>
18288 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18289 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18290 (DEFAULT_ASM_ENDIAN): Define.
18291 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18292 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18293 Update -K PIC clause from sysv4.h.
18294 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18295 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18297 2013-05-06 Alan Modra <amodra@gmail.com>
18299 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18300 twice for little-endian.
18301 (ashrdi3_no_power, ashrdi3): Support little-endian.
18303 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18306 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18307 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18309 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18310 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18311 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18314 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18315 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18317 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18318 * config/pa/pa.opt: Make mbig-switch a no-op.
18319 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18320 (CASE_VECTOR_MODE): Always return SImode.
18321 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18322 for the !TARGET_BIG_SWITCH case.
18323 * config/pa/pa-linux.h: Likewise.
18324 * config/pa/pa-openbsd.h: Likewise.
18325 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18326 * config/pa/pa.md (short_jump): Remove define_insn.
18327 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18328 (casesi0): Remove define_insn.
18329 (type): Remove btable_branch.
18330 (pa_combine_type): Likewise.
18331 (in_nullified_branch_delay): Likewise.
18332 (in_call_delay): Likewise.
18333 (define_delay): Likewise.
18334 (define_insn_reservation "Z3"): Likewise.
18335 (define_insn_reservation "Z4"): Likewise.
18336 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18337 (pa_adjust_insn_length): Remove adjustment for btable branches.
18338 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18341 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18343 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18344 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18345 Add m->r,x alternatives.
18346 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18347 splitters using SWI48x mode iterator.
18348 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18349 TARGET_64BIT. Add m->x alternative.
18350 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18351 Add o->x alternative. Enable for TARGET_SSE.
18352 (sse_storeq): Remove expander.
18353 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18354 with memory input operand.
18355 (*vec_extractv2di_1 splitter): New.
18356 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18357 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18359 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18361 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18362 (INT_LOWPART): Delete.
18363 (extract_MB): Adjust.
18364 (extract_ME): Adjust.
18365 (print_operand): Adjust.
18367 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18369 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18370 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18371 (reg_or_logical_cint_operand, easy_fp_constant,
18372 logical_const_operand): Delete "CONST_DOUBLE" case.
18373 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18374 "HOST_BITS_PER_WIDE_INT == 64" test.
18375 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18376 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18377 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18378 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18380 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18381 CONST_DOUBLE DImode/VOIDmode case.
18382 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18383 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18384 CONST_DOUBLE VOIDmode case.
18385 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18386 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18387 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18388 Delete CONST_DOUBLE case.
18389 (splitters for mov FMOVE64 const_double): Delete
18390 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18391 "HOST_BITS_PER_WIDE_INT >= 64" test.
18392 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18394 (mov DI const_double): Delete.
18396 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18398 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18399 on op shows all bits zero in mode of a lowpart subreg, return zero.
18401 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18404 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18405 to save TFmode registers and DImode to save TImode registers for
18406 caller save operations.
18407 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18408 mark being partially clobbered since they only use the first
18411 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18412 and TDmode only use the upper 64-bits of each VSX register.
18414 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18416 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18417 (find_candidates_in_block): Disable slsr_process_phi.
18419 2013-05-03 Guozhi Wei <carrot@google.com>
18421 * coverage.c (coverage_obj_init): Move the construction of gcov
18423 (build_init_ctor): ... here.
18425 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18427 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18428 (slsr_cand_d): Redefine def_phi.
18429 (stride_status, phi_adjust_status, count_phis_status): New enums.
18430 (find_phi_def): New.
18431 (find_basis_for_base_expr): New.
18432 (find_basis_for_candidate): Handle hidden bases.
18433 (alloc_cand_and_find_basis): Handle phi candidates.
18434 (slsr_process_phi): New.
18435 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18436 (create_mul_imm_cand): Likewise.
18437 (create_add_ssa_cand): Exclude phi base candidates.
18438 (create_add_imm_cand): Likewise.
18439 (slsr_process_cast): Likewise.
18440 (slsr_process_copy): Likewise.
18441 (find_candidates_in_block): Handle phi candidates.
18442 (dump_candidate): Likewise.
18443 (unconditional_cands): Delete.
18444 (unconditional_cands_with_known_stride_p): Delete.
18445 (phi_dependent_cand_p): New.
18446 (cand_increment): Handle phi-dependent candidates.
18447 (replace_dependent): Delete.
18448 (replace_mult_candidate): New.
18449 (replace_unconditional_candidate): New.
18450 (incr_vec_index): Move to avoid forward reference.
18451 (create_add_on_incoming_edge): New.
18452 (create_phi_basis): New.
18453 (replace_dependents): Delete.
18454 (replace_conditional_candidate): New.
18455 (phi_add_costs): New.
18456 (replace_uncond_cands_and_profitable_phis): New.
18457 (record_increment): Handle phi adjustments.
18458 (record_phi_increments): New.
18459 (record_increments): Handle phi adjustments.
18460 (phi_incr_cost): New.
18461 (lowest_cost_path): Handle phis.
18462 (total_savings): Likewise.
18463 (analyze_increments): Likewise.
18464 (ncd_with_phi): New.
18465 (ncd_of_cand_and_phis): New.
18466 (nearest_common_dominator_for_cands): Handle phi increments.
18467 (all_phi_incrs_profitable): New.
18468 (replace_profitable_candidates): Handle phi-dependent candidates.
18469 (analyze_candidates_and_replace): Likewise.
18471 2013-05-03 Teresa Johnson <tejohnson@google.com>
18474 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18475 do not exceed REG_BR_PROB_BASE.
18477 2013-05-03 Jeff Law <law@redhat.com>
18479 PR tree-optimization/57144
18480 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18481 operand of the condition will bit into the new type when eliminating
18482 a cast feeding a condition.
18484 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18486 PR rtl-optimization/57130
18487 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18488 of COMPARE as in_code to the recursive call if needed.
18490 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18492 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18493 (enabled): Handle new members.
18494 * config/i386/sse.md (*vec_concatv2si): Merge from
18495 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18496 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18498 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18500 PR tree-optimization/57027
18501 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18502 for fnms opportunity, check we got the prerequisite kind
18503 of tree / gimple before using accessor functions.
18505 2013-05-03 Richard Biener <rguenther@suse.de>
18507 * double-int.h (lshift): New overload without precision
18508 and arith argument.
18509 (operator *=, operator +=, operator -=): Move ...
18510 * double-int.c (operator *=, operator +=, operator -=): ... here
18511 and implement more efficiently.
18512 (mul_double_with_sign): Remove.
18513 (lshift_double): Adjust to take unsinged shift argument, push
18514 dispatching code to callers.
18515 (mul_double_wide_with_sign): Add early out for callers that
18516 are not interested in high parts or overflow.
18517 (lshift): New function.
18518 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18519 dispatch code here.
18520 (lrotate, rrotate): Use logical shifts.
18521 * expr.c (get_inner_reference): Use lshift.
18522 * fixed-value.c (do_fixed_divide): Likewise.
18523 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18524 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18525 (indirect_refs_may_alias_p): Likewise.
18526 (stmt_kills_ref_p_1): Likewise.
18528 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18530 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18532 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18534 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18535 scalar form of FABD instruction.
18537 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18539 * lra-constraints.c (process_alt_operands): Add checking alt
18540 number to choose the best alternative.
18542 2013-05-02 Richard Biener <rguenther@suse.de>
18544 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18545 bitmap and its handling.
18546 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18548 2013-05-02 Richard Biener <rguenther@suse.de>
18550 PR middle-end/57140
18551 * tree-inline.c (copy_loops): Properly handle removed loops.
18552 (copy_cfg_body): Mark destination loops for fixup if source
18553 loops needed fixup.
18555 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18558 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18559 generating simple_return for naked functions.
18561 2013-05-02 Martin Jambor <mjambor@suse.cz>
18563 PR middle-end/56988
18564 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18565 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18567 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18568 ipa_agg_replacement_value structures.
18569 (known_aggs_to_agg_replacement_list): Likewise.
18570 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18571 (read_agg_replacement_chain): Likewise.
18572 (ipcp_transform_function): Also check that by_ref flags match.
18574 2013-05-02 Richard Biener <rguenther@suse.de>
18576 * graphds.h (struct graph): Add obstack member.
18577 * graphds.c (new_graph): Initialize obstack and allocate
18579 (add_edge): Allocate edge from the obstack.
18580 (free_graph): Free the obstack instead of all edges and vertices.
18582 2013-05-02 Teresa Johnson <tejohnson@google.com>
18584 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18586 * cfg.c (update_bb_profile_for_threading): Ditto.
18587 * tree-inline.c (copy_bb): Ditto.
18588 (copy_edges_for_bb): Ditto.
18589 (initialize_cfun): Ditto.
18590 (copy_cfg_body): Ditto.
18591 (expand_call_inline): Ditto.
18592 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18593 (estimate_node_size_and_time): Ditto.
18594 (inline_merge_summary): Ditto.
18595 * cgraphclones.c (cgraph_clone_edge): Ditto.
18596 (cgraph_clone_node): Ditto.
18597 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18598 (compute_trg_info): Ditto.
18600 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18602 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18603 S reg when fp attribute set.
18604 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18606 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18608 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18610 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18611 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18612 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18614 2013-05-02 Richard Biener <rguenther@suse.de>
18616 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18617 (struct instantiate_cache_entry): New type.
18618 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18619 (struct instantiate_cache_type): New type.
18620 (set_instantiated_value, get_instantiated_value): Remove.
18621 (get_instantiated_value_entry): New function.
18622 (instantiate_scev_name): Use the new cache and adjust.
18623 (instantiate_scev_poly): Adjust.
18624 (instantiate_scev_binary): Likewise.
18625 (instantiate_array_ref): Likewise.
18626 (instantiate_scev_convert): Likewise.
18627 (instantiate_scev_not): Likewise.
18628 (instantiate_scev_3): Likewise.
18629 (instantiate_scev_2): Likewise.
18630 (instantiate_scev_r): Likewise.
18631 (instantiate_scev): Likewise.
18632 (resolve_mixers): Likewise.
18634 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18637 * lra-constraints.c (best_small_class_operands_num): Remove.
18638 (process_alt_operands): Remove small_class_operands_num. Take
18639 small classes operands into losers and only if the operand is not
18640 matched. Modify debugging output.
18641 (curr_insn_transform): Remove best_small_class_operands_num.
18644 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18646 * config/aarch64/aarch64-builtins.c
18647 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18648 * config/aarch64/aarch64-simd-builtins.def
18649 (reduc_splus_): Add new modes.
18650 (reduc_uplus_): New.
18651 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18652 (reduc_uplus_v4sf): Likewise.
18653 (reduc_splus_v4sf): Likewise.
18654 (aarch64_addv<mode>): Likewise.
18655 (reduc_uplus_<mode>): Likewise.
18656 (reduc_splus_<mode>): Likewise.
18657 (aarch64_addvv2di): Likewise.
18658 (reduc_uplus_v2di): Likewise.
18659 (reduc_splus_v2di): Likewise.
18660 (aarch64_addvv2si): Likewise.
18661 (reduc_uplus_v2si): Likewise.
18662 (reduc_splus_v2si): Likewise.
18663 (reduc_<sur>plus_<mode>): New.
18664 (reduc_<sur>plus_v2di): Likewise.
18665 (reduc_<sur>plus_v2si): Likewise.
18666 (reduc_<sur>plus_v4sf): Likewise.
18667 (aarch64_addpv4sf): Likewise.
18668 * config/aarch64/arm_neon.h
18669 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18670 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18671 add UNSPEC_SADDV, UNSPEC_UADDV.
18673 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18675 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18677 * config/aarch64/arm_neon.h
18678 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18680 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18682 * config/aarch64/aarch64-builtins
18683 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18685 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18687 * config/aarch64/aarch64-simd-builtins.def
18688 (reduc_smax_): New.
18689 (reduc_smin_): Likewise.
18690 (reduc_umax_): Likewise.
18691 (reduc_umin_): Likewise.
18692 (reduc_smax_nan_): Likewise.
18693 (reduc_smin_nan_): Likewise.
18696 (smax): Update for V2SF, V4SF and V2DF modes.
18699 (smin_nan): Likewise.
18700 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18701 (<su><maxmin><mode>3): ...This, refactor.
18702 (s<maxmin><mode>3): New.
18703 (<maxmin_uns><mode>3): Likewise.
18704 (reduc_<maxmin_uns>_<mode>): Refactor.
18705 (reduc_<maxmin_uns>_v4sf): Likewise.
18706 (reduc_<maxmin_uns>_v2si): Likewise.
18707 (aarch64_<fmaxmin><mode>: Remove.
18708 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18710 (vmin<q>_f<32,64>): Likewise.
18711 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18713 (su): Add mappings for smax, smin, umax, umin.
18715 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18716 (FMAXMIN): Rename as...
18717 (FMAXMIN_UNS): ...This.
18719 (fmaxminv): Likewise.
18720 (fmaxmin): Likewise.
18722 (maxmin_uns_op): Likewise.
18724 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18726 * config/aarch64/arm_neon.h
18727 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18728 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18729 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18731 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18733 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18734 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18736 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18738 * config/aarch64/aarch64-simd.md
18739 (vcond<mode>_internal): Handle special cases for constant masks.
18740 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18741 (vcondu<mode><mode>): Likewise.
18742 (vcond<v_cmp_result><mode>): New.
18744 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18746 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18747 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18748 * config/aarch64/aarch64-simd-builtins.def
18749 (cmeq): Update to BUILTIN_VALLDI.
18754 * config/aarch64/arm_neon.h
18755 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18756 to builtins or C as appropriate.
18758 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18760 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18762 (cmhi): Rename to...
18764 * config/aarch64/aarch64-simd.md
18765 (simd_mode): Add SF.
18766 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18767 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18768 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18769 (cstore<mode>_neg): ...This.
18770 * config/aarch64/iterators.md
18772 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18773 (COMPARISONS): New.
18774 (UCOMPARISONS): Likewise.
18775 (optab): Add missing comparisons.
18781 (VCMP_S): Likewise.
18782 (VCMP_U): Likewise.
18783 (V_cmp_result): Add DF, SF modes.
18784 (v_cmp_result): Likewise.
18786 (vmtype): Likewise.
18787 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18789 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18791 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18792 define_insn to define_insn_and_split.
18793 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18794 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18795 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18796 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18797 (thumb2_negscc): Likewise.
18799 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18801 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18803 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18805 * config/arm/thumb2.md: Remove trailing whitespaces.
18807 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18809 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18810 Use gen_int_mode rather than GEN_INT.
18812 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18814 * value-prof.c (stream_in_histogram_value): Remove the strayed
18817 2013-04-30 Richard Biener <rguenther@suse.de>
18819 PR middle-end/57122
18820 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18822 2013-04-30 Richard Biener <rguenther@suse.de>
18824 PR middle-end/57107
18825 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18827 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18829 PR rtl-optimization/56957
18830 PR rtl-optimization/57105
18831 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18832 variable. Use just INSN_UID for determining whether an insn
18833 should be only disconnected from the insn stream.
18834 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18836 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18838 PR tree-optimization/57104
18839 * tsan.c (instrument_expr): Don't instrument accesses to
18840 DECL_HARD_REGISTER VAR_DECLs.
18842 2013-04-30 Richard Biener <rguenther@suse.de>
18844 * function.h (loops_for_fn): New inline function.
18845 (set_loops_for_fn): Likewise.
18846 * cfgloop.h (place_new_loop): Add struct function parameter.
18847 (get_loop): Likewise.
18848 (get_loops): Likewise.
18849 (number_of_loops): Likewise.
18850 (fel_next): Adjust.
18851 (fel_init): Likewise.
18852 * cfg.c (get_loop_copy): Adjust.
18853 * cfgloop.c (flow_loops_dump): Likewise.
18854 (record_loop_exits): Likewise.
18855 (verify_loop_structure): Likewise.
18856 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18857 (estimate_reg_pressure_cost): Likewise.
18858 (mark_loop_exit_edges): Likewise.
18859 * cfgloopmanip.c (place_new_loop): Likewise.
18860 (add_loop): Likewise.
18861 (duplicate_loop): Likewise.
18862 * graph.c (draw_cfg_nodes): Likewise.
18863 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18864 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18865 (extract_affine_chrec): Likewise.
18866 (build_scop_iteration_domain): Likewise.
18867 * graphite.c (graphite_initialize): Likewise.
18868 * ira-build.c (create_loop_tree_nodes): Likewise.
18869 (more_one_region_p): Likewise.
18870 (rebuild_regno_allocno_maps): Likewise.
18871 (mark_loops_for_removal): Likewise.
18872 (mark_all_loops_for_removal): Likewise.
18873 (remove_unnecessary_regions): Likewise.
18874 (ira_build): Likewise.
18875 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18876 * loop-init.c (fix_loop_structure): Likewise.
18877 (gate_rtl_move_loop_invariants): Likewise.
18878 (gate_rtl_unswitch): Likewise.
18879 (gate_rtl_unroll_and_peel_loops): Likewise.
18880 (rtl_doloop): Likewise.
18881 * lto-streamer-in.c (input_cfg): Likewise.
18882 * lto-streamer-out.c (output_cfg): Likewise.
18883 * modulo-sched.c (sms_schedule): Likewise.
18884 * predict.c (tree_estimate_probability): Likewise.
18885 (tree_estimate_probability_driver): Likewise.
18886 (estimate_loops): Likewise.
18887 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18888 (move_sese_region_to_fn): Likewise.
18889 (debug_loop_num): Likewise.
18890 * tree-chrec.c (chrec_evaluate): Likewise.
18891 (hide_evolution_in_other_loops_than_loop): Likewise.
18892 (chrec_component_in_loop_num): Likewise.
18893 (reset_evolution_in_loop): Likewise.
18894 (evolution_function_is_invariant_rec_p): Likewise.
18895 * tree-if-conv.c (main_tree_if_conversion): Likewise.
18896 * tree-inline.c (copy_loops): Likewise.
18897 (copy_cfg_body): Likewise.
18898 (tree_function_versioning): Likewise.
18899 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
18900 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
18902 (add_to_evolution_1): Likewise.
18903 (scev_const_prop): Likewise.
18904 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
18905 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
18906 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
18907 (tree_ssa_lim_initialize): Likewise.
18908 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
18909 (verify_loop_closed_ssa): Likewise.
18910 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
18911 (tree_ssa_loop_im): Likewise.
18912 (tree_ssa_loop_unswitch): Likewise.
18913 (tree_vectorize): Likewise.
18914 (check_data_deps): Likewise.
18915 (tree_ssa_loop_ivcanon): Likewise.
18916 (tree_ssa_loop_bounds): Likewise.
18917 (tree_complete_unroll): Likewise.
18918 (tree_complete_unroll_inner): Likewise.
18919 (tree_parallelize_loops): Likewise.
18920 (tree_ssa_loop_prefetch): Likewise.
18921 (tree_ssa_loop_ivopts): Likewise.
18922 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
18923 * tree-vectorizer.c (vectorize_loops): Likewise.
18925 2013-04-29 Mike Frysinger <vapier@gentoo.org>
18927 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
18928 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
18929 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
18930 with EABI_LINK_SPEC.
18932 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18935 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
18938 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
18941 * lra-constraints.c (process_alt_operands): Discourage a bit more
18942 using memory for pseudos. Print cost dump for alternatives.
18943 Modify cost values for conflicts with early clobbers.
18944 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
18946 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18949 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
18951 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18953 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
18954 from/to S register.
18955 (movdi_aarch64): Support LDR/STR from/to D register.
18957 2013-04-29 Ian Bolton <ian.bolton@arm.com>
18959 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
18960 or higher by default.
18962 2013-04-29 Richard Biener <rguenther@suse.de>
18964 PR middle-end/57075
18965 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
18966 even if not adding abnormal edges for calls that can make
18969 2013-04-29 Richard Biener <rguenther@suse.de>
18971 PR middle-end/57103
18972 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18974 (move_stmt_r): Remove redundant checking.
18976 2013-04-29 Teresa Johnson <tejohnson@google.com>
18979 * basic-block.h (apply_scale): New function.
18980 (apply_probability): Use apply_scale.
18981 * gimple-streamer-in.c (input_bb): Ditto.
18982 * lto-streamer-in.c (input_cfg): Ditto.
18983 * lto-cgraph.c (merge_profile_summaries): Ditto.
18984 * tree-optimize.c (execute_fixup_cfg): Ditto.
18985 * tree-inline.c (copy_bb): Update comment to use apply_scale.
18986 (copy_edges_for_bb): Ditto.
18987 (copy_cfg_body): Ditto.
18989 2013-04-29 Tom de Vries <tom@codesourcery.com>
18991 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18992 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18993 (tail_merge_optimize): Handle current_loops == NULL.
18995 2013-04-26 Jeff Law <law@redhat.com>
18997 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18998 (simplify_cond_using_ranges): Generalize code to simplify
18999 COND_EXPRs where one argument is a constant and the other
19000 is an SSA_NAME created by an integral type conversion.
19002 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19004 * config/arm/arm.md (store_minmaxsi): Use only when
19005 optimize_insn_for_size_p.
19007 2013-04-29 Christian Bruel <christian.bruel@st.com>
19010 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19012 2013-04-29 Richard Biener <rguenther@suse.de>
19014 PR middle-end/57089
19015 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19016 loop tree make sure to schedule a fixup for the child as well.
19017 (expand_omp_for_generic): Properly add loops.
19018 (expand_omp_for_static_nochunk): Likewise.
19019 (expand_omp_for_static_chunk): Likewise.
19020 (expand_omp_for): For the degenerate case fixup loops.
19021 (expand_omp_sections): Fix default bb placement in loops.
19022 (expand_omp_atomic_pipeline): Properly add loops.
19024 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19026 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19028 2013-04-29 Tom de Vries <tom@codesourcery.com>
19030 * tree-ssa-tail-merge.c: Update header comment.
19032 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19034 * config/aarch64/arm_neon.h
19035 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19036 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19037 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19038 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19039 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19040 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19041 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19042 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19044 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19046 * config/aarch64/aarch64-simd.md
19047 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19048 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19049 fix_trunc, fixuns_trunc.
19050 (ftrunc<VDQF:mode>2): New.
19051 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19052 (fix_trunc_optab): New.
19054 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19056 * config/aarch64/aarch64-builtins.c
19057 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19058 iceilf, lround, iroundf.
19060 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19063 * config/i386/i386.h (enum ix86_tune_indices)
19064 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19065 New, split from X86_TUNE_INTER_UNIT_MOVES.
19066 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19067 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19068 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19069 (TARGET_INTER_UNIT_MOVES): Remove.
19070 * config/i386/i386.c (initial_ix86_tune_features): Update.
19071 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19072 (ix86_expand_convert_uns_didf_sse): Use
19073 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19074 (ix86_expand_vector_init_one_nonzero): Ditto.
19075 (ix86_expand_vector_init_interleave): Ditto.
19076 (inline_secondary_memory_needed): Return true for moves from SSE class
19077 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19078 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19079 * config/i386/constraints.md (Yi, Ym): Depend on
19080 TARGET_INTER_UNIT_MOVES_TO_VEC.
19081 (Yj, Yn): New constraints.
19082 * config/i386/i386.md (*movdi_internal): Change constraints of
19083 operand 1 from Yi to Yj and from Ym to Yn.
19084 (*movsi_internal): Ditto.
19085 (*movdf_internal): Ditto.
19086 (*movsf_internal): Ditto.
19087 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19088 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19089 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19090 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19091 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19092 * config/i386/sse.md (movdi_to_sse): Ditto.
19093 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19094 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19095 TARGET_INTER_UNIT_MOVES.
19096 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19097 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19098 instead of TARGET_INTER_UNIT_MOVES.
19099 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19100 operand 1 from Yi to Yj and from Ym to Yn.
19102 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19104 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19105 (float_truncate_hi_): Likewise.
19106 (float_extend_lo_): Likewise.
19107 (float_truncate_lo_): Likewise.
19108 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19109 (aarch64_float_extend_lo_v2df): Likewise.
19110 (vec_unpacks_hi_v4sf): Likewise.
19111 (aarch64_float_truncate_lo_v2sf): Likewise.
19112 (aarch64_float_truncate_hi_v4sf): Likewise.
19113 (vec_pack_trunc_v2df): Likewise.
19114 (vec_pack_trunc_df): Likewise.
19116 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19118 * config/aarch64/aarch64-builtins.c
19119 (aarch64_fold_builtin): Fold float conversions.
19120 * config/aarch64/aarch64-simd-builtins.def
19121 (floatv2si, floatv4si, floatv2di): New.
19122 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19123 * config/aarch64/aarch64-simd.md
19124 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19125 * config/aarch64/iterators.md (FLOATUORS): New.
19126 (optab): Add float, floatuns.
19127 (su_optab): Likewise.
19129 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19131 * config/aarch64/aarch64-builtins.c
19132 (aarch64_builtin_vectorized_function): Use new names for
19134 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19135 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19136 (fcvtzu): Split as...
19137 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19138 (fcvtas): Split as...
19139 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19140 (fcvtau): Split as...
19141 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19142 (fcvtps): Split as...
19143 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19144 (fcvtpu): Split as...
19145 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19146 (fcvtms): Split as...
19147 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19148 (fcvtmu): Split as...
19149 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19150 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19151 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19152 (lfrintnusf, lfrintnudf): Likewise.
19153 * config/aarch64/aarch64-simd.md
19154 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19156 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19157 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19158 (fcvt_pattern): Likewise.
19160 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19162 * config/aarch64/aarch64-simd.md
19163 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19164 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19166 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19168 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19169 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19170 (vrnd<a,m,n,p>_f32): Implement using builtins.
19171 (vrnd<i,x><q>_f<32, 64>): New.
19173 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19175 * config/aarch64/aarch64-builtins.c
19176 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19177 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19178 (frintz): Rename to...
19180 (frintp): Rename to...
19182 (frintm): Rename to...
19184 (frinti): Rename to...
19185 (nearbyint): ...this.
19186 (frintx): Rename to...
19188 (frinta): Rename to...
19190 * config/aarch64/aarch64-simd.md
19191 (aarch64_frint<frint_suffix><mode>): Delete.
19192 (<frint_pattern><mode>2): Convert to insn.
19193 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19194 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19195 (frint_pattern): Likewise.
19196 (frint_suffix): Likewise.
19198 2013-04-29 Richard Biener <rguenther@suse.de>
19200 PR tree-optimization/57081
19201 * loop-init.c: Include tree-flow.h.
19202 (loop_optimizer_finalize): Free number of iteration estimates.
19203 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19205 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19207 PR tree-optimization/57083
19208 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19209 non-singleton shift count range, zero extend low_bound for uns case.
19211 * config/i386/predicates.md (general_vector_operand): New predicate.
19212 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19213 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19214 if they aren't nonimmediate operands. If their original values
19215 satisfy const_vector_equal_evenodd_p, don't shift them.
19216 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19217 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19218 if not nonimmediate_operand.
19219 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19220 instead of register_operand.
19221 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19223 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19225 * stor-layout.c (finalize_size_functions): Allocate a structure and
19226 reset cfun before dumping the functions.
19228 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19230 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19233 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19234 use xop_pmacsdqh if uns_p.
19235 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19236 the immediate rotate count.
19238 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19240 * rtl.h (struct rtx_def): Add comment for field jump.
19241 (LRA_SUBREG_P): New macro.
19242 * recog.c (register_operand): Check LRA_SUBREG_P.
19243 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19245 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19246 (lra_final_code_change): Skip subreg change for operators.
19247 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19248 if there are no operand changes.
19249 * lra-constraints.c (curr_insn_set): New.
19250 (match_reload): Set LRA_SUBREG_P.
19251 (emit_spill_move): Ditto.
19252 (check_and_process_move): Use curr_insn_set. Process only single
19253 set insns. Don't initialize sec_mem_p and change_p.
19254 (simplify_operand_subreg): Use LRA_SUBREG_P.
19255 (reg_in_class_p): New function.
19256 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19257 of #ifdef. Add code to remove cycling.
19258 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19259 non-null disp. Reload inner instead of disp when base and index
19260 are null. Try to put lo_sum into register.
19261 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19262 (check_and_process_move): Move code for move cost check to
19263 simple_move_p. Remove equiv_substitution.
19264 (simple_move_p): New function.
19265 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19266 curr_insn_set. Call check_and_process_move only for single set
19267 insns. Use the new function. Move call of check_and_process_move
19268 after operand equiv substitution and address process.
19270 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19273 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19274 with nonlocal goto receivers or returns twice calls, ignore
19275 unininitialized values from abnormal edges to nl goto receiver
19276 or returns twice call.
19278 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19280 PR tree-optimization/57051
19281 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19282 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19285 2013-04-26 Richard Biener <rguenther@suse.de>
19287 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19288 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19289 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19290 (fixup_loop_arrays_after_move): New function.
19291 (move_sese_region_to_fn): Properly outline the loop tree parts
19292 of the SESE region.
19294 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19296 * config/i386/i386.md (type, unit): Fix long lines.
19298 2013-04-26 Richard Biener <rguenther@suse.de>
19300 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19301 (lto-streamer-out.o): Likewise.
19302 * cfgloop.c (init_loops_structure): Export, add struct function
19303 argument and adjust.
19304 (flow_loops_find): Adjust.
19305 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19306 (init_loops_structure): Declare.
19307 * lto-streamer-in.c: Include cfgloop.h.
19308 (input_cfg): Input the loop tree.
19309 * lto-streamer-out.c: Include cfgloop.h.
19310 (output_cfg): Output the loop tree.
19311 (output_struct_function_base): Do not drop PROP_loops.
19313 2013-03-26 Richard Biener <rguenther@suse.de>
19315 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19316 (pass_build_cfg): Provide PROP_loops.
19317 (move_sese_region_to_fn): Remove loops that are outlined into fn
19319 * tree-inline.c: Include cfgloop.h.
19320 (initialize_cfun): Do not drop PROP_loops.
19321 (copy_loops): New function.
19322 (copy_cfg_body): Copy loop structure.
19323 (tree_function_versioning): Initialize destination loop tree.
19324 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19325 (pass_parallelize_loops): Do IL verification.
19326 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19327 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19328 the CFG make sure we fixup loops as well.
19329 * tree-ssa-tail-merge.c: Include cfgloop.h.
19330 (replace_block_by): When merging loop latches mark loops for fixup.
19331 * lto-streamer-out.c (output_struct_function_base): Drop
19332 PROP_loops for now.
19333 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19334 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19335 * ipa-split.c: Include cfgloop.h.
19336 (split_function): Add the new return block to the loop tree root.
19337 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19338 whether we have removed the forwarder block.
19339 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19340 * cfgloop.h (place_new_loop): Declare.
19341 * cfgloopmanip.c (place_new_loop): Export.
19342 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19343 (tree-switch-conversion.o): Likewise.
19344 (tree-complex.o): Likewise.
19345 (tree-inline.o): Likewise.
19346 (tree-ssa-tailmerge.o): Likewise.
19347 (ipa-split.o): Likewise.
19348 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19349 (tree-ssa-copy.o): Likewise.
19350 * tree-switch-conversion.c: Include cfgloop.h
19351 (process_switch): If we emit a bit-test cascade, schedule loops
19353 * tree-complex.c: Include cfgloop.h.
19354 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19355 * asan.c: Include cfgloop.h.
19356 (create_cond_insert_point): Properly add new basic-blocks to
19357 loops, schedule loop fixup.
19358 * cfgloop.c (verify_loop_structure): Check that looks are not
19360 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19362 (expand_omp_for_generic): Likewise.
19363 (expand_omp_sections): Likewise.
19364 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19365 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19366 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19367 is initialized, not when loops are present.
19368 * tree-parloops.c (parallelize_loops): Remove checking here.
19369 * passes.c (init_optimization_passes): Schedule a copy-propagation
19370 pass before complete unrolling of inner loops.
19372 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19374 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19375 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19376 (colorize_init): Add argument to _WIN32 version.
19377 * toplev.c: Include diagnostic-color.h.
19378 (process_options): Default to -fdiagnostics-color=auto if
19379 GCC_COLORS env var is in the environment.
19380 * common.opt (fdiagnostics-color=): Add Var and Init.
19381 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19382 env var is in the environment, the default is auto rather than never.
19384 * diagnostic.h (file_name_as_prefix): Add context argument.
19385 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19386 the string as locus.
19387 * langhooks.c (lhd_print_error_function): Adjust caller.
19389 2013-04-25 Lawrence Crowl <crowl@google.com>
19391 * var-tracking.c (shared_hash_def::htab):
19392 Change type to hash_table. Update dependent calls and types.
19394 2013-04-25 Lawrence Crowl <crowl@google.com>
19396 * Makefile.in: Update as needed below.
19398 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19399 Move declaration to after the type's method definitons.
19401 * attribs.c (htab_t scoped_attributes::attribute_hash):
19402 Change type to hash_table. Update dependent calls and types.
19404 * bitmap.c (htab_t bitmap_desc_hash):
19405 Change type to hash_table. Update dependent calls and types.
19407 * cselib.c (htab_t cselib_hash_table):
19408 Change type to hash_table. Update dependent calls and types.
19410 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19411 (hash_string_slot_node): Move implementation into lto-streamer.h
19412 struct string_slot_hasher.
19413 (eq_string_slot_node): Likewise.
19415 * data-streamer-out.c: Update output_block::string_hash_table
19416 dependent calls and types.
19418 * dwarf2cfi.c (htab_t trace_index):
19419 Change type to hash_table. Update dependent calls and types.
19421 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19422 Change type to hash_table. Update dependent calls and types.
19423 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19424 (htab_t optimize_external_refs::map): Likewise.
19425 (htab_t output_comp_unit::extern_map): Likewise.
19426 (htab_t output_comdat_type_unit::extern_map): Likewise.
19427 (htab_t output_macinfo::macinfo_htab): Likewise.
19428 (htab_t optimize_location_lists::htab): Likewise.
19429 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19431 * except.c (htab_t ehspec_hash_type):
19432 Change type to hash_table. Update dependent calls and types.
19433 (assign_filter_values::ttypes): Likewise.
19434 (assign_filter_values::ehspec): Likewise.
19435 (sjlj_assign_call_site_values::ar_hash): Likewise.
19436 (convert_to_eh_region_ranges::ar_hash): Likewise.
19438 * gcse.c (htab_t pre_ldst_table):
19439 Change type to hash_table. Update dependent calls and types.
19441 * ggc-common.c (htab_t saving_htab):
19442 Change type to hash_table. Update dependent calls and types.
19443 (htab_t loc_hash): Likewise.
19444 (htab_t ptr_hash): Likewise.
19445 (call_count): Rename ggc_call_count.
19446 (call_alloc): Rename ggc_call_alloc.
19447 (loc_descriptor): Rename make_loc_descriptor.
19448 (add_statistics): Rename ggc_add_statistics.
19450 * ggc-common.c (saving_htab):
19451 Change type to hash_table. Update dependent calls and types.
19453 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19454 (push_gimplify_context): Likewise.
19455 (pop_gimplify_context): Likewise.
19456 (struct gimple_temp_hash_elt): Added.
19457 (struct gimplify_hasher): Likewise.
19458 (struct gimplify_ctx.temp_htab):
19459 Change type to hash_table. Update dependent calls and types.
19461 * gimple-fold.c: Include gimplify-ctx.h.
19463 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19464 Change type to hash_table. Update dependent calls and types.
19465 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19466 avoid potential global name collision.
19468 * gimplify.c: Include gimplify-ctx.h.
19469 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19470 (htab_t gimplify_ctx::temp_htab):
19471 Update dependent calls and types for new type hash_table.
19472 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19473 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19475 * gimplify-ctx.h: New.
19476 (struct gimple_temp_hash_elt): Move from gimplify.c.
19477 (class gimplify_hasher): New.
19478 (struct gimplify_ctx): Move from gimple.h.
19479 (htab_t gimplify_ctx::temp_htab):
19480 Change type to hash_table. Update dependent calls and types.
19482 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19483 (htab_t ivs_params::newivs_index):
19484 Change type to hash_table. Update dependent calls and types.
19485 (htab_t ivs_params::params_index): Likewise.
19486 (htab_t print_generated_program::params_index): Likewise.
19487 (htab_t gloog::newivs_index): Likewise.
19488 (htab_t gloog::params_index): Likewise.
19490 * graphite.c: Include graphite-htab.h.
19491 4htab_t graphite_transform_loops::bb_pbb_mapping):
19492 Change type to hash_table. Update dependent calls and types.
19494 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19495 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19496 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19498 * graphite-dependences.c: Include graphite-htab.h.
19499 (loop_is_parallel_p): Change hash table type of parameter.
19501 * graphite-htab.h: New.
19502 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19503 (extern find_pbb_via_hash): Move from graphite-poly.h.
19504 (extern loop_is_parallel_p): Move from graphite-poly.h.
19505 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19507 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19508 (extern loop_is_parallel_p): Move to graphite-htab.h.
19509 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19511 * haifa-sched.c (htab_t delay_htab):
19512 Change type to hash_table. Update dependent calls and types.
19513 (htab_t delay_htab_i2): Likewise.
19515 * ira-color.c (htab_t allocno_hard_regs_htab):
19516 Change type to hash_table. Update dependent calls and types.
19518 * ira-costs.c (htab_t cost_classes_htab):
19519 Change type to hash_table. Update dependent calls and types.
19521 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19522 Change type to hash_table. Update dependent calls and types.
19524 * loop-iv.c (htab_t bivs):
19525 Change type to hash_table. Update dependent calls and types.
19527 * loop-unroll.c (htab_t opt_info::insns_to_split):
19528 Change type to hash_table. Update dependent calls and types.
19529 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19531 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19532 (struct string_slot_hasher): New.
19533 (htab_t output_block::string_hash_table):
19534 Change type to hash_table. Update dependent calls and types.
19536 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19537 (htab_t file_name_hash_table):
19538 Change type to hash_table. Update dependent calls and types.
19540 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19543 * lto-streamer.c (htab_t tree_htab):
19544 Change type to hash_table. Update dependent calls and types.
19546 * omp-low.c: Include gimplify-ctx.h.
19548 * passes.c (htab_t name_to_pass_map):
19549 Change type to hash_table. Update dependent calls and types.
19550 (pass_traverse): Rename to passes_pass_traverse.
19552 * plugin.c (htab_t event_tab):
19553 Change type to hash_table. Update dependent calls and types.
19555 * postreload-gcse.c (htab_t expr_table):
19556 Change type to hash_table. Update dependent calls and types.
19557 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19559 * sese.c (debug_rename_map_1): Make extern.
19560 (htab_t copy_bb_and_scalar_dependences::rename_map):
19561 Change type to hash_table. Update dependent calls and types.
19563 * sese.h (extern debug_rename_map): Move to .c file.
19565 * store-motion.c (htab_t store_motion_mems_table):
19566 Change type to hash_table. Update dependent calls and types.
19568 * trans-mem.c (htab_t tm_new_mem_hash):
19569 Change type to hash_table. Update dependent calls and types.
19571 * tree-browser.c (htab_t TB_up_ht):
19572 Change type to hash_table. Update dependent calls and types.
19574 * tree-cfg.c (htab_t discriminator_per_locus):
19575 Change type to hash_table. Update dependent calls and types.
19577 * tree-complex.c: Include tree-hasher.h
19578 (htab_t complex_variable_components):
19579 Change type to hash_table. Update dependent calls and types.
19581 * tree-eh.c (htab_t finally_tree):
19582 Change type to hash_table. Update dependent calls and types.
19584 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19585 struct int_tree_hasher.
19586 (extern int_tree_map_eq): Likewise.
19587 (uid_decl_map_hash): Removed.
19588 (extern decl_tree_map_eq): Likewise.
19590 * tree-hasher.h: New.
19591 (struct int_tree_hasher): New.
19592 (typedef int_tree_htab_type): New.
19594 * tree-inline.c: Include gimplify-ctx.h.
19596 * tree-mudflap.c: Include gimplify-ctx.h.
19598 * tree-parloops.c: Include tree-hasher.h.
19599 (htab_t eliminate_local_variables_stmt::decl_address):
19600 Change type to hash_table. Update dependent calls and types.
19601 (htab_t separate_decls_in_region::decl_copies): Likewise.
19603 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19604 Change type to hash_table. Update dependent calls and types.
19606 * tree-sra.c (candidates):
19607 Change type to hash_table. Update dependent calls and types.
19609 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19611 (int_tree_map_hash): Likewise.
19613 * tree-ssa-dom.c (htab_t avail_exprs):
19614 Change type to hash_table. Update dependent calls and types.
19616 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19617 Change type to hash_table. Update dependent calls and types.
19619 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19620 Change type to hash_table. Update dependent calls and types.
19622 * tree-ssa-phiopt.c (seen_ssa_names):
19623 Change type to hash_table. Update dependent calls and types.
19625 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19626 Change type to hash_table. Update dependent calls and types.
19628 * tree-ssa-uncprop.c (equiv):
19629 Change type to hash_table. Update dependent calls and types.
19631 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19633 PR rtl-optimization/57003
19634 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19635 call note_stores with kill_clobbered_value callback again after
19636 killing regs_invalidated_by_call.
19638 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19640 * config/aarch64/aarch64-simd.md
19641 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19642 (aarch64_simd_bsl<mode>): Likewise.
19643 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19645 2013-04-25 Marek Polacek <polacek@redhat.com>
19647 PR tree-optimization/57066
19648 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19650 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
19652 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19654 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19656 * config/aarch64/aarch64-builtins.c
19657 (aarch64_fold_builtin): New.
19658 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19659 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19660 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19661 * config/aarch64/arm_neon.h
19662 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19664 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19665 Tejas Belagod <tejas.belagod@arm.com>
19667 * config/aarch64/aarch64-builtins.c
19668 (aarch64_gimple_fold_builtin): New.
19669 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19670 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19671 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19672 (addvv4sf): Update.
19673 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19675 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19677 * config/aarch64/aarch64.md
19678 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19680 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19682 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19683 (*ngcsi_uxtw): New pattern.
19685 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19686 Julian Brown <julian@codesourcery.com>
19688 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19689 (TB_DREG): Add T_V4HF.
19690 (v4hf_UP): New macro.
19691 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19692 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19693 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19695 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19696 (arm_vector_mode_supported_p): Handle V4HF.
19697 (arm_mangle_map): Handle V4HFmode.
19698 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19699 * config/arm/arm_neon_builtins.def: Add entries for
19700 vcvtv4hfv4sf, vcvtv4sfv4hf.
19701 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19702 (neon_vcvtv4hfv4sf): Likewise.
19703 * config/arm/neon-gen.ml: Handle half-precision floating point
19705 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19706 * config/arm/arm_neon.h: Regenerate.
19707 * config/arm/neon.ml (type elts): Add F16.
19708 (type vectype): Add T_float16x4, T_floatHF.
19709 (type vecmode): Add V4HF.
19710 (type features): Add Requires_FP_bit feature.
19711 (elt_width): Handle F16.
19712 (elt_class): Likewise.
19713 (elt_of_class_width): Likewise.
19714 (mode_of_elt): Refactor.
19715 (type_for_elt): Handle F16, fix error messages.
19716 (vectype_size): Handle T_float16x4.
19717 (vcvt_sh): New function.
19718 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19719 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19720 (string_of_mode): Handle V4HF.
19721 * doc/arm-neon-intrinsics.texi: Regenerate.
19723 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19725 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19726 format specifier in 'X' case.
19728 2013-04-25 Alan Modra <amodra@gmail.com>
19731 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19732 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19733 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19734 Repeat for many other rotate/shift and mask patterns using subregs.
19735 Name lshiftrt insns.
19736 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19737 on WORDS_BIG_ENDIAN.
19739 2013-04-25 Alan Modra <amodra@gmail.com>
19741 * config.gcc: Support little-endian powerpc-linux targets.
19742 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19743 (LINK_OS_LINUX_SPEC): Define.
19744 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19745 Preserve MASK_LITTLE_ENDIAN.
19746 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19747 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19748 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19749 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19750 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19751 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19752 Correct fp word order for little-endian. Don't shift toc entries
19753 smaller than a word for little-endian.
19754 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19755 (bswapdi2 splits): Correct low-part subreg for little-endian.
19756 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19757 low/high where such is correct only for be.
19758 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19759 little-endian for -mcall-aixdesc.
19761 2013-04-25 Alan Modra <amodra@gmail.com>
19763 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19764 replace_equiv_address_nv.
19766 2013-04-25 Alan Modra <amodra@gmail.com>
19768 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19770 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19773 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19774 * rtl.h (struct rtx_def): ...
19776 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19778 PR rtl-optimizations/57046
19779 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19780 for multi-reg splits.
19782 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19784 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19786 2013-04-24 Sterling Augustine <saugustine@google.com>
19788 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19789 (comp_dir_string, debug_str_dwo_section): New.
19790 (DEBUG_STR_DWO_SECTION): Rename to ...
19791 (DEBUG_DWO_STR_SECTION): ... this.
19792 (DEBUG_NORM_STR_SECTION): Delete.
19793 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19794 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19795 (find_AT_string): Move most logic to ...
19796 (find_AT_string_in_table): ... here. New.
19797 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19798 add_skeleton_AT_string. Delete logic.
19799 (output_skeleton_debug_sections): Remove call to
19800 add_top_level_skeleton_die_attrs.
19801 (add_comp_dir_attribute): Move logic to comp_dir_string.
19802 (dwarf2out_init): Initialize debug_str_dwo_section.
19803 (output_indirect_string): Call find_string_form.
19804 (output_indirect_strings): Rewrite.
19805 (prune_unused_types): Empty skeleton_debug_str_hash.
19806 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19807 (dwarf2out_finish): Call output_indirect_strings.
19809 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19811 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19813 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19815 * rtl.h (struct rtx_def): Add comment for field jump.
19816 (LRA_SUBREG_P): New macro.
19817 * recog.c (register_operand): Check LRA_SUBREG_P.
19818 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19820 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19821 (lra_final_code_change): Skip subreg change for operators.
19822 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19823 if there are no operand changes.
19824 * lra-constraints.c (curr_insn_set): New.
19825 (match_reload): Set LRA_SUBREG_P.
19826 (emit_spill_move): Ditto.
19827 (check_and_process_move): Use curr_insn_set. Process only single
19828 set insns. Don't initialize sec_mem_p and change_p.
19829 (simplify_operand_subreg): Use LRA_SUBREG_P.
19830 (reg_in_class_p): New function.
19831 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19832 of #ifdef. Add code to remove cycling.
19833 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19834 non-null disp. Reload inner instead of disp when base and index
19835 are null. Try to put lo_sum into register.
19836 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19837 (check_and_process_move): Move code for move cost check to
19838 simple_move_p. Remove equiv_substitution.
19839 (simple_move_p): New function.
19840 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19841 curr_insn_set. Call check_and_process_move only for single set
19842 insns. Use the new function. Move call of check_and_process_move
19843 after operand equiv substitution and address process.
19845 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19847 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19848 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19849 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19851 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19853 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19855 2013-04-24 Marek Polacek <polacek@redhat.com>
19857 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19858 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19859 (select_loops_exit_conditions): Likewise.
19860 (number_of_iterations_for_all_loops): Likewise.
19861 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19862 (scev_analysis): Likewise.
19864 2013-04-02 Catherine Moore <clm@codesourcery.com>
19865 Chao-ying Fu <fu@mips.com>
19867 * config/mips/micromips.md (jraddiusp): New pattern.
19868 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19869 instruction if possible.
19871 2013-04-24 Alan Modra <amodra@gmail.com>
19873 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19875 2013-04-24 Julian Brown <julian@codesourcery.com>
19876 Chung-Lin Tang <cltang@codesourcery.com>
19878 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19879 dependency behavior in enumeration type DIE generation. Add TODO note
19880 to comments about future DW_FORM_sdata/udata re-work of related code.
19882 2013-04-23 Lawrence Crowl <crowl@google.com>
19884 * Makefile.in: Update as needed below.
19886 * hash-table.h (class hash_table):
19887 Correct many methods with parameter types compare_type to the correct
19888 value_type. (Correct code was unlikely to notice the change.)
19889 (hash_table::elements_with_deleted) New.
19890 (class hashtable::iterator): New.
19891 (hashtable::begin()): New.
19892 (hashtable::end()): New.
19893 (FOR_EACH_HASH_TABLE_ELEMENT): New.
19895 * statistics.c (statistics_hashes):
19896 Change type to hash_table. Update dependent calls and types.
19898 * tree-into-ssa.c (var_infos):
19899 Change type to hash_table. Update dependent calls and types.
19901 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
19902 Change type to hash_table. Update dependent calls and types.
19904 * tree-ssa-loop-im.c (struct mem_ref.refs):
19905 Change type to hash_table. Update dependent calls and types.
19907 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
19908 Change type to hash_table. Update dependent calls and types.
19910 * tree-ssa-sccvn.c (vn_tables_s::nary):
19911 Change type to hash_table. Update dependent calls and types.
19912 (vn_tables_s::phis): Likewise.
19913 (vn_tables_s::references): Likewise.
19915 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
19916 (vn_reference_eq): Update parameter and return types.
19918 * tree-ssa-structalias.c (pointer_equiv_class_table):
19919 Change type to hash_table. Update dependent calls and types.
19920 (location_equiv_class_table): Likewise.
19922 * tree-vect-data-refs.c: Consequential changes for making
19923 peeling a hash_table.
19925 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
19926 (destroy_loop_vec_info): Dependent hash_table update.
19928 * tree-vectorizer.h (peeling_htab):
19929 Change type to hash_table. Update dependent calls and types.
19931 2013-04-23 Shiva Chen <shiva0217@gmail.com>
19933 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
19934 to check the register content is equal or not.
19935 * lra-constraints.c (match_reload): Use lra_assign_reg_val
19936 to assign register content record.
19937 * lra-eliminations.c (update_reg_eliminate): Use
19938 lra_update_reg_val_offset to update register content offset.
19939 * lra-int.h (struct lra_reg): Add offset member.
19940 (lra_reg_val_equal_p): New static inline function.
19941 (lra_update_reg_val_offset): New static inline function.
19942 (lra_assign_reg_val): New static inline function.
19943 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
19944 to assign register content record.
19945 (initialize_lra_reg_info_element): Initial offset to zero.
19947 2013-04-23 Catherine Moore <clm@codesourcery.com>
19949 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
19950 operands. Record compression.
19952 2013-04-23 Xinliang David Li <davidxl@google.com>
19954 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
19956 2013-04-23 Richard Biener <rguenther@suse.de>
19958 PR middle-end/57036
19959 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
19960 parameter, only add abnormal goto edges from the copied body
19961 if the call could perform abnormal gotos.
19962 (copy_cfg_body): Adjust.
19964 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
19966 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
19968 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
19970 * coretypes.h (gimple_stmt_iterator): Add struct to make
19973 2013-04-23 Richard Biener <rguenther@suse.de>
19975 PR tree-optimization/57026
19976 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19977 from SSA names occuring in abnormal PHI nodes.
19979 2013-04-22 Andi Kleen <ak@linux.intel.com>
19981 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19983 2013-04-22 Andi Kleen <ak@linux.intel.com>
19985 * lto/lto.c (print_lto_report_1): Declare early.
19986 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19988 2013-04-22 Andi Kleen <ak@linux.intel.com>
19990 * common.opt (-flto-report-wpa): Add.
19991 * doc/invoke.texi (-flto-report-wpa): Add.
19992 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19995 2013-04-22 Xinliang David Li <davidxl@google.com>
19997 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19998 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19999 * Makefile.in: New dependency
20001 David Daney <ddaney.cavm@gmail.com>
20003 * configure.ac (gcc_cv_as_micromips_support): Use the
20004 --fatal-warnings option.
20005 * configure: Regenerate.
20007 2013-04-22 Marek Polacek <polacek@redhat.com>
20010 * tsan.c (instrument_expr): Don't instrument expression
20011 in case its size is zero.
20013 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20017 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20019 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20021 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20023 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20024 (gimple_stmt_iterator): New typedef.
20025 * gimple.h (gimple_stmt_iterator): Rename to...
20026 (gimple_stmt_iterator_d): ... This.
20027 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20028 trees be valid for GIMPLE and GENERIC.
20029 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20030 * gimple-fold.c (gimple_fold_call): Call target hook
20031 gimple_fold_builtin.
20032 * hooks.c (hook_bool_gsiptr_false): New.
20033 * hooks.h (hook_bool_gsiptr_false): New.
20034 * target.def (fold_stmt): New.
20035 * doc/tm.texi: Regenerate.
20037 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20040 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20041 a set sp if no stack realignment.
20043 2013-04-22 Nick Clifton <nickc@redhat.com>
20045 * config.gcc (tilegx-linux): Extend extra_objs rather than
20047 (tilepro-linux): Likewise.
20049 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20051 * config/aarch64/aarch64-builtins.c
20053 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20054 (VAR<1-12>): Add MAP parameter.
20055 (BUILTIN_*): Likewise.
20056 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20057 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20058 (aarch64_ushl_n<mode>): Likewise.
20059 (aarch64_sshr_n<mode>): Likewise.
20060 (aarch64_ushr_n<mode>): Likewise.
20061 (aarch64_<maxmin><mode>): Likewise.
20062 (aarch64_sqrt<mode>): Likewise.
20063 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20064 (vshr<q>_n_*): Likewise.
20066 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20068 * config/aarch64/aarch64-builtins.c
20069 (aarch64_simd_builtin_type_mode): Handle SF types.
20071 (BUILTIN_GPF): Define.
20072 (aarch64_init_simd_builtins): Handle SF types.
20073 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20074 (frecps): Likewise.
20075 (frecpx): Likewise.
20076 * config/aarch64/aarch64-simd.md
20077 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20078 (aarch64_frecpe<mode>): New.
20079 (aarch64_frecps<mode>): Likewise.
20080 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20081 (v8type): Add frecp<esx>.
20082 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20083 (aarch64_frecps<mode>): Likewise.
20084 * config/aarch64/iterators.md (FRECP): New.
20085 (frecp_suffix): Likewise.
20086 * config/aarch64/arm_neon.h
20087 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20089 2013-04-22 Christian Bruel <christian.bruel@st.com>
20092 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20093 (REG_CLASS_NAMES): Idem.
20094 (REG_CLASS_CONTENTS): Idem.
20095 (REGCLASS_HAS_FP_REG): Idem.
20096 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20097 (sh_conditional_register_usage): Idem.
20099 2013-04-21 Jeff Law <law@redhat.com>
20101 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20102 (ssa_forward_propagate_and_combine): Use it.
20104 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20106 * lra.c: Update the flow chart diagram.
20108 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20110 PR rtl-optimization/56847
20111 * lra-constraints.c (process_alt_operands): Discourage alternative
20112 with non-matche doffsettable memory constraint fro memory with
20115 2013-04-19 Richard Biener <rguenther@suse.de>
20117 PR tree-optimization/56982
20118 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20120 * gimplify.c (gimplify_call_expr): Notice special calls.
20121 (gimplify_modify_expr): Likewise.
20122 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20123 abnormal control flow receivers.
20124 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20125 in the same way as cfun->has_nonlocal_labels.
20126 (gimple_purge_dead_abnormal_call_edges): Likewise.
20127 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20128 receivers start a basic-block.
20130 2013-04-19 Richard Biener <rguenther@suse.de>
20132 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20134 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20135 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20136 (SLP_TREE_LOAD_PERMUTATION): Add.
20137 (vect_transform_slp_perm_load): Adjust prototype.
20138 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20139 (vect_free_slp_instance): Likewise.
20140 (vect_create_new_slp_node): Likewise.
20141 (vect_supported_slp_permutation_p): Remove.
20142 (vect_slp_rearrange_stmts): Adjust.
20143 (vect_supported_load_permutation_p): Likewise. Inline
20144 vect_supported_slp_permutation_p here.
20145 (vect_analyze_slp_instance): Compute load permutations per
20146 slp node instead of per instance.
20147 (vect_get_slp_defs): Adjust.
20148 (vect_transform_slp_perm_load): Likewise.
20149 (vect_schedule_slp_instance): Remove redundant code.
20150 (vect_schedule_slp): Remove hack for PR56270, add it ...
20151 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20152 CSE loads for SLP. Adjust.
20154 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20156 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20157 spelling in two comments.
20159 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20162 * config/arm/arm.c (load_multiple_sequence): Require SP
20163 as base register for loads if SP is in the register list.
20165 2013-04-19 Martin Jambor <mjambor@suse.cz>
20167 PR tree-optimization/56718
20168 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20169 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20170 and made public. Adjusted all callers.
20171 (ipa_intraprocedural_devirtualization): New function.
20172 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20173 (ipa_intraprocedural_devirtualization): Likewise.
20174 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20176 2013-04-19 Richard Biener <rguenther@suse.de>
20178 PR tree-optimization/57000
20179 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20181 2013-04-19 Terry Guo <terry.guo@arm.com>
20183 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20185 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20186 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20187 (cortex_m4_fmacs): Use new reservations.
20188 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20190 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20192 PR rtl-optimization/56999
20193 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20195 (lra_coalesce): Remove split_origin_bitmap and related code.
20196 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20197 ranges if necessary.
20199 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20201 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20203 (ix86_expand_call): Remove clobbered_registers array and use
20204 x86_64_ms_sysv_extra_clobbered_registers instead.
20205 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20207 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20209 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20210 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20211 (*call_value_rex64_ms_sysv): Ditto.
20213 2013-04-18 Cary Coutant <ccoutant@google.com>
20215 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20216 parent instead of die_mark.
20218 2013-04-18 Diego Novillo <dnovillo@google.com>
20220 * gimple.c (create_gimple_tmp): New.
20221 (get_expr_type): New.
20222 (build_assign): New.
20223 (build_type_cast): New.
20224 * gimple.h (enum ssa_mode): Define.
20225 (gimple_seq_set_location): New.
20226 * asan.c (build_check_stmt): Change some gimple_build_* calls
20227 to use build_assign and build_type_cast.
20229 2013-04-18 Richard Biener <rguenther@suse.de>
20231 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20232 handle negative step. Remove redundant checks.
20233 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20234 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20235 for negative step and grouped loads fail to vectorize.
20237 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20239 * emit-rtl.c (reset_insn_used_flags): New function.
20240 (reset_all_used_flags): Use it.
20241 (verify_insn_sharing): New function.
20242 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20244 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20246 PR tree-optimization/56984
20247 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20248 and (x >> M) >= N don't register any assertion if N << M is the
20251 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20253 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20254 do not delete_insn insns that have not yet been emitted, only
20255 unlink them with remove_insn.
20256 * df-scan.c (df_insn_delete): Revert r197492.
20258 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20260 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20261 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20263 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20265 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20266 define_insn_and_split.
20267 (and_scc,ior_scc,negscc): Likewise.
20268 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20270 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20272 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20273 can benefit from using a sequence of LDRD instructions in epilogue
20274 instead of a single LDM instruction.
20276 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20279 * doc/extend.texi: Fix typo.
20281 2013-04-17 Richard Biener <rguenther@suse.de>
20283 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20284 (vect_build_slp_tree): ... here.
20285 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20286 match. Remove special-casing of mismatched complex loads.
20287 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20288 re-try the match with swapped commutative operands.
20289 (vect_supported_load_permutation_p): Remove special-casing of
20290 mismatched complex loads.
20291 (vect_analyze_slp_instance): Adjust.
20293 2013-04-17 Richard Biener <rguenther@suse.de>
20295 PR rtl-optimization/56921
20296 * cfgloop.h (struct loop): Add simple_loop_desc member.
20297 (struct niter_desc): Mark with GTY(()).
20298 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20299 * loop-iv.c (get_simple_loop_desc): Likewise.
20300 (free_simple_loop_desc): Likewise.
20303 2013-04-16 Richard Biener <rguenther@suse.de>
20305 PR rtl-optimization/56921
20306 * loop-init.c (pass_rtl_move_loop_invariants): Add
20307 TODO_do_not_ggc_collect to todo_flags_finish.
20308 (pass_rtl_unswitch): Same.
20309 (pass_rtl_unroll_and_peel_loops): Same.
20310 (pass_rtl_doloop): Same.
20312 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20314 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20315 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20316 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20318 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20319 * tree-streamer.c (record_common_node): Adjust reference in comment.
20321 2013-04-17 Terry Guo <terry.guo@arm.com>
20323 * config/arm/cortex-m4.md: Add a new bypass.
20325 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20327 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20329 (*subs_<optab><mode>_multp2): New pattern.
20330 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20331 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20333 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20335 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20336 (*subs_mul_imm_<mode>): New pattern.
20338 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20341 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20342 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20343 (vsx_movti_32bit): Change j->wa to O->wa.
20345 2013-04-16 Richard Biener <rguenther@suse.de>
20347 PR rtl-optimization/56921
20348 * loop-init.c (pass_rtl_move_loop_invariants): Add
20349 TODO_do_not_ggc_collect to todo_flags_finish.
20350 (pass_rtl_unswitch): Same.
20351 (pass_rtl_unroll_and_peel_loops): Same.
20352 (pass_rtl_doloop): Same.
20354 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20356 * config/arm/arm.c (emit_multi_reg_push): New declaration
20357 for an existing function.
20358 (arm_emit_strd_push): New function.
20359 (arm_expand_prologue): Used here.
20360 (arm_emit_ldrd_pop): New function.
20361 (arm_expand_epilogue): Used here.
20362 (arm_get_frame_offsets): Update condition.
20363 (arm_emit_multi_reg_pop): Add a special case for load of a single
20364 register with writeback.
20366 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20368 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20371 2013-04-16 Richard Biener <rguenther@suse.de>
20373 PR tree-optimization/56756
20374 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20375 (first_mem_ref_loc): New.
20376 (execute_sm): Place the load temporarily before a previous
20377 access instead of in the latch edge to ensure its SSA dependencies
20378 are defined at points dominating the load.
20380 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20382 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20383 correct fix by moving header and footer insn to the footer of
20384 the merged basic block. Clear BB_END of the merged-away block.
20386 PR middle-end/43631
20387 * emit-rtl.c (make_note_raw): New function.
20388 (link_insn_into_chain): New static inline function.
20389 (add_insn): Use it.
20390 (add_insn_before, add_insn_after): Factor insn chain linking code...
20391 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20392 using link_insn_into_chain.
20393 (note_outside_basic_block_p): New helper function for emit_note_after
20394 and emit_note_before.
20395 (emit_note_after): Use nobb variant of add_insn_after if the note
20396 should not be contained in a basic block.
20397 (emit_note_before): Use nobb variant of add_insn_before if the note
20398 should not be contained in a basic block.
20399 (emit_note_copy): Use make_note_raw.
20400 (emit_note): Likewise.
20401 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20402 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20403 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20404 the moved barrier the tail of the basic block it follows.
20405 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20407 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20409 PR tree-optimization/56962
20410 * gimple-ssa-strength-reduction.c (record_increment): Only set
20411 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20412 either rhs1 or rhs2 is equal to c->base_expr.
20414 2013-04-15 Richard Biener <rguenther@suse.de>
20416 PR tree-optimization/56933
20417 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20419 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20420 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20421 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20422 dependence check ...
20423 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20425 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20426 GROUP_READ_WRITE_DEPENDENCE.
20428 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20430 * emit-rtl.c (reset_all_used_flags): New function.
20431 (verify_rtl_sharing): Call reset_all_used_flags before and after
20432 performing the checks.
20434 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20436 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20437 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20438 * config/arm/constraints.md (De): New constraint.
20439 * config/arm/neon.md (anddi3_neon): Delete.
20440 (neon_vand<mode>): Expand to standard anddi3 pattern.
20441 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20442 Move earlier in the file.
20443 (neon_inv_logic_op2): Likewise.
20444 (arm_anddi_operand_neon): New predicate.
20446 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20448 * configure.ac (gcc_cv_ld_as_needed): Set
20449 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20450 Use -z ignore, -z record on *-*-solaris2*.
20451 (HAVE_LD_AS_NEEDED): Update comment.
20452 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20453 * configure: Regenerate.
20454 * config.in: Regenerate.
20455 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20456 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20457 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20458 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20459 equivalents. Fix markup.
20460 * doc/tm.texi: Regenerate.
20462 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20464 * config/i386/i386.opt: New option mstack-protector-guard=.
20465 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20466 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20467 TARGET_SSP_TLS_GUARD.
20468 * config/i386/i386.c (ix86_option_override_internal): Set
20469 ix86_stack_protector_guard.
20470 * config/i386/i386.md (stack_protect_set): Enable for
20471 TARGET_SSP_TLS_GUARD only.
20472 (stack_protect_set_<mode>): Ditto.
20473 (stack_protect_test): Ditto.
20474 (stack_protect_test_<mode>): Ditto.
20475 * doc/invoke.texi (i386 Option): Document.
20477 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20480 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20481 (S_MODES): Set H_MODE bit.
20482 (SF_MODES): Set only S_MODE and SF_MODE bits.
20483 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20484 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20485 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20486 <MODE_FLOAT>: Likewise.
20488 2013-04-15 Joey Ye <joey.ye@arm.com>
20490 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20492 2013-04-15 Joey Ye <joey.ye@arm.com>
20494 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20496 (thumb_far_jump_used_p): Count instruction size and set
20499 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20501 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20502 * resource.c (mark_target_live_regs): Reformat conditional block.
20504 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20506 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20507 notes, they are emitted only just before final.
20508 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20510 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20512 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20513 * cfgrtl.c (delete_insn): Call it here instead.
20514 * lra-spills.c (lra_final_code_change): Use delete_insn.
20515 * haifa-sched.c (sched_remove_insn): Likewise.
20516 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20517 returning to the nop pool.
20518 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20519 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20521 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20523 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20524 * doc/tm.texi: Regenerated.
20526 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20528 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20531 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20533 * df-core.c (df_find_def): Compare register numbers.
20534 (df_find_use): Likewise.
20536 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20539 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20540 lra_in_progress for return.
20542 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20544 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20545 define_insn into define_insn_and_split and emit movsicc patterns.
20547 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20549 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20551 2013-04-12 Richard Biener <rguenther@suse.de>
20553 * tree-pass.h (TODO_do_not_ggc_collect): New.
20554 * passes.c (execute_one_ipa_transform_pass): Honor
20555 TODO_do_not_ggc_collect.
20556 (execute_one_pass): Likewise.
20559 2013-04-10 Richard Biener <rguenther@suse.de>
20561 * passes.c (init_optimization_passes): Remove reload pass.
20562 * ira.c (do_reload): Merge into ...
20564 (rest_of_handle_reload): Remove.
20565 (pass_reload): Likewise.
20566 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20567 of reload for vzeroupper pass placement.
20569 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20571 PR tree-optimization/56918
20572 PR tree-optimization/56920
20573 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20574 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20575 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20576 use wide_mul_with_sign method.
20578 2013-04-12 Richard Biener <rguenther@suse.de>
20580 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20581 not be considered a gimple constant.
20583 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20585 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20586 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20589 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20590 Jakub Jelinek <jakub@redhat.com>
20592 * opts.c: Include diagnostic-color.h.
20593 (common_handle_option): Handle OPT_fdiagnostics_color_.
20594 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20595 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20596 (diagnostic-color.o): New.
20597 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20598 (diagnostic_color_rule): New enum.
20599 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20600 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20601 the location string.
20602 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20603 either NULL, or color kind.
20604 * diagnostic-color.c: New file.
20605 * diagnostic-color.h: New file.
20606 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20608 * doc/invoke.texi (-fdiagnostics-color): Document.
20609 * pretty-print.h (pp_show_color): Define.
20610 (struct pretty_print_info): Add show_color field.
20611 * diagnostic.c: Include diagnostic-color.h.
20612 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20613 macros. Colorize error:, warning: etc. strings and also the location
20615 (diagnostic_show_locus): Colorize the caret line.
20616 * pretty-print.c: Include diagnostic-color.h.
20617 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20618 inside of %< %> quotes or quoted through q format modifier.
20620 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20622 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20624 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20626 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20627 code in CC_NZ mode.
20628 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20631 2013-04-11 Marek Polacek <polacek@redhat.com>
20633 PR tree-optimization/48184
20634 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20636 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20638 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20639 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20640 (skip_simple_arithmetic): Tidy up.
20641 * tree.h (skip_simple_constant_arithmetic): Declare.
20643 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20645 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20647 2013-04-11 Richard Biener <rguenther@suse.de>
20649 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20650 generate vector constants.
20652 2013-04-11 Richard Biener <rguenther@suse.de>
20654 PR tree-optimization/56878
20655 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20656 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20658 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20659 Prefer to align the DR with the most invariant base address.
20661 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20663 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20666 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20668 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20669 floating-point vector comparisons against 0.
20671 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20673 PR tree-optimization/56899
20674 * fold-const.c (extract_muldiv_1): Apply distributive law
20675 only if TYPE_OVERFLOW_WRAPS (ctype).
20677 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20680 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20681 loading has side effect.
20683 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20685 * config/sparc/sparc.c: Include tree-pass.h.
20686 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20687 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20688 head of file. Change return type. Split off gate function.
20689 (sparc_gate_work_around_errata): New function.
20690 (pass_work_around_errata): New pass definition.
20691 (insert_pass_work_around_errata) New pass insert definition to
20692 insert pass_work_around_errata just after delayed-branch scheduling.
20693 (sparc_option_override): Insert the pass.
20694 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20696 2013-04-10 David S. Miller <davem@davemloft.net>
20698 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20699 or -mcpu=hypersparc.
20701 * target.def (cstore_mode): New hook.
20702 * target.h: Include insn-codes.h
20703 * targhooks.c: Likewise.
20704 (default_cstore_mode): New function.
20705 * targhooks.h: Declare it.
20706 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20707 * doc/tm.texi: Rebuild.
20708 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20709 target hook, rather than inspecting the insn_data.
20710 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20711 (TARGET_CSTORE_MODE): Redefine.
20712 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20714 * config/sparc/predicates.md (cstore_result_operand): New special
20716 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20717 Use it for operand 0.
20718 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20719 (*snesi_special): Likewise.
20720 (*snesi_zero): Likewise.
20721 (*seqsi_zero): Likewise.
20722 (*sltu_insn): Likewise.
20723 (*sgeu_insn): Likewise.
20724 (*seqdi_special): Make operand 0 and comparison operation be of
20726 (*snedi_special): Likewise.
20727 (*snedi_special_vis3): Likewise.
20728 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20729 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20730 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20731 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20732 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20733 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20734 (*sltu_extend_sp64): Likewise.
20735 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20736 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20737 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20738 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20739 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20741 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20743 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20744 (aarch64_start_file): Use the new function.
20746 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20747 Jason Merrill <jason@redhat.com>
20749 * common.opt: Add -gdwarf.
20750 * opts.c (common_handle_option): Handle it.
20751 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20753 2013-04-10 Richard Biener <rguenther@suse.de>
20755 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20756 (execute_one_ipa_transform_pass): But unconditionally here.
20757 (execute_one_pass): And here.
20758 (init_optimization_passes): Remove reload pass.
20759 * tree-pass.h (TODO_ggc_collect): Remove.
20760 (pass_reload): Likewise.
20761 * ira.c (do_reload): Merge into ...
20763 (rest_of_handle_reload): Remove.
20764 (pass_reload): Likewise.
20765 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20766 of reload for vzeroupper pass placement.
20767 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20768 and todo_flags_finish of all passes.
20770 2013-04-10 Richard Biener <rguenther@suse.de>
20772 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20773 first_const_oprnd field, rename first_def_type to first_op_type.
20774 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20775 (vect_get_and_check_slp_defs): Always use the type of the
20776 operand. Allow mixed vect_external_def, vect_constant_def types.
20777 (vect_get_constant_vectors): Handle mixed vect_external_def,
20778 vect_constant_def types.
20780 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20782 PR tree-optimization/55524
20783 * tree-ssa-math-opts.c
20784 (convert_mult_to_fma): Don't use an fms construct
20785 when we don't have an fms operation, but fnma, and it looks
20786 likely that we'll be able to use the latter.
20788 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20790 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20792 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20793 inline fail caused by overwritable functions.
20795 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20797 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20798 unnecessary bits in the constant power of two case.
20800 2013-04-10 Richard Biener <rguenther@suse.de>
20802 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20803 broken code swapping operands.
20804 (vect_build_slp_tree): Do not compute load permutations here.
20805 (vect_analyze_slp_instance): Compute load permutations here,
20806 after building the SLP tree.
20808 2013-04-09 Christian Bruel <christian.bruel@st.com>
20810 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20811 of next/prev_real_insn.
20813 2013-04-09 Jan Hubicka <jh@suse.cz>
20815 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20816 Drop aliased parameter.
20817 (function_and_variable_visibility): Do not handle alias pairs.
20818 * cgraph.c (varpool_externally_visible_p): Update prototype.
20819 * varpool.c (varpool_add_new_variable): Update.
20821 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20823 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20825 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20827 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20829 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20831 2013-04-09 Marek Polacek <polacek@redhat.com>
20833 PR tree-optimization/48762
20834 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20836 2013-04-09 Richard Biener <rguenther@suse.de>
20838 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20840 (vect_build_slp_tree): Likewise.
20841 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20842 calculating the cost of a SLP instance.
20843 (vect_analyze_slp_instance): Use it from here, after building
20846 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20848 PR middle-end/56883
20849 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20850 expand_omp_for_static_chunk): Use simple_p = true in
20851 force_gimple_operand_gsi calls when assigning to addressable decls.
20853 2013-04-09 Jeff Law <law@redhat.com>
20855 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20856 when the boolean was created by converting a wider object which
20857 had a boolean range.
20859 2013-04-09 Richard Biener <rguenther@suse.de>
20861 * tree-vectorizer.h (slp_void_p): Remove.
20862 (slp_tree): Typedef before _slp_tree declaration.
20863 (struct _slp_tree): Use a vector of slp_tree as children.
20864 (vect_get_place_in_interleaving_chain): Remove.
20865 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20867 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20869 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20870 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20871 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20872 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20873 Use slp_node instead of slp_void_p and adjust.
20875 2013-04-09 Richard Biener <rguenther@suse.de>
20877 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20878 work that is not necessary.
20880 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20882 PR tree-optimization/56854
20883 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20884 forward into clobber stmts if it would change MEM_REF lhs into
20887 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20889 * tree.c (type_hash_lookup, type_hash_add): Make static.
20890 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20892 2013-04-09 Richard Biener <rguenther@suse.de>
20894 * tree.h (unsave_expr_now): Remove.
20895 * tree-inline.c (mark_local_for_remap_r): Remove.
20896 (unsave_expr_1): Likewise.
20897 (unsave_r): Likewise.
20898 (unsave_expr_now): Likewise.
20899 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
20900 (propagate_tree_value): Likewise.
20902 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
20904 * doc/rtl.texi (sequence): Rewrite documentation to match the
20905 current use of SEQUENCE rtl objects.
20906 * rtl.def (SEQUENCE): Likewise.
20908 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
20909 Update documentation.
20910 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
20911 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
20913 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
20915 2013-04-08 Teresa Johnson <tejohnson@google.com>
20917 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
20918 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
20920 (estimate_edge_size_and_time): Add comment to suggest using rounding
20922 (estimate_node_size_and_time): Ditto.
20923 (remap_edge_change_prob): Use helper rounding divide methods.
20924 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
20925 (gimple_mod_pow2_value_transform): Ditto.
20926 (gimple_mod_subtract_transform): Ditto.
20927 (gimple_ic_transform): Ditto.
20928 (gimple_stringops_transform): Ditto.
20929 * stmt.c (conditional_probability): Ditto.
20930 (emit_case_dispatch_table): Ditto.
20931 * lto-cgraph.c (merge_profile_summaries): Ditto.
20932 * tree-optimize.c (execute_fixup_cfg): Ditto.
20933 * cfgcleanup.c (try_forward_edges): Ditto.
20934 * cfgloopmanip.c (scale_loop_profile): Ditto.
20936 (duplicate_loop_to_header_edge): Ditto.
20937 (lv_adjust_loop_entry_edge): Ditto.
20938 * tree-vect-loop.c (vect_transform_loop): Ditto.
20939 * profile.c (compute_branch_probabilities): Ditto.
20940 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
20941 * lto-streamer-in.c (input_cfg): Ditto.
20942 * gimple-streamer-in.c (input_bb): Ditto.
20943 * ipa-cp.c (update_profiling_info): Ditto.
20944 (update_specialized_profile): Ditto.
20945 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
20946 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
20948 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20949 (compute_trg_info): Ditto.
20950 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
20951 (purge_dead_edges): Ditto.
20952 * loop-unswitch.c (unswitch_loop): Ditto.
20953 * cgraphclones.c (cgraph_clone_edge): Ditto.
20954 (cgraph_clone_node): Ditto.
20955 * tree-inline.c (copy_bb): Ditto.
20956 (copy_edges_for_bb): Ditto.
20957 (initialize_cfun): Ditto.
20958 (copy_cfg_body): Ditto.
20959 (expand_call_inline): Ditto.
20961 2013-04-08 Kai Tietz <ktietz@redhat.com>
20963 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
20964 TARGET_CYGWIN64 by TARGET_64BIT.
20966 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20968 * config/epiphany/epiphany.md (GPR_1): New constant.
20969 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
20970 * config/epiphany/epiphany.c (gen_compare_reg):
20971 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
20972 is already in place.
20973 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20974 Don't require being called during rtl expansion; If y operlaps r0,
20976 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20977 (epiphany_expand_epilogue): Likewise.
20979 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20980 Don't use CC_FPmode for ORDERED / UNORDERED.
20981 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
20983 * config/epiphany/constraints.md (CnL): New constraint.
20984 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20985 * config/epiphany/predicates.md (add_operand): Allow 1024.
20987 * config/epiphany/epiphany.md (logical_op): New code iterator.
20988 (op_mnc): New code attribute.
20989 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20990 (mov_f+1, mov_f+2): New peephole2 patterns.
20992 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20993 (cstoresi4): Also allow re-use of zero result when doing a NE
20994 comparison to a non-zero operand.
20995 Use (clobber (scratch)) for first insn if the gpr output is not needed.
20997 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20998 Use gen_addsi3_i / gen_subsi3_i.
21000 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21004 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21005 contain anything but clobbers, at most one __builtin_stack_restore,
21006 optionally debug stmts and final resx, and if it has at least one
21007 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21008 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21009 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21010 which isn't defaut definition, remove them.
21011 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21012 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21013 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21014 with MEM_REF LHS with SSA_NAME address.
21016 2013-04-08 Jeff Law <law@redhat.com>
21018 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21020 2013-04-08 Richard Biener <rguenther@suse.de>
21022 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21024 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21025 determined vector type.
21026 (vect_analyze_data_refs): Likewise.
21027 (vect_get_new_vect_var): Adjust.
21028 (vect_create_destination_var): Preserve SSA name versions.
21029 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21030 not dump anything here.
21032 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21034 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21035 Add member lr_slot_known.
21036 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21038 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21039 Remove code that sets lr_slot_offset according to what a previous
21040 version of epiphany_emit_save_restore used to do.
21041 (epiphany_emit_save_restore): When doing an lr save or restore,
21042 set/verify lr_slot_known and lr_slot_offset.
21044 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21047 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21050 2013-04-08 Richard Biener <rguenther@suse.de>
21052 * alias.c (find_base_term): Fix thinko in previous change.
21054 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21056 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21057 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21058 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21059 if possible to compute val.
21060 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21061 For QImode integers don't require anything about precision. Use
21062 const_with_all_bytes_same to find out if the constant doesn't have
21063 repeated bytes in it.
21065 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21067 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21070 2013-04-08 Marek Polacek <polacek@redhat.com>
21072 PR rtl-optimization/48182
21073 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21076 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21079 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21080 nuses, make sure we have a label.
21082 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21085 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21086 (rs6000_emit_swdiv_low_precision): Remove.
21087 (rs6000_emit_swdiv): Rewrite to handle between one and four
21088 iterations of Newton-Raphson generally; modify required number of
21089 iterations for some cases.
21090 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21092 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21094 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21095 set-but-unused variable.
21097 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21098 basic blocks of released function bodies garbage-collectable.
21100 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21101 (struct rtl_opt_pass): Add TODO_df_finish.
21103 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21105 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21107 * config/arm/constraints.md (q): New constraint.
21108 * config/arm/ldrdstrd.md: New file.
21109 * config/arm/arm.md (ldrdstrd.md) New include.
21110 (arm_movdi): Use "q" instead of "r" constraint
21111 for double-word memory access.
21112 (movdf_soft_insn): Likewise.
21113 * config/arm/vfp.md (movdi_vfp): Likewise.
21114 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21115 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21116 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21117 (mem_ok_for_ldrd_strd): Likewise.
21118 (output_move_double): Update assertion.
21120 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21122 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21124 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21126 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21127 define_insn_and_split.
21128 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21130 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21132 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21133 define_insn_and_split.
21134 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21135 (shiftsi3_compare): New pattern.
21136 (rrx): New pattern.
21137 * config/arm/unspecs.md (UNSPEC_RRX): New.
21139 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21141 * config/arm/arm.md (negdi_extendsidi): New pattern.
21142 (negdi_zero_extendsidi): Likewise.
21144 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21146 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21147 define_insn_and_split.
21148 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21149 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21151 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21153 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21154 define_insn_and_split.
21155 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21156 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21158 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21160 * config/arm/arm.md (subsi3_carryin): New pattern.
21161 (subsi3_carryin_const): Likewise.
21162 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21163 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21165 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21167 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21169 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21171 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21172 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21174 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21176 * config/arm/arm.c (arm_expand_builtin): Change fcode
21177 type to unsigned int.
21179 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21181 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21183 2013-04-04 Ian Lance Taylor <iant@google.com>
21185 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21188 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21190 PR middle-end/56729
21191 * df-scan.c (df_insn_delete): Disable failing assert.
21193 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21195 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21196 New function prototype.
21197 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21198 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21199 (arm_builtin_vectorized_function): New function.
21201 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21203 * config/arm/arm_neon_builtins.def: New file.
21204 * config/arm/arm.c (neon_builtin_data): Move contents to
21205 arm_neon_builtins.def.
21206 (enum arm_builtins): Include neon builtin definitions.
21207 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21208 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21210 2013-04-04 Marek Polacek <polacek@redhat.com>
21212 PR tree-optimization/48186
21213 * predict.c (maybe_hot_frequency_p): Return false if
21214 HOT_BB_FREQUENCY_FRACTION is 0.
21215 (cgraph_maybe_hot_edge_p): Likewise.
21217 2013-04-04 Richard Biener <rguenther@suse.de>
21219 PR tree-optimization/56826
21220 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21223 2013-04-04 Richard Biener <rguenther@suse.de>
21225 PR tree-optimization/56213
21226 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21227 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21229 2013-04-04 Richard Biener <rguenther@suse.de>
21231 PR tree-optimization/56837
21232 * tree-loop-distribution.c (classify_partition): For non-zero
21233 values require that the value has the same precision as its
21234 mode to be useful as memset value.
21236 2013-04-03 Nick Clifton <nickc@redhat.com>
21238 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21239 (fmssf4): Use fmsf.s on E3V5 architectures.
21240 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21241 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21243 2013-04-03 Jeff Law <law@redhat.com>
21245 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21246 (lra-eliminations.o): Likewise.
21248 2013-04-03 Teresa Johnson <tejohnson@google.com>
21250 * gcov-io.c (compute_working_sets): Moved most of body of old
21251 compute_working_sets here from profile.c.
21252 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21253 (gcov_working_set_t): Moved typedef here from basic-block.h
21254 (compute_working_set): Declare.
21255 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21256 (get_working_sets): Renamed from compute_working_set,
21257 replace most of body with call to new compute_working_sets.
21258 (get_exec_counts): Replace call to compute_working_sets
21259 to get_working_sets.
21260 * profile.h (get_working_sets): Renamed from compute_working_set.
21261 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21262 to get_working_sets.
21263 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21264 * gcov-dump.c (dump_working_sets): New function.
21266 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21268 * hwint.c (sext_hwi, zext_hwi): New functions.
21269 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21270 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21271 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21272 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21273 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21274 (sext_hwi, zext_hwi): New functions.
21276 2013-04-03 Jeff Law <law@redhat.com>
21278 PR tree-optimization/56799
21279 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21280 back test for widening conversion erroneously dropped in prior change.
21282 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21285 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21286 instead of next_real_insn.
21288 2013-04-03 Marek Polacek <polacek@redhat.com>
21291 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21293 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21296 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21298 (thumb1_output_casesi): Likewise.
21299 (thumb2_output_casesi): Likewise.
21301 2013-04-03 Richard Biener <rguenther@suse.de>
21303 PR tree-optimization/56817
21304 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21306 (tree_unroll_loops_completely_1): ... new function to manually
21307 walk the loop tree, properly defering outer loops of unrolled
21308 loops to later iterations.
21310 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21312 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21313 (vectorizable_load): Likewise.
21314 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21315 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21317 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21319 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21322 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21324 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21326 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21328 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21330 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21332 PR tree-optimization/56790
21333 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21336 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21338 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21340 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21341 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21344 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21347 * tree.h (force_folding_builtin_constant_p): New decl.
21348 * builtins.c (force_folding_builtin_constant_p): New variable.
21349 (fold_builtin_constant_p): Fold immediately also if
21350 force_folding_builtin_constant_p.
21352 2013-04-03 Richard Biener <rguenther@suse.de>
21354 PR tree-optimization/56812
21355 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21356 DRs of the same interleaving chain are independent.
21358 2013-04-02 Jason Merrill <jason@redhat.com>
21360 * gdbinit.in (pbb): Use debug fn.
21362 2013-04-02 Lawrence Crowl <crowl@google.com>
21364 * sese.h (struct ivtype_map_elt_s): Remove unused.
21365 (extern debug_ivtype_map): Remove unused.
21366 (extern eq_ivtype_map_elts): Remove unused.
21367 * sese.c (debug_ivtype_map): Removed unused.
21368 (debug_ivtype_map_1): Removed unused.
21369 (debug_ivtype_elt): Remove unused.
21370 (eq_ivtype_map_elts): Remove unused.
21373 2013-04-02 Kai Tietz <ktietz@redhat.com>
21376 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21377 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21378 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21380 (legitimize_pe_coff_symbol): Likewise.
21381 (is_imported_p): New helper-function.
21382 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21383 for Windows x64 targets.
21384 (ix86_expand_prologue): Optimize for pe-coff targets.
21385 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21386 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21387 medium/large code-model.
21388 (legitimize_pic_address): Likewise.
21389 (legitimize_tls_address): Likewise.
21390 (ix86_expand_call): Likewise.
21391 (x86_output_mi_thunk): Likewise.
21392 (get_dllimport_decl): Add new beimport argument.
21393 (construct_plt_address): Don't assert for x64 pe-coff targets.
21394 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21396 (SYMBOL_FLAG_STUBVAR): New macro.
21397 (SYMBOL_REF_STUBVAR_P): Likewise.
21398 * config/i386/winnt.c (stub_list): New structure.
21399 (stub_head): New local variable.
21400 (i386_pe_record_stub): New function.
21401 (i386_pe_file_end): Emit refptr-stubs.
21403 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21405 PR rtl-optimization/56745
21406 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21407 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21410 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21411 and both of them are MEM_REFs, just compare first argument for
21412 equality and attempt to deal even with differing offsets.
21415 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21416 of gimple_clobber_p to be MEM_REF.
21417 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21418 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21419 after gimplification.
21420 * asan.c (get_mem_ref_of_assignment): Don't instrument
21421 gimple_clobber_p stmts.
21422 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21423 gimple_clobber_p stmt if they have MEM_REF lhs and
21424 are dead because of another gimple_clobber_p stmt.
21425 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21426 gimple_clobber_p stmts like debug stmts.
21427 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21428 that refer to unused VAR_DECLs or uninitialized values.
21429 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21430 gimple_clobber_p stmts if they refer to removed parameters.
21431 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21434 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21436 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21437 using SWI48 mode attribute.
21439 2013-04-02 Wei Mi <wmi@google.com>
21441 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21442 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21443 *<rotate_insn><mode>3_mask in i386.md.
21445 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21447 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21449 2013-04-02 Richard Biener <rguenther@suse.de>
21451 PR tree-optimization/56778
21452 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21453 Runtime alias tests are not supported for gather loads.
21454 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21455 stmts referenced from SSA operands before updating SSA form.
21457 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21458 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21460 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21461 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21462 * config/arm/cortex-a53.md: New file.
21463 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21464 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21465 * config/arm/arm.c (arm_issue_rate): Likewise.
21466 * config/arm/arm-tune.md: Regenerate
21467 * config/arm/arm-tables.opt: Regenerate.
21468 * config/arm/arm-cores.def: Add cortex-a53.
21470 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21472 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21475 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21477 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21478 scalar load/store operations using B/H registers.
21479 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21481 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21483 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21485 * config/aarch64/aarch64.c
21486 (aarch64_simd_scalar_immediate_valid_for_move): New.
21487 * config/aarch64/aarch64-protos.h
21488 (aarch64_simd_scalar_immediate_valid_for_move): New.
21489 * config/aarch64/constraints.md (Dh, Dq): New.
21490 * config/aarch64/iterators.md (hq): New.
21492 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21494 * reorg.c (get_branch_condition): Deal with conditional returns.
21495 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21497 2013-04-01 Wei Mi <wmi@google.com>
21499 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21500 Truncate operand 2 using %b asm operand modifier.
21501 (*<shift_insn><mode>3_mask): Ditto.
21502 (*<rotate_insn><mode>3_mask): Ditto.
21504 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21506 PR middle-end/56798
21507 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21509 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21511 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21513 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21515 2013-03-30 Lawrence Crowl <crowl@google.com>
21517 * dse.c (clear_alias_sets): Remove never set.
21518 (disqualified_clear_alias_sets): Remove never set.
21519 (clear_alias_mode_pool): Remove never set.
21520 (dse_step0): Remove condition that is never true.
21521 (canon_address): Remove condition that is never true.
21522 (dse_step7): Remove condition that is never true.
21523 (rest_of_handle_dse): Remove condition that is never true.
21524 (rest_of_handle_dse::did_global): Remove never read from above.
21525 (dse_step2_spill): Remove never called from above.
21526 (dse_step5_spill): Remove never called from above.
21528 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21530 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21531 JUMP_TABLE_DATA changes.
21532 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21533 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21534 (Insns) <jump_table_data>: New entry.
21535 * doc/tm.texi: Regenerate.
21537 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21539 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21540 for table jump at the end of a basic block using tablejump_p.
21541 * targhooks.c (default_invalid_within_doloop): Likewise.
21542 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21543 target hook implementation that is identical to the default hook.
21544 (rs6000_invalid_within_doloop): Remove.
21546 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21547 unused variable from tablejump_p call.
21549 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21550 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21551 (INSN_DELETED_P): Likewise.
21552 (emit_jump_table_data): New prototype.
21553 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21554 after 4th as unused.
21555 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21556 * sched-vis.c (print_insn): Likewise.
21557 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21558 insn for compatibility with back ends that use next_active_insn to
21559 identify jump table data.
21560 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21561 (remove_insn): Likewise.
21562 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21564 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21565 (emit_jump_table_data): New function.
21567 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21568 basic block, a JUMP_TABLE_DATA never is.
21569 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21570 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21571 off from code handling real insns.
21572 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21573 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21574 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21575 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21576 is not a NONDEBUG_INSN_P.
21577 * ira-costs.c (scan_one_insn): Likewise.
21578 * jump.c (mark_all_labels): Likewise.
21579 (mark_jump_label_1): Likewise.
21580 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21581 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21583 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21584 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21585 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21586 JUMP_TABLE_DATA_P insns.
21587 (calculate_elim_costs_all_insns): Likewise.
21588 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21589 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21590 (delete_output_reload): Code style fixups.
21591 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21592 insn flags on this non-insn.
21593 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21594 as scheduling barriers, for pre-change compatibility.
21595 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21596 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21598 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21599 redundant JUMP_TABLE_DATA_P test.
21600 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21601 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21602 (frv_for_each_packet): Likewise.
21603 * config/i386/i386.c (min_insn_size): Likewise.
21604 (ix86_avoid_jump_mispredicts): Likewise.
21605 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21606 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21607 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21608 (mips16_insn_length): Robustify.
21609 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21610 (mips16_split_long_branches): Likewise.
21611 * config/pa/pa.c (pa_combine_instructions): Likewise.
21612 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21613 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21614 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21615 as contributing to pool range lengths.
21616 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21617 Remove redundant JUMP_TABLE_DATA_P test.
21618 (sh_loop_align): Likewise.
21619 (split_branches): Likewise.
21620 (sh_insn_length_adjustment): Likewise.
21621 * config/spu/spu.c (get_branch_target): Likewise.
21623 2013-03-29 Jan Hubicka <jh@suse.cz>
21625 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21626 gcov streaming; stream hot bb threshold to ltrans.
21627 * predict.c (get_hot_bb_threshold): Break out from ....
21628 (maybe_hot_count_p): ... here.
21629 (set_hot_bb_threshold): New function.
21630 * lto-section-in.c (lto_section_name): Add profile.
21631 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21632 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21633 and data-streamer.h
21634 (histogram_entry): New structure.
21635 (histogram, histogram_pool): New global vars.
21636 (histogram_hash): New structure.
21637 (histogram_hash::hash): New method.
21638 (histogram_hash::equal): Likewise.
21639 (account_time_size): New function.
21640 (cmp_counts): New function.
21641 (dump_histogram): New function.
21642 (ipa_profile_generate_summary): New function.
21643 (ipa_profile_write_summary): New function.
21644 (ipa_profile_read_summary): New function.
21645 (ipa_profile): Decide on threshold.
21646 (pass_ipa_profile): Add ipa_profile_write_summary and
21647 ipa_profile_read_summary.
21648 * Makefile.in (ipa.o): Update dependencies.
21649 * lto-streamer.h (LTO_section_ipa_profile): New section.
21651 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21653 * tree.h (VAR_P): New.
21655 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21658 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21660 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21662 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21663 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21664 (control_flow_insn_p): Likewise.
21665 * cfgrtl.c (duplicate_insn_chain): Likewise.
21666 * final.c (get_attr_length_1): Likewise.
21667 (shorten_branches): Likewise.
21668 (final_scan_insn): Likewise.
21669 * function.c (instantiate_virtual_regs): Likewise.
21670 * gcse.c (insert_insn_end_basic_block): Likewise.
21671 * ira-costs.c (scan_one_insn): Likewise.
21672 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21673 * lra.c (check_rtl): Likewise.
21674 * reload1.c (elimination_costs_in_insn): Likewise.
21675 * reorg.c (follow_jumps): Likewise.
21677 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21678 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21679 (thumb_far_jump_used_p): Likewise.
21680 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21681 (workaround_speculation): Likewise.
21682 (add_sched_insns_for_speculation): Likewise.
21683 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21684 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21685 (frv_for_each_packet): Likewise.
21686 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21687 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21688 (final_emit_insn_group_barriers): Likewise.
21689 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21690 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21691 (mips16_insn_length): Likewise.
21692 * config/pa/pa.c (pa_reorg): Likewise.
21693 (pa_combine_instructions): Likewise.
21694 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21695 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21696 (sh_reorg): Likewise.
21697 (split_branches): Likewise.
21698 * config/spu/spu.c (get_branch_target): Likewise.
21700 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21703 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21705 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21706 Fix declaration name.
21708 2013-03-28 Lawrence Crowl <crowl@google.com>
21710 * graphds.h (struct graph.indicies): Remove unused.
21711 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21712 (SCOP_ORIGINAL_PDDRS): Remove unused.
21713 * sese.h (extern insert_loop_close_phis): Removed unused.
21714 (extern insert_guard_phis): Removed unused.
21715 (extern ivtype_map_elt_info): Removed unused.
21716 (new_ivtype_map_elt): Removed unused.
21717 * sese.c (ivtype_map_elt_info): Removed unused.
21719 2013-03-28 Lawrence Crowl <crowl@google.com>
21721 * Makefile.in: Add several missing include dependences.
21723 (test-dump.o): New. This object is not added to any executable,
21724 but is present for ad-hoc testing.
21726 (debug (const bitmap_head_def &)): New.
21727 (debug (const bitmap_head_def *)): New.
21729 (extern debug (const bitmap_head_def &)): New.
21730 (extern debug (const bitmap_head_def *)): New.
21732 (debug (edge_def &)): New.
21733 (debug (edge_def *)): New.
21735 (debug (basic_block_def &)): New.
21736 (debug (basic_block_def *)): New.
21738 (dump_node (const_tree, int, FILE *)): Correct source file.
21740 (debug (die_struct &)): New.
21741 (debug (die_struct *)): New.
21743 (extern debug (die_struct &)): New.
21744 (extern debug (die_struct *)): New.
21745 * gimple-pretty-print.c
21746 (debug (gimple_statement_d &)): New.
21747 (debug (gimple_statement_d *)): New.
21748 * gimple-pretty-print.h
21749 (extern debug (gimple_statement_d &)): New.
21750 (extern debug (gimple_statement_d *)): New.
21752 (debug (ira_allocno_copy &)): New.
21753 (debug (ira_allocno_copy *)): New.
21754 (debug (ira_allocno &)): New.
21755 (debug (ira_allocno *)): New.
21757 (extern debug (ira_allocno_copy &)): New.
21758 (extern debug (ira_allocno_copy *)): New.
21759 (extern debug (ira_allocno &)): New.
21760 (extern debug (ira_allocno *)): New.
21762 (debug (live_range &)): New.
21763 (debug (live_range *)): New.
21765 (debug (lra_live_range &)): New.
21766 (debug (lra_live_range *)): New.
21768 (debug (lra_live_range &)): New.
21769 (debug (lra_live_range *)): New.
21771 (debug (omega_pb_d &)): New.
21772 (debug (omega_pb_d *)): New.
21774 (extern debug (omega_pb_d &)): New.
21775 (extern debug (omega_pb_d *)): New.
21777 (debug (const rtx_def &)): New.
21778 (debug (const rtx_def *)): New.
21780 (debug_tree (tree): Move within file.
21781 (debug_raw (const tree_node &)): New.
21782 (debug_raw (const tree_node *)): New.
21783 (dump_tree_via_hooks (const tree_node *, int)): New.
21784 (debug (const tree_node &)): New.
21785 (debug (const tree_node *)): New.
21786 (debug_verbose (const tree_node &)): New.
21787 (debug_verbose (const tree_node *)): New.
21788 (debug_head (const tree_node &)): New.
21789 (debug_head (const tree_node *)): New.
21790 (debug_body (const tree_node &)): New.
21791 (debug_body (const tree_node *)): New.
21792 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21793 (debug (vec<tree, va_gc> &)): New.
21794 (debug (vec<tree, va_gc> *)): New.
21796 (extern debug (const rtx_def &)): New.
21797 (extern debug (const rtx_def *)): New.
21799 (debug_raw (simple_bitmap_def &)): New.
21800 (debug_raw (simple_bitmap_def *)): New.
21801 (debug (simple_bitmap_def &)): New.
21802 (debug (simple_bitmap_def *)): New.
21804 (extern debug (simple_bitmap_def &)): New.
21805 (extern debug (simple_bitmap_def *)): New.
21806 (extern debug_raw (simple_bitmap_def &)): New.
21807 (extern debug_raw (simple_bitmap_def *)): New.
21809 (debug (vinsn_def &)): New.
21810 (debug (vinsn_def *)): New.
21811 (debug_verbose (vinsn_def &)): New.
21812 (debug_verbose (vinsn_def *)): New.
21813 (debug (expr_def &)): New.
21814 (debug (expr_def *)): New.
21815 (debug_verbose (expr_def &)): New.
21816 (debug_verbose (expr_def *)): New.
21817 (debug (vec<rtx> &)): New.
21818 (debug (vec<rtx> *)): New.
21820 (extern debug (vinsn_def &)): New.
21821 (extern debug (vinsn_def *)): New.
21822 (extern debug_verbose (vinsn_def &)): New.
21823 (extern debug_verbose (vinsn_def *)): New.
21824 (extern debug (expr_def &)): New.
21825 (extern debug (expr_def *)): New.
21826 (extern debug_verbose (expr_def &)): New.
21827 (extern debug_verbose (expr_def *)): New.
21828 (extern debug (vec<rtx> &)): New.
21829 (extern debug (vec<rtx> *)): New.
21831 (_list_iter_cond_expr): Make inline instead of static.
21833 (debug (sreal &)): New.
21834 (debug (sreal *)): New.
21836 (extern debug (sreal &)): New.
21837 (extern debug (sreal *)): New.
21839 (extern debug_raw (const tree_node &)): New.
21840 (extern debug_raw (const tree_node *)): New.
21841 (extern debug (const tree_node &)): New.
21842 (extern debug (const tree_node *)): New.
21843 (extern debug_verbose (const tree_node &)): New.
21844 (extern debug_verbose (const tree_node *)): New.
21845 (extern debug_head (const tree_node &)): New.
21846 (extern debug_head (const tree_node *)): New.
21847 (extern debug_body (const tree_node &)): New.
21848 (extern debug_body (const tree_node *)): New.
21849 (extern debug (vec<tree, va_gc> &)): New.
21850 (extern debug (vec<tree, va_gc> *)): New.
21852 (debug (struct loop &)): New.
21853 (debug (struct loop *)): New.
21854 (debug_verbose (struct loop &)): New.
21855 (debug_verbose (struct loop *)): New.
21856 * tree-dump.c: Add header dependence.
21858 (extern debug (struct loop &)): New.
21859 (extern debug (struct loop *)): New.
21860 (extern debug_verbose (struct loop &)): New.
21861 (extern debug_verbose (struct loop *)): New.
21863 (debug (data_reference &)): New.
21864 (debug (data_reference *)): New.
21865 (debug (vec<data_reference_p> &)): New.
21866 (debug (vec<data_reference_p> *)): New.
21867 (debug (vec<ddr_p> &)): New.
21868 (debug (vec<ddr_p> *)): New.
21870 (extern debug (data_reference &)): New.
21871 (extern debug (data_reference *)): New.
21872 (extern debug (vec<data_reference_p> &)): New.
21873 (extern debug (vec<data_reference_p> *)): New.
21874 (extern debug (vec<ddr_p> &)): New.
21875 (extern debug (vec<ddr_p> *)): New.
21877 (debug (pt_solution &)): New.
21878 (debug (pt_solution *)): New.
21880 (extern debug (pt_solution &)): New.
21881 (extern debug (pt_solution *)): New.
21883 (debug (_var_map &)): New.
21884 (debug (_var_map *)): New.
21885 (debug (tree_live_info_d &)): New.
21886 (debug (tree_live_info_d *)): New.
21888 (extern debug (_var_map &)): New.
21889 (extern debug (_var_map *)): New.
21890 (extern debug (tree_live_info_d &)): New.
21891 (extern debug (tree_live_info_d *)): New.
21893 2013-03-28 Jan Hubicka <jh@suse.cz>
21895 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
21897 2013-03-28 Ian Bolton <ian.bolton@arm.com>
21899 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
21900 record only when desired or required.
21902 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
21904 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
21905 *vec_extractv2di_1_rex64. Use x64 isa attribute.
21907 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21909 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
21910 (*andsi3_compare0_uxtw): New pattern.
21911 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
21912 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
21914 2013-03-28 Jan Hubicka <jh@suse.cz>
21916 * data-streamer-in.c (streamer_read_gcov_count): New function.
21917 * gimple-streamer-out.c: Include value-prof.h.
21918 (output_gimple_stmt): Output histogram.
21919 (output_bb): Use streamer_write_gcov_count.
21920 * value-prof.c: Include data-streamer.h
21921 (dump_histogram_value): Add HIST_TYPE_MAX.
21922 (stream_out_histogram_value): New function.
21923 (stream_in_histogram_value): New function.
21924 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
21925 (stream_out_histogram_value, stream_in_histogram_value): Declare.
21926 * data-streamer-out.c (streamer_write_gcov_count): New function.
21927 (streamer_write_gcov_count_stream): New function.
21928 * lto-cgraph.c (lto_output_edge): Update counter streaming.
21929 (lto_output_node): Likewise.
21930 (input_node, input_edge): Likewise.
21931 * lto-streamer-out.c (output_cfg): Update streaming.
21932 * lto-streamer-in.c (input_cfg): Likewise.
21933 * data-streamer.h (streamer_write_gcov_count,
21934 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
21935 * gimple-streamer-in.c: Include value-prof.h
21936 (input_gimple_stmt): Input histograms.
21937 (input_bb): Update profile streaming.
21939 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21941 * genmodes.c (emit_max_int): New function.
21942 (emit_insn_modes_h): Added call to emit_max_function.
21943 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
21945 * machmode.def: Fixed comment.
21947 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
21949 * combine.c (try_combine): Removed useless assert.
21950 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
21952 2013-03-28 Marek Polacek <polacek@redhat.com>
21953 Richard Biener <rguenther@suse.de>
21955 PR tree-optimization/56695
21956 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
21957 build signed result of a vector comparison.
21958 * tree-cfg.c (verify_gimple_comparison): Check that a result
21959 of a vector comparison has signed type.
21961 2013-03-28 Richard Biener <rguenther@suse.de>
21963 PR tree-optimization/37021
21964 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
21965 do not restrict gaps between groups.
21966 * tree-vect-stmts.c (vectorizable_load): Properly account for
21967 a gap between groups.
21969 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
21971 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
21972 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21975 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21977 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21978 * doc/extend.texi (Named Address Spaces): Ditto.
21979 (Variable Attributes): Ditto.
21981 2013-03-27 Kai Tietz <ktietz@redhat.com>
21983 * config.build: Add support for cygwin x64 target.
21984 * config.gcc: Likewise.
21985 * config.host: Likewise.
21986 * configure.ac: Likewise
21987 * configure: Regenerated.
21989 2013-03-27 Kai Tietz <ktietz@redhat.com>
21991 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21992 * config/i386/t-cygwin-w64: New file.
21993 * config/i386/cygwin-w64.h: New file.
21994 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21995 and add support for x64-cygwin target.
21996 (CPP_SPEC): Likewise.
21997 (CXX_WRAP_SPEC_LIST): Undefine before define.
21998 (LIBGCJ_SONAME): Use 15 as version.
22000 2013-03-27 Richard Biener <rguenther@suse.de>
22002 PR tree-optimization/56716
22003 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22004 dumping for ref nodes.
22006 2013-03-27 Martin Jambor <mjambor@suse.cz>
22008 PR tree-optimization/55334
22009 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22010 restricted pointers to arrays.
22012 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22014 * Makefile.in (.SUFFIXES): Add .cc.
22015 (.c.o): Apply same recipe for implicit rule .cc.o.
22017 2013-03-27 Richard Biener <rguenther@suse.de>
22019 PR tree-optimization/37021
22020 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22021 REALPART/IMAGPART_EXPRs around the supported refs.
22022 * tree-ssa-structalias.c (find_func_aliases): Assume that
22023 floating-point values are not used to transfer pointers.
22025 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22027 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22028 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22029 * doc/tm.texi: Regenerate.
22030 * targhooks.h (default_has_ifunc_p): New.
22031 * targhooks.c (default_has_ifunc_p): Ditto.
22032 * config/linux-protos.h: New file.
22033 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22034 hook for linux which disables support of indirect functions in android.
22035 * config/linux-android.c: New file.
22036 * config/t-linux-android.c: Ditto.
22037 * config.gcc: Added new object file linux-android.o.
22038 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22039 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22040 * varasm.c (do_assemble_alias): Likewise.
22041 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22042 doesn't support indirect functions.
22043 * configure: Regenerate.
22045 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22048 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22049 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22052 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22054 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22056 2013-03-27 Terry Guo <terry.guo@arm.com>
22058 * config/arm/arm-cores.def: Added core cortex-r7.
22059 * config/arm/arm-tune.md: Regenerated.
22060 * config/arm/arm-tables.opt: Regenerated.
22061 * doc/invoke.texi: Added entry for core cortex-r7.
22063 2013-03-27 Walter Lee <walt@tilera.com>
22065 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22066 double-decrement of next_scratch_regno.
22068 2013-03-27 Walter Lee <walt@tilera.com>
22070 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22072 (insn_v1mulus): Ditto.
22073 (insn_v2muls): Ditto.
22075 2013-03-27 Walter Lee <walt@tilera.com>
22077 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22078 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22080 2013-03-27 Walter Lee <walt@tilera.com>
22082 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22083 (*sibcall_value): Ditto.
22085 2013-03-27 Walter Lee <walt@tilera.com>
22087 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22088 (insn_mnz_v8qi): ... this ...
22089 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22091 (insn_v<n>mnz): Replaced by ...
22092 (insn_v1mnz): ... this ...
22093 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22095 (insn_mz_<mode>): Replaced by ...
22096 (insn_mz_v8qi): ... this ...
22097 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22099 (insn_v<n>mz): Replaced by ...
22100 (insn_v1mz): ... this ...
22101 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22104 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22106 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22108 2013-03-26 Roland McGrath <mcgrathr@google.com>
22110 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22111 than fprintf with a non-constant, non-format string.
22113 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22115 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22116 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22117 operand 0 predicate.
22118 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22119 attribute. Use general_x64nomem_operand as operand 1 predicate.
22120 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22121 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22122 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22123 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22124 (mov<mode>_insv_1): Remove expander. Merge insn with
22125 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22126 Use general_x64nomem_operand as operand 1 predicate.
22127 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22128 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22129 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22130 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22131 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22132 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22133 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22134 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22135 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22136 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22137 (general_x64nomem_operand): Ditto.
22139 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22141 * config/rtems.opt: Add -pthread option.
22143 2013-03-26 Richard Biener <rguenther@suse.de>
22145 * alias.c (find_base_term): Avoid redundant and not used recursion.
22146 (base_alias_check): Get the initial base term from the caller.
22147 (true_dependence_1): Compute and pass base terms to base_alias_check.
22148 (write_dependence_p): Likewise.
22149 (may_alias_p): Likewise.
22151 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22153 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22154 PC-relative load in SI modes and above only.
22156 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22158 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22159 * config/arm/iwmmxt.md (WCGR0): Update.
22160 (WCGR1, WCGR2, WCGR3): Likewise.
22162 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22164 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22165 Use x64 and nox64 isa attributes.
22167 2013-03-26 Richard Biener <rguenther@suse.de>
22169 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22170 alignment computations and rely on get_object_alignment_1
22171 for the !TYPE_P case.
22172 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22174 2013-03-26 Walter Lee <walt@tilera.com>
22176 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22177 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22179 2013-03-25 Jeff Law <law@redhat.com>
22181 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22182 check for INTEGRAL_TYPE_P that was missing due to checking in
22183 wrong version of prior patch.
22185 2013-03-25 Walter Lee <walt@tilera.com>
22187 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22188 TILEGX_INSN_SHUFFLEBYTES1.
22189 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22191 (tilegx_builtins): Ditto.
22192 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22194 2013-03-25 Walter Lee <walt@tilera.com>
22196 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22197 (floatunssisf2): New pattern.
22198 (floatsidf2): New pattern.
22199 (floatunssidf2): New pattern.
22201 2013-03-25 Walter Lee <walt@tilera.com>
22203 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22204 tests for constraint J, K, N, P.
22206 2013-03-25 Walter Lee <walt@tilera.com>
22208 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22209 Use indirect/pcrel encoding.
22210 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22213 2013-03-25 Steve Ellcey <sellcey@mips.com>
22215 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22216 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22217 * config/mips/mips.c (mips_option_override): Set IMADD default.
22218 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22219 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22220 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22221 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22222 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22224 2013-03-25 Jeff Law <law@redhat.com>
22226 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22227 slightly to avoid creating and folding useless trees. Simplify
22228 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22230 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22232 * config/i386/i386.md (*zero_extendsidi2): Merge with
22233 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22234 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22235 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22236 targets. Clarify comment.
22238 2013-03-25 Martin Jambor <mjambor@suse.cz>
22240 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22241 pass-through jump functions differently.
22242 (ipa_read_jump_function): Likewise. Also use setter functions to set
22245 2013-03-25 Martin Jambor <mjambor@suse.cz>
22247 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22248 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22250 (ipa_get_indirect_edge_target): New function.
22251 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22252 ipa_get_indirect_edge_target. Update all callers.
22253 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22254 ipa_get_indirect_edge_target_1 instead of calling
22255 ipa_get_indirect_edge_target.
22256 (create_specialized_node): Pass aggvlas to
22257 ipcp_discover_new_direct_edges.
22259 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22261 * config/arm/arm.md (f_sels, f_seld): New types.
22262 (*cmov<mode>): New pattern.
22263 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22266 2013-03-25 Kai Tietz <ktietz@redhat.com>
22268 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22269 POSIX-printf for mingw-hosted builds.
22271 2013-03-25 Richard Biener <rguenther@suse.de>
22273 PR middle-end/56694
22274 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22275 must-not-throw stmt location.
22277 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22279 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22280 Emit load-acquire versions when acq is true.
22281 (arm_emit_store_exclusive): Add rel parameter.
22282 Emit store-release versions when rel is true.
22283 (arm_split_compare_and_swap): Use acquire-release instructions
22285 of barriers when appropriate.
22286 (arm_split_atomic_op): Likewise.
22287 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22288 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22289 (VUNSPEC_SLX): Likewise.
22290 (VUNSPEC_LDA): Likewise.
22291 (VUNSPEC_STL): Likewise.
22292 * config/arm/sync.md (atomic_load<mode>): New pattern.
22293 (atomic_store<mode>): Likewise.
22294 (arm_load_acquire_exclusive<mode>): Likewise.
22295 (arm_load_acquire_exclusivesi): Likewise.
22296 (arm_load_acquire_exclusivedi): Likewise.
22297 (arm_store_release_exclusive<mode>): Likewise.
22299 2013-03-25 Catherine Moore <clm@codesourcery.com>
22301 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22302 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22303 * config/mip/predicates.md (lwsp_swsp_operand,
22304 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22305 sb16_operand, db4_operand, db7_operand, ib3_operand,
22306 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22307 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22308 andi16_operand): New predicates.
22309 * config/mips/mips.md (compression): New attribute.
22310 (enabled): New attribute.
22311 (length): Consider compression in computing length.
22312 (shift_compression): New code attribute.
22313 (*add<mode>3): New operands. Record compression.
22314 (sub<mode>3): Likewise.
22315 (one_cmpl<mode>2): Likewise.
22316 (*and<mode>3): Likewise.
22317 (*ior<mode>3): Likewise.
22318 (unnamed pattern for xor): Likewise.
22319 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22320 (*<optab><mode>3): Likewise.
22321 (*mov<mode>_internal: Likewise.
22322 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22323 (mips_unsigned_immediate_p): New.
22324 (umips_lwsp_swsp_address_p): New.
22325 (m16_based_address_p): New.
22326 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22327 (mips_unsigned_immediate_p): New prototype.
22328 (lwsp_swsp_address_p): New prototype.
22329 (m16_based_address_p): New prototype.
22330 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22331 (mips_signed_immediate_p): New function.
22332 (m16_based_address_p): New function.
22333 (lwsp_swsp_address_p): New function.
22334 (mips_print_operand_punctuation): Recognize short delay slot insns
22335 for microMIPS.add<mode>3"
22337 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22340 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22341 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22343 2013-03-25 Richard Biener <rguenther@suse.de>
22345 PR tree-optimization/56689
22346 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22349 2013-03-25 Richard Biener <rguenther@suse.de>
22351 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22353 (memory_references): Likewise.
22354 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22355 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22356 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22357 (gather_mem_refs_in_loops): Fold into ...
22358 (analyze_memory_references): ... this. Move initialization
22359 to tree_ssa_lim_initialize.
22360 (fill_always_executed_in): Rename to ...
22361 (fill_always_executed_in_1): ... this.
22362 (fill_always_executed_in): Move contains_call computation to
22363 this new function from ...
22364 (tree_ssa_lim_initialize): ... here.
22365 (tree_ssa_lim): Call fill_always_executed_in.
22367 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22369 * postreload.c (reload_combine): Fix code detecting returns.
22371 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22373 * function.c (emit_use_return_register_into_block): On cc0 targets,
22374 do not emit the sequence between cc0 setter and user.
22376 2013-03-25 Kai Tietz <ktietz@redhat.com>
22378 * config/i386/predicates.md (local_symbolic_operand): Interpret
22379 dll-imported symbols as none-local.
22381 2013-03-25 Richard Biener <rguenther@suse.de>
22383 * tree-ssa-loop-im.c (struct depend): Remove.
22384 (struct lim_aux_data): Make depends a vec of gimples.
22385 (free_lim_aux_data): Adjust.
22386 (add_dependency): Likewise.
22387 (set_level): Likewise.
22389 2013-03-25 Richard Biener <rguenther@suse.de>
22391 PR middle-end/56434
22392 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22393 the pointer returned by calls with ECF_MALLOC set.
22395 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22397 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22399 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22401 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22402 using MMXMODE mode iterator.
22403 (*move<mode>_internal): Merge with *movv2sf_internal and
22404 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22406 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22408 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22409 (record_last_mem_set_info): Likewise.
22411 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22412 of XNEWVEC followed by memset.
22413 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22415 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22417 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22418 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22419 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22420 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22421 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22422 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22423 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22424 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22425 BARRIER_P instead of GET_CODE.
22427 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22429 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22430 inaccuracy in the probing code.
22432 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22433 (ctrapdi4): Likewise.
22435 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22437 * calls.c (expand_call): Add missing guard to code handling return
22438 of non-BLKmode structures in MSB.
22439 * function.c (expand_function_end): Likewise.
22441 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22443 * combine.c (try_combine): Adjust comment. Do not add the set of
22444 insn #0 if the destination indirectly is set or dies in insn #2.
22445 Tidy up code to distribute a new note.
22447 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22449 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22450 also for alternatives 16 and 17.
22452 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22454 * config/i386/sse.md (*mov<mode>_internal): Merge with
22455 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22456 Emit insn template depending on type attribute. Use
22457 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22458 movd instead of movq mnemonic for interunit moves. Rewrite mode
22459 attribute calculation. Remove unit attribute calculation.
22460 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22461 Set prefix_data16 attribute for DImode ssemov types.
22462 Use Ym instead of y for SSE-MMX conversion alternatives.
22463 Reorder operand constraints.
22465 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22467 * df.h (df_insn_delete): Adjust prototype.
22468 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22469 and let it decide whether mark the basic block dirty.
22470 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22471 * df-scan.c (df_insn_info_delete): New helper function, split
22472 off from df_insn_delete.
22473 (df_scan_free_bb_info): Use it.
22474 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22476 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22477 that the insn is actually an insn and it has a non-NULL basic block.
22478 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22480 2013-03-22 Richard Biener <rguenther@suse.de>
22482 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22484 (mem_ref_alloc): Do not allocate them.
22485 (refs_independent_p): Do not query or maintain a cache.
22487 2013-03-22 Richard Biener <rguenther@suse.de>
22489 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22490 (gather_mem_refs_in_loops): Do not compute it.
22491 (analyze_memory_references): Do not allocate it.
22492 (tree_ssa_lim_finalize): Do not free it.
22493 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22495 2013-03-22 Richard Biener <rguenther@suse.de>
22497 * is-a.h (as_a): Use gcc_checking_assert.
22499 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22501 * config/aarch64/aarch64.c (aarch64_print_operand): New
22502 format specifier for printing a constant in hex.
22503 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22504 format specifier for printing second operand.
22506 2013-03-22 Richard Biener <rguenther@suse.de>
22508 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22510 (gather_mem_refs_in_loops): Perform store accumulation here.
22511 (create_vop_ref_mapping_loop): Remove.
22512 (create_vop_ref_mapping): Likewise.
22513 (analyze_memory_references): Initialize refs_stored_in_loop.
22514 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22515 (record_indep_loop): Remove.
22516 (record_dep_loop): New function.
22517 (ref_indep_loop_p_1): Adjust to only walk over references
22518 in the loop, not its subloops.
22519 (ref_indep_loop_p): Rename to ...
22520 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22521 maintaining a more fine-grained cache.
22522 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22523 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22525 2013-03-22 Richard Biener <rguenther@suse.de>
22527 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22528 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22529 aggregate mem_ref_loc.
22530 (free_mem_ref_locs): Inline into ...
22531 (memref_free): ... this and adjust.
22532 (mem_ref_alloc): Adjust.
22533 (mem_ref_locs_alloc): Remove.
22534 (record_mem_ref_loc): Adjust.
22535 (get_all_locs_in_loop): Rewrite into ...
22536 (for_all_locs_in_loop): ... this iterator.
22537 (rewrite_mem_ref_loc): New functor.
22538 (rewrite_mem_refs): Use for_all_locs_in_loop.
22539 (sm_set_flag_if_changed): New functor.
22540 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22541 (ref_always_accessed): New functor.
22542 (ref_always_accessed_p): Use for_all_locs_in_loop.
22544 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22546 * tree-pass.h (PROP_gimple_lvec): New.
22547 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22548 (init_optimization_passes): Move pass_lower_vector.
22549 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22551 (pass_lower_vector): Provide PROP_gimple_lvec.
22552 (pass_lower_vector_ssa): Likewise.
22553 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22555 2013-03-21 Mark Wielaard <mjw@redhat.com>
22557 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22559 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22561 * config/i386/i386.md (*movdi_internal): Disparage slightly
22562 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22563 conversion alternatives.
22565 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22567 PR middle-end/48087
22568 * diagnostic.def (DK_WERROR): New kind.
22569 * diagnostic.h (werrorcount): Define.
22570 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22571 promoted to DK_ERROR, increment DK_WERROR counter instead of
22573 * toplev.c (toplev_main): Call print_ignored_options even if
22574 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22575 even if just werrorcount is non-zero.
22578 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22580 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22581 (string_cst_pool_decl): New function.
22582 (optimize_one_addr_into_implicit_ptr): New function.
22583 (resolve_addr_in_expr): Optimize DWARF location expression
22584 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22585 which doesn't live in memory, but has DW_AT_location or
22586 DW_AT_const_value, or refers to a string literal, into
22587 DW_OP_GNU_implicit_pointer.
22588 (optimize_location_into_implicit_ptr): New function.
22589 (resolve_addr): If removing DW_AT_location of a variable because
22590 it was DW_OP_addr of address of the variable, but the variable doesn't
22591 live in memory, try to emit const value attribute for the initializer.
22593 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22595 * tree.h (VECTOR_TYPE_P): New macro.
22596 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22597 TYPE_MODE): Use it.
22598 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22599 VEC_COND_EXPR cannot be lvalues.
22600 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22602 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22604 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22605 Restrict the transformation to equal modes.
22607 2013-03-21 Richard Biener <rguenther@suse.de>
22609 PR tree-optimization/39326
22610 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22611 (MEM_ANALYZABLE): Adjust.
22612 (record_mem_ref_loc): Move bitmap ops ...
22613 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22614 unanalyzable refs, do not record locations for it.
22615 (analyze_memory_references): Allocate ref zero as shared
22617 (refs_independent_p): Do not test for unanalyzed mems here.
22618 (ref_indep_loop_p_1): Special-case disambiguation against
22619 the unanalyzed ref.
22620 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22622 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22624 * config/arm/arm-protos.h (tune_params): Add
22625 prefer_neon_for_64bits field.
22626 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22627 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22628 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22629 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22630 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22631 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22632 (arm_option_override): Handle -mneon-for-64bits new option.
22633 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22634 (prefer_neon_for_64bits): Declare new variable.
22635 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22636 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22637 (arch_enabled): Handle new arch types. Remove support for onlya8
22639 (one_cmpldi2): Use new arch names.
22640 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22641 * config/arm/arm.opt (mneon-for-64bits): Add option.
22642 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22643 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22644 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22646 * doc/invoke.texi (-mneon-for-64bits): Document.
22648 2013-03-21 Richard Biener <rguenther@suse.de>
22650 PR tree-optimization/39326
22651 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22652 (sort_bbs_in_loop_postorder_cmp): New function.
22653 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22655 2013-03-21 Richard Biener <rguenther@suse.de>
22657 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22658 (vect_insert_into_interleaving_chain): Likewise.
22659 (vect_drs_dependent_in_basic_block): Inline ...
22660 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22662 (vect_analyze_data_ref_dependence): ... here. Simplify.
22663 (vect_check_interleaving): Simplify.
22664 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22665 (vect_slp_analyze_data_ref_dependences): ... this new function.
22666 (dr_group_sort_cmp): New function.
22667 (vect_analyze_data_ref_accesses): Compute data-reference groups
22668 here instead of in vect_analyze_data_ref_dependence. Use
22669 a more efficient algorithm.
22670 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22671 vect_slp_analyze_data_ref_dependences. Call
22672 vect_analyze_data_ref_accesses earlier.
22673 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22674 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22675 (vect_slp_analyze_data_ref_dependences): New prototype.
22677 2013-03-21 Richard Biener <rguenther@suse.de>
22679 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22680 ref is stored in the loop.
22681 (find_refs_for_sm): Walk only over all stores.
22682 (store_motion_loop): Allocate from lim_bitmap_obstack.
22683 (store_motion): Likewise.
22685 2013-03-21 Richard Biener <rguenther@suse.de>
22687 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22688 Update virtual SSA form.
22690 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22692 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22693 * configure: Regenerate.
22694 * config.in: Regenerate.
22695 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22696 if !HAVE_LD_EH_FRAME_CIEV3.
22698 2013-03-21 Richard Biener <rguenther@suse.de>
22700 * tree-cfg.c (verify_expr_no_block): New function.
22701 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22702 nor DECL_VALUE_EXPR have locations with associated blocks.
22703 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22704 (clear_unused_block_pointer): Remove code dealing with
22705 blocks in DECL_DEBUG_EXPR locations.
22707 2013-03-21 Richard Biener <rguenther@suse.de>
22709 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22710 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22711 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22712 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22713 instead of DECL_DEBUG_EXPR_IS_FROM.
22714 * gimplify.c (gimplify_modify_expr): Likewise.
22715 * tree-cfg.c (verify_expr_location_1): Likewise.
22716 * tree-complex.c (create_one_component_var): Likewise.
22717 * tree-sra.c (create_access_replacement): Likewise.
22718 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22719 (clear_unused_block_pointer): Likewise.
22720 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22721 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22722 * var-tracking.c (var_debug_decl): Likewise.
22723 (track_expr_p): Likewise.
22724 * tree-inline.c (add_local_variables): Likewise. Set
22725 DECL_HAS_DEBUG_EXPR_P after copying it.
22726 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22727 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22729 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22732 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22733 * configure: Regenerate.
22734 * config.in: Regenerate.
22735 * config/i386/i386.md (*movdf_internal): Use
22736 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22737 movd instead of movq mnemonic for interunit moves.
22738 (*movdi_internal): Ditto.
22740 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22742 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22743 (abd<mode>_3): New pattern.
22744 (aba<mode>_3): New pattern.
22745 (fabd<mode>_3): New pattern.
22747 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22749 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22750 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22751 occurrence of REGISTER_PREFIX as its empty string.
22753 2013-03-20 Jeff Law <law@redhat.com>
22755 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22756 addititional equivalences for equality comparisons between an SSA_NAME
22757 and a constant where the SSA_NAME was set from a widening conversion.
22759 2013-03-20 Walter Lee <walt@tilera.com>
22761 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22763 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22765 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22766 depending on type attribute.
22767 (*movti_internal): Ditto.
22768 (*movtf_internal): Ditto.
22769 (*movxf_internal): Ditto.
22770 (*movdf_internal): Ditto.
22771 (*movsf_internal): Ditto.
22773 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22775 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22776 maybe_vex for sselog1 and ssemov types.
22777 (*movdi_internal): Reorder operand constraints.
22778 (*movsi_internal): Ditto. Set prefix attribute to
22779 maybe_vex for sselog1 and ssemov types.
22780 (*movtf_internal): Set prefix attribute to maybe_vex
22781 for sselog1 and ssemov types.
22782 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22783 DImode ssemov types. Reorder operand constraints.
22784 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22785 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22786 attribute for SImode ssemov types. Reorder operand constraints.
22788 2013-03-20 Martin Jambor <mjambor@suse.cz>
22790 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22791 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22793 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22795 * config/rs6000/predicates.md (indexed_address, update_address_mem
22796 update_indexed_address_mem): New predicates.
22797 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22798 attribute for load/store instructions.
22799 * config/rs6000/dfp.md (movsd_store): Likewise.
22800 (movsd_load): Likewise.
22801 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22802 (unnamed HI->DI extend define_insn): Likewise.
22803 (unnamed SI->DI extend define_insn): Likewise.
22804 (unnamed QI->SI extend define_insn): Likewise.
22805 (unnamed QI->HI extend define_insn): Likewise.
22806 (unnamed HI->SI extend define_insn): Likewise.
22807 (unnamed HI->SI extend define_insn): Likewise.
22808 (extendsfdf2_fpr): Likewise.
22809 (movsi_internal1): Likewise.
22810 (movsi_internal1_single): Likewise.
22811 (movhi_internal): Likewise.
22812 (movqi_internal): Likewise.
22813 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22814 attribute for load/store instructions.
22815 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22817 (mov<mode>_softfloat): Likewise.
22818 (mov<mode>_hardfloat32): Likewise.
22819 (mov<mode>_hardfloat64): Likewise.
22820 (mov<mode>_softfloat64): Likewise.
22821 (movdi_internal32): Likewise.
22822 (movdi_internal64): Likewise.
22823 (probe_stack_<mode>): Likewise.
22825 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22827 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22828 floating point, and decimal floating point to reload iterator.
22830 * config/rs6000/constraints.md (wl constraint): New constraints to
22831 return FLOAT_REGS if certain options are used to reduce the number
22832 of separate patterns that exist in the file.
22833 (wx constraint): Likewise.
22834 (wz constraint): Likewise.
22836 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22837 -mdebug=reg, print wg, wl, wx, and wz constraints.
22838 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22839 Initialize the reload functions for 64-bit binary/decimal floating
22841 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22842 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22843 create the buffer on the stack to overcome not having a 32-bit
22845 (rs6000_emit_move): Likewise.
22846 (rs6000_secondary_memory_needed_rtx): Likewise.
22847 (rs6000_alloc_sdmode_stack_slot): Likewise.
22848 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22849 via xxlxor, just like DFmode 0.0.
22851 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22852 define as 1 if we are running on a power7 or newer.
22853 (enum r6000_reg_class_enum): Add new constraints.
22855 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22856 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22857 with other moves by using conditional constraits (wg). Use LFIWZX
22858 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22859 (movsd splitter): Likewise.
22860 (movsd_hardfloat): Likewise.
22861 (movsd_softfloat): Likewise.
22863 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22864 binary and decimal floating point moves.
22865 (fmove_ok): New attributes to combine binary and decimal floating
22866 point moves, and to combine power6x (mfpgpr) moves along normal
22868 (real_value_to_target): Likewise.
22869 (f32_lr): Likewise.
22870 (f32_lm): Likewise.
22871 (f32_li): Likewise.
22872 (f32_sr): Likewise.
22873 (f32_sm): Likewise.
22874 (f32_si): Likewise.
22875 (movsf): Combine binary and decimal floating point moves. Combine
22876 power6x (mfpgpr) moves with other moves by using conditional
22877 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22878 (mov<mode> for SFmode/SDmode); Likewise.
22879 (SFmode/SDmode splitters): Likewise.
22880 (movsf_hardfloat): Likewise.
22881 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22882 (movsf_softfloat): Likewise.
22883 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22885 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22886 wx and wz constraints.
22888 * config/rs6000/constraints.md (wg constraint): New constraint to
22889 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22891 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
22894 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22895 -mdebug=reg, print wg, wl, wx, and wz constraints.
22896 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22897 Initialize the reload functions for 64-bit binary/decimal floating
22899 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22900 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22901 create the buffer on the stack to overcome not having a 32-bit
22903 (rs6000_emit_move): Likewise.
22904 (rs6000_secondary_memory_needed_rtx): Likewise.
22905 (rs6000_alloc_sdmode_stack_slot): Likewise.
22906 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22907 via xxlxor, just like DFmode 0.0.
22909 * config/rs6000/dfp.md (movdd): Delete, combine with binary
22910 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22911 with other moves by using conditional constraits (wg). Use LFIWZX
22912 and STFIWX for loading SDmode on power7.
22913 (movdd splitters): Likewise.
22914 (movdd_hardfloat32): Likewise.
22915 (movdd_softfloat32): Likewise.
22916 (movdd_hardfloat64_mfpgpr): Likewise.
22917 (movdd_hardfloat64): Likewise.
22918 (movdd_softfloat64): Likewise.
22920 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
22921 64-bit binary and decimal floating point moves.
22922 (FMOVE64X): Likewise.
22923 (movdf): Combine 64-bit binary and decimal floating point moves.
22924 Combine power6x (mfpgpr) moves with other moves by using
22925 conditional constraits (wg).
22926 (mov<mode> for DFmode/DDmode): Likewise.
22927 (DFmode/DDmode splitters): Likewise.
22928 (movdf_hardfloat32): Likewise.
22929 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
22930 (movdf_softfloat32): Likewise.
22931 (movdf_hardfloat64_mfpgpr): Likewise.
22932 (movdf_hardfloat64): Likewise.
22933 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
22934 (movdf_softfloat64): Likewise.
22935 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
22936 (reload_<mode>_load): Move to later in the file so they aren't in
22937 the middle of the floating point move insns.
22938 (reload_<mode>_store): Likewise.
22940 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
22943 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
22944 constraint if -mdebug=reg.
22945 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
22946 Enable using dd reload support if needed.
22948 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
22949 binary and decimal floating point moves in rs6000.md.
22950 (movtd_internal): Likewise.
22952 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
22953 decimal floating point moves.
22955 (movtf_internal): Likewise.
22956 (mov<mode>_internal, TDmode/TFmode): Likewise.
22957 (movtf_softfloat): Likewise.
22958 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
22960 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
22961 movdi_internal64, using wg constraint for move direct operations.
22962 (movdi_internal64): Likewise.
22964 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
22965 MODES_TIEABLE_P for selected modes. Print the numerical value of
22966 the various virtual registers. Use GPR/FPR first/last values,
22967 instead of hard coding the register numbers. Print which modes
22968 have reload functions registered.
22969 (rs6000_option_override_internal): If -mdebug=reg, trace the options
22970 settings before/after setting cpu, target and subtarget settings.
22971 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
22972 and for secondary reload failures in rs6000_secondary_reload_inner.
22973 (rs6000_secondary_reload_fail): Likewise.
22974 (rs6000_secondary_reload_inner): Likewise.
22976 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22977 macros for first/last GPR and FPR registers.
22978 (LAST_GPR_REGNO): Likewise.
22979 (FIRST_FPR_REGNO): Likewise.
22980 (LAST_FPR_REGNO): Likewise.
22982 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22983 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22984 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22985 (vcond<mode><mode>): Likewise.
22986 (vcondu<mode><mode>): Likewise.
22987 (vector_gtu<mode>): Likewise.
22988 (vector_gte<mode>): Likewise.
22989 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
22990 to prevent the compiler from converting DImode operations to TImode.
22991 (ior<mode>3): Likewise.
22992 (and<mode>3): Likewise.
22993 (one_cmpl<mode>2): Likewise.
22994 (nor<mode>3): Likewise.
22995 (andc<mode>3): Likewise.
22997 * config/rs6000/constraints.md (wt constraint): New constraint
22998 that returns VSX_REGS if TImode is allowed in VSX registers.
23000 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23001 constant under VSX.
23003 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23004 similar to TImode, but it is restricted to being in the GPRs.
23006 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23007 TImode to occupy a single VSX register.
23009 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23010 -mvsx-timode for power7/power8.
23011 (power7 cpu): Likewise.
23012 (power8 cpu): Likewise.
23014 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23015 sure that TFmode/TDmode take up two registers if they are ever
23016 allowed in the upper VSX registers.
23017 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23019 (rs6000_init_hard_regno_mode_ok): Likewise.
23020 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23021 constraint. Print if LRA is turned on.
23022 (rs6000_option_override_internal): Give an error if -mvsx-timode
23023 and VSX is not enabled.
23024 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23025 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23026 to reg+offset addressing. Use PTImode when checking offset
23027 addresses for validity.
23028 (reg_offset_addressing_ok_p): Likewise.
23029 (rs6000_legitimate_offset_address_p): Likewise.
23030 (rs6000_legitimize_address): Likewise.
23031 (rs6000_legitimize_reload_address): Likewise.
23032 (rs6000_legitimate_address_p): Likewise.
23033 (rs6000_eliminate_indexed_memrefs): Likewise.
23034 (rs6000_emit_move): Likewise.
23035 (rs6000_secondary_reload): Likewise.
23036 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23037 reloads to fpr registers to continue to use reg+offset addressing,
23038 but 64-bit reloads to altivec registers need reg+reg addressing.
23039 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23040 it. Treat LO_SUM like a PLUS operation.
23041 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23042 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23043 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23044 registers to share a register with a smaller sized type, since VSX
23045 puts scalars in the upper 64-bits.
23046 (print_operand): Add support for PTImode.
23047 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23048 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23049 registers, but don't have arithmetic support.
23050 (rs6000_memory_move_cost): Add test for VSX.
23051 (rs6000_opt_masks): Add -mvsx-timode.
23053 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23056 (VSr): Use wt constraint for TImode.
23057 (VSv): Drop TImode support.
23058 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23059 (vsx_movti_64bit): Likewise.
23060 (vsx_movti_32bit): Likewise.
23061 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23062 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23063 one '?' on the appropriate output constraint. Do not allow TImode
23064 logical operations on 32-bit systems.
23065 (vsx_ior<mode>3): Likewise.
23066 (vsx_xor<mode>3): Likewise.
23067 (vsx_one_cmpl<mode>2): Likewise.
23068 (vsx_nor<mode>3): Likewise.
23069 (vsx_andc<mode>3): Likewise.
23070 (vsx_concat_<mode>): Likewise.
23071 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23073 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23074 OPTION_MASK_VSX_TIMODE.
23075 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23076 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23078 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23079 (TI2 iterator): New iterator for TImode, PTImode.
23080 (wd mode attribute): Add values for vector types.
23081 (movti_string): Replace TI move operations with operations for TImode
23082 and PTImode. Add support for TImode being allowed in VSX registers.
23083 (mov<mode>_string, TImode/PTImode): Likewise.
23084 (movti_ppc64): Likewise.
23085 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23086 (TI mode splitters): Likewise.
23088 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23091 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23093 PR tree-optimization/56355
23094 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23095 Also handle integers with undefined overflow.
23097 2013-03-20 Catherine Moore <clm@codesourcery.com>
23098 Maciej W. Rozycki <macro@codesourcery.com>
23099 Tom de Vries <tom@codesourcery.com>
23100 Nathan Sidwell <nathan@codesourcery.com>
23101 Iain Sandoe <iain@codesourcery.com>
23102 Nathan Froyd <froydnj@codesourcery.com>
23103 Chao-ying Fu <fu@mips.com>
23105 * doc/extend.texi: (micromips, nomicromips, nocompression):
23106 Document new function attributes.
23107 * doc/invoke.texi (minterlink-compressed, mmicromips,
23108 m14k, m14ke, m14kec): Document new options.
23109 (minterlink-mips16): Update documentation.
23110 * doc/md.texi (ZC, ZD): Document new constraints.
23111 * configure.ac (gcc_cv_as_micromips): Check if linker
23112 supports the .set micromips directive.
23113 * configure: Regenerate.
23114 * config.in: Regenerate.
23115 * config/mips/mips-tables.opt: Regenerate.
23116 * config/mips/micromips.md: New file.
23117 * constraints.md (ZC, ZD): New constraints.
23118 * config/mips/predicates.md (movep_src_register): New predicate.
23119 (movep_src_operand): New predicate.
23120 (non_volatile_mem_operand): New predicate.
23121 * config/mips/mips.md (multimem): New type.
23122 (length): Differentiate between 17-bit and 18-bit branch offsets.
23123 (MOVEP1, MOVEP2): New mode iterator.
23124 (mov_<load>l): Use ZC constraint.
23125 (mov_<load>r): Likewise.
23126 (mov_<store>l): Likewise.
23127 (mov_<store>r): Likewise.
23128 (*branch_equality<mode>_inverted): Add microMIPS support.
23129 (*branch_equality<mode>): Likewise.
23130 (*jump_absolute): Likewise.
23131 (indirect_jump_<mode>): Likewise.
23132 (tablejump_<mode>): Likewise.
23133 (<optab>_internal): Likewise.
23134 (sibcall_internal): Likewise.
23135 (sibcall_value_internal): Likewise.
23136 (prefetch): Use constraint ZD.
23137 * config/mips/mips.opt (minterlink-compressed): New option.
23138 (minterlink-mips16): Now an alias for minterlink-compressed.
23139 (mmicromips): New option.
23140 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23141 (compare_and_swap_12): Likewise.
23142 (sync_add<mode>): Likewise.
23143 (sync_<optab>_12): Likewise.
23144 (sync_old_<optab>_12): Likewise.
23145 (sync_new_<optab>_12): Likewise.
23146 (sync_nand_12): Likewise.
23147 (sync_old_nand_12): Likewise.
23148 (sync_new_nand_12): Likewise.
23149 (sync_sub<mode>): Likewise.
23150 (sync_old_add<mode>): Likewise.
23151 (sync_old_sub<mode>): Likewise.
23152 (sync_new_add<mode>): Likewise.
23153 (sync_new_sub<mode>): Likewise.
23154 (sync_<optab><mode>): Likewise.
23155 (sync_old_<optab><mode>): Likewise.
23156 (sync_new_<optab><mode>): Likewise.
23157 (sync_nand<mode>): Likewise.
23158 (sync_old_nand<mode>): Likewise.
23159 (sync_new_nand<mode>): Likewise.
23160 (sync_lock_test_and_set<mode>): Likewise.
23161 (test_and_set_12): Likewise.
23162 (atomic_compare_and_swap<mode>): Likewise.
23163 (atomic_exchange<mode>_llsc): Likewise.
23164 (atomic_fetch_add<mode>_llsc): Likewise.
23165 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23166 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23167 (umips_save_restore_pattern_p): Likewise.
23168 (umips_load_store_pair_p): Likewise.
23169 (umips_output_load_store_pair): Likewise.
23170 (umips_movep_target_p): Likewise.
23171 (umips_12bit_offset_address_p): Likewise.
23172 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23173 (mips_base_mips16): Rename this...
23174 (mips_base_compression_flags): ...to this. Update all uses.
23175 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23176 (mips_mips16_decl_p): Delete.
23177 (mips_nomips16_decl_p): Delete.
23178 (mips_get_compress_on_flags): New function.
23179 (mips_get_compress_off_flags): New function.
23180 (mips_get_compress_mode): New function.
23181 (mips_get_compress_on_name): New function.
23182 (mips_get_compress_off_name): New function.
23183 (mips_insert_attributes): Support multiple compression types.
23184 (mips_merge_decl_attributes): Likewise.
23185 (umips_12bit_offset_address_p): New function.
23186 (mips_start_function_definition): Emit .set micromips directive.
23187 (mips_call_may_need_jalx_p): New function.
23188 (mips_function_ok_for_sibcall): Add microMIPS support.
23189 (mips_print_operand_punctuation): Support short delay slots and
23191 (umips_swm_mask, umips_swm_encoding): New.
23192 (umips_build_save_restore): New function.
23193 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23194 (was_mips16_p): Remove.
23195 (old_compression_mode): New.
23196 (mips_set_compression_mode): New function.
23197 (mips_set_current_function): Add microMIPS support.
23198 (mips_option_override): Likewise.
23199 (umips_save_restore_pattern_p): New function.
23200 (umips_output_save_restore): New function.
23201 (umips_load_store_pair_p_1): New function.
23202 (umips_load_store_pair_p): New function.
23203 (umips_output_load_store_pair_1): New function.
23204 (umips_output_load_store_pair): New function.
23205 (umips_movep_target_p) New function.
23206 (mips_prepare_pch_save): Add microMIPS support.
23207 * config/mips/mips.h (TARGET_COMPRESSION): New.
23208 (TARGET_CPU_CPP_BUILTINS): Update macro
23209 to use new compression flags and to support microMIPS.
23210 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23211 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23212 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23213 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23214 (ASM_SPEC): Support mmicromips and mno-micromips.
23215 (M16STORE_REG_P): New macro.
23216 (MIPS_CALL): Support TARGET_MICROMIPS.
23217 (MICROMIPS_J): New macro.
23218 (mips_base_mips16): Rename this...
23219 (mips_base_compression_flags): ...to this.
23220 (UMIPS_12BIT_OFFSET_P): New macro.
23221 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23222 (MULTILIB_DIRNAMES): Likewise.
23223 2013-03-20 Richard Biener <rguenther@suse.de>
23225 PR tree-optimization/56661
23226 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23227 the result does not have to be distinct.
23229 2013-03-20 Richard Biener <rguenther@suse.de>
23231 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23234 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23235 Steven Bosscher <steven@gcc.gnu.org>
23237 PR rtl-optimization/56605
23238 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23240 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23243 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23244 that require movd instead of movq.
23246 2013-03-20 Richard Biener <rguenther@suse.de>
23248 * tree-ssa-structalias.c (struct variable_info): Add pointer
23249 to the first field of an aggregate with sub-vars. Make
23250 this and the pointer to the next subfield its ID.
23251 (vi_next): New function.
23252 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23253 storedanything_id, integer_id): Increment by one.
23254 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23255 get_call_clobber_vi): Adjust.
23256 (solution_set_expand): Simplify and speedup.
23257 (solution_set_add): Inline into ...
23258 (set_union_with_increment): ... this. Adjust accordingly.
23259 (do_sd_constraint): Likewise.
23260 (do_ds_constraint): Likewise.
23261 (do_complex_constraint): Simplify.
23262 (build_pred_graph): Adjust.
23263 (solve_graph): Likewise. Simplify and speedup.
23264 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23265 get_constraint_for_component_ref, get_constraint_for_1,
23266 first_vi_for_offset, first_or_preceding_vi_for_offset,
23267 create_function_info_for, create_variable_info_for_1,
23268 create_variable_info_for, intra_create_variable_infos): Adjust.
23269 (init_base_vars): Push NULL for ID zero.
23270 (compute_points_to_sets): Adjust.
23272 2013-03-20 Richard Biener <rguenther@suse.de>
23274 * cfgloop.c (verify_loop_structure): Streamline and avoid
23275 ICEing on corrupt loop tree.
23276 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23279 2013-03-20 Richard Biener <rguenther@suse.de>
23281 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23282 check whether an SSA update is needed.
23284 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23286 * config/mips/constraints.md (T): Rename to...
23290 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23291 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23293 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23295 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23296 (*subsi3_carryin_uxtw): Likewise.
23298 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23300 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23301 (*rorsi3_insn_uxtw): Likewise.
23303 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23305 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23306 (*extrsi5_insn_uxtw): Likewise.
23308 2013-03-19 Richard Biener <rguenther@suse.de>
23310 PR tree-optimization/56273
23311 * passes.c (init_optimization_passes): Move second VRP after DOM.
23313 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23315 * config/i386/i386.md (*movti_internal): Merge from
23316 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23317 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23318 nox64 isa attributes.
23320 2013-03-18 Richard Biener <rguenther@suse.de>
23322 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23324 (merge_node_constraints): Likewise.
23325 (build_succ_graph): Likewise.
23326 (valid_graph_edge): Inline into single caller.
23327 (unify_nodes): Likewise. Use bitmap_set_bit return value
23329 (scc_visit): Fix formatting and variable use.
23330 (do_sd_constraint): Use gcc_checking_assert.
23331 (do_ds_constraint): Likewise.
23332 (do_complex_constraint): Likewise.
23333 (condense_visit): Likewise. Cleanup.
23334 (dump_pred_graph): New function.
23335 (perform_var_substitution): Dump the pred-graph before
23336 variable substitution.
23337 (find_equivalent_node): Use gcc_checking_assert.
23338 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23340 2013-03-18 Richard Biener <rguenther@suse.de>
23342 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23343 Remove cond_expr_stmt_list argument and do not gimplify the
23345 (vect_loop_versioning): Adjust.
23346 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23347 Cleanup to use less temporaries.
23348 (vect_create_data_ref_ptr): Cleanup.
23350 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23352 PR tree-optimization/56635
23353 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23354 require types_compatible_p types.
23356 2013-03-18 Nick Clifton <nickc@redhat.com>
23358 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23359 spurious backslash.
23361 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23362 Add missing line to comment describing function.
23364 2013-03-18 Richard Biener <rguenther@suse.de>
23366 PR tree-optimization/56210
23367 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23368 Handle string / character search functions.
23369 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23371 2013-03-18 Richard Biener <rguenther@suse.de>
23373 PR middle-end/56483
23374 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23375 and implement properly.
23376 * gimple.h (gimple_cond_single_var_p): Remove.
23378 2013-03-18 Richard Biener <rguenther@suse.de>
23380 * tree-data-ref.h (find_data_references_in_loop): Declare.
23381 * tree-data-ref.c (get_references_in_stmt): Use a stack
23382 vector pre-allocated in the callers.
23383 (find_data_references_in_stmt): Adjust.
23384 (graphite_find_data_references_in_stmt): Likewise.
23385 (create_rdg_vertices): Likewise.
23386 (find_data_references_in_loop): Export.
23387 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23388 Compute dependences here...
23389 (vect_analyze_data_refs): ...not here. When we encounter
23390 a non-vectorizable data reference in basic-block vectorization
23391 truncate the data reference vector. Do not bother to
23392 fixup data-dependence information for gather loads.
23393 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23394 of data references, as reported.
23396 2013-03-18 Richard Biener <rguenther@suse.de>
23398 PR tree-optimization/3713
23399 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23400 has_constants and expr.
23401 (stmt_has_constants): Properly valueize SSA names when deciding
23402 whether the stmt has constants.
23404 2013-03-18 Richard Biener <rguenther@suse.de>
23406 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23407 whole function when there is nothing to do.
23408 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23409 * tree-vectorizer.c (vectorize_loops): Update virtual and
23410 loop-closed SSA once.
23411 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23413 2013-03-18 Richard Biener <rguenther@suse.de>
23415 PR middle-end/56113
23416 * domwalk.c (bb_postorder): New global static.
23417 (cmp_bb_postorder): New function.
23418 (walk_dominator_tree): Replace scheme imposing an order for
23419 visiting dominator sons by one sorting them at the time they
23420 are pushed on the stack.
23422 2013-03-18 Richard Biener <rguenther@suse.de>
23424 PR tree-optimization/39326
23425 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23426 (struct mem_ref): Replace mem member with ao_ref typed member.
23427 (MEM_ANALYZABLE): Adjust.
23428 (memref_eq): Likewise.
23429 (mem_ref_alloc): Likewise.
23430 (gather_mem_refs_stmt): Likewise.
23431 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23432 (execute_sm_if_changed_flag_set): Adjust.
23433 (execute_sm): Likewise.
23434 (ref_always_accessed_p): Likewise.
23435 (refs_independent_p): Likewise.
23436 (can_sm_ref_p): Likewise.
23438 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23441 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23442 return 1 even for !unsignedp.
23444 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23446 * config/i386/i386.md (isa): Add x64 and nox64.
23447 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23448 (*pushtf): Enable *roF alternative for x64 isa only.
23449 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23450 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23451 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23452 (*movtf_internal): Merge from *movtf_internal_rex64 and
23453 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23454 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23455 nox64 isa attributes.
23456 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23457 nox64 isa attributes.
23458 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23460 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23462 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23464 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23467 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23468 class names. Remove trailing comma after "ALL_REGS".
23470 2013-03-16 Jan Hubicka <jh@suse.cz>
23472 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23473 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23474 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23475 of cgraph_get_create_node.
23476 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23478 2013-03-16 Jason Merrill <jason@redhat.com>
23481 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23482 with DW_AT_default_value.
23484 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23486 * BASE-VER: Set to 4.9.0.
23488 2013-03-14 Andi Kleen <ak@linux.intel.com>
23491 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23492 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23493 Document _x* TSX intrinsics.
23495 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23496 David Holsgrove <david.holsgrove@xilinx.com>
23498 * configure.ac: Add MicroBlaze TLS support detection.
23499 * configure: Regenerate.
23500 * config/microblaze/microblaze-protos.h
23501 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23502 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23503 * config/microblaze/microblaze.c (microblaze_address_type): Add
23504 ADDRESS_TLS and tls_reloc address types.
23505 (microblaze_address_info): Add tls_reloc.
23506 (TARGET_HAVE_TLS): Define.
23507 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23508 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23509 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23510 load_tls_operand, microblaze_call_tls_get_addr,
23511 microblaze_legitimize_tls_address): New functions.
23512 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23513 (get_base_reg): Use microblaze_tls_symbol_p.
23514 (microblaze_classify_address): Handle TLS.
23515 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23516 label_mentioned_p and microblaze_tls_referenced_p.
23517 (microblaze_legitimize_address): Handle TLS.
23518 (microblaze_address_insns): Handle ADDRESS_TLS.
23519 (pic_address_needs_scratch): Handle TLS.
23520 (print_operand_address): Handle TLS.
23521 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23522 (microblaze_expand_move): Handle TLS.
23523 (microblaze_legitimate_constant_p): Check
23524 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23525 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23526 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23527 (PIC_OFFSET_TABLE_REGNUM): Set.
23528 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23529 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23530 (addsi3, movsi_internal2, movdf_internal): Update constraints
23531 * config/microblaze/predicates.md (arith_plus_operand): Define
23532 (move_operand): Redefine as move_src_operand,
23533 check microblaze_tls_referenced_p.
23535 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23537 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23538 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23540 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23542 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23545 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23547 PR tree-optimization/53265
23548 * common.opt (Waggressive-loop-optimizations): New option.
23549 * tree-ssa-loop-niter.c: Include tree-pass.h.
23550 (do_warn_aggressive_loop_optimizations): New function.
23551 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23552 if number_of_latch_executions returned constant.
23553 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23554 early. If number_of_latch_executions returned constant, set
23555 nb_iterations_upper_bound back to it.
23556 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23558 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23559 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23561 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23562 (MULTILIB_OSDIRNAMES): Set.
23563 * genmultilib: If defaultosdirname doesn't start with :: , set
23564 defaultosdirname2 instead, clear it and emit two . multilib_raw
23565 entries instead of just one.
23567 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23569 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23570 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23571 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23572 (SUBTARGET_OVERRIDE_OPTIONS): New.
23574 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23577 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23578 (musermode): Convert to Var(TARGET_USERMODE).
23579 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23580 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23581 * config/sh/sh.c (sh_option_override): Use
23582 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23583 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23585 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23587 (udivsi3_i4_single, divsi3_i4_single): Use
23588 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23590 2013-03-13 Dave Korn <dave.korn.cygwin@....>
23592 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23595 2013-03-13 Richard Biener <rguenther@suse.de>
23597 PR tree-optimization/56608
23598 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23599 calls when vectorizing basic-blocks.
23601 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23604 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23607 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23609 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23611 2013-03-11 Jan Hubicka <jh@suse.cz>
23614 * lto-streamer-out.c (output_symbol_p): Skip references from
23615 constructors of external variables.
23617 2013-03-11 Jan Hubicka <jh@suse.cz>
23619 PR middle-end/56571
23620 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23622 * emit-rtl.c (verify_rtx_sharing): Likewise.
23623 (copy_insn_1): Likewise.
23624 * rtl.c (copy_rtx): Likewise.
23626 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23629 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23630 output_operand_lossage message.
23632 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23635 * arm.c (shift_op): Validate RTL pattern on the fly.
23636 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23639 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23642 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23643 new call patterns that clobber %fr12.
23644 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23645 split and postreload patterns.
23646 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23647 registers %fr12 and %fr12R as call used.
23649 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23651 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23652 (canon_address, record_store, replace_read, check_mem_read_rtx,
23653 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23654 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23655 rest_of_handle_dse): Likewise.
23657 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23659 PR middle-end/56524
23660 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23662 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23663 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23664 (save_optabs_if_changed): Replace with...
23665 (init_tree_optimization_optabs): ...this.
23666 * optabs.c (save_optabs_if_changed): Rename to...
23667 (init_tree_optimization_optabs): ...this. Take the optimization node
23668 as argument. Do nothing if the base optabs are already correct.
23669 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23670 to recompute optabs.
23671 * function.h (function): Remove optabs field.
23672 * function.c (invoke_set_current_function_hook): Call
23673 init_tree_optimization_optabs. Use the result to initialize
23676 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23678 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23679 if GTMA_HAS_NO_INSTRUMENTATION.
23680 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23681 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23682 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23683 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23684 GTMA_HAS_NO_INSTRUMENTATION.
23686 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23688 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23691 2013-03-08 Marek Polacek <polacek@redhat.com>
23692 Jakub Jelinek <jakub@redhat.com>
23694 PR tree-optimization/56478
23695 * predict.c (is_comparison_with_loop_invariant_p): Change the
23696 type of loop_step to tree.
23697 (predict_loops): Adjust.
23698 (predict_iv_comparison): Perform the computations on double_ints.
23700 2013-03-08 Richard Biener <rguenther@suse.de>
23702 PR tree-optimization/56570
23703 * tree-cfg.c (verify_expr_location_1): Verify locations for
23705 * tree-sra.c (create_access_replacement): Strip locations
23706 from DECL_DEBUG_EXPRs.
23708 2013-03-08 Richard Biener <rguenther@suse.de>
23710 * tree-inline.c (expand_call_inline): Do not associate
23711 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23712 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23714 2013-03-08 Richard Biener <rguenther@suse.de>
23716 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23717 or block changes with -Og. Fix for location / block encoding
23718 changes and PHI arguments with locations.
23720 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23722 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23724 (struct output_info): Likewise.
23725 (register_overhead): Remove bad gcc_assert.
23726 (bitmap_find_bit): If there is only a single bitmap element, do not
23727 count a miss as a search.
23728 (print_statistics): Update for counter type changes.
23729 (dump_bitmap_statistics): Likewise. Print headers such that they
23730 are properly lined up with the printed counters.
23732 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23734 PR tree-optimization/56559
23735 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23736 check that it has only a single use.
23738 2013-03-07 Richard Biener <rguenther@suse.de>
23740 * doc/invoke.texi (fwhole-program): Discourage use in combination
23743 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23745 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23747 PR tree-optimization/56539
23748 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23749 instead of GSI_CONTINUE_LINKING as last argument to
23750 force_gimple_operand_gsi. Adjust function comment.
23752 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23755 PR middle-end/56548
23756 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23757 promoted mode, convert the result back to the original mode.
23759 2013-03-06 Richard Biener <rguenther@suse.de>
23761 PR middle-end/56294
23762 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23763 (insert_updated_phi_nodes_compare_uids): New function.
23764 (update_ssa): Sort symbols_to_rename after UID before
23765 traversing it to insert PHI nodes.
23767 2013-03-06 Richard Biener <rguenther@suse.de>
23769 PR middle-end/50494
23770 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23771 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23774 2013-02-13 Richard Biener <rguenther@suse.de>
23777 * varasm.c (output_constant_def_1): Get the decl representing
23778 the constant as argument.
23779 (output_constant_def): Wrap output_constant_def_1.
23780 (make_decl_rtl): Use output_constant_def_1 with the decl
23781 representing the constant.
23782 (build_constant_desc): Optionally re-use a decl already
23783 representing the constant.
23784 (tree_output_constant_def): Adjust.
23786 2013-03-06 Joey Ye <joey.ye@arm.com>
23789 * gcc.c (convert_white_space): New function.
23790 (main): Handles white space in function name.
23792 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23795 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23796 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23797 to SH_DIV_CALL_TABLE for TARGET_SH2.
23798 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23800 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23801 call-table options.
23803 2013-03-05 Sterling Augustine <saugustine@google.com>
23804 Cary Coutant <ccoutant@google.com>
23807 * dwarf2out.c (resolve_addr): Don't call
23808 remove_loc_list_addr_table_entries a second time for the same
23811 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23814 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23815 (avoid_complex_debug_insns): New function.
23816 (expand_debug_locations): Call it.
23818 PR rtl-optimization/56484
23819 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23820 lifetimes of hard registers on small register class machines.
23822 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23824 * config/microblaze/microblaze-protos.h: Rename
23825 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23826 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23828 (microblaze_fast_interrupt_function_p): New function.
23829 (microblaze_is_interrupt_handler): Rename to
23830 microblaze_is_interrupt_variant and add fast_interrupt check.
23831 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23832 (save_restore_insns): Likewise.
23833 (compute_frame_size): Likewise.
23834 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23835 (microblaze_globalize_label): Likewise.
23836 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23837 * config/microblaze/microblaze.md: Use wrapper
23838 microblaze_is_interrupt_variant.
23840 2013-03-05 Kai Tietz <ktietz@redhat.com>
23842 * sdbout.c (sdbout_one_type): Switch to current function's section
23843 supporting cold/hot.
23845 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23847 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23850 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23852 PR middle-end/56461
23853 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23854 if VALGRIND_GET_VBITS is defined, temporarily make object
23855 memory all defined, and restore previous valgrind addressability
23856 and definability afterwards. Free this_object at the end.
23858 PR middle-end/56461
23859 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23860 right before calling lra_create_live_ranges, also call it
23861 when clearing live_p. Only call lra_clear_live_ranges
23862 at the end if live_p.
23864 PR middle-end/56461
23865 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23867 2013-03-05 Richard Biener <rguenther@suse.de>
23869 PR tree-optimization/56521
23870 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23873 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23876 * except.h (remove_unreachable_eh_regions): New prototype.
23877 * except.c (remove_eh_handler_splicer): New function, split out
23878 of remove_eh_handler.
23879 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23880 warning about running it on many EH regions one at a time.
23881 (remove_unreachable_eh_regions_worker): New function, walk the
23882 EH tree in depth-first order and remove non-marked regions.
23883 (remove_unreachable_eh_regions): New function.
23884 * tree-eh.c (mark_reachable_handlers): New function, split out
23885 from remove_unreachable_handlers.
23886 (remove_unreachable_handlers): Use mark_reachable_handlers and
23887 remove_unreachable_eh_regions.
23888 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23889 and remove_unreachable_eh_regions.
23891 2013-03-05 Richard Biener <rguenther@suse.de>
23893 PR middle-end/56525
23894 * loop-init.c (fix_loop_structure): Remove loops in two stages,
23895 not freeing them until the end.
23897 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23899 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
23901 2013-03-05 Richard Biener <rguenther@suse.de>
23903 PR tree-optimization/56270
23904 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
23905 of loads after scheduling an SLP instance.
23907 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23909 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
23911 (check_gcc_parallelize): Run guality.exp as a separate job from
23912 vect.exp with unsorted.exp and $(dg_target_exps) separately from
23913 struct-layout-1.exp with stackalign.exp.
23915 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
23917 PR middle-end/56461
23918 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
23919 load_index sbitmap even if some bit in it isn't set.
23921 PR middle-end/56461
23922 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
23923 (discover_iteration_bound_by_body_walk): Change queues to
23924 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
23925 spelling in comment. Call safe_push on queues[bound_index] directly.
23926 Release queues[queue_index] in every iteration unconditionally.
23927 Release bounds vector.
23929 PR middle-end/56461
23930 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
23931 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
23932 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
23935 2013-03-05 Richard Biener <rguenther@suse.de>
23938 * tree-inline.c (remap_blocks_to_null): New function.
23939 (expand_call_inline): When expanding a call stmt without
23940 an associated block inline remap all callee blocks to NULL.
23942 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23944 PR rtl-optimization/56494
23945 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
23946 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
23947 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
23949 PR middle-end/56461
23950 * sel-sched-ir.c (free_sched_pools): Release
23951 succs_info_pool.stack[succs_info_pool.max_top] vectors too
23952 if succs_info_pool.max_top isn't -1.
23955 * opts.c (opts_obstack, opts_concat): Moved to...
23956 * opts-common.c (opts_obstack, opts_concat): ... here.
23958 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23960 PR middle-end/56461
23961 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
23963 2013-03-04 Martin Jambor <mjambor@suse.cz>
23965 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
23966 all appropriate places.
23968 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
23970 PR tree-optimization/56424
23971 * ipa-split.c (split_function): Do not set the RSO flag if result is
23972 not by reference and its type is a register type.
23974 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
23976 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
23977 (microblaze_legitimate_pic_operand): Likewise
23978 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
23979 new function microblaze_legitimate_pic_operand
23980 * config/microblaze/microblaze-protos.h
23981 (microblaze_legitimate_pic_operand): Declare.
23983 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23985 * config/microblaze/predicates.md (call_insn_simple_operand):
23986 New predicate for supported rtx code types.
23987 * config/microblaze/microblaze.md (call_internal1): Use
23988 call_insn_simple_operand predicate.
23990 2013-03-04 Jakub Jelinek <jakub@redhat.com>
23992 PR middle-end/56461
23993 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23994 partitions.ordered_remove.
23996 PR middle-end/56461
23997 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23998 vec_oprnds0.create (1) for modifier == NONE.
24000 PR middle-end/56461
24001 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24002 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24003 vec_oprnds1 right before pushing anything to it for
24006 PR middle-end/56461
24007 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24008 set nbbs to 0 instead of having separate code path.
24009 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24010 instead of false as last argument if returning NULL.
24012 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24014 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24015 the attribute is now called "target" instead of "option".
24016 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24017 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24018 attribute/pragma name for TARGET_OPTION_VALID_P and
24019 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24020 * doc/tm.texi: Regenerated.
24022 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24024 * config/microblaze/microblaze.c:
24025 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24026 * config/microblaze/microblaze.h: Add -mxl-reorder to
24028 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24029 instructions emitted if TARGET_REORDER.
24030 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24031 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24034 2013-03-01 Xinliang David Li <davidxl@google.com>
24036 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24039 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24041 PR middle-end/56461
24042 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24043 vector even when returning true. Fix up function comment formatting.
24045 PR middle-end/56461
24046 * ira-build.c (ira_loop_nodes_count): New variable.
24047 (create_loop_tree_nodes): Initialize it.
24048 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24050 PR middle-end/56461
24051 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24052 method on dr_chain and result_chain.
24053 * tree-vect-stmts.c (vectorizable_store): Only call
24054 result_chain.create if j == 0.
24056 PR middle-end/56461
24057 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24058 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24059 before overwriting it.
24061 2013-03-01 Tobias Burnus <burnus@net-b.de>
24063 * doc/extended.texi (C Extensions): Change order in @menu
24065 (Other MIPS Built-in Functions): Move last MIPS entry before
24066 "picoChip Built-in Functions".
24067 (SH Built-in Functions): Move after RX Built-in Functions.
24068 * doc/gcc.texi (Introduction): Change order in @menu
24070 * doc/md.texi (Constraints): Ditto.
24071 * gty.texi (Type Information): Ditto.
24072 (User-provided marking routines for template types): Make
24074 * doc/invoke.texi (AArch64 Options): Move before
24075 "Adapteva Epiphany Options".
24077 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24078 Jakub Jelinek <jakub@redhat.com>
24081 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24082 no_address_safety_analysis attribute.
24083 * doc/extend.texi (no_address_safety_attribute): Rename to
24084 no_sanitize_address attribute, mention no_address_safety_analysis
24085 attribute as deprecated alias.
24087 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24089 PR middle-end/56461
24090 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24091 type to vec<vec<tree> > *.
24092 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24093 to be vec<tree> instead of vec<tree> *, set vec_defs
24094 to vNULL and call vec_defs.create (number_of_vects), adjust other
24096 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24097 vectorizable_condition): Adjust vect_get_slp_defs callers.
24099 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24101 * config/aarch64/aarch64.c
24102 (aarch64_float_const_representable): Remove unused variable.
24104 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24106 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24108 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24110 * config/aarch64/aarch64-builtins.c
24111 (aarch64_init_simd_builtins): Make static.
24113 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24115 * config/aarch64/aarch64.c
24116 (aarch64_simd_make_constant): Make static.
24118 2013-02-28 Martin Jambor <mjambor@suse.cz>
24120 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24121 with no initialization to the RHS of debug statements.
24123 2013-02-28 Martin Jambor <mjambor@suse.cz>
24125 PR tree-optimization/56294
24126 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24128 (get_access_replacement): Do not call create_access_replacement.
24129 Assert a replacement exists.
24130 (get_repl_default_def_ssa_name): Create the replacement declaration
24133 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24135 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24136 final_end_function.
24138 2013-02-28 Marek Polacek <polacek@redhat.com>
24140 PR rtl-optimization/56466
24141 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24142 if we're changing a loop.
24143 (peel_loops_completely): Likewise.
24145 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24148 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24150 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24153 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24154 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24155 INTX_FTYPE_FX, FX_FTYPE_INTX.
24156 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24158 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24160 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24161 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24162 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24163 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24164 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24165 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24166 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24167 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24168 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24169 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24170 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24171 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24172 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24173 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24174 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24175 (avrxmega6): Increase max flash segments from 5 to 6.
24176 * config/avr/t-multilib: Regenerate.
24177 * config/avr/avr-tables.opt: Regenerate.
24178 * doc/avr-mmcu.texi: Regenerate.
24180 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24182 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24183 (avr_device_to_arch): Rename to avr_device_to_ld.
24184 (avr_device_to_as): New prototype.
24185 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24186 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24187 * config/avr/driver-avr.c (avr_device_to_as): New.
24188 (avr_device_to_arch): Rename to avr_device_to_ld.
24190 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24192 PR middle-end/56461
24193 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24194 method on dr_chain and result_chain.
24196 PR middle-end/56461
24197 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24198 pointer_set_destroy on not_executed_last_iteration.
24200 PR middle-end/56461
24201 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24203 PR middle-end/56461
24204 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24205 FOR_EACH_DEFINED_FUNCTION when freeing state.
24207 PR middle-end/56461
24208 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24210 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24213 PR middle-end/56461
24214 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24215 known_aggs[i].items and release known_aggs vector.
24217 PR middle-end/56461
24218 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24220 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24222 * config/microblaze/microblaze.c (microblaze_emit_compare):
24223 Use xor for EQ/NE comparisions.
24224 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24225 (cbranchsf4): Adjust operator to comparison_operator.
24227 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24229 PR middle-end/56461
24230 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24232 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24233 vec_safe_push, always update *slot.
24234 (redirect_edge_var_map_clear): Use vec_free.
24235 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24236 (free_var_map_entry): Use vec_free.
24237 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24238 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24240 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24242 PR middle-end/45472
24243 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24244 when the may_trap_p bit of the exprs being merged differs.
24245 Reorder tests for speculativeness in the logical and operator.
24247 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24249 * incpath.c (add_standard_paths): Use reconcat instead of concat
24250 where appropriate and avoid leaking memory.
24252 * opts.h: Include obstack.h.
24253 (opts_concat): New prototype.
24254 (opts_obstack): New declaration.
24255 * opts.c (opts_concat): New function.
24256 (opts_obstack): New variable.
24257 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24258 (finish_options): Use opts_concat instead of concat
24259 and XOBNEWVEC instead of XNEWVEC.
24260 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24261 generate_option): Likewise.
24262 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24263 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24266 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24267 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24269 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24271 PR middle-end/56461
24272 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24274 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24276 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24277 (arm_block_move_unaligned_straight): Likewise.
24278 (arm_adjust_block_mem): Likewise.
24280 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24283 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24284 temp, cond and label.
24285 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24288 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24289 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24292 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24293 of prologue/epilogue functions.
24296 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24297 Remove unused variable cfa_offset.
24298 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24301 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24302 type promotion to unsigned.
24305 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24306 for HOST_WIDE_INT of 32 bit / same size as int.
24307 (arm_block_move_unaligned_straight): Likewise.
24308 (arm_adjust_block_mem): Likewise.
24311 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24314 2013-02-26 Marek Polacek <polacek@redhat.com>
24316 PR tree-optimization/56426
24317 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24319 2013-02-26 Richard Biener <rguenther@suse.de>
24322 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24323 unused variable loops.
24325 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24327 PR tree-optimization/56448
24328 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24329 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24330 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24331 later operands of the references, or even first operand for
24332 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24334 PR tree-optimization/56443
24335 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24336 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24337 to type_for_mode langhook.
24339 2013-02-25 Matt Turner <mattst88@gmail.com>
24341 * doc/invoke.texi: Document r4700.
24343 2013-02-25 Richard Biener <rguenther@suse.de>
24345 PR tree-optimization/56175
24346 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24348 (simplify_bitwise_binary): ... here. Also guard the conversion
24349 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24351 2013-02-25 Catherine Moore <clm@codesourcery.com>
24354 2013-02-24 Catherine Moore <clm@codesourcery.com>
24355 Maciej W. Rozycki <macro@codesourcery.com>
24356 Tom de Vries <tom@codesourcery.com>
24357 Nathan Sidwell <nathan@codesourcery.com>
24358 Iain Sandoe <iain@codesourcery.com>
24359 Nathan Froyd <froydnj@codesourcery.com>
24360 Chao-ying Fu <fu@mips.com>
24362 * doc/extend.texi: (micromips, nomicromips, nocompression):
24363 Document new function attributes.
24364 * doc/invoke.texi (minterlink-compressed, mmicromips,
24365 m14k, m14ke, m14kec): Document new options.
24366 (minterlink-mips16): Update documentation.
24367 * doc/md.texi (ZC, ZD): Document new constraints.
24368 * configure.ac (gcc_cv_as_micromips): Check if linker
24369 supports the .set micromips directive.
24370 * configure: Regenerate.
24371 * config.in: Regenerate.
24372 * config/mips/mips-tables.opt: Regenerate.
24373 * config/mips/micromips.md: New file.
24374 * constraints.md (ZC, AD): New constraints.
24375 * config/mips/predicates.md (movep_src_register): New predicate.
24376 (movep_src_operand): New predicate.
24377 (non_volatile_mem_operand): New predicate.
24378 * config/mips/mips.md (multimem): New type.
24379 (length): Differentiate between 17-bit and 18-bit branch offsets.
24380 (MOVEP1, MOVEP2): New mode iterator.
24381 (mov_<load>l): Use ZC constraint.
24382 (mov_<load>r): Likewise.
24383 (mov_<store>l): Likewise.
24384 (mov_<store>r): Likewise.
24385 (*branch_equality<mode>_inverted): Add microMIPS support.
24386 (*branch_equality<mode>): Likewise.
24387 (*jump_absolute): Likewise.
24388 (indirect_jump_<mode>): Likewise.
24389 (tablejump_<mode>): Likewise.
24390 (<optab>_internal): Likewise.
24391 (sibcall_internal): Likewise.
24392 (sibcall_value_internal): Likewise.
24393 (prefetch): Use constraint ZD.
24394 * config/mips/mips.opt (minterlink-compressed): New option.
24395 (minterlink-mips16): Now an alias for minterlink-compressed.
24396 (mmicromips): New option.
24397 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24398 (compare_and_swap_12): Likewise.
24399 (sync_add<mode>): Likewise.
24400 (sync_<optab>_12): Likewise.
24401 (sync_old_<optab>_12): Likewise.
24402 (sync_new_<optab>_12): Likewise.
24403 (sync_nand_12): Likewise.
24404 (sync_old_nand_12): Likewise.
24405 (sync_new_nand_12): Likewise.
24406 (sync_sub<mode>): Likewise.
24407 (sync_old_add<mode>): Likewise.
24408 (sync_old_sub<mode>): Likewise.
24409 (sync_new_add<mode>): Likewise.
24410 (sync_new_sub<mode>): Likewise.
24411 (sync_<optab><mode>): Likewise.
24412 (sync_old_<optab><mode>): Likewise.
24413 (sync_new_<optab><mode>): Likewise.
24414 (sync_nand<mode>): Likewise.
24415 (sync_old_nand<mode>): Likewise.
24416 (sync_new_nand<mode>): Likewise.
24417 (sync_lock_test_and_set<mode>): Likewise.
24418 (test_and_set_12): Likewise.
24419 (atomic_compare_and_swap<mode>): Likewise.
24420 (atomic_exchange<mode>_llsc): Likewise.
24421 (atomic_fetch_add<mode>_llsc): Likewise.
24422 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24423 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24424 (umips_save_restore_pattern_p): Likewise.
24425 (umips_load_store_pair_p): Likewise.
24426 (umips_output_load_store_pair): Likewise.
24427 (umips_movep_target_p): Likewise.
24428 (umips_12bit_offset_address_p): Likewise.
24429 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24430 (mips_base_mips16): Rename this...
24431 (mips_base_compression_flags): ...to this. Update all uses.
24432 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24433 (mips_mips16_decl_p): Delete.
24434 (mips_nomips16_decl_p): Delete.
24435 (mips_get_compress_on_flags): New function.
24436 (mips_get_compress_off_flags): New function.
24437 (mips_get_compress_mode): New function.
24438 (mips_get_compress_on_name): New function.
24439 (mips_get_compress_off_name): New function.
24440 (mips_insert_attributes): Support multiple compression types.
24441 (mips_merge_decl_attributes): Likewise.
24442 (umips_12bit_offset_address_p): New function.
24443 (mips_start_function_definition): Emit .set micromips directive.
24444 (mips_call_may_need_jalx_p): New function.
24445 (mips_function_ok_for_sibcall): Add microMIPS support.
24446 (mips_print_operand_punctuation): Support short delay slots and
24448 (umips_swm_mask, umips_swm_encoding): New.
24449 (umips_build_save_restore): New function.
24450 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24451 (was_mips16_p): Remove.
24452 (old_compression_mode): New.
24453 (mips_set_compression_mode): New function.
24454 (mips_set_current_function): Add microMIPS support.
24455 (mips_option_override): Likewise.
24456 (umips_save_restore_pattern_p): New function.
24457 (umips_output_save_restore): New function.
24458 (umips_load_store_pair_p_1): New function.
24459 (umips_load_store_pair_p): New function.
24460 (umips_output_load_store_pair_1): New function.
24461 (umips_output_load_store_pair): New function.
24462 (umips_movep_target_p) New function.
24463 (mips_prepare_pch_save): Add microMIPS support.
24464 * config/mips/mips.h (TARGET_COMPRESSION): New.
24465 (TARGET_CPU_CPP_BUILTINS): Update macro
24466 to use new compression flags and to support microMIPS.
24467 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24468 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24469 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24470 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24471 (ASM_SPEC): Support mmicromips and mno-micromips.
24472 (M16STORE_REG_P): New macro.
24473 (MIPS_CALL): Support TARGET_MICROMIPS.
24474 (MICROMIPS_J): New macro.
24475 (mips_base_mips16): Rename this...
24476 (mips_base_compression_flags): ...to this.
24477 (UMIPS_12BIT_OFFSET_P): New macro.
24478 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24479 (MULTILIB_DIRNAMES): Likewise.
24481 2013-02-25 Tom de Vries <tom@codesourcery.com>
24483 PR rtl-optimization/56131
24484 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24485 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24486 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24488 2013-02-25 Tobias Burnus <burnus@net-b.de>
24490 * doc/invoke.texi (-fsanitize=): Move from optimization
24491 to debugging options.
24493 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24495 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24497 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24498 Alexander Monakov <amonakov@ispras.ru>
24500 PR middle-end/56077
24501 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24502 flush pending lists also on non-jumps. Adjust comment.
24504 2013-02-24 Catherine Moore <clm@codesourcery.com>
24505 Maciej W. Rozycki <macro@codesourcery.com>
24506 Tom de Vries <tom@codesourcery.com>
24507 Nathan Sidwell <nathan@codesourcery.com>
24508 Iain Sandoe <iain@codesourcery.com>
24509 Nathan Froyd <froydnj@codesourcery.com>
24510 Chao-ying Fu <fu@mips.com>
24512 * doc/extend.texi: (micromips, nomicromips, nocompression):
24513 Document new function attributes.
24514 * doc/invoke.texi (minterlink-compressed, mmicromips,
24515 m14k, m14ke, m14kec): Document new options.
24516 (minterlink-mips16): Update documentation.
24517 * doc/md.texi (ZC, ZD): Document new constraints.
24518 * configure.ac (gcc_cv_as_micromips): Check if linker
24519 supports the .set micromips directive.
24520 * configure: Regenerate.
24521 * config.in: Regenerate.
24522 * config/mips/mips-tables.opt: Regenerate.
24523 * config/mips/micromips.md: New file.
24524 * constraints.md (ZC, AD): New constraints.
24525 * config/mips/predicates.md (movep_src_register): New predicate.
24526 (movep_src_operand): New predicate.
24527 (non_volatile_mem_operand): New predicate.
24528 * config/mips/mips.md (multimem): New type.
24529 (length): Differentiate between 17-bit and 18-bit branch offsets.
24530 (MOVEP1, MOVEP2): New mode iterator.
24531 (mov_<load>l): Use ZC constraint.
24532 (mov_<load>r): Likewise.
24533 (mov_<store>l): Likewise.
24534 (mov_<store>r): Likewise.
24535 (*branch_equality<mode>_inverted): Add microMIPS support.
24536 (*branch_equality<mode>): Likewise.
24537 (*jump_absolute): Likewise.
24538 (indirect_jump_<mode>): Likewise.
24539 (tablejump_<mode>): Likewise.
24540 (<optab>_internal): Likewise.
24541 (sibcall_internal): Likewise.
24542 (sibcall_value_internal): Likewise.
24543 (prefetch): Use constraint ZD.
24544 * config/mips/mips.opt (minterlink-compressed): New option.
24545 (minterlink-mips16): Now an alias for minterlink-compressed.
24546 (mmicromips): New option.
24547 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24548 (compare_and_swap_12): Likewise.
24549 (sync_add<mode>): Likewise.
24550 (sync_<optab>_12): Likewise.
24551 (sync_old_<optab>_12): Likewise.
24552 (sync_new_<optab>_12): Likewise.
24553 (sync_nand_12): Likewise.
24554 (sync_old_nand_12): Likewise.
24555 (sync_new_nand_12): Likewise.
24556 (sync_sub<mode>): Likewise.
24557 (sync_old_add<mode>): Likewise.
24558 (sync_old_sub<mode>): Likewise.
24559 (sync_new_add<mode>): Likewise.
24560 (sync_new_sub<mode>): Likewise.
24561 (sync_<optab><mode>): Likewise.
24562 (sync_old_<optab><mode>): Likewise.
24563 (sync_new_<optab><mode>): Likewise.
24564 (sync_nand<mode>): Likewise.
24565 (sync_old_nand<mode>): Likewise.
24566 (sync_new_nand<mode>): Likewise.
24567 (sync_lock_test_and_set<mode>): Likewise.
24568 (test_and_set_12): Likewise.
24569 (atomic_compare_and_swap<mode>): Likewise.
24570 (atomic_exchange<mode>_llsc): Likewise.
24571 (atomic_fetch_add<mode>_llsc): Likewise.
24572 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24573 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24574 (umips_save_restore_pattern_p): Likewise.
24575 (umips_load_store_pair_p): Likewise.
24576 (umips_output_load_store_pair): Likewise.
24577 (umips_movep_target_p): Likewise.
24578 (umips_12bit_offset_address_p): Likewise.
24579 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24580 (mips_base_mips16): Rename this...
24581 (mips_base_compression_flags): ...to this. Update all uses.
24582 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24583 (mips_mips16_decl_p): Delete.
24584 (mips_nomips16_decl_p): Delete.
24585 (mips_get_compress_on_flags): New function.
24586 (mips_get_compress_off_flags): New function.
24587 (mips_get_compress_mode): New function.
24588 (mips_get_compress_on_name): New function.
24589 (mips_get_compress_off_name): New function.
24590 (mips_insert_attributes): Support multiple compression types.
24591 (mips_merge_decl_attributes): Likewise.
24592 (umips_12bit_offset_address_p): New function.
24593 (mips_start_function_definition): Emit .set micromips directive.
24594 (mips_call_may_need_jalx_p): New function.
24595 (mips_function_ok_for_sibcall): Add microMIPS support.
24596 (mips_print_operand_punctuation): Support short delay slots and
24598 (umips_swm_mask, umips_swm_encoding): New.
24599 (umips_build_save_restore): New function.
24600 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24601 (was_mips16_p): Remove.
24602 (old_compression_mode): New.
24603 (mips_set_compression_mode): New function.
24604 (mips_set_current_function): Add microMIPS support.
24605 (mips_option_override): Likewise.
24606 (umips_save_restore_pattern_p): New function.
24607 (umips_output_save_restore): New function.
24608 (umips_load_store_pair_p_1): New function.
24609 (umips_load_store_pair_p): New function.
24610 (umips_output_load_store_pair_1): New function.
24611 (umips_output_load_store_pair): New function.
24612 (umips_movep_target_p) New function.
24613 (mips_prepare_pch_save): Add microMIPS support.
24614 * config/mips/mips.h (TARGET_COMPRESSION): New.
24615 (TARGET_CPU_CPP_BUILTINS): Update macro
24616 to use new compression flags and to support microMIPS.
24617 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24618 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24619 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24620 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24621 (ASM_SPEC): Support mmicromips and mno-micromips.
24622 (M16STORE_REG_P): New macro.
24623 (MIPS_CALL): Support TARGET_MICROMIPS.
24624 (MICROMIPS_J): New macro.
24625 (mips_base_mips16): Rename this...
24626 (mips_base_compression_flags): ...to this.
24627 (UMIPS_12BIT_OFFSET_P): New macro.
24628 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24629 (MULTILIB_DIRNAMES): Likewise.
24631 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24634 * target-globals.c (save_target_globals): For init_reg_sets and
24635 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24637 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
24639 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24640 * config/aarch64/t-aarch64
24641 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24643 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24645 PR inline-asm/56148
24646 * lra-constraints.c (process_alt_operands): Reload operand
24647 conflicting with earlier clobber only if no more other conflicting
24650 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24653 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24654 if not linking a shared library.
24656 2013-02-22 Seth LaForge <sethml@google.com>
24658 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24660 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24662 * config/arm/arm.md (split for extendsidi): Update condition.
24663 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24664 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24665 (qhs_zextenddi_cstr): Likewise.
24667 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24669 PR middle-end/56420
24670 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24671 avoid signed wrapping.
24672 (expand_mult): Handle properly multiplication by
24673 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24674 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24675 in the compiler if coeff is HOST_WIDE_INT_MIN.
24676 (expand_divmod): Don't make ext_op1 static, change it's type to
24677 uhwi. Avoid undefined behavior in -INTVAL (op1).
24679 PR rtl-optimization/50339
24680 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24682 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24683 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24684 into splitting_ashiftrt field.
24685 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24687 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24690 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24692 PR middle-end/56108
24693 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24694 are sure to go irrevocable.
24696 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24698 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24699 scalars are valid operands.
24701 2013-02-21 Martin Jambor <mjambor@suse.cz>
24703 PR tree-optimization/56310
24704 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24705 only matching indices and non-negative final offsets.
24706 (intersect_aggregates_with_edge): Pass src_idx to
24707 agg_replacements_to_vector. Pass src_idx insstead of index to
24708 intersect_with_agg_replacements.
24710 2013-02-21 Martin Jambor <mjambor@suse.cz>
24712 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24713 instead of hard-wired defaults.
24715 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24717 * doc/invoke.texi (MIPS Options): Update documentation of the
24718 floating-point multiply-accumulate instruction restrictions.
24720 2013-02-21 Kostya Serebryany <kcc@google.com>
24722 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24723 asan_shadow_offset on x86_64 linux.
24725 2013-02-21 Richard Biener <rguenther@suse.de>
24727 PR tree-optimization/56415
24729 2013-02-11 Richard Biener <rguenther@suse.de>
24731 PR tree-optimization/56273
24732 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24735 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24738 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24741 PR inline-asm/56405
24742 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24743 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24745 2013-02-20 Jan Hubicka <jh@suse.cz>
24747 PR tree-optimization/56265
24748 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24749 when target is referenced for first time.
24751 2013-02-20 Richard Biener <rguenther@suse.de>
24753 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24754 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24755 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24756 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24757 not return anything.
24758 (rename_ssa_copies): Do not remove unused locals.
24759 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24760 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24761 * passes.c (execute_function_todo): Do not schedule unused locals
24762 removal if cleanup_tree_cfg did something.
24763 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24764 about the number of removed locals.
24766 2013-02-20 Richard Biener <rguenther@suse.de>
24768 PR tree-optimization/56398
24769 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24771 2013-02-20 Martin Jambor <mjambor@suse.cz>
24773 PR tree-optimization/55334
24774 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24775 restricted pointers to arrays.
24777 2013-02-20 Richard Biener <rguenther@suse.de>
24778 Jakub Jelinek <jakub@redhat.com>
24780 PR tree-optimization/56396
24781 * tree-ssa-ccp.c (n_const_val): New static variable.
24782 (get_value): Return NULL for SSA names we don't have a lattice
24784 (ccp_initialize): Initialize n_const_val.
24785 * tree-ssa-copy.c (n_copy_of): New static variable.
24786 (init_copy_prop): Initialize n_copy_of.
24787 (get_value): Return NULL_TREE for SSA names we don't have a
24790 2013-02-20 Martin Jambor <mjambor@suse.cz>
24792 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24794 2013-02-20 Richard Biener <rguenther@suse.de>
24796 * genpreds.c (write_lookup_constraint): Do not compare first
24797 letter of the constraint again.
24799 2013-02-20 Richard Biener <rguenther@suse.de>
24801 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24803 (get_use_iv_cost): Terminate hashtable walk when coming across
24806 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24808 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24809 reassociation for avx2 targets.
24811 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24813 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24814 Add version check for v8.10.a to enable microblaze_has_clz
24815 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24816 version and TARGET_PATTERN_COMPARE check
24817 * config/microblaze/microblaze.md: New clzsi2 instruction
24819 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24821 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24822 function before branching.
24824 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24826 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24828 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24830 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24832 PR middle-end/55889
24833 * sel-sched.c: Include ira.h.
24834 (implicit_clobber_conflict_p): New function.
24835 (moveup_expr): Use it.
24836 * Makefile.in (sel-sched.o): Depend on ira.h.
24838 2013-02-19 Richard Biener <rguenther@suse.de>
24840 PR tree-optimization/56384
24841 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24842 (vn_hash_type): Split out from ...
24843 (vn_hash_constant_with_type): ... here.
24844 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24845 (vn_phi_eq): Compare types from vn_phi_s structure.
24846 (vn_phi_lookup): Populate vn_phi_s type.
24847 (vn_phi_insert): Likewise.
24849 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24851 PR tree-optimization/56350
24852 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24853 if haven't found reduction or nested cycle operand, rather than
24854 asserting we must find it.
24856 PR tree-optimization/56381
24857 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24860 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24861 Jakub Jelinek <jakub@redhat.com>
24864 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24865 (swap_optab_enable): Same.
24866 (init_all_optabs): Use argument instead of global.
24867 * tree.h (struct tree_optimization_option): New field target_optabs.
24868 * expr.h (init_all_optabs): Add argument to prototype.
24869 (TREE_OPTIMIZATION_OPTABS): New.
24870 (save_optabs_if_changed): Protoize.
24871 * optabs.h: Declare this_fn_optabs.
24872 * optabs.c (save_optabs_if_changed): New.
24873 Declare this_fn_optabs.
24874 (init_optabs): Add argument to init_all_optabs() call.
24875 * function.c (invoke_set_current_function_hook): Handle per
24877 * function.h (struct function): New field optabs.
24878 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24879 optimization_current_node has changed.
24880 * target-globals.h (save_target_globals_default_opts): Protoize.
24881 * target-globals.c (save_target_globals_default_opts): New.
24883 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24886 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24887 registers %fr12 and %fr12R as call used.
24890 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24891 and HImode, require all displacements to be an integer multiple of
24893 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
24894 only allow QImode and HImode when reload is in progress and strict is
24895 true. Likewise for symbolic addresses. Use base14_operand to check
24896 displacements in REG+BASE addresses.
24898 2013-02-18 Richard Biener <rguenther@suse.de>
24900 PR tree-optimization/56366
24901 * tree-vect-loop.c (get_initial_def_for_induction): Properly
24902 handle sign-conversion of outer-loop initial induction value.
24904 2013-02-18 Richard Biener <rguenther@suse.de>
24906 PR middle-end/56349
24907 * cfghooks.c (merge_blocks): If we merge a latch into another
24908 block adjust references to it.
24909 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
24910 (verify_loop_structure): Verify that a recorded latch is in fact
24913 2013-02-18 Richard Biener <rguenther@suse.de>
24915 PR tree-optimization/56321
24916 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
24917 order SSA name release and virtual operand unlinking.
24919 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24921 * config/microblaze/microblaze.md (save_stack_block): Define.
24922 (restore_stack_block): Likewise.
24924 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24926 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
24927 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
24928 * config/microblaze/microblaze.c (microblaze_option_override):
24929 Bail out early for PIC modes when target does not support PIC.
24931 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24933 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
24934 Replace with a microblaze version.
24935 (microblaze_trampoline_init): Adapt for microblaze.
24936 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
24939 2013-02-16 Jakub Jelinek <jakub@redhat.com>
24940 Dodji Seketeli <dodji@redhat.com>
24943 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
24944 (instrument_mem_region_access): Do not forget to always put
24945 instrumentation of the of 'base' and 'base + len' in a "if (len !=
24946 0) statement, even for cases where either 'base' or 'base + len'
24947 are not instrumented -- because they have been previously
24948 instrumented. Simplify the logic by putting all the statements
24949 instrument 'base + len' inside a sequence, and then insert that
24950 sequence right before the current insertion point. Then, to
24951 instrument 'base + len', just get an iterator on that statement.
24952 And do not forget to update the pointer to iterator the function
24953 received as argument.
24955 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
24957 PR rtl-optimization/56348
24958 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
24960 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
24962 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
24963 (clean_graph_dump_file): Pass base to start_graph_dump.
24965 2013-02-14 Richard Henderson <rth@redhat.com>
24968 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
24970 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
24972 * collect2-aix.h: Define F_LOADONLY.
24974 2013-02-14 Richard Biener <rguenther@suse.de>
24977 * varasm.c (output_constant_def_1): Get the decl representing
24978 the constant as argument.
24979 (output_constant_def): Wrap output_constant_def_1.
24980 (make_decl_rtl): Use output_constant_def_1 with the decl
24981 representing the constant.
24982 (build_constant_desc): Optionally re-use a decl already
24983 representing the constant.
24984 (tree_output_constant_def): Adjust.
24986 2013-02-14 Dodji Seketeli <dodji@redhat.com>
24989 * asan.c (instrument_builtin_call): Really put the length of the
24990 second source argument into src1_len.
24992 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24994 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24995 argument. If it is false, don't create edge from then_bb to
24997 (insert_if_then_before_iter): Pass true to it.
24998 (build_check_stmt): Pass false to it.
24999 (transform_statements): Flush hash table only on extended basic
25000 block boundaries, rather than at the beginning of every bb.
25001 Don't flush hash table on nonfreeing_call_p calls.
25002 * tree-flow.h (nonfreeing_call_p): New prototype.
25003 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25005 2013-02-13 David S. Miller <davem@davemloft.net>
25007 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25009 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25012 * ira.c (max_regno_before_ira): Move from ...
25014 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25017 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25019 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25021 2013-02-13 Richard Biener <rguenther@suse.de>
25024 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25025 globals in MEM_REFs.
25027 2013-02-13 Richard Biener <rguenther@suse.de>
25029 * loop-init.c (loop_optimizer_init): Clear loop state when
25030 re-initializing preserved loops.
25031 * loop-unswitch.c (unswitch_single_loop): Return whether
25032 we unswitched the loop. Do not verify loop state here.
25033 (unswitch_loops): When we unswitched a loop discover new loops.
25035 2013-02-13 Kostya Serebryany <kcc@google.com>
25037 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25039 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25041 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25043 Avoid instrumenting duplicated memory access in the same basic block
25044 * Makefile.in (asan.o): Add new dependency on hash-table.h
25045 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25046 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25047 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25048 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25049 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25050 (get_mem_ref_of_assignment): New functions.
25051 (get_mem_refs_of_builtin_call): Extract from
25052 instrument_builtin_call and tweak a little bit to make it fit with
25054 (instrument_builtin_call): Use the new
25055 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25056 of is_gimple_builtin_call.
25057 (instrument_derefs, instrument_mem_region_access): Insert the
25058 instrumented memory reference into the hash table.
25059 (maybe_instrument_assignment): Renamed instrument_assignment into
25060 this, and change it to advance the iterator when instrumentation
25061 actually happened and return true in that case. This makes it
25062 homogeneous with maybe_instrument_assignment, and thus give a
25063 chance to callers to be more 'regular'.
25064 (transform_statements): Clear the memory reference hash table
25065 whenever we enter a new BB, when we cross a function call, or when
25066 we are done transforming statements. Use
25067 maybe_instrument_assignment instead of instrumentation. No more
25068 need to special case maybe_instrument_assignment and advance the
25069 iterator after calling it; it's now handled just like
25070 maybe_instrument_call. Update comment.
25072 2013-02-13 Richard Biener <rguenther@suse.de>
25074 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25075 Fix loop discovery code.
25077 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25079 PR inline-asm/56148
25080 * lra-constraints.c (process_alt_operands): Match early clobber
25081 operand with itself. Check conflicts with earlyclobber only if
25082 the operand is not reloaded. Prefer to reload conflicting operand
25083 if earlyclobber and matching operands are the same.
25085 2013-02-12 Richard Biener <rguenther@suse.de>
25088 * lto-streamer-out.c (write_symbol): Do not output symbols
25089 for hard register variables.
25091 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25094 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25095 (umulsidi3_insn, mulsidi3_insn): New insns.
25097 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25099 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25100 (struct tune_params): Add vec_costs field.
25101 * config/arm/arm.c (arm_builtin_vectorization_cost)
25102 (arm_add_stmt_cost): New functions.
25103 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25104 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25105 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25106 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25107 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25108 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25109 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25111 2013-02-12 Richard Biener <rguenther@suse.de>
25114 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25115 decls again if possible.
25117 2013-02-12 Richard Biener <rguenther@suse.de>
25119 PR middle-end/56288
25120 * tree-ssa.c (verify_ssa_name): Fix check, move
25121 SSA_NAME_IN_FREE_LIST check up.
25123 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25124 Steven Bosscher <steven@gcc.gnu.org>
25126 PR rtl-optimization/56151
25127 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25128 equal to op0 or op1, and last_insn pattern is CODE operation
25129 with MEM dest and one of the operands matches that MEM.
25131 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
25133 * doc/extend.texi: Document Function Multiversioning and "default"
25134 parameter string to target attribute.
25135 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25136 target attribute parameter is "default".
25137 (ix86_compare_version_priority): Remove checks for target attribute.
25138 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25139 Remove check for target attribute equal to NULL. Add assert.
25140 (ix86_generate_version_dispatcher_body): Change error to sorry.
25142 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25143 Jack Howarth <howarth@bromo.med.uc.edu>
25144 Patrick Marlier <patrick.marlier@gmail.com>
25147 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25148 define ENDFILE_SPEC as TM_DESTRUCTOR.
25149 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25151 2013-02-11 Alexander Potapenko <glider@google.com>
25152 Jack Howarth <howarth@bromo.med.uc.edu>
25153 Jakub Jelinek <jakub@redhat.com>
25156 * config/darwin.c (cdtor_record): Rename ctor_record.
25157 (sort_cdtor_records): Rename sort_ctor_records.
25158 (finalize_dtors): New routine to sort destructors by
25159 priority before use in assemble_integer.
25160 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25162 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25164 PR rtl-optimization/56275
25165 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25166 offset is non-negative and less than cmode size before
25167 calling simplify_subreg.
25169 2013-02-11 Richard Biener <rguenther@suse.de>
25171 PR tree-optimization/56264
25172 * cfgloop.h (fix_loop_structure): Adjust prototype.
25173 * loop-init.c (fix_loop_structure): Return the number of
25174 newly discovered loops.
25175 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25176 are discovered, do a full loop-closed SSA rewrite.
25178 2013-02-11 Richard Biener <rguenther@suse.de>
25180 PR tree-optimization/56273
25181 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25183 (check_array_ref): Fix missing newline in dumps.
25184 (search_for_addr_array): Likewise.
25186 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25188 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25190 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25193 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25195 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25197 PR rtl-optimization/56246
25198 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25200 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25201 constraints are satisfied.
25203 2013-02-08 Jeff Law <law@redhat.com>
25206 * emit-rtl.c (reg_is_parm_p): New function.
25207 * regs.h (reg_is_parm_p): New prototype.
25208 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25209 callee-clobbered registers.
25211 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25214 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25215 If there is no implicit builtin declaration, just return NULL.
25217 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25219 * config/i386/sse.md (FMAMODEM): New mode iterator.
25220 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25221 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25223 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25225 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25226 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25227 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25229 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25231 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25232 (microblaze*-*-elf): Likewise.
25233 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25235 * config/microblaze/microblaze-c.c: Add builtin defines for
25236 _LITTLE_ENDIAN and _BIG_ENDIAN.
25237 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25238 add to TARGET_DEFAULT flags.
25239 Expand ASM_SPEC and LINK_SPEC.
25240 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25241 * config/microblaze/microblaze.md: Update extendsidi2 and
25242 movdi_internal instructions to use low-order / high-order reg
25244 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25245 options and inversemask / mask of LITTLE_ENDIAN.
25246 * config/microblaze/t-microblaze: Expand multilib options to
25247 include mlittle-endian (le) and update exceptions patterns.
25249 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25251 PR rtl-optimization/56195
25252 * lra-constraints.c (get_reload_reg): Don't reuse regs
25253 if they have smaller mode than requested, if they have
25254 wider mode than requested, try to return a SUBREG.
25256 PR tree-optimization/56250
25257 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25258 if type is unsigned and code isn't MULT_EXPR.
25260 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25262 PR tree-optimization/56064
25263 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25264 bits according to mode.
25265 * fixed-value.h (fixed_from_double_int)
25266 (const_fixed_from_double_int): Adjust comments.
25268 2013-02-08 Richard Biener <rguenther@suse.de>
25271 * lto-streamer.h (struct data_in): Remove current_file, current_line
25272 and current_col members.
25273 * lto-streamer-out.c (lto_output_location): Stream changed bits
25274 en-block for efficiency.
25275 * lto-streamer-in.c (clear_line_info): Remove.
25276 (lto_input_location): Cache current file, line and column
25277 globally via local statics. Read changed bits en-block.
25278 (input_function): Do not call clear_line_info.
25279 (lto_read_body): Likewise.
25280 (lto_input_toplevel_asms): Likewise.
25282 2013-02-08 Michael Matz <matz@suse.de>
25284 PR tree-optimization/52448
25285 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25286 (nt_call_phase): New static.
25287 (add_or_mark_expr): Only mark accesses with newer phase than any
25289 (nonfreeing_call_p): New.
25290 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25291 (nt_fini_block): Keep blocks marked as visited.
25292 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25294 2013-02-08 Richard Biener <rguenther@suse.de>
25296 * ira.c (ira): Free broken dominator information.
25298 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25300 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25302 2013-02-08 Marek Polacek <polacek@redhat.com>
25304 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25306 2013-02-08 Richard Biener <rguenther@suse.de>
25308 PR middle-end/56181
25309 * cfgloop.h (flow_loops_find): Adjust.
25310 (bb_loop_header_p): Declare.
25311 * cfgloop.c (bb_loop_header_p): New function split out from ...
25312 (flow_loops_find): ... here. Adjust function signature,
25313 support incremental loop structure update.
25314 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25315 * cfgloopmanip.c (fix_loop_structure): Move ...
25316 * loop-init.c (fix_loop_structure): ... here.
25317 (apply_loop_flags): Split out from ...
25318 (loop_optimizer_init): ... here.
25319 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25320 in incremental mode, only remove dead loops here.
25322 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25325 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25326 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25327 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25328 (*round<mode>3.libgcc): New insns for fixed-modes.
25329 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25330 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25331 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25332 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25333 implementations. Define to __builtin_avr_absFX,
25334 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25335 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25336 __builtin_avr_countlsFX, respectively.
25337 * config/avr/avr-c.c (target.h): Include it.
25338 (enum avr_builtin_id): New enum.
25339 (avr_resolve_overloaded_builtin): New static function.
25340 (avr_register_target_pragmas): Use it to set
25341 targetm.resolve_overloaded_builtin.
25342 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25343 tree nodes used by DEF_BUILTIN.
25344 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25345 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25346 <AVR_BUILTIN_xxBITS>: Same.
25348 2013-02-08 Richard Biener <rguenther@suse.de>
25350 * cfgloop.c (verify_loop_structure): Properly handle
25351 a loop exiting to another loop header.
25352 * ira-int.h (ira_loops): Remove.
25353 * ira.c (ira_loops): Remove.
25354 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25355 (do_reload): Use loop_optimizer_finalize.
25356 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25357 number_of_loops to access the loop tree.
25358 (more_one_region_p): Likewise.
25359 (finish_loop_tree_nodes): Likewise.
25360 (rebuild_regno_allocno_maps): Likewise.
25361 (mark_loops_for_removal): Likewise.
25362 (mark_all_loops_for_removal): Likewise.
25363 (remove_unnecessary_regions): Likewise.
25364 (ira_build): Likewise.
25365 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25367 2013-02-08 Richard Biener <rguenther@suse.de>
25369 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25370 * ipa-pure-const.c (analyze_function): Avoid calling
25371 mark_irreducible_loops twice.
25372 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25374 2013-02-07 David S. Miller <davem@davemloft.net>
25376 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25378 * var-tracking.c (vt_add_function_parameter): Test the presence of
25379 HAVE_window_save properly and do not remap argument registers when
25380 we have a leaf function.
25382 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25385 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25387 * config/i386/i386.c (ix86_print_operand): Ditto.
25389 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25391 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25393 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25395 PR rtl-optimization/56225
25396 * lra-constraints.c (process_alt_operands): Check that reload hard
25397 reg can hold value for strict_low_part.
25399 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25402 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25403 dwarf2out_end_function.
25404 (in_first_function_p, maybe_at_text_label_p,
25405 first_loclabel_num_not_at_text_label): New variables.
25406 (dwarf2out_var_location): In the first function find out
25407 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25408 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25411 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25413 PR rtl-optimization/56178
25414 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25415 SUBREG of a register. Tidy up related block of code.
25416 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25417 note if the source is a register or a SUBREG of a register.
25419 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25422 * config/rs6000/rs6000.md (ptrm): New mode attr.
25423 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25424 call_value_indirect_aix<pttrsize>,
25425 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25428 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25430 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25431 if -bnortl. Convert to strcmp and strncmp.
25433 2013-02-07 Alan Modra <amodra@gmail.com>
25436 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25437 addresses won't wrap when offsetting.
25438 (rs6000_secondary_reload): Provide secondary reloads needed for
25439 wrapping LO_SUM addresses.
25441 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25443 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25444 MACH, just __MACH__.
25446 2013-02-06 Richard Biener <rguenther@suse.de>
25448 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25449 instead of calling fix_loop_structure.
25451 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25453 PR middle-end/56217
25454 * omp-low.c (use_pointer_for_field): Return false if
25455 lower_send_shared_vars doesn't generate any copy-out code.
25457 2013-02-06 Tom de Vries <tom@codesourcery.com>
25459 PR rtl-optimization/56131
25460 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25461 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25462 of the label is NULL. Add comment.
25464 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25466 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25469 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25470 (STATIC_LIBTSAN_LIBS): Likewise.
25471 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25472 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25473 is defined, don't add anything else beyond that.
25474 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25475 (LINK_COMMAND_SPEC): Use them.
25477 PR tree-optimization/56205
25478 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25479 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25480 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25482 2013-02-05 Richard Biener <rguenther@suse.de>
25484 PR tree-optimization/53342
25485 PR tree-optimization/53185
25486 * tree-vectorizer.h (vect_check_strided_load): Remove.
25487 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25488 not disallow peeling for vectorized strided loads.
25489 (vect_check_strided_load): Make static and simplify.
25490 (vect_analyze_data_refs): Adjust.
25491 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25492 correctly when vectorizing strided loads.
25494 2013-02-05 Richard Biener <rguenther@suse.de>
25496 * doc/install.texi: Refer to ISL, not PPL.
25498 2013-02-05 Jan Hubicka <jh@suse.cz>
25500 PR tree-optimization/55789
25501 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25503 2013-02-05 Jan Hubicka <jh@suse.cz>
25505 PR tree-optimization/55789
25506 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25507 the dead call anyway.
25509 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25512 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25514 2013-02-04 Alexander Potapenko <glider@google.com>
25515 Jack Howarth <howarth@bromo.med.uc.edu>
25516 Jakub Jelinek <jakub@redhat.com>
25519 * config/darwin.c (sort_ctor_records): Stabilized qsort
25520 on constructor priority by using original position.
25521 (finalize_ctors): New routine to sort constructors by
25522 priority before use in assemble_integer.
25523 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25525 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25528 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25529 about visibility on artificial decls.
25530 * config/sol2.c (solaris_assemble_visibility): Likewise.
25532 2013-02-04 Kai Tietz <ktietz@redhat.com>
25535 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25536 argument and improve checking of return-argument types for 16-byte
25538 (ix86_function_value_1): Add additional valtype argument on call
25539 of function_value_64.
25540 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25541 handling infunction_value_64 function.
25543 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25545 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25547 2013-02-04 Richard Biener <rguenther@suse.de>
25549 PR tree-optimization/56188
25550 * tree-ssa-structalias.c (label_visit): Consider case with
25551 initially non-empty points-to set.
25552 (perform_var_substitution): Dump node mapping and clean up.
25554 2013-02-04 Richard Guenther <rguenther@suse.de>
25557 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25558 node prevail as last resort.
25559 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25560 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25562 2013-02-04 Richard Biener <rguenther@suse.de>
25564 PR tree-optimization/56113
25565 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25567 (equiv_class_lookup_or_add): ... this.
25568 (label_visit): Adjust and fix error in previous patch.
25569 (perform_var_substitution): Adjust.
25571 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25573 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25574 * config/sh/sh4-300.md: Likewise.
25575 * config/sh/sh4a.md: Likewise.
25576 * config/sh/constraints.md: Likewise.
25577 * config/sh/sh.md: Likewise.
25578 * config/sh/netbsd-elf.h: Likewise.
25579 * config/sh/predicates.md: Likewise.
25580 * config/sh/sh-protos.h: Likewise.
25581 * config/sh/ushmedia.h: Likewise.
25582 * config/sh/linux.h: Likewise.
25583 * config/sh/sh.c: Likewise.
25584 * config/sh/superh.h: Likewise.
25585 * config/sh/elf.h: Likewise.
25586 * config/sh/sh4.md: Likewise.
25587 * config/sh/sh.h: Likewise.
25589 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25591 * config/pa/constraints.md: Adjust unused letters. Change "T"
25592 constraint to match_test floating_point_store_memory_operand().
25593 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25594 (base14_operand): New.
25595 (floating_point_store_memory_operand): New.
25596 (integer_store_memory_operand): Revise to use base14_operand and
25597 reg_plus_base_memory_operand.
25598 (move_dest_operand): Allow symbolic_memory_operands.
25599 (symbolic_memory_operand): Check for LO_SOM.
25600 (symbolic_operand): Change default case to break.
25601 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25602 CONST_DOUBLE values to be reloaded by putting them into memory when
25603 the destination is a floating point register.
25604 (movdf): Remove code to handle CONST_DOUBLE.
25606 (reload_indf_r1): New.
25607 (reload_insf_r1): New.
25608 Consistently use "Q" and "T" constraints with integer and floating
25609 point move instructions, respectively.
25610 (movdi): Remove FAIL.
25611 Change predicate for source operand unamed DImode move from
25612 general_operand to move_src_operand.
25613 (umulsidi3): Change predicate for destination operand to
25615 Likewise for similar unamed patterns.
25616 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25617 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25618 (hppa_legitimize_address): Simplify mask calculation.
25619 (pa_emit_move_sequence): Revised handling of secondary reloads from
25620 REG+D addresses for floating point loads and stores. Directly handle
25621 loading CONST0_RTX (mode) to a floating point register.
25622 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25623 to floating point registers. Don't restrict secondary reloads to
25624 floating point registers to integer modes. Revise some comments and
25626 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25627 (pa_legitimate_address_p): New.
25628 (pa_legitimize_reload_address): New.
25629 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25630 (STRICT_REG_OK_FOR_BASE_P): New.
25631 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25632 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25634 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25635 Andrew Dixie <andrewd@gentrack.com>
25637 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25640 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25642 * expmed.c (extract_bit_field_1): Pass the full width of the
25643 structure to get_best_reg_extraction_insn.
25645 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25648 * configure.ac (use_cxa_atexit): Add AIX.
25649 * configure: Regenerate.
25651 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25653 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25656 * final.c (need_profile_function): New variable.
25657 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25658 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25659 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25660 notes, targetm.asm_out.function_prologue doesn't emit anything,
25661 HAVE_prologue and profiler should be emitted before prologue,
25662 set need_profile_function instead of emitting it.
25663 (final_scan_insn): If need_profile_function, emit
25664 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25665 NOTE_INSN_FUNCTION_BEG note.
25667 2013-02-01 Richard Henderson <rth@redhat.com>
25669 * config/rs6000/rs6000.md (smulditi3): New.
25672 * config/alpha/alpha.md (umulditi3): New.
25674 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25676 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25677 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25679 2013-02-01 Richard Biener <rguenther@suse.de>
25681 PR tree-optimization/56113
25682 * tree-ssa-structalias.c (label_visit): Reduce work for
25683 single-predecessor nodes.
25685 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25687 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25688 range isn't testing for zero.
25690 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25692 PR middle-end/56113
25693 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25695 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25696 Nick Clifton <nickc@redhat.com>
25698 * config/v850/constraints.md (Q): Define as a memory constraint.
25699 * config/v850/predicates.md (label_ref_operand): New predicate.
25700 (e3v5_shift_operand): New predicate.
25701 (ior_operator): New predicate.
25702 * config/v850/t-v850: Add e3v5 multilib.
25703 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25704 (v850_gen_movdi): Prototype.
25705 * config/v850/v850.c: Add support for e3v5 architecture.
25706 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25708 (construct_save_jarl): Add e3v5 long JARL support.
25709 (v850_adjust_insn_length): New function. Adjust length of call
25710 insns when using e3v5 instructions.
25711 (v850_gen_movdi): New function: Generate instructions to move a
25713 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25714 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25715 (TARGET_USE_FPU): Enable for e3v5.
25716 (CONST_OK_FOR_W): New macro.
25717 (ADJUST_INSN_LENGTH): Define.
25718 * config/v850/v850.md (UNSPEC_LOOP): Define.
25719 (attr cpu): Add v850e3v5.
25720 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25721 (movdi): New pattern.
25722 (movdi_internal): New pattern.
25723 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25724 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25725 (cstoresf4): Likewise.
25726 (cstoredf4): Likewise.
25727 (insv): New pattern.
25728 (rotlso3_a): New pattern.
25729 (rotlsi3_b): New pattern
25730 (rotlsi3_v850e3v5): New pattern.
25731 (doloop_begin): New pattern.
25732 (fix_loop_counter): New pattern.
25733 (doloop_end): New pattern.
25734 (branch_normal): Add e3v5 long branch support.
25735 (branch_invert): Likewise.
25736 (branch_z_normal): Likewise.
25737 (branch_z_invert): Likewise.
25738 (branch_nz_normal): Likewise.
25739 (branch_nz_invert): Likewise.
25740 (call_internal_short): Add e3v5 register-indirect JARL support.
25741 (call_internal_long): Likewise.
25742 (call_value_internal_short): Likewise.
25743 (call_value_internal_long): Likewise.
25744 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25745 (mloop): New option.
25746 * config.gcc: Add support for configuring v840e3v5 target.
25747 * doc/invoke.texi: Document new v850 specific command line options.
25749 2013-01-31 Paul Koning <ni1d@arrl.net>
25753 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25754 children if parent is a class.
25755 (prune_unused_types_prune): Don't add DW_AT_declaration.
25757 2013-01-31 Richard Biener <rguenther@suse.de>
25759 PR tree-optimization/56157
25760 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25761 match up operand with SLP child.
25763 2013-01-31 Jason Merrill <jason@redhat.com>
25766 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25767 parameters the first time.
25768 (gen_scheduled_generic_parms_dies): Check completeness here.
25770 2013-01-31 Richard Biener <rguenther@suse.de>
25772 PR middle-end/53073
25773 * common.opt (faggressive-loop-optimizations): New flag,
25774 enabled by default.
25775 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25776 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25777 infer_loop_bounds_from_undefined by it.
25779 2013-01-31 Richard Biener <rguenther@suse.de>
25781 PR tree-optimization/56150
25782 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25783 visit virtual operands.
25784 (find_uses_to_rename_bb): Likewise.
25786 2013-01-31 Richard Biener <rguenther@suse.de>
25788 PR tree-optimization/56150
25789 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25790 mixed store non-store stmts.
25792 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25795 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25796 LIBASAN_EARLY_SPEC is defined.
25797 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25798 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25800 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25803 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25804 invalid args instead of ICEing on it.
25805 (ix86_valid_target_attribute_tree): Return error_mark_node if
25806 ix86_valid_target_attribute_inner_p failed.
25807 (ix86_valid_target_attribute_p): Return false only if
25808 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25809 target("default") attribute.
25810 (sorted_attr_string): Change argument from const char * to tree,
25811 merge in all target attribute arguments rather than just one.
25812 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25813 instead of free. Avoid using strcat.
25814 (ix86_mangle_function_version_assembler_name): Mangle
25815 target("default") as if no target attribute is present. Adjust
25816 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25817 instead of xmalloc and XDELETEVEC instead of free.
25818 (ix86_function_versions): Don't return true if one of the decls
25819 doesn't have target attribute. If they don't and one of the decls
25820 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25821 sorted_attr_string caller. Use XDELETEVEC instead of free.
25822 (ix86_supports_function_versions): Remove.
25823 (make_name): Fix up formatting.
25824 (make_dispatcher_decl): Remove resolver_name and its initialization.
25825 Avoid leaking memory.
25826 (is_function_default_version): Return true if there is
25827 target("default") attribute rather than no target attribute at all.
25828 (make_resolver_func): Avoid leaking memory.
25829 (ix86_generate_version_dispatcher_body): Likewise.
25830 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25831 * target.def (supports_function_versions): Remove.
25832 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25833 * doc/tm.texi: Regenerated.
25835 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25837 PR rtl-optimization/56144
25838 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25839 for values with side effects.
25841 2013-01-30 Richard Biener <rguenther@suse.de>
25843 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25844 (sparseset_pop): Likewise.
25845 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25846 to be able to use quick_push in the worker loop.
25848 2013-01-30 Marek Polacek <polacek@redhat.com>
25850 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25852 2013-01-30 Richard Biener <rguenther@suse.de>
25855 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25857 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25859 PR tree-optimization/56064
25860 * fixed-value.c (fixed_from_double_int): New function.
25861 * fixed-value.h (fixed_from_double_int): New prototype.
25862 (const_fixed_from_double_int): New static inline function.
25863 * fold-const.c (native_interpret_fixed): New static function.
25864 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25865 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25866 (native_encode_fixed): New static function.
25867 (native_encode_expr) <FIXED_CST>: Use it.
25868 (native_interpret_int): Move double_int worker code to...
25869 * double-int.c (double_int::from_buffer): ...this new static method.
25870 * double-int.h (double_int::from_buffer): Prototype it.
25872 2013-01-30 Richard Biener <rguenther@suse.de>
25874 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25875 New pointer-map and obstack.
25876 (init_alias_vars): Allocate pointer-map and obstack.
25877 (delete_points_to_sets): Free them.
25878 (find_what_var_points_to): Cache result.
25879 (find_what_p_points_to): Adjust for changed interface of
25880 find_what_var_points_to.
25881 (compute_points_to_sets): Likewise.
25882 (ipa_pta_execute): Likewise.
25884 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25886 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25887 * configure: Regenerate.
25888 * config.in: Regenerate.
25889 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25890 #nobits/#progbits if supported.
25892 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
25895 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
25896 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
25897 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
25899 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25901 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
25902 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
25904 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25906 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
25907 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
25909 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25911 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
25913 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
25914 * config/arm/cortex-a7.md: New bypasses using
25915 arm_mac_accumulator_is_result.
25917 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25919 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
25920 (cortex_a7_neon_mla): Likewise.
25921 (cortex_a7_fpfmad): New reservation.
25922 (cortex_a7_fpmacs): Use ffmas and update required units.
25923 (cortex_a7_fpmuld): Update required units and latency.
25924 (cortex_a7_fpmacd): Likewise.
25925 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
25926 (cortex_a7_neon). Likewise.
25927 (bypass) Update participating units.
25929 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
25931 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
25932 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
25934 * config/arm/vfp11.md (vfp_farith): Use ffmas.
25935 (vfp_fmul): Use ffmad.
25936 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
25937 (cortex_r4_fmacd): Use ffmad.
25938 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
25939 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
25940 (cortex_a9_fmacd): Use ffmad.
25941 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
25942 (cortex_a8_vfp_macd): Use ffmad.
25943 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
25944 (cortex_a5_fpmacd): Use ffmad.
25945 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
25946 (cortex_a15_vfp_macd): Use ffmad.
25947 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
25949 2013-01-29 Jason Merrill <jason@redhat.com>
25952 * varasm.c (default_assemble_visibility): Don't warn about
25953 visibility on artificial decls.
25955 2013-01-29 Richard Biener <rguenther@suse.de>
25957 PR tree-optimization/56113
25958 * tree-ssa-structalias.c (equiv_class_lookup): Also return
25960 (label_visit): Free duplicate bitmaps and record the leader instead.
25961 (perform_var_substitution): Adjust.
25963 2013-01-29 Richard Biener <rguenther@suse.de>
25965 PR tree-optimization/55270
25966 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
25967 the CFG, schedule loops for fixup.
25969 2013-01-29 Nick Clifton <nickc@redhat.com>
25971 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
25974 2013-01-28 Leif Ekblad <leif@rdos.net>
25976 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25977 * config/i386/i386.h (TARGET_RDOS): New macro.
25978 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25979 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25980 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25981 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25982 DEFAULT_LARGE_SECTION_THRESHOLD.
25983 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25984 * config/i386/rdos.h: New file.
25985 * config/i386/rdos64.h: New file.
25987 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25990 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25993 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25995 PR rtl-optimization/56117
25996 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25997 call cselib_lookup_from_insn on the MEM before calling
25998 add_insn_mem_dependence.
26000 2013-01-28 Richard Biener <rguenther@suse.de>
26002 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26003 to a stmt that didn't have one.
26004 (copy_phis_for_bb): Likewise for PHI arguments.
26005 (copy_debug_stmt): Likewise for debug stmts.
26007 2013-01-28 Richard Biener <rguenther@suse.de>
26009 PR tree-optimization/56034
26010 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26011 (partition_builtin_p): Adjust.
26012 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26013 it is the last partition.
26014 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26015 up the vertex for the definition.
26016 (classify_partition): Classify whether a partition is a
26017 PKIND_REDUCTION, thus has uses outside of the loop.
26018 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26019 Merge all PKIND_REDUCTION partitions into the last partition.
26020 (tree_loop_distribution): Seed partitions from reductions as well.
26022 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26024 PR tree-optimization/56125
26025 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26026 pow(x,c) into sqrt(x) * powi(x, n/2) or
26027 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26028 optimizing for size.
26029 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26030 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26033 PR tree-optimization/56094
26034 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26035 to UNKNOWN_LOCATION while gimplifying expr.
26037 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26040 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26041 operand 0 in movabs insn template for -masm=intel asm alternative.
26042 (*movabs<mode>_2): Ditto for operand 1.
26044 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26047 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26050 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26052 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26055 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26057 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26058 Undef to avoid warning.
26060 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26062 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26063 * configure: Regenerate.
26065 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26067 PR tree-optimization/56098
26068 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26069 for stmts with volatile ops.
26070 (cond_store_replacement): Don't optimize if assign has volatile ops.
26071 (cond_if_else_store_replacement_1): Don't optimize if either
26072 then_assign or else_assign have volatile ops.
26073 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26076 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26078 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26080 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26082 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26083 missing ':' in asm example.
26085 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26087 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26088 entries into lane and laneq entries.
26089 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26090 Remove AdvSIMD scalar modes.
26091 (aarch64_sq<r>dmulh_laneq<mode>): New.
26092 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26094 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26095 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26096 * config/aarch64/iterators.md (VCOND): New.
26099 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26102 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26103 Add NULL LIBNAME argument to existing definitions.
26104 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26105 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26106 * config/avr/avr.c (DEF_BUILTIN): Same.
26107 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26108 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26109 implementation is available (DECL_ASSEMBLER_NAME is set).
26110 (avr_fold_absfx): New static function.
26111 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26112 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26113 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26114 AVR_BUILTIN_ABSLLK.
26115 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26116 (abshk, absk, abslk, absllk): Provide as static inline functions.
26118 2013-01-25 Marek Polacek <polacek@redhat.com>
26120 PR tree-optimization/56035
26121 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26123 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26125 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26126 (*movtf_internal_rex64): Add (!o,C) alternative
26127 (*movxf_internal_rex64): Ditto.
26128 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26130 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26132 * doc/invoke.texi: fix typo.
26133 * doc/objc.texi: fix typo.
26135 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26137 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26138 for the first two alternatives.
26140 2013-01-24 Diego Novillo <dnovillo@google.com>
26142 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26143 (ggc-zone.o): Remove.
26144 * configure.ac: Remove option --with-gc.
26145 * configure: Re-generate.
26146 * doc/install.texi: Remove documentation for --with-gc.
26147 * gengtype.c (write_enum_defn): Remove. Update all users.
26148 (write_Types_process_field): Remove generation of gt_e_* argument.
26149 (output_type_enum): Remove. Update all users.
26150 (write_enum_defn): Remove. Update all users.
26151 (enum alloc_zone): Remove. Update all users.
26152 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26153 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26154 Update all callers.
26155 (struct ptr_data): Remove field TYPE. Update all users.
26156 (gt_pch_note_object): Remove argument TYPE. Update all users.
26157 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26159 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26160 (struct alloc_zone): Remove.
26161 (ggc_internal_alloc_zone_stat): Remove.
26162 (ggc_internal_cleared_alloc_zone_stat): Remove.
26163 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26164 (ggc_pch_count_object): Remove last argument. Update all users.
26165 (ggc_pch_alloc_object): Remove last argument. Update all users.
26166 (struct alloc_zone): Remove.
26167 * ggc-zone.c: Remove.
26168 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26169 (struct alloc_zone): Remove.
26170 (ggc_alloc_typed_stat): Remove.
26171 (ggc_alloc_typed): Remove.
26172 (ggc_splay_alloc): Remove first argument.
26173 (rtl_zone): Remove. Update all users.
26174 (tree_zone): Remove. Update all users.
26175 (tree_id_zone): Remove. Update all users.
26176 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26177 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26178 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26179 * tree-ssanames.c: Remove references to zone allocator in comments.
26181 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26183 * config/avr/avr.c (avr_out_fract): Make register numbers that
26184 might be outside of source operand signed.
26186 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26188 * config/i386/constraints.md (Yf): New constraint.
26189 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26190 of f constraint to conditionaly disable x87 register preferences.
26191 (*movdf_internal): Ditto.
26192 (*movsf_internal): Ditto.
26194 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26196 PR inline-asm/55934
26197 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26198 that have operands with impossible constraints.
26199 Add a FIXME for a speed-up opportunity.
26200 * lra-constraints.c (process_alt_operands): Verify that a class
26201 selected from constraints on asms is valid for the operand mode.
26202 (curr_insn_transform): Remove incorrect comment.
26204 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26206 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26207 TOC operand is a valid symbol ref in the constant pool.
26209 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26211 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26213 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26216 * config/avr/stdfix.h: New file.
26217 * t-avr (stdfix-gcc.h): New rule to build it.
26218 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26220 2013-01-23 Kostya Serebryany <kcc@google.com>
26222 * config/darwin.h: remove dependency on
26223 CoreFoundation (asan on Mac OS).
26225 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26228 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26229 instead of cmpdi_operand for first comparison operand.
26230 Don't assert that comparison operands aren't both constants.
26232 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26234 * doc/install.texi (Downloading the Source): Update references to
26235 downloading separate components.
26237 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26239 * doc/extend.texi (__int128): Improve grammar.
26241 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26244 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26245 alternative to (o,r).
26246 (*movdi_internal_rex64): Remove (!o,n) alternative.
26247 (DImode immediate->memory splitter): Remove.
26248 (DImode immediate->memory peephole2): Remove.
26249 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26250 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26251 alternative to (!o,*r).
26252 (*movtf_internal_sse): New pattern.
26253 (*movxf_internal_rex64): New pattern.
26254 (*movxf_internal): Disable for TARGET_64BIT.
26255 (*movdf_internal_rex64): Remove (!o,F) alternative.
26257 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26259 PR middle-end/56074
26260 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26261 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26262 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26263 stmt locations where LOCATION_LOCUS of the stmt location is
26264 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26267 * config/i386/i386.md (UNSPEC_STOS): New.
26268 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26269 *strsetqi_1): Add UNSPEC_STOS.
26271 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26274 * doc/invoke.texi: Remove left over -Wsynth example.
26276 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26278 PR tree-optimization/56051
26279 * fold-const.c (fold_binary_loc): Don't fold
26280 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26281 a narrowing conversion, or widening conversion from signed
26284 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26286 PR rtl-optimization/56023
26287 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26288 dependent on debug instruction.
26290 2013-01-21 Martin Jambor <mjambor@suse.cz>
26292 PR middle-end/56022
26293 * function.c (allocate_struct_function): Call
26294 invoke_set_current_function_hook earlier.
26296 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26298 * reload1.c (init_reload): Only initialize reload_obstack
26299 during the first call.
26301 2013-01-21 Marek Polacek <polacek@redhat.com>
26303 * cfgloop.c (verify_loop_structure): Fix up grammar.
26305 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26307 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26308 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26310 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26313 * config/arm/marvell-pj4.md: Update copyright year.
26314 Fix up use of alu to alu_reg and simple_alu_imm.
26316 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26318 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26320 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26323 * lra-constraints.c (curr_insn_transform): Don't reuse original
26324 insn for secondary memory move when memory mode should be different.
26326 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26328 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26329 atomic_storedi_1): New patterns.
26331 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26333 btver2 pipeline descriptions.
26334 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26336 * config/i386/i386.md (btver2_decode): New type attributes.
26337 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26339 * config/i386/btver2.md: New file describing btver2 pipelines.
26341 2013-01-19 Andrew Pinski <apinski@cavium.com>
26343 PR tree-optimization/52631
26344 * tree-ssa-sccvn (visit_use): Before looking up the original
26345 statement, try looking up the simplified expression.
26347 2013-01-19 Anthony Green <green@moxielogic.com>
26349 * config/moxie/moxie.c (moxie_expand_prologue): Set
26350 current_function_static_stack_size.
26352 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26354 PR tree-optimization/56029
26355 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26356 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26358 2013-01-18 Sharad Singhai <singhai@google.com>
26360 PR tree-optimization/55995
26361 * dumpfile.c (dump_loc): Print location only if available.
26362 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26364 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26367 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26368 secondary memory move.
26369 (inherit_reload_reg): Use rclass instead of cl for
26370 check_secondary_memory_needed_p.
26372 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26374 PR middle-end/56015
26375 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26376 the case where writing real complex part of target modifies op1.
26378 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26380 * config/aarch64/aarch64-simd.md
26381 (aarch64_vcond_internal<mode>): Handle unordered cases.
26382 * config/aarch64/iterators.md (v_cmp_result): New.
26384 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26385 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26387 * config/arm/marvell-pj4.md: New file.
26388 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26389 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26390 (generic_vfp): Likewise.
26391 * config/arm/arm-cores.def: Add marvell-pj4.
26392 * config/arm/arm-tune.md: Regenerate.
26393 * config/arm/arm-tables.opt: Regenerate.
26394 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26395 * doc/invoke.texi: Document marvell-pj4.
26397 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26399 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26401 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26406 * var-tracking.c (negative_power_of_two_p): New.
26407 (global_get_addr_cache, local_get_addr_cache): New.
26408 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26409 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26411 (vt_stack_offset_p): Remove.
26412 (vt_canon_true_dep): Always canonicalize loc's address.
26413 (clobber_overlapping_mems): Make sure we have a MEM.
26414 (local_get_addr_clear_given_value): New.
26415 (val_reset): Clear local cached entries.
26416 (compute_bb_dataflow): Create and release the local cache.
26417 Disable duplicate MEMs clobbering.
26418 (emit_notes_in_bb): Clobber MEMs likewise.
26419 (vt_emit_notes): Create and release the local cache.
26420 (vt_initialize, vt_finalize): Create and release the global
26421 cache, respectively.
26422 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26424 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26426 PR libmudflap/53359
26427 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26428 not found in the symtab.
26430 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26433 PR rtl-optimization/55547
26434 PR rtl-optimization/53827
26437 * alias.c (offset_overlap_p): New, factored out of...
26438 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26439 the conservative special case for symbolic constants. Don't
26440 adjust zero sizes on alignment.
26442 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26444 PR rtl-optimization/52573
26445 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26446 REG_UNUSED for the same register.
26448 2013-01-17 Richard Biener <rguenther@suse.de>
26449 Marek Polacek <polacek@redhat.com>
26451 PR rtl-optimization/55833
26452 * loop-unswitch.c (unswitch_loops): Move loop verification...
26453 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26454 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26455 Set it to true when we're removing a loop from hierarchy tree in
26456 an irreducible region.
26457 (fix_bb_placements): Adjust caller.
26458 (fix_loop_placements): Likewise.
26460 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26462 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26463 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26464 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26465 Remove ID. Adjust comments.
26466 * config/avr/avr-c.c (avr_builtin_name): Remove.
26467 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26468 * config/avr/avr.c (avr_tolower): New static function.
26469 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26470 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26471 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26474 2013-01-17 Jan Hubicka <jh@suse.cz>
26476 PR tree-optimization/55273
26477 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26479 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26482 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26483 store through atomic_store<mode>_1.
26484 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26486 2013-01-17 Martin Jambor <mjambor@suse.cz>
26488 PR tree-optimizations/55264
26489 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26490 for virtual methods.
26491 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26492 virtual methods before inlining is over.
26493 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26495 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26498 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26500 PR rtl-optimization/56005
26501 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26502 pending reads for prefetch.
26504 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26506 * config/aarch64/aarch64.md
26507 (*cstoresi_neg_uxtw): New pattern.
26508 (*cmovsi_insn_uxtw): New pattern.
26509 (*<optab>si3_uxtw): New pattern.
26510 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26511 (*<optab>si3_insn_uxtw): New pattern.
26512 (*bswapsi2_uxtw): New pattern.
26514 2013-01-16 Richard Biener <rguenther@suse.de>
26516 * tree-inline.c (tree_function_versioning): Remove set but
26517 never used variable.
26519 2013-01-16 Richard Biener <rguenther@suse.de>
26521 PR tree-optimization/55964
26522 * tree-flow.h (rename_variables_in_loop): Remove.
26523 (rename_variables_in_bb): Likewise.
26524 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26525 (copy_loop_before): Adjust and delete update-ssa status.
26526 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26527 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26528 (rename_variables_in_loop): Remove.
26529 (slpeel_update_phis_for_duplicate_loop): Likewise.
26530 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26531 use available cfg machinery instead of duplicating it.
26532 Update PHI nodes and perform poor-mans SSA update here.
26533 (slpeel_tree_peel_loop_to_edge): Adjust.
26535 2013-01-16 Richard Biener <rguenther@suse.de>
26537 PR tree-optimization/54767
26538 PR tree-optimization/53465
26539 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26540 (vrp_visit_phi_node): For PHI arguments coming via backedges
26541 drop all symbolical range information.
26542 (execute_vrp): Compute backedges.
26544 2013-01-16 Richard Biener <rguenther@suse.de>
26546 * doc/install.texi: Update CLooG and ISL requirements to
26549 2013-01-16 Christian Bruel <christian.bruel@st.com>
26552 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26553 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26554 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26556 2013-01-16 DJ Delorie <dj@redhat.com>
26558 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26559 (UNSPECV_SP_SWITCH_E): New.
26560 (sp_switch_1): Change to an unspec.
26561 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26564 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26566 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26567 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26568 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26569 (expand_mem_thread_fence): Ditto.
26570 (expand_mem_signal_fence): Ditto.
26571 (expand_atomic_load): Ditto.
26572 (expand_atomic_store): Ditto.
26574 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26576 PR rtl-optimization/55547
26577 PR rtl-optimization/53827
26580 * alias.c (memrefs_conflict_p): Set sizes to negative after
26583 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26586 * function.c (thread_prologue_and_epilogue_insns): Always
26587 add crtl->drap_reg to set_up_by_prologue.set, even if
26588 stack_realign_drap is false.
26590 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26592 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26593 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26594 *call): Fix indention.
26596 2013-01-15 Tom de Vries <tom@codesourcery.com>
26599 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26602 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26604 PR rtl-optimization/55153
26605 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26607 2013-01-15 Martin Jambor <mjambor@suse.cz>
26609 PR tree-optimization/55920
26610 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26611 accesses as grp_to_be_debug_replaced.
26613 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26615 PR tree-optimization/55920
26616 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26617 there is non-useless type conversion needed from debug rhs to lhs,
26618 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26620 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26621 Mikael Pettersson <mikpe@it.uu.se>
26624 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26626 (ASM_OUTPUT_CASE_LABEL): Remove.
26627 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26628 * final.c (shorten_branches): Update alignment of labels before
26629 jump tables if CASE_VECTOR_SHORTEN_MODE.
26631 2013-01-15 Richard Biener <rguenther@suse.de>
26634 * system.h: Do not include gmp.h for building host tools.
26636 2013-01-15 Richard Biener <rguenther@suse.de>
26638 PR middle-end/55882
26639 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26640 account for bitpos when computing alignment.
26642 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26644 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26645 (ix86_target_macros_internal): Likewise.
26647 * config/i386/i386.c (m_CORE2I7): Removed.
26648 (m_CORE_HASWELL): New macro.
26649 (m_CORE_ALL): Likewise.
26650 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26651 (initial_ix86_arch_features): Likewise.
26652 (processor_target_table): Initializations for Core avx2.
26653 (cpu_names): New names "core-avx2".
26654 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26655 PROCESSOR_CORE_HASWELL.
26656 (ix86_issue_rate): New case.
26657 (ia32_multipass_dfa_lookahead): Likewise.
26658 (ix86_sched_init_global): Likewise.
26660 * config/i386/i386.h (TARGET_HASWELL): New macro.
26661 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26662 (processor_type): New PROCESSOR_HASWELL.
26664 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26666 PR tree-optimization/55955
26667 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26668 *SHIFT_EXPR and *ROTATE_EXPR codes.
26670 PR tree-optimization/48766
26671 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26672 -ftrapv disable -fwrapv.
26674 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26677 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26678 etc. to 1 and not to __flash.
26679 Use LL suffix for __INT24_MAX__ with -mint8.
26680 Use ULL suffix for __UINT24_MAX__ with -mint8.
26682 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26684 * config/avr/avr-arch.h
26685 (struct base_arch_s): Use typedef avr_arch_t instead.
26686 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26687 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26688 * config/avr/avr.c: Same.
26689 * config/avr/avr-devices.c: Same.
26690 * config/avr/driver-avr.c: Same.
26691 * config/avr/gen-avr-mmcu-texi.c: Same.
26692 * config/avr/avr-mcus.def: Adjust comment.
26694 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26696 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26697 * config/aarch64/iterators.md (VALLDI): New.
26699 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26700 Andi Kleen <ak@linux.intel.com>
26703 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26704 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26707 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26709 * config/avr/avr-stdint.h: Remove trailing blanks.
26710 * config/avr/avr-log.h: Same.
26711 * config/avr/avr-arch.h: Same.
26712 * config/avr/avr-devices.c: Same.
26713 * config/avr/avr-dimode.md: Same.
26714 * config/avr/predicates.md: Same.
26715 * config/avr/avr-c.c: Same. And fix typo.
26717 * config/avr/avr-protos.h: Same. And:
26718 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26719 (init_cumulative_args): Rename to avr_init_cumulative_args.
26720 (expand_prologue): Rename to avr_expand_prologue.
26721 (expand_epilogue): Rename to avr_expand_epilogue.
26722 (adjust_insn_length): Rename to avr_adjust_insn_length.
26723 (notice_update_cc): Rename to avr_notice_update_cc.
26724 (final_prescan_insn): Rename to avr_final_prescan_insn.
26725 * config/avr/avr.c: Same.
26726 * config/avr/avr.h: Same.
26727 * config/avr/avr.md: Remove trailing blanks.
26728 (prologue): Use avr_expand_prologue.
26729 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26731 2013-01-14 Richard Biener <rguenther@suse.de>
26733 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26734 verify_location, collect_subblocks): New functions.
26735 (verify_gimple_in_cfg): Verify that locations only reference
26736 BLOCKs in the functions BLOCK tree.
26738 2013-01-14 Richard Biener <rguenther@suse.de>
26740 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26742 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26744 (insert_out_of_ssa_copy_on_edge): Likewise.
26745 (rewrite_close_phi_out_of_ssa): Likewise.
26746 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26748 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26749 propagated constants.
26750 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26753 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26755 * config/avr/avr-modes.def: Add GPL copyright notice.
26757 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26759 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26760 MEMMODEL_MASK to determine memory model.
26761 (atomic_store<mode>): Ditto from operands[2].
26762 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26764 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26767 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26768 (fold_gimple_assign): Don't call unshare_expr here.
26769 (fold_ctor_reference): Call unshare_expr.
26771 2013-01-13 Terry Guo <terry.guo@arm.com>
26773 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26774 * doc/fragments.texi: Document MULTILIB_REUSE.
26775 * gcc.c (multilib_reuse): New internal spec.
26776 (set_multilib_dir): Also search multilib from multilib_reuse.
26777 * genmultilib (tmpmultilib3): Refactor code.
26778 (tmpmultilib4): Ditto.
26779 (multilib_reuse): New multilib argument.
26781 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26783 * Makefile.in: Update copyright.
26785 2013-01-12 Tom de Vries <tom@codesourcery.com>
26787 PR middle-end/55890
26788 * calls.c (expand_call): Check if arg_nr is valid.
26790 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26792 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26793 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26794 documentation. Add missing '__' in front of
26795 __builtin_ia32_packssdw256.
26797 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26800 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26801 NO_REGS for larl operands.
26802 (s390_reload_larl_operand): Use s390_load_address instead of
26805 2013-01-11 Richard Biener <rguenther@suse.de>
26807 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26808 (verify_node_sharing): ... here.
26809 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26811 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26813 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26814 Substitute TREECHECKING.
26815 * configure: Regenerate.
26816 * Makefile.in (TREECHECKING): New.
26818 2013-01-11 Richard Guenther <rguenther@suse.de>
26820 PR tree-optimization/44061
26821 * tree-vrp.c (extract_range_basic): Compute zero as
26822 value-range for __builtin_constant_p of function parameters.
26824 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26826 Update copyright years.
26828 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26830 PR rtl-optimization/55672
26831 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26832 const to be eliminable.
26834 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26836 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26837 * configure: Regenerate.
26839 2013-01-10 Richard Biener <rguenther@suse.de>
26841 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26843 2013-01-10 Richard Biener <rguenther@suse.de>
26846 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26847 locations for virtual PHI arguments.
26848 (rewrite_update_phi_arguments): Likewise.
26850 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26852 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26855 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26857 PR tree-optimization/55921
26858 * tree-complex.c (expand_complex_asm): New function.
26859 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26861 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26864 * config/s390/s390.c (s390_symref_operand_p)
26865 (s390_loadrelative_operand_p): Merge the two functions.
26866 (s390_check_qrst_address, print_operand_address): Add parameters
26867 to s390_loadrelative_operand_p invokation.
26868 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26869 (s390_reload_larl_operand, s390_secondary_reload): Use
26870 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26871 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26873 2013-01-09 Mike Stump <mikestump@comcast.net>
26875 * dse.c (record_store): Remove unnecessary assert.
26877 2013-01-09 Jan Hubicka <jh@suse.cz>
26879 PR tree-optimization/55569
26880 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26881 * cfgloop.h (scale_loop_profile): Likewise.
26883 2013-01-09 Jan Hubicka <jh@suse.cz>
26886 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26888 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26890 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26892 PR middle-end/55114
26893 * expr.h (maybe_emit_group_store): Declare.
26894 * expr.c (maybe_emit_group_store): New function.
26895 * builtins.c (expand_builtin_int_roundingfn): Call it.
26896 (expand_builtin_int_roundingfn_2): Likewise.
26898 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
26900 PR rtl-optimization/55829
26901 * lra-constraints.c (match_reload): Add code for absent output.
26902 (curr_insn_transform): Add code for reloads of matched inputs
26905 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
26907 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
26908 attribute of movddup insn to DF.
26909 (*vec_interleave_lowv2df): Ditto.
26910 (vec_dupv2df): Ditto.
26912 2013-01-09 Jan Hubicka <jh@suse.cz>
26914 PR tree-optimiation/55875
26915 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
26916 EVERY_ITERATION parameter.
26917 (number_of_iterations_exit): Check if exit is executed every iteration.
26918 (idx_infer_loop_bounds): Similarly here.
26919 (n_of_executions_at_most): Simplify
26920 to only test for cases where statement is dominated by the
26921 particular bound; handle correctly the "postdominance" test.
26922 (scev_probably_wraps_p): Use max loop iterations info
26923 as a global bound first.
26925 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
26926 Nick Clifton <nickc@redhat.com>
26928 * config/v850/v850.md (cbranchsf4): New pattern.
26929 (cstoresf4): New pattern.
26930 (cbranchdf4): New pattern.
26931 (cstoredf4): New pattern.
26932 (movsicc): Disallow floating point comparisons.
26933 (cmpsf_le_insn): Fix order of operators.
26934 (cmpsf_lt_insn): Likewise.
26935 (cmpsf_eq_insn): Likewise.
26936 (cmpdf_le_insn): Likewise.
26937 (cmpdf_lt_insn): Likewise.
26938 (cmpdf_eq_insn): Likewise.
26939 (cmpsf_ge_insn): Use LE comparison.
26940 (cmpdf_ge_insn): Likewise.
26941 (cmpsf_gt_insn): Use LT comparison.
26942 (cmpdf_gt_insn): Likewise.
26943 (cmpsf_ne_insn): Delete pattern.
26944 (cmpdf_ne_insn): Delete pattern.
26945 * config/v850/v850.c (v850_gen_float_compare): Use
26946 gen_cmpdf_eq_insn for NE comparison.
26947 (v850_float_z_comparison_operator)
26948 (v850_float_nz_comparison_operator): Move from here ...
26949 * config/v850/predicates.md: ... to here. Move GT and GE
26950 comparisons into v850_float_z_comparison_operator.
26951 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
26953 (v850_float_nz_comparison_operator): Likewise.
26955 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26957 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
26958 with calls to gen_insvsi/gen_insvdi.
26960 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26962 * config/i386/i386.c (initial_ix86_tune_features): Set up
26963 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
26965 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
26966 Jakub Jelinek <jakub@redhat.com>
26968 PR tree-optimization/48189
26969 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
26970 If nitercst is 0, don't predict the exit edge.
26972 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26974 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26975 in asm_fprintf with reg_names.
26976 (aarch64_print_operand_address): Likewise.
26977 (aarch64_return_addr): Likewise.
26978 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26980 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26982 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26983 (INT_U6_BITS): Likewise.
26984 * config/pa/predicates.md (uint6_operand): New predicate.
26985 (shift5_operand, shift6_operand): Likewise.
26986 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26988 (lshrdi3): Use shift6_operand.
26989 (shrpsi4, shrpdi4): New insn patterns.
26990 (extzv): Delete expander.
26991 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26992 predicates in unamed zero extract patterns. Tighten common constraint.
26993 (extv): Delete expander.
26994 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26995 predicates in unamed sign extract patterns. Tighten common constraint.
26996 (insv): Delete expander.
26997 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26998 predicates in unamed insert patterns. Tighten common constraint.
26999 Change uint32_operand predicate to uint6_operand predicate in unamed
27000 DImode pattern to insert constant values of type 1...1xxxx.
27002 2013-01-04 Jan Hubicka <jh@suse.cz>
27004 PR tree-optimization/55823
27005 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27008 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27009 Uros Bizjak <ubizjak@gmail.com>
27011 PR rtl-optimization/55845
27012 * df-problems.c (can_move_insns_across): Stop scanning at
27013 volatile_insn_p source instruction or give up if
27014 across_from .. across_to range contains any volatile_insn_p
27017 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27019 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27020 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27022 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27023 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27025 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27028 * asan.c (asan_clear_shadow): New function.
27029 (asan_emit_stack_protection): Use it.
27031 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27033 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27034 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27035 with tab instead of space.
27037 2013-01-08 Nick Clifton <nickc@redhat.com>
27039 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27040 register bank 0 at the start of an interrupt handler.
27041 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27044 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27046 * config/aarch64/aarch64-simd.md
27047 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27048 (aarch64_simd_bsl): Likewise.
27049 (aarch64_vcond_internal<mode>): Likewise.
27050 (vcond<mode><mode>): Likewise.
27051 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27052 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27054 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27056 * config/aarch64/aarch64-builtins.c
27057 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27059 2013-01-08 Martin Jambor <mjambor@suse.cz>
27062 * tree-sra.c (analyze_access_subtree): Return true also after
27063 potentially creating a debug-only replacement.
27065 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27067 PR middle-end/55890
27068 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27070 PR tree-optimization/54120
27071 * tree-vrp.c (range_fits_type_p): Don't allow
27072 src_precision < precision from signed vr to unsigned_p
27073 if vr->min or vr->max is negative.
27074 (simplify_float_conversion_using_ranges): Test can_float_p
27075 against CODE_FOR_nothing.
27077 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27078 Richard Biener <rguenther@suse.de>
27080 PR middle-end/55851
27081 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27082 types instead of just INTEGER_TYPE types.
27084 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27086 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27089 2013-01-07 Steve Ellcey <sellcey@mips.com>
27092 * config/mips/mips.opt: Change mad to mmad to match documentation.
27094 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27097 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27098 .progmemx.data now.
27100 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27103 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27104 (avr_addrspace_t): Add .section_name field.
27105 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27107 (avr_addrspace): Same. Initialize .section_name. Remove last
27108 NULL entry. Put __memx into .progmemx.data.
27109 (progmem_section_prefix): Remove.
27110 (avr_asm_init_sections): No need to initialize progmem_section.
27111 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27112 section name prefix.
27113 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27114 retrieve the progmem section.
27115 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27116 boundary to run over avr_addrspace[].
27117 (avr_register_target_pragmas): Ditto.
27119 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27121 * varasm.c (output_constant_def_contents): For asan_protect_global
27122 protected strings, adjust DECL_ALIGN if needed, before testing for
27124 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27125 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27127 (output_object_block): For asan protected decls, emit asan padding
27128 after their contents.
27129 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27130 (asan_finish_file): Test it here instead.
27132 2013-01-07 Nick Clifton <nickc@redhat.com>
27133 Matthias Klose <doko@debian.org>
27134 Doug Kwan <dougkwan@google.com>
27135 H.J. Lu <hongjiu.lu@intel.com>
27138 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27140 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27142 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27144 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27146 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27148 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27151 * doc/install.texi (Cross-Compiler-Specific Options): Document
27154 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27156 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27157 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27158 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27159 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27160 vqmovun_high_s64): Fix source operand number and update copyright.
27162 2013-01-07 Richard Biener <rguenther@suse.de>
27164 PR middle-end/55890
27165 * gimple.h (gimple_call_builtin_p): New overload.
27166 * gimple.c (validate_call): New function.
27167 (gimple_call_builtin_p): Likewise.
27168 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27169 Use gimple_call_builtin_p.
27170 (find_func_clobbers): Likewise.
27171 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27172 (strlen_optimize_stmt): Likewise.
27174 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27176 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27177 (vld1q_dup_*): Likewise.
27178 (vld1_*): Likewise.
27179 (vld1q_*): Likewise.
27180 (vld1_lane_*): Likewise.
27181 (vld1q_lane_*): Likewise.
27183 2013-01-07 Richard Biener <rguenther@suse.de>
27185 * lto-streamer.h (LTO_minor_version): Bump to 2.
27187 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27189 * config/aarch64/aarch64-protos.h
27190 (aarch64_const_double_zero_rtx_p): Rename to...
27191 (aarch64_float_const_zero_rtx_p): ...this.
27192 (aarch64_float_const_representable_p): New.
27193 (aarch64_output_simd_mov_immediate): Likewise.
27194 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27195 move immediate case.
27196 * config/aarch64/aarch64.c
27197 (aarch64_const_double_zero_rtx_p): Rename to...
27198 (aarch64_float_const_zero_rtx_p): ...this.
27199 (aarch64_print_operand): Allow printing of new constants.
27200 (aarch64_valid_floating_const): New.
27201 (aarch64_legitimate_constant_p): Check for valid floating-point
27203 (aarch64_simd_valid_immediate): Likewise.
27204 (aarch64_vect_float_const_representable_p): New.
27205 (aarch64_float_const_representable_p): Likewise.
27206 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27207 (aarch64_output_simd_mov_immediate): New.
27208 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27209 (*movdf_aarch64): Likewise.
27210 * config/aarch64/constraints.md (Ufc): New.
27211 (Y): call aarch64_float_const_zero_rtx.
27212 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27214 2013-01-07 Richard Biener <rguenther@suse.de>
27216 PR tree-optimization/55888
27217 PR tree-optimization/55862
27218 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27219 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27220 not if it is contained therein.
27222 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27224 * config/avr/t-avr: Typo.
27226 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27229 * config/avr/t-avr: Don't automatically rebuild
27230 $(srcdir)/config/avr/t-multilib
27231 $(srcdir)/config/avr/avr-tables.opt
27232 $(srcdir)/doc/avr-mmcu.texi
27233 (avr-mcus): New phony target to build them on request.
27234 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27235 * avr/avr-mcus.def: Adjust comments.
27237 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27239 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27241 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27243 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27245 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27247 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27249 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27251 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27252 to generate profiling.
27253 * config/rs6000/aix64.h (LIB_SPEC): Same.
27255 2013-01-04 Andrew Pinski <apinski@cavium.com>
27257 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27259 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27261 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27263 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27265 (ix86_expand_move): Ditto.
27266 (ix86_zero_extend_to_Pmode): Ditto.
27267 (ix86_expand_call): Ditto.
27268 (ix86_expand_special_args_builtin): Ditto.
27269 (ix86_expand_builtin): Ditto.
27271 2013-01-04 Richard Biener <rguenther@suse.de>
27273 PR tree-optimization/55862
27274 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27275 translating them through PHI nodes.
27277 2013-01-04 Martin Jambor <mjambor@suse.cz>
27279 PR tree-optimization/55755
27280 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27281 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27283 2013-01-04 Marek Polacek <polacek@redhat.com>
27285 PR middle-end/55859
27286 * opts.c (default_options_optimization): Clarify error message.
27288 2013-01-04 Richard Biener <rguenther@suse.de>
27290 PR middle-end/55863
27291 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27294 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27297 * config/pa/pa.md (movsi): Revert previous change.
27298 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27301 2013-01-03 Richard Henderson <rth@redhat.com>
27303 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27304 after eliminating TLS symbols.
27306 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27309 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27310 * graphite-poly.c (debug_gmp_value): Likewise.
27312 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27315 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27316 selected code model, define __code_mode_small__, __code_model_medium__,
27317 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27318 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27319 xchg temporary register with %k. Declare temporary register as
27321 [__x86_64__]: For medium and large code models, preserve %rbx register.
27323 2013-01-03 Richard Biener <rguenther@suse.de>
27325 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27326 (dump_subscript): Adjust.
27327 (finalize_ddr_dependent): Do not dump redundant info.
27328 (analyze_siv_subscript): Adjust.
27329 (subscript_dependence_tester): Likewise.
27330 (compute_affine_dependence): Likewise.
27332 2013-01-03 Richard Biener <rguenther@suse.de>
27335 2013-01-03 Richard Biener <rguenther@suse.de>
27337 PR tree-optimization/55857
27338 * tree-vect-stmts.c (vectorizable_load): Do not setup
27339 re-alignment for invariant loads.
27341 2013-01-02 Richard Biener <rguenther@suse.de>
27343 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27344 invariant load do not generate a vector load from the scalar location.
27346 2013-01-03 Richard Biener <rguenther@suse.de>
27348 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27349 for not vectorizing.
27350 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27351 not build INDIRECT_REFs, call get_name once only.
27352 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27353 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27355 2013-01-03 Richard Biener <rguenther@suse.de>
27357 PR tree-optimization/55857
27358 * tree-vect-stmts.c (vectorizable_load): Do not setup
27359 re-alignment for invariant loads.
27361 2013-01-03 Richard Biener <rguenther@suse.de>
27364 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27365 prefer a built-in decl.
27367 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27369 * gcc.c (process_command): Update copyright notice dates.
27370 * gcov.c (print_version): Likewise.
27371 * gcov-dump.c (print_version): Likewise.
27373 PR rtl-optimization/55838
27374 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27375 iv0.step, iv1.step and step.
27377 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27378 Marc Glisse <marc.glisse@inria.fr>
27380 PR tree-optimization/55832
27381 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27382 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27383 integer_{one,zero}_node.
27385 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27388 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27389 * var-tracking.c (reverse_op): Don't add reverse ops to
27390 VALUEs that have already
27391 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27393 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27395 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27397 2013-01-02 Teresa Johnson <tejohnson@google.com>
27399 * dumpfile.c (dump_loc): Print filename with location.
27400 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27401 new location_t parameter to emit complete unroll message with
27402 new dump framework.
27403 (canonicalize_loop_induction_variables): Compute loops location
27404 and pass to try_unroll_loop_completely.
27405 * loop-unroll.c (report_unroll_peel): New function.
27406 (peel_loops_completely): Use new dump format with location
27407 for main dumpfile message, and invoke report_unroll_peel on success.
27408 (decide_unrolling_and_peeling): Ditto.
27409 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27410 by report_unroll_peel.
27411 (decide_peel_completely): Ditto.
27412 (decide_unroll_constant_iterations): Ditto.
27413 (decide_unroll_runtime_iterations): Ditto.
27414 (decide_peel_simple): Ditto.
27415 (decide_unroll_stupid): Ditto.
27416 * cfgloop.c (get_loop_location): New function.
27417 * cfgloop.h (get_loop_location): Declare.
27419 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27421 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27424 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27426 PR middle-end/55198
27427 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27428 BLKmode objects when EXPAND_MEMORY is specified.
27430 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27432 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27434 (fold_builtin_cpu): Do not share cpu model decls across statements.
27436 2013-01-02 Jason Merrill <jason@redhat.com>
27439 * tree.c (build_array_type_1): Revert earlier change.
27441 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27443 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27445 * config/aarch64/aarch64-tune.md: Re-generate.
27447 2013-01-02 Richard Biener <rguenther@suse.de>
27449 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27450 invariant load do not generate a vector load from the scalar location.
27452 2013-01-02 Richard Biener <rguenther@suse.de>
27455 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27456 * configure: Regenerate.
27458 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27460 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27461 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27462 (expand_builtin_int_roundingfn_2): Keep the original target around
27463 for the fallback case.
27465 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27467 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27468 to be clear for sign changes.
27470 2013-01-01 Jan Hubicka <jh@suse.cz>
27472 * ipa-inline-analysis.c: Fix formatting.
27474 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27476 PR tree-optimization/55831
27477 * tree-vect-loop.c (get_initial_def_for_induction): Use
27478 gsi_after_labels instead of gsi_start_bb.
27480 Copyright (C) 2013 Free Software Foundation, Inc.
27482 Copying and distribution of this file, with or without modification,
27483 are permitted in any medium without royalty provided the copyright
27484 notice and this notice are preserved.