1 2013-10-23 Sriraman Tallam <tmsriram@google.com>
4 * config/i386/i386.c (ix86_option_override_internal):
5 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
6 (ix86_valid_target_attribute_tree):
7 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
8 Change TARGET_SSE to TARGET_SSE_P (opts->...)
10 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
12 * tree-ssa-loop.h: Remove include files.
13 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
14 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
15 * graphite-clast-to-gimple.c: Likewise.
16 * graphite-scop-detection.c: Likewise.
17 * graphite-sese-to-poly.c: Likewise.
18 * ipa-inline-analysis.c: Likewise.
19 * ipa-pure-const.c: Likewise.
20 * loop-init.c: Likewise.
22 * predict.c: Likewise.
23 * tree-cfg.c: Likewise.
24 * tree-cfgcleanup.c: Likewise.
25 * tree-chrec.c: Likewise.
26 * tree-data-ref.c: Likewise.
27 * tree-loop-distribution.c: Likewise.
28 * tree-parloops.c: Likewise.
29 * tree-predcom.c: Likewise.
30 * tree-scalar-evolution.c: Likewise.
31 * tree-ssa-address.c: Likewise.
32 * tree-ssa.c: Likewise.
33 * tree-ssa-dce.c: Likewise.
34 * tree-ssa-loop.c: Likewise.
35 * tree-ssa-loop-im.c: Likewise.
36 * tree-ssa-loop-ivcanon.c: Likewise.
37 * tree-ssa-loop-ivopts.c: Likewise.
38 * tree-ssa-loop-manip.c: Likewise.
39 * tree-ssa-loop-niter.c: Likewise.
40 * tree-ssa-loop-prefetch.c: Likewise.
41 * tree-ssa-loop-unswitch.c: Likewise.
42 * tree-ssa-reassoc.c: Likewise.
43 * tree-vect-data-refs.c: Likewise.
44 * tree-vect-loop.c: Likewise.
45 * tree-vect-loop-manip.c: Likewise.
46 * tree-vectorizer.c: Likewise.
47 * tree-vect-stmts.c: Likewise.
48 * tree-vrp.c: Likewise.
50 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
52 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
54 2013-10-23 Jakub Jelinek <jakub@redhat.com>
56 PR tree-optimization/58775
57 PR tree-optimization/58791
58 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
59 (insert_stmt_after): Rewritten, don't move the stmt, but really
61 (get_stmt_uid_with_default): Remove.
62 (build_and_add_sum): Use insert_stmt_after and
63 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only
65 (update_range_test): Set uid on stmts added by
66 force_gimple_operand_gsi. Don't immediately modify statements
67 in inter-bb optimization, just update oe->op values.
68 (optimize_range_tests): Return bool whether any changed have
70 (update_ops): New function.
71 (struct inter_bb_range_test_entry): New type.
72 (maybe_optimize_range_tests): Perform statement changes here.
73 (not_dominated_by, appears_later_in_bb, get_def_stmt,
74 ensure_ops_are_available): Remove.
75 (find_insert_point): Rewritten.
76 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
77 return LHS of the (new resp. old) stmt. Don't call
78 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
79 instead of last, move new stmt at the right place.
80 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
81 (negate_value): Likewise. Set uids.
82 (break_up_subtract_bb): Initialize uids.
83 (reassociate_bb): Adjust rewrite_expr_tree caller.
84 (do_reassoc): Don't call renumber_gimple_stmt_uids.
86 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
89 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
90 TARGET_32BIT final condition.
91 (mulsi3_internal2 and splitter): Same.
93 2013-10-23 Jeff Law <law@redhat.com>
95 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
96 through joiner blocks with abnormal outgoing edges.
98 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
99 Add parameter JOINERS, to allow/disallow threading through joiner
101 (thread_block): New. Call thread_block_1.
102 (mark_threaded_blocks): Remove code to filter out certain cases
103 of threading through joiner blocks.
104 (thread_through_all_blocks): Document how we can have a dangling
105 edge AUX field and clear it.
107 2013-10-23 Ian Lance Taylor <iant@google.com>
109 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
110 (C++ Dialect Options): Likewise.
111 (Optimize Options): Likewise.
113 2013-10-23 Tom de Vries <tom@codesourcery.com>
115 PR tree-optimization/58805
116 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
118 2013-10-23 Jakub Jelinek <jakub@redhat.com>
120 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
121 sequence based on get_range_info returned range.
123 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
125 * tree-ssa.h: Remove all #include's
126 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
127 * alias.c: Move required includes from tree-ssa.h.
129 * builtins.c: Likewise.
131 * cfgexpand.c: Likewise.
132 * cfghooks.c: Likewise.
133 * cfgloop.c: Likewise.
134 * cfgloopmanip.c: Likewise.
135 * cgraph.c: Likewise.
136 * cgraphbuild.c: Likewise.
137 * cgraphclones.c: Likewise.
138 * cgraphunit.c: Likewise.
140 * except.c: Likewise.
143 * fold-const.c: Likewise.
144 * ggc-page.c: Likewise.
145 * gimple-builder.c: Likewise.
146 * gimple-fold.c: Likewise.
147 * gimple-iterator.c: Likewise.
148 * gimple-low.c: Likewise.
149 * gimple-pretty-print.c: Likewise.
150 * gimple-ssa-strength-reduction.c: Likewise.
151 * gimple-streamer-in.c: Likewise.
152 * gimple-streamer-out.c: Likewise.
153 * gimplify.c: Likewise.
154 * graphite-blocking.c: Likewise.
155 * graphite-clast-to-gimple.c: Likewise.
156 * graphite-dependences.c: Likewise.
157 * graphite-interchange.c: Likewise.
158 * graphite-optimize-isl.c: Likewise.
159 * graphite-poly.c: Likewise.
160 * graphite-scop-detection.c: Likewise.
161 * graphite-sese-to-poly.c: Likewise.
162 * graphite.c: Likewise.
163 * ipa-cp.c: Likewise.
164 * ipa-inline-analysis.c: Likewise.
165 * ipa-inline-transform.c: Likewise.
166 * ipa-inline.c: Likewise.
167 * ipa-prop.c: Likewise.
168 * ipa-pure-const.c: Likewise.
169 * ipa-reference.c: Likewise.
170 * ipa-split.c: Likewise.
171 * ipa-utils.c: Likewise.
172 * loop-init.c: Likewise.
173 * lto-cgraph.c: Likewise.
174 * lto-section-in.c: Likewise.
175 * lto-section-out.c: Likewise.
176 * lto-streamer-in.c: Likewise.
177 * lto-streamer-out.c: Likewise.
178 * lto-streamer.c: Likewise.
179 * omp-low.c: Likewise.
180 * passes.c: Likewise.
181 * predict.c: Likewise.
182 * print-tree.c: Likewise.
183 * profile.c: Likewise.
185 * targhooks.c: Likewise.
186 * tracer.c: Likewise.
187 * trans-mem.c: Likewise.
188 * tree-call-cdce.c: Likewise.
189 * tree-cfg.c: Likewise.
190 * tree-cfgcleanup.c: Likewise.
191 * tree-chrec.c: Likewise.
192 * tree-complex.c: Likewise.
193 * tree-data-ref.c: Likewise.
194 * tree-dfa.c: Likewise.
195 * tree-eh.c: Likewise.
196 * tree-emutls.c: Likewise.
197 * tree-if-conv.c: Likewise.
198 * tree-inline.c: Likewise.
199 * tree-into-ssa.c: Likewise.
200 * tree-loop-distribution.c: Likewise.
201 * tree-mudflap.c: Likewise.
202 * tree-nested.c: Likewise.
203 * tree-nrv.c: Likewise.
204 * tree-object-size.c: Likewise.
205 * tree-outof-ssa.c: Likewise.
206 * tree-parloops.c: Likewise.
207 * tree-phinodes.c: Likewise.
208 * tree-predcom.c: Likewise.
209 * tree-pretty-print.c: Likewise.
210 * tree-profile.c: Likewise.
211 * tree-scalar-evolution.c: Likewise.
212 * tree-sra.c: Likewise.
213 * tree-ssa-address.c: Likewise.
214 * tree-ssa-alias.c: Likewise.
215 * tree-ssa-ccp.c: Likewise.
216 * tree-ssa-coalesce.c: Likewise.
217 * tree-ssa-copy.c: Likewise.
218 * tree-ssa-copyrename.c: Likewise.
219 * tree-ssa-dce.c: Likewise.
220 * tree-ssa-dom.c: Likewise.
221 * tree-ssa-dse.c: Likewise.
222 * tree-ssa-forwprop.c: Likewise.
223 * tree-ssa-ifcombine.c: Likewise.
224 * tree-ssa-live.c: Likewise.
225 * tree-ssa-loop-ch.c: Likewise.
226 * tree-ssa-loop-im.c: Likewise.
227 * tree-ssa-loop-ivcanon.c: Likewise.
228 * tree-ssa-loop-ivopts.c: Likewise.
229 * tree-ssa-loop-manip.c: Likewise.
230 * tree-ssa-loop-niter.c: Likewise.
231 * tree-ssa-loop-prefetch.c: Likewise.
232 * tree-ssa-loop-unswitch.c: Likewise.
233 * tree-ssa-loop.c: Likewise.
234 * tree-ssa-math-opts.c: Likewise.
235 * tree-ssa-operands.c: Likewise.
236 * tree-ssa-phiopt.c: Likewise.
237 * tree-ssa-phiprop.c: Likewise.
238 * tree-ssa-pre.c: Likewise.
239 * tree-ssa-propagate.c: Likewise.
240 * tree-ssa-reassoc.c: Likewise.
241 * tree-ssa-sccvn.c: Likewise.
242 * tree-ssa-sink.c: Likewise.
243 * tree-ssa-strlen.c: Likewise.
244 * tree-ssa-structalias.c: Likewise.
245 * tree-ssa-tail-merge.c: Likewise.
246 * tree-ssa-ter.c: Likewise.
247 * tree-ssa-threadedge.c: Likewise.
248 * tree-ssa-threadupdate.c: Likewise.
249 * tree-ssa-uncprop.c: Likewise.
250 * tree-ssa-uninit.c: Likewise.
251 * tree-ssa.c: Likewise.
252 * tree-ssanames.c: Likewise.
253 * tree-stdarg.c: Likewise.
254 * tree-streamer-in.c: Likewise.
255 * tree-switch-conversion.c: Likewise.
256 * tree-tailcall.c: Likewise.
257 * tree-vect-data-refs.c: Likewise.
258 * tree-vect-generic.c: Likewise.
259 * tree-vect-loop-manip.c: Likewise.
260 * tree-vect-loop.c: Likewise.
261 * tree-vect-patterns.c: Likewise.
262 * tree-vect-slp.c: Likewise.
263 * tree-vect-stmts.c: Likewise.
264 * tree-vectorizer.c: Likewise.
265 * tree-vrp.c: Likewise.
268 * value-prof.c: Likewise.
269 * var-tracking.c: Likewise.
270 * varpool.c: Likewise.
271 * vtable-verify.c: Likewise.
273 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
275 * config/tilegx/tilegx.c: Include "tree.h".
277 2013-10-23 Jakub Jelinek <jakub@redhat.com>
279 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
280 the info, not after it.
281 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
282 dump_ssaname_info call.
283 (pp_gimple_stmt_1): Adjust caller.
284 (dump_phi_nodes): Likewise. Don't print "# " here.
286 2013-10-22 Jan Hubicka <jh@suse.cz>
288 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
289 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
290 * i386.c (expand_small_movmem_or_setmem): New function.
291 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
292 (alg_usable_p): Add support for value ranges; cleanup.
293 (ix86_expand_set_or_movmem): Add support for misaligned moves.
295 2013-10-22 Sterling Augustine <saugustine@google.com>
297 * doc/invoke.texi: Document -ggnu-pubnames.
298 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
300 * dwarf2out.c: Include gdb/gdb-index.h.
301 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
302 debug_generate_pub_sections.
303 (is_java, output_pubtables, output_pubname): New functions.
304 (include_pubname_in_output): Handle debug_generate_pub_sections at
306 (size_of_pubnames): Use new local space_for_flags based on
307 debug_generate_pub_sections.
308 (output_pubnames): Unify pubnames and pubtypes output logic.
309 Genericize comments. Call output_pubname.
310 (dwarf2out_finish): Move logic to output_pubnames and call it.
312 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
315 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
316 Remove CCCmode handling.
317 <case LTU>: Return 'c' suffix for CCCmode.
318 <case GEU>: Return 'nc' suffix for CCCmode.
319 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
320 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
321 (*sub<mode>3_cc_overflow): Ditto.
322 (*subsi3_zext_cc_overflow): Ditto.
324 2013-10-22 Steve Ellcey <sellcey@mips.com>
326 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
327 (AND (NOT OP1) (NOT OP2)).
329 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
331 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
332 meaning of merge-high and merge-low masks for little endian; avoid
333 use of vector-pack masks for little endian for mismatched modes.
335 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
337 * config/tilepro/tilepro.c: Include "tree.h".
339 2013-10-22 Andreas Schwab <schwab@suse.de>
341 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
344 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
346 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
347 [Ed Smith-Rowland]): New entries.
348 ([Stephen M. Webb]): Update.
350 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
352 * tree-ssa-ter.h: Remove duplicate copy of file contents.
354 2013-10-21 Marek Polacek <polacek@redhat.com>
357 * fold-const.c (fold_range_test): Return 0 if the type is not
360 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
362 * system.h: Move hwint.h include further down.
363 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
364 gcc_checking_asserts.
365 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
367 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
369 Fix volatile issues in optimize_bit_field_compare.
370 * fold-const.c (optimize_bit_field_compare): Bail out if
371 lvolatilep or rvolatilep.
373 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
375 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
376 and get_inner_reference returning different pmode for non-volatile
377 bit-field members dependent on flag_strict_volatile_bitfields.
378 * stor-layout.c (layout_decl): Remove special handling of
379 flag_strict_volatile_bitfields.
380 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
381 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
383 2013-10-21 Paulo Matos <pmatos@broadcom.com>
385 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
388 2013-10-21 Jeff Law <law@redhat.com>
390 * tree-ssa-threadedge.c (thread_through_normal_block): New
391 argument VISITED. Remove VISISTED as a local variable. When we
392 have a threadable jump, verify the destination of the jump has not
394 (thread_across_edge): Allocate VISITED bitmap once at function
395 scope and use it throughout. Make sure to set appropriate bits in
396 VISITED for E (start of jump thread path).
397 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
398 through a joiner if any edge on the path has a recorded jump thread.
400 2013-10-21 Ian Lance Taylor <iant@google.com>
402 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
403 don't imply that attributes can solve all problems.
404 (Directory Options): Fix typo.
406 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
408 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
409 extend and extend_add.
411 2013-10-21 Richard Biener <rguenther@suse.de>
413 PR tree-optimization/58794
414 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
415 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
417 2013-10-21 Richard Biener <rguenther@suse.de>
420 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
421 to (T) X for sign-changing conversions (or no conversion).
423 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
425 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
427 2013-10-20 Jan Hubicka <jh@suse.cz>
429 * config/i386/i386-tune.def: Add comment; organize into categories
431 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
433 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
434 argument. Update function comment.
435 (expand_set_or_movmem_via_rep): New function combining
436 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
437 (expand_movmem_via_rep_mov): Remove.
438 expand_setmem_via_rep_stos): Remove.
439 (expand_movmem_epilogue): Update calls correspondingly.
440 (expand_setmem_epilogue_via_loop): Likewise.
442 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
443 (expand_set_or_movmem_prologue): New function combining
444 expand_movmem_prologue and expand_setmem_prologue.
445 (expand_movmem_prologue): Remove.
446 (expand_setmem_prologue): Remove.
447 (expand_set_or_movmem_constant_prologue): New function combining
448 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
449 (expand_constant_movmem_prologue): Remove.
450 (expand_constant_setmem_prologue): Remove.
451 (promote_duplicated_reg): Allow vector-const0 value.
452 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
453 and ix86_expand_setmem.
454 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
455 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
457 2013-10-21 Diego Novillo <dnovillo@google.com>
459 * asan.c: Include tree.h
460 * bb-reorder.c: Likewise.
461 * cfgcleanup.c: Likewise.
462 * cfgloopmanip.c: Likewise.
463 * data-streamer-in.c: Likewise.
464 * data-streamer-out.c: Likewise.
465 * data-streamer.c: Likewise.
466 * dwarf2cfi.c: Likewise.
467 * graphite-blocking.c: Likewise.
468 * graphite-clast-to-gimple.c: Likewise.
469 * graphite-dependences.c: Likewise.
470 * graphite-interchange.c: Likewise.
471 * graphite-optimize-isl.c: Likewise.
472 * graphite-poly.c: Likewise.
473 * graphite-scop-detection.c: Likewise.
474 * graphite-sese-to-poly.c: Likewise.
475 * graphite.c: Likewise.
476 * ipa-devirt.c: Likewise.
477 * ipa-profile.c: Likewise.
480 * loop-init.c: Likewise.
481 * loop-unroll.c: Likewise.
482 * lower-subreg.c: Likewise.
483 * lto/lto-object.c: Likewise.
485 * reginfo.c: Likewise.
486 * tree-loop-distribution.c: Likewise.
487 * tree-parloops.c: Likewise.
488 * tree-ssa-strlen.c: Likewise.
489 * tree-streamer.c: Likewise.
490 * value-prof.c: Likewise.
491 * target-globals.c: Likewise.
492 * expr.h: Include tree-core.h instead of tree.h.
493 * gimple.h: Likewise.
494 * ipa-prop.h: Likewise.
495 * ipa-utils.h: Likewise.
496 * lto-streamer.h: Likewise.
497 * streamer-hooks.h: Likewise.
498 * ipa-reference.h: Include cgraph.h instead of tree.h.
499 * cgraph.h: Include basic-block.h instead of tree.h.
500 * tree-streamer.h: Do not include tree.h.
501 * genattrtab.c (write_header): Generate inclusion of
503 * genautomata.c (main): Likewise.
504 * genemit.c: Likewise.
505 * genopinit.c: Likewise.
506 * genoutput.c (output_prologue): Likewise.
507 * genpeep.c: Likewise.
509 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
511 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
513 (vec_unpacku_hi_v8hi): Likewise.
514 (vec_unpacku_lo_v16qi): Likewise.
515 (vec_unpacku_lo_v8hi): Likewise.
517 2013-10-20 Jan Hubicka <jh@suse.cz>
519 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
520 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
521 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
523 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
525 * config/mips/mips.h (ISA_HAS_WSBH): Define.
526 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
528 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
530 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
533 * system.h: Undef m_slot.
535 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
537 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
538 all elements for both endian flavors.
540 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
543 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
544 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
545 and SI_REG for 64bit SYSV ABI targets.
547 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
549 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
550 to multi_reg_return. Clarify that we are skipping USEs of multiple
551 return registers. Use bool type where appropriate.
553 2013-10-18 Jan Hubicka <jh@suse.cz>
555 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
557 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
558 (X86_TUNE_PUSH_MEMORY): Likewise.
559 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
560 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
561 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
563 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
564 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
566 (ix86_option_override_internal): Update to use tune features instead
569 2013-10-18 Cong Hou <congh@google.com>
571 PR tree-optimization/58508
572 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
573 statement that contains data refs with zero-step.
575 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
577 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
578 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
579 * gimple-low.c (gimple_check_call_arg,
580 gimple_check_call_matching_types): Move to cgraph.c.
581 * gimple-low.h: Remove prototype.
582 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
583 Relocate from gimple-low.c.
584 * cgraph.h: Add prototype. Don't include basic-block.h.
585 * gimplify.c: Add gimple-low to include list.
586 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
587 * tree-eh.c: Add gimple-low to include list.
588 * tree-nested.c: Likewise.
589 * cfgexpand.c: Add tree-ssa-address.h to include list.
591 * gimple-fold.c: Likewise.
592 * gimple-ssa-strength-reduction.c: Likewise.
593 * trans-mem.c: Likewise.
594 * tree-mudflap.c: Likewise.
595 * tree-ssa-loop-ivopts.c: Likewise.
596 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
597 (degenerate_phi_result): Move to tree-phinodes.c.
598 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
599 * tree-ssa-threadedge.c: Likewise.
600 * tree-vrp.c: Likewise.
601 * tree-phinodes.c (degenerate_phi_result): Relocate here.
602 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
603 * tree-phinodes.h (degenerate_phi_result): Add prototype.
604 * tree-ssa-copy.c: Include tree-ssa-dom.h.
605 * tree-ssa-forwprop.c: Likewise.
606 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
607 pass_data_cleanup_cfg_post_optimizing,
608 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
609 * tree-optimize.c: Delete File.
610 * graphite.c: Include tree-cfgcleanup.h.
611 * passes.c: Likewise.
612 * tree-cfg.c: Likewise.
613 * tree-profile.c: Likewise.
614 * tree-ssa-dse.c: Likewise.
615 * tree-ssa-loop-ivcanon.c: Likewise.
616 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
617 * tree-outof-ssa.c: Include sbitmap.h.
618 * tree-ssa-live.c: Likewise.
619 * tree-ssa-propagate.c: Likewise.
620 * tree-ssa-structalias.c: Likewise.
621 * tree-stdarg.c: Likewise.
622 * Makefile.in (OBJS): Delete tree-optimize.o.
623 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
624 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
625 * varasm.c: Include basic-block.h.
626 * cfgloop.h: Include function.h instead of basic-block.h
627 (bb_loop_depth): Move to cfgloop.c.
628 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
630 2013-10-18 Teresa Johnson <tejohnson@google.com>
632 * predict.c (probably_never_executed): Compare frequency-based
633 count to number of training runs.
634 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
636 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
638 * config/arm/arm.c (cortexa9_extra_costs): New table.
639 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
641 2013-10-18 Jeff Law <law@redhat.com>
643 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
645 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
646 (register_jump_thread): Add "registered_jump_thread" debug
648 * dbgcnt.def (registered_jump_thread): New debug counter.
650 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
652 * config/rs6000/rs6000.c: Include cgraph.h.
654 2013-10-18 Teresa Johnson <tejohnson@google.com>
656 * tree-ssa-tail-merge.c (replace_block_by): Update edge
657 weights during merging.
659 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
661 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
662 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
663 * tree-ssa-operands.h: Remove prototype.
664 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
665 * gimple.c (virtual_operand_p): Relocate from gimple.c.
666 * gimple.h: Add prototype.
667 * gimple-ssa.h: Include tree-ssa-operands.h.
668 * tree-dump.c: Add tree-cfg.h to include list.
669 * tree-ssa-alias.c: Add ipa-reference.h to include list.
670 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
671 * config/i386/i386.c: Don't include tree-flow.h.
672 * config/rs6000/rs6000.c: Likewise.
674 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
676 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
678 2013-10-18 Richard Biener <rguenther@suse.de>
680 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
681 or TYPE_UNSIGNED of integral types.
682 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
683 NULL_TREE for zero-precision integral types.
685 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
687 * config/aarch64/arm_neon.h
688 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
690 2013-10-17 Sriraman Tallam <tmsriram@google.com>
693 * opth-gen.awk: Define target_flags_explicit.
695 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
697 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
698 fields to the reg_addr array that describes the valid addressing
699 mode for any register, general purpose registers, floating point
700 registers, and Altivec registers.
701 (FIRST_RELOAD_REG_CLASS): Likewise.
702 (LAST_RELOAD_REG_CLASS): Likewise.
703 (struct reload_reg_map_type): Likewise.
704 (reload_reg_map_type): Likewise.
705 (RELOAD_REG_VALID): Likewise.
706 (RELOAD_REG_MULTIPLE): Likewise.
707 (RELOAD_REG_INDEXED): Likewise.
708 (RELOAD_REG_OFFSET): Likewise.
709 (RELOAD_REG_PRE_INCDEC): Likewise.
710 (RELOAD_REG_PRE_MODIFY): Likewise.
711 (reg_addr): Likewise.
712 (mode_supports_pre_incdec_p): New helper functions to say whether
713 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
714 (mode_supports_pre_modify_p): Likewise.
715 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
716 print the valid address mode bits for each mode.
717 (rs6000_debug_print_mode): Likewise.
718 (rs6000_debug_reg_global): Likewise.
719 (rs6000_setup_reg_addr_masks): New function to set up the address
720 mask bits for each type.
721 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
722 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
723 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
724 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
725 PRE_MODIFY are supported.
726 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
727 registers, instead of {src,dest}_av_p.
728 (rs6000_print_options_internal): Tweak the debug output slightly.
730 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
732 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
735 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
737 * tree-flow.h (struct omp_region): Move to omp-low.c.
738 Remove omp_ prototypes and variables.
739 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
740 (copy_var_decl): Relocate prototype from tree-flow.h.
741 * gimple.c (copy_var_decl): Relocate from omp-low.c.
742 * tree.h: Move prototype to omp-low.h.
743 * omp-low.h: New File. Relocate prototypes here.
744 * omp-low.c (struct omp_region): Make local here.
745 (root_omp_region): Make static.
746 (copy_var_decl) Move to gimple.c.
747 (new_omp_region): Make static.
748 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
749 * tree-cfg.c: Include omp-low.h.
750 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
751 * gimplify.c: Include omp-low.h.
752 * tree-parloops.c: Likewise.
754 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
756 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
757 operands are in memory, prefer to force non-matched operand 1 to
760 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
763 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
764 restrict TImode addresses to single indirect registers if both
765 -mquad-memory and -mvsx-timode are used.
766 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
767 we should emit load/store quad. Remove using %y for quad memory
770 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
771 constraints to allow load/store quad on machines where TImode is
772 not allowed in VSX registers. Use 'n' instead of 'F' constraint
773 for TImode to load integer constants.
775 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
777 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
779 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
781 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
782 handling of STACK_REG.
784 2013-10-17 Richard Biener <rguenther@suse.de>
786 PR tree-optimization/58143
787 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
789 (rewrite_to_defined_overflow): Likewise.
790 (move_computations_dom_walker::before_dom): Rewrite stmts
791 with undefined signed overflow that are not always executed
792 into unsigned arithmetic.
794 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
797 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
798 explicit isa flag to be an options variable, instead of using
799 global_options_set. Remove define from rs6000.h.
800 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
802 * config/rs6000/rs6000.c (rs6000_option_override_internal):
803 Initialize rs6000_isa_flags_explicit.
804 (rs6000_function_specific_save): Add gcc_options* parameter, so
805 that the powerpc builds after the 2013-10-15 changes.
806 (rs6000_function_specific_restore): Likewise.
808 2013-10-16 DJ Delorie <dj@redhat.com>
810 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
811 op is a REG before checking REGNO.
812 (rl78_alloc_physical_registers): Verify pattern is a SET before
815 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
817 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
819 (vec_unpacks_lo_v4sf): Likewise.
820 (vec_unpacks_float_hi_v4si): Likewise.
821 (vec_unpacks_float_lo_v4si): Likewise.
822 (vec_unpacku_float_hi_v4si): Likewise.
823 (vec_unpacku_float_lo_v4si): Likewise.
825 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
827 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
828 (vsx_concat_v2sf): Likewise.
830 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
832 * config/aarch64/aarch64.md
833 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
835 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
837 PR tree-optimization/58697
838 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
839 estimated_loop_iterations_int.
840 (max_stmt_executions_int): Call get_max_loop_iterations_int.
841 (get_max_loop_iterations_int): New. HWINT version of
842 get_max_loop_iterations.
843 * cfgloop.h: Add prototypes.
844 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
845 * loop-unroll.c (decide_peel_once_rolling): Call
846 get_estimated_loop_iterations_int.
847 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
848 * tree-ssa-loop-niter.h: Tweak prototypes.
850 2013-10-16 David Malcolm <dmalcolm@redhat.com>
852 * gengtype-parse.c (struct_field_seq): Ignore access-control
853 keywords ("public:" etc).
855 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
857 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
858 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
860 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
862 * config/arm/arm.opt (mlra): New option.
863 * config/arm/arm.c (arm_lra_p): New function.
864 (TARGET_LRA_P): Define.
866 2013-10-16 Paulo Matos <pmatos@broadcom.com>
868 * tree-core.h (tree_code_name): Remove.
869 * tree.h (get_tree_code_name): New prototype.
870 * tree.c (tree_code_name): Make static.
871 (get_tree_code_name): New function.
872 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
873 tree_class_check_failed, tree_range_check_failed,
874 tree_not_class_check_failed, omp_clause_check_failed,
875 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
876 wrapper get_tree_code_name instead of calling tree_code_name directly.
877 * tree-vrp.c (dump_asserts_for): Likewise.
878 * tree-dump.c (dequeue_and_dump): Likewise.
879 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
880 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
881 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
882 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
883 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
884 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
885 dump_gimple_omp_for): Likewise.
886 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
887 * tree-ssa-pre.c (print_pre_expr): Likewise.
888 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
889 * print-tree.c (print_node_brief, print_node): Likewise.
890 * gimple.c (gimple_check_failed): Likewise.
891 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
892 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
893 * config/mep/mep.c (mep_validate_vliw): Likewise.
894 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
895 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
897 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
899 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
902 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
904 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
905 (MULTILIB_MATCHES): Add multilib for -march=v8.
907 2013-10-15 Sriraman Tallam <tmsriram@google.com>
910 * optc-save-gen.awk: Add extra parameter to the save and restore
912 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
913 * tree.c (build_optimization_node): New parameter. Add extra parameter
914 to call to cl_optimization_save.
915 (build_target_option_node): New parameter. Add extra parameter
916 to call to cl_target_option_save.
917 * tree.h (build_optimization_node): New parameter.
918 (build_target_option_node): New parameter.
919 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
920 build_optimization_node and build_target_option_node.
921 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
922 (handle_pragma_push_options): Ditto.
923 * toplev.c (process_options): Ditto.
924 * opts.c (init_options_struct): Check for opts_set non-null.
925 * target.def (target_option.save): New parameter.
926 (target_option.restore): New parameter.
928 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
929 (ix86_pragma_target_parse): Ditto.
930 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
932 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
933 to build_optimization_node and build_target_option_node.
934 (rs6000_valid_attribute_p): Ditto.
935 (rs6000_pragma_target_parse): Ditto.
936 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
938 * config/i386/i386.h:
939 TARGET_64BIT_P: New Macro
940 TARGET_MMX_P: New Macro.
941 TARGET_3DNOW_P: New Macro.
942 TARGET_3DNOW_A_P: New Macro.
943 TARGET_SSE_P: New Macro.
944 TARGET_SSE2_P: New Macro.
945 TARGET_SSE3_P: New Macro.
946 TARGET_SSSE3_P: New Macro.
947 TARGET_SSE4_1_P: New Macro.
948 TARGET_SSE4_2_P: New Macro.
949 TARGET_AVX_P: New Macro.
950 TARGET_AVX2_P: New Macro.
951 TARGET_AVX512F_P: New Macro.
952 TARGET_AVX512PF_P: New Macro.
953 TARGET_AVX512ER_P: New Macro.
954 TARGET_AVX512CD_P: New Macro.
955 TARGET_FMA_P: New Macro.
956 TARGET_SSE4A_P: New Macro.
957 TARGET_FMA4_P: New Macro.
958 TARGET_XOP_P: New Macro.
959 TARGET_LWP_P: New Macro.
960 TARGET_ABM_P: New Macro.
961 TARGET_BMI_P: New Macro.
962 TARGET_BMI2_P: New Macro.
963 TARGET_LZCNT_P: New Macro.
964 TARGET_TBM_P: New Macro.
965 TARGET_POPCNT_P: New Macro.
966 TARGET_SAHF_P: New Macro.
967 TARGET_MOVBE_P: New Macro.
968 TARGET_CRC32_P: New Macro.
969 TARGET_AES_P: New Macro.
970 TARGET_PCLMUL_P: New Macro.
971 TARGET_CMPXCHG16B_P: New Macro.
972 TARGET_FSGSBASE_P: New Macro.
973 TARGET_RDRND_P: New Macro.
974 TARGET_F16C_P: New Macro.
975 TARGET_RTM_P: New Macro.
976 TARGET_HLE_P: New Macro.
977 TARGET_RDSEED_P: New Macro.
978 TARGET_PRFCHW_P: New Macro.
979 TARGET_ADX_P: New Macro.
980 TARGET_FXSR_P: New Macro.
981 TARGET_XSAVE_P: New Macro.
982 TARGET_XSAVEOPT_P: New Macro.
983 TARGET_LP64_P: New Macro.
984 TARGET_X32_P: New Macro.
985 TARGET_FPMATH_DEFAULT_P: New Macro.
986 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
987 * config/i386/i386.c (ix86_option_override_internal): New parameters.
989 Change ix86_tune_string to access opts->x_ix86_tune_string.
990 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
991 Change ix86_arch_string to access opts->x_ix86_arch_string.
992 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
993 Change ix86_pmode to access opts->x_ix86_pmode.
994 Change ix86_abi to access opts->x_ix86_abi.
995 Change ix86_cmodel to access opts->x_ix86_cmodel.
996 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
997 Change ix86_isa_flags_explicit to access
998 opts->x_ix86_isa_flags_explicit.
999 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
1000 Change ix86_regparm to access opts->x_ix86_regparm.
1001 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
1002 Change ix86_preferred_stack_boundary_arg to access
1003 opts->x_ix86_preferred_stack_boundary_arg.
1004 Change ix86_force_align_arg_pointer to access
1005 opts->x_ix86_force_align_arg_pointer.
1006 Change ix86_incoming_stack_boundar_arg to access
1007 opts->x_ix86_incoming_stack_boundar_arg.
1008 Change ix86_fpmath to access opts->x_ix86_fpmath.
1009 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
1010 Change ix86_recip_name to access opts->x_ix86_recip_name.
1011 Change ix86_stack_protector_guard to access
1012 opts->x_ix86_stack_protector_guard.
1013 Change ix86_tune_memcpy_strategy to access
1014 opts->x_ix86_tune_memcpy_strategy.
1015 Change ix86_tune_memset_strategy to access
1016 opts->x_ix86_tune_memset_strategy.
1017 Change global_options to access opts.
1018 Change global_options_set to access opts_set.
1019 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
1020 Change TARGET_MMX to TARGET_MMX_P (opts->...).
1021 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
1022 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
1023 Change TARGET_SSE to TARGET_SSE_P (opts->...).
1024 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
1025 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
1026 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
1027 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
1028 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
1029 Change TARGET_AVX to TARGET_AVX_P (opts->...).
1030 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
1031 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
1032 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
1033 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
1034 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
1035 Change TARGET_FMA to TARGET_FMA_P (opts->...).
1036 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
1037 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
1038 Change TARGET_XOP to TARGET_XOP_P (opts->...).
1039 Change TARGET_LWP to TARGET_LWP_P (opts->...).
1040 Change TARGET_ABM to TARGET_ABM_P (opts->...).
1041 Change TARGET_BMI to TARGET_BMI_P (opts->...).
1042 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
1043 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
1044 Change TARGET_TBM to TARGET_TBM_P (opts->...).
1045 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
1046 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
1047 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
1048 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
1049 Change TARGET_AES to TARGET_AES_P (opts->...).
1050 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
1051 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
1052 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
1053 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
1054 Change TARGET_F16C to TARGET_F16C_P (opts->...).
1055 Change TARGET_RTM to TARGET_RTM_P (opts->...).
1056 Change TARGET_HLE to TARGET_HLE_P (opts->...).
1057 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
1058 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
1059 Change TARGET_ADX to TARGET_ADX_P (opts->...).
1060 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
1061 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
1062 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
1063 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
1064 Change TARGET_X32 to TARGET_X32_P (opts->...).
1065 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
1066 Change TARGET_FLOAT_RETURNS_IN_80387 to
1067 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
1068 (ix86_function_specific_save): New parameter. Use opts-> fields
1069 to replace global fields.
1070 (ix86_function_specific_restore): Ditto.
1071 (ix86_valid_target_attribute_inner_p): New parameters.
1073 Fix call to ix86_handle_option and set_option.
1074 (ix86_valid_target_attribute_tree): New parameters.
1075 Change global_options to access opts.
1076 Change global_options_set to access opts_set.
1077 Fix call to ix86_valid_target_attribute_inner_p.
1078 Change ix86_tune_string to access opts->x_ix86_tune_string.
1079 Change ix86_arch_string to access opts->x_ix86_arch_string.
1080 Change ix86_fpmath to access opts->x_ix86_fpmath
1081 Fix call to ix86_option_override_internal.
1082 Fix call to ix86_add_new_builtins.
1083 Fix calls to build_optimization_node and build_target_option_node.
1084 (ix86_valid_target_attribute_p): Remove access to global_options.
1085 Use new gcc_options structure func_options.
1086 Fix call to ix86_valid_target_attribute_tree.
1087 Fix call to build_optimization_node.
1088 (get_builtin_code_for_version): Fix call to
1089 ix86_valid_target_attribute_tree.
1091 2013-10-15 David Malcolm <dmalcolm@redhat.com>
1093 * Makefile.in (PICFLAG): New.
1094 (enable_host_shared): New.
1095 (INTERNAL_CFLAGS): Use PICFLAG.
1096 (LIBIBERTY): Use pic build of libiberty.a if configured with
1097 --enable-host-shared.
1098 * configure.ac: Add --enable-host-shared, setting up new
1100 * configure: Regenerate.
1101 * doc/install.texi (--enable-shared): Add note contrasting it with ...
1102 (--enable-host-shared): New option.
1104 2013-10-15 Richard Biener <rguenther@suse.de>
1106 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1107 for built-in functions.
1109 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1111 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1112 (optimize_range_tests_1): New function,
1113 extracted from optimize_range_tests.
1114 (optimize_range_tests_xor): Similarly.
1115 (optimize_range_tests_diff): New function.
1116 (optimize_range_tests): Use optimize_range_tests_1.
1118 2013-10-15 Cong Hou <congh@google.com>
1120 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1121 requirement of the reduction pattern so that one operand of the
1122 reduction operation can come from outside of the loop.
1124 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1126 * config/arm/neon-schedgen.ml: Remove.
1127 * config/arm/cortex-a9-neon.md: Remove comment regarding
1130 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1132 * config/arm/types: Remove old neon types.
1134 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1136 * config/arm/cortex-a7.md
1137 (cortex_a7_neon_type): New.
1138 (cortex_a7_neon_mul): Update for new types.
1139 (cortex_a7_neon_mla): Likewise.
1140 (cortex_a7_neon): Likewise.
1142 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1144 * config/arm/cortex-a15-neon.md
1145 (cortex_a15_neon_type): New,
1147 (cortex_a15_neon_int_1): Remove.
1148 (cortex_a15_neon_int_2): Likewise.
1149 (cortex_a15_neon_int_3): Likewise.
1150 (cortex_a15_neon_int_4): Likewise.
1151 (cortex_a15_neon_int_5): Likewise.
1152 (cortex_a15_neon_vqneg_vqabs): Likewise.
1153 (cortex_a15_neon_vmov): Likewise.
1154 (cortex_a15_neon_vaba): Likewise.
1155 (cortex_a15_neon_vaba_qqq): Likewise.
1156 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1157 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1158 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1160 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1161 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1162 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1163 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1164 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1165 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1166 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1167 (cortex_a15_neon_shift_1): Likewise.
1168 (cortex_a15_neon_shift_2): Likewise.
1169 (cortex_a15_neon_shift_3): Likewise.
1170 (cortex_a15_neon_vshl_ddd): Likewise.
1171 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1172 (cortex_a15_neon_vsra_vrsra): Likewise.
1173 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1174 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1175 (cortex_a15_neon_bp_3cycle): Likewise.
1176 (cortex_a15_neon_ldm_2): Likewise.
1177 (cortex_a15_neon_stm_2): Likewise.
1178 (cortex_a15_neon_mcr): Likewise.
1179 (cortex_a15_neon_mrc): Likewise.
1180 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1181 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1182 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1183 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1184 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1185 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1186 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1187 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1188 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1189 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1190 (cortex_a15_neon_bp_simple): Likewise.
1191 (cortex_a15_neon_bp_2cycle): Likewise.
1192 (cortex_a15_neon_bp_3cycle): Likewise.
1193 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1194 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1195 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1196 (cortex_a15_neon_vld2_4_regs): Likewise.
1197 (cortex_a15_neon_vld3_vld4): Likewise.
1198 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1199 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1200 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1201 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1202 (cortex_a15_neon_vst3_vst4): Rename to...
1203 (cortex_a15_neon_vst4): This, update for new attributes.
1204 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1205 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1206 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1207 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1208 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1209 (cortex_a15_neon_ldm_2): Likewise.
1210 (cortex_a15_neon_stm_2): Likewise.
1211 (cortex_a15_neon_mcr): Likewise.
1212 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1213 (cortex_a15_neon_mrc): Likewise.
1214 (cortex_a15_neon_mrrc): Likewise.
1216 (cortex_a15_neon_abd): New.
1217 (cortex_a15_neon_abd_q): Likewise.
1218 (cortex_a15_neon_aba): Likewise.
1219 (cortex_a15_neon_aba_q): Likewise.
1220 (cortex_a15_neon_acc): Likewise.
1221 (cortex_a15_neon_acc_q): Likewise.
1222 (cortex_a15_neon_arith_basic): Likewise.
1223 (cortex_a15_neon_arith_complex): Likewise.
1224 (cortex_a15_neon_multiply): Likewise.
1225 (cortex_a15_neon_multiply_q): Likewise.
1226 (cortex_a15_neon_mla): Likewise.
1227 (cortex_a15_neon_mla_q): Likewise.
1228 (cortex_a15_neon_sat_mla_long): Likewise.
1229 (cortex_a15_neon_shift_acc): Likewise.
1230 (cortex_a15_neon_shift_imm_basic): Likewise.
1231 (cortex_a15_neon_shift_imm_complex): Likewise.
1232 (cortex_a15_neon_shift_reg_basic): Likewise.
1233 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1234 (cortex_a15_neon_shift_reg_complex): Likewise.
1235 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1236 (cortex_a15_neon_fp_negabs): Likewise
1237 (cortex_a15_neon_fp_arith): Likewise
1238 (cortex_a15_neon_fp_arith_q): Likewise
1239 (cortex_a15_neon_fp_cvt_int): Likewise
1240 (cortex_a15_neon_fp_cvt_int_q): Likewise
1241 (cortex_a15_neon_fp_cvt_16): Likewise
1242 (cortex_a15_neon_fp_mul): Likewise
1243 (cortex_a15_neon_fp_mul_q): Likewise
1244 (cortex_a15_neon_fp_mla): Likewise
1245 (cortex_a15_neon_fp_mla_q): Likewise
1246 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1247 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1248 (cortex_a15_neon_bitops): Likewise.
1249 (cortex_a15_neon_bitops_q): Likewise.
1250 (cortex_a15_neon_from_gp): Likewise.
1251 (cortex_a15_neon_from_gp_q): Likewise.
1252 (cortex_a15_neon_tbl3_tbl4): Likewise.
1253 (cortex_a15_neon_zip_q): Likewise.
1254 (cortex_a15_neon_to_gp): Likewise.
1255 (cortex_a15_neon_load_a): Likewise.
1256 (cortex_a15_neon_load_b): Likewise.
1257 (cortex_a15_neon_load_c): Likewise.
1258 (cortex_a15_neon_load_d): Likewise.
1259 (cortex_a15_neon_load_e): Likewise.
1260 (cortex_a15_neon_load_f): Likewise.
1261 (cortex_a15_neon_store_a): Likewise.
1262 (cortex_a15_neon_store_b): Likewise.
1263 (cortex_a15_neon_store_c): Likewise.
1264 (cortex_a15_neon_store_d): Likewise.
1265 (cortex_a15_neon_store_e): Likewise.
1266 (cortex_a15_neon_store_f): Likewise.
1267 (cortex_a15_neon_store_g): Likewise.
1268 (cortex_a15_neon_store_h): Likewise.
1269 (cortex_a15_vfp_to_from_gp): Likewise.
1271 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1273 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1275 (cortex_a9_neon_vshl_ddd): Remove.
1276 (cortex_a9_neon_vst3_vst4): Likewise.
1277 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1279 (cortex_a9_neon_bit_ops_q): New.
1281 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1282 (cortex_a9_neon_int_2): Likewise.
1283 (cortex_a9_neon_int_3): Likewise.
1284 (cortex_a9_neon_int_4): Likewise.
1285 (cortex_a9_neon_int_5): Likewise.
1286 (cortex_a9_neon_vqneg_vqabs): Likewise.
1287 (cortex_a9_neon_vmov): Likewise.
1288 (cortex_a9_neon_vaba): Likewise.
1289 (cortex_a9_neon_vaba_qqq): Likewise.
1290 (cortex_a9_neon_shift_1): Likewise.
1291 (cortex_a9_neon_shift_2): Likewise.
1292 (cortex_a9_neon_shift_3): Likewise.
1293 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1294 (cortex_a9_neon_vsra_vrsra): Likewise.
1295 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1296 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1297 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1299 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1300 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1301 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1303 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1304 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1305 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1306 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1307 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1308 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1309 (cortex_a9_neon_fp_vsum): Likewise.
1310 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1311 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1312 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1313 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1314 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1315 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1316 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1317 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1318 (cortex_a9_neon_bp_simple): Likewise.
1319 (cortex_a9_neon_bp_2cycle): Likewise.
1320 (cortex_a9_neon_bp_3cycle): Likewise.
1321 (cortex_a9_neon_ldr): Likewise.
1322 (cortex_a9_neon_str): Likewise.
1323 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1324 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1325 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1326 (cortex_a9_neon_vld2_4_regs): Likewise.
1327 (cortex_a9_neon_vld3_vld4): Likewise.
1328 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1329 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1330 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1331 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1332 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1333 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1334 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1335 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1336 (cortex_a9_neon_mcr): Likewise.
1337 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1338 (cortex_a9_neon_mrc): Likewise.
1339 (cortex_a9_neon_mrrc): Likewise.
1341 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1343 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1345 (cortex_a8_neon_vshl_ddd): Remove.
1346 (cortex_a8_neon_vst3_vst4): Likewise.
1347 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1349 (cortex_a8_neon_bit_ops_q): New.
1351 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1352 (cortex_a8_neon_int_2): Likewise..
1353 (cortex_a8_neon_int_3): Likewise.
1354 (cortex_a8_neon_int_5): Likewise.
1355 (cortex_a8_neon_vqneg_vqabs): Likewise.
1356 (cortex_a8_neon_int_4): Likewise.
1357 (cortex_a8_neon_vaba): Likewise.
1358 (cortex_a8_neon_vaba_qqq): Likewise.
1359 (cortex_a8_neon_shift_1): Likewise.
1360 (cortex_a8_neon_shift_2): Likewise.
1361 (cortex_a8_neon_shift_3): Likewise.
1362 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1363 (cortex_a8_neon_vsra_vrsra): Likewise.
1364 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1365 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1366 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1368 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1369 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1370 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1372 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1373 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1374 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1375 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1376 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1377 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1378 (cortex_a8_neon_fp_vsum): Likewise.
1379 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1380 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1381 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1382 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1383 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1384 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1385 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1386 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1387 (cortex_a8_neon_bp_simple): Likewise.
1388 (cortex_a8_neon_bp_2cycle): Likewise.
1389 (cortex_a8_neon_bp_3cycle): Likewise.
1390 (cortex_a8_neon_ldr): Likewise.
1391 (cortex_a8_neon_str): Likewise.
1392 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1393 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1394 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1395 (cortex_a8_neon_vld2_4_regs): Likewise.
1396 (cortex_a8_neon_vld3_vld4): Likewise.
1397 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1398 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1399 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1400 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1401 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1402 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1403 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1404 (cortex_a8_neon_mcr): Likewise.
1405 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1406 (cortex_a8_neon_mrc): Likewise.
1407 (cortex_a8_neon_mrrc): Likewise.
1409 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1411 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1413 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1414 (aarch64_simd_dup<mode>): Add "type" attribute.
1415 (aarch64_dup_lane<mode>): Likewise.
1416 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1417 (*aarch64_simd_mov<mode>): Likewise.
1418 (aarch64_simd_mov_from_<mode>low): Likewise.
1419 (aarch64_simd_mov_from_<mode>high): Likewise.
1420 (orn<mode>3): Likewise.
1421 (bic<mode>3): Likewise.
1422 (add<mode>3): Likewise.
1423 (sub<mode>3): Likewise.
1424 (mul<mode>3): Likewise.
1425 (*aarch64_mul3_elt<mode>): Likewise.
1426 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1427 (*aarch64_mul3_elt_to_128df): Likewise.
1428 (*aarch64_mul3_elt_to_64v2df): Likewise.
1429 (neg<mode>2): Likewise.
1430 (abs<mode>2): Likewise.
1431 (abd<mode>_3): Likewise.
1432 (aba<mode>_3): Likewise.
1433 (fabd<mode>_3): Likewise.
1434 (*fabd_scalar<mode>3): Likewise.
1435 (and<mode>3): Likewise.
1436 (ior<mode>3): Likewise.
1437 (xor<mode>3): Likewise.
1438 (one_cmpl<mode>2): Likewise.
1439 (aarch64_simd_vec_set<mode>): Likewise.
1440 (aarch64_simd_lshr<mode>): Likewise.
1441 (aarch64_simd_ashr<mode>): Likewise.
1442 (aarch64_simd_imm_shl<mode>): Likewise.
1443 (aarch64_simd_reg_sshl<mode): Likewise.
1444 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1445 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1446 (aarch64_simd_vec_setv2di): Likewise.
1447 (aarch64_simd_vec_set<mode>): Likewise.
1448 (aarch64_mla<mode>): Likewise.
1449 (*aarch64_mla_elt<mode>): Likewise.
1450 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1451 (aarch64_mls<mode>): Likewise.
1452 (*aarch64_mls_elt<mode>): Likewise.
1453 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1454 (<su><maxmin><mode>3): Likewise.
1455 (move_lo_quad_<mode>): Likewise.
1456 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1457 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1458 (vec_pack_trunc_<mode>): Likewise.
1459 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1460 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1461 (*aarch64_<su>mlal_lo<mode>): Likewise.
1462 (*aarch64_<su>mlal_hi<mode>): Likewise.
1463 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1464 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1465 (*aarch64_<su>mlal<mode>): Likewise.
1466 (*aarch64_<su>mlsl<mode>): Likewise.
1467 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1468 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1469 (add<mode>3): Likewise.
1470 (sub<mode>3): Likewise.
1471 (mul<mode>3): Likewise.
1472 (div<mode>3): Likewise.
1473 (neg<mode>2): Likewise.
1474 (abs<mode>2): Likewise.
1475 (fma<mode>4): Likewise.
1476 (*aarch64_fma4_elt<mode>): Likewise.
1477 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1478 (*aarch64_fma4_elt_to_128df): Likewise.
1479 (*aarch64_fma4_elt_to_64v2df): Likewise.
1480 (fnma<mode>4): Likewise.
1481 (*aarch64_fnma4_elt<mode>): Likewise.
1482 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1483 (*aarch64_fnma4_elt_to_128df): Likewise.
1484 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1485 (<frint_pattern><mode>2): Likewise.
1486 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1487 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1488 (vec_unpacks_lo_v4sf): Likewise.
1489 (aarch64_float_extend_lo_v2df): Likewise.
1490 (vec_unpacks_hi_v4sf): Likewise.
1491 (aarch64_float_truncate_lo_v2sf): Likewise.
1492 (aarch64_float_truncate_hi_v4sf): Likewise.
1493 (aarch64_vmls<mode>): Likewise.
1494 (<su><maxmin><mode>3): Likewise.
1495 (<maxmin_uns><mode>3): Likewise.
1496 (reduc_<sur>plus_<mode>): Likewise.
1497 (reduc_<sur>plus_v2di): Likewise.
1498 (reduc_<sur>plus_v2si): Likewise.
1499 (reduc_<sur>plus_<mode>): Likewise.
1500 (aarch64_addpv4sf): Likewise.
1501 (clz<mode>2): Likewise.
1502 (reduc_<maxmin_uns>_<mode>): Likewise.
1503 (reduc_<maxmin_uns>_v2di): Likewise.
1504 (reduc_<maxmin_uns>_v2si): Likewise.
1505 (reduc_<maxmin_uns>_<mode>): Likewise.
1506 (reduc_<maxmin_uns>_v4sf): Likewise.
1507 (aarch64_simd_bsl<mode>_internal): Likewise.
1508 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1509 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1510 (aarch64_get_lane<mode>): Likewise.
1511 (*aarch64_combinez<mode>): Likewise.
1512 (aarch64_combine<mode>): Likewise.
1513 (aarch64_simd_combine<mode>): Likewise.
1514 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1515 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1516 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1517 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1518 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1519 (aarch64_<sur>h<addsub><mode>): Likewise.
1520 (aarch64_<sur><addsub>hn<mode>): Likewise.
1521 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1522 (aarch64_pmul<mode>): Likewise.
1523 (aarch64_<su_optab><optab><mode>): Likewise.
1524 (aarch64_<sur>qadd<mode>): Likewise.
1525 (aarch64_sqmovun<mode>): Likewise.
1526 (aarch64_<sur>qmovn<mode>): Likewise.
1527 (aarch64_s<optab><mode>): Likewise.
1528 (aarch64_sq<r>dmulh<mode>): Likewise.
1529 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1530 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1531 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1532 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1533 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1534 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1535 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1536 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1537 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1538 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1539 (aarch64_sqdmull<mode>): Likewise.
1540 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1541 (aarch64_sqdmull_n<mode>): Likewise.
1542 (aarch64_sqdmull2<mode>_internal): Likewise.
1543 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1544 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1545 (aarch64_<sur>shl<mode>): Likewise.
1546 (aarch64_<sur>q<r>shl<mode>
1547 (aarch64_<sur>shll_n<mode>): Likewise.
1548 (aarch64_<sur>shll2_n<mode>): Likewise.
1549 (aarch64_<sur>shr_n<mode>): Likewise.
1550 (aarch64_<sur>sra_n<mode>): Likewise.
1551 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1552 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1553 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1554 (aarch64_cm<optab><mode>): Likewise.
1555 (aarch64_cm<optab>di): Likewise.
1556 (aarch64_cm<optab><mode>): Likewise.
1557 (aarch64_cm<optab>di): Likewise.
1558 (aarch64_cmtst<mode>): Likewise.
1559 (aarch64_cmtstdi): Likewise.
1560 (aarch64_cm<optab><mode>): Likewise.
1561 (*aarch64_fac<optab><mode>): Likewise.
1562 (aarch64_addp<mode>): Likewise.
1563 (aarch64_addpdi): Likewise.
1564 (sqrt<mode>2): Likewise.
1565 (vec_load_lanesoi<mode>): Likewise.
1566 (vec_store_lanesoi<mode>): Likewise.
1567 (vec_load_lanesci<mode>): Likewise.
1568 (vec_store_lanesci<mode>): Likewise.
1569 (vec_load_lanesxi<mode>): Likewise.
1570 (vec_store_lanesxi<mode>): Likewise.
1571 (*aarch64_mov<mode>): Likewise.
1572 (aarch64_ld2<mode>_dreg): Likewise.
1573 (aarch64_ld2<mode>_dreg): Likewise.
1574 (aarch64_ld3<mode>_dreg): Likewise.
1575 (aarch64_ld3<mode>_dreg): Likewise.
1576 (aarch64_ld4<mode>_dreg): Likewise.
1577 (aarch64_ld4<mode>_dreg): Likewise.
1578 (aarch64_tbl1<mode>): Likewise.
1579 (aarch64_tbl2v16qi): Likewise.
1580 (aarch64_combinev16qi): Likewise.
1581 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1582 (aarch64_st2<mode>_dreg): Likewise.
1583 (aarch64_st2<mode>_dreg): Likewise.
1584 (aarch64_st3<mode>_dreg): Likewise.
1585 (aarch64_st3<mode>_dreg): Likewise.
1586 (aarch64_st4<mode>_dreg): Likewise.
1587 (aarch64_st4<mode>_dreg): Likewise.
1588 (*aarch64_simd_ld1r<mode>): Likewise.
1589 (aarch64_frecpe<mode>): Likewise.
1590 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1591 (aarch64_frecps<mode>): Likewise.
1593 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1595 * config/arm/iterators.md (V_elem_ch): New.
1597 (VQH_type): Likewise.
1598 * config/arm/arm.md (is_neon_type): New.
1599 (conds): Use is_neon_type.
1600 (anddi3_insn): Update type attribute.
1601 (xordi3_insn): Likewise.
1602 (one_cmpldi2): Likewise.
1603 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1604 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1605 (*movmisalign<mode>_neon_store): Likewise.
1606 (*movmisalign<mode>_neon_load): Likewise.
1607 (vec_set<mode>_internal): Likewise.
1608 (vec_set<mode>_internal): Likewise.
1609 (vec_setv2di_internal): Likewise.
1610 (vec_extract<mode>): Likewise.
1611 (vec_extract<mode>): Likewise.
1612 (vec_extractv2di): Likewise.
1613 (*add<mode>3_neon): Likewise.
1614 (adddi3_neon): Likewise.
1615 (*sub<mode>3_neon): Likewise.
1616 (subdi3_neon): Likewise.
1617 (fma<VCVTF:mode>4): Likewise.
1618 (fma<VCVTF:mode>4_intrinsic): Likewise.
1619 (*fmsub<VCVTF:mode>4): Likewise.
1620 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1621 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1622 (ior<mode>3): Likewise.
1623 (and<mode>3): Likewise.
1624 (orn<mode>3_neon): Likewise.
1625 (orndi3_neon): Likewise.
1626 (bic<mode>3_neon): Likewise.
1627 (bicdi3_neon): Likewise.
1628 (xor<mode>3): Likewise.
1629 (one_cmpl<mode>2): Likewise.
1630 (abs<mode>2): Likewise.
1631 (neg<mode>2): Likewise.
1632 (negdi2_neon): Likewise.
1633 (*umin<mode>3_neon): Likewise.
1634 (*umax<mode>3_neon): Likewise.
1635 (*smin<mode>3_neon): Likewise.
1636 (*smax<mode>3_neon): Likewise.
1637 (vashl<mode>3): Likewise.
1638 (vashr<mode>3_imm): Likewise.
1639 (vlshr<mode>3_imm): Likewise.
1640 (ashl<mode>3_signed): Likewise.
1641 (ashl<mode>3_unsigned): Likewise.
1642 (neon_load_count): Likewise.
1643 (ashldi3_neon_noclobber): Likewise.
1644 (ashldi3_neon): Likewise.
1645 (signed_shift_di3_neon): Likewise.
1646 (unsigned_shift_di3_neon): Likewise.
1647 (ashrdi3_neon_imm_noclobber): Likewise.
1648 (lshrdi3_neon_imm_noclobber): Likewise.
1649 (<shift>di3_neon): Likewise.
1650 (widen_ssum<mode>3): Likewise.
1651 (widen_usum<mode>3): Likewise.
1652 (quad_halves_<code>v4si): Likewise.
1653 (quad_halves_<code>v4sf): Likewise.
1654 (quad_halves_<code>v8hi): Likewise.
1655 (quad_halves_<code>v16qi): Likewise.
1656 (reduc_splus_v2di): Likewise.
1657 (neon_vpadd_internal<mode>): Likewise.
1658 (neon_vpsmin<mode>): Likewise.
1659 (neon_vpsmax<mode>): Likewise.
1660 (neon_vpumin<mode>): Likewise.
1661 (neon_vpumax<mode>): Likewise.
1662 (*ss_add<mode>_neon): Likewise.
1663 (*us_add<mode>_neon): Likewise.
1664 (*ss_sub<mode>_neon): Likewise.
1665 (*us_sub<mode>_neon): Likewise.
1666 (neon_vadd<mode>_unspec): Likewise.
1667 (neon_vaddl<mode>): Likewise.
1668 (neon_vaddw<mode>): Likewise.
1669 (neon_vhadd<mode>): Likewise.
1670 (neon_vqadd<mode>): Likewise.
1671 (neon_vaddhn<mode>): Likewise.
1672 (neon_vmul<mode>): Likewise.
1673 (neon_vfms<VCVTF:mode>): Likewise.
1674 (neon_vmlal<mode>): Likewise.
1675 (neon_vmls<mode>): Likewise.
1676 (neon_vmlsl<mode>): Likewise.
1677 (neon_vqdmulh<mode>): Likewise.
1678 (neon_vqdmlal<mode>): Likewise.
1679 (neon_vqdmlsl<mode>): Likewise.
1680 (neon_vmull<mode>): Likewise.
1681 (neon_vqdmull<mode>): Likewise.
1682 (neon_vsub<mode>_unspec): Likewise.
1683 (neon_vsubl<mode>): Likewise.
1684 (neon_vsubw<mode>): Likewise.
1685 (neon_vqsub<mode>): Likewise.
1686 (neon_vhsub<mode>): Likewise.
1687 (neon_vsubhn<mode>): Likewise.
1688 (neon_vceq<mode>): Likewise.
1689 (neon_vcge<mode>): Likewise.
1690 (neon_vcgeu<mode>): Likewise.
1691 (neon_vcgt<mode>): Likewise.
1692 (neon_vcgtu<mode>): Likewise.
1693 (neon_vcle<mode>): Likewise.
1694 (neon_vclt<mode>): Likewise.
1695 (neon_vcage<mode>): Likewise.
1696 (neon_vcagt<mode>): Likewise.
1697 (neon_vtst<mode>): Likewise.
1698 (neon_vabd<mode>): Likewise.
1699 (neon_vabdl<mode>): Likewise.
1700 (neon_vaba<mode>): Likewise.
1701 (neon_vabal<mode>): Likewise.
1702 (neon_vmax<mode>): Likewise.
1703 (neon_vmin<mode>): Likewise.
1704 (neon_vpaddl<mode>): Likewise.
1705 (neon_vpadal<mode>): Likewise.
1706 (neon_vpmax<mode>): Likewise.
1707 (neon_vpmin<mode>): Likewise.
1708 (neon_vrecps<mode>): Likewise.
1709 (neon_vrsqrts<mode>): Likewise.
1710 (neon_vqabs<mode>): Likewise.
1711 (neon_vqneg<mode>): Likewise.
1712 (neon_vcls<mode>): Likewise.
1713 (clz<mode>2): Likewise.
1714 (popcount<mode>2): Likewise.
1715 (neon_vrecpe<mode>): Likewise.
1716 (neon_vrsqrte<mode>): Likewise.
1717 (neon_vget_lane<mode>_sext_internal): Likewise.
1718 (neon_vget_lane<mode>_zext_internal): Likewise.
1719 (neon_vdup_n<mode>): Likewise.
1720 (neon_vdup_n<mode>): Likewise.
1721 (neon_vdup_nv2di): Likewise.
1722 (neon_vdup_lane<mode>_interal): Likewise.
1723 (*neon_vswp<mode>): Likewise.
1724 (neon_vcombine<mode>): Likewise.
1725 (float<mode><V_cvtto>2): Likewise.
1726 (floatuns<mode><V_cvtto>2): Likewise.
1727 (fix_trunc<mode><V_cvtto>2): Likewise.
1728 (fixuns_trunc<mode><V_cvtto>2
1729 (neon_vcvt<mode>): Likewise.
1730 (neon_vcvt<mode>): Likewise.
1731 (neon_vcvtv4sfv4hf): Likewise.
1732 (neon_vcvtv4hfv4sf): Likewise.
1733 (neon_vcvt_n<mode>): Likewise.
1734 (neon_vcvt_n<mode>): Likewise.
1735 (neon_vmovn<mode>): Likewise.
1736 (neon_vqmovn<mode>): Likewise.
1737 (neon_vqmovun<mode>): Likewise.
1738 (neon_vmovl<mode>): Likewise.
1739 (neon_vmul_lane<mode>): Likewise.
1740 (neon_vmul_lane<mode>): Likewise.
1741 (neon_vmull_lane<mode>): Likewise.
1742 (neon_vqdmull_lane<mode>): Likewise.
1743 (neon_vqdmulh_lane<mode>): Likewise.
1744 (neon_vqdmulh_lane<mode>): Likewise.
1745 (neon_vmla_lane<mode>): Likewise.
1746 (neon_vmla_lane<mode>): Likewise.
1747 (neon_vmlal_lane<mode>): Likewise.
1748 (neon_vqdmlal_lane<mode>): Likewise.
1749 (neon_vmls_lane<mode>): Likewise.
1750 (neon_vmls_lane<mode>): Likewise.
1751 (neon_vmlsl_lane<mode>): Likewise.
1752 (neon_vqdmlsl_lane<mode>): Likewise.
1753 (neon_vext<mode>): Likewise.
1754 (neon_vrev64<mode>): Likewise.
1755 (neon_vrev32<mode>): Likewise.
1756 (neon_vrev16<mode>): Likewise.
1757 (neon_vbsl<mode>_internal): Likewise.
1758 (neon_vshl<mode>): Likewise.
1759 (neon_vqshl<mode>): Likewise.
1760 (neon_vshr_n<mode>): Likewise.
1761 (neon_vshrn_n<mode>): Likewise.
1762 (neon_vqshrn_n<mode>): Likewise.
1763 (neon_vqshrun_n<mode>): Likewise.
1764 (neon_vshl_n<mode>): Likewise.
1765 (neon_vqshl_n<mode>): Likewise.
1766 (neon_vqshlu_n<mode>): Likewise.
1767 (neon_vshll_n<mode>): Likewise.
1768 (neon_vsra_n<mode>): Likewise.
1769 (neon_vsri_n<mode>): Likewise.
1770 (neon_vsli_n<mode>): Likewise.
1771 (neon_vtbl1v8qi): Likewise.
1772 (neon_vtbl2v8qi): Likewise.
1773 (neon_vtbl3v8qi): Likewise.
1774 (neon_vtbl4v8qi): Likewise.
1775 (neon_vtbl1v16qi): Likewise.
1776 (neon_vtbl2v16qi): Likewise.
1777 (neon_vcombinev16qi): Likewise.
1778 (neon_vtbx1v8qi): Likewise.
1779 (neon_vtbx2v8qi): Likewise.
1780 (neon_vtbx3v8qi): Likewise.
1781 (neon_vtbx4v8qi): Likewise.
1782 (*neon_vtrn<mode>_insn): Likewise.
1783 (*neon_vzip<mode>_insn): Likewise.
1784 (*neon_vuzp<mode>_insn): Likewise.
1785 (neon_vld1<mode>): Likewise.
1786 (neon_vld1_lane<mode>): Likewise.
1787 (neon_vld1_lane<mode>): Likewise.
1788 (neon_vld1_dup<mode>): Likewise.
1789 (neon_vld1_dup<mode>): Likewise.
1790 (neon_vld1_dupv2di): Likewise.
1791 (neon_vst1<mode>): Likewise.
1792 (neon_vst1_lane<mode>): Likewise.
1793 (neon_vst1_lane<mode>): Likewise.
1794 (neon_vld2<mode>): Likewise.
1795 (neon_vld2<mode>): Likewise.
1796 (neon_vld2_lane<mode>): Likewise.
1797 (neon_vld2_lane<mode>): Likewise.
1798 (neon_vld2_dup<mode>): Likewise.
1799 (neon_vst2<mode>): Likewise.
1800 (neon_vst2<mode>): Likewise.
1801 (neon_vst2_lane<mode>): Likewise.
1802 (neon_vst2_lane<mode>): Likewise.
1803 (neon_vld3<mode>): Likewise.
1804 (neon_vld3qa<mode>): Likewise.
1805 (neon_vld3qb<mode>): Likewise.
1806 (neon_vld3_lane<mode>): Likewise.
1807 (neon_vld3_lane<mode>): Likewise.
1808 (neon_vld3_dup<mode>): Likewise.
1809 (neon_vst3<mode>): Likewise.
1810 (neon_vst3qa<mode>): Likewise.
1811 (neon_vst3qb<mode>): Likewise.
1812 (neon_vst3_lane<mode>): Likewise.
1813 (neon_vst3_lane<mode>): Likewise.
1814 (neon_vld4<mode>): Likewise.
1815 (neon_vld4qa<mode>): Likewise.
1816 (neon_vld4qb<mode>): Likewise.
1817 (neon_vld4_lane<mode>): Likewise.
1818 (neon_vld4_lane<mode>): Likewise.
1819 (neon_vld4_dup<mode>): Likewise.
1820 (neon_vst4<mode>): Likewise.
1821 (neon_vst4qa<mode>): Likewise.
1822 (neon_vst4qb<mode>): Likewise.
1823 (neon_vst4_lane<mode>): Likewise.
1824 (neon_vst4_lane<mode>): Likewise.
1825 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1826 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1827 (neon_vec_<US>mult_lo_<mode>): Likewise.
1828 (neon_vec_<US>mult_hi_<mode>): Likewise.
1829 (neon_vec_<US>shiftl_<mode>): Likewise.
1830 (neon_unpack<US>_<mode>): Likewise.
1831 (neon_vec_<US>mult_<mode>): Likewise.
1832 (vec_pack_trunc_<mode>): Likewise.
1833 (neon_vec_pack_trunc_<mode>): Likewise.
1834 (neon_vabd<mode>_2): Likewise.
1835 (neon_vabd<mode>_3): Likewise.
1837 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1839 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1840 (load_pair): Update type attribute.
1841 (store_pair): Update type attribute.
1842 * config/aarch64/iterators.md (q): New.
1844 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1846 * config/arm/types.md: Add new types for Neon insns.
1848 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1849 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1850 Sergey Lega <sergey.s.lega@intel.com>
1851 Anna Tikhonova <anna.tikhonova@intel.com>
1852 Ilya Tocar <ilya.tocar@intel.com>
1853 Andrey Turetskiy <andrey.turetskiy@intel.com>
1854 Ilya Verbin <ilya.verbin@intel.com>
1855 Kirill Yukhin <kirill.yukhin@intel.com>
1856 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1858 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1859 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1860 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1862 (srcp14<mode>): Ditto.
1863 (rsqrt14<mode>): Ditto.
1864 (rsqrt14<mode>): Ditto.
1865 (avx512f_vmscalef<mode>): Ditto.
1866 (avx512f_scalef<mode>): Ditto.
1867 (avx512f_getexp<mode>): Ditto.
1868 (avx512f_sgetexp<mode>): Ditto.
1869 (avx512f_fixupimm<mode>): Ditto.
1870 (avx512f_sfixupimm<mode>): Ditto.
1871 (avx512f_rndscale<mode>): Ditto.
1872 (*avx512er_exp2<mode>): Ditto.
1873 (*avx512er_rcp28<mode>): Ditto.
1874 (avx512er_rsqrt28<mode>): Ditto.
1875 (avx512f_getmant<mode>): Ditto.
1876 (avx512f_getmant<mode>): Ditto.
1877 (avx512f_rndscale<mode>): Fix formatting.
1879 2013-10-15 Martin Jambor <mjambor@suse.cz>
1881 * ipa-utils.h (ipa_edge_within_scc): Declare.
1882 * ipa-cp.c (edge_within_scc): Moved...
1883 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1885 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1886 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1887 Sergey Lega <sergey.s.lega@intel.com>
1888 Anna Tikhonova <anna.tikhonova@intel.com>
1889 Ilya Tocar <ilya.tocar@intel.com>
1890 Andrey Turetskiy <andrey.turetskiy@intel.com>
1891 Ilya Verbin <ilya.verbin@intel.com>
1892 Kirill Yukhin <kirill.yukhin@intel.com>
1893 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1895 * config/i386/predicates.md (const_8_to_15_operand): New.
1896 (const_16_to_31_operand): Ditto.
1897 * config/i386/sse.md (V8FI): New.
1899 (reduc_splus_v8df): Ditto.
1900 (reduc_splus_v16sf): Ditto.
1901 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1902 (vec_extract_hi_<mode>): Ditto.
1903 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1904 (vec_set_lo_<mode>): Ditto.
1905 (vec_set_hi_<mode>): Ditto.
1906 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1907 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1908 (avx512f_pshufd_1): Ditto.
1909 (avx512f_broadcast<mode>): Ditto.
1910 (avx512f_broadcast<mode>): Ditto.
1911 (define_split): Split vec_extract_lo into move.
1912 (ssequartermode): Ditto.
1913 (ssedoublemode): Extened with wider modes.
1914 (vec_extract_lo_<mode>): Ditto.
1916 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1917 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1918 Sergey Lega <sergey.s.lega@intel.com>
1919 Anna Tikhonova <anna.tikhonova@intel.com>
1920 Ilya Tocar <ilya.tocar@intel.com>
1921 Andrey Turetskiy <andrey.turetskiy@intel.com>
1922 Ilya Verbin <ilya.verbin@intel.com>
1923 Kirill Yukhin <kirill.yukhin@intel.com>
1924 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1926 * config/i386/predicates.md (register_or_constm1_operand): New.
1927 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1928 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1929 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1930 UNSPEC_SCATTER_PREFETCH
1932 (avx512f_ucmp<mode>3): Ditto.
1933 (avx512f_vternlog<mode>): Ditto.
1934 (avx512f_align<mode>): Ditto.
1935 (<shift_insn><mode>3): Ditto.
1936 (avx512f_<rotate>v<mode>): Ditto.
1937 (avx512f_<rotate><mode>): Ditto.
1938 (avx512f_eq<mode>3): Ditto.
1939 (avx512f_eq<mode>3_1): Ditto.
1940 (avx512f_gt<mode>3): Ditto.
1941 (avx512f_testm<mode>3): Ditto.
1942 (avx512f_testnm<mode>3): Ditto.
1943 (avx512pf_gatherpf<mode>): Ditto.
1944 (*avx512pf_gatherpf<mode>_mask): Ditto.
1945 (*avx512pf_gatherpf<mode>): Ditto.
1946 (avx512pf_scatterpf<mode>): Ditto.
1947 (*avx512pf_scatterpf<mode>_mask): Ditto.
1948 (*avx512pf_scatterpf<mode>): Ditto.
1949 (avx512f_vec_dup_gpr<mode>): Ditto.
1950 (clz<mode>2): Ditto.
1951 (conflict<mode>): Ditto.
1952 (REDUC_SMINMAX_MODE): Extened with wider modes.
1953 (reduc_<code>_<mode>): Ditto.
1954 (vlshr<mode>3): Ditto.
1955 (vashl<mode>3): Ditto.
1957 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1958 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1959 Sergey Lega <sergey.s.lega@intel.com>
1960 Anna Tikhonova <anna.tikhonova@intel.com>
1961 Ilya Tocar <ilya.tocar@intel.com>
1962 Andrey Turetskiy <andrey.turetskiy@intel.com>
1963 Ilya Verbin <ilya.verbin@intel.com>
1964 Kirill Yukhin <kirill.yukhin@intel.com>
1965 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1967 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1970 (avx512fmaskmode): Ditto.
1971 (bcstscalarsuff): Ditto.
1972 (avx512f_blendm<mode>): Ditto.
1973 (cmp_imm_predicate): Ditto.
1974 (avx512f_cmp<mode>3): Ditto.
1975 (avx512f_vec_dup<mode>): Ditto.
1976 (avx512f_vec_dup_mem<mode>): Ditto.
1977 (avx512f_vpermi2var<mode>3): Ditto.
1978 (avx512f_vpermt2var<mode>3): Ditto.
1979 (vec_init<mode>): Ditto.
1980 (avx512f_gathersi<mode>): Ditto.
1981 (*avx512f_gathersi<mode>): Ditto.
1982 (*avx512f_gathersi<mode>_2): Ditto.
1983 (avx512f_gatherdi<mode>): Ditto.
1984 (*avx512f_gatherdi<mode>): Ditto.
1985 (*avx512f_gatherdi<mode>_2): Ditto.
1986 (avx512f_scattersi<mode>): Ditto.
1987 (*avx512f_scattersi<mode>): Ditto.
1988 (avx512f_scatterdi<mode>): Ditto.
1989 (*avx512f_scatterdi<mode>): Ditto.
1990 (sseintprefix): Extened with wider modes.
1991 (VEC_GATHER_IDXSI): Ditto.
1992 (VEC_GATHER_IDXDI): Ditto.
1993 (VEC_GATHER_SRCDI): Ditto.
1995 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
1996 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1998 * config/arm/t-aprofile: New file.
1999 * config.gcc: Handle --with-multilib-list option.
2001 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
2003 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
2006 2013-10-15 Richard Biener <rguenther@suse.de>
2008 * tree-loop-distribution.c (build_empty_rdg): Inline into
2010 (rdg_flag_vertex): Inline into single user.
2011 (rdg_flag_vertex_and_dependent): Likewise.
2012 (build_rdg_partition_for_vertex): Remove processed bitmap.
2013 (rdg_build_partitions): Simplify.
2015 2013-10-15 Richard Biener <rguenther@suse.de>
2017 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
2018 Restructure forwarding through conversions and copies to
2019 avoid performing copy-propagation the wrong way. Adjust
2020 recursion invocations.
2021 (forward_propagate_addr_expr): Add argument stating if we
2022 are recursing from a single-use.
2023 (ssa_forward_propagate_and_combine): Adjust.
2025 2013-10-14 David Malcolm <dmalcolm@redhat.com>
2027 * dumpfile.h (gcc::dump_manager): New class, to hold state
2028 relating to dumpfile management.
2029 (get_dump_file_name): Remove in favor of method of dump_manager.
2030 (dump_initialized_p): Likewise.
2031 (dump_start): Likewise.
2032 (dump_finish): Likewise.
2033 (dump_switch_p): Likewise.
2034 (dump_register): Likewise.
2035 (get_dump_file_info): Likewise.
2036 * context.c (gcc::context::context): Construct the dump_manager
2038 * context.h (gcc::context::get_dumps): New.
2039 (gcc::context::m_dumps): New.
2040 * coverage.c (coverage_init): Port to dump_manager API.
2041 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
2042 (extra_dump_files_in_use): Likewise.
2043 (extra_dump_files_alloced): Likewise.
2044 (gcc::dump_manager::dump_manager): New.
2045 (dump_register): Convert to...
2046 (gcc::dump_manager::dump_register): ...method, replacing
2047 function-static next_dump with m_next_dump field.
2048 (get_dump_file_info): Convert to...
2049 (gcc::dump_manager::get_dump_file_info): ...method.
2050 (get_dump_file_name): Convert to...
2051 (gcc::dump_manager::get_dump_file_name): ...method.
2052 (dump_start): Convert to...
2053 (gcc::dump_manager::dump_start): ...method.
2054 (dump_finish): Convert to...
2055 (gcc::dump_manager::dump_finish): ...method.
2056 (dump_begin): Replace body with...
2057 (gcc::dump_manager::dump_begin): ...new method.
2058 (dump_phase_enabled_p): Convert to...
2059 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2060 (dump_phase_enabled_p): Convert to...
2061 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2062 (dump_initialized_p): Convert to...
2063 (gcc::dump_manager::dump_initialized_p): ...method.
2064 (dump_flag_name): Replace body with...
2065 (gcc::dump_manager::dump_flag_name): ...new method.
2066 (dump_enable_all): Convert to...
2067 (gcc::dump_manager::dump_enable_all): ...new method.
2068 (opt_info_enable_passes): Convert to...
2069 (gcc::dump_manager::opt_info_enable_passes): ...new method.
2070 (dump_switch_p_1): Convert to...
2071 (gcc::dump_manager::dump_switch_p_1): ...new method.
2072 (dump_switch_p): Convert to...
2073 (gcc::dump_manager::dump_switch_p): ...new method.
2074 (opt_info_switch_p): Port to dump_manager API.
2075 (enable_rtl_dump_file): Likewise.
2076 * opts-global.c (handle_common_deferred_options): Port to new
2078 * passes.c (pass_manager::finish_optimization_passes): Likewise.
2079 (pass_manager::register_one_dump_file): Likewise.
2080 (pass_manager::register_pass): Likewise.
2081 (pass_init_dump_file): Likewise.
2082 (pass_fini_dump_file): Likewise.
2083 * statistics.c (statistics_early_init): Likewise.
2085 2013-10-14 Richard Biener <rguenther@suse.de>
2087 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
2088 iterative_hash_canonical_type, gimple_canonical_type_hash,
2089 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
2090 gimple_register_canonical_type, print_gimple_types_stats,
2091 free_gimple_type_tables): Move to lto/lto.c
2092 (gt-gimple.h): Do not include.
2093 * gimple.h (gimple_register_canonical_type,
2094 print_gimple_types_stats, free_gimple_type_tables): Remove.
2095 * Makefile.in (GTFILES): Remove gimple.c.
2097 2013-10-14 Travis Snoozy <quandary@remstate.com>
2100 * config/msp430/msp430.c (msp430_option_override): Correct thinko
2101 scanning for msp430x targets.
2103 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
2106 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2107 (registers_ok_for_ldd_peep): Move around.
2108 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2109 * config/sparc/sparc.md (widening peepholes): Use it.
2111 2013-10-14 Richard Biener <rguenther@suse.de>
2115 * gimple.c (iterative_hash_canonical_type): Make sure to
2116 record the hash into the correct hashtable slot.
2118 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2120 PR rtl-optimization/58662
2121 * combine.c (try_combine): Take into account death nodes on I2 when
2122 splitting a PARALLEL of two independent SETs. Fix dump message.
2124 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2127 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2128 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2129 * config/sh/t-sh (sh_treg_combine.o): New entry.
2130 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2131 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2132 (register_sh_passes): New function. Register sh_treg_combine pass.
2133 (sh_option_override): Invoke it.
2134 (sh_canonicalize_comparison): Handle op0_preserve_value.
2135 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2136 opportunities. Canonicalize branch condition.
2137 (nott): Allow only if pseudos can be created for non-SH2A.
2139 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2142 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2143 (ix86_expand_movmem): Replace copy_addr_to_reg with
2144 ix86_copy_addr_to_reg.
2145 (ix86_expand_setmem): Likewise.
2147 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2149 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2150 provided by ix86_fp_compare_mode instead of CCFPUmode.
2152 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2154 * config/aarch64/arm_neon.h
2155 (vtbx<1,3>_<psu>8): Fix register constriants.
2157 2013-10-11 Jeff Law <law@redhat.com>
2159 PR tree-optimization/58640
2160 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2161 threading paths that cross over two loop entry points.
2163 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2165 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2166 handle vector float as well.
2167 (*vsx_le_perm_load_v4si): Likewise.
2168 (*vsx_le_perm_store_v2di): Likewise.
2169 (*vsx_le_perm_store_v4si): Likewise.
2171 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2173 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2174 directly to circumvent subtract from splat{31} workaround.
2175 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2177 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2178 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2179 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2180 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2181 separate big and little endian logic.
2182 (*altivec_vperm_<mode>_internal): New define_insn.
2183 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2184 separate big and little endian logic.
2185 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2186 (vec_permv16qi): Add little endian logic.
2188 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2190 * doc/extend.texi (returns_nonnull): Remove arguments.
2192 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2193 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2194 Sergey Lega <sergey.s.lega@intel.com>
2195 Anna Tikhonova <anna.tikhonova@intel.com>
2196 Ilya Tocar <ilya.tocar@intel.com>
2197 Andrey Turetskiy <andrey.turetskiy@intel.com>
2198 Ilya Verbin <ilya.verbin@intel.com>
2199 Kirill Yukhin <kirill.yukhin@intel.com>
2200 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2202 * config/i386/sse.md (VI48F_256_512): New.
2203 (avx2_permvar<mode>): Change to ...
2204 (<avx2_avx512f>_permvar<mode>): This.
2206 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2207 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2208 Sergey Lega <sergey.s.lega@intel.com>
2209 Anna Tikhonova <anna.tikhonova@intel.com>
2210 Ilya Tocar <ilya.tocar@intel.com>
2211 Andrey Turetskiy <andrey.turetskiy@intel.com>
2212 Ilya Verbin <ilya.verbin@intel.com>
2213 Kirill Yukhin <kirill.yukhin@intel.com>
2214 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2216 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2217 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2218 * config/i386/sse.md (VI4_AVX): New.
2219 (sf2simodelower): Ditto.
2220 (sse2_cvtps2dq): Change to ...
2221 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2223 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2224 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2225 Sergey Lega <sergey.s.lega@intel.com>
2226 Anna Tikhonova <anna.tikhonova@intel.com>
2227 Ilya Tocar <ilya.tocar@intel.com>
2228 Andrey Turetskiy <andrey.turetskiy@intel.com>
2229 Ilya Verbin <ilya.verbin@intel.com>
2230 Kirill Yukhin <kirill.yukhin@intel.com>
2231 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2233 * config/i386/sse.md (V_512): New.
2235 (vcond<V_512:mode><VF_512:mode>): Ditto.
2236 (vcond<V_512:mode><VI_512:mode>): Ditto.
2237 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2239 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2240 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2241 Sergey Lega <sergey.s.lega@intel.com>
2242 Anna Tikhonova <anna.tikhonova@intel.com>
2243 Ilya Tocar <ilya.tocar@intel.com>
2244 Andrey Turetskiy <andrey.turetskiy@intel.com>
2245 Ilya Verbin <ilya.verbin@intel.com>
2246 Kirill Yukhin <kirill.yukhin@intel.com>
2247 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2249 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2250 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2252 (fma<mode>4): Removed condition.
2253 (fms<mode>4): Ditto.
2254 (fnma<mode>4): Ditto.
2255 (fnms<mode>4): Ditto.
2256 (fma4i_fmadd_<mode>): Ditto.
2257 (*fma_fmadd_<mode>): Ditto.
2258 (*fma_fmsub_<mode>): Ditto.
2259 (*fma_fnmadd_<mode>): Ditto.
2260 (*fma_fnmsub_<mode>): Ditto.
2261 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2262 (*fma_fmaddsub_<mode>): Ditto.
2263 (*fma_fmsubadd_<mode>): Ditto.
2264 (*fmai_fmadd_<mode>): Ditto.
2265 (*fmai_fmsub_<mode>): Ditto.
2266 (*fmai_fnmadd_<mode>): Ditto.
2267 (*fmai_fnmsub_<mode>): Ditto.
2269 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2270 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2271 Sergey Lega <sergey.s.lega@intel.com>
2272 Anna Tikhonova <anna.tikhonova@intel.com>
2273 Ilya Tocar <ilya.tocar@intel.com>
2274 Andrey Turetskiy <andrey.turetskiy@intel.com>
2275 Ilya Verbin <ilya.verbin@intel.com>
2276 Kirill Yukhin <kirill.yukhin@intel.com>
2277 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2279 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2280 (VI124_256): Changed to ...
2281 (VI124_256_48_512): This.
2282 (ssepackmode): Extended with wider modes.
2283 (<code><mode>3): Changed iterator.
2284 (*avx2_<code><mode>3): Ditto.
2285 (vec_pack_trunc_<mode>): Ditto.
2287 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2288 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2289 Sergey Lega <sergey.s.lega@intel.com>
2290 Anna Tikhonova <anna.tikhonova@intel.com>
2291 Ilya Tocar <ilya.tocar@intel.com>
2292 Andrey Turetskiy <andrey.turetskiy@intel.com>
2293 Ilya Verbin <ilya.verbin@intel.com>
2294 Kirill Yukhin <kirill.yukhin@intel.com>
2295 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2297 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2298 (VI8F_256_512): Ditto.
2299 (abs<mode>2): Changed iterator.
2300 (avx2_perm<mode>): Changed to ...
2301 (<avx2_avx512f>_perm<mode>): This.
2302 (avx2_perm<mode>_1): Changed to ...
2303 (<avx2_avx512f>_perm<mode>_1): This.
2305 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2306 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2307 Sergey Lega <sergey.s.lega@intel.com>
2308 Anna Tikhonova <anna.tikhonova@intel.com>
2309 Ilya Tocar <ilya.tocar@intel.com>
2310 Andrey Turetskiy <andrey.turetskiy@intel.com>
2311 Ilya Verbin <ilya.verbin@intel.com>
2312 Kirill Yukhin <kirill.yukhin@intel.com>
2313 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2315 * config/i386/sse.md (VI48_AVX512F): New.
2316 (VI48_AVX2): Changed to ...
2317 (VI48_AVX2_48_AVX512F): This.
2318 (avx2_ashrv<mode>): Changed to ...
2319 (<avx2_avx512f>_ashrv<mode>): This.
2320 (avx2_<shift_insn>v<mode>): Changed to ...
2321 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2323 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2324 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2325 Sergey Lega <sergey.s.lega@intel.com>
2326 Anna Tikhonova <anna.tikhonova@intel.com>
2327 Ilya Tocar <ilya.tocar@intel.com>
2328 Andrey Turetskiy <andrey.turetskiy@intel.com>
2329 Ilya Verbin <ilya.verbin@intel.com>
2330 Kirill Yukhin <kirill.yukhin@intel.com>
2331 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2333 * config/i386/sse.md (VI4_AVX512F): New.
2334 (VI8_AVX2_AVX512F): Ditto.
2335 (mul<mode>3): Extended with wider modes.
2336 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2337 (mul<mode>3): Ditto.
2338 (vec_widen_<s>mult_odd_<mode>): Ditto.
2340 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2341 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2342 Sergey Lega <sergey.s.lega@intel.com>
2343 Anna Tikhonova <anna.tikhonova@intel.com>
2344 Ilya Tocar <ilya.tocar@intel.com>
2345 Andrey Turetskiy <andrey.turetskiy@intel.com>
2346 Ilya Verbin <ilya.verbin@intel.com>
2347 Kirill Yukhin <kirill.yukhin@intel.com>
2348 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2350 * config/i386/sse.md (VI2_AVX512F): New.
2351 (VI124_AVX512F): Ditto.
2352 (sseunpackmode): Extended with wider modes.
2353 (sseunpackfltmode): Ditto.
2354 (vec_unpacks_float_hi_<mode>): Ditto.
2355 (vec_unpacks_float_lo_<mode>): Ditto.
2356 (vec_unpacku_float_hi_<mode>): Ditto.
2357 (vec_unpacku_float_lo_<mode>): Ditto.
2358 (vec_unpacks_lo_<mode>): Ditto.
2359 (vec_unpacks_hi_<mode>): Ditto.
2360 (vec_unpacku_lo_<mode>): Ditto.
2361 (vec_unpacku_hi_<mode>): Ditto.
2363 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2364 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2365 Sergey Lega <sergey.s.lega@intel.com>
2366 Anna Tikhonova <anna.tikhonova@intel.com>
2367 Ilya Tocar <ilya.tocar@intel.com>
2368 Andrey Turetskiy <andrey.turetskiy@intel.com>
2369 Ilya Verbin <ilya.verbin@intel.com>
2370 Kirill Yukhin <kirill.yukhin@intel.com>
2371 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2373 * config/i386/i386.md (multdiv): New.
2374 (multdiv_mnemonic): Ditto.
2375 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2376 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2377 (<sse>_vmdiv<mode>3): Removed.
2379 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2380 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2381 Sergey Lega <sergey.s.lega@intel.com>
2382 Anna Tikhonova <anna.tikhonova@intel.com>
2383 Ilya Tocar <ilya.tocar@intel.com>
2384 Andrey Turetskiy <andrey.turetskiy@intel.com>
2385 Ilya Verbin <ilya.verbin@intel.com>
2386 Kirill Yukhin <kirill.yukhin@intel.com>
2387 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2389 * config/i386/sse.md (V): Extended with wider modes.
2391 (ssehalfvecmode): Ditto.
2393 (ssepackfltmode): Ditto.
2394 (avx_vec_concat<mode>): Ditto.
2395 (V_256_512): New iterator.
2396 (VF2_512_256): Ditto.
2397 (si2dfmode): New attribute.
2398 (si2dfmodelower): Ditto.
2400 (concat_tg_mode): Ditto.
2401 (floatv4siv4df2): Changed to ...
2402 (float<si2dfmodelower><mode>2): This.
2403 (avx_cvtps2pd256): Changed to ...
2404 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2405 (vec_pack_trunc_v4df): Changed to ...
2406 (vec_pack_trunc_<mode>): This.
2407 (avx_vpermil<mode>): Changed to ...
2408 (<sse2_avx_avx512f>_vpermil<mode>): This.
2409 (<fixsuffix>fix_truncv8dfv8si2): New.
2410 (vec_pack_sfix_trunc_v8df): Ditto.
2411 (avx512f_rndscale<mode>): Ditto.
2412 (avx512f_roundpd512): Ditto.
2413 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2415 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2416 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2417 Sergey Lega <sergey.s.lega@intel.com>
2418 Anna Tikhonova <anna.tikhonova@intel.com>
2419 Ilya Tocar <ilya.tocar@intel.com>
2420 Andrey Turetskiy <andrey.turetskiy@intel.com>
2421 Ilya Verbin <ilya.verbin@intel.com>
2422 Kirill Yukhin <kirill.yukhin@intel.com>
2423 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2425 * config/i386/i386.md (any_fix): New iterator.
2426 (fixsuffix): New attribute.
2427 * config/i386/sse.md (VF1): Extened with wider modes.
2431 (sseintvecmodelower): Ditto.
2432 (ssescalarmode): Ditto.
2433 (ssescalarnum): Ditto.
2435 (ssexmmmode): Ditto.
2436 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2437 (<sse>_rcp<mode>2): Change iterator.
2438 (rsqrt<mode>2): Ditto.
2439 (<sse>_rsqrt<mode>2): Ditto.
2440 (avx2_vec_dup<mode>): Ditto.
2441 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2442 (round<mode>2_sfix): Ditto.
2443 (avx2_pbroadcast<mode>): Ditto.
2444 (*andnot<mode>3): Handle XI mode.
2445 (*<code><mode>3): Ditto.
2446 (AVXTOSSEMODE): Removed.
2447 (avx_vpermil<mode>): Changed to ...
2448 (<sse2_avx_avx512f>_vpermil<mode>): This.
2450 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2451 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2452 Sergey Lega <sergey.s.lega@intel.com>
2453 Anna Tikhonova <anna.tikhonova@intel.com>
2454 Ilya Tocar <ilya.tocar@intel.com>
2455 Andrey Turetskiy <andrey.turetskiy@intel.com>
2456 Ilya Verbin <ilya.verbin@intel.com>
2457 Kirill Yukhin <kirill.yukhin@intel.com>
2458 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2460 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2461 (<sse>_comi): Ditto.
2462 (<sse>_ucomi): Ditto.
2463 (sse_cvtss2siq_2): Ditto.
2464 (sse2_cvtsd2si): Ditto.
2465 (sse2_cvtsd2siq): Ditto.
2466 (sse2_cvttsd2si): Ditto.
2467 (sse2_cvttsd2siq): Ditto.
2468 (<shift_insn><mode>3): Ditto.
2469 (sse2_cvtsi2sdq): Update constraint and prefix.
2470 (sse_cvtsi2ss): Update prefix.
2471 (sse_cvtsi2ssq): Ditto.
2473 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2475 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2476 ignore internal calls.
2478 2013-10-11 Richard Biener <rguenther@suse.de>
2480 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2481 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2482 TREE_OVERFLOW is set.
2484 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2486 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2488 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2489 * gimple.h: Likewise.
2491 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2493 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2496 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2498 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2499 * configure: Regenerate.
2501 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2503 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2504 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2505 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2506 similarly pp_right_paren.
2507 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2508 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2509 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2510 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2512 (omp_declare_simd_clauses_equal,
2513 omp_remove_redundant_declare_simd_attrs): New functions.
2514 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2515 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2516 * tree.h (OMP_LOOP_CHECK): Define.
2517 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2518 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2519 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2520 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2521 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2522 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2523 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2524 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2525 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2526 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2527 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2528 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2529 (omp_remove_redundant_declare_simd_attrs): New prototype.
2530 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2531 GIMPLE_OMP_TEAMS): New codes.
2532 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2533 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2535 (target_nesting_level): New variable.
2536 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2537 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2538 for collapse > 1 static schedule unless ordered.
2539 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2540 (determine_parallel_type): Adjust get_ws_args_for caller.
2541 (install_var_field): Handle mask & 4 for double indirection.
2542 (scan_sharing_clauses): Ignore shared clause on teams construct.
2543 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2544 (create_omp_child_function): If inside target or declare target
2545 constructs, set "omp declare target" attribute on the child function.
2546 (find_combined_for): New function.
2547 (scan_omp_parallel): Handle combined loops.
2548 (scan_omp_target, scan_omp_teams): New functions.
2549 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2550 restrictions and set ctx->cancellable for cancellable constructs.
2551 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2552 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2553 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2554 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2555 (omp_clause_aligned_alignment): New function.
2556 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2557 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2558 on teams constructs. Handle user defined reductions and new
2560 (lower_reduction_clauses): Don't set placeholder to address of ref
2561 if it has already the right type.
2562 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2563 (expand_parallel_call): Use the new non-_start suffixed builtins,
2564 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2565 and GOMP_parallel_end after the call.
2566 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2567 (expand_omp_for_init_counts): Handle combined loops.
2568 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2570 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2571 end of cancellable loops.
2572 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2573 Likewise. Handle collapse > 1 loops.
2574 (expand_omp_simd): Handle combined loops.
2575 (expand_omp_for): Add inner_stmt argument, adjust callers of
2576 expand_omp_for* functions, use expand_omp_for_static*chunk even
2577 for collapse > 1 unless ordered.
2578 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2579 of cancellable sections.
2580 (expand_omp_single): Remove need_barrier variable, just rely on
2581 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2582 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2583 (expand_omp_atomic_load, expand_omp_atomic_store,
2584 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2585 (expand_omp_target): New function.
2586 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2587 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2588 (build_omp_regions_1): Immediately close region for
2589 GF_OMP_TARGET_KIND_UPDATE.
2590 (maybe_add_implicit_barrier_cancel): New function.
2591 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2593 (lower_omp_single): Likewise. Add clobber after the barrier.
2594 (lower_omp_taskgroup): New function.
2595 (lower_omp_for): Handle combined loops. Adjust
2596 lower_rec_input_clauses caller. Handle cancellation.
2597 (lower_depend_clauses): New function.
2598 (lower_omp_taskreg): Lower depend clauses. Adjust
2599 lower_rec_input_clauses caller. Add clobber after the call. Handle
2601 (lower_omp_target, lower_omp_teams): New functions.
2602 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2603 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2604 and GOMP_cancellation_point calls.
2605 (lower_omp): Fold stmts inside of target region.
2606 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2607 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2608 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2609 (BT_FN_VOID_OMPFN_PTR_UINT,
2610 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2611 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2612 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2613 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2614 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2615 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2616 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2617 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2618 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2619 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2620 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2621 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2622 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2623 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2624 BUILT_IN_GOMP_PARALLEL_END.
2625 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2626 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2627 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2628 GF_OMP_FOR_KIND_DISTRIBUTE.
2629 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2630 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2631 (dump_gimple_omp_return): Print lhs if it has any.
2632 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2633 gimple_omp_atomic_seq_cst_p.
2634 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2635 and GIMPLE_OMP_TEAMS.
2636 * langhooks.c (lhd_omp_mappable_type): New function.
2637 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2638 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2640 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2641 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2642 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2643 (struct gimplify_omp_ctx): Add combined_loop field.
2644 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2645 on stmts inside of target region.
2646 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2647 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2648 ORT_TARGET and ORT_TARGET_DATA.
2649 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2651 (omp_notice_threadprivate_variable): Complain about threadprivate
2652 variables in target region.
2653 (omp_notice_variable): Complain about vars with non-mappable type
2654 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2655 (omp_check_private): Ignore ORT_TARGET* regions.
2656 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2657 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2658 (find_combined_omp_for): New function.
2659 (gimplify_omp_for): Handle gimplification of combined loops.
2660 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2662 (gimplify_omp_target_update): New function.
2663 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2664 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2665 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2666 (gimplify_body): If fndecl has "omp declare target" attribute, add
2667 implicit ORT_TARGET context around it.
2668 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2669 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2670 * tree-nested.c (convert_nonlocal_reference_stmt,
2671 convert_local_reference_stmt, convert_gimple_call): Handle
2672 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2673 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2674 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2675 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2676 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2677 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2678 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2679 BUILT_IN_GOMP_LOOP_END_CANCEL,
2680 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2681 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2682 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2683 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2684 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2685 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2686 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2687 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2688 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2689 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2690 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2691 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2692 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2693 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2694 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2695 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2696 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2697 gimple_build_omp_teams): New functions.
2698 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2699 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2700 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2701 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2702 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2703 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2704 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2705 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2706 GF_OMP_ATOMIC_SEQ_CST): New.
2707 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2708 gimple_build_omp_teams): New prototypes.
2709 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2710 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2711 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2712 GIMPLE_OMP_SINGLE as end of range.
2713 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2714 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2715 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2716 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2717 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2718 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2719 gimple_omp_target_kind, gimple_omp_target_set_kind,
2720 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2721 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2722 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2723 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2724 gimple_omp_teams_set_clauses): New inlines.
2725 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2726 and GIMPLE_OMP_TASKGROUP.
2727 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2728 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2729 enum omp_clause_proc_bind_kind): New.
2730 (union omp_clause_subcode): Add depend_kind, map_kind and
2731 proc_bind_kind fields.
2732 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2733 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2734 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2735 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2736 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2738 2013-10-10 Teresa Johnson <tejohnson@google.com>
2740 * predict.c (tree_estimate_probability): Add new parameter
2741 for estimate_bb_frequencies.
2742 (estimate_bb_frequencies): Add new parameter to force estimation.
2743 (rebuild_frequencies): When max frequency in function is small,
2744 recompute counts from frequencies.
2745 * predict.h (estimate_bb_frequencies): New parameter.
2747 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2749 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2750 optimizations are disabled.
2752 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2754 * coverage.c (coverage_finish): Fix leak of da_file_name.
2756 2013-10-10 Jan Hubicka <jh@suse.cz>
2758 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2759 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2762 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2765 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2766 if any labels are in FALLTHRU_BB, use a special label emitted
2767 immediately after the asm goto insn rather than label_rtx
2769 (expand_asm_stmt): Adjust caller.
2770 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2771 edge if the last insn in predecessor is a jump with single successor,
2772 but it isn't simplejump_p.
2774 2013-10-10 Richard Biener <rguenther@suse.de>
2776 PR tree-optimization/58656
2777 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2779 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2781 * gimplify.c: Include expr.h and tm_p.h for targets with special
2782 va-arg padding requirements.
2784 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2786 * tree-flow.h: Move some prototypes to gimple.h.
2787 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2788 * gimple.h: Relocate some prototypes from tree-flow.h
2789 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2791 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2792 (build_va_arg_indirect_ref): Relocate and make static.
2793 (std_gimplify_va_arg_expr): Relocate here.
2794 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2795 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2797 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2799 * doc/md.texi: Document the mnemonic attribute.
2801 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2804 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2805 (set_attr_alternative x ...) when searching for user defined
2808 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2810 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2812 2013-10-09 Easwaran Raman <eraman@google.com>
2814 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2815 * cfgexpand.c (defer_stack_allocation): ...use here
2816 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2818 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2820 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2821 (operand_equal_for_value_replacement): New function, extracted from
2822 value_replacement and enhanced to catch more cases.
2823 (value_replacement): Use operand_equal_for_value_replacement.
2825 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2827 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2828 get_max_loop_iterations.
2830 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2832 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2833 Place comment above function.
2835 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2837 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2838 are not related to tree-cfg.c.
2839 * tree-ssa-address.h: New file. Relocate prototypes.
2840 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2841 (addr_for_mem_ref): New. Combine call to get_address_description and
2842 return addr_for_mem_ref.
2843 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2844 * tree-ssa-live.h: Adjust prototypes.
2845 * passes.c: Include tree-ssa-live.h.
2846 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2847 * graphite.c (graphite_transform_loops): Make static.
2848 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2849 make_pass_graphite, pass_data_graphite_transforms,
2850 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2851 * ipa-pure-const.c (warn_function_noreturn): Make static.
2852 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2853 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2854 Relocate from tree-cfg.c
2855 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2857 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2858 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2859 Move to ipa-pure-const.c.
2860 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2861 Relocate from tree-optimize.c.
2862 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2863 make_pass_fixup_cfg): Move to tree-cfg.c.
2864 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2865 Relocate some prototypes.
2866 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2867 * tree-dump.c (dump_function_to_file): Remove prototype.
2868 Add tree-flow.h to the include file.
2869 * tree-dump.h: Remove prototype.
2870 * tree-parloops.h: New File. Add prototypes.
2871 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2872 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2873 from tree-ssa-loop.c.
2874 * tree-predcom.c (run_tree_predictive_commoning,
2875 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2876 Relocate here from tree-ssa-loop.c.
2877 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2878 ssa_name_values.release ().
2879 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2880 (ssa_name_values): Relocate from tree-flow.h.
2881 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2882 * tree-ssa-loop.c (run_tree_predictive_commoning,
2883 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2884 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2885 make_pass_graphite, pass_data_graphite_transforms,
2886 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2887 tree_parallelize_loops, pass_data_parallelize_loops,
2888 make_pass_parallelize_loops): Move to other files.
2889 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2891 * tree.h: Remove prototypes from tree-address.c.
2893 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2895 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2896 (struct int_tree_map): Move to tree-hasher.h
2897 (SCALE, LABEL, PERCENT): Move to gimple.h
2898 * tree-flow-inline.h: Delete. Move functions to other files.
2899 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2900 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2901 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2902 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2903 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2904 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2905 * tree-hasher.h: Don't include tree-flow.h.
2906 (struct int_tree_map): Relocate from tree-flow.h.
2907 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2908 tree-flow-inline.h and make static.
2909 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2911 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2912 tree-flow-inline.h and make static.
2913 * tree.h (is_global_var, may_be_aliased): Relocate from
2915 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2916 * value-prof.c: No longer include tree-flow-inline.h.
2917 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2919 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2921 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2922 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2923 (enum move_pos): Move to tree-ssa-loop-im.h
2924 * cfgloop.h: Move some prototypes.
2925 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2926 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2927 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2928 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2929 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2930 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2931 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2932 (enum move_pos): Relocate here.
2933 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2935 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2936 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2937 make_pass_lim): Relocate here from tree-ssa-loop.c.
2938 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2940 (loop_edge_to_cancel, unloop_loops): Make static.
2941 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2942 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2943 (tree_complete_unroll, gate_tree_complete_unroll,
2944 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2945 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2946 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2947 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2948 (stmt_invariant_in_loop_p): Remove unused function.
2949 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2950 * tree-ssa-loop-manip.h: New file. Add prototypes.
2951 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2952 (gcov_type_to_double_int): Move to cfgloop.h.
2953 (double_int_cmp, bound_index,
2954 estimate_numbers_of_iterations_loop): Make static.
2955 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2956 (max_loop_iterations): Factor out get_max_loop_iterations.
2957 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2959 * tree-ssa-loop-niter.h: New file. Add prototypes.
2960 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2961 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2962 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2963 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2964 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2965 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2966 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2967 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2968 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2969 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2970 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2971 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2972 pass_data_complete_unroll, make_pass_complete_unroll,
2973 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2974 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
2975 tree-ssa-loop-ivcanon.c.
2976 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
2977 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
2978 tree-ssa-loop-prefetch.c.
2979 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
2981 (get_lsm_tmp_name): Relocate and add suffix parameter.
2982 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
2983 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
2984 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
2985 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
2986 (get_estimated_loop_iterations): Factor out accessor from
2987 estimated_loop_iterations in tree-ssa-loop-niter.c.
2988 (get_max_loop_iterations): Factor out accessor from
2989 _max_loop_iterations in tree-ssa-niter.c.
2990 * loop-unroll.c (decide_unroll_constant_iterations,
2991 decide_unroll_runtime_iterations, decide_peel_simple,
2992 decide_unroll_stupid): Use new get_* accessors.
2994 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
2996 PR tree-optimization/20318
2997 * doc/extend.texi (returns_nonnull): New function attribute.
2998 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
3000 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
3001 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
3003 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
3006 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
3007 false if both components are bitfields.
3009 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3011 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
3012 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
3013 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
3014 * config/aarch64/aarch64.h
3015 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
3016 * config/aarch64/aarch64-simd-builtins.def
3017 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
3019 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3021 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
3023 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3025 * config/aarch64/arm_neon.h (vdiv_f64): Added.
3027 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3029 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
3030 (vneg_f64): New intrinsic.
3031 (vneg_s8): Asm replaced with C.
3032 (vneg_s16): Likewise.
3033 (vneg_s32): Likewise.
3034 (vneg_s64): New intrinsic.
3035 (vnegq_f32): Asm replaced with C.
3036 (vnegq_f64): Likewise.
3037 (vnegq_s8): Likewise.
3038 (vnegq_s16): Likewise.
3039 (vnegq_s32): Likewise.
3040 (vnegq_s64): Likewise.
3042 2013-10-09 Renlin Li <Renlin.Li@arm.com>
3044 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
3046 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3048 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
3049 packed stack special handling.
3050 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
3051 back to fixed stack slots for FPRs saved due to stdarg.
3053 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3055 * config/s390/s390.c (s390_frame_info): Restructure function.
3057 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3059 * config/s390/s390.c (struct s390_frame_layout): New field
3061 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
3062 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
3063 regs_ever_clobbered to char*.
3064 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
3065 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
3066 (s390_register_info): Enable FPR save slots. Move/Copy some
3067 functionality into ...
3068 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
3069 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
3071 (s390_frame_info): Do gpr slot allocation here now. stdarg does
3072 not imply a stack frame.
3073 (s390_init_frame_layout): Remove variable clobbered_regs.
3074 (s390_update_register_info): Remove function.
3075 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
3076 cannot be used for register renaming.
3077 (s390_hard_regno_scratch_ok): New function.
3078 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
3079 (s390_initial_elimination_offset): Change offset calculation of
3080 the return address pointer.
3081 (save_gprs): Deal with only r6 being saved from the call-saved regs.
3082 (restore_gprs): Set frame related flag.
3083 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
3084 (s390_emit_prologue): Call s390_register_info instead of
3085 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
3086 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
3087 (s390_optimize_prologue): Call s390_optimize_register_info.
3088 Try to remove also FPR slot save/restore INSNs. Remove frame
3089 related flags from restore INSNs.
3091 2013-10-08 DJ Delorie <dj@redhat.com>
3093 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
3096 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
3097 avoid conflict with the MI use of %c.
3098 * config/rl78/rl78-real.md: change %c to %C throughout.
3099 * config/rl78/rl78-virt.md: Likewise.
3101 2013-10-08 Jan Hubicka <jh@suse.cz>
3103 * config/i386/i386.c (ix86_option_override_internal): Switch
3104 to SSE math for -ffast-math when target ISA supports SSE2.
3106 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3108 * tree-flow.h: Remove some prototypes.
3109 * tree.h: Remove some protypes, add a couple.
3110 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3111 using_eh_for_cleanups_p): Add interface routines for front ends.
3112 * tree-eh.h: New file. Add protoptyes.
3113 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3114 (add_stmt_to_eh_lp_fn): Make static.
3115 (lower_try_finally): Use new using_eh_for_cleanups_p.
3116 * emit-rtl.c: Include tree-eh.h.
3117 * gimple.h: Include tree-eh.h.
3119 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3121 PR tree-optimization/58480
3122 * tree-vrp.c (infer_nonnull_range): New function.
3123 (infer_value_range): Call infer_nonnull_range.
3125 2013-10-08 Dehao Chen <dehao@google.com>
3127 PR tree-optimization/58619
3128 * tree-inline.c (copy_phis_for_bb): Combine location data
3131 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3134 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3135 RTX_FRAME_RELATED_P on INSN.
3137 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3139 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3140 (altivec_expand_vec_perm_const): Call it.
3142 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3144 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3145 sequences for LE VSX loads and stores at expand time.
3146 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3148 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3149 (rs6000_gen_le_vsx_permute): New.
3150 (rs6000_gen_le_vsx_load): New.
3151 (rs6000_gen_le_vsx_store): New.
3152 (rs6000_gen_le_vsx_move): New.
3153 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3154 (*vsx_le_perm_load_v4si): New.
3155 (*vsx_le_perm_load_v8hi): New.
3156 (*vsx_le_perm_load_v16qi): New.
3157 (*vsx_le_perm_store_v2di): New.
3158 (*vsx_le_perm_store_v4si): New.
3159 (*vsx_le_perm_store_v8hi): New.
3160 (*vsx_le_perm_store_v16qi): New.
3161 (*vsx_xxpermdi2_le_<mode>): New.
3162 (*vsx_xxpermdi4_le_<mode>): New.
3163 (*vsx_xxpermdi8_le_V8HI): New.
3164 (*vsx_xxpermdi16_le_V16QI): New.
3165 (*vsx_lxvd2x2_le_<mode>): New.
3166 (*vsx_lxvd2x4_le_<mode>): New.
3167 (*vsx_lxvd2x8_le_V8HI): New.
3168 (*vsx_lxvd2x16_le_V16QI): New.
3169 (*vsx_stxvd2x2_le_<mode>): New.
3170 (*vsx_stxvd2x4_le_<mode>): New.
3171 (*vsx_stxvd2x8_le_V8HI): New.
3172 (*vsx_stxvd2x16_le_V16QI): New.
3174 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3176 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3178 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3180 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3181 loop to work also for 31bit ABI.
3182 Save the stack pointer for frame_size > 0.
3184 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3186 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3187 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3188 constraint letters from expanders.
3189 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3190 retry count to general_operand.
3191 ("tabort"): Give operand 0 a mode.
3192 ("tabort_1"): Add mode and constraint letter for operand 0.
3193 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3195 2013-10-04 Jeff Law <law@redhat.com>
3197 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3199 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3201 (thread_across_edge): Here. Call it.
3203 2013-10-04 Cary Coutant <ccoutant@google.com>
3205 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3206 discarding a location list expression (or a piece of one).
3208 2013-10-03 Jan Hubicka <jh@suse.cz>
3210 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3211 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3212 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3214 2013-10-03 Jan Hubicka <jh@suse.cz>
3216 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3217 accumulate-outgoing-args when producing unwind info.
3219 2013-10-03 Wei Mi <wmi@google.com>
3221 * lra-constraints.c (insert_move_for_subreg): New function
3222 extracted from simplify_operand_subreg.
3223 (simplify_operand_subreg): Add reload for paradoxical subreg.
3225 2013-10-03 Rong Xu <xur@google.com>
3227 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3228 the candidate of builtin_expect such that we should fix the
3229 size/time estimation.
3230 (estimate_function_body_sizes): Do the acutally size/time fix-up
3233 2013-10-03 Rong Xu <xur@google.com>
3235 * predict.c (tree_predict_by_opcode): Get the probability
3236 for builtin_expect from param builtin_expect_probability.
3237 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3238 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3239 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3241 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3244 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3245 * calls.c (alloca_call_p): Use get_callee_fndecl.
3246 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3247 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3249 (vrp_visit_stmt): Remove duplicated code.
3251 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3253 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3254 ceildf2, btruncdf2, instead of vsx_* name.
3256 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3257 iterators to only do V2DF and V4SF here. Move the DF code to
3258 rs6000.md where it is combined with SF mode. Replace <VSv> with
3259 just 'v' since only vector operations are handled with these insns
3260 after moving the DF support to rs6000.md.
3261 (vsx_sub<mode>3): Likewise.
3262 (vsx_mul<mode>3): Likewise.
3263 (vsx_div<mode>3): Likewise.
3264 (vsx_fre<mode>2): Likewise.
3265 (vsx_neg<mode>2): Likewise.
3266 (vsx_abs<mode>2): Likewise.
3267 (vsx_nabs<mode>2): Likewise.
3268 (vsx_smax<mode>3): Likewise.
3269 (vsx_smin<mode>3): Likewise.
3270 (vsx_sqrt<mode>2): Likewise.
3271 (vsx_rsqrte<mode>2): Likewise.
3272 (vsx_fms<mode>4): Likewise.
3273 (vsx_nfma<mode>4): Likewise.
3274 (vsx_copysign<mode>3): Likewise.
3275 (vsx_btrunc<mode>2): Likewise.
3276 (vsx_floor<mode>2): Likewise.
3277 (vsx_ceil<mode>2): Likewise.
3278 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3279 (vsx_sminsf3): Likewise.
3280 (vsx_fmadf4): Likewise.
3281 (vsx_fmsdf4): Likewise.
3282 (vsx_nfmadf4): Likewise.
3283 (vsx_nfmsdf4): Likewise.
3284 (vsx_cmpdf_internal1): Likewise.
3286 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3287 simpler to select whether a target has SPE or traditional floating
3288 point support in iterators.
3289 (TARGET_DF_SPE): Likewise.
3290 (TARGET_SF_FPR): Likewise.
3291 (TARGET_DF_FPR): Likewise.
3292 (TARGET_SF_INSN): Macros to say whether floating point support
3293 exists for a given operation for expanders.
3294 (TARGET_DF_INSN): Likewise.
3296 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3297 combining of SF/DF mode operations, using both traditional and VSX
3305 (abs<mode>2): Combine SF/DF modes using traditional floating point
3306 instructions. Add support for using the upper DF registers with
3307 VSX support, and SF registers with power8-vector support. Update
3308 expanders for operations supported by both the SPE and traditional
3309 floating point units.
3310 (abs<mode>2_fpr): Likewise.
3311 (nabs<mode>2): Likewise.
3312 (nabs<mode>2_fpr): Likewise.
3313 (neg<mode>2): Likewise.
3314 (neg<mode>2_fpr): Likewise.
3315 (add<mode>3): Likewise.
3316 (add<mode>3_fpr): Likewise.
3317 (sub<mode>3): Likewise.
3318 (sub<mode>3_fpr): Likewise.
3319 (mul<mode>3): Likewise.
3320 (mul<mode>3_fpr): Likewise.
3321 (div<mode>3): Likewise.
3322 (div<mode>3_fpr): Likewise.
3323 (sqrt<mode>3): Likewise.
3324 (sqrt<mode>3_fpr): Likewise.
3325 (fre<Fs>): Likewise.
3326 (rsqrt<mode>2): Likewise.
3327 (cmp<mode>_fpr): Likewise.
3328 (smax<mode>3): Likewise.
3329 (smin<mode>3): Likewise.
3330 (smax<mode>3_vsx): Likewise.
3331 (smin<mode>3_vsx): Likewise.
3332 (negsf2): Delete SF operations that are merged with DF.
3339 (fmasf4_fpr): Likewise.
3340 (fmssf4_fpr): Likewise.
3341 (nfmasf4_fpr): Likewise.
3342 (nfmssf4_fpr): Likewise.
3343 (sqrtsf2): Likewise.
3344 (rsqrtsf_internal1): Likewise.
3345 (smaxsf3): Likewise.
3346 (sminsf3): Likewise.
3347 (cmpsf_internal1): Likewise.
3348 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3349 (negdf2): Delete DF operations that are merged with SF.
3351 (nabsdf2): Likewise.
3357 (rsqrtdf_internal1): Likewise.
3358 (fmadf4_fpr): Likewise.
3359 (fmsdf4_fpr): Likewise.
3360 (nfmadf4_fpr): Likewise.
3361 (nfmsdf4_fpr): Likewise.
3362 (sqrtdf2): Likewise.
3363 (smaxdf3): Likewise.
3364 (smindf3): Likewise.
3365 (cmpdf_internal1): Likewise.
3366 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3367 (btrunc<mode>2): Delete separate expander, and combine with the
3368 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3369 (btrunc<mode>2_fpr): Likewise.
3370 (ceil<mode>2): Likewise.
3371 (ceil<mode>2_fpr): Likewise.
3372 (floor<mode>2): Likewise.
3373 (floor<mode>2_fpr): Likewise.
3374 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3375 Add support for using the upper registers with VSX and
3376 power8-vector. Move insns to be closer to the define_expands. On
3377 VSX systems, prefer the traditional form of FMA over the VSX
3378 version, since the traditional form allows the target not to
3379 overlap with the inputs.
3380 (fms<mode>4_fpr): Likewise.
3381 (nfma<mode>4_fpr): Likewise.
3382 (nfms<mode>4_fpr): Likewise.
3384 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3385 Richard Earnshaw <richard.earnshaw@arm.com>
3387 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3388 (struct mult_cost_table): Likewise.
3389 (struct mem_cost_table): Likewise.
3390 (struct fp_cost_table): Likewise.
3391 (struct vector_cost_table): Likewise.
3392 (cpu_cost_table): Likewise.
3393 * config/arm/arm.opt (mold-rts-costs): New option.
3394 (mnew-generic-costs): Likewise.
3395 * config/arm/arm.c (generic_extra_costs): New table.
3396 (cortexa15_extra_costs): Likewise.
3397 (arm_slowmul_tune): Use NULL as new costs.
3398 (arm_fastmul_tune): Likewise.
3399 (arm_strongarm_tune): Likewise.
3400 (arm_xscale_tune): Likewise.
3401 (arm_9e_tune): Likewise.
3402 (arm_v6t2_tune): Likewise.
3403 (arm_cortex_a5_tune): Likewise.
3404 (arm_cortex_a9_tune): Likewise.
3405 (arm_v6m_tune): Likewise.
3406 (arm_fa726te_tune): Likewise.
3407 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3408 (arm_cortex_tune): Use generict_extra_costs.
3409 (shifter_op_p): New function.
3410 (arm_unspec_cost): Likewise.
3411 (LIBCALL_COST): Define.
3412 (arm_new_rtx_costs): New function.
3413 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3414 table is available. Use old costs otherwise unless mnew-generic-costs
3416 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3417 (cpu_cost_table): Declare.
3419 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3422 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3423 (*subs_mul_imm_<mode>)
3424 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3425 (*sub_<shift>_<mode>)
3426 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3427 Remove k constraint.
3429 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3431 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3433 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3435 2013-10-02 Teresa Johnson <tejohnson@google.com>
3437 * predict.c (probably_never_executed): New function.
3438 (probably_never_executed_bb_p): Invoke probably_never_executed.
3439 (probably_never_executed_edge_p): Ditto.
3440 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3441 Treat profile insanities conservatively.
3443 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3445 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3447 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3449 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3450 use it. Use smaller increase for scratch. Don't increase reject
3451 for early clobber scratch.
3452 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3453 setting eliminated regs except setting fp from hfp.
3454 (lra_eliminate): Check lra_insn_recog_data on NULL.
3456 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3459 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3460 setting -mvsx-timode by default until the underlying problem is fixed.
3461 (RS6000_CPU, power7 defaults): Likewise.
3463 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3465 * config/x-linux (host-linux.o): Remove header dependencies.
3466 Use $(COMPILE) and $(POSTCOMPILE).
3467 * config/t-linux-android (linux-android.o): Ditto.
3469 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3471 * Makefile.in (expmed.o-warn): Remove.
3473 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3475 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3476 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3478 2013-10-02 Teresa Johnson <tejohnson@google.com>
3480 * dojump.c (do_jump_1): Divide probability between
3481 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3483 2013-10-02 Tom Tromey <tromey@redhat.com>
3485 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3487 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3489 * tree-flow.h: Remove some prototypes.
3490 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3491 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3492 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3493 mark_virtual_phi_result_for_renaming): Relocate here.
3494 * tree-into-ssa.h: Add prototypes.
3495 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3496 single_pred_before_succ_order.
3497 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3498 (nonfreeing_call_p) Move to gimple.c.
3499 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3501 * basic-block.h (single_pred_before_succ_order): Add prototype.
3502 * gimple.c (nonfreeing_call_p): Relocate here.
3503 * gimple.h: Add prototype.
3504 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3505 * tree-ssa-dom.h: New file. Relocate prototypes here.
3506 * tree-ssa.h: Include tree-ssa-dom.h.
3508 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3510 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3511 Use $(COMPILE) and $(POSTCOMPILE).
3513 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3515 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3517 * tree-flow.h: Remove some prototypes.
3518 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3519 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3520 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3521 propagate_tree_value*): Move from here to...
3522 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3523 propagate_tree_value*): Relocate here.
3524 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3525 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3526 * gimple-fold.c: Remove gimple-fold.h from include list.
3527 * tree-vrp.c: Remove gimple-fold.h from include list.
3528 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3529 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3530 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3531 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3532 * sese.c: Add tree-ssa-propagate.h to include list.
3534 2013-10-02 Richard Biener <rguenther@suse.de>
3536 * tree-loop-distribution.c: Include tree-vectorizer.h for
3538 (enum partition_kind): Remove PKIND_REDUCTION.
3539 (struct partition_s): Remove has_writes member, add reduction_p member.
3540 (partition_alloc): Adjust.
3541 (partition_builtin_p): Likewise.
3542 (partition_has_writes): Remove.
3543 (partition_reduction_p): New function.
3544 (partition_merge_into): Likewise.
3545 (generate_code_for_partition): Commonize builtin partition
3547 (rdg_cannot_recompute_vertex_p): Remove.
3548 (already_processed_vertex_p): Likewise.
3549 (rdg_flag_vertex): Do not set has_writes.
3550 (classify_partition): Adjust.
3551 (rdg_build_partitions): Do not set has_writes, treat all
3552 partitions as useful.
3553 (distribute_loop): Record number of library calls generated. Adjust.
3554 (tree_loop_distribution): Report number of loops and library
3555 calls generated as opt-info.
3557 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3559 * tree-flow.h: Include new .h files. Move prototypes.
3560 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3561 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3562 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3563 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3564 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3565 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3566 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3567 * tree.h (get_ref_base_and_extent): Move prototype out.
3568 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3570 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3571 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3572 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3573 * tree-scalar-evolution.c: Include tree.h.
3574 * sese.c: Include tree.h.
3575 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3576 * dwarf2out.c: Include tree-dfa.h.
3577 * tree-chrec.c: Include tree.h.
3578 * tree-data-ref.c: Include tree.h.
3580 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3582 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3585 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3587 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3588 Use $(COMPILE) and $(POSTCOMPILE).
3589 (sol2-cxx.o): Likewise.
3590 (sol2-stubs.o): Likewise.
3592 * config/x-solaris (host-solaris.o): Likewise.
3594 * config/sparc/t-sparc (sparc.o): Remove.
3595 (sparc-c.o): Remove header dependencies.
3596 Use $(COMPILE) and $(POSTCOMPILE).
3597 * config/sparc/x-sparc: Likewise.
3599 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3601 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3602 * config/arc/arc700.md: Likewise.
3603 * config/arc/arc-modes.def: Likewise.
3604 * config/arc/arc-simd.h: Likewise.
3605 * config/arc/t-arc-uClibc: Likewise.
3606 * config/arc/t-arc-newlib: Likewise.
3608 2013-10-02 Renlin Li <renlin.li@arm.com>
3610 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3612 (aarch64_expand_epilogue): Likewise.
3614 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3615 Yufeng Zhang <yufeng.zhang@arm.com>
3617 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3619 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3620 'base_in' represent a conversion and legal_cast_p_1 holds; set
3621 'base_in' with the returned value from get_unwidened.
3623 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3625 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3626 plus_constant is not used.
3628 2013-10-01 Wei Mi <wmi@google.com>
3630 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3631 * config/i386/i386.md: Add define_peephole2 to
3632 break partial reg stall for cvtss2sd/cvtsd2ss.
3634 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3636 * config/arc/arc.c (pass_arc_ifcvt::clone):
3637 Update for ctxt_ -> m_ctxt change.
3639 2013-10-01 Jeff Law <law@redhat.com>
3641 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3642 outgoing_edge and intermediate_edge fields. Instead store the path.
3643 (redirection_data::hash): Hash on the last edge's destination index.
3644 (redirection_data::equal): Check the entire thread path.
3645 (lookup_redirectio_data): Corresponding changes.
3646 (create_edge_and_update_destination_phis): Likewise.
3647 (thread_single_edge): Likewise.
3649 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3650 Diego Novillo <dnovillo@google.com>
3652 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3653 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3654 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3655 versions of these patterns.
3657 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3658 (__builtin_arc_aligned): Likewise.
3660 * config/arc/arc.md: Expand adc_0 comment stating the intended
3661 purpose and why it isn't ready.
3662 Replace commented out call_value_via_label_mixed with a
3663 plain comment about bl_s.
3665 * config/arc/arc.c (stdio.h): Don't include directly.
3666 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3667 Remove if (1) condition.
3668 (arc_encode_section_info): Fix comment.
3670 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3672 * config/arc/arc.c (arc_conditional_register_usage):
3673 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3674 Also set reg_alloc_order for DMA config regs.
3676 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3677 Jeremy Bennett <jeremy.bennett@embecosm.com>
3679 * doc/install.texi (--with-cpu): Mention ARC.
3680 (arc-*-elf32): New paragraph.
3681 (arc-linux-uclibc): Likewise.
3682 * doc/md.texi (Machine Constraints): Add ARC part.
3683 * doc/invoke.texi: (menu): Add ARC Options.
3684 (Machine Dependent Options) <ARC Options>: Add synopsis.
3685 (node ARC Options): Add.
3686 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3687 (ARC Built-in Functions): New section defining
3688 generic ARC built-in functions.
3689 (ARC SIMD Built-in Functions): New section defining SIMD specific
3691 (Declaring Attributes of Functions): Extended
3692 description of short_call and long_call attributes for ARC and
3693 added index entries.
3695 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3696 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3697 Joern Rennecke <joern.rennecke@embecosm.com>
3698 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3699 Brendan Kehoe <brendan@zen.org>
3700 Michael Eager <eager@eagercon.com>
3701 Simon Cook <simon.cook@embecosm.com>
3702 Jeremy Bennett <jeremy.bennett@embecosm.com>
3704 * config/arc, common/config/arc: New directories.
3706 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3707 Brendan Kehoe <brendan@zen.org>
3708 Simon Cook <simon.cook@embecosm.com>
3710 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3712 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3714 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3715 * tree-ssa-coalesce.h: New. Move prototype to here.
3716 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3717 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3718 (gimple_can_coalesce_p): Move to...
3719 * gimple.c (gimple_can_coalesce_p): Here.
3721 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3723 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3724 (dump_decl_set): Move to gimple.c.
3725 * gimple.h: Don't include tree-ssa-operands.h.
3726 (dump_decl_set): Add prototype.
3727 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3728 Move to gimple-ssa.h.
3729 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3730 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3731 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3732 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3733 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3735 (dump_decl_set): Relocate here.
3736 * gimple-ssa.h: New file.
3737 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3738 Relocate from gimple.h.
3739 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3740 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3741 swap_tree_operands and remove non-ssa path.
3742 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3743 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3745 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3746 vect_is_simple_reduction_1): Use swap_ssa_operands.
3747 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3748 (enum need_phi_state): Move to tree-into-ssa.c.
3749 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3750 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3751 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3752 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3753 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3754 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3755 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3756 num_imm_uses): Move to ssa-iterators.h.
3757 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3758 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3760 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3761 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3762 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3763 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3764 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3765 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3766 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3767 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3768 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3769 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3770 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3771 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3772 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3773 (set_phi_nodes): Move to tree-phinodes.h.
3774 * tree-ssa-operands.h (enum ssa_op_iter_type,
3775 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3776 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3777 (dump_decl_set): Remove prototype.
3778 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3779 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3780 (set_phi_nodes): Relocate from tree-flow-inline.h.
3781 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3783 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3784 include list. Temporarily add gimple.h to include list.
3785 * ssa-iterators.h: New file.
3786 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3787 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3788 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3789 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3790 Relocate from tree-ssa-operands.h.
3791 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3792 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3793 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3794 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3795 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3796 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3797 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3798 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3799 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3800 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3801 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3802 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3803 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3804 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3805 Relocate from tree-flow-inline.h.
3806 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3808 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3811 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3812 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3813 Insert '\t' to output template.
3814 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3815 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3816 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3817 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3819 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3821 * doc/install.texi (Host/target specific installation notes for GCC):
3822 Put @anchor before @heading.
3823 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3824 Use @email for email addresses.
3826 2013-10-01 Jeff Law <law@redhat.com>
3828 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3829 a vec. Only delete the path if we create one without successfully
3830 registering a jump thread.
3831 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3833 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3834 (paths): New vector of jump threading paths.
3835 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3836 (THREAD_PATH): New accessor macro for the entire thread path.
3837 (lookup_redirection_data): Get intermediate and final outgoing edge
3838 from the thread path.
3839 (create_edge_and_update_destination_phis): Copy the threading path.
3840 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3841 jump threading path.
3842 (ssa_redirect_edges): Get edges and block types from the jump threading
3843 path. Free the path vector.
3844 (thread_block): Get edges from the jump threading path. Look at the
3845 entire path to see if we thread to a loop exit. If we cancel a jump
3846 thread request, then free the path vector.
3847 (thread_single_edge): Get edges and block types from the jump threading
3848 path. Free the path vector.
3849 (thread_through_loop_header): Get edges and block types from the jump
3850 threading path. Free the path vector.
3851 (mark_threaded_blocks): Iterate over the vector of paths and store
3852 the path on the appropriate edge. Get edges and block types from the
3853 jump threading path.
3854 (mark_threaded_blocks): Get edges and block types from the jump
3855 threading path. Free the path vector.
3856 (thread_through_all_blocks): Use the vector of paths rather than
3857 a vector of 3-edge sets.
3858 (register_jump_thread): Accept pointer to a path vector rather
3859 than the path vector itself. Store the path vector for later use.
3862 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3863 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3866 * config/s390/s390.c (s390_split_branches): Modify check for table
3868 (s390_chunkify_start): Rearrange table jump insn check in order to
3869 deal with compare and branch insns correctly.
3871 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3875 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3876 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3877 define_insn_and_split.
3878 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3879 (shiftsi3_compare): New pattern.
3881 * config/arm/unspecs.md (UNSPEC_RRX): New.
3883 2013-10-01 Alan Modra <amodra@gmail.com>
3885 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3886 casing inout operands.
3888 2013-10-01 Richard Biener <rguenther@suse.de>
3890 PR tree-optimization/58553
3891 * tree-loop-distribution.c (struct partition_s): Add niter member.
3892 (classify_partition): Populate niter member for the partition
3893 and properly identify whether the relevant store happens before
3894 or after the loop exit.
3895 (generate_memset_builtin): Use niter member from the partition.
3896 (generate_memcpy_builtin): Likewise.
3898 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3900 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3901 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3903 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3905 * basic-block.h (edge_list): Prefix member names with "m_".
3906 * context.h (context): Likewise.
3907 * domwalk.h (dom_walker): Likewise.
3908 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3909 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3910 * hash-table.h (hash_table): Likewise.
3911 * machmode.h (bit_field_mode_iterator): Likewise.
3912 * pass_manager.h (pass_list): Likewise.
3913 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3914 * tree-pass.h (pass_data): Likewise.
3915 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3916 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3917 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3918 * asan.c (pass_data_asan): Update accordingly.
3919 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3920 (control_dependences::control_dependences): Likewise.
3921 (control_dependences::~control_dependences): Likewise.
3922 (control_dependences::~control_dependences): Likewise.
3923 (control_dependences::get_edges_dependent_on): Likewise.
3924 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3925 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3926 * context.c (gcc::context::context): Likewise.
3927 * cprop.c (pass_rtl_cprop::clone): Likewise.
3928 * domwalk.c (dom_walker::walk): Likewise.
3929 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3930 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3931 * mode-switching.c (pass_mode_switching::clone): Likewise.
3932 * passes.c (opt_pass::opt_pass): Likewise.
3933 (pass_manager::pass_manager): Likewise.
3934 * predict.c (pass_strip_predict_hints::clone): Likewise.
3935 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3936 (pass_split_all_insns::clone): Likewise.
3937 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3939 (bit_field_mode_iterator::next_mode): Likewise.
3940 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3941 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3942 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3943 * tree-complex.c (pass_lower_complex::clone): Likewise.
3944 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3945 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3946 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3947 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3948 (pass_fold_builtins::clone): Likewise.
3949 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3950 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3951 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3952 (pass_cd_dce::clone): Likewise.
3953 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3954 (pass_phi_only_cprop::clone): Likewise.
3955 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3956 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3957 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3958 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3959 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3960 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3961 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3962 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3963 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3964 * tree-vrp.c (pass_vrp::clone): Likewise.
3965 * tsan.c (pass_tsan::clone): Likewise.
3967 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3970 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3971 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
3974 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
3975 optimization, punt if sign_bit_p looked through any zero extension.
3977 2013-09-30 Teresa Johnson <tejohnson@google.com>
3979 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
3980 Update redirected out edge count in joiner case.
3981 (ssa_redirect_edges): Common the joiner and non-joiner cases
3982 so that joiner case gets profile updates.
3984 2013-09-30 Richard Biener <rguenther@suse.de>
3986 PR tree-optimization/58554
3987 * tree-loop-distribution.c (classify_partition): Require
3988 unconditionally executed stores for memcpy and memset recognition.
3989 (tree_loop_distribution): Calculate dominance info.
3991 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
3993 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
3994 (NO_PROFILE_COUNTERS): Likewise.
3995 (PROFILE_HOOK): Likewise.
3996 (FUNCTION_PROFILER): Likewise.
3997 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
3999 2013-09-30 Iain Sandoe <iain@codesourcery.com>
4001 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
4002 calls and defines in TARGET_MACHO conditional.
4003 (load_macho_picbase_di): Likewise.
4004 (reload_macho_picbase): Likewise.
4005 (reload_macho_picbase_si): Likewise.
4006 (reload_macho_picbase_di): Likewise.
4007 (nonlocal_goto_receiver): Likewise.
4009 2013-09-30 Nick Clifton <nickc@redhat.com>
4011 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
4012 that use the MSP430X ISA.
4013 (msp430_option_override): Scan -mmcu command line option for any
4014 MCU name that supports the MSP430X ISA.
4015 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
4016 -mmcu options which enable the MSP430X ISA.
4018 2013-09-30 Richard Biener <rguenther@suse.de>
4021 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
4022 before looking for setjmp-like calls.
4024 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4027 * config/darwin-protos.h (machopic_get_function_picbase): New.
4028 * config/darwin.c (machopic_get_function_picbase): New.
4029 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
4030 label for a new func. (load_macho_picbase_di): Likewise.
4031 (reload_macho_picbase): New expand.
4032 (reload_macho_picbase_si): New insn.
4033 (reload_macho_picbase_di): New insn.
4034 (nonlocal_goto_receiver): New define and split.
4035 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
4036 (unspecv enum): Add UNSPECV_NLGR.
4038 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4040 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
4041 that altivec registers are correctly sized on Darwin.
4043 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4045 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
4046 darwin-driver.o): Use COMPILE and POSTCOMPILE.
4047 * config/x-darwin (host-darwin.o): Likewise.
4048 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
4049 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
4050 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
4052 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
4054 * doc/invoke.texi: Fix usage of @tie{} command.
4056 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
4058 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
4060 2013-09-28 Jan Hubicka <jh@suse.cz>
4062 * config/i386/x86-tune.def: Add documentation for each of the options;
4065 2013-09-28 Jan Hubicka <jh@suse.cz>
4067 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
4069 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
4070 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
4071 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
4072 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
4073 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
4075 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
4077 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
4078 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
4079 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
4080 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
4081 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
4082 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
4083 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
4084 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
4085 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
4086 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
4087 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
4088 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
4089 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
4090 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
4091 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
4092 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
4093 gimple.h, godump.c, graphite-clast-to-gimple.c,
4094 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
4095 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
4096 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
4097 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4098 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
4099 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
4100 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
4101 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
4102 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
4103 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
4104 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
4105 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
4106 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4107 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4108 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4109 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4110 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4111 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4112 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4113 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4114 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4115 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4116 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4117 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4118 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4119 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4120 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4121 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4122 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4123 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4124 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4125 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4126 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4127 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4128 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4129 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4130 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4131 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4132 whitespace before "(".
4134 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4136 * expr.h (extract_bit_field): Remove packedp parameter.
4137 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4138 from forward declaration.
4139 (store_split_bit_field): Remove packedp arg from calls to
4140 extract_fixed_bit_field.
4141 (extract_bit_field_1): Remove packedp parameter and packedp
4142 argument from recursive calls and calls to extract_fixed_bit_field.
4143 (extract_bit_field): Remove packedp parameter and corresponding
4144 arg to extract_bit_field_1.
4145 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4147 (extract_split_bit_field): Remove packedp arg from call to
4148 extract_fixed_bit_field.
4149 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4150 (copy_blkmode_from_reg): Likewise.
4151 (copy_blkmode_to_reg): Likewise.
4152 (read_complex_part): Likewise.
4153 (store_field): Likewise.
4154 (expand_expr_real_1): Likewise.
4155 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4156 to extract_bit_field.
4157 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4158 call to extract_bit_field.
4159 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4160 call to extract_bit_field.
4161 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4162 and special packedp behavior from -fstrict-volatile-bitfields
4165 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4167 * lra-eliminations.c (init_elim_table): Guard value_p.
4169 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4171 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4172 DFmode, DImode, and SFmode in the upper VSX registers based on the
4173 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4174 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4175 of the VSX register handling.
4177 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4179 (zero_extendsidi2_lfiwzx): Add support for loading into the
4180 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4181 only do VSX memory options on Altivec registers.
4182 (extendsidi2_lfiwax): Likewise.
4183 (extendsfdf2_fpr): Likewise.
4184 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4185 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4186 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4187 (movdi_internal64): Likewise.
4189 2013-09-27 Xinliang David Li <davidxl@google.com>
4191 * opts.c (finish_options): Adjust parameters
4192 according to vect cost model.
4193 (common_handle_option): Set dynamic vect cost
4195 targhooks.c (default_add_stmt_cost): Compute stmt cost
4197 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4198 Use helper function.
4199 * tree-vectorizer.h (unlimited_cost_model): New function.
4200 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4201 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4203 (vect_enhance_data_refs_alignment): Ditto.
4204 * flag-types.h: New enum.
4205 * common/config/i386/i386-common.c (ix86_option_init_struct):
4206 No need to initialize vect_cost_model flag.
4207 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4210 2013-09-27 Diego Novillo <dnovillo@google.com>
4212 * gimple.h (enum ssa_mode): Remove.
4214 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4216 * cfgloop.h (number_of_loops): Fix typo in check for null.
4218 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4221 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4222 are children of outermost saved_cfun's loop, and set it up to
4223 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4224 if loop != loop0 and SESE region contains bbs that belong to loop0.
4226 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4228 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4229 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4230 (set_address_segment, set_address_base, set_address_index)
4231 (set_address_disp): Accept the argument unconditionally.
4232 (baseness): Remove must_be_base_p and must_be_index_p checks.
4233 (decompose_normal_address): Classify as much as possible in the
4236 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4238 * cse.c (count_reg_usage): Handle INT_LIST.
4239 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4240 * reginfo.c (reg_scan_mark_refs): Likewise.
4241 * reload1.c (eliminate_regs_1): Likewise.
4243 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4246 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4249 2013-09-27 Richard Biener <rguenther@suse.de>
4251 PR tree-optimization/58459
4252 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4253 restriction not propagating into loops.
4255 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4257 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4258 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4259 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4261 2013-09-26 Richard Biener <rguenther@suse.de>
4263 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4265 2013-09-26 Richard Biener <rguenther@suse.de>
4267 * alias.h (component_uses_parent_alias_set): Rename to ...
4268 (component_uses_parent_alias_set_from): ... this.
4269 * alias.c (component_uses_parent_alias_set): Rename to ...
4270 (component_uses_parent_alias_set_from): ... this and return
4272 (reference_alias_ptr_type_1): Use the result from
4273 component_uses_parent_alias_set_from instead of stripping
4274 components one at a time.
4275 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4277 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4279 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4280 Move prototypes to...
4281 * tree-ssa-ter.h: New File. Move prototypes here.
4282 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4283 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4284 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4285 from is_replaceable_p.
4286 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4287 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4288 (process_replaceable): Use ter_is_replaceable_p.
4289 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4290 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4291 newly refactored ssa_is_replaceable_p.
4292 * cfgexpand.c: Include tree-outof-ssa.h.
4293 * ssaexpand.h: Delete.
4295 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4297 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4298 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4300 (create_gimple_tmp): Delete.
4301 (get_expr_type, build_assign, build_type_cast): Move to...
4302 * gimple-builder.c: New File.
4303 (get_expr_type): Relocate from gimple.c.
4304 (build_assign, build_type_cast): Change to only create ssanames.
4305 * gimple.h: Move prototypes to...
4306 * gimple-builder.h: New File. Here.
4307 * tree-ssa.h: And here.
4308 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4309 count_uses_and_derefs): Relocate from gimple.c.
4310 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4311 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4312 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4313 gimple_replace_ssa_lhs.
4314 * asan.c: Include gimple-builder.h.
4315 * Makefile.in: Add gimple-builder.o.
4317 2013-09-26 Richard Biener <rguenther@suse.de>
4319 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4320 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4321 (loe_visit_block): Use gcc_checking_assert.
4322 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4323 gimple_assign_ssa_name_copy_p.
4324 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4327 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4329 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4330 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4332 2013-09-26 Richard Biener <rguenther@suse.de>
4334 PR tree-optimization/58539
4335 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4336 the fact that debug statements are not taking part in loop-closed
4339 2013-09-26 Nick Clifton <nickc@redhat.com>
4341 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4342 time warning message.
4343 (msp430_print_operand_raw): Delete unused letter parameter.
4344 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4345 (msp430_print_operand_address): New function.
4346 (msp430_print_operand): Move address printing code from here to
4348 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4350 (zero_extendpsisi2): Likewise.
4351 (extendpsisi2): New pattern.
4352 (andneghi3): New pattern.
4354 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4356 * config/aarch64/aarch64.opt (mlra): New option.
4357 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4358 (TARGET_LRA_P): Define.
4360 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4362 * expr.c (expand_assignment): Remove obsolete comment.
4364 2013-09-25 Jeff Law <law@redhat.com>
4366 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4367 tree-ssa-threadupdate.h.
4368 (register_jump_thread): Similarly.
4369 * tree-ssa-threadupdate.h: New header file.
4370 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4371 * tree-vrp.c: Likewise.
4372 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4373 (thread_around_empty_blocks): Change type of path vector argument to
4374 an edge,type pair from just an edge. Initialize both elements when
4375 appending to a jump threading path. Tweak references to elements
4377 (thread_across_edge): Similarly. Release memory for the elements
4379 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4380 (dump_jump_thread_path): New function broken out from
4381 register_jump_thread.
4382 (register_jump_thread): Use dump_jump_thread_path. Change type of
4383 path vector entries. Search the path for NULL edges and dump
4384 the path if one is found. Tweak the conversion of path to 3-edge
4385 form to use the block copy type information embedded in the path.
4387 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4389 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4391 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4392 Vladimir Makarov <vmakarov@redhat.com>
4394 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4395 from the least significant bit.
4396 (strip_address_mutations): Add bitfield operations handling.
4397 (must_be_index_p): Add shifting and rotate operations handling.
4398 (set_address_base): Use must_be_base_p predicate.
4399 (set_address_index): Use must_be_index_p predicate.
4401 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4402 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4403 Sergey Lega <sergey.s.lega@intel.com>
4404 Anna Tikhonova <anna.tikhonova@intel.com>
4405 Ilya Tocar <ilya.tocar@intel.com>
4406 Andrey Turetskiy <andrey.turetskiy@intel.com>
4407 Ilya Verbin <ilya.verbin@intel.com>
4408 Kirill Yukhin <kirill.yukhin@intel.com>
4409 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4411 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4413 (ix86_expand_vector_move_misalign): Support new unaligned load and
4414 stores and use new names.
4415 (CODE_FOR_sse2_storedqu): Rename to ...
4416 (CODE_FOR_sse2_storedquv16qi): ... this.
4417 (CODE_FOR_sse2_loaddqu): Rename to ...
4418 (CODE_FOR_sse2_loaddquv16qi): ... this.
4419 (CODE_FOR_avx_loaddqu256): Rename to ...
4420 (CODE_FOR_avx_loaddquv32qi): ... this.
4421 (CODE_FOR_avx_storedqu256): Rename to ...
4422 (CODE_FOR_avx_storedquv32qi): ... this.
4423 * config/i386/i386.md (fpint_logic): New.
4424 * config/i386/sse.md (VMOVE): Extend for AVX512.
4428 (VI_UNALIGNED_LOADSTORE): Ditto.
4429 (sse2_avx_avx512f): Ditto.
4430 (sse2_avx2): Extend for AVX512.
4431 (sse4_1_avx2): Ditto.
4432 (avx2_avx512f): New.
4433 (sse): Extend for AVX512.
4436 (avxsizesuffix): Ditto.
4437 (sseintvecmode): Ditto.
4439 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4440 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4441 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4442 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4443 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4444 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4445 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4446 (STORENT_MODE): Extend for AVX512.
4447 (*absneg<mode>2): Replace constraint "x" with "v".
4448 (*mul<mode>3): Ditto.
4449 (*ieee_smin<mode>3): Ditto.
4450 (*ieee_smax<mode>3): Ditto.
4451 (avx_cmp<mode>3): Replace VF with VF_128_256.
4452 (*<sse>_maskcmp<mode>3_comm): Ditto.
4453 (<sse>_maskcmp<mode>3): Ditto.
4454 (<sse>_andnot<mode>3): Extend for AVX512.
4455 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4456 (<code><mode>3, fpint_logic): New.
4457 (*<code><mode>3): Extend for AVX512.
4458 (avx512flogicsuff): New.
4459 (avx512f_<logic><mode>): Ditto.
4460 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4462 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4463 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4464 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4465 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4466 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4467 (xop_vpermil2<mode>3): Ditto.
4468 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4469 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4470 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4471 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4473 2013-09-25 Tom Tromey <tromey@redhat.com>
4475 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4476 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4477 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4478 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4479 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4480 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4481 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4482 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4483 (GRAPHITE_HTAB_H): Remove.
4485 2013-09-25 Tom Tromey <tromey@redhat.com>
4487 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4489 2013-09-25 Tom Tromey <tromey@redhat.com>
4491 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4493 2013-09-25 Tom Tromey <tromey@redhat.com>
4495 * config/i386/t-i386 (i386.o): Remove.
4496 (i386-c.o): Use COMPILE and POSTCOMPILE.
4498 2013-09-25 Tom Tromey <tromey@redhat.com>
4500 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4502 2013-09-25 Tom Tromey <tromey@redhat.com>
4504 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4505 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4506 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4507 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4508 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4509 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4510 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4511 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4512 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4513 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4514 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4515 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4516 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4517 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4518 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4519 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4520 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4521 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4522 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4523 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4524 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4525 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4526 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4527 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4528 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4529 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4530 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4531 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4532 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4533 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4534 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4535 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4536 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4537 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4538 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4539 (graphite-clast-to-gimple.o, graphite-dependences.o)
4540 (graphite-interchange.o, graphite-poly.o)
4541 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4542 (graphite-optimize-isl.o, tree-vect-loop.o)
4543 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4544 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4545 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4546 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4547 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4548 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4549 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4550 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4551 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4552 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4553 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4554 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4555 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4556 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4557 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4558 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4559 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4560 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4561 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4562 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4563 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4564 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4565 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4566 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4567 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4568 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4569 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4570 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4571 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4572 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4573 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4574 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4575 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4576 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4577 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4578 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4579 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4580 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4581 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4582 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4583 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4584 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4585 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4586 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4587 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4588 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4589 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4590 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4591 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4592 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4593 (gcov-dump.o): Remove.
4594 (default-c.o): Use COMPILE and POSTCOMPILE.
4595 (CFLAGS-gcc.o): New variable.
4596 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4598 2013-09-25 Tom Tromey <tromey@redhat.com>
4600 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4601 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4602 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4603 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4604 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4605 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4606 (c-family/array-notation-common.o, c-family/stub-objc.o)
4607 (c-family/c-ubsan.o): Remove.
4609 2013-09-25 Tom Tromey <tromey@redhat.com>
4611 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4613 2013-09-25 Tom Tromey <tromey@redhat.com>
4615 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4616 to add -DENABLE_SHARED_LIBGCC.
4617 (gcc.o): Don't use subshell.
4619 2013-09-25 Tom Tromey <tromey@redhat.com>
4621 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4622 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4623 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4624 * configure, config.in: Rebuild.
4626 2013-09-25 Tom Tromey <tromey@redhat.com>
4628 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4629 (COMPILE, POSTCOMPILE): New variables.
4630 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4631 (DEPFILES): New variable.
4632 Include ".Po" files.
4633 * configure.ac: Add checks for dependency checking.
4634 * configure, aclocal.m4: Regenerate.
4636 2013-09-25 Tom Tromey <tromey@redhat.com>
4638 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4639 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4641 2013-09-25 Tom Tromey <tromey@redhat.com>
4643 * Makefile.in (generated_files): Add options.h,
4644 target-hooks-def.h, insn-opinit.h,
4645 common/common-target-hooks-def.h, pass-instances.def,
4646 c-family/c-target-hooks-def.h.
4648 2013-09-25 Jeff Law <law@redhat.com>
4650 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4651 than foo[foo.length () - 1] to access last member in a vec.
4652 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4654 2013-09-25 Richard Biener <rguenther@suse.de>
4657 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4659 2013-09-25 Jan Hubicka <jh@suse.cz>
4661 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4664 2013-09-25 Marek Polacek <polacek@redhat.com>
4667 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4668 when determining the type name.
4670 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4672 * config/sh/sh.md: Fix formatting.
4674 2013-09-24 Xinliang David Li <davidxl@google.com>
4676 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4677 max peel iterations parameter.
4678 * param.def: New parameter.
4679 * doc/invoke.texi: Document New parameter.
4681 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4683 * gimple-pretty-print.c: Various whitespace tweaks.
4684 * tree-core.h: Likewise.
4685 * tree-pretty-print.c: Likewise.
4686 * tree-ssa-alias.c: Likewise.
4687 * tree-ssa-copy.c: Likewise.
4688 * tree-ssanames.c: Likewise.
4689 * tree-ssanames.h: Likewise.
4690 * tree-vrp.c: Likewise.
4692 2013-09-24 Alan Modra <amodra@gmail.com>
4696 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4697 for output operands that disallow regs. Don't use EXPAND_WRITE on
4700 2013-09-24 Richard Biener <rguenther@suse.de>
4703 * tree.c (reference_alias_ptr_type): Move ...
4704 * alias.c (reference_alias_ptr_type): ... here and implement
4705 in terms of the new reference_alias_ptr_type_1.
4706 (ref_all_alias_ptr_type_p): New helper.
4707 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4708 use ref_all_alias_ptr_type_p.
4709 (get_deref_alias_set): Add flag_strict_aliasing check here.
4710 (reference_alias_ptr_type_1): New function, split out from ...
4711 (get_alias_set): ... here.
4712 (alias_ptr_types_compatible_p): New function.
4713 * alias.h (reference_alias_ptr_type): Declare.
4714 (alias_ptr_types_compatible_p): Likewise.
4715 * tree.h (reference_alias_ptr_type): Remove.
4716 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4717 to compare MEM_REF alias types.
4719 2013-09-24 Richard Biener <rguenther@suse.de>
4721 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4723 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4725 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4726 reload helper function arrays into a single array reg_addr.
4727 (reload_fpr_gpr): Likewise.
4728 (reload_gpr_vsx): Likewise.
4729 (reload_vsx_gpr): Likewise.
4730 (struct rs6000_reg_addr): Likewise.
4731 (reg_addr): Likewise.
4732 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4733 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4734 (rs6000_init_hard_regno_mode_ok): Likewise.
4735 (rs6000_secondary_reload_direct_move): Likewise.
4736 (rs6000_secondary_reload): Likewise.
4738 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4739 constraints: wu, ww, and wy. Repurpose wv constraint added during
4740 power8 changes. Put wg constraint in alphabetical order.
4742 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4743 for future work to add ISA 2.07 VSX single precision support.
4744 (-mvsx-scalar-double): Change default from -1 to 1, update
4745 documentation comment.
4746 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4747 (-mupper-regs-df): New debug switch to control whether DF values
4748 can go in the traditional Altivec registers.
4749 (-mupper-regs-sf): New debug switch to control whether SF values
4750 can go in the traditional Altivec registers.
4752 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4754 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4755 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4756 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4757 (rs6000_debug_legitimate_address_p): Print if we are running
4758 before, during, or after reload.
4759 (rs6000_secondary_reload): Add a comment.
4760 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4762 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4763 constraints. Update documentation string.
4764 (wd constraint): Likewise.
4765 (wf constraint): Likewise.
4766 (wg constraint): Likewise.
4767 (wn constraint): Likewise.
4768 (ws constraint): Likewise.
4769 (wt constraint): Likewise.
4770 (wx constraint): Likewise.
4771 (wz constraint): Likewise.
4772 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4774 (ww constraint): Likewise.
4775 (wy constraint): Likewise.
4776 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4777 in the previous submissions.
4778 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4780 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4782 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4783 (REG_BR_PROB): Say that the probability is stored in an int_list.
4784 * reg-notes.def: Update commentary to mention INT_LIST.
4785 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4786 (INT_LIST): New rtx.
4787 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4788 * rtlanal.c (int_reg_note_p): New function.
4789 (alloc_reg_note): Assert that the note does not have an int argument.
4790 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4791 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4792 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4793 rather than an INSN_LIST. Handle INT_LIST.
4794 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4795 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4796 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4797 Manipulate them as ints rather than rtxes.
4798 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4799 * regmove.c (copy_src_to_dest): Likewise.
4800 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4802 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4803 into the cases that need it.
4804 * config/arm/arm.c (arm_unwind_emit): Likewise.
4806 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4807 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4808 * loop-doloop.c (add_test, doloop_modify): Likewise.
4809 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4810 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4811 * predict.c (combine_predictions_for_insn): Likewise.
4812 * print-rtl.c (print_rtx): Handle INT_LIST.
4813 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4814 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4815 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4816 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4817 (ix86_split_fp_branch, predict_jump): Likewise.
4818 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4819 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4820 * config/spu/spu.c (ea_load_store_inline): Likewise.
4822 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4823 value of a REG_BR_PROB note.
4824 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4825 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4826 * emit-rtl.c (try_split): Likewise.
4827 * predict.c (br_prob_note_reliable_p): Likewise.
4828 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4829 * reorg.c (mostly_true_jump): Likewise.
4830 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4831 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4832 * config/i386/i386.c (ix86_print_operand): Likewise.
4833 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4834 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4835 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4836 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4837 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4838 * config/sparc/sparc.c (output_cbranch): Likewise.
4839 * config/spu/spu.c (get_branch_target): Likewise.
4840 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4841 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4843 2013-09-23 Jan Hubicka <jh@suse.cz>
4845 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4847 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4848 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4850 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4851 (gimple_fold_call): Dump inconsistent devirtualizations; add
4852 sanity check for type based devirtualizations.
4853 * ipa-prop.c: Include ipa-utils.h
4854 (ipa_intraprocedural_devirtualization): Add sanity check.
4855 (try_make_edge_direct_virtual_call): Likewise.
4857 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4859 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4860 assignment statements.
4862 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4864 * gimple-pretty-print.c (dump_ssaname_info): New function.
4865 (dump_gimple_phi): Call it.
4866 (pp_gimple_stmt_1): Likewise.
4867 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4868 (range_info_def): Declare.
4869 * tree-pretty-print.c (pp_double_int): New function.
4870 (dump_generic_node): Call it.
4871 * tree-pretty-print.h (pp_double_int): Declare.
4872 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4873 * tree-ssanames.h (range_info_def): New structure.
4874 (value_range_type): Move definition here.
4875 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4877 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4879 (set_range_info): New function.
4880 (get_range_info): Likewise.
4881 (duplicate_ssa_name_range_info): Likewise.
4882 (duplicate_ssa_name_fn): Check pointer type and call
4883 duplicate_ssa_name_range_info.
4884 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4885 * tree-vrp.c (value_range_type): Remove definition, now in
4887 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4888 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4889 (SSA_NAME_RANGE_INFO): New macro.
4891 2013-09-23 Richard Biener <rguenther@suse.de>
4893 PR tree-optimization/58464
4894 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4895 (phi_trans_add): Change to add conditionally on being not present.
4896 (phi_translate_1): Remove recursion detection here.
4897 (phi_translate): Pre-seed the cache with NULL to catch
4898 recursion here in a more generic way.
4899 (bitmap_find_leader): Adjust comment.
4900 (get_representative_for): Dump value-numbers.
4901 (create_expression_by_pieces): Likewise.
4902 (insert_into_preds_of_block): Likewise.
4904 2013-09-23 Christian Bruel <christian.bruel@st.com>
4907 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4908 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4910 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4913 2013-09-20 Renlin Li <renlin.li@arm.com>
4915 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4917 (aarch64_expand_epilogue): Likewise.
4918 (aarch64_legitimize_reload_address): Likewise.
4920 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4922 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4923 NULL_TREE before pushing them onto the vector. Likewise for labels.
4925 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4927 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4928 comparison operators when -fno-trapping-math is in effect.
4929 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4930 comparison operators in TFmode and assert that unsupported operators
4932 (ia64_print_operand): Likewise.
4934 2013-09-21 Jan Hubicka <jh@suse.cz>
4936 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4937 (sse_typeless_stores): Enable for core
4938 (sse_load0_by_pxor): Likewise.
4939 (four_jump_limit): Disable for core.
4940 (pad_returns): Likewise.
4941 (avoid_vector_decode): Likewise.
4942 (fuse_cmp_and_branch): Enable for cores.
4943 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4945 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4948 * config/pa/pa.c (pa_option_override): Disable auto increment and
4949 decrement instructions until reload is completed.
4951 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4952 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4953 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4955 2013-09-20 DJ Delorie <dj@redhat.com>
4956 Nick Clifton <nickc@redhat.com>
4958 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4959 (need_to_save): Save bank 0 on interrupts.
4960 (characterize_address): Strip far address wrappers.
4961 (rl78_as_legitimate_address): Likewise.
4962 (transcode_memory_rtx): Likewise.
4963 (rl78_peep_movhi_p): Disable this peephole after devirt.
4964 (rl78_propogate_register_origins): Forget all origins when a
4966 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4967 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4969 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4970 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4971 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4973 * config/rl78/constraints.md: Various whitespace and paren tweaks.
4975 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4977 * config/pa/pa.md: In "scc" insn patterns, change output template to
4978 handle const0_rtx in reg_or_0_operand operands.
4980 2013-09-20 Martin Husemann <martin@NetBSD.org>
4983 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
4984 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
4986 2013-09-20 Richard Biener <rguenther@suse.de>
4989 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
4990 remembering SSA name version and block index.
4991 (new_scev_info_str): Adjust.
4992 (hash_scev_info): Likewise. Also hash the block index.
4993 (eq_scev_info): Adjust.
4994 (find_var_scev_info): Likewise.
4995 (struct instantiate_cache_entry): Remove.
4996 (struct instantiate_cache_type): Use a htab to map name, block
4998 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4999 (get_instantiated_value_entry): Likewise.
5000 (hash_idx_scev_info, eq_idx_scev_info): New functions.
5001 (instantiate_scev_name): Adjust.
5003 2013-09-20 Jeff Law <law@redhat.com>
5005 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
5007 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
5009 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
5010 Call aarch64_simd_expand_args to update op[argc].
5012 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5014 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
5017 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5019 * gengtype.c (file_rules): Added rule for *.cc files.
5020 (get_output_file_with_visibility): Give fatal message when no
5023 2013-09-20 Renlin Li <renlin.li@arm.com>
5025 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
5026 (aarch64_expand_epilogue): Likewise.
5027 (aarch64_legitimize_reload_address): Likewise.
5029 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
5032 * expr.c (expand_assignment): Remove misalignp code path.
5034 2013-09-20 Marek Polacek <polacek@redhat.com>
5037 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
5038 TYPE_PRECISION. Add asserts.
5040 2013-09-20 Richard Biener <rguenther@suse.de>
5042 PR tree-optimization/58453
5043 * tree-loop-distribution.c (distribute_loop): Apply the cost
5044 model for -ftree-loop-distribute-patterns, too.
5046 2013-09-20 Richard Biener <rguenther@suse.de>
5049 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
5050 make type comparison less strict.
5052 2013-09-20 Alan Modra <amodra@gmail.com>
5054 * configure: Regenerate.
5055 * aclocal.m4: Regenerate.
5057 2013-09-20 Marek Polacek <polacek@redhat.com>
5060 * doc/extend.texi: Document that attribute used is meant to be used
5061 on variables with static storage duration.
5063 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5065 PR tree-optimization/58472
5066 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
5067 simd_lane_access set inv_p = false.
5068 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
5069 the simduid magic VAR_DECL.
5071 2013-09-19 Jan Hubicka <jh@suse.cz>
5073 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
5075 2013-09-17 Jeff Law <law@redhat.com>
5077 * tree-ssa-dom.c (record_temporary_equivalences): New function
5078 split out of dom_opt_dom_walker::after_dom_children.
5079 (dom_opt_dom_walker::thread_across_edge): Move common code
5080 in here from dom_opt_dom_walker::after_dom_children.
5081 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
5083 2013-09-19 Jan Hubicka <jh@suse.cz>
5085 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
5086 (TARGET_GENERIC): Use PROCESOR_GENERIC
5087 (enum processor_type): Unify generic32 and 64.
5088 * i386.md (cpu): Likewise.
5089 * x86-tune.def (use_leave): Enable for generic32.
5090 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
5091 * athlon.md: Change generic64 to generic in all occurences.
5092 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
5093 (ix86_target_macros_internal): Likewise.
5094 * driver-i386.c (host_detect_local_cpu): Likewise.
5095 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
5097 (generic_memcpy, generic_memset, generic_cost): This one.
5098 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
5099 (m_GENERIC32, m_GENERIC64): Remove.
5100 (m_GENERIC): Turn into one flag.
5101 (processor_target): Unify generic tunnings.
5102 (ix86_option_override_internal): Replace generic32/64 by generic.
5103 (ix86_issue_rate): Likewise.
5104 (ix86_adjust_cost): Likewise.
5106 2013-09-19 Jan Hubicka <jh@suse.cz>
5108 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5109 of speculative flag.
5111 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5113 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5114 GOMP_sections_start, even if !exit_reachable.
5116 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5118 * lra-constraints.c (need_for_all_save_p): Use macro
5119 HARD_REGNO_CALL_PART_CLOBBERED.
5120 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5121 set up pseudo conflict hard regs.
5123 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5126 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5129 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5131 PR rtl-optimization/58438
5132 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5133 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5134 from previous subpasses.
5136 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5138 * arm.c (arm_get_frame_offsets): Validate architecture supports
5139 LDRD/STRD before accepting the tuning preference.
5140 (arm_expand_prologue): Likewise.
5141 (arm_expand_epilogue): Likewise.
5143 2013-09-18 Richard Biener <rguenther@suse.de>
5145 PR tree-optimization/58417
5146 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5147 have chrecs with symbols defined in the loop as operands.
5148 (chrec_fold_multiply): Likewise.
5149 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5150 parameters before folding binary operations.
5151 (struct instantiate_cache_entry_hasher): Remove.
5152 (struct instantiate_cache_type): Use a pointer-map.
5153 (instantiate_cache_type::instantiate_cache_type): New function.
5154 (instantiate_cache_type::get): Likewise.
5155 (instantiate_cache_type::set): Likewise.
5156 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5157 (get_instantiated_value_entry): Likewise.
5158 (global_cache): New global.
5159 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5160 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5161 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5162 (instantiate_scev_name): Adjust.
5163 (instantiate_scev): Construct global instead of local cache.
5164 (resolve_mixers): Likewise.
5166 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5167 Paolo Carlini <paolo.carlini@oracle.com>
5170 * doc/implement-cxx.texi: Fix references to the C++ standards.
5172 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5174 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5175 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5176 simd_lane_access drs, update also DR_ALIGNED_TO.
5178 2013-09-18 Marek Polacek <polacek@redhat.com>
5181 * doc/extend.texi: Document no_sanitize_undefined attribute.
5182 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5183 no_sanitize_undefined attribute.
5185 2013-09-18 Nick Clifton <nickc@redhat.com>
5187 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5188 (ASM_DECLARE_FUNCTION_NAME): Define.
5190 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5192 * compare-elim.c (find_comparison_dom_walker): New class
5193 (find_comparisons_in_bb): Rename to
5194 find_comparison_dom_walker::before_dom_children
5195 (find_comparisons): Adjust
5196 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5198 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5199 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5200 (init_walk_dominator_tree): Remove declaration.
5201 (fini_walk_dominator_tree): Remove declaration.
5202 * fwprop.c (single_def_use_dom_walker): New class
5203 (single_def_use_enter_block): Convert to
5204 single_def_use_dom_walker::before_dom_children.
5205 (single_def_use_leave_block): Convert to
5206 single_def_use_dom_walker::after_dom_children.
5207 (build_single_def_use_links): Adjust.
5208 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5210 (find_candidates_in_block): Convert to
5211 find_candidates_dom_walker::before_dom_children.
5212 (execute_strength_reduction): Adjust.
5213 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5214 (sese_dom_walker): New class.
5215 (sese_dom_walker::sese_dom_walker): New constructor.
5216 (sese_dom_walker::~sese_dom_walker): New destructor.
5217 (build_sese_conditions_before): Convert to
5218 sese_dom_walker::before_dom_children.
5219 (build_sese_conditions_after): Convert to
5220 sese_dom_walker::after_dom_children.
5221 (build_poly_scop): Adjust
5222 * tree-into-ssa.c (rewrite_dom_walker): New class
5223 (rewrite_enter_block): Convert to
5224 rewrite_dom_walker::before_dom_children.
5225 (rewrite_leave_block): Convert to
5226 rewrite_dom_walker::after_dom_children.
5227 (rewrite_update_dom_walker): New class.
5228 (rewrite_update_enter_block): Convert to
5229 rewrite_update_dom_walker::before_dom_children.
5230 (rewrite_update_leave_block): Convert to
5231 rewrite_update_dom_walker::after_dom_children.
5232 (rewrite_blocks, rewrite_into_ssa): Adjust.
5233 (mark_def_dom_walker): New class.
5234 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5235 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5236 (mark_def_sites_blocks): Convert to
5237 mark_def_dom_walker::before_dom_children.
5238 (mark_def_site_blocks): Remove.
5239 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5240 (tree_ssa_dominator_optimize): Adjust.
5241 (dom_thread_across_edge): Convert to method
5242 dom_opt_dom_walker::thread_across_edge.
5243 (dom_opt_enter_block): Convert to member function
5244 dom_opt_dom_walker::before_dom_children.
5245 (dom_opt_leave_block): Convert to member function
5246 dom_opt_dom_walker::after_dom_children.
5247 * tree-ssa-dse.c (dse_dom_walker): New class.
5248 (dse_enter_block): Convert to member function
5249 dse_dom_walker::before_dom_children.
5250 (tree_ssa_dse): Adjust.
5251 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5252 (determine_invariantness_stmt): Convert to method
5253 invariantness_dom_walker::before_dom_children.
5254 (determine_invariantness): Remove
5255 (move_computations_dom_walker): New class.
5256 (move_computations_stmt): Convert to method
5257 move_computations_dom_walker::before_dom_children.
5258 (move_computations, tree_ssa_lim): Adjust.
5259 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
5260 (nt_init_block): Make method
5261 notrappping_dom_walker::before_dom_children.
5262 (nt_fini_block): Make
5263 method nontrapping_dom_walker::after_dom_children.
5264 (get_non_trapping): Adjust.
5265 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5266 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
5267 (eliminate_leave_block): Make method.
5268 eliminate_dom_walker::after_dom_children.
5270 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5271 (strlen_enter_block): Make method
5272 strlen_dom_walker::before_dom_children.
5273 (strlen_leave_block): Make
5274 method strlen_dom_walker::after_dom_children.
5275 (tree_ssa_strlen): Adjust.
5276 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5277 (tree_ssa_uncprop): Adjust.
5278 (uncprop_leave_block): Make method
5279 uncprop_dom_walker::after_dom_children.
5280 (uncprop_leave_block): Make method
5281 uncprop_dom_walker::before_dom_children.
5283 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5285 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5286 branch in same basic block. Check both src and dest of the move insn.
5288 2013-09-17 Nick Clifton <nickc@redhat.com>
5290 * config/rl78/rl78-real.md (bf): New pattern.
5292 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5293 (rl78_print_operand): Do not put a # before a %B.
5294 * config/rl78/rl78.opt: Tweak doc strings.
5296 2013-09-17 DJ Delorie <dj@redhat.com>
5298 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5299 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5300 (rl78_option_override): Likewise, if -mallregs.
5301 (is_virtual_register): Likewise.
5302 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5303 (REGNO_OK_FOR_BASE_P): Likewise.
5304 * config/rl78/rl78.opt (-mallregs): New.
5306 2013-09-17 Nick Clifton <nickc@redhat.com>
5308 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5309 For interrupt functions: save all call clobbered registers if the
5310 interrupt handler is not a leaf function.
5311 (rl78_expand_prologue): Always recompute the frame information.
5312 For interrupt functions: only select bank 0 if one of the bank 0
5313 registers is going to be psuhed.
5315 2013-09-17 DJ Delorie <dj@redhat.com>
5317 * config/rl78/constraints.md: For each W* constraint, rename to C*
5318 and create a W* constraint that checks for an optional ES: prefix
5320 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5321 (es_addr): New. Used to wrap far addresses.
5322 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5323 (rl78_es_base): New.
5324 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5325 wrapped far addresses.
5326 (rl78_print_operand_1): Unwrap far addresses before processing.
5327 (rl78_lo16): Wrap far addresses in unspecs.
5328 (rl78_es_addr): New.
5329 (rl78_es_base): New.
5330 (insn_ok_now): Check for not-yet-wrapped far addresses.
5331 (transcode_memory_rtx): Properly re-wrap far addresses.
5333 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5335 * config/sparc/t-rtems: Add leon3 multilibs.
5337 2013-09-17 Cong Hou <congh@google.com>
5339 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5340 when checking the dot production pattern. The type of rhs operand
5341 of multiply is now checked correctly.
5343 2013-09-17 Jeff Law <law@redhat.com>
5345 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5346 edge implied equivalences into successor phis.
5347 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5348 here from tree-ssa-threadedge.c.
5349 (mark_threaded_blocks): When threading through a joiner, if both
5350 successors of the joiner's clone reach the same block, verify the
5351 PHI arguments are equal. If not, cancel the jump threading request.
5352 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5353 tree-ssa-threadupdate.c
5354 (thread_across_edge): Don't check PHI argument equality when
5355 threading through joiner block here.
5357 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5359 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5360 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5361 (warn_uninit, warn_uninitialized_vars,
5362 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5363 Move to tree-ssa-uninit.c.
5364 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5365 (has_undefined_value_p): New. Pass dependant parts of
5366 ssa_undefined_value_p.
5367 (uninit_undefined_value_p): Use has_undefined_value_p.
5368 (warn_uninit, warn_uninitialized_vars,
5369 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5370 Move from tree-ssa.c.
5371 * tree-ssa.h: Adjust prototypes.
5373 2013-09-17 Jan Hubicka <jh@suse.cz>
5376 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5377 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5378 FUNCTION_NOT_OPTIMIZED.
5379 * ipa-inline-analysis.c (compute_inline_parameters): Function
5380 not optimized is not inlinable unless it is alwaysinline.
5381 (inline_analyze_function): Force calls in not optimized
5382 function not inlinable.
5384 2013-09-17 Jan Hubicka <jh@suse.cz>
5387 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5389 * ipa.c (function_and_variable_visibility): Likewise.
5390 * ipa-profile.c (ipa_profile): Likewise.
5392 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5395 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5396 attribute, and don't inline the resolver in this case.
5398 2013-09-17 Teresa Johnson <tejohnson@google.com>
5400 * coverage.c (get_coverage_counts): Add missing newline.
5402 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5404 PR tree-optimization/58088
5405 * fold-const.c (mask_with_trailing_zeros): New function.
5406 (fold_binary_loc): Make sure we don't recurse infinitely
5407 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5408 Use mask_with_trailing_zeros where appropriate.
5410 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5412 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5413 of 'prev' var to get better distance estimation.
5415 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5417 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5418 * tree-inline.c (is_parameter_of): New predicate.
5419 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5420 a parameter has been remapped.
5421 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5422 (optimize_inline_calls): Initialize transform_parameter.
5423 (copy_gimple_seq_and_replace_locals): Likewise.
5424 (tree_function_versioning): Likewise.
5425 (maybe_inline_call_in_expr): Likewise.
5427 2013-09-17 Nick Clifton <nickc@redhat.com>
5429 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5430 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5432 (is_attr_func): New function.
5433 (msp430_is_interrupt_func): New function.
5434 (is_naked_func): New function.
5435 (is_reentrant_func): New function.
5436 (is_critical_func): New function.
5437 (msp430_start_function): Add annotations for function attributes.
5438 (msp430_attr): New function.
5439 (msp430_attribute_table): New.
5440 (msp430_function_section): New function.
5441 (TARGET_ASM_FUNCTION_SECTION): Define.
5442 (msp430_builtin): New enum.
5443 (msp430_init_builtins): New function.
5444 (msp430_builtin_devl): New function.
5445 (msp430_expand_builtin): New function.
5446 (TARGET_INIT_BUILTINS): Define.
5447 (TARGET_EXPAND_BUILTINS): Define.
5448 (TARGET_BUILTIN_DECL): Define.
5449 (msp430_expand_prologue): Add support for naked, interrupt,
5450 critical and reentrant functions.
5451 (msp430_expand_epilogue): Likewise.
5452 (msp430_print_operand): Handle 'O' character.
5453 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5455 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5456 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5457 (pushm): Use a 'n' rather than an 'i' constraint.
5458 (msp_return): Add generation of the interrupt return instruction.
5459 (disable_interrupts): New pattern.
5460 (enable_interrupts): New pattern.
5461 (push_intr_state): New pattern.
5462 (pop_intr_state): New pattern.
5463 (bic_SR): New pattern.
5464 (bis_SR): New pattern.
5465 * doc/extend.texi: Document MSP430 function attributes and builtin
5468 2013-09-17 Richard Biener <rguenther@suse.de>
5470 PR tree-optimization/58432
5471 * tree-loop-distribution.c (tree_loop_distribution): Also
5472 scan PHIs for outside loop uses and seed a partition from them.
5474 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5476 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5477 (restructure_reference): Call backtrace_base_for_ref.
5479 2013-09-17 Alan Modra <amodra@gmail.com>
5482 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5485 2013-09-16 DJ Delorie <dj@redhat.com>
5487 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5488 vregs location for RL78/G10.
5489 (rl78_expand_prologue): Avoid SEL on G10.
5490 (rl78_expand_epilogue): Likewise.
5491 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5492 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5493 __RL78_G10__ when appropriate.
5494 (ASM_SPEC): Pass -mg10 along to the assembler.
5495 * config/rl78/rl78.md (sel_rb): Disable for G10.
5496 * config/rl78/rl78.opt: Add -mg10 option.
5497 * config/rl78/t-rl78: Add -mg10 multilib.
5499 2013-09-16 Xinliang David Li <davidxl@google.com>
5501 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5502 * omp-low.c (omp_max_vf): Ditto.
5503 (expand_omp_simd): Ditto.
5504 * tree-vectorizer.c (vectorize_loops): Ditto.
5505 (gate_vect_slp): Ditto.
5506 (gate_increase_alignment): Ditto.
5507 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5508 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5509 (gate_tree_vectorize): Name change.
5510 (tree_vectorize): Ditto.
5511 (pass_vectorize::gate): Call new function.
5512 (pass_vectorize::execute): Ditto.
5513 * opts.c: O3 default setting change.
5514 (finish_options): Check new flag.
5515 * doc/invoke.texi: Document new flags.
5516 * common.opt: New flags.
5518 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5520 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5521 (Condition Code): Don't reference it.
5522 * doc/tm.texi: Regenerate.
5524 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5527 * lra-constraints.c (undo_optional_reloads): Consider all optional
5528 reload even if it did not get a hard reg.
5530 2013-09-16 Teresa Johnson <tejohnson@google.com>
5532 * dumpfile.c (dump_loc): Remove newline emission.
5533 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5534 emission to dump_printf_loc calls where missing.
5535 (vect_mark_for_runtime_alias_test): Ditto.
5536 (vect_analyze_data_ref_dependence): Ditto.
5537 (vect_analyze_data_ref_dependences): Ditto.
5538 (vect_slp_analyze_data_ref_dependence): Ditto.
5539 (vect_slp_analyze_data_ref_dependences): Ditto.
5540 (vect_compute_data_ref_alignment): Ditto.
5541 (vect_update_misalignment_for_peel): Ditto.
5542 (vect_verify_datarefs_alignment): Ditto.
5543 (vector_alignment_reachable_p): Ditto.
5544 (vect_get_data_access_cost): Ditto.
5545 (vect_enhance_data_refs_alignment): Ditto.
5546 (vect_find_same_alignment_drs): Ditto.
5547 (vect_analyze_data_refs_alignment): Ditto.
5548 (vect_analyze_group_access): Ditto.
5549 (vect_analyze_data_ref_access): Ditto.
5550 (vect_analyze_data_ref_accesses): Ditto.
5551 (vect_prune_runtime_alias_test_list): Ditto.
5552 (vect_analyze_data_refs): Ditto.
5553 (vect_create_addr_base_for_vector_ref): Ditto.
5554 (vect_create_data_ref_ptr): Ditto.
5555 (vect_grouped_store_supported): Ditto.
5556 (vect_grouped_load_supported): Ditto.
5557 * value-prof.c (check_counter): Ditto.
5558 (check_ic_target): Ditto.
5559 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5560 (vect_recog_widen_mult_pattern): Ditto.
5561 (vect_recog_widen_sum_pattern): Ditto.
5562 (vect_recog_over_widening_pattern): Ditto.
5563 (vect_recog_widen_shift_pattern): Ditto.
5564 (vect_recog_rotate_pattern): Ditto.
5565 (vect_recog_vector_vector_shift_pattern): Ditto.
5566 (vect_recog_divmod_pattern): Ditto.
5567 (vect_recog_mixed_size_cond_pattern): Ditto.
5568 (vect_recog_bool_pattern): Ditto.
5569 (vect_pattern_recog_1): Ditto.
5570 (vect_pattern_recog): Ditto.
5571 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5572 (vect_is_simple_iv_evolution): Ditto.
5573 (vect_analyze_scalar_cycles_1): Ditto.
5574 (vect_get_loop_niters): Ditto.
5575 (vect_analyze_loop_1): Ditto.
5576 (vect_analyze_loop_form): Ditto.
5577 (vect_analyze_loop_operations): Ditto.
5578 (vect_analyze_loop_2): Ditto.
5579 (vect_analyze_loop): Ditto.
5580 (report_vect_op): Ditto.
5581 (vect_is_slp_reduction): Ditto.
5582 (vect_is_simple_reduction_1): Ditto.
5583 (vect_get_known_peeling_cost): Ditto.
5584 (vect_estimate_min_profitable_iters): Ditto.
5585 (vect_model_reduction_cost): Ditto.
5586 (vect_model_induction_cost): Ditto.
5587 (get_initial_def_for_induction): Ditto.
5588 (vect_create_epilog_for_reduction): Ditto.
5589 (vectorizable_reduction): Ditto.
5590 (vectorizable_induction): Ditto.
5591 (vectorizable_live_operation): Ditto.
5592 (vect_loop_kill_debug_uses): Ditto.
5593 (vect_transform_loop): Ditto.
5594 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5595 (vect_stmt_relevant_p): Ditto.
5596 (process_use): Ditto.
5597 (vect_mark_stmts_to_be_vectorized): Ditto.
5598 (vect_model_simple_cost): Ditto.
5599 (vect_model_promotion_demotion_cost): Ditto.
5600 (vect_model_store_cost): Ditto.
5601 (vect_get_store_cost): Ditto.
5602 (vect_model_load_cost): Ditto.
5603 (vect_get_load_cost): Ditto.
5604 (vect_init_vector_1): Ditto.
5605 (vect_get_vec_def_for_operand): Ditto.
5606 (vect_finish_stmt_generation): Ditto.
5607 (vectorizable_call): Ditto.
5608 (vectorizable_conversion): Ditto.
5609 (vectorizable_assignment): Ditto.
5610 (vectorizable_shift): Ditto.
5611 (vectorizable_operation): Ditto.
5612 (vectorizable_store): Ditto.
5613 (vectorizable_load): Ditto.
5614 (vectorizable_condition): Ditto.
5615 (vect_analyze_stmt): Ditto.
5616 (vect_transform_stmt): Ditto.
5617 (vect_is_simple_use): Ditto.
5618 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5619 (vect_can_advance_ivs_p): Ditto.
5620 (vect_update_ivs_after_vectorizer): Ditto.
5621 (vect_do_peeling_for_loop_bound): Ditto.
5622 (vect_gen_niters_for_prolog_loop): Ditto.
5623 (vect_update_inits_of_drs): Ditto.
5624 (vect_create_cond_for_alias_checks): Ditto.
5625 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5626 (vect_build_slp_tree_1): Ditto.
5627 (vect_supported_load_permutation_p): Ditto.
5628 (vect_analyze_slp_instance): Ditto.
5629 (vect_analyze_slp): Ditto.
5630 (vect_make_slp_decision): Ditto.
5631 (vect_detect_hybrid_slp): Ditto.
5632 (vect_bb_vectorization_profitable_p): Ditto.
5633 (vect_slp_analyze_bb_1): Ditto.
5634 (vect_update_slp_costs_according_to_vf): Ditto.
5635 (vect_get_mask_element): Ditto.
5636 (vect_transform_slp_perm_load): Ditto.
5637 (vect_schedule_slp_instance): Ditto.
5638 (vect_schedule_slp): Ditto.
5639 (vect_slp_transform_bb): Ditto.
5640 * profile.c (read_profile_edge_counts): Ditto.
5641 (compute_branch_probabilities): Ditto.
5642 * coverage.c (get_coverage_counts): Ditto.
5644 2013-09-16 Diego Novillo <dnovillo@google.com>
5646 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5648 2013-09-16 Jan Hubicka <jh@suse.cz>
5650 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5651 abstract functions; for static functions check the presence of body.
5653 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5655 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5656 * config/aarch64/aarch64-simd.md
5657 (aarch64_mla_elt<mode>): New.
5658 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5659 (aarch64_mls_elt<mode>): Likewise.
5660 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5661 (aarch64_fma4_elt<mode>): Likewise.
5662 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5663 (aarch64_fma4_elt_to_128v2df): Likewise.
5664 (aarch64_fma4_elt_to_64df): Likewise.
5665 (fnma<mode>4): Likewise.
5666 (aarch64_fnma4_elt<mode>): Likewise.
5667 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5668 (aarch64_fnma4_elt_to_128v2df): Likewise.
5669 (aarch64_fnma4_elt_to_64df): Likewise.
5670 * config/aarch64/iterators.md (VDQSF): New.
5671 * config/aarch64/arm_neon.h
5672 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5673 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5675 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5677 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5678 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5679 (aarch64_mul3_elt_to_128df): Likewise.
5680 (aarch64_mul3_elt_to_64v2df): Likewise.
5681 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5683 (VMUL_CHANGE_NLANES) Likewise.
5686 * config/aarch64/arm_neon.h
5687 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5689 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5691 * config/aarch64/arm_neon.h
5692 (vcvtx_high_f32_f64): Fix parameters.
5694 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5695 Uros Bizjak <ubizjak@gmail.com>
5697 * config/alpha.c: Include tree-ssanames.h.
5699 2013-09-16 Richard Biener <rguenther@suse.de>
5701 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5702 (dot_rdg_1): Handle control_dd.
5703 (create_edge_for_control_dependence): New function.
5704 (create_rdg_edges): Add control dependences if asked for.
5705 (build_rdg): Likewise.
5706 (generate_loops_for_partition): If there are not necessary
5707 control stmts remove all their dependencies.
5708 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5709 (distribute_loop): Pass on control dependences.
5710 (tree_loop_distribution): Compute control dependences and remove
5711 restriction on number of loop nodes.
5713 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5715 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5718 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5720 * cse.c (try_const_anchors): Punt on CC modes.
5722 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5724 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5726 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5729 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5732 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5735 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5737 (darwin_objc1_section): Likewise.
5738 (darwin_file_end): Emit Image Info section when required.
5740 2013-09-14 Jan Hubicka <jh@suse.cz>
5742 * tree-into-ssa.c (gate_into_ssa): New.
5743 (pass_data_build_ssa): Use it.
5744 * cgraph.h (expand_thunk): Update prototype.
5745 * cgraphunit.c (analyze_function): Expand thunks early.
5746 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5747 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5748 set lowered flag; do not add new function.
5749 (assemble_thunks_and_aliases): Update.
5750 * tree-ssa.c (gate_init_datastructures): New gate.
5751 (pass_data_init_datastructures): Use it.
5753 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5756 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5757 xmm register set described in the psABI.
5759 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5761 * dwarf2out.c (should_emit_struct_debug): Add check
5762 for type_decl variable is not NULL.
5764 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5766 * config.gcc: Use new winnt-c.c target hooks
5767 * config/t-winnt: New file
5768 * config/winnt-c.c: New file
5769 * doc/tm.texi.in: Document new hook
5770 * doc/tm.texi: Regenerated
5772 2013-09-13 Jan Hubicka <jh@suse.cz>
5775 * ipa-inline.c (check_callers): New function.
5776 (check_caller_edge): Remove.
5777 (want_inline_function_to_all_callers_p): Also permit alises that are
5779 (inline_to_all_callers): Terminate the walk when devirtualization
5780 introduce new calls.
5782 2013-09-13 Jan Hubicka <jh@suse.cz>
5784 * ipa-inline-analysis.c (struct growth_data): Add node.
5785 (do_estimate_growth_1): Fix detection of recursion.
5787 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5789 PR tree-optimization/58392
5790 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5791 to avoid shadowing of outer loop variable. If
5792 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5793 replace_by_duplicate_decl simduid of loops that have it set and
5794 set dest_cfun->has_simduid_loops and/or
5795 dest_cfun->has_force_vect_loops.
5796 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5797 instead of maybe_lookup_decl.
5798 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5799 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5800 call. Copy over force_vect and copy and remap simduid. Set
5801 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5802 (copy_cfg_body): Remove blocks_to_copy argument. Use
5803 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5804 caller. Don't set cfun->has_simduid_loops and/or
5805 cfun->has_force_vect_loops here.
5806 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5808 (expand_call_inline, tree_function_versioning): Adjust copy_body
5811 2013-09-13 Martin Jambor <mjambor@suse.cz>
5814 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5815 the assert if the edge was a speculative one.
5817 2013-09-13 Richard Biener <rguenther@suse.de>
5819 * tree-data-ref.h (known_dependences_p): Move here ...
5820 * tree-loop-distribution.c (known_dependences_p): ... from here.
5821 (dump_rdg_component, debug_rdg_component): Remove.
5823 (generate_loops_for_partition): Use gimple_uid instead of
5824 relying on matching stmt visit order.
5825 (rdg_build_partitions): Take starting stmt vector.
5826 (ldist_gen): Merge into ...
5827 (distribute_loop): ... this function. Do not compute starting
5829 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5831 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5833 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5834 Set type attribute correctly. Set predicable_short_it attribute.
5835 (cmpsi_shiftsi): Remove %? from output template.
5837 2013-09-13 Richard Biener <rguenther@suse.de>
5839 * tree-loop-distribution.c (struct rdg_component,
5840 rdg_defs_used_in_other_loops_p, free_rdg_components,
5841 rdg_build_components): Remove.
5842 (stmts_from_loop): Do not record virtual PHIs.
5843 (generate_loops_for_partition): Skip virtual PHIs.
5844 (build_rdg_partition_for_component): Rename to ...
5845 (build_rdg_partition_for_vertex): ... this and adjust.
5846 (rdg_build_partitions): Take a vector of starting vertices
5847 instead of components. Remove unnecessary leftover handling.
5848 (ldist_gen): Do not build components or record other stores.
5849 (distribute_loop): Do not distribute loops containing stmts
5852 2013-09-13 Christian Bruel <christian.bruel@st.com>
5855 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5857 2013-09-13 Kai Tietz <ktietz@redhat.com>
5859 * config.gcc: Separate cases for mingw and cygwin targets,
5860 and add 64-bit cygwin target case.
5862 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5863 dll-export inline-functions.
5864 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5867 2013-09-13 Jeff Law <law@redhat.com>
5871 2013-09-06 Jeff Law <law@redhat.com>
5873 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5874 edge implied equivalences into successor phis.
5876 2013-09-12 DJ Delorie <dj@redhat.com>
5878 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5881 2013-09-12 Brooks Moses <bmoses@google.com>
5884 * Makefile.in: Do not install driver binaries in $(target)/bin.
5886 2013-09-12 DJ Delorie <dj@redhat.com>
5888 * config/rl78/rl78.opt (mrelax): New.
5889 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5890 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5892 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5894 (rl78_expand_epilogue): Likewise.
5896 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5899 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5900 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5901 pointer elimination with using remove_reg_equal_offset_note.
5903 2013-09-12 DJ Delorie <dj@redhat.com>
5905 * config/msp430/: New port.
5906 * config.gcc (msp430): Added.
5907 * doc/invoke.texi: Document MSP430 options.
5908 * doc/install.texi: Document msp430-elf
5909 * doc/md.texi: Document msp430-elf
5910 * doc/contrib.texi: Document msp430-elf
5912 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5915 2013-09-12 Martin Jambor <mjambor@suse.cz>
5918 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5919 reference descriptor is NULL.
5920 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5923 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5925 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5926 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5927 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5928 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5929 make_temp_ssa_name): move to tree-ssanames.h
5930 * tree-ssa-alias.h: Move prototype.
5931 * tree-ssa.h: Include tree-ssanames.h.
5932 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5933 * tree-ssanames.h: New. Move items from tree-flow*.h
5934 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5936 2013-09-12 Richard Biener <rguenther@suse.de>
5938 PR tree-optimization/58404
5939 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5940 propagate non-invariant addresses into dereferences wrapped
5941 in component references.
5943 2013-09-12 Richard Biener <rguenther@suse.de>
5945 PR tree-optimization/58402
5946 * passes.def: Move pass_late_warn_uninitialized later.
5948 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5950 * tree-ssa.h: New. Move content from tree-flow.h and
5952 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5953 Move prototypes belonging to tree-ssa.c.
5954 * tree-flow-inline.h (redirect_edge_var_map_def,
5955 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5957 * gimple.h: Adjust prototypes.
5958 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5960 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5961 * tree.h: Move prototype to tree-ssa.h.
5962 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5963 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5964 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5965 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5966 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5967 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5968 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5969 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5970 graphite-blocking.c, graphite-clast-to-gimple.c,
5971 graphite-dependences.c, graphite-interchange.c,
5972 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5973 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
5974 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
5975 ipa-reference.c, ipa-split.c, ipa-utils.c,
5976 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
5977 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
5978 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
5979 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
5980 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
5981 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
5982 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
5983 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
5984 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
5985 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
5986 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
5987 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
5988 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
5989 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
5990 value-prof.c, var-tracking.c,
5991 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
5993 2013-09-12 Richard Biener <rguenther@suse.de>
5995 PR tree-optimization/58396
5996 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
5997 (build_rdg): Take a loop-nest parameter, fix memleaks.
5998 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
6000 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6002 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
6005 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
6007 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
6008 The FMA instruction names should have a 'v' prefix.
6010 2013-09-12 Richard Biener <rguenther@suse.de>
6012 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
6013 (dot_rdg): Use popen instead of system in optional code.
6014 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
6015 (already_processed_vertex_p): Adjust.
6016 (has_anti_or_output_dependence, predecessor_has_mem_write,
6017 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
6018 rdg_flag_uses): Remove.
6019 (rdg_flag_vertex): Simplify.
6020 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
6022 (build_rdg_partition_for_component): Process the first vertex
6023 of a component only.
6024 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
6026 2013-09-12 Alan Modra <amodra@gmail.com>
6028 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
6030 2013-09-11 DJ Delorie <dj@redhat.com>
6031 Nick Clifton <nickc@redhat.com>
6033 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
6034 (rl78_stack_based_mem): New.
6035 * config/rl78/constraints.md (Iv08): New.
6044 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
6046 (movsi): Change from expand to insn-and-split.
6047 (ashrsi3): Clobber AX.
6051 * config/rl78/rl78.md (CC_REG): Fix.
6052 (addsi3): Allow memory and immediate operands.
6053 (addsi3_internal): Split into...
6054 (addsi3_internal_virt): ...new, and ...
6055 (addsi3_internal_real): ...new.
6057 (subsi3_internal_virt): New.
6058 (subsi3_internal_real): New.
6059 (mulsi3): Add memory operand.
6060 (mulsi3_rl78): Likewise.
6061 (mulsi3_g13): Likewise.
6062 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
6063 (cbranchqi4_real): Add more constraint options.
6064 (cbranchhi4_real): Expand pattern.
6065 (cbranchhi4_real_signed): New.
6066 (cbranchhi4_real_inverted): New.
6067 (cbranchsi4_real_lt): New.
6068 (cbranchsi4_real_ge): New.
6069 (cbranchsi4_real_signed): New.
6070 (cbranchsi4_real): New.
6072 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
6074 (lshrsi3_virt): Likewise.
6075 (ashlsi3_virt): Likewise.
6076 (cbranchqi4_virt_signed): New.
6077 (cbranchhi4_virt_signed): New.
6078 (cbranchsi4_virt): New.
6079 * config/rl78/rl78.c: Whitespace fixes throughout.
6080 (move_elim_pass): New.
6081 (pass_data_rl78_move_elim): New.
6082 (pass_rl78_move_elim): New.
6083 (make_pass_rl78_move_elim): New.
6084 (rl78_devirt_info): Run devirt earlier.
6085 (rl78_move_elim_info): New.
6086 (rl78_asm_file_start): Register it.
6087 (rl78_split_movsi): New.
6088 (rl78_as_legitimate_address): Allow virtual base registers when
6090 (rl78_addr_space_convert): Remove spurious debug stuff.
6091 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
6092 (rl78_print_operand): More cases for not printing '#'.
6093 (rl78_expand_compare): Remove most of the logic.
6094 (content_memory): New.
6095 (clear_content_memory): New.
6096 (get_content_index): New.
6097 (get_content_name): New.
6098 (display_content_memory): New.
6099 (update_content): New.
6100 (record_content): New.
6101 (already_contains): New.
6102 (insn_ok_now): Re-recog insns with virtual registers.
6103 (add_postponed_content_update): New.
6104 (process_postponed_content_update): New.
6105 (gen_and_emit_move): New.
6106 (transcode_memory_rtx): Record new location content.
6107 Use gen_and_emit_move.
6108 (force_into_acc): New.
6109 (move_to_acc): Use gen_and_emit_move.
6110 (move_from_acc): Likewise.
6111 (move_acc_to_reg): Likewise.
6112 (move_to_x): Likewise.
6113 (move_to_hl): Likewise.
6114 (move_to_de): Likewise.
6115 (rl78_alloc_physical_registers_op1): Record location content.
6116 (has_constraint): New.
6117 (rl78_alloc_physical_registers_op2): Record location content.
6119 (rl78_alloc_physical_registers_ro1): Likewise.
6120 (rl78_alloc_physical_registers_cmp): Likewise.
6121 (rl78_alloc_physical_registers_umul): Likewise.
6122 (rl78_alloc_address_registers_macax): New.
6123 (rl78_alloc_physical_registers): Initialize and set location
6124 content memory as needed.
6125 (rl78_reorg): Make sure split2 is called.
6126 (rl78_rtx_costs): New.
6128 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6130 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6131 for (not (neg ...)) and (neg (not ...)) cases.
6133 2013-09-11 Richard Biener <rguenther@suse.de>
6136 * passes.def: Split critical edges before late uninit warning passes.
6137 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6139 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6141 PR tree-optimization/58385
6142 * fold-const.c (build_range_check): If both low and high are NULL,
6143 use omit_one_operand_loc to preserve exp side-effects.
6145 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6147 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6149 2013-09-11 Richard Biener <rguenther@suse.de>
6151 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6152 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6153 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6154 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6155 stmts_from_loop, known_dependences_p, build_empty_rdg,
6156 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6157 * tree-loop-distribution.c: ... here.
6158 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6159 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6160 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6161 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6162 * tree-loop-distribution.c: ... here.
6163 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6164 (struct partition_s): Add loops member.
6165 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6166 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6167 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6169 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6170 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6171 Sergey Lega <sergey.s.lega@intel.com>
6172 Anna Tikhonova <anna.tikhonova@intel.com>
6173 Ilya Tocar <ilya.tocar@intel.com>
6174 Andrey Turetskiy <andrey.turetskiy@intel.com>
6175 Ilya Verbin <ilya.verbin@intel.com>
6176 Kirill Yukhin <kirill.yukhin@intel.com>
6177 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6179 * config/i386/constraints.md (k): New.
6181 * config/i386/i386.c (const regclass_map): Add new mask registers.
6182 (dbx_register_map): Ditto.
6183 (dbx64_register_map): Ditto.
6184 (svr4_dbx_register_map): Ditto.
6185 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6187 (ix86_preferred_reload_class): Disable constants for mask registers.
6188 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6189 (ix86_hard_regno_mode_ok): Support new mask registers.
6190 (x86_order_regs_for_local_alloc): Ditto.
6191 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6192 (FIXED_REGISTERS): Add new mask registers.
6193 (CALL_USED_REGISTERS): Ditto.
6194 (REG_ALLOC_ORDER): Ditto.
6195 (VALID_MASK_REG_MODE): New.
6196 (FIRST_MASK_REG): Ditto.
6197 (LAST_MASK_REG): Ditto.
6198 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6199 (MAYBE_MASK_CLASS_P): New.
6200 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6201 (REG_CLASS_CONTENTS): Ditto.
6202 (MASK_REGNO_P): New.
6203 (ANY_MASK_REG_P): Ditto.
6204 (HI_REGISTER_NAMES): Add new mask registers.
6205 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6206 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6208 (attribute "type"): Add mskmov, msklog.
6209 (attribute "length_immediate"): Support them.
6210 (attribute "memory"): Ditto.
6211 (attribute "prefix_0f"): Ditto.
6212 (*movhi_internal): Support new mask registers.
6213 (*movqi_internal): Ditto.
6214 (define_split): Split out clobber pattern is a logic
6215 insn on mask registers.
6216 (*k<logic><mode>): New.
6217 (*andhi_1): Extend to support mask regs.
6218 (*andqi_1): Extend to support mask regs.
6220 (define_split): Split and-not to and and not if operands
6222 (*<code><mode>_1): Separate HI mode to new pattern...
6223 (*<code>hi_1): This.
6224 (*<code>qi_1): Extend to support mask regs.
6226 (kortestzhi): Ditto.
6227 (kortestchi): Ditto.
6229 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6230 (*one_cmplhi2_1): ...Here, now with mask registers support.
6231 (*one_cmplqi2_1): Support new mask registers.
6232 (HI/QImode arithmetics splitter): Don't split if mask registers
6234 (HI/QImode not splitter): Ditto.
6235 * config/i386/predicated.md (mask_reg_operand): New.
6236 (general_reg_operand): Ditto.
6238 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6240 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6241 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6243 2013-09-10 Jeff Law <law@redhat.com>
6245 PR tree-optimization/58380
6246 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6247 of threading through a buried loop header.
6249 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6250 return value for single successor case.
6252 2013-09-10 Jan Hubicka <jh@suse.cz>
6254 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6256 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6259 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6260 support conditional execution.
6261 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6263 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6265 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6267 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6268 Check destination too to check move insn.
6269 (undo_optional_reloads): Add check that the original peudo did not
6270 changed its allocation and the optional reload was inherited on last
6271 inheritance pass. Break loop after deciding to keep optional reload.
6272 (lra_undo_inheritance): Add check that inherited pseudo still in
6275 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6277 * config/aarch64/aarch64.md (generic_sched): New.
6278 * config/aarch64/aarch64-generic.md (load): Make conditional
6279 on generic_sched attribute.
6280 (nonload): Likewise.
6282 2013-09-10 Jan Hubicka <jh@suse.cz>
6284 * lto-cgraph.c: Include ipa-utils.h.
6285 (compute_ltrans_boundary): Also add possible targets into the boundary.
6287 2013-09-10 Jan Hubicka <jh@suse.cz>
6289 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6290 VAR_DECL of vtable rather than full expression.
6292 2013-09-10 Jan Hubicka <jh@suse.cz>
6293 Paolo Carlini <paolo.carlini@oracle.com>
6295 * cgraphunit.c (analyze_functions): Save input_location, set it
6296 to UNKNOWN_LOCATION and restore it at the end.
6298 2013-09-10 Martin Jambor <mjambor@suse.cz>
6300 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6301 represented by a thunk.
6303 2013-09-10 Jeff Law <law@redhat.com>
6305 PR tree-optimization/58343
6306 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6307 jump threading requests through loop headers buried in the middle
6308 of a jump threading path.
6310 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6311 in return value/type.
6313 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6315 PR rtl-optimization/58365
6316 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6317 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6320 2013-09-10 Richard Biener <rguenther@suse.de>
6322 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6323 * tree-data-ref.c (create_rdg_vertices): Collect all data
6324 references, signal failure to the caller, use data-ref API.
6325 (build_rdg): Compute data references only once. Maintain lifetime
6326 of data references and data dependences from within RDG.
6327 (free_rdg): Free dependence relations.
6328 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6329 inventing extra dependences.
6330 (distribute_loop): Update for RDG API changes.
6332 2013-09-10 Kai Tietz <ktietz@redhat.com>
6334 * doc/invoke.texi (fms-extensions): Document changed
6335 behavior for ms-abi targets.
6336 * config/i386/i386.c (ix86_option_override_internal):
6337 Set default value of option -fms-extension for ms-abi targets.
6339 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6341 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6343 2013-09-10 Alan Modra <amodra@gmail.com>
6346 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6348 2013-09-10 Alan Modra <amodra@gmail.com>
6350 * config/rs6000/predicates.md (add_cint_operand): New.
6351 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6352 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6353 using add_cint_operand.
6354 (largetoc_high_plus_aix): Likewise.
6356 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6358 PR tree-optimization/58364
6359 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6360 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6361 the current range can't be an unconditional true or false.
6363 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6365 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6367 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6369 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6371 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6374 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6376 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6378 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6379 NO_REGS for immediate that can't be moved directly into FP_REGS.
6381 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6383 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6384 comparison with negated operand.
6385 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6388 2013-09-09 Richard Biener <rguenther@suse.de>
6391 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6392 of a subloop record all its block as affecting loop-closed SSA form.
6394 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6396 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6397 of an rtx/bitpos pair.
6398 (store_fixed_bit_field): Update accordingly.
6400 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6402 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6404 * builtins.c (expand_errno_check): Likewise.
6405 * dwarf2cfi.c (init_return_column_size): Likewise.
6406 * except.c (sjlj_mark_call_sites): Likewise.
6407 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6408 * lra-constraints.c (emit_inc): Likewise.
6409 * ree.c (combine_set_extension): Likewise.
6410 * regmove.c (fixup_match_2): Likewise.
6411 * reload1.c (inc_for_reload): Likewise.
6413 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6415 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6416 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6417 Use gen_int_mode with the mode of the associated simplify_* call.
6418 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6419 * expmed.c (expand_shift_1): Likewise.
6420 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6421 * loop-iv.c (iv_number_of_iterations): Likewise.
6422 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6423 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6425 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6427 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6428 of the associated expand_* call.
6429 (asan_emit_stack_protection): Likewise.
6430 * builtins.c (round_trampoline_addr): Likewise.
6431 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6432 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6433 (emit_store_flag): Likewise.
6434 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6435 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6437 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6438 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6439 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6440 * modulo-sched.c (generate_prolog_epilog): Likewise.
6441 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6442 (expand_ctz, expand_ffs, expand_unop): Likewise.
6444 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6446 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6447 of the associated gen_rtx_* call.
6448 * caller-save.c (init_caller_save): Likewise.
6449 * combine.c (find_split_point, make_extraction): Likewise.
6450 (make_compound_operation): Likewise.
6451 * dwarf2out.c (mem_loc_descriptor): Likewise.
6452 * explow.c (plus_constant, probe_stack_range): Likewise.
6453 * expmed.c (expand_mult_const): Likewise.
6454 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6455 * reload1.c (init_reload): Likewise.
6456 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6457 * var-tracking.c (adjust_mems): Likewise.
6458 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6459 rather than gen_rtx_fmt_ee.
6461 2013-09-09 Jan Hubicka <jh@suse.cz>
6464 * value-prof.c (gimple_ic): Copy also abnormal edges.
6466 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6468 * asan.c (asan_shadow_cst): Use gen_int_mode.
6470 2013-09-08 Jan Hubicka <jh@suse.cz>
6472 * ipa-profile.c: Add toplevel comment.
6473 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6474 (contains_hot_call_p): New function.
6475 (ipa_propagate_frequency): Set frequencies based on counts when
6477 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6478 profile; do not tamper with profile after inlining if it is read.
6480 2013-09-08 Jan Hubicka <jh@suse.cz>
6482 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6485 2013-09-08 Jan Hubicka <jh@suse.cz>
6487 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6490 2013-09-08 Jeff Law <law@redhat.com>
6493 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6496 2013-09-08 Andi Kleen <ak@linux.intel.com>
6498 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6500 2013-09-08 Jan Hubicka <jh@suse.cz>
6502 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6504 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6506 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6507 for non-debug insns.
6508 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6509 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6510 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6513 2013-09-08 Jan Hubicka <jh@suse.cz>
6515 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6516 targets and devirtualize to BUILT_IN_UNREACHABLE.
6517 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6518 * ipa.c (walk_polymorphic_call_targets): New function.
6519 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6520 functions; use the new timevar.
6521 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6522 was removed from the program.
6523 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6524 not consider it in the walk when its vtable is dead.
6525 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6527 (devirt_variable_node_removal_hook): New function.
6528 (possible_polymorphic_call_targets): Also register
6529 devirt_variable_node_removal_hook.
6530 (ipa_devirt): Do not do non-speculative devirtualization.
6531 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6533 2013-09-08 Jan Hubicka <jh@suse.cz>
6535 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6536 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6537 varpool_remove_variable_insertion_hook): Declare.
6538 * varpool.c (varpool_node_hook_list): New structure.
6539 (first_varpool_node_removal_hook,
6540 first_varpool_variable_insertion_hook): New variables.
6541 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6542 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6543 varpool_remove_variable_insertion_hook,
6544 varpool_call_variable_insertion_hooks): New functions.
6545 (varpool_remove_node): Use it.
6547 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6550 * diagnostic.c (diagnostic_build_prefix): When s.file is
6551 "<built-in>" don't output line and column numbers.
6553 2013-09-06 Jan Hubicka <jh@suse.cz>
6555 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6556 * lto-streamer-out.c: Stream thunks, too.
6557 * lto-streamer-in.c (input_function): Pop cfun here
6558 (lto_read_body): Instead of here.
6560 2013-09-06 Caroline Tice <cmtice@google.com>
6562 * doc/install.texi: Add documentation for the --enable-vtable-verify
6563 and the --disable-libvtv configure options.
6565 2013-09-06 Jeff Law <law@redhat.com>
6567 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6568 edge implied equivalences into successor phis.
6570 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6572 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6574 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6576 * resource.c (mark_target_live_regs): Compute resources taking
6577 into account if a call is predicated or not.
6579 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6581 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6582 the compiler in the function names.
6584 2013-09-06 Jan Hubicka <jh@suse.cz>
6587 * ipa-inline.c (has_caller_p): New function.
6588 (want_inline_function_to_all_callers_p): Use it.
6589 (sum_callers, inline_to_all_callers): Break out from ...
6590 (ipa_inline): ... here.
6592 2013-09-06 Jan Hubicka <jh@suse.cz>
6594 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6595 only when AVX is enabled.
6597 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6599 * config/aarch64/aarch64.md
6600 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6601 is fpsimd_<load/store>2.
6602 (load_pair<mode>): Likewise.
6603 (store_pair<mode>): Likewise.
6605 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6607 * config/arm/types.md (type): Add "mrs" type.
6608 * config/aarch64/aarch64.md
6609 (aarch64_load_tp_hard): Make type "mrs".
6611 (load_tp_hard): Make type "mrs".
6612 * config/arm/cortex-a15.md: Update with new attributes.
6613 * config/arm/cortex-a5.md: Update with new attributes.
6614 * config/arm/cortex-a53.md: Update with new attributes.
6615 * config/arm/cortex-a7.md: Update with new attributes.
6616 * config/arm/cortex-a8.md: Update with new attributes.
6617 * config/arm/cortex-a9.md: Update with new attributes.
6618 * config/arm/cortex-m4.md: Update with new attributes.
6619 * config/arm/cortex-r4.md: Update with new attributes.
6620 * config/arm/fa526.md: Update with new attributes.
6621 * config/arm/fa606te.md: Update with new attributes.
6622 * config/arm/fa626te.md: Update with new attributes.
6623 * config/arm/fa726te.md: Update with new attributes.
6625 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6627 * config/aarch64/aarch64.md
6628 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6629 (*movtf_aarch64): Likewise.
6631 (thumb1_movdi_insn): Use "multiple" for type where more than one
6632 instruction is used for a move.
6633 (*arm32_movhf): Likewise.
6634 (*thumb_movdf_insn): Likewise.
6636 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6638 * config/arm/types.md (type): Rename fcpys to fmov.
6640 (*arm_movsi_vfp): Rename type fcpys as fmov.
6641 (*thumb2_movsi_vfp): Likewise
6642 (*movhf_vfp_neon): Likewise
6643 (*movhf_vfp): Likewise
6644 (*movsf_vfp): Likewise
6645 (*thumb2_movsf_vfp): Likewise
6646 (*movsfcc_vfp): Likewise
6647 (*thumb2_movsfcc_vfp): Likewise
6648 * config/aarch64/aarch64-simd.md
6649 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6650 * config/aarch64/aarch64.md
6651 (*movsi_aarch64): Replace type mov_reg with fmovs.
6652 (*movdi_aarch64): Likewise
6653 (*movsf_aarch64): Likewise
6654 (*movdf_aarch64): Likewise
6656 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6657 * config/arm/iwmmxt.md
6658 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6659 * config/arm/arm1020e.md: Update with new attributes.
6660 * config/arm/cortex-a15-neon.md: Update with new attributes.
6661 * config/arm/cortex-a5.md: Update with new attributes.
6662 * config/arm/cortex-a53.md: Update with new attributes.
6663 * config/arm/cortex-a7.md: Update with new attributes.
6664 * config/arm/cortex-a8-neon.md: Update with new attributes.
6665 * config/arm/cortex-a9.md: Update with new attributes.
6666 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6667 * config/arm/cortex-r4f.md: Update with new attributes.
6668 * config/arm/marvell-pj4.md: Update with new attributes.
6669 * config/arm/vfp11.md: Update with new attributes.
6671 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6673 * config/aarch64/aarch64.md
6674 (*madd<mode>): Fix type attribute.
6675 (*maddsi_uxtw): Likewise.
6676 (*msub<mode>): Likewise.
6677 (*msubsi_uxtw): Likewise.
6678 (<su_optab>maddsidi4): Likewise.
6679 (<su_optab>msubsidi4): Likewise.
6681 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6683 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6684 * config/arm/arm.md (core_cycles): Remove fdiv.
6685 * config/arm/vfp.md:
6686 (*sqrtsf2_vfp): Update for attribute changes.
6687 (*sqrtdf2_vfp): Likewise.
6688 * config/aarch64/aarch64.md:
6689 (sqrt<mode>2): Update for attribute changes.
6690 * config/arm/arm1020e.md: Update with new attributes.
6691 * config/arm/cortex-a15-neon.md: Update with new attributes.
6692 * config/arm/cortex-a5.md: Update with new attributes.
6693 * config/arm/cortex-a53.md: Update with new attributes.
6694 * config/arm/cortex-a7.md: Update with new attributes.
6695 * config/arm/cortex-a8-neon.md: Update with new attributes.
6696 * config/arm/cortex-a9.md: Update with new attributes.
6697 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6698 * config/arm/cortex-r4f.md: Update with new attributes.
6699 * config/arm/marvell-pj4.md: Update with new attributes.
6700 * config/arm/vfp11.md: Update with new attributes.
6702 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6704 * config/arm/types.md
6705 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6706 * config/aarch64/aarch64.md
6707 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6709 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6710 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6711 (float<GPI:mode><GPF:mode>2): Likewise.
6713 (*truncsisf2_vfp): Update with new attributes.
6714 (*truncsidf2_vfp): Likewise.
6715 (fixuns_truncsfsi2): Likewise.
6716 (fixuns_truncdfsi2): Likewise.
6717 (*floatsisf2_vfp): Likewise.
6718 (*floatsidf2_vfp): Likewise.
6719 (floatunssisf2): Likewise.
6720 (floatunssidf2): Likewise.
6721 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6722 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6723 * config/arm/arm1020e.md: Update with new attributes.
6724 * config/arm/cortex-a15-neon.md: Update with new attributes.
6725 * config/arm/cortex-a5.md: Update with new attributes.
6726 * config/arm/cortex-a53.md: Update with new attributes.
6727 * config/arm/cortex-a7.md: Update with new attributes.
6728 * config/arm/cortex-a8-neon.md: Update with new attributes.
6729 * config/arm/cortex-a9.md: Update with new attributes.
6730 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6731 * config/arm/cortex-r4f.md: Update with new attributes.
6732 * config/arm/marvell-pj4.md: Update with new attributes.
6733 * config/arm/vfp11.md: Update with new attributes.
6735 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6737 * config/aarch64/arm_neon.h
6738 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6739 (vqtbx<1,2,3,4><q>_s8): Likewise.
6741 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6743 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6744 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6745 (add<mode>3): Add type attribute.
6746 (add<mode>3): Likewise.
6747 (usadd<mode>3): Likewise.
6748 (ssadd<mode>3): Likewise.
6749 (sub<mode>3): Likewise.
6750 (sub<mode>3): Likewise.
6751 (ussub<mode>3): Likewise.
6752 (sssub<mode>3): Likewise.
6753 (ssmulsa3): Likewise.
6754 (usmulusa3): Likewise.
6755 (arm_usatsihi): Likewise.
6757 (*movdi_vfp): Add types for all instructions.
6758 (*movdi_vfp_cortexa8): Likewise.
6759 (*movhf_vfp_neon): Likewise.
6760 (*movhf_vfp): Likewise.
6761 (*movdf_vfp): Likewise.
6762 (*thumb2_movdf_vfp): Likewise.
6763 (*thumb2_movdfcc_vfp): Likewise.
6764 * config/arm/arm.md: Add type attribute to all insn patterns.
6765 (*thumb1_adddi3): Add type attribute.
6766 (*arm_adddi3): Likewise.
6767 (*adddi_sesidi_di): Likewise.
6768 (*adddi_zesidi_di): Likewise.
6769 (*thumb1_addsi3): Likewise.
6770 (addsi3_compare0): Likewise.
6771 (*addsi3_compare0_scratch): Likewise.
6772 (*compare_negsi_si): Likewise.
6773 (cmpsi2_addneg): Likewise.
6774 (*addsi3_carryin_<optab>): Likewise.
6775 (*addsi3_carryin_alt2_<optab>): Likewise.
6776 (*addsi3_carryin_clobercc_<optab>): Likewise.
6777 (*subsi3_carryin): Likewise.
6778 (*subsi3_carryin_const): Likewise.
6779 (*subsi3_carryin_compare): Likewise.
6780 (*subsi3_carryin_compare_const): Likewise.
6781 (*arm_subdi3): Likewise.
6782 (*thumb_subdi3): Likewise.
6783 (*subdi_di_zesidi): Likewise.
6784 (*subdi_di_sesidi): Likewise.
6785 (*subdi_zesidi_di): Likewise.
6786 (*subdi_sesidi_di): Likewise.
6787 (*subdi_zesidi_ze): Likewise.
6788 (thumb1_subsi3_insn): Likewise.
6789 (*arm_subsi3_insn): Likewise.
6790 (*anddi3_insn): Likewise.
6791 (*anddi_zesidi_di): Likewise.
6792 (*anddi_sesdi_di): Likewise.
6793 (*ne_zeroextracts): Likewise.
6794 (*ne_zeroextracts): Likewise.
6795 (*ite_ne_zeroextr): Likewise.
6796 (*ite_ne_zeroextr): Likewise.
6797 (*anddi_notdi_di): Likewise.
6798 (*anddi_notzesidi): Likewise.
6799 (*anddi_notsesidi): Likewise.
6800 (andsi_notsi_si): Likewise.
6801 (thumb1_bicsi3): Likewise.
6802 (*iordi3_insn): Likewise.
6803 (*iordi_zesidi_di): Likewise.
6804 (*iordi_sesidi_di): Likewise.
6805 (*thumb1_iorsi3_insn): Likewise.
6806 (*xordi3_insn): Likewise.
6807 (*xordi_zesidi_di): Likewise.
6808 (*xordi_sesidi_di): Likewise.
6809 (*arm_xorsi3): Likewise.
6810 (*andsi_iorsi3_no): Likewise.
6811 (*smax_0): Likewise.
6812 (*smax_m1): Likewise.
6813 (*arm_smax_insn): Likewise.
6814 (*smin_0): Likewise.
6815 (*arm_smin_insn): Likewise.
6816 (*arm_umaxsi3): Likewise.
6817 (*arm_uminsi3): Likewise.
6818 (*minmax_arithsi): Likewise.
6819 (*minmax_arithsi_): Likewise.
6820 (*satsi_<SAT:code>): Likewise.
6821 (arm_ashldi3_1bit): Likewise.
6822 (arm_ashrdi3_1bit): Likewise.
6823 (arm_lshrdi3_1bit): Likewise.
6824 (*arm_negdi2): Likewise.
6825 (*thumb1_negdi2): Likewise.
6826 (*arm_negsi2): Likewise.
6827 (*thumb1_negsi2): Likewise.
6828 (*negdi_extendsid): Likewise.
6829 (*negdi_zero_extend): Likewise.
6830 (*arm_abssi2): Likewise.
6831 (*thumb1_abssi2): Likewise.
6832 (*arm_neg_abssi2): Likewise.
6833 (*thumb1_neg_abss): Likewise.
6834 (one_cmpldi2): Likewise.
6835 (extend<mode>di2): Likewise.
6836 (*compareqi_eq0): Likewise.
6837 (*arm_extendhisi2addsi): Likewise.
6838 (*arm_movdi): Likewise.
6839 (*thumb1_movdi_insn): Likewise.
6840 (*arm_movt): Likewise.
6841 (*thumb1_movsi_insn): Likewise.
6842 (pic_add_dot_plus_four): Likewise.
6843 (pic_add_dot_plus_eight): Likewise.
6844 (tls_load_dot_plus_eight): Likewise.
6845 (*thumb1_movhi_insn): Likewise.
6846 (*thumb1_movsf_insn): Likewise.
6847 (*movdf_soft_insn): Likewise.
6848 (*thumb_movdf_insn): Likewise.
6849 (cbranchsi4_insn): Likewise.
6850 (cbranchsi4_scratch): Likewise.
6851 (*negated_cbranchsi4): Likewise.
6852 (*tbit_cbranch): Likewise.
6853 (*tlobits_cbranch): Likewise.
6854 (*tstsi3_cbranch): Likewise.
6855 (*cbranchne_decr1): Likewise.
6856 (*addsi3_cbranch): Likewise.
6857 (*addsi3_cbranch_scratch): Likewise.
6858 (*arm_cmpdi_insn): Likewise.
6859 (*arm_cmpdi_unsig): Likewise.
6860 (*arm_cmpdi_zero): Likewise.
6861 (*thumb_cmpdi_zero): Likewise.
6862 (*deleted_compare): Likewise.
6863 (*mov_scc): Likewise.
6864 (*mov_negscc): Likewise.
6865 (*mov_notscc): Likewise.
6866 (*cstoresi_eq0_thumb1_insn): Likewise.
6867 (cstoresi_nltu_thumb1): Likewise.
6868 (cstoresi_ltu_thu): Likewise.
6869 (thumb1_addsi3_addgeu): Likewise.
6870 (*arm_jump): Likewise.
6871 (*thumb_jump): Likewise.
6872 (*check_arch2): Likewise.
6873 (arm_casesi_internal): Likewise.
6874 (thumb1_casesi_dispatch): Likewise.
6875 (*arm_indirect_jump): Likewise.
6876 (*thumb1_indirect_jump): Likewise.
6878 (*and_scc): Likewise.
6879 (*ior_scc): Likewise.
6880 (*compare_scc): Likewise.
6881 (*cond_move): Likewise.
6882 (*cond_arith): Likewise.
6883 (*cond_sub): Likewise.
6884 (*cmp_ite0): Likewise.
6885 (*cmp_ite1): Likewise.
6886 (*cmp_and): Likewise.
6887 (*cmp_ior): Likewise.
6888 (*ior_scc_scc): Likewise.
6889 (*ior_scc_scc_cmp): Likewise.
6890 (*and_scc_scc): Likewise.
6891 (*and_scc_scc_cmp): Likewise.
6892 (*and_scc_scc_nod): Likewise.
6893 (*negscc): Likewise.
6894 (movcond_addsi): Likewise.
6895 (movcond): Likewise.
6896 (*ifcompare_plus_move): Likewise.
6897 (*if_plus_move): Likewise.
6898 (*ifcompare_move_plus): Likewise.
6899 (*if_move_plus): Likewise.
6900 (*ifcompare_arith_arith): Likewise.
6901 (*if_arith_arith): Likewise.
6902 (*ifcompare_arith_move): Likewise.
6903 (*if_arith_move): Likewise.
6904 (*ifcompare_move_arith): Likewise.
6905 (*if_move_arith): Likewise.
6906 (*ifcompare_move_not): Likewise.
6907 (*if_move_not): Likewise.
6908 (*ifcompare_not_move): Likewise.
6909 (*if_not_move): Likewise.
6910 (*ifcompare_shift_move): Likewise.
6911 (*if_shift_move): Likewise.
6912 (*ifcompare_move_shift): Likewise.
6913 (*if_move_shift): Likewise.
6914 (*ifcompare_shift_shift): Likewise.
6915 (*ifcompare_not_arith): Likewise.
6916 (*ifcompare_arith_not): Likewise.
6917 (*if_arith_not): Likewise.
6918 (*ifcompare_neg_move): Likewise.
6919 (*if_neg_move): Likewise.
6920 (*ifcompare_move_neg): Likewise.
6921 (*if_move_neg): Likewise.
6922 (prologue_thumb1_interwork): Likewise.
6923 (*cond_move_not): Likewise.
6924 (*sign_extract_onebit): Likewise.
6925 (*not_signextract_onebit): Likewise.
6926 (stack_tie): Likewise.
6927 (align_4): Likewise.
6928 (align_8): Likewise.
6929 (consttable_end): Likewise.
6930 (consttable_1): Likewise.
6931 (consttable_2): Likewise.
6932 (consttable_4): Likewise.
6933 (consttable_8): Likewise.
6934 (consttable_16): Likewise.
6935 (*thumb1_tablejump): Likewise.
6936 (prefetch): Likewise.
6937 (force_register_use): Likewise.
6938 (thumb_eh_return): Likewise.
6939 (load_tp_hard): Likewise.
6940 (load_tp_soft): Likewise.
6941 (tlscall): Likewise.
6942 (*arm_movtas_ze): Likewise.
6943 (*arm_rev): Likewise.
6944 (*arm_revsh): Likewise.
6945 (*arm_rev16): Likewise.
6946 * config/arm/thumb2.md
6947 (*thumb2_smaxsi3): Likewise.
6948 (*thumb2_sminsi3): Likewise.
6949 (*thumb32_umaxsi3): Likewise.
6950 (*thumb2_uminsi3): Likewise.
6951 (*thumb2_negdi2): Likewise.
6952 (*thumb2_abssi2): Likewise.
6953 (*thumb2_neg_abss): Likewise.
6954 (*thumb2_movsi_insn): Likewise.
6955 (tls_load_dot_plus_four): Likewise.
6956 (*thumb2_movhi_insn): Likewise.
6957 (*thumb2_mov_scc): Likewise.
6958 (*thumb2_mov_negs): Likewise.
6959 (*thumb2_mov_negs): Likewise.
6960 (*thumb2_mov_nots): Likewise.
6961 (*thumb2_mov_nots): Likewise.
6962 (*thumb2_movsicc_): Likewise.
6963 (*thumb2_movsfcc_soft_insn): Likewise.
6964 (*thumb2_indirect_jump): Likewise.
6965 (*thumb2_and_scc): Likewise.
6966 (*thumb2_ior_scc): Likewise.
6967 (*thumb2_ior_scc_strict_it): Likewise.
6968 (*thumb2_cond_move): Likewise.
6969 (*thumb2_cond_arith): Likewise.
6970 (*thumb2_cond_ari): Likewise.
6971 (*thumb2_cond_sub): Likewise.
6972 (*thumb2_negscc): Likewise.
6973 (*thumb2_movcond): Likewise.
6974 (thumb2_casesi_internal): Likewise.
6975 (thumb2_casesi_internal_pic): Likewise.
6976 (*thumb2_alusi3_short): Likewise.
6977 (*thumb2_mov<mode>_shortim): Likewise.
6978 (*thumb2_addsi_short): Likewise.
6979 (*thumb2_subsi_short): Likewise.
6980 (thumb2_addsi3_compare0): Likewise.
6981 (*thumb2_cbz): Likewise.
6982 (*thumb2_cbnz): Likewise.
6983 (*thumb2_one_cmplsi2_short): Likewise.
6984 (*thumb2_negsi2_short): Likewise.
6985 (*orsi_notsi_si): Likewise.
6986 * config/arm/arm1020e.md: Update with new attributes.
6987 * config/arm/arm1026ejs.md: Update with new attributes.
6988 * config/arm/arm1136jfs.md: Update with new attributes.
6989 * config/arm/arm926ejs.md: Update with new attributes.
6990 * config/arm/cortex-a15.md: Update with new attributes.
6991 * config/arm/cortex-a5.md: Update with new attributes.
6992 * config/arm/cortex-a53.md: Update with new attributes.
6993 * config/arm/cortex-a7.md: Update with new attributes.
6994 * config/arm/cortex-a8.md: Update with new attributes.
6995 * config/arm/cortex-a9.md: Update with new attributes.
6996 * config/arm/cortex-m4.md: Update with new attributes.
6997 * config/arm/cortex-r4.md: Update with new attributes.
6998 * config/arm/fa526.md: Update with new attributes.
6999 * config/arm/fa606te.md: Update with new attributes.
7000 * config/arm/fa626te.md: Update with new attributes.
7001 * config/arm/fa726te.md: Update with new attributes.
7003 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7005 * config/aarch64/aarch64-simd.md
7006 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
7007 <vwx> iterator to ensure correct register choice.
7008 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7009 (aarch64_sqdmull_n<mode>): Likewise.
7010 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7011 * config/aarch64/arm_neon.h
7012 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
7013 (vml<as><q>_n_<su>16): Likewise.
7014 (vml<as>l_high_lane<q>_<su>16): Likewise.
7015 (vml<as>l_high_n_<su>16): Likewise.
7016 (vml<as>l_lane<q>_<su>16): Likewise.
7017 (vml<as>l_n_<su>16): Likewise.
7018 (vmul<q>_lane<q>_<su>16): Likewise.
7019 (vmul<q>_n_<su>16): Likewise.
7020 (vmull_lane<q>_<su>16): Likewise.
7021 (vmull_n_<su>16): Likewise.
7022 (vmull_high_lane<q>_<su>16): Likewise.
7023 (vmull_high_n_<su>16): Likewise.
7024 (vqrdmulh<q>_n_s16): Likewise.
7026 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
7028 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
7029 have the correct lane parameter.
7031 2013-09-06 Richard Biener <rguenther@suse.de>
7033 * cfganal.c (control_dependences::~control_dependences):
7034 Properly free all of the vector.
7036 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
7039 * config/i386/i386.c (ix86_conditional_register_usage):
7040 Proper initialize extended SSE registers.
7042 2013-09-06 Jan Hubicka <jh@suse.cz>
7044 PR tree-optimization/58311
7045 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
7047 2013-09-06 Jan Hubicka <jh@suse.cz>
7049 * Makefile.in (tree-sra.o): Update dependencies.
7050 * tree-sra.c: Include ipa-utils.h
7051 (scan_function): Use recursive_call_p.
7052 (has_caller_p): New function.
7053 (cgraph_for_node_and_aliases): Count also callers of aliases.
7055 2013-09-06 Jan Hubicka <jh@suse.cz>
7058 * cgraph.h (symtab_semantically_equivalent_p): Declare.
7059 * tree-tailcall.c: Include ipa-utils.h.
7060 (find_tail_calls): Use it.
7061 * ipa-pure-const.c (check_call): Likewise.
7062 * ipa-utils.c (recursive_call_p): New function.
7063 * ipa-utils.h (recursive_call_p): Dclare.
7064 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
7065 (symtab_semantically_equivalent_p): New function.
7066 * Makefile.in (tree-tailcall.o): Update dependencies.
7068 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7070 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
7073 2013-09-06 Richard Biener <rguenther@suse.de>
7075 * lto-streamer.h (lto_global_var_decls): Remove.
7076 * Makefile.in (OBJS): Remove lto-symtab.o.
7077 (lto-symtab.o): Remove.
7078 (GTFILES): Remove lto-symtab.c
7079 * lto-symtab.c: Move to lto/
7081 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7083 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
7084 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
7085 (UNSPEC_FPINT_RINT): New constant definitions.
7086 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
7087 definition with 2 attributes.
7088 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
7089 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
7092 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7094 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
7095 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
7096 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
7097 attribute to "z196_alone".
7098 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
7101 2013-09-06 Richard Biener <rguenther@suse.de>
7103 * basic-block.h (class control_dependences): New.
7104 * tree-ssa-dce.c (control_dependence_map): Remove.
7106 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7107 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7108 find_pdom, find_control_dependence, find_all_control_dependences):
7110 (mark_control_dependent_edges_necessary,
7111 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7112 tree_dce_done, perform_tree_ssa_dce): Adjust.
7113 * cfganal.c (set_control_dependence_map_bit,
7114 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7115 find_all_control_dependences): Move from tree-ssa-dce.c and
7116 implement as methods of control_dependences class.
7117 (control_dependences::control_dependences): New.
7118 (control_dependences::~control_dependences): Likewise.
7119 (control_dependences::get_edges_dependent_on): Likewise.
7120 (control_dependences::get_edge): Likewise.
7122 2013-09-04 Jan Hubicka <jh@suse.cz>
7124 * tree.c (types_same_for_odr): Drop overactive check.
7125 * ipa-devirt.c (hash_type_name): Likewise.
7127 2013-09-04 Jan Hubicka <jh@suse.cz>
7129 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7130 (analyze_functions): ... here.
7132 2013-09-04 Jan Hubicka <jh@suse.cz>
7135 * cgraphunit.c (analyze_functions): Clear AUX fields
7136 after processing; initialize assembler name has.
7138 2013-09-05 Jeff Law <law@redhat.com>
7140 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7141 from thread_around_empty_block. Record threading path into PATH.
7142 Recurse if threading through the initial block is successful.
7143 (thread_across_edge): Corresponding changes to slightly simplify.
7145 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7147 * config/aarch64/aarch64.md
7148 (type): Remove frecpe, frecps, frecpx.
7149 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7150 fix to be a TARGET_SIMD instruction.
7151 (aarch64_frecps): Remove.
7152 * config/aarch64/aarch64-simd.md
7153 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7154 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7156 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7157 Sofiane Naci <sofiane.naci@arm.com>
7159 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7160 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7161 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7162 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7163 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7164 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7165 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7166 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7167 into "clz, "rbit". Rename "shift" to "shift_imm".
7168 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7169 changes. Update for attribute changes all occurrences of arlo_* and
7171 * config/arm/arm-fixed.md: Update for attribute changes
7172 all occurrences of arlo_* types.
7173 * config/arm/thumb2.md: Update for attribute changes all occurrences
7175 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7176 (cortexa7_older_only): Likewise.
7177 (cortexa7_younger): Likewise.
7178 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7179 (1020alu_shift_op): Likewise.
7180 (1020alu_shift_reg_op): Likewise.
7181 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7182 (alu_shift_op): Likewise.
7183 (alu_shift_reg_op): Likewise.
7184 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7185 (11_alu_shift_op): Likewise.
7186 (11_alu_shift_reg_op): Likewise.
7187 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7188 (9_alu_shift_reg_op): Likewise.
7189 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7191 (cortex_a15_alu_shift): Likewise.
7192 (cortex_a15_alu_shift_reg): Likewise.
7193 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7195 (cortex_a5_alu_shift): Likewise.
7196 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7198 (cortex_a53_alu_shift): Likewise.
7199 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7201 (cortex_a7_alu_reg): Likewise.
7202 (cortex_a7_alu_shift): Likewise.
7203 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7205 (cortex_a8_alu_shift): Likewise.
7206 (cortex_a8_alu_shift_reg): Likewise.
7207 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7208 (cortex_a9_dp_shift): Likewise.
7209 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7211 * config/arm/cortex-r4.md
7212 (cortex_r4_alu): Update for attribute changes.
7213 (cortex_r4_mov): Likewise.
7214 (cortex_r4_alu_shift_reg): Likewise.
7215 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7216 (526_alu_shift_op): Likewise.
7217 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7218 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7219 (626te_alu_shift_op): Likewise.
7220 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7221 (726te_alu_shift_op): Likewise.
7222 (726te_alu_shift_reg_op): Likewise.
7223 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7224 (mp626_alu_shift_op): Likewise.
7225 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7226 (pj4_alu_conds): Likewise.
7227 (pj4_shift): Likewise.
7228 (pj4_shift_conds): Likewise.
7229 (pj4_alu_shift): Likewise.
7230 (pj4_alu_shift_conds): Likewise.
7231 * config/aarch64/aarch64.md: Update for attribute change
7232 all occurrences of arlo_* and shift* types.
7234 2013-09-05 Mike Stump <mikestump@comcast.net>
7236 * tree.h: Move documentation for tree_function_decl to tree-core.h
7237 with the declaration.
7239 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7242 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7243 looking for widest mode.
7245 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7247 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7249 2013-09-05 Richard Biener <rguenther@suse.de>
7251 PR tree-optimization/58137
7252 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7253 Do not create vectors of pointers.
7254 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7255 types for the components of the vector initializer.
7256 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7257 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7259 2013-09-05 Martin Jambor <mjambor@suse.cz>
7261 * ipa-prop.c (remove_described_reference): Accept missing references,
7262 return false if that hppens, otherwise return true.
7263 (cgraph_node_for_jfunc): New function.
7264 (try_decrement_rdesc_refcount): Likewise.
7265 (try_make_edge_direct_simple_call): Use them.
7266 (ipa_edge_removal_hook): Remove references from rdescs.
7267 (ipa_edge_duplication_hook): Clone rdescs and their references
7268 when the new edge has the same caller as the old one.
7269 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7270 reference before removing any edges.
7272 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7274 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7276 * thumb2.md (thumb2_storewb_parisi): New pattern.
7278 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7280 * config/aarch64/aarch64-option-extensions.def: Add
7281 AARCH64_OPT_EXTENSION of 'crc'.
7282 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7283 (AARCH64_ISA_CRC): Ditto.
7284 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7285 description of the CRC extension.
7287 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7289 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7290 * config/rs6000/linux.h: Ditto.
7291 * alpha/linux.h: Ditto.
7292 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7293 no_c99_libc_has_function.
7294 * config/c6x/uclinux-elf.h: Ditto.
7295 * config/lm32/uclinux-elf.h: Ditto.
7296 * config/m68k/uclinux.h: Ditto.
7297 * config/moxie/uclinux.h: Ditto.
7298 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7299 (crisv32-*-linux*, cris-*-linux*): Ditto.
7300 * config/bfin/bfin.c: Include "tm_p.h".
7302 2013-09-05 Richard Biener <rguenther@suse.de>
7304 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7305 check for a definition without a basic-block.
7307 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7308 Sofiane Naci <sofiane.naci@arm.com>
7310 * config/aarch64/aarch64.md
7311 (*movti_aarch64): Rename r_2_f and f_2_r.
7312 (*movsf_aarch64): Likewise.
7313 (*movdf_aarch64): Likewise.
7314 (*movtf_aarch64): Likewise.
7315 (aarch64_movdi_<mode>low): Likewise.
7316 (aarch64_movdi_<mode>high): Likewise.
7317 (aarch64_mov<mode>high_di): Likewise.
7318 (aarch64_mov<mode>low_di): Likewise.
7319 (aarch64_movtilow_tilow): Likewise.
7320 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7321 values to config/arm/types.md
7322 (attribute "conds"): Update for attribute change.
7323 (anddi3_insn): Likewise.
7324 (iordi3_insn): Likewise.
7325 (xordi3_insn): Likewise.
7326 (one_cmpldi2): Likewise.
7327 * config/arm/types.md (type): Add Neon types.
7328 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7329 use "type" attribute.
7330 (movmisalign<mode>_neon_store): Likewise.
7331 (movmisalign<mode>_neon_load): Likewise.
7332 (vec_set<mode>_internal): Likewise.
7333 (vec_setv2di_internal): Likewise.
7334 (vec_extract<mode>): Likewise.
7335 (vec_extractv2di): Likewise.
7336 (add<mode>3_neon): Likewise.
7337 (adddi3_neon): Likewise.
7338 (sub<mode>3_neon): Likewise.
7339 (subdi3_neon): Likewise.
7340 (mul<mode>3_neon): Likewise.
7341 (mul<mode>3add<mode>_neon): Likewise.
7342 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7343 (fma<VCVTF:mode>4)): Likewise.
7344 (fma<VCVTF:mode>4_intrinsic): Likewise.
7345 (fmsub<VCVTF:mode>4)): Likewise.
7346 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7347 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7348 (ior<mode>3): Likewise.
7349 (and<mode>3): Likewise.
7350 (anddi3_neon): Likewise.
7351 (orn<mode>3_neon): Likewise.
7352 (orndi3_neon): Likewise.
7353 (bic<mode>3_neon): Likewise.
7354 (bicdi3_neon): Likewise.
7355 (xor<mode>3): Likewise.
7356 (one_cmpl<mode>2): Likewise.
7357 (abs<mode>2): Likewise.
7358 (neg<mode>2): Likewise.
7359 (umin<mode>3_neon): Likewise.
7360 (umax<mode>3_neon): Likewise.
7361 (smin<mode>3_neon): Likewise.
7362 (smax<mode>3_neon): Likewise.
7363 (vashl<mode>3): Likewise.
7364 (vashr<mode>3_imm): Likewise.
7365 (vlshr<mode>3_imm): Likewise.
7366 (ashl<mode>3_signed): Likewise.
7367 (ashl<mode>3_unsigned): Likewise.
7368 (neon_load_count): Likewise.
7369 (ashldi3_neon_noclobber): Likewise.
7370 (signed_shift_di3_neon): Likewise.
7371 (unsigned_shift_di3_neon): Likewise.
7372 (ashrdi3_neon_imm_noclobber): Likewise.
7373 (lshrdi3_neon_imm_noclobber): Likewise.
7374 (widen_ssum<mode>3): Likewise.
7375 (widen_usum<mode>3): Likewise.
7376 (quad_halves_<code>v4si): Likewise.
7377 (quad_halves_<code>v4sf): Likewise.
7378 (quad_halves_<code>v8hi): Likewise.
7379 (quad_halves_<code>v16qi): Likewise.
7380 (reduc_splus_v2di): Likewise.
7381 (neon_vpadd_internal<mode>): Likewise.
7382 (neon_vpsmin<mode>): Likewise.
7383 (neon_vpsmax<mode>): Likewise.
7384 (neon_vpumin<mode>): Likewise.
7385 (neon_vpumax<mode>): Likewise.
7386 (ss_add<mode>_neon): Likewise.
7387 (us_add<mode>_neon): Likewise.
7388 (ss_sub<mode>_neon): Likewise.
7389 (us_sub<mode>_neon): Likewise.
7390 (neon_vadd<mode>_unspec): Likewise.
7391 (neon_vaddl<mode>): Likewise.
7392 (neon_vaddw<mode>): Likewise.
7393 (neon_vhadd<mode>): Likewise.
7394 (neon_vqadd<mode>): Likewise.
7395 (neon_vaddhn<mode>): Likewise.
7396 (neon_vmul<mode>): Likewise.
7397 (neon_vmla<mode>): Likewise.
7398 (neon_vmlal<mode>): Likewise.
7399 (neon_vmls<mode>): Likewise.
7400 (neon_vmlsl<mode>): Likewise.
7401 (neon_vqdmulh<mode>): Likewise.
7402 (neon_vqdmlal<mode>): Likewise.
7403 (neon_vqdmlsl<mode>): Likewise.
7404 (neon_vmull<mode>): Likewise.
7405 (neon_vqdmull<mode>): Likewise.
7406 (neon_vsub<mode>_unspec): Likewise.
7407 (neon_vsubl<mode>): Likewise.
7408 (neon_vsubw<mode>): Likewise.
7409 (neon_vqsub<mode>): Likewise.
7410 (neon_vhsub<mode>): Likewise.
7411 (neon_vsubhn<mode>): Likewise.
7412 (neon_vceq<mode>): Likewise.
7413 (neon_vcge<mode>): Likewise.
7414 (neon_vcgeu<mode>): Likewise.
7415 (neon_vcgt<mode>): Likewise.
7416 (neon_vcgtu<mode>): Likewise.
7417 (neon_vcle<mode>): Likewise.
7418 (neon_vclt<mode>): Likewise.
7419 (neon_vcage<mode>): Likewise.
7420 (neon_vcagt<mode>): Likewise.
7421 (neon_vtst<mode>): Likewise.
7422 (neon_vabd<mode>): Likewise.
7423 (neon_vabdl<mode>): Likewise.
7424 (neon_vaba<mode>): Likewise.
7425 (neon_vabal<mode>): Likewise.
7426 (neon_vmax<mode>): Likewise.
7427 (neon_vmin<mode>): Likewise.
7428 (neon_vpaddl<mode>): Likewise.
7429 (neon_vpadal<mode>): Likewise.
7430 (neon_vpmax<mode>): Likewise.
7431 (neon_vpmin<mode>): Likewise.
7432 (neon_vrecps<mode>): Likewise.
7433 (neon_vrsqrts<mode>): Likewise.
7434 (neon_vqabs<mode>): Likewise.
7435 (neon_vqneg<mode>): Likewise.
7436 (neon_vcls<mode>): Likewise.
7437 (clz<mode>2): Likewise.
7438 (popcount<mode>2): Likewise.
7439 (neon_vrecpe): Likewise.
7440 (neon_vrsqrte): Likewise.
7441 (neon_vget_lane<mode>_sext_internal): Likewise.
7442 (neon_vget_lane<mode>_zext_internal): Likewise.
7443 (neon_vdup_n<mode>): Likewise.
7444 (neon_vdup_nv2di): Likewise.
7445 (neon_vdpu_lane<mode>_internal): Likewise.
7446 (neon_vswp<mode>): Likewise.
7447 (float<mode><V_cvtto>2): Likewise.
7448 (floatuns<mode><V_cvtto>2): Likewise.
7449 (fix_trunc<mode><V_cvtto>)2): Likewise
7450 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7451 (neon_vcvt<mode>): Likewise.
7452 (neon_vcvtv4sfv4hf): Likewise.
7453 (neon_vcvtv4hfv4sf): Likewise.
7454 (neon_vcvt_n<mode>): Likewise.
7455 (neon_vmovn<mode>): Likewise.
7456 (neon_vqmovn<mode>): Likewise.
7457 (neon_vqmovun<mode>): Likewise.
7458 (neon_vmovl<mode>): Likewise.
7459 (neon_vmul_lane<mode>): Likewise.
7460 (neon_vmull_lane<mode>): Likewise.
7461 (neon_vqdmull_lane<mode>): Likewise.
7462 (neon_vqdmulh_lane<mode>): Likewise.
7463 (neon_vmla_lane<mode>): Likewise.
7464 (neon_vmlal_lane<mode>): Likewise.
7465 (neon_vqdmlal_lane<mode>): Likewise.
7466 (neon_vmls_lane<mode>): Likewise.
7467 (neon_vmlsl_lane<mode>): Likewise.
7468 (neon_vqdmlsl_lane<mode>): Likewise.
7469 (neon_vext<mode>): Likewise.
7470 (neon_vrev64<mode>): Likewise.
7471 (neon_vrev32<mode>): Likewise.
7472 (neon_vrev16<mode>): Likewise.
7473 (neon_vbsl<mode>_internal): Likewise.
7474 (neon_vshl<mode>): Likewise.
7475 (neon_vqshl<mode>): Likewise.
7476 (neon_vshr_n<mode>): Likewise.
7477 (neon_vshrn_n<mode>): Likewise.
7478 (neon_vqshrn_n<mode>): Likewise.
7479 (neon_vqshrun_n<mode>): Likewise.
7480 (neon_vshl_n<mode>): Likewise.
7481 (neon_vqshl_n<mode>): Likewise.
7482 (neon_vqshlu_n<mode>): Likewise.
7483 (neon_vshll_n<mode>): Likewise.
7484 (neon_vsra_n<mode>): Likewise.
7485 (neon_vsri_n<mode>): Likewise.
7486 (neon_vsli_n<mode>): Likewise.
7487 (neon_vtbl1v8qi): Likewise.
7488 (neon_vtbl2v8qi): Likewise.
7489 (neon_vtbl3v8qi): Likewise.
7490 (neon_vtbl4v8qi): Likewise.
7491 (neon_vtbx1v8qi): Likewise.
7492 (neon_vtbx2v8qi): Likewise.
7493 (neon_vtbx3v8qi): Likewise.
7494 (neon_vtbx4v8qi): Likewise.
7495 (neon_vtrn<mode>_internal): Likewise.
7496 (neon_vzip<mode>_internal): Likewise.
7497 (neon_vuzp<mode>_internal): Likewise.
7498 (neon_vld1<mode>): Likewise.
7499 (neon_vld1_lane<mode>): Likewise.
7500 (neon_vld1_dup<mode>): Likewise.
7501 (neon_vld1_dupv2di): Likewise.
7502 (neon_vst1<mode>): Likewise.
7503 (neon_vst1_lane<mode>): Likewise.
7504 (neon_vld2<mode>): Likewise.
7505 (neon_vld2_lane<mode>): Likewise.
7506 (neon_vld2_dup<mode>): Likewise.
7507 (neon_vst2<mode>): Likewise.
7508 (neon_vst2_lane<mode>): Likewise.
7509 (neon_vld3<mode>): Likewise.
7510 (neon_vld3qa<mode>): Likewise.
7511 (neon_vld3qb<mode>): Likewise.
7512 (neon_vld3_lane<mode>): Likewise.
7513 (neon_vld3_dup<mode>): Likewise.
7514 (neon_vst3<mode>): Likewise.
7515 (neon_vst3qa<mode>): Likewise.
7516 (neon_vst3qb<mode>): Likewise.
7517 (neon_vst3_lane<mode>): Likewise.
7518 (neon_vld4<mode>): Likewise.
7519 (neon_vld4qa<mode>): Likewise.
7520 (neon_vld4qb<mode>): Likewise.
7521 (neon_vld4_lane<mode>): Likewise.
7522 (neon_vld4_dup<mode>): Likewise.
7523 (neon_vst4<mode>): Likewise.
7524 (neon_vst4qa<mode>): Likewise.
7525 (neon_vst4qb<mode>): Likewise.
7526 (neon_vst4_lane<mode>): Likewise.
7527 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7528 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7529 (neon_vec_<US>mult_lo_<mode>): Likewise.
7530 (neon_vec_<US>mult_hi_<mode>): Likewise.
7531 (neon_vec_<US>shiftl_<mode>): Likewise.
7532 (neon_unpack<US>_<mode>): Likewise.
7533 (neon_vec_<US>mult_<mode>): Likewise.
7534 (vec_pack_trunc_<mode>): Likewise.
7535 (neon_vec_pack_trunk_<mode>): Likewise.
7536 (neon_vabd<mode>_2): Likewise.
7537 (neon_vabd<mode>_3): Likewise.
7538 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7539 (thumb2_movsi_vfp): Likewise.
7540 (movdi_vfp): Likewise.
7541 (movdi_vfp_cortexa8): Likewise.
7542 (movhf_vfp_neon): Likewise.
7543 (movhf_vfp): Likewiwse.
7544 (movsf_vfp): Likewiwse.
7545 (thumb2_movsf_vfp): Likewiwse.
7546 (movdf_vfp): Likewise.
7547 (thumb2_movdf_vfp): Likewise.
7548 (movsfcc_vfp): Likewise.
7549 (thumb2_movsfcc_vfp): Likewise.
7550 (movdfcc_vfp): Likewise.
7551 (thumb2_movdfcc_vfp): Likewise.
7552 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7553 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7554 (v10_v2c): Likewise.
7555 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7557 (cortex_a15_neon_int_2): Likewise.
7558 (cortex_a15_neon_int_3): Likewise.
7559 (cortex_a15_neon_int_4): Likewise.
7560 (cortex_a15_neon_int_5): Likewise.
7561 (cortex_a15_neon_vqneg_vqabs): Likewise.
7562 (cortex_a15_neon_vmov): Likewise.
7563 (cortex_a15_neon_vaba): Likewise.
7564 (cortex_a15_neon_vaba_qqq): Likewise.
7565 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7566 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7567 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7568 scalar_64_32_long_scalar): Likewise.
7569 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7570 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7571 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7572 lotype_qdd_64_32_long): Likewise.
7573 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7574 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7575 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7576 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7577 (cortex_a15_neon_shift_1): Likewise.
7578 (cortex_a15_neon_shift_2): Likewise.
7579 (cortex_a15_neon_shift_3): Likewise.
7580 (cortex_a15_neon_vshl_ddd): Likewise.
7581 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7582 (cortex_a15_neon_vsra_vrsra): Likewise.
7583 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7584 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7585 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7586 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7587 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7588 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7589 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7590 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7591 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7592 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7593 (cortex_a15_neon_bp_simple): Likewise.
7594 (cortex_a15_neon_bp_2cycle): Likewise.
7595 (cortex_a15_neon_bp_3cycle): Likewise.
7596 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7597 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7598 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7599 (cortex_a15_neon_vld2_4_regs): Likewise.
7600 (cortex_a15_neon_vld3_vld4): Likewise.
7601 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7602 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7603 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7604 (cortex_a15_neon_vst3_vst4): Likewise.
7605 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7606 (cortex_a15_neon_vld3_vld4_lane" 10
7607 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7608 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7609 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7610 (cortex_a15_neon_ldm_2): Likewise.0
7611 (cortex_a15_neon_stm_2): Likewise.
7612 (cortex_a15_neon_mcr): Likewise.
7613 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7614 (cortex_a15_neon_mrc): Likewise.
7615 (cortex_a15_neon_mrrc): Likewise.
7616 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7618 (cortex_a15_alu_shift): Likewise.
7619 (cortex_a15_alu_shift_reg): Likewise.
7620 (cortex_a15_mult32): Likewise.
7621 (cortex_a15_mult64): Likewise.
7622 (cortex_a15_block): Likewise.
7623 (cortex_a15_branch): Likewise.
7624 (cortex_a15_load1): Likewise.
7625 (cortex_a15_load3): Likewise.
7626 (cortex_a15_store1): Likewise.
7627 (cortex_a15_store3): Likewise.
7628 (cortex_a15_call): Likewise.
7629 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7630 (cortex_a5_f2r): Likewise.
7631 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7633 (cortex_a53_f2r): Likewise.
7634 * config/arm/cortex-a7.md
7635 (cortex_a7_branch): Update for attribute change.
7636 (cortex_a7_call): Likewise.
7637 (cortex_a7_alu_imm): Likewise.
7638 (cortex_a7_alu_reg): Likewise.
7639 (cortex_a7_alu_shift): Likewise.
7640 (cortex_a7_mul): Likewise.
7641 (cortex_a7_load1): Likewise.
7642 (cortex_a7_store1): Likewise.
7643 (cortex_a7_load2): Likewise.
7644 (cortex_a7_store2): Likewise.
7645 (cortex_a7_load3): Likewise.
7646 (cortex_a7_store3): Likewise.
7647 (cortex_a7_load4): Likewise.
7648 (cortex_a7_store4): Likewise.
7649 (cortex_a7_fpalu): Likewise.
7650 (cortex_a7_fconst): Likewise.
7651 (cortex_a7_fpmuls): Likewise.
7652 (cortex_a7_neon_mul): Likewise.
7653 (cortex_a7_fpmacs): Likewise.
7654 (cortex_a7_neon_mla: Likewise.
7655 (cortex_a7_fpmuld: Likewise.
7656 (cortex_a7_fpmacd: Likewise.
7657 (cortex_a7_fpfmad: Likewise.
7658 (cortex_a7_fdivs: Likewise.
7659 (cortex_a7_fdivd: Likewise.
7660 (cortex_a7_r2f: Likewise.
7661 (cortex_a7_f2r: Likewise.
7662 (cortex_a7_f_flags: Likewise.
7663 (cortex_a7_f_loads: Likewise.
7664 (cortex_a7_f_loadd: Likewise.
7665 (cortex_a7_f_stores: Likewise.
7666 (cortex_a7_f_stored: Likewise.
7667 (cortex_a7_neon): Likewise.
7668 * config/arm/cortex-a8-neon.md
7669 (cortex_a8_neon_mrc): Update for attribute change.
7670 (cortex_a8_neon_mrrc): Likewise.
7671 (cortex_a8_neon_int_1): Likewise.
7672 (cortex_a8_neon_int_2): Likewise.
7673 (cortex_a8_neon_int_3): Likewise.
7674 (cortex_a8_neon_int_4): Likewise.
7675 (cortex_a8_neon_int_5): Likewise.
7676 (cortex_a8_neon_vqneg_vqabs): Likewise.
7677 (cortex_a8_neon_vmov): Likewise.
7678 (cortex_a8_neon_vaba): Likewise.
7679 (cortex_a8_neon_vaba_qqq): Likewise.
7680 (cortex_a8_neon_vsma): Likewise.
7681 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7682 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7683 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7684 long_scalar): Likewise.
7685 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7686 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7687 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7688 long_scalar_qdd_64_32_long): Likewise.
7689 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7690 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7691 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7692 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7693 (cortex_a8_neon_shift_1): Likewise.
7694 (cortex_a8_neon_shift_2): Likewise.
7695 (cortex_a8_neon_shift_3): Likewise.
7696 (cortex_a8_neon_vshl_ddd): Likewise.
7697 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7698 (cortex_a8_neon_vsra_vrsra): Likewise.
7699 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7700 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7701 (cortex_a8_neon_fp_vsum): Likewise.
7702 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7703 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7704 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7705 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7706 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7707 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7708 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7709 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7710 (cortex_a8_neon_bp_simple): Likewise.
7711 (cortex_a8_neon_bp_2cycle): Likewise.
7712 (cortex_a8_neon_bp_3cycle): Likewise.
7713 (cortex_a8_neon_ldr): Likewise.
7714 (cortex_a8_neon_str): Likewise.
7715 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7716 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7717 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7718 (cortex_a8_neon_vld2_4_regs): Likewise.
7719 (cortex_a8_neon_vld3_vld4): Likewise.
7720 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7721 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7722 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7723 (cortex_a8_neon_vst3_vst4): Likewise.
7724 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7725 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7726 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7727 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7728 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7729 (cortex_a8_neon_mcr): Likewise.
7730 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7731 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7732 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7734 (ca9_neon_mrrc): Likewise.
7735 (cortex_a9_neon_int_1): Likewise.
7736 (cortex_a9_neon_int_2): Likewise.
7737 (cortex_a9_neon_int_3): Likewise.
7738 (cortex_a9_neon_int_4): Likewise.
7739 (cortex_a9_neon_int_5): Likewise.
7740 (cortex_a9_neon_vqneg_vqabs): Likewise.
7741 (cortex_a9_neon_vmov): Likewise.
7742 (cortex_a9_neon_vaba): Likewise.
7743 (cortex_a9_neon_vaba_qqq): Likewise.
7744 (cortex_a9_neon_vsma): Likewise.
7745 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7746 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7747 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7748 long_scalar): Likewise.
7749 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7750 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7751 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7752 long_scalar_qdd_64_32_long): Likewise.
7753 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7754 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7755 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7756 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7757 (cortex_a9_neon_shift_1): Likewise.
7758 (cortex_a9_neon_shift_2): Likewise.
7759 (cortex_a9_neon_shift_3): Likewise.
7760 (cortex_a9_neon_vshl_ddd): Likewise.
7761 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7762 (cortex_a9_neon_vsra_vrsra): Likewise.
7763 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7764 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7765 (cortex_a9_neon_fp_vsum): Likewise.
7766 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7767 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7768 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7769 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7770 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7771 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7772 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7773 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7774 (cortex_a9_neon_bp_simple): Likewise.
7775 (cortex_a9_neon_bp_2cycle): Likewise.
7776 (cortex_a9_neon_bp_3cycle): Likewise.
7777 (cortex_a9_neon_ldr): Likewise.
7778 (cortex_a9_neon_str): Likewise.
7779 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7780 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7781 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7782 (cortex_a9_neon_vld2_4_regs): Likewise.
7783 (cortex_a9_neon_vld3_vld4): Likewise.
7784 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7785 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7786 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7787 (cortex_a9_neon_vst3_vst4): Likewise.
7788 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7789 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7790 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7791 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7792 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7793 (cortex_a9_neon_mcr): Likewise.
7794 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7795 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7796 (cortex_a9_fps): Likewise.
7797 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7799 (cortex_m4_fmuls): Likewise.
7800 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7802 (cortex_r4_mrc): Likewise.
7803 * config/arm/iterators.md: Update comment referring to neon_type.
7804 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7805 (iwmmxt_movsi_insn): Likewise.
7806 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7808 (pj4_core_to_vfp): Likewise.
7809 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7811 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7812 (vfp_fstore): Likewise.
7813 * doc/md.texi: Change references to neon_type to refer to type.
7815 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7817 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7819 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7821 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7824 2013-09-04 Easwaran Raman <eraman@google.com>
7827 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7828 (build_and_add_sum): Use it.
7829 (appears_later_in_bb): Simplify code.
7831 2013-09-04 Teresa Johnson <tejohnson@google.com>
7833 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7835 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7837 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7839 2013-09-04 Jan Hubicka <jh@suse.cz>
7841 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7842 * ipa-devirt.c: Include diganostic.h
7843 (odr_type_d): Add types and types_set.
7844 (hash_type_name): Work for types with vtables during LTO.
7845 (odr_hasher::remove): Fix comment; destroy types_set.
7846 (add_type_duplicate): New function,
7847 (get_odr_type): Use it.
7848 (dump_type_inheritance_graph): Dump type duplicates.
7849 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7851 * tree.c (types_same_for_odr): Give exact answers on types with
7854 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7856 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7857 explaining their differences.
7859 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7861 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7863 2013-09-03 Jeff Law <law@redhat.com>
7865 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7866 when not threading through a joiner block. Pass joiner/no joiner
7867 state to register_jump_thread.
7868 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7869 state from argument rather than implying on path length.
7870 Dump the entire jump thread path into debugging dump.
7871 * tree-flow.h (register_jump_thread): Update prototype.
7873 2013-08-29 Xinliang David Li <davidxl@google.com>
7875 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7876 Remove a trivial gcc_assert.
7878 2013-08-29 Xinliang David Li <davidxl@google.com>
7880 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7881 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7882 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7883 Delay base decl alignment adjustment.
7884 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7885 * tree-vectorizer.h: New data structure.
7886 (set_dr_misalignment): New function.
7887 (dr_misalignment): Ditto.
7888 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7889 (vectorizable_load): Ditto.
7890 (ensure_base_align): New function.
7891 (vectorize_loops): Add dbg_cnt support.
7892 (execute_vect_slp): Ditto.
7893 * dbgcnt.def: New debug counter.
7894 * Makefile: New dependency.
7896 2013-09-03 Meador Inge <meadori@codesourcery.com>
7900 2013-08-30 Meador Inge <meadori@codesourcery.com>
7902 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7904 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7906 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7907 function descriptor.
7909 2013-09-03 Richard Biener <rguenther@suse.de>
7911 * tree-affine.c (add_elt_to_tree): Fix association issue,
7912 avoid useless converts and make sure to always return a
7913 properly typed result.
7915 2013-09-03 Richard Biener <rguenther@suse.de>
7918 * fold-const.c (negate_expr_p): Fix division case.
7919 (negate_expr): Likewise.
7921 2013-09-03 Richard Biener <rguenther@suse.de>
7924 * tree-streamer-out.c: Include tm.h.
7925 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7927 2013-09-03 Jan Hubicka <jh@suse.cz>
7929 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7931 2013-09-03 Alan Modra <amodra@gmail.com>
7933 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7934 multilibs to big-endian target and vice versa.
7935 * config/rs6000/t-linux64: Use := assignment on all vars.
7936 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7937 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7938 * config/rs6000/t-linux64le: New file.
7939 * config/rs6000/t-linux64bele: New file.
7940 * config/rs6000/t-linux64lebe: New file.
7942 2013-09-02 Jan Hubicka <jh@suse.cz>
7944 * ipa-inline-transform.c (inline_transform): Do not
7945 optimize_inline_calls when not optimizing.
7947 2013-09-02 Jan Hubicka <jh@suse.cz>
7949 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7950 duplicated nodes for assembler names.
7951 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7952 hard registers from assembler name hash.
7954 2013-09-02 Jan Hubicka <jh@suse.cz>
7956 * ipa-split.c (execute_split_functions): Split externally visible
7957 functions called once.
7959 2013-09-02 Martin Jambor <mjambor@suse.cz>
7962 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7963 linked list. When finding the correct duplicate, also consider also
7964 the caller in additon to its inlined_to node.
7966 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7968 * config/aarch64/aarch64-simd-builtins.def
7969 (dup_lane_scalar): Remove.
7970 * config/aarch64/aarch64-simd.md
7971 (aarch64_simd_dup): Add 'w->w' alternative.
7972 (aarch64_dup_lane<mode>): Allow for VALL.
7973 (aarch64_dup_lane_scalar<mode>): Remove.
7974 (aarch64_dup_lane_<vswap_width_name><mode>): New.
7975 (aarch64_get_lane_signed<mode>): Add w->w altenative.
7976 (aarch64_get_lane_unsigned<mode>): Likewise.
7977 (aarch64_get_lane<mode>): Likewise.
7978 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
7979 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
7980 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
7981 (VCON): Change container of V2SF.
7982 (vswap_width_name): Likewise.
7983 * config/aarch64/arm_neon.h
7984 (__aarch64_vdup_lane_any): New.
7985 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7986 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
7987 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
7989 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
7992 * expr.c (emit_move_complex): Do not move complex FP values as parts if
7993 the source or the destination is a single hard register.
7995 2013-09-02 Richard Biener <rguenther@suse.de>
7998 * tree-scalar-evolution.c (instantiate_scev_name): Allow
8001 2013-09-02 Richard Biener <rguenther@suse.de>
8003 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
8004 arithmetic to sizetype.
8006 2013-09-02 Bin Cheng <bin.cheng@arm.com>
8008 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
8009 Find auto-increment use both before and after candidate.
8011 2013-09-02 Marek Polacek <polacek@redhat.com>
8013 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
8015 2013-09-01 Jan Hubicka <jh@suse.cz>
8017 * Makefile.in: Add ipa-profile.o
8018 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
8019 * cgraph.c (struct cgraph_propagate_frequency_data,
8020 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
8021 ipa-profile.c; replace cgraph_ by ipa_ prefix.
8022 * cgraph.h (cgraph_propagate_frequency): Remove.
8023 * ipa-inline-analysis.c: Include ipa-utils.h;
8024 drop duplicated cfgloop.h.
8025 (inline_update_callee_summaries): Update.
8026 * ipa-profile.c: New file.
8027 * ipa-utils.h (ipa_propagate_frequency): Declare.
8028 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
8029 data-streamer.h, value-prof.h.
8030 (symtab_remove_unreachable_nodes): Update profile.
8031 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
8032 account_time_size, cmp_counts, dump_histogram,
8033 ipa_profile_generate_summary, ipa_profile_write_summary,
8034 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
8035 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
8036 Move to ipa-profile.c.
8038 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
8040 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
8042 2013-09-01 Jan Hubicka <jh@suse.cz>
8044 * common.opt (fdevirtualize-speculatively): New function.
8045 * invoke.texi (fdevirtualize-speculatively): Document.
8046 * ipa-devirt.c: Include ipa-inline.h
8047 (likely_target_p): New function.
8048 (ipa_devirt): New function.
8049 (gate_ipa_devirt): New function.
8050 (pass_data_ipa_devirt): New static var.
8051 (pass_ipa_devirt): Likewise.
8052 (make_pass_ipa_devirt): New function.
8053 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
8054 (common_handle_option): Disable devirtualization when
8055 value range profiling is available.
8056 * passes.def (pass_ipa_devirt): Add.
8057 * timever.def (TV_IPA_DEVIRT): New timevar.
8058 * tree-pass.h (make_pass_ipa_devirt):
8060 2013-09-01 Iain Sandoe <iain@codesourcery.com>
8062 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
8063 include sanitize(undefined).
8065 2013-08-31 Diego Novillo <dnovillo@google.com>
8067 * Makefile.in (TREE_CORE_H): Define.
8069 (GTFILES): Add tree-core.h.
8070 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
8072 * tree-core.h: New file.
8073 Move all data structures, enum, typedefs, global
8074 declarations and constants from ...
8077 2013-08-31 Jan Hubicka <jh@suse.cz>
8079 * bulitins.c (expand_builtin): Do not early exit for gcov
8080 instrumented functions.
8082 2013-08-31 Marek Polacek <polacek@redhat.com>
8084 * ubsan.c: Include tm_p.h.
8086 2013-08-31 Jan Hubicka <jh@suse.cz>
8088 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
8091 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
8092 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
8093 * tree-cfg.c (verify_gimple_label): ... here.
8094 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
8095 (ipa_merge_profiles): New function.
8096 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
8097 (lto_input_function_body): Likewise.
8098 * ipa-utils.h (ipa_merge_profiles): Declare.
8099 * lto-streamer.h (lto_input_function_body): Update prototype.
8100 (emit_label_in_global_context_p): Remove.
8101 * lto-symtab.c: Include ipa-utils.h
8102 (lto_cgraph_replace_node): Use ipa_merge_profiles.
8104 2013-08-31 Jan Hubicka <jh@suse.cz>
8106 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8108 2013-08-31 Jan Hubicka <jh@suse.cz>
8110 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8112 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8114 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8115 "cmp" RTX before signed_comparison_operator check to account
8118 2013-08-30 Jan Hubicka <jh@suse.cz>
8120 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8121 (detect_type_change_1): Rename to ...
8122 (detect_type_change): ... this one; early return on non-polymorphic
8124 (detect_type_change_ssa): Add comp_type parameter; update
8125 use of detect_type_change.
8126 (compute_complex_assign_jump_func): Add param_type parameter;
8127 update use of detect_type_change_ssa.
8128 (compute_complex_ancestor_jump_func): Likewise.
8129 (ipa_get_callee_param_type): New function.
8130 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8131 update calls to the jump function computation functions.
8133 2013-08-30 Teresa Johnson <tejohnson@google.com>
8134 Steven Bosscher <steven@gcc.gnu.org>
8136 * cfgrtl.c (fixup_new_cold_bb): New routine.
8137 (commit_edge_insertions): Invoke fixup_partitions.
8138 (find_partition_fixes): New routine.
8139 (fixup_partitions): Ditto.
8140 (verify_hot_cold_block_grouping): Update comments.
8141 (rtl_verify_edges): Invoke find_partition_fixes.
8142 (rtl_verify_bb_pointers): Update comments.
8143 (rtl_verify_bb_layout): Ditto.
8144 * basic-block.h (probably_never_executed_edge_p): Declare.
8145 (fixup_partitions): Ditto.
8146 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8147 * bb-reorder.c (sanitize_hot_paths): New function.
8148 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8150 * predict.c (probably_never_executed_edge_p): New routine.
8151 * cfg.c (check_bb_profile): Add partition insanity warnings.
8153 2013-08-30 Meador Inge <meadori@codesourcery.com>
8155 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8157 2013-08-30 Marek Polacek <polacek@redhat.com>
8159 * Makefile.in (ubsan.o): Add.
8160 (c-family/c-ubsan.o): Add.
8161 (builtins.o): Add ubsan.h dependency.
8162 * ubsan.h: New file.
8163 * ubsan.c: New file.
8164 * common.opt: Add -fsanitize=undefined option.
8165 (flag_sanitize): Add variable.
8166 (fsanitize=): Add option. Add Driver.
8167 (fsanitize=thread): Remove option.
8168 (fsanitize=address): Likewise.
8169 (static-libubsan): New option.
8170 * doc/invoke.texi: Document the new flag and -static-libubsan.
8171 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8172 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8173 * builtin-attrs.def (ATTR_COLD): Define.
8174 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8175 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8176 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8177 * flag-types.h (sanitize_code): New enum.
8178 * opts.c (common_handle_option): Parse command line arguments
8179 of -fsanitize=. Add -fsanitize=unreachable option.
8180 * varasm.c (get_variable_section): Adjust.
8181 (assemble_noswitch_variable): Likewise.
8182 (assemble_variable): Likewise.
8183 (output_constant_def_contents): Likewise.
8184 (categorize_decl_for_section): Likewise.
8185 (place_block_symbol): Likewise.
8186 (output_object_block): Likewise.
8187 * builtins.def: Likewise.
8188 * toplev.c (compile_file): Likewise.
8189 (process_options): Likewise.
8190 * cppbuiltin.c: Likewise.
8191 * tsan.c (tsan_pass): Likewise.
8192 (tsan_gate): Likewise.
8193 (tsan_gate_O0): Likewise.
8194 * cfgexpand.c (partition_stack_vars): Likewise.
8195 (expand_stack_vars): Likewise.
8196 (defer_stack_allocation): Likewise.
8197 (expand_used_vars): Likewise.
8198 * cfgcleanup.c (old_insns_match_p): Likewise.
8199 * asan.c (asan_finish_file): Likewise.
8200 (asan_instrument): Likewise.
8201 (gate_asan): Likewise.
8202 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8203 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8204 (asan_global_struct): Use pointer_sized_int_node instead
8205 calling build_nonstandard_integer_type.
8206 (initialize_sanitizer_builtins): Likewise.
8207 (asan_finish_file): Likewise.
8208 * gcc.c: Document %{%:function(args):X}.
8209 (static_spec_functions): Add sanitize.
8210 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8211 store funcval != NULL there.
8212 (do_spec_1): Adjust handle_spec_function caller.
8213 (handle_braces): Allow %:function(args) as condition.
8214 (sanitize_spec_function): New function.
8215 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8216 (LIBUBSAN_SPEC): Likewise.
8217 (LIBUBSAN_EARLY_SPEC): Likewise.
8218 (SANITIZER_SPEC): Handle libubsan.
8219 (SANITIZER_EARLY_SPEC): Likewise.
8220 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8221 instead of fsanitize=address.
8222 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8223 instead of fsanitize=address*.
8224 * builtins.c: Include ubsan.h.
8225 (fold_builtin_0): Instrument __builtin_unreachable.
8226 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8227 instead of flag_asan.
8228 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8229 (pointer_sized_int_node): Define.
8230 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8232 2013-08-30 Mike Stump <mikestump@comcast.net>
8234 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8237 2013-08-29 Jan Hubicka <jh@suse.cz>
8239 * cgraph.c (cgraph_function_body_availability): Handle weakref
8241 * passes.def: Remove pass_fixup_cfg.
8242 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8243 track when we need to remove functions.
8244 (gate_ipa_inline): Execute inlining always; add comment why.
8245 (pass_data_ipa_inline): Remove TODO_remove_functions.
8246 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8247 do not produce summaries.
8248 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8249 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8250 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8251 constant pool and vtable.
8253 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8255 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8256 New arm_neon.h's internal macros to specify 64-bit constants.
8257 Avoid using stdint.h's macros.
8259 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8261 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8262 before applying REG_P.
8264 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8266 PR tree-optimization/58277
8267 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8268 after seeing too many stmts with vdef in between dombb and current
8269 bb, invalidate everything.
8271 2013-08-30 Richard Biener <rguenther@suse.de>
8273 * fold-const.c (fold_single_bit_test): Fix overflow test.
8275 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8277 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8278 and which can live in a register, always retrieve the value on entry.
8279 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8280 passed by invisible reference specially.
8281 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8282 (vt_add_function_parameter): Correctly deal with a parameter passed by
8283 invisible reference.
8285 2013-08-30 Jan Hubicka <jh@suse.cz>
8287 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8289 2013-08-30 Richard Biener <rguenther@suse.de>
8291 PR tree-optimization/58228
8292 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8293 allow invariant loads in nested loop vectorization.
8295 2013-08-30 Richard Biener <rguenther@suse.de>
8297 PR tree-optimization/58223
8298 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8299 (has_anti_or_output_dependence): ... this and adjust to also
8300 look for output dependences.
8301 (mark_nodes_having_upstream_mem_writes): Adjust.
8302 (rdg_flag_uses): Likewise.
8304 2013-08-30 Richard Biener <rguenther@suse.de>
8306 PR tree-optimization/58010
8307 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8308 assert that we have a loop-closed PHI.
8310 2013-08-29 Jan Hubicka <jh@suse.cz>
8312 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8313 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8314 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8315 (lto_free_function_in_decl_state_for_node): New function.
8317 2013-08-29 Xinliang David Li <davidxl@google.com>
8319 * loop-unroll.c (report_unroll_peel): Minor message change.
8320 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8321 Emit alignment peeling message with default -fopt-info.
8322 (vect_loop_versioning): Emit loop version info message.
8323 * tree-vectorizer.c (vectorize_loops): Minor message change.
8324 (execute_vect_slp): Ditto.
8326 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8328 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8329 of the clone from the DECL_NAME of the original function.
8331 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8333 * passes.c (register_pass): Add overload.
8334 * tree-pass.h (register_pass): Forward declare it. Add comment.
8336 2013-08-29 Jan Hubicka <jh@suse.cz>
8338 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8339 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8340 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8341 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8342 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8343 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8344 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8345 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8347 2013-08-29 Teresa Johnson <tejohnson@google.com>
8349 * dumpfile.c (dump_loc): Output column number.
8350 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8351 * doc/invoke.texi: Document optall -fopt-info flag.
8352 * profile.c (read_profile_edge_counts): Use new dump framework.
8353 (compute_branch_probabilities): Ditto.
8354 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8355 when pass not in any opt group.
8356 * pass_manager.h (pass_manager::get_pass_profile): New method.
8357 * value-prof.c (check_counter): Use new dump framework.
8358 (check_ic_target): Ditto.
8359 * coverage.c (get_coverage_counts): Ditto.
8360 (coverage_init): Setup new dump framework.
8362 2013-08-29 Richard Biener <rguenther@suse.de>
8364 PR tree-optimization/58246
8365 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8366 handle the dominance check inside a basic-block.
8368 2013-08-29 Richard Biener <rguenther@suse.de>
8371 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8372 of default defs that appear in abnormal PHI nodes.
8374 2013-08-29 Richard Biener <rguenther@suse.de>
8376 PR tree-optimization/57685
8377 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8378 single-use operands to avoid exponential complexity.
8380 2013-08-28 Dehao Chen <dehao@google.com>
8382 * ipa-inline.c (edge_badness): Fix integer underflow.
8384 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8386 * gtm-builtins.def (_ITM_free): Declare leaf.
8388 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8391 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8392 (*tls_local_dynamic_base_64_largepic): Likewise.
8393 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8394 Remove predicate from call operand.
8395 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8396 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8398 2013-08-28 Jeff Law <law@redhat.com>
8400 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8401 checks for the number of predecessors and successors allowed.
8402 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8403 which require copying a joiner block if there is a request which
8404 is a subpath that requires no joiner block copying.
8406 2013-08-28 Jan Hubicka <jh@suse.cz>
8408 * lto-streamer-out.c (DFS_write_tree_body): Drop
8409 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8410 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8411 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8412 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8413 Do not read DECL_ERROR_ISSUED.
8414 (unpack_ts_decl_with_vis_value_fields): Do not read
8416 (lto_input_ts_binfo_tree_pointers): Do not read
8417 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8418 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8419 write DECL_ERROR_ISSUED..
8420 (pack_ts_decl_with_vis_value_fields): Do not write
8422 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8423 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8424 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8425 * tree.h (tree_decl_common): Update comment.
8426 (DECL_ERROR_ISSUED): Remove.
8428 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8431 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8433 2013-08-28 Jan Hubicka <jh@suse.cz>
8435 * builtins.def (free): Declare leaf.
8437 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8440 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8441 * configure: Regenerate.
8443 2013-08-27 Martin Jambor <mjambor@suse.cz>
8445 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8446 (ipa_ancestor_jf_data): Likewise.
8447 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8448 (ipa_get_jf_pass_through_type_preserved): New function.
8449 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8450 (ipa_get_jf_ancestor_type_preserved): New function.
8451 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8452 (ipa_get_jf_ancestor_result): Likewise.
8453 (propagate_vals_accross_pass_through): Use
8454 ipa_get_jf_pass_through_result to do all the value mappings.
8455 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8456 type_preserved flag.
8457 (ipa_set_jf_cst_copy): New function.
8458 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8459 (ipa_set_jf_arith_pass_through): Likewise.
8460 (ipa_set_ancestor_jf): Likewise.
8461 (compute_complex_assign_jump_func): Set type_preserved instead of
8463 (ipa_compute_jump_functions_for_edge): Likewise.
8464 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8465 (update_jump_functions_after_inlining): Update type_preserved.
8466 Explicitely create jump functions when combining one with pass_through.
8467 (ipa_write_jump_function): Stream the type_preserved flags.
8468 (ipa_read_jump_function): Likewise.
8470 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8471 Aldy Hernandez <aldyh@redhat.com>
8473 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8474 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8475 * function.h (struct function): Add has_force_vect_loops and
8477 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8478 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8480 * gimple.def: Update CLAUSES comments.
8481 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8482 (gimple_build_omp_for): Add argument to prototype.
8483 (gimple_omp_for_kind): New.
8484 (gimple_omp_for_set_kind): New.
8485 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8486 GOVD_DATA_SHARE_CLASS.
8487 (enum omp_region_type): Add ORT_SIMD.
8488 (gimple_add_tmp_var): Handle ORT_SIMD.
8489 (gimplify_var_or_parm_decl): Same.
8490 (is_gimple_stmt): Same.
8491 (omp_firstprivatize_variable): Same.
8492 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8493 (omp_notice_variable): Handle ORT_SIMD.
8494 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8495 (omp_check_private): Handle ORT_SIMD.
8496 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8498 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8499 Handle OMP_CLAUSE_LASTPRIVATE.
8500 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8502 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8503 (gimplify_expr): Handle OMP_SIMD.
8504 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8505 (expand_GOMP_SIMD_VF): New.
8506 (expand_GOMP_SIMD_LAST_LANE): New.
8507 * internal-fn.def (GOMP_SIMD_LANE): New.
8508 (GOMP_SIMD_VF): New.
8509 (GOMP_SIMD_LAST_LANE): New.
8510 * omp-low.c: Include target.h.
8511 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8513 (check_omp_nesting_restrictions): Same.
8515 (lower_rec_simd_input_clauses): New.
8516 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8518 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8519 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8520 (expand_omp_build_assign): New.
8521 (expand_omp_for_init_counts): New.
8522 (expand_omp_for_init_vars): New.
8523 (extract_omp_for_update_vars): New.
8524 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8525 and rewrite accordingly.
8526 (expand_omp_simd): New.
8527 (expand_omp_for): Use expand_omp_simd.
8528 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8529 (lower_omp_for): Do not lower the body.
8530 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8532 * tree-flow.h (find_omp_clause): Remove prototype.
8533 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8534 forcing vectorization of the loop, or if flag_tree_vectorize.
8535 (gate_tree_if_conversion): Similarly.
8536 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8537 gimple_build_omp_for.
8538 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8539 * tree-parloops (create_parallel_loop): Pass kind argument to
8540 gimple_build_omp_for.
8541 * tree-pretty-print.c (dump_omp_clause): Add cases for
8542 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8543 OMP_CLAUSE__SIMDUID_.
8544 (dump_generic_node): Handle OMP_SIMD.
8545 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8546 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8547 unroll OMP_SIMD loops here.
8548 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8549 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8551 (vect_analyze_data_refs): Handle simd loops.
8552 * tree-vect-loop.c (vectorizable_live_operation): Handle
8554 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8555 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8556 (vectorizable_load): Same.
8557 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8558 (struct simduid_to_vf): New.
8559 (simduid_to_vf::hash): New.
8560 (simduid_to-vf::equal): New.
8561 (struct simd_array_to_simduid): New.
8562 (simd_array_to_simduid::hash): New.
8563 (simd_array_to_simduid::equal): New.
8564 (adjust_simduid_builtins): New.
8565 (struct note_simd_array_uses_struct): New.
8566 (note_simd_array_uses_cb): New.
8567 (note_simd_array_uses): New.
8568 (vectorize_loops): Handle simd hints and adjust simd builtins
8570 * tree-vectorizer.h (struct _stmt_vec_info): Add
8571 simd_lane_access_p field.
8572 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8573 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8574 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8575 (omp_clause_code_name): Same.
8576 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8577 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8578 * tree.def (OMP_SIMD): New entry.
8579 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8580 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8581 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8582 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8583 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8584 (OMP_CLAUSE_LINEAR_STEP): New.
8585 (OMP_CLAUSE_SAFELEN_EXPR): New.
8586 (OMP_CLAUSE__SIMDUID__DECL): New.
8587 (find_omp_clause): New prototype.
8589 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8591 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8592 Haswell processor detection.
8594 2013-08-27 Christian Widmer <shadow@umbrox.de>
8597 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8598 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8599 AVX2 capable processors.
8601 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8603 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8604 of vget_low_* with implementations in terms of other intrinsics.
8606 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8609 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8610 values to -1, 0 and 1.
8612 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8614 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8615 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8616 (<optab><mode>3_insn): Remove.
8617 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8618 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8619 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8620 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8621 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8622 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8623 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8624 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8625 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8626 (ror<mode>3_insn): Likewise.
8627 * config/aarch64/predicates.md (aarch64_simd_register): New.
8629 2013-08-27 Richard Biener <rguenther@suse.de>
8631 PR tree-optimization/57521
8632 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8633 one edge is non-critical.
8634 (find_phi_replacement_condition): Make sure to use a non-critical
8635 edge. Cleanup and remove old bug workarounds.
8636 (bb_postdominates_preds): Remove.
8637 (if_convertible_loop_p_1): Do not compute post-dominators.
8638 (combine_blocks): Do not free post-dominators.
8639 (main_tree_if_conversion): Likewise.
8640 (pass_data_if_conversion): Add TODO_verify_ssa.
8642 2013-08-27 DJ Delorie <dj@redhat.com>
8644 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8646 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8648 * function.c (assign_parm_find_data_types): Set passed_mode and
8649 nominal_mode to the TYPE_MODE of nominal_type for the built
8650 pointer type in case of the struct-pass-by-reference.
8652 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8654 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8655 (UINT16_TYPE): Change default to "unsigned int".
8657 * config/avr/avr.opt (mfract-convert-truncate): New option.
8658 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8659 is set, round negative fractional integers according to n1169
8660 when converting to integer types.
8662 2013-08-26 Jan Hubicka <jh@suse.cz>
8664 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8665 methods can not be called indirectly when their address is not taken.
8667 2013-08-26 Jan Hubicka <jh@suse.cz>
8669 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8672 2013-08-26 Jan Hubicka <jh@suse.cz>
8674 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8677 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8679 * reload.c (find_valid_class): Allow classes that do not include
8680 FIRST_PSEUDO_REGISTER - 1.
8682 2013-08-26 Jan Hubicka <jh@suse.cz>
8684 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8685 fix edge count/frequency when speculation failed; fix type check
8686 for the direct call.
8688 2013-08-26 Jan Hubicka <jh@suse.cz>
8690 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8692 2013-08-26 Jan Hubicka <jh@suse.cz>
8694 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8695 to be changed by edge redirection.
8697 2013-08-26 Jan Hubicka <jh@suse.cz>
8699 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8700 formating; add sanity check.
8701 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8702 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8703 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8704 (resolve_noninline_speculation): Update callee keys, too.
8706 2013-08-26 Jan Hubicka <jh@suse.cz>
8708 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8709 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8711 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8713 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8716 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8718 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8719 (x86_64_elf_select_section): Ditto.
8720 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8722 (ix86_pass_by_reference): Ditto.
8723 (output_set_got): Ditto.
8724 (ix86_unary_operator_ok): Ditto.
8725 (ix86_expand_builtin): Ditto.
8727 2013-08-23 Jan Hubicka <jh@suse.cz>
8729 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8731 2013-08-23 Jan Hubicka <jh@suse.cz>
8733 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8734 (tree_decl_with_vis): Add FINAL field.
8736 2013-08-23 Jeff Law <law@redhat.com>
8738 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8739 the debugging dump when the expression is fully redundant.
8741 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8743 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8744 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8745 * pretty-print.c (pp_formatted_text_data): Likewise.
8746 (pp_write_text_to_stream): Likewise.
8747 (pp_write_text_as_dot_label_to_stream): Likewise.
8748 (pp_append_r): Likewise.
8749 (pp_format): Likewise.
8750 (pp_flush): Likewise.
8751 (pp_clear_output_area): Likewise.
8752 (pp_append_text): Likewise.
8753 (pp_formatted_text): Likewise.
8754 (pp_remaining_character_count_for_line): Likewise.
8755 (pp_newline): Likewise.
8756 (pp_character): Likewise.
8757 (output_buffer::~output_buffer): Define.
8758 (pretty_printer::~pretty_printer): Destruct output buffer.
8759 * pretty-print.h (output_buffer::~output_buffer): Declare.
8760 (pretty_printer::~pretty_printer): Declare virtual.
8762 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8765 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8766 HOST_WIDE_INT_M1U): New macros.
8767 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8768 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8769 unsigned -1 for lshift.
8770 * cse.c (cse_insn): Likewise.
8771 * double-int.c (rshift_double, lshift_double): Likewise.
8772 * builtins.c (fold_builtin_bitop): Likewise.
8773 * combine.c (force_to_mode): Likewise.
8774 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8775 * simplify-rtx.c (simplify_const_unary_operation,
8776 simplify_const_binary_operation): Likewise.
8777 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8778 check_va_list_escapes): Likewise.
8779 * rtlanal.c (nonzero_bits1): Likewise.
8780 * expmed.c (expand_smod_pow2): Likewise.
8781 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8783 2013-08-23 Jan Hubicka <jh@suse.cz>
8785 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8786 as having address taken.
8788 2013-08-23 Jan Hubicka <jh@suse.cz>
8790 * ipa-utils.h (method_class_type): Declare.
8791 * ipa-devirt.c (method_class_type): Export.
8793 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8794 do not walk base classes of anonymous types.
8796 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8798 PR rtl-optimization/58220
8800 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8801 handle SEQUENCE insns properly.
8803 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8805 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8807 (pp_newline_and_indent): Likewise.
8808 (pp_separate_with): Likewise.
8809 * pretty-print.c (pp_newline_and_flush): Define.
8810 (pp_newline_and_indent): Likewise.
8811 (pp_separate_with): Likewise.
8813 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8816 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8817 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8819 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8821 * config/i386/predicates.md (ext_sse_reg_operand): New.
8822 * config/i386/i386.md (*movti_internal): Use
8823 predicate to determine if EVEX is needed.
8824 (*movsi_internal): Ditto.
8825 (*movdf_internal): Ditto.
8826 (*movsf_internal): Ditto.
8827 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8829 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8831 PR tree-optimization/58209
8832 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8833 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8834 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8835 emit POINTER_PLUS_EXPR.
8836 (create_tailcall_accumulator): For pointer result type accumulate in
8839 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8841 * configure.ac: Add backslashes missing from the last change.
8842 * configure: Regenerate.
8844 2013-08-22 Jan Hubicka <jh@susue.cz>
8846 * ipa.c (function_and_variable_visibility): First remember function
8847 was global and then make it local.
8849 2013-08-22 Julian Brown <julian@codesourcery.com>
8851 * configure.ac: Add aarch64 to list of arches which use "nop" in
8853 * configure: Regenerate.
8855 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8857 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8858 gnu_libc_has_function.
8859 * config/s390/tpf.h: Likewise.
8861 2013-08-22 Jan Hubicka <jh@susue.cz>
8863 * timevar.c (validate_phases): Add cast.
8865 2013-08-22 Jan Hubicka <jh@susue.cz>
8867 * timevar.c (validate_phases): Use size_t for memory.
8868 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8870 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8872 * pretty-print.h (output_buffer::output_buffer): Declare.
8873 (pretty_printer::pretty_printer): Likewise.
8874 (pp_construct): Remove.
8875 * pretty-print.c (output_buffer::output_buffer): Define.
8876 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8877 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8878 (print_gimple_expr): Likewise.
8879 (print_gimple_seq): Likewise.
8880 (gimple_dump_bb): Likewise.
8881 * sched-vis.c (dump_value_slim): Likewise.
8882 (dump_insn_slim): Likewise.
8883 (dump_rtl_slim): Likewise.
8884 (str_pattern_slim): Likewise.
8885 * tree-mudflap.c (mf_varname_tree): Likewise.
8886 * graph.c (print_graph_cfg): Likewise.
8887 (start_graph_dump): Likewise.
8888 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8890 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8891 pretty printer initialization.
8892 * coretypes.h (diagnostic_context): Remove superflous type alias
8894 (pretty_printer): Likewise. Declare directly as a class.
8895 (pretty_print_info): Remove declaration as class.
8896 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8897 and pp_clear_output_area.
8898 (asan_add_global): Likewise.
8900 2013-08-22 Jan Hubicka <jh@suse.cz>
8902 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8903 * ipa-utils.h (update_type_inheritance_graph): Declare.
8904 (possible_polymorphic_call_target_p): Declare.
8905 (possible_polymorphic_call_target_p): New.
8906 * ipa-devirt.c: Update toplevel comments.
8907 (cached_polymorphic_call_targets): Move up.
8908 (odr_type_d): Move ID down.
8909 (polymorphic_type_binfo_p): Update comment.
8910 (odr_hasher::remove): Likewise;
8911 (get_odr_type): Set anonymous_namespace.
8912 (dump_odr_type): Dump it.
8913 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8914 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8915 (record_binfo): Add comment.
8916 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8918 (devirt_node_removal_hook): Do not iCE when cache is freed.
8919 (possible_polymorphic_call_target_p): New predicate.
8920 (update_type_inheritance_graph): New function.
8922 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8923 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8924 Sergey Lega <sergey.s.lega@intel.com>
8925 Anna Tikhonova <anna.tikhonova@intel.com>
8926 Ilya Tocar <ilya.tocar@intel.com>
8927 Andrey Turetskiy <andrey.turetskiy@intel.com>
8928 Ilya Verbin <ilya.verbin@intel.com>
8929 Kirill Yukhin <kirill.yukhin@intel.com>
8930 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8932 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8933 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8934 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8935 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8936 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8937 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8938 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8939 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8940 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8941 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8942 OPT_mavx512pf, OPT_mavx512er cases.
8943 * config/i386/constraints.md (v): New constraint.
8944 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8945 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8946 (bit_AVX512CD): New.
8947 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8948 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8949 * config/i386/i386-c.c (ix86_target_macros_internal):
8950 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8952 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8953 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8954 * config/i386/i386.c (regclass_map, dbx_register_map)
8955 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8956 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8957 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8959 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8960 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8961 -mavx512pf options. Fix formatting.
8962 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8963 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8964 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8965 -mavx512cd, -mavx512pf options.
8966 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8967 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8968 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8970 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8971 (ix86_set_reg_reg_cost): Ditto.
8972 (x86_order_regs_for_local_alloc): Ditto.
8973 (MAX_VECT_LEN): Extend to 64-byte.
8974 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
8975 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
8976 (TARGET_AVX512ER, TARGET_AVX512CD): New.
8977 (BIGGEST_ALIGNMENT): Extend to 512-bits.
8978 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
8979 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
8980 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
8981 (SSE_REG_MODE_P): Support new modes.
8982 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
8983 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
8984 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
8985 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
8986 (REG_CLASS_CONTENTS): Add new registers.
8987 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
8988 (EXT_REX_SSE_REGNO_P): New.
8989 (HI_REGISTER_NAMES): Add new registers.
8990 * config/i386/i386.md: Define constants for new registers.
8991 (mode): Add new 512-bit modes.
8992 (prefix): Support evex prefix.
8993 (isa): Support avx512f, noavx512f, fma_avx512f.
8994 (ssemodesuffix): Add new 512-bit modes.
8996 (*movxi_internal_avx512f): Ditto.
8997 (*movdi_internal): Replace constraint "x" with the new constraint "v".
8999 (*movsi_internal): Likewise.
9000 (*movdf_internal): Likewise.
9001 (*movsf_internal): Likewise.
9002 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
9003 (<code><mode>3): Likewise.
9004 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
9006 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
9007 with the new constraint "v".
9008 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
9010 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
9011 with the new constraint "v".
9012 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
9013 (<sse2>_storedqu<avxsizesuffix>): Likewise.
9014 (*<plusminus_insn><mode>3): Likewise.
9015 (<sse>_vm<plusminus_insn><mode>3): Likewise.
9016 (*mul<mode>3): Likewise.
9017 (<sse>_vmmul<mode>3): Likewise.
9018 (<sse>_div<mode>3): Likewise.
9019 (<sse>_vmdiv<mode>3): Likewise.
9020 (<sse>_sqrt<mode>2): Likewise.
9021 (<sse>_vmsqrt<mode>2): Likewise.
9022 (*<code><mode>3_finite): Likewise.
9023 (*<code><mode>3) <smaxmin>: Likewise.
9024 (<sse>_vm<code><mode>3): Likewise.
9025 (*<code><mode>3) <any_logic>: Likewise.
9026 (*fma_fmadd_<mode>): Likewise.
9027 (*fma_fmsub_<mode>): Likewise.
9028 (*fma_fnmadd_<mode>): Likewise.
9029 (*fma_fnmsub_<mode>): Likewise.
9030 (*fma_fmaddsub_<mode>): Likewise.
9031 (*fma_fmsubadd_<mode>): Likewise.
9032 (*fmai_fmadd_<mode>): Likewise.
9033 (*fmai_fmsub_<mode>): Likewise.
9034 (*fmai_fnmadd_<mode>): Likewise.
9035 (*fmai_fnmsub_<mode>): Likewise.
9036 (sse_cvtsi2ss): Likewise.
9037 (sse_cvtsi2ssq): Likewise.
9038 (sse_cvtss2si): Likewise.
9039 (sse_cvtss2si_2): Likewise.
9040 (sse_cvtss2siq): Likewise.
9041 (sse_cvtss2siq_2): Likewise.
9042 (sse_cvttss2si): Likewise.
9043 (sse_cvtss2siq_2): Likewise.
9044 (float<sseintvecmodelower><mode>2): Likewise.
9045 (sse2_cvtsd2si_2): Likewise.
9046 (sse2_cvtsd2siq_2): Likewise.
9047 (*<plusminus_insn><mode>3): Likewise.
9048 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
9049 (*<sse4_1_avx2>_mul<mode>3): Likewise.
9050 (ashr<mode>3): Likewise.
9051 (<shift_insn><mode>3): Likewise.
9052 (avx2_<code><mode>3): Likewise.
9053 (*avx2_<code><mode>3): Likewise.
9054 (*andnot<mode>3): Likewise.
9055 (*<code><mode>3) <any_logic>: Likewise.
9056 (abs<mode>2): Likewise.
9057 (avx2_permvar<mode>): Likewise.
9058 (avx2_perm<mode>_1): Likewise.
9059 (*avx_vpermilp<mode>): Likewise.
9060 (avx_vpermilvar<mode>3): Likewise.
9061 (avx2_ashrv<mode>): Likewise.
9062 (avx2_<shift_insn>v<mode>): Likewise.
9063 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
9065 * doc/rtl.texi: Document XImode.
9067 2013-08-21 Jeff Law <law@redhat.com>
9069 * tree-flow.h (register_jump_thread): Pass vector of edges
9070 instead of each important edge.
9071 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
9072 thread path into a vector and pass that to register_jump_thread.
9073 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
9074 passed in edge vector to the current 3-edge form.
9077 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9079 * dce.c (fini_dce): Call df_analyze again just in case
9080 delete_unmarked_insns removed anything.
9082 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
9084 * reload.h (struct reg_equivs): Rename to ..
9085 (struct reg_equivs_s): .. this.
9087 2013-08-20 Martin Liska <marxin.liska@gmail.com>
9089 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
9091 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9093 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
9095 2013-08-21 Jeff Law <law@redhat.com>
9097 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
9098 simplify assignments too. If the RHS collapses to a singleton
9099 range, then return the value for the range.
9101 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
9103 * config/i386/sse.md (V16): Rename to...
9105 (mov<mode>): Update iterator name.
9106 (*mov<mode>_internal): Ditto.
9107 (push<mode>1): Ditto.
9108 (movmisalign<mode>): Ditto.
9110 2013-08-20 Jan Hubicka <jh@suse.cz>
9113 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9114 entry for direct edges.
9115 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9117 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9119 Revert my last two changes, r201865 and r201864:
9122 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9124 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9125 instances can own GC refs.
9127 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9128 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9129 (gcc::context::gt_pch_nx): Likewise.
9130 (gcc::context::gt_pch_nx): Likewise.
9131 * ggc.h (gt_ggc_mx <T>): New.
9132 (gt_pch_nx_with_op <T>): New.
9133 (gt_pch_nx <T>): New.
9134 * passes.c (opt_pass::gt_ggc_mx): New.
9135 (opt_pass::gt_pch_nx): New.
9136 (opt_pass::gt_pch_nx_with_op): New.
9137 (pass_manager::gt_ggc_mx): New.
9138 (pass_manager::gt_pch_nx): New.
9139 (pass_manager::gt_pch_nx_with_op): New.
9140 (pass_manager::operator new): Use
9141 ggc_internal_cleared_alloc_stat rather than xcalloc.
9142 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9143 (pass_manager::gt_ggc_mx): New.
9144 (pass_manager::gt_pch_nx): New.
9145 (pass_manager::gt_pch_nx_with_op): New.
9146 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9147 (opt_pass::operator new): New.
9148 (opt_pass::gt_ggc_mx): New.
9149 (opt_pass::gt_pch_nx): New.
9150 (opt_pass::gt_pch_nx_with_op): New.
9153 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9155 * Makefile.in (GTFILES): Add context.h.
9156 * context.c (gcc::context::operator new): New.
9157 (gcc::context::gt_ggc_mx): New.
9158 (gcc::context::gt_pch_nx): New.
9159 (gcc::context::gt_pch_nx): New.
9160 * context.h (gcc::context): Add GTY((user)) marking.
9161 (gcc::context::operator new): New.
9162 (gcc::context::gt_ggc_mx): New.
9163 (gcc::context::gt_pch_nx): New.
9164 (gcc::context::gt_pch_nx): New.
9165 (g): Add GTY marking.
9166 (gt_ggc_mx (gcc::context *)): New.
9167 (gt_pch_nx (gcc::context *)): New.
9168 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9169 void *cookie)): New.
9170 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9172 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9174 * dce.c (fini_dce): Call df_analyze again just in case
9175 delete_unmarked_insns removed anything.
9177 2013-08-20 Teresa Johnson <tejohnson@google.com>
9179 PR rtl-optimizations/57451
9180 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9181 from crossing split section boundaries.
9183 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9185 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9186 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9187 with MULTLIB_DEFAULTS.
9189 2013-08-20 Nick Clifton <nickc@redhat.com>
9191 * target.def (narrow_volatile_bitfield): Note that the default
9192 value is false, not !TARGET_STRICT_ALIGN.
9193 * doc/tm.texi: Regenerate.
9195 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9197 Fix LIB_SPEC for systems without libpthread.
9199 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9200 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9202 * config/i386/linux-common.h: Likewise.
9203 * config/mips/linux-common.h: Likewise.
9205 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9207 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9210 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9212 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9213 instances can own GC refs.
9215 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9216 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9217 (gcc::context::gt_pch_nx): Likewise.
9218 (gcc::context::gt_pch_nx): Likewise.
9219 * ggc.h (gt_ggc_mx <T>): New.
9220 (gt_pch_nx_with_op <T>): New.
9221 (gt_pch_nx <T>): New.
9222 * passes.c (opt_pass::gt_ggc_mx): New.
9223 (opt_pass::gt_pch_nx): New.
9224 (opt_pass::gt_pch_nx_with_op): New.
9225 (pass_manager::gt_ggc_mx): New.
9226 (pass_manager::gt_pch_nx): New.
9227 (pass_manager::gt_pch_nx_with_op): New.
9228 (pass_manager::operator new): Use
9229 ggc_internal_cleared_alloc_stat rather than xcalloc.
9230 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9231 (pass_manager::gt_ggc_mx): New.
9232 (pass_manager::gt_pch_nx): New.
9233 (pass_manager::gt_pch_nx_with_op): New.
9234 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9235 (opt_pass::operator new): New.
9236 (opt_pass::gt_ggc_mx): New.
9237 (opt_pass::gt_pch_nx): New.
9238 (opt_pass::gt_pch_nx_with_op): New.
9240 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9242 * Makefile.in (GTFILES): Add context.h.
9243 * context.c (gcc::context::operator new): New.
9244 (gcc::context::gt_ggc_mx): New.
9245 (gcc::context::gt_pch_nx): New.
9246 (gcc::context::gt_pch_nx): New.
9247 * context.h (gcc::context): Add GTY((user)) marking.
9248 (gcc::context::operator new): New.
9249 (gcc::context::gt_ggc_mx): New.
9250 (gcc::context::gt_pch_nx): New.
9251 (gcc::context::gt_pch_nx): New.
9252 (g): Add GTY marking.
9253 (gt_ggc_mx (gcc::context *)): New.
9254 (gt_pch_nx (gcc::context *)): New.
9255 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9256 void *cookie)): New.
9257 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9259 2013-08-20 Alan Modra <amodra@gmail.com>
9262 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9263 (rs6000_emit_epilogue): Likewise.
9265 2013-08-19 Dehao Chen <dehao@google.com>
9267 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9269 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9270 Jakub Jelinek <jakub@redhat.com>
9272 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9273 (BUILT_IN_FABSD64): Likewise.
9274 (BUILT_IN_FABSD128): Likewise.
9275 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9276 (fold_builtin_1): Likewise.
9277 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9278 destination and source operands.
9279 (*abstd2_fpr): Likewise.
9280 (*nabstd2_fpr): Likewise.
9282 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9284 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9287 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9289 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9292 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9294 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9295 * builtins.c (default_libc_has_function): New.
9296 (gnu_libc_has_function): Ditto.
9297 (no_c99_libc_has_function): Ditto.
9298 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9299 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9300 (fold_builtin_sincos): Likewise.
9301 (fold_builtin_cexp): Likewise.
9302 * builtins.def (DEF_C94_BUILTIN): Likewise.
9303 (DEF_C99_BUILTIN): Likewise.
9304 (DEF_C99_C90RES_BUILTIN): Likewise.
9305 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9306 definitions to using this define.
9307 * config/darwin-protos.h (darwin_libc_has_function): New.
9308 * config/darwin.c (darwin_libc_has_function): Ditto.
9309 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9310 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9311 * config/darwin.h: Ditto.
9312 * config/elfos.h: Ditto.
9313 * config/freebsd.h: Ditto.
9314 * config/i386/cygming.h: Ditto.
9315 * config/i386/djgpp.h: Ditto.
9316 * config/i386/i386-interix.h: Ditto.
9317 * config/microblaze/microblaze.h: Ditto.
9318 * config/mmix/mmix.h: Ditto.
9319 * config/gnu-user.h: Ditto.
9320 * config/ia64/hpux.h: Ditto.
9321 * config/pa/pa-hpux.h: Ditto.
9322 * config/pdp11/pdp11.h: Ditto.
9323 * config/picochip/picochip.h: Ditto.
9324 * config/linux.h: Ditto.
9325 * config/netbsd.h: Ditto.
9326 * config/openbsd.h: Ditto.
9327 * config/rs6000/aix43.h: Ditto.
9328 * config/rs6000/aix51.h: Ditto.
9329 * config/rs6000/aix52.h: Ditto.
9330 * config/rs6000/aix53.h: Ditto.
9331 * config/rs6000/aix61.h: Ditto.
9332 * config/rs6000/darwin.h: Ditto.
9333 * config/rs6000/linux.h: Ditto.
9334 * config/rs6000/linux64.h: Ditto.
9335 * config/s390/tpf.h: Ditto.
9336 * config/sol2-10.h: Ditto.
9337 * config/sol2.h: Ditto.
9338 * config/vms/vms.h: Ditto.
9339 * config/vxworks.h: Ditto.
9340 * config/linux-android.c (linux_android_libc_has_function):
9341 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9342 * config/linux-protos.h (linux_android_libc_has_function):
9344 * config/i386/i386.c (ix86_libc_has_function): New.
9345 * config/i386/i386-protos.h
9346 (ix86_libc_has_function): New declaration.
9347 * config/i386/i386.md
9348 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9349 ("isinf<mode>2): Likewise.
9350 * convert.c (convert_to_integer): Using new target hook
9351 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9352 TARGET_C99_FUNCTIONS.
9353 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9354 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9355 * coretypes.h (function_class): New enum for different
9356 classes of functions.
9357 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9358 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9359 (TARGET_HAS_SINCOS): Likewise.
9360 (TARGET_LIBC_HAS_FUNCTION): New.
9361 * doc/tm.texi: Regenerated.
9362 * targhooks.h (default_libc_has_function): New declaration.
9363 (no_c99_libc_has_function): Ditto.
9364 (gnu_libc_has_function): Ditto.
9365 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9366 and TARGET_HAS_SINCOS.
9368 2013-08-18 Jan Hubicka <jh@suse.cz>
9370 * Makeifle-in (ipa-devirt.o): New.
9371 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9372 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9373 (analyze_functions): Look into possible targets of polymorphic call.
9374 * dumpfile.c (dump_files): Add type-inheritance dump.
9375 * dumpfile.h (TDI_inheritance): New.
9376 * ipa-devirt.c: New file.
9377 * ipa-utils.h (odr_type_d): Forward declare.
9378 (odr_type): New type.
9379 (build_type_inheritance_graph): Declare.
9380 (possible_polymorphic_call_targets): Declare and introduce inline
9381 variant when only edge is pased.
9382 (dump_possible_polymorphic_call_targets): Likewise.
9383 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9384 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9385 (types_same_for_odr): ... here.
9386 * tree.h (type_in_anonymous_namespace_p): Declare.
9388 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9390 PR tree-optimization/58006
9391 * tree-parloops.c (take_address_of): Don't ICE if get_name
9393 (eliminate_local_variables_stmt): Remove clobber stmts.
9395 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9397 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9398 error message is issued for an alias to undefined symbol.
9400 2013-08-18 Jan Hubicka <jh@suse.cz>
9402 * cgraph.c (cgraph_create_indirect_edge): Discover
9403 polymorphic calls and record basic info into indirect_info.
9404 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9405 devirtualization, ignore objc function calls.
9406 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9407 call with no parm index info.
9408 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9409 * tree.c (virtual_method_call_p): New function.
9410 * tree.h (virtual_method_call_p): Declare.
9412 2013-08-16 Jan Hubicka <jh@suse.cz>
9415 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9417 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9419 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9422 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9424 * gengtype.c (type_for_name): Add special-case support for
9425 locating types within the "gcc::" namespace.
9426 (open_base_files): Emit a "using namespace gcc" directive.
9428 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9431 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9432 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9434 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9435 array instead of each individual operand as a separate argument.
9436 (emit_fusion_gpr_load): Likewise.
9437 (expand_fusion_gpr_load): Add new function declaration.
9439 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9440 signature to have the operands passed as an array, instead of as
9441 separate arguments. Allow ZERO_EXTEND to be in the memory
9442 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9443 depend on the register live/dead flags when peepholes are run.
9444 (expand_fusion_gpr_load): New function to be called from the
9445 peephole2 pass, to change the register that addis sets to be the
9447 (emit_fusion_gpr_load): Change the calling signature to have the
9448 operands passed as an array, instead of as separate arguments.
9449 Allow ZERO_EXTEND to be in the memory address, and also
9450 SIGN_EXTEND if -mpower8-fusion-sign.
9452 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9454 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9455 adjust the register addis loads up in the peephole2 pass. Do not
9456 depend on the register live/dead state when the peephole pass is done.
9458 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9460 * gengtype.c (create_user_defined_type): Ensure that the kind
9461 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9462 declaration is seen before the GTY((user)) marking.
9464 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9467 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9469 2013-08-16 Jan Hubicka <jh@suse.cz>
9471 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9472 arugment expected_type.
9473 (gimple_fold_call): Use it.
9474 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9475 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9476 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9477 (try_make_edge_direct_virtual_call): Likewise.
9478 * tree.c (obj_type_ref_class): New.
9479 * tree.h (obj_type_ref_class): Use it.
9481 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9483 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9484 (rtl_slim_pp): Likewise.
9485 (init_rtl_slim_pretty_print): Likewise.
9486 (dump_value_slim): Don't call it. Use local pretty printer.
9487 (dump_insn_slim): Likewise.
9488 (dump_rtl_slim): Likewise.
9489 (str_pattern_slim): Likewise.
9490 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9493 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9495 PR tree-optimization/58164
9496 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9497 walk gimple_goto_dest of GIMPLE_GOTO.
9499 PR tree-optimization/58165
9500 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9501 bi_call must be the last stmt in a bb, don't split_block, instead
9502 use fallthru edge from it and give up if there is none.
9503 Release conds vector when returning early.
9505 2013-08-14 Xinliang David Li <davidxl@google.com>
9507 * config/i386/i386.c (ix86_option_override_internal):
9508 Remove unused variable and field.
9510 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9513 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9514 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9515 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9516 and Linux, correct BLKmode alignment when 128-bit alignment is
9517 required and compatibility flag is not set.
9518 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9519 for zero-size arguments when compatibility flag is not set.
9521 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9523 PR tree-optimization/58145
9524 * tree-sra.c (build_ref_for_offset): If prev_base has
9525 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9527 2013-08-14 Xinliang David Li <davidxl@google.com>
9529 * config/i386/i386.c (ix86_option_override_internal):
9530 Fix uninitialized variable error.
9532 2013-08-14 Xinliang David Li <davidxl@google.com>
9534 * config/i386/i386.opt: Define two new options.
9535 * config/i386/x86-tune.def: Add arch selector field in macros.
9536 * config/i386/i386.h: Adjust macro definition.
9537 * config/i386/i386.c (ix86_option_override_internal):
9539 (parse_mtune_ctrl_str): New function.
9540 (set_ix86_tune_features): New function.
9541 (ix86_function_specific_restore): Call the new helper function.
9543 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9545 PR rtl-optimization/57662
9546 * sel-sched.c (code_motion_process_successors): When the current insn
9547 is removed after the recursive traversal, break from the loop.
9548 Add comments and debug printouts.
9550 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9551 Alexandre Oliva <aoliva@redhat.com>
9554 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9555 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9556 there also UNSPEC_PLTOFF.
9558 2013-08-14 Marek Polacek <polacek@redhat.com>
9560 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9561 NUM_CONDITIONS bit positions.
9563 2013-08-13 Cary Coutant <ccoutant@google.com>
9565 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9566 (attr_checksum): Hash vector contents instead of pointer.
9567 (attr_checksum_ordered): Likewise.
9569 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9571 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9572 when Pmode != word_mode. Add length_address attribute.
9573 (sse3_monitor_<mode>): Merge from sse3_monitor and
9574 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9575 Pmode != word_mode. Update insn length attribute.
9576 * config/i386/i386.c (ix86_option_override_internal): Update
9577 ix86_gen_monitor selection for merged sse3_monitor insn.
9579 2013-08-13 Julian Brown <julian@codesourcery.com>
9581 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9582 perform invalid legitimization on greater-than-word-size modes for
9585 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9587 * ira.c (setup_class_translate_array): Use aclass instead of cl
9588 for classes not fully covered by allocno classes.
9590 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9592 PR tree-optimization/57661
9593 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9594 * tree-inline.c (tree_function_versioning): Initialize it.
9595 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9596 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9597 that is not being copied.
9600 * asan.c (instrument_strlen_call): Fix typo in comment.
9601 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9603 2013-08-13 Steve Ellcey <sellcey@mips.com>
9605 * config/mips/mips.md (prefetch): Use lw instead of ld on
9606 loongson in 32bit mode.
9608 2013-08-13 Nick Clifton <nickc@redhat.com>
9610 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9612 2013-08-13 Jan Hubicka <jh@suse.cz>
9614 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9615 introduced edge; fix typo in sanity check.
9616 (cgraph_resolve_speculation): Export; improve diagnostic.
9617 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9618 speculation at type mismatch.
9619 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9620 (cgraph_resolve_speculation): Declare.
9621 (symtab_can_be_discarded): New function.
9622 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9623 * ipa-inline-transform.c (speculation_removed): New global var.
9624 (clone_inlined_nodes): See if speculation can be removed.
9625 (inline_call): If speculations was removed, we growths may not match.
9626 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9627 (speculation_useful_p): New function.
9628 (resolve_noninline_speculation): New function.
9629 (inline_small_functions): Resolve useless speculations.
9630 * ipa-inline.h (speculation_useful_p): Declare
9631 * ipa.c (can_replace_by_local_alias): Simplify.
9632 (ipa_profile): Produce speculative calls in non-lto, too;
9633 add simple cost model; produce local aliases.
9635 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9637 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9640 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9642 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9644 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9646 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9647 iterator instead of X87MODEF.
9649 2013-08-12 Perez Read <netfirewall@gmail.com>
9652 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9653 operand 0 for intel asm alternative.
9654 (*movabs<mode>_2): Ditto for operand 1.
9656 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9658 * config/aarch64/arm_none.h
9659 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9661 2013-08-12 Nick Clifton <nickc@redhat.com>
9663 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9665 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9667 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9668 expand for QI/HImode operand to produce more effictive code for
9669 unsigned char(short) --> float(double) conversion.
9671 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9673 * doc/invoke.texi: Mention that -ftls-model does not force the final
9676 2013-08-12 Marek Polacek <polacek@redhat.com>
9677 Marc Glisse <marc.glisse@inria.fr>
9679 PR tree-optimization/57980
9680 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9681 when creating -1 constant.
9683 2013-08-10 Jan Hubicka <jh@suse.cz>
9685 Workaround binutils PR14342.
9686 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9687 (gimple_init_edge_profiler): Likewise.
9688 (gimple_gen_ic_func_profiler): Likewise.
9690 2013-08-09 Jan Hubicka <jh@suse.cz>
9692 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9694 2013-08-09 Xinliang David Li <davidxl@google.com>
9696 * config/i386/stringop.def: New file.
9697 * config/i386/stringop.opt: New file.
9698 * config/i386/i386-opts.h: Include stringopt.def.
9699 * config/i386/i386.opt: Include stringopt.opt.
9700 * config/i386/i386.c (ix86_option_override_internal):
9701 Override default size based stringop inline strategies with options.
9702 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9705 2013-08-09 Jan Hubicka <jh@suse.cz>
9707 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9709 2013-08-09 Jan Hubicka <jh@suse.cz>
9711 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9713 (dump_cgraph_node): Dump profile-id.
9714 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9715 and common_target_probability.
9716 * lto-cgraph.c (lto_output_edge): Stream common targets.
9717 (lto_output_node): Stream profile ids.
9718 (input_node): Stream profile ids.
9719 (input_edge): Stream common targets.
9720 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9721 * ipa.c: Include value-prof.h
9722 (ipa_profile_generate_summary): Turn indirect call statement histograms
9723 into common targets.
9724 (ipa_profile): Turn common targets into speculative edges.
9726 2013-08-09 Jan Hubicka <jh@suse.cz>
9728 * cgraph.h (cgraph_node): Add profile_id.
9729 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9730 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9731 (del_node_map): Update.
9732 (find_func_by_funcdef_no): Replace by ...
9733 (find_func_by_profile_id): ... this one.
9734 (gimple_ic_transform): Do not remove useful histograms when
9735 speculation is not done; dump info when indirect call removal
9737 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9738 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9739 (__gcov_indirect_call_profiler_v2): .. this one.
9740 * profile.h (init_node_map): Update.
9741 * coverage.c (coverage_compute_profile_id): New function.
9742 * coverage.h (coverage_compute_profile_id): Declare.
9743 * tree-profile.c (init_ic_make_global_vars): Make
9744 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9745 (gimple_init_edge_profiler): Update prototype of
9746 __gcov_indirect_call_profiler.
9747 (gimple_gen_ic_func_profiler): Simplify.
9748 (tree_profiling): Use init_node_map
9750 2013-08-09 Jan Hubicka <jh@suse.cz>
9752 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9753 non-speculative refs.
9754 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9755 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9756 (cgraph_set_call_stmt): Likewise.
9757 (cgraph_create_edge_1): Fix release checking compilatoin;
9759 (cgraph_free_edge): Free indirect info.
9760 (cgraph_turn_edge_to_speculative): New function.
9761 (cgraph_speculative_call_info): New function.
9762 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9763 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9764 (dump_cgraph_node): Dump speculation.
9765 (verify_edge_count_and_frequency): Accept speculative edges.
9766 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9767 (verify_cgraph_node): Handle speculation.
9768 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9769 (cgraph_set_call_stmt): Update prototype.
9770 (cgraph_make_edge_direct): Update prototype.
9771 (cgraph_speculative_call_info): Declare.
9772 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9773 to change; update call of ipa_find_references.
9774 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9775 lto_stmt_uid and speculative flags.
9776 (ipa_dump_references): Dump speculation.
9777 (ipa_clone_references): Clone speculative flag.
9778 (ipa_clone_referring): Likewise.
9779 (ipa_clone_ref): New function.
9780 (ipa_find_reference): Look into lto_stmt_uids
9781 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9782 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9783 (ipa_find_reference): Update declaration.
9784 (ipa_clone_ref): Declare.
9785 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9786 stream speculative flag.
9787 (lto_output_ref): Stream statements uids and speculation.
9788 (input_ref): Likewise.
9789 (input_edge): Stream speuclation.
9790 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9791 (cgraph_set_call_stmt_including_clones): Handle speculation.
9792 * ipa-inline.c (heap_edge_removal_hook): New function.
9793 (inline_small_functions): Register it.
9794 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9795 also initialize refs.
9796 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9798 (try_make_edge_direct_simple_call): Likewise.
9799 (try_make_edge_direct_simple_call): Likewise.
9800 (update_indirect_edges_after_inlining): Likewise.
9801 (remove_described_reference): Look proper lto_stmt_uid.
9802 (propagate_controlled_uses): Likewise.
9803 (propagate_controlled_uses): Liekwise.
9804 * tree-inline.c (copy_bb): Copy speculative edges.
9805 (redirect_all_calls): New function.
9806 (copy_cfg_body): Do redirection after loop info is updated.
9807 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9809 2013-08-09 Jan Hubicka <jh@suse.cz>
9811 * lto-streamer-out.c (output_function): Renumber PHIs.
9812 * lto-streamer-in.c (input_function): Likewise.
9814 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9816 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9817 (get_lane_unsigned): Likewise.
9818 (dup_lane_scalar): Likewise.
9819 (get_lane): enable for VALL.
9820 * config/aarch64/aarch64-simd.md
9821 (aarch64_dup_lane_scalar<mode>): Remove.
9822 (aarch64_get_lane_signed<mode>): Likewise.
9823 (aarch64_get_lane_unsigned<mode>): Likewise.
9824 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9825 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9826 (aarch64_get_lane<mode>): Enable for all vector modes.
9827 (aarch64_get_lanedi): Remove misleading constraints.
9828 * config/aarch64/arm_neon.h
9829 (__aarch64_vget_lane_any): Define.
9830 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9831 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9832 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9833 * config/aarch64/iterators.md (VDQQH): New.
9837 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9839 * configure.ac: Add GAS check for LEON instructions on SPARC.
9840 * configure: Regenerate.
9841 * config.in: Likewise.
9842 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9844 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9845 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9847 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9848 (AS_LEON_FLAG): New macro.
9849 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9850 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9851 Deal with LEON and LEON3 for the memory model.
9852 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9853 (atomic_compare_and_swap<mode>_1): Likewise.
9854 (*atomic_compare_and_swap<mode>_1): Likewise.
9856 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9858 * config/arm/neon.md (vcond): Fix floating-point vector
9859 comparisons against 0.
9861 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9863 * lra-constraints.c (emit_spill_move): Remove assert.
9864 (process_alt_operands): Add more debugging
9865 output. Increase reject for spilling into memory. Decrease
9866 reject for reloading scratch.
9867 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9869 2013-08-08 Steve Ellcey <sellcey@mips.com>
9871 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9872 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9873 micromips incompatible. Add nan2008.
9874 (MULTILIB_DIRNAMES): Add nan2008.
9875 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9876 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9877 and micromips incompatible. Add nan2008.
9878 (MULTILIB_DIRNAMES): Add nan2008.
9879 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9881 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9883 PR rtl-optimization/58079
9884 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9885 simplify_comparison has widened a comparison with an integer.
9887 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9889 * config/arm/neon.md (movmisalign<mode>): Disable when we
9890 don't allow unaligned accesses.
9891 (*movmisalign<mode>_neon_store): Likewise.
9892 (*movmisalign<mode>_neon_load): Likewise.
9893 (*movmisalign<mode>_neon_store): Likewise.
9894 (*movmisalign<mode>_neon_load): Likewise.
9896 2013-08-08 Jan Hubicka <jh@suse.cz>
9898 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9899 (make_pass_rebuild_cgraph_edges): Also clear references.
9900 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9901 * ipa-inline-transform.c (inline_transform): Remove all references
9903 * cgraphunit.c (expand_function): Remove all references after
9905 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9906 (ipa_find_reference): Rewrite to iterator.
9907 (remove_stmt_references): Likewise.
9908 (ipa_clear_stmts_in_references): New function.
9909 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9910 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9912 * ipa-split.c (split_function): Remove references in split function.
9914 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9917 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9918 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9920 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9923 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9924 suggested mode for the assignment isn't compatible with the
9927 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9930 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9932 2013-08-07 Xinliang David Li <davidxl@google.com>
9934 * config/i386/i386.opt: New option -mtune-ctrl=.
9935 * config/i386/x86-tune.def: New file.
9936 * config/i386/i386.h: include x86-tune.def.
9937 * config/i386/i386.c (ix86_option_override_internal):
9938 Parsing -mtune-ctrl= option and set tune features.
9940 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9943 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9944 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9947 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9949 * rtl.h (update_alignments): Declare.
9950 * final.c (grow_label_align): New function extracted from...
9951 (shorten_branches): ...here. Call it.
9952 (update_alignments): New function.
9953 * reorg.c (sibling_labels): New variable.
9954 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9955 the new label along with it onto the sibling_labels vector.
9956 (fill_simple_delay_slots): Adjust call to get_label_before.
9957 (fill_slots_from_thread): Likewise.
9958 (relax_delay_slots): Likewise.
9959 (make_return_insns): Likewise.
9960 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9962 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9964 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9965 document its semantics.
9966 (diagnostic_report_diagnostic): Adjust accordingly.
9968 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9970 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9971 (sparc_option_override): ...and port to new C++ pass API.
9972 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
9974 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
9976 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
9978 2013-08-06 Caroline Tice <cmtice@google.com>
9980 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
9981 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
9982 * tree-pass.h: Add pass_vtable_verify.
9983 * varasm.c (assemble_variable): Add code to properly set the comdat
9984 section and name for the .vtable_map_vars section.
9985 (assemble_vtyv_preinit_initializer): New function.
9986 (default_sectin_type_flags): Make sure .vtable_map_vars section has
9988 * output.h: Add function decl for assemble_vtv_preinit_initializer.
9989 * vtable-verify.c: New file.
9990 * vtable-verify.h: New file.
9991 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
9992 initialiation levels.
9993 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
9994 * passes.def: Insert pass_vtable_verify.
9995 * aclocal.m4: Reorder includes.
9996 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
9997 -fvtv-counts options.
9998 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
9999 as appropriate, if -fvtable-verify=... is used.
10000 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
10001 -fvtable-verify=... is used.
10002 * Makefile.in (OBJS): Add vtable-verify.o to list.
10003 (vtable-verify.o): Add new build rule.
10004 (GTFILES): Add vtable-verify.c to list.
10005 * common.opt (fvtable-verify=): New flag.
10006 (vtv_priority): Values for fvtable-verify= flag.
10007 (fvtv-counts): New flag.
10008 (fvtv-debug): New flag.
10009 * tree.h (save_vtable_map_decl): New extern function decl.
10011 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10013 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
10014 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
10015 (pass_data_rl78_devirt): ...new pass_data instance and...
10016 (make_pass_rl78_devirt): ...new function.
10017 (rl78_asm_file_start): Port pass registration to new C++ API.
10019 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10021 * coretypes.h (rtl_opt_pass): Add.
10022 (gcc::context): Add.
10023 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
10024 (epiphany_init): Port to new C++ pass API.
10025 (epiphany_optimize_mode_switching): Likewise.
10026 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
10027 (pass_manager::get_pass_mode_switching): New.
10028 (pass_manager::get_pass_peephole2): New.
10029 * mode-switching.c (pass_mode_switching): Add clone method.
10030 * recog.c (pass_peephole2): Add clone method.
10031 (pass_split_all_insns): Add clone method.
10033 2013-08-06 David Malcolm <dmalcolm@redhat.com>
10035 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
10036 (mips_option_override): ...here, porting to new C++ API for passes.
10038 2013-08-06 Jan Hubicka <jh@suse.cz>
10040 * cgraph.c (cgraph_get_body): New function based on lto.c
10042 * cgraph.h (cgraph_get_body): Declare.
10043 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
10045 * cgraphunit.c (expand_function): Get body prior expanding.
10046 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
10047 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
10049 * passes.c (do_per_function_toporder): Get body.
10050 * tree-inline.c (expand_call_inline): Get body prior inlining it.
10051 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
10053 2013-08-06 Martin Jambor <mjambor@suse.cz>
10056 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
10057 is not re-finalized. Rename second parameter to no_collect.
10059 2013-08-06 Martin Jambor <mjambor@suse.cz>
10061 PR middle-end/58041
10062 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
10063 MEM_REF has proper alignment information.
10065 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
10068 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
10070 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
10071 rtx (*) (rtx, ...) with insn_gen_fn.
10072 * genoutput.c (output_insn_data): Cast gen_? function pointers to
10073 insn_gen_fn::stored_funcptr. Add initializer braces.
10075 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10077 Rewrite how instances of passes are cloned to remove assumptions
10078 about their sizes (thus allowing pass subclasses to have
10079 additional data fields, albeit non-GC-managed ones at this point).
10081 * passes.c (make_pass_instance): Now that passes have clone
10082 methods, rewrite this function to eliminate XNEW and memcpy
10083 calls that used hardcoded sizes. Since this function no longer
10084 creates pass instances, rename it to...
10085 (add_pass_instance): ...this. Document the old way that passes were
10086 numbered and flagged, and rework this function to continue using it.
10087 (next_pass_1): Add an initial_pass argument for use by
10089 (position_pass): When adding multiple instances of a pass, use
10090 the pass's clone method, rather than relying on the XNEW/memcpy
10091 within the former make_pass_instance (now add_pass_instance).
10092 (pass_manager::pass_manager): When invoking next_pass_1, also supply
10093 the initial instance of the current pass within the pass manager.
10095 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10097 This is the automated part of the conversion of passes from C
10098 structs to C++ classes.
10100 Patch autogenerated by refactor_passes.py from
10101 https://github.com/davidmalcolm/gcc-refactoring-scripts
10102 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
10104 * asan.c (pass_asan): Convert from a global struct to a subclass of
10105 gimple_opt_pass along with...
10106 (pass_data_asan): ...new pass_data instance and...
10107 (make_pass_asan): ...new function.
10108 (pass_asan_O0): Convert from a global struct to a subclass of
10109 gimple_opt_pass along with...
10110 (pass_data_asan_O0): ...new pass_data instance and...
10111 (make_pass_asan_O0): ...new function.
10112 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10113 subclass of rtl_opt_pass along with...
10114 (pass_data_inc_dec): ...new pass_data instance and...
10115 (make_pass_inc_dec): ...new function.
10116 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10117 a subclass of rtl_opt_pass along with...
10118 (pass_data_reorder_blocks): ...new pass_data instance and...
10119 (make_pass_reorder_blocks): ...new function.
10120 (pass_duplicate_computed_gotos): Convert from a global struct to a
10121 subclass of rtl_opt_pass along with...
10122 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10123 (make_pass_duplicate_computed_gotos): ...new function.
10124 (pass_partition_blocks): Convert from a global struct to a subclass of
10125 rtl_opt_pass along with...
10126 (pass_data_partition_blocks): ...new pass_data instance and...
10127 (make_pass_partition_blocks): ...new function.
10128 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10129 struct to a subclass of rtl_opt_pass along with...
10130 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10132 (make_pass_branch_target_load_optimize1): ...new function.
10133 (pass_branch_target_load_optimize2): Convert from a global struct to a
10134 subclass of rtl_opt_pass along with...
10135 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10137 (make_pass_branch_target_load_optimize2): ...new function.
10138 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10139 of rtl_opt_pass along with...
10140 (pass_data_jump): ...new pass_data instance and...
10141 (make_pass_jump): ...new function.
10142 (pass_jump2): Convert from a global struct to a subclass of
10143 rtl_opt_pass along with...
10144 (pass_data_jump2): ...new pass_data instance and...
10145 (make_pass_jump2): ...new function.
10146 * cfgexpand.c (pass_expand): Convert from a global struct to a
10147 subclass of rtl_opt_pass along with...
10148 (pass_data_expand): ...new pass_data instance and...
10149 (make_pass_expand): ...new function.
10150 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10151 of rtl_opt_pass along with...
10152 (pass_data_free_cfg): ...new pass_data instance and...
10153 (make_pass_free_cfg): ...new function.
10154 (pass_into_cfg_layout_mode): Convert from a global struct to a
10155 subclass of rtl_opt_pass along with...
10156 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10157 (make_pass_into_cfg_layout_mode): ...new function.
10158 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10159 subclass of rtl_opt_pass along with...
10160 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10161 (make_pass_outof_cfg_layout_mode): ...new function.
10162 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10163 struct to a subclass of gimple_opt_pass along with...
10164 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10165 (make_pass_build_cgraph_edges): ...new function.
10166 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10167 subclass of gimple_opt_pass along with...
10168 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10169 (make_pass_rebuild_cgraph_edges): ...new function.
10170 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10171 subclass of gimple_opt_pass along with...
10172 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10174 (make_pass_remove_cgraph_callee_edges): ...new function.
10175 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10176 struct to a subclass of rtl_opt_pass along with...
10177 (pass_data_stack_adjustments): ...new pass_data instance and...
10178 (make_pass_stack_adjustments): ...new function.
10179 * combine.c (pass_combine): Convert from a global struct to a subclass
10180 of rtl_opt_pass along with...
10181 (pass_data_combine): ...new pass_data instance and...
10182 (make_pass_combine): ...new function.
10183 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10184 global struct to a subclass of rtl_opt_pass along with...
10185 (pass_data_compare_elim_after_reload): ...new pass_data instance
10187 (make_pass_compare_elim_after_reload): ...new function.
10188 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10189 of rtl_opt_pass along with...
10190 (pass_data_rtl_cprop): ...new pass_data instance and...
10191 (make_pass_rtl_cprop): ...new function.
10192 * cse.c (pass_cse): Convert from a global struct to a subclass of
10193 rtl_opt_pass along with...
10194 (pass_data_cse): ...new pass_data instance and...
10195 (make_pass_cse): ...new function.
10196 (pass_cse2): Convert from a global struct to a subclass of
10197 rtl_opt_pass along with...
10198 (pass_data_cse2): ...new pass_data instance and...
10199 (make_pass_cse2): ...new function.
10200 (pass_cse_after_global_opts): Convert from a global struct to a
10201 subclass of rtl_opt_pass along with...
10202 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10203 (make_pass_cse_after_global_opts): ...new function.
10204 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10205 of rtl_opt_pass along with...
10206 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10207 (make_pass_ud_rtl_dce): ...new function.
10208 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10209 rtl_opt_pass along with...
10210 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10211 (make_pass_fast_rtl_dce): ...new function.
10212 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10213 a subclass of rtl_opt_pass along with...
10214 (pass_data_df_initialize_opt): ...new pass_data instance and...
10215 (make_pass_df_initialize_opt): ...new function.
10216 (pass_df_initialize_no_opt): Convert from a global struct to a
10217 subclass of rtl_opt_pass along with...
10218 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10219 (make_pass_df_initialize_no_opt): ...new function.
10220 (pass_df_finish): Convert from a global struct to a subclass of
10221 rtl_opt_pass along with...
10222 (pass_data_df_finish): ...new pass_data instance and...
10223 (make_pass_df_finish): ...new function.
10224 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10225 rtl_opt_pass along with...
10226 (pass_data_rtl_dse1): ...new pass_data instance and...
10227 (make_pass_rtl_dse1): ...new function.
10228 (pass_rtl_dse2): Convert from a global struct to a subclass of
10229 rtl_opt_pass along with...
10230 (pass_data_rtl_dse2): ...new pass_data instance and...
10231 (make_pass_rtl_dse2): ...new function.
10232 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10233 subclass of rtl_opt_pass along with...
10234 (pass_data_dwarf2_frame): ...new pass_data instance and...
10235 (make_pass_dwarf2_frame): ...new function.
10236 * except.c (pass_set_nothrow_function_flags): Convert from a global
10237 struct to a subclass of rtl_opt_pass along with...
10238 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10240 (make_pass_set_nothrow_function_flags): ...new function.
10241 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10242 subclass of rtl_opt_pass along with...
10243 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10245 (make_pass_convert_to_eh_region_ranges): ...new function.
10246 * final.c (pass_compute_alignments): Convert from a global struct to a
10247 subclass of rtl_opt_pass along with...
10248 (pass_data_compute_alignments): ...new pass_data instance and...
10249 (make_pass_compute_alignments): ...new function.
10250 (pass_final): Convert from a global struct to a subclass of
10251 rtl_opt_pass along with...
10252 (pass_data_final): ...new pass_data instance and...
10253 (make_pass_final): ...new function.
10254 (pass_shorten_branches): Convert from a global struct to a subclass of
10255 rtl_opt_pass along with...
10256 (pass_data_shorten_branches): ...new pass_data instance and...
10257 (make_pass_shorten_branches): ...new function.
10258 (pass_clean_state): Convert from a global struct to a subclass of
10259 rtl_opt_pass along with...
10260 (pass_data_clean_state): ...new pass_data instance and...
10261 (make_pass_clean_state): ...new function.
10262 * function.c (pass_instantiate_virtual_regs): Convert from a global
10263 struct to a subclass of rtl_opt_pass along with...
10264 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10265 (make_pass_instantiate_virtual_regs): ...new function.
10266 (pass_leaf_regs): Convert from a global struct to a subclass of
10267 rtl_opt_pass along with...
10268 (pass_data_leaf_regs): ...new pass_data instance and...
10269 (make_pass_leaf_regs): ...new function.
10270 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10271 subclass of rtl_opt_pass along with...
10272 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10274 (make_pass_thread_prologue_and_epilogue): ...new function.
10275 (pass_match_asm_constraints): Convert from a global struct to a
10276 subclass of rtl_opt_pass along with...
10277 (pass_data_match_asm_constraints): ...new pass_data instance and...
10278 (make_pass_match_asm_constraints): ...new function.
10279 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10280 subclass of rtl_opt_pass along with...
10281 (pass_data_rtl_fwprop): ...new pass_data instance and...
10282 (make_pass_rtl_fwprop): ...new function.
10283 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10284 rtl_opt_pass along with...
10285 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10286 (make_pass_rtl_fwprop_addr): ...new function.
10287 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10288 rtl_opt_pass along with...
10289 (pass_data_rtl_pre): ...new pass_data instance and...
10290 (make_pass_rtl_pre): ...new function.
10291 (pass_rtl_hoist): Convert from a global struct to a subclass of
10292 rtl_opt_pass along with...
10293 (pass_data_rtl_hoist): ...new pass_data instance and...
10294 (make_pass_rtl_hoist): ...new function.
10295 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10296 subclass of gimple_opt_pass along with...
10297 (pass_data_lower_cf): ...new pass_data instance and...
10298 (make_pass_lower_cf): ...new function.
10299 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10300 from a global struct to a subclass of gimple_opt_pass along with...
10301 (pass_data_strength_reduction): ...new pass_data instance and...
10302 (make_pass_strength_reduction): ...new function.
10303 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10304 of rtl_opt_pass along with...
10305 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10306 (make_pass_rtl_ifcvt): ...new function.
10307 (pass_if_after_combine): Convert from a global struct to a subclass of
10308 rtl_opt_pass along with...
10309 (pass_data_if_after_combine): ...new pass_data instance and...
10310 (make_pass_if_after_combine): ...new function.
10311 (pass_if_after_reload): Convert from a global struct to a subclass of
10312 rtl_opt_pass along with...
10313 (pass_data_if_after_reload): ...new pass_data instance and...
10314 (make_pass_if_after_reload): ...new function.
10315 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10316 a subclass of rtl_opt_pass along with...
10317 (pass_data_initialize_regs): ...new pass_data instance and...
10318 (make_pass_initialize_regs): ...new function.
10319 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10320 of ipa_opt_pass_d along with...
10321 (pass_data_ipa_cp): ...new pass_data instance and...
10322 (make_pass_ipa_cp): ...new function.
10323 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10324 global struct to a subclass of gimple_opt_pass along with...
10325 (pass_data_inline_parameters): ...new pass_data instance and...
10326 (make_pass_inline_parameters): ...new function.
10327 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10328 subclass of gimple_opt_pass along with...
10329 (pass_data_early_inline): ...new pass_data instance and...
10330 (make_pass_early_inline): ...new function.
10331 (pass_ipa_inline): Convert from a global struct to a subclass of
10332 ipa_opt_pass_d along with...
10333 (pass_data_ipa_inline): ...new pass_data instance and...
10334 (make_pass_ipa_inline): ...new function.
10335 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10336 struct to a subclass of gimple_opt_pass along with...
10337 (pass_data_local_pure_const): ...new pass_data instance and...
10338 (make_pass_local_pure_const): ...new function.
10339 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10340 ipa_opt_pass_d along with...
10341 (pass_data_ipa_pure_const): ...new pass_data instance and...
10342 (make_pass_ipa_pure_const): ...new function.
10343 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10344 to a subclass of ipa_opt_pass_d along with...
10345 (pass_data_ipa_reference): ...new pass_data instance and...
10346 (make_pass_ipa_reference): ...new function.
10347 * ipa-split.c (pass_split_functions): Convert from a global struct to
10348 a subclass of gimple_opt_pass along with...
10349 (pass_data_split_functions): ...new pass_data instance and...
10350 (make_pass_split_functions): ...new function.
10351 (pass_feedback_split_functions): Convert from a global struct to a
10352 subclass of gimple_opt_pass along with...
10353 (pass_data_feedback_split_functions): ...new pass_data instance and...
10354 (make_pass_feedback_split_functions): ...new function.
10355 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10356 global struct to a subclass of simple_ipa_opt_pass along with...
10357 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10359 (make_pass_ipa_function_and_variable_visibility): ...new function.
10360 (pass_ipa_free_inline_summary): Convert from a global struct to a
10361 subclass of simple_ipa_opt_pass along with...
10362 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10363 (make_pass_ipa_free_inline_summary): ...new function.
10364 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10365 subclass of ipa_opt_pass_d along with...
10366 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10368 (make_pass_ipa_whole_program_visibility): ...new function.
10369 (pass_ipa_profile): Convert from a global struct to a subclass of
10370 ipa_opt_pass_d along with...
10371 (pass_data_ipa_profile): ...new pass_data instance and...
10372 (make_pass_ipa_profile): ...new function.
10373 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10374 ipa_opt_pass_d along with...
10375 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10376 (make_pass_ipa_cdtor_merge): ...new function.
10377 * ira.c (pass_ira): Convert from a global struct to a subclass of
10378 rtl_opt_pass along with...
10379 (pass_data_ira): ...new pass_data instance and...
10380 (make_pass_ira): ...new function.
10381 (pass_reload): Convert from a global struct to a subclass of
10382 rtl_opt_pass along with...
10383 (pass_data_reload): ...new pass_data instance and...
10384 (make_pass_reload): ...new function.
10385 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10386 subclass of rtl_opt_pass along with...
10387 (pass_data_cleanup_barriers): ...new pass_data instance and...
10388 (make_pass_cleanup_barriers): ...new function.
10389 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10390 of rtl_opt_pass along with...
10391 (pass_data_loop2): ...new pass_data instance and...
10392 (make_pass_loop2): ...new function.
10393 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10394 rtl_opt_pass along with...
10395 (pass_data_rtl_loop_init): ...new pass_data instance and...
10396 (make_pass_rtl_loop_init): ...new function.
10397 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10398 rtl_opt_pass along with...
10399 (pass_data_rtl_loop_done): ...new pass_data instance and...
10400 (make_pass_rtl_loop_done): ...new function.
10401 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10402 subclass of rtl_opt_pass along with...
10403 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10404 (make_pass_rtl_move_loop_invariants): ...new function.
10405 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10406 rtl_opt_pass along with...
10407 (pass_data_rtl_unswitch): ...new pass_data instance and...
10408 (make_pass_rtl_unswitch): ...new function.
10409 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10410 subclass of rtl_opt_pass along with...
10411 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10413 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10414 (pass_rtl_doloop): Convert from a global struct to a subclass of
10415 rtl_opt_pass along with...
10416 (pass_data_rtl_doloop): ...new pass_data instance and...
10417 (make_pass_rtl_doloop): ...new function.
10418 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10419 a subclass of rtl_opt_pass along with...
10420 (pass_data_lower_subreg): ...new pass_data instance and...
10421 (make_pass_lower_subreg): ...new function.
10422 (pass_lower_subreg2): Convert from a global struct to a subclass of
10423 rtl_opt_pass along with...
10424 (pass_data_lower_subreg2): ...new pass_data instance and...
10425 (make_pass_lower_subreg2): ...new function.
10426 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10427 struct to a subclass of ipa_opt_pass_d along with...
10428 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10429 (make_pass_ipa_lto_gimple_out): ...new function.
10430 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10431 of ipa_opt_pass_d along with...
10432 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10433 (make_pass_ipa_lto_finish_out): ...new function.
10434 * mode-switching.c (pass_mode_switching): Convert from a global struct
10435 to a subclass of rtl_opt_pass along with...
10436 (pass_data_mode_switching): ...new pass_data instance and...
10437 (make_pass_mode_switching): ...new function.
10438 * modulo-sched.c (pass_sms): Convert from a global struct to a
10439 subclass of rtl_opt_pass along with...
10440 (pass_data_sms): ...new pass_data instance and...
10441 (make_pass_sms): ...new function.
10442 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10443 subclass of gimple_opt_pass along with...
10444 (pass_data_expand_omp): ...new pass_data instance and...
10445 (make_pass_expand_omp): ...new function.
10446 (pass_lower_omp): Convert from a global struct to a subclass of
10447 gimple_opt_pass along with...
10448 (pass_data_lower_omp): ...new pass_data instance and...
10449 (make_pass_lower_omp): ...new function.
10450 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10451 of gimple_opt_pass along with...
10452 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10453 (make_pass_diagnose_omp_blocks): ...new function.
10454 * passes.c (pass_early_local_passes): Convert from a global struct to
10455 a subclass of simple_ipa_opt_pass along with...
10456 (pass_data_early_local_passes): ...new pass_data instance and...
10457 (make_pass_early_local_passes): ...new function.
10458 (pass_all_early_optimizations): Convert from a global struct to a
10459 subclass of gimple_opt_pass along with...
10460 (pass_data_all_early_optimizations): ...new pass_data instance and...
10461 (make_pass_all_early_optimizations): ...new function.
10462 (pass_all_optimizations): Convert from a global struct to a subclass
10463 of gimple_opt_pass along with...
10464 (pass_data_all_optimizations): ...new pass_data instance and...
10465 (make_pass_all_optimizations): ...new function.
10466 (pass_all_optimizations_g): Convert from a global struct to a subclass
10467 of gimple_opt_pass along with...
10468 (pass_data_all_optimizations_g): ...new pass_data instance and...
10469 (make_pass_all_optimizations_g): ...new function.
10470 (pass_rest_of_compilation): Convert from a global struct to a subclass
10471 of rtl_opt_pass along with...
10472 (pass_data_rest_of_compilation): ...new pass_data instance and...
10473 (make_pass_rest_of_compilation): ...new function.
10474 (pass_postreload): Convert from a global struct to a subclass of
10475 rtl_opt_pass along with...
10476 (pass_data_postreload): ...new pass_data instance and...
10477 (make_pass_postreload): ...new function.
10478 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10479 subclass of rtl_opt_pass along with...
10480 (pass_data_gcse2): ...new pass_data instance and...
10481 (make_pass_gcse2): ...new function.
10482 * postreload.c (pass_postreload_cse): Convert from a global struct to
10483 a subclass of rtl_opt_pass along with...
10484 (pass_data_postreload_cse): ...new pass_data instance and...
10485 (make_pass_postreload_cse): ...new function.
10486 * predict.c (pass_profile): Convert from a global struct to a subclass
10487 of gimple_opt_pass along with...
10488 (pass_data_profile): ...new pass_data instance and...
10489 (make_pass_profile): ...new function.
10490 (pass_strip_predict_hints): Convert from a global struct to a subclass
10491 of gimple_opt_pass along with...
10492 (pass_data_strip_predict_hints): ...new pass_data instance and...
10493 (make_pass_strip_predict_hints): ...new function.
10494 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10495 of rtl_opt_pass along with...
10496 (pass_data_peephole2): ...new pass_data instance and...
10497 (make_pass_peephole2): ...new function.
10498 (pass_split_all_insns): Convert from a global struct to a subclass of
10499 rtl_opt_pass along with...
10500 (pass_data_split_all_insns): ...new pass_data instance and...
10501 (make_pass_split_all_insns): ...new function.
10502 (pass_split_after_reload): Convert from a global struct to a subclass
10503 of rtl_opt_pass along with...
10504 (pass_data_split_after_reload): ...new pass_data instance and...
10505 (make_pass_split_after_reload): ...new function.
10506 (pass_split_before_regstack): Convert from a global struct to a
10507 subclass of rtl_opt_pass along with...
10508 (pass_data_split_before_regstack): ...new pass_data instance and...
10509 (make_pass_split_before_regstack): ...new function.
10510 (pass_split_before_sched2): Convert from a global struct to a subclass
10511 of rtl_opt_pass along with...
10512 (pass_data_split_before_sched2): ...new pass_data instance and...
10513 (make_pass_split_before_sched2): ...new function.
10514 (pass_split_for_shorten_branches): Convert from a global struct to a
10515 subclass of rtl_opt_pass along with...
10516 (pass_data_split_for_shorten_branches): ...new pass_data instance
10518 (make_pass_split_for_shorten_branches): ...new function.
10519 * ree.c (pass_ree): Convert from a global struct to a subclass of
10520 rtl_opt_pass along with...
10521 (pass_data_ree): ...new pass_data instance and...
10522 (make_pass_ree): ...new function.
10523 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10524 subclass of rtl_opt_pass along with...
10525 (pass_data_stack_regs): ...new pass_data instance and...
10526 (make_pass_stack_regs): ...new function.
10527 (pass_stack_regs_run): Convert from a global struct to a subclass of
10528 rtl_opt_pass along with...
10529 (pass_data_stack_regs_run): ...new pass_data instance and...
10530 (make_pass_stack_regs_run): ...new function.
10531 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10532 subclass of rtl_opt_pass along with...
10533 (pass_data_cprop_hardreg): ...new pass_data instance and...
10534 (make_pass_cprop_hardreg): ...new function.
10535 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10536 subclass of rtl_opt_pass along with...
10537 (pass_data_reginfo_init): ...new pass_data instance and...
10538 (make_pass_reginfo_init): ...new function.
10539 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10540 of rtl_opt_pass along with...
10541 (pass_data_regmove): ...new pass_data instance and...
10542 (make_pass_regmove): ...new function.
10543 * regrename.c (pass_regrename): Convert from a global struct to a
10544 subclass of rtl_opt_pass along with...
10545 (pass_data_regrename): ...new pass_data instance and...
10546 (make_pass_regrename): ...new function.
10547 * reorg.c (pass_delay_slots): Convert from a global struct to a
10548 subclass of rtl_opt_pass along with...
10549 (pass_data_delay_slots): ...new pass_data instance and...
10550 (make_pass_delay_slots): ...new function.
10551 (pass_machine_reorg): Convert from a global struct to a subclass of
10552 rtl_opt_pass along with...
10553 (pass_data_machine_reorg): ...new pass_data instance and...
10554 (make_pass_machine_reorg): ...new function.
10555 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10556 of rtl_opt_pass along with...
10557 (pass_data_sched): ...new pass_data instance and...
10558 (make_pass_sched): ...new function.
10559 (pass_sched2): Convert from a global struct to a subclass of
10560 rtl_opt_pass along with...
10561 (pass_data_sched2): ...new pass_data instance and...
10562 (make_pass_sched2): ...new function.
10563 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10564 to a subclass of rtl_opt_pass along with...
10565 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10566 (make_pass_stack_ptr_mod): ...new function.
10567 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10568 to a subclass of rtl_opt_pass along with...
10569 (pass_data_rtl_store_motion): ...new pass_data instance and...
10570 (make_pass_rtl_store_motion): ...new function.
10571 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10572 of gimple_opt_pass along with...
10573 (pass_data_tracer): ...new pass_data instance and...
10574 (make_pass_tracer): ...new function.
10575 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10576 to a subclass of gimple_opt_pass along with...
10577 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10578 (make_pass_diagnose_tm_blocks): ...new function.
10579 (pass_lower_tm): Convert from a global struct to a subclass of
10580 gimple_opt_pass along with...
10581 (pass_data_lower_tm): ...new pass_data instance and...
10582 (make_pass_lower_tm): ...new function.
10583 (pass_tm_init): Convert from a global struct to a subclass of
10584 gimple_opt_pass along with...
10585 (pass_data_tm_init): ...new pass_data instance and...
10586 (make_pass_tm_init): ...new function.
10587 (pass_tm_mark): Convert from a global struct to a subclass of
10588 gimple_opt_pass along with...
10589 (pass_data_tm_mark): ...new pass_data instance and...
10590 (make_pass_tm_mark): ...new function.
10591 (pass_tm_edges): Convert from a global struct to a subclass of
10592 gimple_opt_pass along with...
10593 (pass_data_tm_edges): ...new pass_data instance and...
10594 (make_pass_tm_edges): ...new function.
10595 (pass_tm_memopt): Convert from a global struct to a subclass of
10596 gimple_opt_pass along with...
10597 (pass_data_tm_memopt): ...new pass_data instance and...
10598 (make_pass_tm_memopt): ...new function.
10599 (pass_ipa_tm): Convert from a global struct to a subclass of
10600 simple_ipa_opt_pass along with...
10601 (pass_data_ipa_tm): ...new pass_data instance and...
10602 (make_pass_ipa_tm): ...new function.
10603 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10604 subclass of gimple_opt_pass along with...
10605 (pass_data_call_cdce): ...new pass_data instance and...
10606 (make_pass_call_cdce): ...new function.
10607 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10608 subclass of gimple_opt_pass along with...
10609 (pass_data_build_cfg): ...new pass_data instance and...
10610 (make_pass_build_cfg): ...new function.
10611 (pass_split_crit_edges): Convert from a global struct to a subclass of
10612 gimple_opt_pass along with...
10613 (pass_data_split_crit_edges): ...new pass_data instance and...
10614 (make_pass_split_crit_edges): ...new function.
10615 (pass_warn_function_return): Convert from a global struct to a
10616 subclass of gimple_opt_pass along with...
10617 (pass_data_warn_function_return): ...new pass_data instance and...
10618 (make_pass_warn_function_return): ...new function.
10619 (pass_warn_function_noreturn): Convert from a global struct to a
10620 subclass of gimple_opt_pass along with...
10621 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10622 (make_pass_warn_function_noreturn): ...new function.
10623 (pass_warn_unused_result): Convert from a global struct to a subclass
10624 of gimple_opt_pass along with...
10625 (pass_data_warn_unused_result): ...new pass_data instance and...
10626 (make_pass_warn_unused_result): ...new function.
10627 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10628 a subclass of gimple_opt_pass along with...
10629 (pass_data_merge_phi): ...new pass_data instance and...
10630 (make_pass_merge_phi): ...new function.
10631 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10632 a subclass of gimple_opt_pass along with...
10633 (pass_data_lower_complex): ...new pass_data instance and...
10634 (make_pass_lower_complex): ...new function.
10635 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10636 gimple_opt_pass along with...
10637 (pass_data_lower_complex_O0): ...new pass_data instance and...
10638 (make_pass_lower_complex_O0): ...new function.
10639 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10640 subclass of gimple_opt_pass along with...
10641 (pass_data_lower_eh): ...new pass_data instance and...
10642 (make_pass_lower_eh): ...new function.
10643 (pass_refactor_eh): Convert from a global struct to a subclass of
10644 gimple_opt_pass along with...
10645 (pass_data_refactor_eh): ...new pass_data instance and...
10646 (make_pass_refactor_eh): ...new function.
10647 (pass_lower_resx): Convert from a global struct to a subclass of
10648 gimple_opt_pass along with...
10649 (pass_data_lower_resx): ...new pass_data instance and...
10650 (make_pass_lower_resx): ...new function.
10651 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10652 of gimple_opt_pass along with...
10653 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10654 (make_pass_lower_eh_dispatch): ...new function.
10655 (pass_cleanup_eh): Convert from a global struct to a subclass of
10656 gimple_opt_pass along with...
10657 (pass_data_cleanup_eh): ...new pass_data instance and...
10658 (make_pass_cleanup_eh): ...new function.
10659 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10660 to a subclass of simple_ipa_opt_pass along with...
10661 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10662 (make_pass_ipa_lower_emutls): ...new function.
10663 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10664 a subclass of gimple_opt_pass along with...
10665 (pass_data_if_conversion): ...new pass_data instance and...
10666 (make_pass_if_conversion): ...new function.
10667 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10668 subclass of gimple_opt_pass along with...
10669 (pass_data_build_ssa): ...new pass_data instance and...
10670 (make_pass_build_ssa): ...new function.
10671 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10672 global struct to a subclass of gimple_opt_pass along with...
10673 (pass_data_loop_distribution): ...new pass_data instance and...
10674 (make_pass_loop_distribution): ...new function.
10675 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10676 subclass of gimple_opt_pass along with...
10677 (pass_data_mudflap_1): ...new pass_data instance and...
10678 (make_pass_mudflap_1): ...new function.
10679 (pass_mudflap_2): Convert from a global struct to a subclass of
10680 gimple_opt_pass along with...
10681 (pass_data_mudflap_2): ...new pass_data instance and...
10682 (make_pass_mudflap_2): ...new function.
10683 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10684 subclass of gimple_opt_pass along with...
10685 (pass_data_mudflap_1): ...new pass_data instance and...
10686 (make_pass_mudflap_1): ...new function.
10687 (pass_mudflap_2): Convert from a global struct to a subclass of
10688 gimple_opt_pass along with...
10689 (pass_data_mudflap_2): ...new pass_data instance and...
10690 (make_pass_mudflap_2): ...new function.
10691 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10692 gimple_opt_pass along with...
10693 (pass_data_nrv): ...new pass_data instance and...
10694 (make_pass_nrv): ...new function.
10695 (pass_return_slot): Convert from a global struct to a subclass of
10696 gimple_opt_pass along with...
10697 (pass_data_return_slot): ...new pass_data instance and...
10698 (make_pass_return_slot): ...new function.
10699 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10700 to a subclass of gimple_opt_pass along with...
10701 (pass_data_object_sizes): ...new pass_data instance and...
10702 (make_pass_object_sizes): ...new function.
10703 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10704 global struct to a subclass of gimple_opt_pass along with...
10705 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10707 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10708 (pass_fixup_cfg): Convert from a global struct to a subclass of
10709 gimple_opt_pass along with...
10710 (pass_data_fixup_cfg): ...new pass_data instance and...
10711 (make_pass_fixup_cfg): ...new function.
10712 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10713 (make_pass_mudflap_1): ...new function.
10714 (pass_mudflap_2): Replace declaration with that of...
10715 (make_pass_mudflap_2): ...new function.
10716 (pass_asan): Replace declaration with that of...
10717 (make_pass_asan): ...new function.
10718 (pass_asan_O0): Replace declaration with that of...
10719 (make_pass_asan_O0): ...new function.
10720 (pass_tsan): Replace declaration with that of...
10721 (make_pass_tsan): ...new function.
10722 (pass_tsan_O0): Replace declaration with that of...
10723 (make_pass_tsan_O0): ...new function.
10724 (pass_lower_cf): Replace declaration with that of...
10725 (make_pass_lower_cf): ...new function.
10726 (pass_refactor_eh): Replace declaration with that of...
10727 (make_pass_refactor_eh): ...new function.
10728 (pass_lower_eh): Replace declaration with that of...
10729 (make_pass_lower_eh): ...new function.
10730 (pass_lower_eh_dispatch): Replace declaration with that of...
10731 (make_pass_lower_eh_dispatch): ...new function.
10732 (pass_lower_resx): Replace declaration with that of...
10733 (make_pass_lower_resx): ...new function.
10734 (pass_build_cfg): Replace declaration with that of...
10735 (make_pass_build_cfg): ...new function.
10736 (pass_early_tree_profile): Replace declaration with that of...
10737 (make_pass_early_tree_profile): ...new function.
10738 (pass_cleanup_eh): Replace declaration with that of...
10739 (make_pass_cleanup_eh): ...new function.
10740 (pass_sra): Replace declaration with that of...
10741 (make_pass_sra): ...new function.
10742 (pass_sra_early): Replace declaration with that of...
10743 (make_pass_sra_early): ...new function.
10744 (pass_early_ipa_sra): Replace declaration with that of...
10745 (make_pass_early_ipa_sra): ...new function.
10746 (pass_tail_recursion): Replace declaration with that of...
10747 (make_pass_tail_recursion): ...new function.
10748 (pass_tail_calls): Replace declaration with that of...
10749 (make_pass_tail_calls): ...new function.
10750 (pass_tree_loop): Replace declaration with that of...
10751 (make_pass_tree_loop): ...new function.
10752 (pass_tree_loop_init): Replace declaration with that of...
10753 (make_pass_tree_loop_init): ...new function.
10754 (pass_lim): Replace declaration with that of...
10755 (make_pass_lim): ...new function.
10756 (pass_tree_unswitch): Replace declaration with that of...
10757 (make_pass_tree_unswitch): ...new function.
10758 (pass_predcom): Replace declaration with that of...
10759 (make_pass_predcom): ...new function.
10760 (pass_iv_canon): Replace declaration with that of...
10761 (make_pass_iv_canon): ...new function.
10762 (pass_scev_cprop): Replace declaration with that of...
10763 (make_pass_scev_cprop): ...new function.
10764 (pass_empty_loop): Replace declaration with that of...
10765 (make_pass_empty_loop): ...new function.
10766 (pass_record_bounds): Replace declaration with that of...
10767 (make_pass_record_bounds): ...new function.
10768 (pass_graphite): Replace declaration with that of...
10769 (make_pass_graphite): ...new function.
10770 (pass_graphite_transforms): Replace declaration with that of...
10771 (make_pass_graphite_transforms): ...new function.
10772 (pass_if_conversion): Replace declaration with that of...
10773 (make_pass_if_conversion): ...new function.
10774 (pass_loop_distribution): Replace declaration with that of...
10775 (make_pass_loop_distribution): ...new function.
10776 (pass_vectorize): Replace declaration with that of...
10777 (make_pass_vectorize): ...new function.
10778 (pass_slp_vectorize): Replace declaration with that of...
10779 (make_pass_slp_vectorize): ...new function.
10780 (pass_complete_unroll): Replace declaration with that of...
10781 (make_pass_complete_unroll): ...new function.
10782 (pass_complete_unrolli): Replace declaration with that of...
10783 (make_pass_complete_unrolli): ...new function.
10784 (pass_parallelize_loops): Replace declaration with that of...
10785 (make_pass_parallelize_loops): ...new function.
10786 (pass_loop_prefetch): Replace declaration with that of...
10787 (make_pass_loop_prefetch): ...new function.
10788 (pass_iv_optimize): Replace declaration with that of...
10789 (make_pass_iv_optimize): ...new function.
10790 (pass_tree_loop_done): Replace declaration with that of...
10791 (make_pass_tree_loop_done): ...new function.
10792 (pass_ch): Replace declaration with that of...
10793 (make_pass_ch): ...new function.
10794 (pass_ccp): Replace declaration with that of...
10795 (make_pass_ccp): ...new function.
10796 (pass_phi_only_cprop): Replace declaration with that of...
10797 (make_pass_phi_only_cprop): ...new function.
10798 (pass_build_ssa): Replace declaration with that of...
10799 (make_pass_build_ssa): ...new function.
10800 (pass_build_alias): Replace declaration with that of...
10801 (make_pass_build_alias): ...new function.
10802 (pass_build_ealias): Replace declaration with that of...
10803 (make_pass_build_ealias): ...new function.
10804 (pass_dominator): Replace declaration with that of...
10805 (make_pass_dominator): ...new function.
10806 (pass_dce): Replace declaration with that of...
10807 (make_pass_dce): ...new function.
10808 (pass_dce_loop): Replace declaration with that of...
10809 (make_pass_dce_loop): ...new function.
10810 (pass_cd_dce): Replace declaration with that of...
10811 (make_pass_cd_dce): ...new function.
10812 (pass_call_cdce): Replace declaration with that of...
10813 (make_pass_call_cdce): ...new function.
10814 (pass_merge_phi): Replace declaration with that of...
10815 (make_pass_merge_phi): ...new function.
10816 (pass_split_crit_edges): Replace declaration with that of...
10817 (make_pass_split_crit_edges): ...new function.
10818 (pass_pre): Replace declaration with that of...
10819 (make_pass_pre): ...new function.
10820 (pass_profile): Replace declaration with that of...
10821 (make_pass_profile): ...new function.
10822 (pass_strip_predict_hints): Replace declaration with that of...
10823 (make_pass_strip_predict_hints): ...new function.
10824 (pass_lower_complex_O0): Replace declaration with that of...
10825 (make_pass_lower_complex_O0): ...new function.
10826 (pass_lower_complex): Replace declaration with that of...
10827 (make_pass_lower_complex): ...new function.
10828 (pass_lower_vector): Replace declaration with that of...
10829 (make_pass_lower_vector): ...new function.
10830 (pass_lower_vector_ssa): Replace declaration with that of...
10831 (make_pass_lower_vector_ssa): ...new function.
10832 (pass_lower_omp): Replace declaration with that of...
10833 (make_pass_lower_omp): ...new function.
10834 (pass_diagnose_omp_blocks): Replace declaration with that of...
10835 (make_pass_diagnose_omp_blocks): ...new function.
10836 (pass_expand_omp): Replace declaration with that of...
10837 (make_pass_expand_omp): ...new function.
10838 (pass_expand_omp_ssa): Replace declaration with that of...
10839 (make_pass_expand_omp_ssa): ...new function.
10840 (pass_object_sizes): Replace declaration with that of...
10841 (make_pass_object_sizes): ...new function.
10842 (pass_strlen): Replace declaration with that of...
10843 (make_pass_strlen): ...new function.
10844 (pass_fold_builtins): Replace declaration with that of...
10845 (make_pass_fold_builtins): ...new function.
10846 (pass_stdarg): Replace declaration with that of...
10847 (make_pass_stdarg): ...new function.
10848 (pass_early_warn_uninitialized): Replace declaration with that of...
10849 (make_pass_early_warn_uninitialized): ...new function.
10850 (pass_late_warn_uninitialized): Replace declaration with that of...
10851 (make_pass_late_warn_uninitialized): ...new function.
10852 (pass_cse_reciprocals): Replace declaration with that of...
10853 (make_pass_cse_reciprocals): ...new function.
10854 (pass_cse_sincos): Replace declaration with that of...
10855 (make_pass_cse_sincos): ...new function.
10856 (pass_optimize_bswap): Replace declaration with that of...
10857 (make_pass_optimize_bswap): ...new function.
10858 (pass_optimize_widening_mul): Replace declaration with that of...
10859 (make_pass_optimize_widening_mul): ...new function.
10860 (pass_warn_function_return): Replace declaration with that of...
10861 (make_pass_warn_function_return): ...new function.
10862 (pass_warn_function_noreturn): Replace declaration with that of...
10863 (make_pass_warn_function_noreturn): ...new function.
10864 (pass_cselim): Replace declaration with that of...
10865 (make_pass_cselim): ...new function.
10866 (pass_phiopt): Replace declaration with that of...
10867 (make_pass_phiopt): ...new function.
10868 (pass_forwprop): Replace declaration with that of...
10869 (make_pass_forwprop): ...new function.
10870 (pass_phiprop): Replace declaration with that of...
10871 (make_pass_phiprop): ...new function.
10872 (pass_tree_ifcombine): Replace declaration with that of...
10873 (make_pass_tree_ifcombine): ...new function.
10874 (pass_dse): Replace declaration with that of...
10875 (make_pass_dse): ...new function.
10876 (pass_nrv): Replace declaration with that of...
10877 (make_pass_nrv): ...new function.
10878 (pass_rename_ssa_copies): Replace declaration with that of...
10879 (make_pass_rename_ssa_copies): ...new function.
10880 (pass_sink_code): Replace declaration with that of...
10881 (make_pass_sink_code): ...new function.
10882 (pass_fre): Replace declaration with that of...
10883 (make_pass_fre): ...new function.
10884 (pass_check_data_deps): Replace declaration with that of...
10885 (make_pass_check_data_deps): ...new function.
10886 (pass_copy_prop): Replace declaration with that of...
10887 (make_pass_copy_prop): ...new function.
10888 (pass_vrp): Replace declaration with that of...
10889 (make_pass_vrp): ...new function.
10890 (pass_uncprop): Replace declaration with that of...
10891 (make_pass_uncprop): ...new function.
10892 (pass_return_slot): Replace declaration with that of...
10893 (make_pass_return_slot): ...new function.
10894 (pass_reassoc): Replace declaration with that of...
10895 (make_pass_reassoc): ...new function.
10896 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10897 (make_pass_rebuild_cgraph_edges): ...new function.
10898 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10899 (make_pass_remove_cgraph_callee_edges): ...new function.
10900 (pass_build_cgraph_edges): Replace declaration with that of...
10901 (make_pass_build_cgraph_edges): ...new function.
10902 (pass_local_pure_const): Replace declaration with that of...
10903 (make_pass_local_pure_const): ...new function.
10904 (pass_tracer): Replace declaration with that of...
10905 (make_pass_tracer): ...new function.
10906 (pass_warn_unused_result): Replace declaration with that of...
10907 (make_pass_warn_unused_result): ...new function.
10908 (pass_diagnose_tm_blocks): Replace declaration with that of...
10909 (make_pass_diagnose_tm_blocks): ...new function.
10910 (pass_lower_tm): Replace declaration with that of...
10911 (make_pass_lower_tm): ...new function.
10912 (pass_tm_init): Replace declaration with that of...
10913 (make_pass_tm_init): ...new function.
10914 (pass_tm_mark): Replace declaration with that of...
10915 (make_pass_tm_mark): ...new function.
10916 (pass_tm_memopt): Replace declaration with that of...
10917 (make_pass_tm_memopt): ...new function.
10918 (pass_tm_edges): Replace declaration with that of...
10919 (make_pass_tm_edges): ...new function.
10920 (pass_split_functions): Replace declaration with that of...
10921 (make_pass_split_functions): ...new function.
10922 (pass_feedback_split_functions): Replace declaration with that of...
10923 (make_pass_feedback_split_functions): ...new function.
10924 (pass_strength_reduction): Replace declaration with that of...
10925 (make_pass_strength_reduction): ...new function.
10926 (pass_ipa_lower_emutls): Replace declaration with that of...
10927 (make_pass_ipa_lower_emutls): ...new function.
10928 (pass_ipa_function_and_variable_visibility): Replace declaration with
10930 (make_pass_ipa_function_and_variable_visibility): ...new function.
10931 (pass_ipa_tree_profile): Replace declaration with that of...
10932 (make_pass_ipa_tree_profile): ...new function.
10933 (pass_early_local_passes): Replace declaration with that of...
10934 (make_pass_early_local_passes): ...new function.
10935 (pass_ipa_whole_program_visibility): Replace declaration with that
10937 (make_pass_ipa_whole_program_visibility): ...new function.
10938 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10939 (make_pass_ipa_lto_gimple_out): ...new function.
10940 (pass_ipa_increase_alignment): Replace declaration with that of...
10941 (make_pass_ipa_increase_alignment): ...new function.
10942 (pass_ipa_inline): Replace declaration with that of...
10943 (make_pass_ipa_inline): ...new function.
10944 (pass_ipa_free_lang_data): Replace declaration with that of...
10945 (make_pass_ipa_free_lang_data): ...new function.
10946 (pass_ipa_free_inline_summary): Replace declaration with that of...
10947 (make_pass_ipa_free_inline_summary): ...new function.
10948 (pass_ipa_cp): Replace declaration with that of...
10949 (make_pass_ipa_cp): ...new function.
10950 (pass_ipa_reference): Replace declaration with that of...
10951 (make_pass_ipa_reference): ...new function.
10952 (pass_ipa_pure_const): Replace declaration with that of...
10953 (make_pass_ipa_pure_const): ...new function.
10954 (pass_ipa_pta): Replace declaration with that of...
10955 (make_pass_ipa_pta): ...new function.
10956 (pass_ipa_lto_finish_out): Replace declaration with that of...
10957 (make_pass_ipa_lto_finish_out): ...new function.
10958 (pass_ipa_tm): Replace declaration with that of...
10959 (make_pass_ipa_tm): ...new function.
10960 (pass_ipa_profile): Replace declaration with that of...
10961 (make_pass_ipa_profile): ...new function.
10962 (pass_ipa_cdtor_merge): Replace declaration with that of...
10963 (make_pass_ipa_cdtor_merge): ...new function.
10964 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10966 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10967 (pass_init_datastructures): Replace declaration with that of...
10968 (make_pass_init_datastructures): ...new function.
10969 (pass_fixup_cfg): Replace declaration with that of...
10970 (make_pass_fixup_cfg): ...new function.
10971 (pass_expand): Replace declaration with that of...
10972 (make_pass_expand): ...new function.
10973 (pass_instantiate_virtual_regs): Replace declaration with that of...
10974 (make_pass_instantiate_virtual_regs): ...new function.
10975 (pass_rtl_fwprop): Replace declaration with that of...
10976 (make_pass_rtl_fwprop): ...new function.
10977 (pass_rtl_fwprop_addr): Replace declaration with that of...
10978 (make_pass_rtl_fwprop_addr): ...new function.
10979 (pass_jump): Replace declaration with that of...
10980 (make_pass_jump): ...new function.
10981 (pass_jump2): Replace declaration with that of...
10982 (make_pass_jump2): ...new function.
10983 (pass_lower_subreg): Replace declaration with that of...
10984 (make_pass_lower_subreg): ...new function.
10985 (pass_cse): Replace declaration with that of...
10986 (make_pass_cse): ...new function.
10987 (pass_fast_rtl_dce): Replace declaration with that of...
10988 (make_pass_fast_rtl_dce): ...new function.
10989 (pass_ud_rtl_dce): Replace declaration with that of...
10990 (make_pass_ud_rtl_dce): ...new function.
10991 (pass_rtl_dce): Replace declaration with that of...
10992 (make_pass_rtl_dce): ...new function.
10993 (pass_rtl_dse1): Replace declaration with that of...
10994 (make_pass_rtl_dse1): ...new function.
10995 (pass_rtl_dse2): Replace declaration with that of...
10996 (make_pass_rtl_dse2): ...new function.
10997 (pass_rtl_dse3): Replace declaration with that of...
10998 (make_pass_rtl_dse3): ...new function.
10999 (pass_rtl_cprop): Replace declaration with that of...
11000 (make_pass_rtl_cprop): ...new function.
11001 (pass_rtl_pre): Replace declaration with that of...
11002 (make_pass_rtl_pre): ...new function.
11003 (pass_rtl_hoist): Replace declaration with that of...
11004 (make_pass_rtl_hoist): ...new function.
11005 (pass_rtl_store_motion): Replace declaration with that of...
11006 (make_pass_rtl_store_motion): ...new function.
11007 (pass_cse_after_global_opts): Replace declaration with that of...
11008 (make_pass_cse_after_global_opts): ...new function.
11009 (pass_rtl_ifcvt): Replace declaration with that of...
11010 (make_pass_rtl_ifcvt): ...new function.
11011 (pass_into_cfg_layout_mode): Replace declaration with that of...
11012 (make_pass_into_cfg_layout_mode): ...new function.
11013 (pass_outof_cfg_layout_mode): Replace declaration with that of...
11014 (make_pass_outof_cfg_layout_mode): ...new function.
11015 (pass_loop2): Replace declaration with that of...
11016 (make_pass_loop2): ...new function.
11017 (pass_rtl_loop_init): Replace declaration with that of...
11018 (make_pass_rtl_loop_init): ...new function.
11019 (pass_rtl_move_loop_invariants): Replace declaration with that of...
11020 (make_pass_rtl_move_loop_invariants): ...new function.
11021 (pass_rtl_unswitch): Replace declaration with that of...
11022 (make_pass_rtl_unswitch): ...new function.
11023 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
11024 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11025 (pass_rtl_doloop): Replace declaration with that of...
11026 (make_pass_rtl_doloop): ...new function.
11027 (pass_rtl_loop_done): Replace declaration with that of...
11028 (make_pass_rtl_loop_done): ...new function.
11029 (pass_web): Replace declaration with that of...
11030 (make_pass_web): ...new function.
11031 (pass_cse2): Replace declaration with that of...
11032 (make_pass_cse2): ...new function.
11033 (pass_df_initialize_opt): Replace declaration with that of...
11034 (make_pass_df_initialize_opt): ...new function.
11035 (pass_df_initialize_no_opt): Replace declaration with that of...
11036 (make_pass_df_initialize_no_opt): ...new function.
11037 (pass_reginfo_init): Replace declaration with that of...
11038 (make_pass_reginfo_init): ...new function.
11039 (pass_inc_dec): Replace declaration with that of...
11040 (make_pass_inc_dec): ...new function.
11041 (pass_stack_ptr_mod): Replace declaration with that of...
11042 (make_pass_stack_ptr_mod): ...new function.
11043 (pass_initialize_regs): Replace declaration with that of...
11044 (make_pass_initialize_regs): ...new function.
11045 (pass_combine): Replace declaration with that of...
11046 (make_pass_combine): ...new function.
11047 (pass_if_after_combine): Replace declaration with that of...
11048 (make_pass_if_after_combine): ...new function.
11049 (pass_ree): Replace declaration with that of...
11050 (make_pass_ree): ...new function.
11051 (pass_partition_blocks): Replace declaration with that of...
11052 (make_pass_partition_blocks): ...new function.
11053 (pass_match_asm_constraints): Replace declaration with that of...
11054 (make_pass_match_asm_constraints): ...new function.
11055 (pass_regmove): Replace declaration with that of...
11056 (make_pass_regmove): ...new function.
11057 (pass_split_all_insns): Replace declaration with that of...
11058 (make_pass_split_all_insns): ...new function.
11059 (pass_fast_rtl_byte_dce): Replace declaration with that of...
11060 (make_pass_fast_rtl_byte_dce): ...new function.
11061 (pass_lower_subreg2): Replace declaration with that of...
11062 (make_pass_lower_subreg2): ...new function.
11063 (pass_mode_switching): Replace declaration with that of...
11064 (make_pass_mode_switching): ...new function.
11065 (pass_sms): Replace declaration with that of...
11066 (make_pass_sms): ...new function.
11067 (pass_sched): Replace declaration with that of...
11068 (make_pass_sched): ...new function.
11069 (pass_ira): Replace declaration with that of...
11070 (make_pass_ira): ...new function.
11071 (pass_reload): Replace declaration with that of...
11072 (make_pass_reload): ...new function.
11073 (pass_clean_state): Replace declaration with that of...
11074 (make_pass_clean_state): ...new function.
11075 (pass_branch_prob): Replace declaration with that of...
11076 (make_pass_branch_prob): ...new function.
11077 (pass_value_profile_transformations): Replace declaration with that
11079 (make_pass_value_profile_transformations): ...new function.
11080 (pass_postreload_cse): Replace declaration with that of...
11081 (make_pass_postreload_cse): ...new function.
11082 (pass_gcse2): Replace declaration with that of...
11083 (make_pass_gcse2): ...new function.
11084 (pass_split_after_reload): Replace declaration with that of...
11085 (make_pass_split_after_reload): ...new function.
11086 (pass_branch_target_load_optimize1): Replace declaration with that
11088 (make_pass_branch_target_load_optimize1): ...new function.
11089 (pass_thread_prologue_and_epilogue): Replace declaration with that
11091 (make_pass_thread_prologue_and_epilogue): ...new function.
11092 (pass_stack_adjustments): Replace declaration with that of...
11093 (make_pass_stack_adjustments): ...new function.
11094 (pass_peephole2): Replace declaration with that of...
11095 (make_pass_peephole2): ...new function.
11096 (pass_if_after_reload): Replace declaration with that of...
11097 (make_pass_if_after_reload): ...new function.
11098 (pass_regrename): Replace declaration with that of...
11099 (make_pass_regrename): ...new function.
11100 (pass_cprop_hardreg): Replace declaration with that of...
11101 (make_pass_cprop_hardreg): ...new function.
11102 (pass_reorder_blocks): Replace declaration with that of...
11103 (make_pass_reorder_blocks): ...new function.
11104 (pass_branch_target_load_optimize2): Replace declaration with that
11106 (make_pass_branch_target_load_optimize2): ...new function.
11107 (pass_leaf_regs): Replace declaration with that of...
11108 (make_pass_leaf_regs): ...new function.
11109 (pass_split_before_sched2): Replace declaration with that of...
11110 (make_pass_split_before_sched2): ...new function.
11111 (pass_compare_elim_after_reload): Replace declaration with that of...
11112 (make_pass_compare_elim_after_reload): ...new function.
11113 (pass_sched2): Replace declaration with that of...
11114 (make_pass_sched2): ...new function.
11115 (pass_stack_regs): Replace declaration with that of...
11116 (make_pass_stack_regs): ...new function.
11117 (pass_stack_regs_run): Replace declaration with that of...
11118 (make_pass_stack_regs_run): ...new function.
11119 (pass_df_finish): Replace declaration with that of...
11120 (make_pass_df_finish): ...new function.
11121 (pass_compute_alignments): Replace declaration with that of...
11122 (make_pass_compute_alignments): ...new function.
11123 (pass_duplicate_computed_gotos): Replace declaration with that of...
11124 (make_pass_duplicate_computed_gotos): ...new function.
11125 (pass_variable_tracking): Replace declaration with that of...
11126 (make_pass_variable_tracking): ...new function.
11127 (pass_free_cfg): Replace declaration with that of...
11128 (make_pass_free_cfg): ...new function.
11129 (pass_machine_reorg): Replace declaration with that of...
11130 (make_pass_machine_reorg): ...new function.
11131 (pass_cleanup_barriers): Replace declaration with that of...
11132 (make_pass_cleanup_barriers): ...new function.
11133 (pass_delay_slots): Replace declaration with that of...
11134 (make_pass_delay_slots): ...new function.
11135 (pass_split_for_shorten_branches): Replace declaration with that of...
11136 (make_pass_split_for_shorten_branches): ...new function.
11137 (pass_split_before_regstack): Replace declaration with that of...
11138 (make_pass_split_before_regstack): ...new function.
11139 (pass_convert_to_eh_region_ranges): Replace declaration with that
11141 (make_pass_convert_to_eh_region_ranges): ...new function.
11142 (pass_shorten_branches): Replace declaration with that of...
11143 (make_pass_shorten_branches): ...new function.
11144 (pass_set_nothrow_function_flags): Replace declaration with that of...
11145 (make_pass_set_nothrow_function_flags): ...new function.
11146 (pass_dwarf2_frame): Replace declaration with that of...
11147 (make_pass_dwarf2_frame): ...new function.
11148 (pass_final): Replace declaration with that of...
11149 (make_pass_final): ...new function.
11150 (pass_rtl_seqabstr): Replace declaration with that of...
11151 (make_pass_rtl_seqabstr): ...new function.
11152 (pass_release_ssa_names): Replace declaration with that of...
11153 (make_pass_release_ssa_names): ...new function.
11154 (pass_early_inline): Replace declaration with that of...
11155 (make_pass_early_inline): ...new function.
11156 (pass_inline_parameters): Replace declaration with that of...
11157 (make_pass_inline_parameters): ...new function.
11158 (pass_update_address_taken): Replace declaration with that of...
11159 (make_pass_update_address_taken): ...new function.
11160 (pass_convert_switch): Replace declaration with that of...
11161 (make_pass_convert_switch): ...new function.
11162 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11163 to a subclass of simple_ipa_opt_pass along with...
11164 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11165 (make_pass_ipa_tree_profile): ...new function.
11166 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11167 subclass of gimple_opt_pass along with...
11168 (pass_data_sra_early): ...new pass_data instance and...
11169 (make_pass_sra_early): ...new function.
11170 (pass_sra): Convert from a global struct to a subclass of
11171 gimple_opt_pass along with...
11172 (pass_data_sra): ...new pass_data instance and...
11173 (make_pass_sra): ...new function.
11174 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11175 gimple_opt_pass along with...
11176 (pass_data_early_ipa_sra): ...new pass_data instance and...
11177 (make_pass_early_ipa_sra): ...new function.
11178 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11179 subclass of gimple_opt_pass along with...
11180 (pass_data_ccp): ...new pass_data instance and...
11181 (make_pass_ccp): ...new function.
11182 (pass_fold_builtins): Convert from a global struct to a subclass of
11183 gimple_opt_pass along with...
11184 (pass_data_fold_builtins): ...new pass_data instance and...
11185 (make_pass_fold_builtins): ...new function.
11186 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11187 subclass of gimple_opt_pass along with...
11188 (pass_data_copy_prop): ...new pass_data instance and...
11189 (make_pass_copy_prop): ...new function.
11190 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11191 global struct to a subclass of gimple_opt_pass along with...
11192 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11193 (make_pass_rename_ssa_copies): ...new function.
11194 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11195 subclass of gimple_opt_pass along with...
11196 (pass_data_dce): ...new pass_data instance and...
11197 (make_pass_dce): ...new function.
11198 (pass_dce_loop): Convert from a global struct to a subclass of
11199 gimple_opt_pass along with...
11200 (pass_data_dce_loop): ...new pass_data instance and...
11201 (make_pass_dce_loop): ...new function.
11202 (pass_cd_dce): Convert from a global struct to a subclass of
11203 gimple_opt_pass along with...
11204 (pass_data_cd_dce): ...new pass_data instance and...
11205 (make_pass_cd_dce): ...new function.
11206 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11207 subclass of gimple_opt_pass along with...
11208 (pass_data_dominator): ...new pass_data instance and...
11209 (make_pass_dominator): ...new function.
11210 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11211 gimple_opt_pass along with...
11212 (pass_data_phi_only_cprop): ...new pass_data instance and...
11213 (make_pass_phi_only_cprop): ...new function.
11214 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11215 subclass of gimple_opt_pass along with...
11216 (pass_data_dse): ...new pass_data instance and...
11217 (make_pass_dse): ...new function.
11218 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11219 a subclass of gimple_opt_pass along with...
11220 (pass_data_forwprop): ...new pass_data instance and...
11221 (make_pass_forwprop): ...new function.
11222 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11223 struct to a subclass of gimple_opt_pass along with...
11224 (pass_data_tree_ifcombine): ...new pass_data instance and...
11225 (make_pass_tree_ifcombine): ...new function.
11226 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11227 subclass of gimple_opt_pass along with...
11228 (pass_data_ch): ...new pass_data instance and...
11229 (make_pass_ch): ...new function.
11230 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11231 subclass of gimple_opt_pass along with...
11232 (pass_data_tree_loop): ...new pass_data instance and...
11233 (make_pass_tree_loop): ...new function.
11234 (pass_tree_loop_init): Convert from a global struct to a subclass of
11235 gimple_opt_pass along with...
11236 (pass_data_tree_loop_init): ...new pass_data instance and...
11237 (make_pass_tree_loop_init): ...new function.
11238 (pass_lim): Convert from a global struct to a subclass of
11239 gimple_opt_pass along with...
11240 (pass_data_lim): ...new pass_data instance and...
11241 (make_pass_lim): ...new function.
11242 (pass_tree_unswitch): Convert from a global struct to a subclass of
11243 gimple_opt_pass along with...
11244 (pass_data_tree_unswitch): ...new pass_data instance and...
11245 (make_pass_tree_unswitch): ...new function.
11246 (pass_predcom): Convert from a global struct to a subclass of
11247 gimple_opt_pass along with...
11248 (pass_data_predcom): ...new pass_data instance and...
11249 (make_pass_predcom): ...new function.
11250 (pass_vectorize): Convert from a global struct to a subclass of
11251 gimple_opt_pass along with...
11252 (pass_data_vectorize): ...new pass_data instance and...
11253 (make_pass_vectorize): ...new function.
11254 (pass_graphite): Convert from a global struct to a subclass of
11255 gimple_opt_pass along with...
11256 (pass_data_graphite): ...new pass_data instance and...
11257 (make_pass_graphite): ...new function.
11258 (pass_graphite_transforms): Convert from a global struct to a subclass
11259 of gimple_opt_pass along with...
11260 (pass_data_graphite_transforms): ...new pass_data instance and...
11261 (make_pass_graphite_transforms): ...new function.
11262 (pass_check_data_deps): Convert from a global struct to a subclass of
11263 gimple_opt_pass along with...
11264 (pass_data_check_data_deps): ...new pass_data instance and...
11265 (make_pass_check_data_deps): ...new function.
11266 (pass_iv_canon): Convert from a global struct to a subclass of
11267 gimple_opt_pass along with...
11268 (pass_data_iv_canon): ...new pass_data instance and...
11269 (make_pass_iv_canon): ...new function.
11270 (pass_scev_cprop): Convert from a global struct to a subclass of
11271 gimple_opt_pass along with...
11272 (pass_data_scev_cprop): ...new pass_data instance and...
11273 (make_pass_scev_cprop): ...new function.
11274 (pass_record_bounds): Convert from a global struct to a subclass of
11275 gimple_opt_pass along with...
11276 (pass_data_record_bounds): ...new pass_data instance and...
11277 (make_pass_record_bounds): ...new function.
11278 (pass_complete_unroll): Convert from a global struct to a subclass of
11279 gimple_opt_pass along with...
11280 (pass_data_complete_unroll): ...new pass_data instance and...
11281 (make_pass_complete_unroll): ...new function.
11282 (pass_complete_unrolli): Convert from a global struct to a subclass of
11283 gimple_opt_pass along with...
11284 (pass_data_complete_unrolli): ...new pass_data instance and...
11285 (make_pass_complete_unrolli): ...new function.
11286 (pass_parallelize_loops): Convert from a global struct to a subclass
11287 of gimple_opt_pass along with...
11288 (pass_data_parallelize_loops): ...new pass_data instance and...
11289 (make_pass_parallelize_loops): ...new function.
11290 (pass_loop_prefetch): Convert from a global struct to a subclass of
11291 gimple_opt_pass along with...
11292 (pass_data_loop_prefetch): ...new pass_data instance and...
11293 (make_pass_loop_prefetch): ...new function.
11294 (pass_iv_optimize): Convert from a global struct to a subclass of
11295 gimple_opt_pass along with...
11296 (pass_data_iv_optimize): ...new pass_data instance and...
11297 (make_pass_iv_optimize): ...new function.
11298 (pass_tree_loop_done): Convert from a global struct to a subclass of
11299 gimple_opt_pass along with...
11300 (pass_data_tree_loop_done): ...new pass_data instance and...
11301 (make_pass_tree_loop_done): ...new function.
11302 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11303 struct to a subclass of gimple_opt_pass along with...
11304 (pass_data_cse_reciprocals): ...new pass_data instance and...
11305 (make_pass_cse_reciprocals): ...new function.
11306 (pass_cse_sincos): Convert from a global struct to a subclass of
11307 gimple_opt_pass along with...
11308 (pass_data_cse_sincos): ...new pass_data instance and...
11309 (make_pass_cse_sincos): ...new function.
11310 (pass_optimize_bswap): Convert from a global struct to a subclass of
11311 gimple_opt_pass along with...
11312 (pass_data_optimize_bswap): ...new pass_data instance and...
11313 (make_pass_optimize_bswap): ...new function.
11314 (pass_optimize_widening_mul): Convert from a global struct to a
11315 subclass of gimple_opt_pass along with...
11316 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11317 (make_pass_optimize_widening_mul): ...new function.
11318 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11319 subclass of gimple_opt_pass along with...
11320 (pass_data_phiopt): ...new pass_data instance and...
11321 (make_pass_phiopt): ...new function.
11322 (pass_cselim): Convert from a global struct to a subclass of
11323 gimple_opt_pass along with...
11324 (pass_data_cselim): ...new pass_data instance and...
11325 (make_pass_cselim): ...new function.
11326 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11327 subclass of gimple_opt_pass along with...
11328 (pass_data_phiprop): ...new pass_data instance and...
11329 (make_pass_phiprop): ...new function.
11330 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11331 subclass of gimple_opt_pass along with...
11332 (pass_data_pre): ...new pass_data instance and...
11333 (make_pass_pre): ...new function.
11334 (pass_fre): Convert from a global struct to a subclass of
11335 gimple_opt_pass along with...
11336 (pass_data_fre): ...new pass_data instance and...
11337 (make_pass_fre): ...new function.
11338 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11339 subclass of gimple_opt_pass along with...
11340 (pass_data_reassoc): ...new pass_data instance and...
11341 (make_pass_reassoc): ...new function.
11342 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11343 subclass of gimple_opt_pass along with...
11344 (pass_data_sink_code): ...new pass_data instance and...
11345 (make_pass_sink_code): ...new function.
11346 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11347 subclass of gimple_opt_pass along with...
11348 (pass_data_strlen): ...new pass_data instance and...
11349 (make_pass_strlen): ...new function.
11350 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11351 struct to a subclass of gimple_opt_pass along with...
11352 (pass_data_build_alias): ...new pass_data instance and...
11353 (make_pass_build_alias): ...new function.
11354 (pass_build_ealias): Convert from a global struct to a subclass of
11355 gimple_opt_pass along with...
11356 (pass_data_build_ealias): ...new pass_data instance and...
11357 (make_pass_build_ealias): ...new function.
11358 (pass_ipa_pta): Convert from a global struct to a subclass of
11359 simple_ipa_opt_pass along with...
11360 (pass_data_ipa_pta): ...new pass_data instance and...
11361 (make_pass_ipa_pta): ...new function.
11362 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11363 subclass of gimple_opt_pass along with...
11364 (pass_data_uncprop): ...new pass_data instance and...
11365 (make_pass_uncprop): ...new function.
11366 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11367 global struct to a subclass of gimple_opt_pass along with...
11368 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11369 (make_pass_late_warn_uninitialized): ...new function.
11370 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11371 to a subclass of gimple_opt_pass along with...
11372 (pass_data_init_datastructures): ...new pass_data instance and...
11373 (make_pass_init_datastructures): ...new function.
11374 (pass_early_warn_uninitialized): Convert from a global struct to a
11375 subclass of gimple_opt_pass along with...
11376 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11377 (make_pass_early_warn_uninitialized): ...new function.
11378 (pass_update_address_taken): Convert from a global struct to a
11379 subclass of gimple_opt_pass along with...
11380 (pass_data_update_address_taken): ...new pass_data instance and...
11381 (make_pass_update_address_taken): ...new function.
11382 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11383 struct to a subclass of gimple_opt_pass along with...
11384 (pass_data_release_ssa_names): ...new pass_data instance and...
11385 (make_pass_release_ssa_names): ...new function.
11386 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11387 subclass of gimple_opt_pass along with...
11388 (pass_data_stdarg): ...new pass_data instance and...
11389 (make_pass_stdarg): ...new function.
11390 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11391 global struct to a subclass of gimple_opt_pass along with...
11392 (pass_data_convert_switch): ...new pass_data instance and...
11393 (make_pass_convert_switch): ...new function.
11394 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11395 to a subclass of gimple_opt_pass along with...
11396 (pass_data_tail_recursion): ...new pass_data instance and...
11397 (make_pass_tail_recursion): ...new function.
11398 (pass_tail_calls): Convert from a global struct to a subclass of
11399 gimple_opt_pass along with...
11400 (pass_data_tail_calls): ...new pass_data instance and...
11401 (make_pass_tail_calls): ...new function.
11402 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11403 struct to a subclass of gimple_opt_pass along with...
11404 (pass_data_lower_vector): ...new pass_data instance and...
11405 (make_pass_lower_vector): ...new function.
11406 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11407 gimple_opt_pass along with...
11408 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11409 (make_pass_lower_vector_ssa): ...new function.
11410 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11411 to a subclass of gimple_opt_pass along with...
11412 (pass_data_slp_vectorize): ...new pass_data instance and...
11413 (make_pass_slp_vectorize): ...new function.
11414 (pass_ipa_increase_alignment): Convert from a global struct to a
11415 subclass of simple_ipa_opt_pass along with...
11416 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11417 (make_pass_ipa_increase_alignment): ...new function.
11418 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11419 gimple_opt_pass along with...
11420 (pass_data_vrp): ...new pass_data instance and...
11421 (make_pass_vrp): ...new function.
11422 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11423 subclass of simple_ipa_opt_pass along with...
11424 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11425 (make_pass_ipa_free_lang_data): ...new function.
11426 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11427 gimple_opt_pass along with...
11428 (pass_data_tsan): ...new pass_data instance and...
11429 (make_pass_tsan): ...new function.
11430 (pass_tsan_O0): Convert from a global struct to a subclass of
11431 gimple_opt_pass along with...
11432 (pass_data_tsan_O0): ...new pass_data instance and...
11433 (make_pass_tsan_O0): ...new function.
11434 * var-tracking.c (pass_variable_tracking): Convert from a global
11435 struct to a subclass of rtl_opt_pass along with...
11436 (pass_data_variable_tracking): ...new pass_data instance and...
11437 (make_pass_variable_tracking): ...new function.
11438 * web.c (pass_web): Convert from a global struct to a subclass of
11439 rtl_opt_pass along with...
11440 (pass_data_web): ...new pass_data instance and...
11441 (make_pass_web): ...new function.
11442 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11443 declaration with that of...
11444 (make_pass_mode_switch_use): ...new function.
11445 (pass_resolve_sw_modes): Replace declaration with that of...
11446 (make_pass_resolve_sw_modes): ...new function.
11447 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11448 from a global struct to a subclass of rtl_opt_pass along with...
11449 (pass_data_mode_switch_use): ...new pass_data instance and...
11450 (make_pass_mode_switch_use): ...new function.
11451 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11452 from a global struct to a subclass of rtl_opt_pass along with...
11453 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11454 (make_pass_resolve_sw_modes): ...new function.
11455 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11456 struct to a subclass of rtl_opt_pass along with...
11457 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11458 (make_pass_insert_vzeroupper): ...new function.
11459 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11460 global struct to a subclass of rtl_opt_pass along with...
11461 (pass_data_work_around_errata): ...new pass_data instance and...
11462 (make_pass_work_around_errata): ...new function.
11463 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11464 struct to a subclass of rtl_opt_pass along with...
11465 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11466 (make_pass_mips_machine_reorg2): ...new function.
11468 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11470 * passes.c (pass_manager::operator new): New.
11472 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11474 Handwritten part of conversion of passes to C++ classes.
11476 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11477 (toplev.o): Add dep on PASS_MANAGER_H.
11478 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11479 of early local pases to reflect this moving from a global to a
11480 member of gcc::pass_manager.
11481 (cgraph_add_new_function): Likewise.
11482 * lto-cgraph.c (lto_output_node): Update for conversion of
11483 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11484 * passes.c (opt_pass::clone): New.
11485 (opt_pass::gate): New.
11486 (opt_pass::execute): New.
11487 (opt_pass::opt_pass): New.
11488 (pass_manager::execute_early_local_passes): New.
11489 (pass_manager::execute_pass_mode_switching): new.
11490 (finish_optimization_passes): Convert to...
11491 (pass_manager::finish_optimization_passes): ...this.
11492 (finish_optimization_passes): Update for conversion of passes to
11494 (register_dump_files_1): Use has_gate since we cannot portably
11495 check a vtable entry against NULL.
11496 (dump_one_pass): Likewise.
11497 (ipa_write_summaries_2): Likewise.
11498 (ipa_write_optimization_summaries_1): Likewise.
11499 (ipa_read_summaries_1): Likewise.
11500 (ipa_read_optimization_summaries_1): Likewise.
11501 (execute_ipa_stmt_fixups): Likewise.
11502 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11503 pass-creation functions rather than wiring up globals, and
11504 storing the results in fields of pass_manager generated using
11505 pass-instances.def.
11506 (pass_manager::dump_profile_report): Update for conversion of
11507 passes to C++ classes.
11508 (pass_manager::execute_ipa_summary_passes): Likewise.
11509 (execute_one_ipa_transform_pass): Likewise.
11510 (execute_one_pass): Use has_gate and has_execute since we cannot
11511 portably check a vtable entry against NULL.
11512 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11513 (pass_manager): Use pass-instances.def to add fields for the
11514 various pass instances.
11515 * toplev.c (finalize): Update for move of
11516 finish_optimization_passes to a method of gcc::pass_manager.
11517 * toplev.h (finish_optimization_passes): Move to method of class
11519 * tree-pass.h (struct pass_data): New.
11520 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11521 (opt_pass::gate): Convert to virtual function.
11522 (opt_pass::~opt_pass): New.
11523 (opt_pass::clone): New.
11524 (opt_pass::execute): Convert to virtual function.
11525 (opt_pass::opt_pass): New.
11526 (opt_pass::ctxt_): new.
11527 (gimple_opt_pass): Convert to subclass of opt_pass.
11528 (gimple_opt_pass::gimple_opt_pass): New.
11529 (rtl_opt_pass): Convert to subclass of opt_pass.
11530 (rtl_opt_pass::rtl_opt_pass): New.
11531 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11532 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11533 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11534 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11535 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11536 invocation of pass_mode_switching to reflect this moving from a
11537 global to a member of gcc::pass_manager.
11538 (ix86_option_override): Rework how pass_insert_vzeroupper is
11539 added to the pass_manager to reflect autogenerated changes.
11540 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11542 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11544 PR rtl-optimization/57708
11545 * recog.c (peep2_find_free_register): Validate all regs in a
11548 2013-08-05 Jan Hubicka <jh@suse.cz>
11551 * cgraph.c (verify_cgraph_node): Accept local flags from other
11553 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11554 (function_and_variable_visibility): Likewise.
11555 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11557 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11559 * graph.c (init_graph_slim_pretty_print): Remove.
11560 (print_graph_cfg): Do not call it. Use local pretty printer.
11561 (start_graph_dump): Likewise.
11563 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11565 * gimple-pretty-print.c (buffer): Remove.
11566 (initialized): Likewise.
11567 (maybe_init_pretty_print): Likewise.
11568 (print_gimple_stmt): Do not call it. Use non-static local
11569 pretty_printer variable.
11570 (print_gimple_expr): Likewise.
11571 (print_gimple_seq): Likewise.
11572 (gimple_dump_bb): Likewise.
11574 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11576 * asan.c (asan_pp): Remove.
11577 (asan_pp_initialized): Likewise.
11578 (asan_pp_initialize): Likewise.
11579 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11580 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11581 (asan_add_global): Likewise.
11583 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11585 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11586 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11587 * pretty-print.c (pp_formatted_text_data): Likewise.
11588 (pp_indent): Rename from pp_base_indent.
11589 (pp_format): Rename from pp_base_format.
11590 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11591 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11592 (pp_flush): Rename from pp_base_flush.
11593 (pp_set_line_maximum_length): Rename from
11594 pp_base_set_line_maximum_length.
11595 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11596 (pp_set_prefix): Rename from pp_base_set_prefix.
11597 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11598 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11599 (pp_append_text): Rename from pp_base_append_text.
11600 (pp_formatted_text): Rename from pp_base_formatted_text.
11601 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11602 (pp_remaining_character_count_for_line): Rename from
11603 pp_base_remaining_character_count_for_line.
11604 (pp_newline): Rename from pp_base_newline.
11605 (pp_character): Rename from pp_base_character.
11606 (pp_string): Rename from pp_base_string.
11607 (pp_maybe_space): Rename from pp_base_maybe_space.
11608 * asan.c (asan_pp_string): Adjust.
11609 (asan_emit_stack_protection): Likewise.
11610 (asan_add_global): Likewise.
11611 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11612 * tree-mudflap.c (mf_varname_tree): Likewise.
11613 * tree-pretty-print.c (pp_tree_identifier): Rename from
11614 pp_base_tree_identifier.
11615 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11616 Declare as function.
11618 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11620 * pretty-print.h (pp_bar_bar): New.
11621 (pp_ampersand_ampersand): Likewise.
11622 (pp_less_equal): Likewise.
11623 (pp_greater_equal): Likewise.
11624 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11625 printer functions instead of pp_string or operators and punctuators.
11626 (dump_gimple_call): Likewise.
11627 (dump_gimple_omp_for): Likewise.
11628 (dump_gimple_transaction): Likewise.
11629 (dump_gimple_phi): Likewise.
11630 (pp_gimple_stmt_1): Likewise.
11631 * sched-vis.c (print_insn): Likewise.
11632 * tree-mudflap.c (mf_varname_tree): Likewise.
11633 * tree-pretty-print.c (dump_block_node): Likewise.
11634 (dump_generic_node): Likewise.
11636 2013-08-02 Jan Hubicka <jh@suse.cz>
11638 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11640 * lto-streamer-out.c (tree_is_indexable): Results decls and
11641 parm decls are not indexable.
11642 (DFS_write_tree_body): Do not follow args and results.
11643 (hash_tree): Likewise.
11644 (output_functions): Rearrange so struct function is needed
11645 only when real body is output; be able to also ouptut abstract
11646 functions; output DECL_ARGUMENTS and DECL_RESULT.
11647 (lto_output): When not in WPA, ale store abstract functions.
11648 (write_symbol): Do not care about RESULT_DECL.
11649 (output_symbol_p): Handle correctly sbtract decls.
11650 * lto-streamer-in.c (input_function): Rearrange so struct
11651 function can be NULL at entry; allow streaming of
11652 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11653 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11654 sanity check during LTO.
11655 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11656 RESULT_DECl and DECL_ARGUMENTS.
11657 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11660 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11662 * pretty-print.h (pp_underscore): New.
11664 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11665 printer functions instead of pp_character.
11666 (dump_binary_rhs): Likewise.
11667 (dump_ternary_rhs): Likewise.
11668 (dump_gimple_call_args): Likewise.
11669 (pp_points_to_solution): Likewise.
11670 (dump_gimple_call): Likewise.
11671 (dump_gimple_switch): Likewise.
11672 (dump_gimple_cond): Likewise.
11673 (dump_gimple_bind): Likewise.
11674 (dump_gimple_try): Likewise.
11675 (dump_gimple_omp_for): Likewise.
11676 (dump_gimple_omp_continue): Likewise.
11677 (dump_gimple_omp_single): Likewise.
11678 (dump_gimple_omp_sections): Likewise.
11679 (dump_gimple_omp_block): Likewise.
11680 (dump_gimple_omp_critical): Likewise.
11681 (dump_gimple_transaction): Likewise.
11682 (dump_gimple_asm): Likewise.
11683 (dump_gimple_phi): Likewise.
11684 (dump_gimple_omp_parallel): Likewise.
11685 (dump_gimple_omp_task): Likewise.
11686 (dump_gimple_omp_atomic_load): Likewise.
11687 (dump_gimple_omp_atomic_store): Likewise.
11688 (dump_gimple_mem_ops): Likewise.
11689 (pp_gimple_stmt_1): Likewise.
11690 (pp_cfg_jump): Likewise.
11691 (dump_implicit_edges): Likewise.
11692 (gimple_dump_bb_for_graph): Likewise.
11693 * graph.c (draw_cfg_node): Likewise.
11694 * langhooks.c (lhd_print_error_function): Likewise.
11695 * sched-vis.c (print_exp): Likewise.
11696 (print_value): Likewise.
11697 (print_pattern): Likewise.
11698 (print_insn): Likewise.
11699 (rtl_dump_bb_for_graph): Likewise.
11700 * tree-pretty-print.c (dump_function_declaration): Likewise.
11701 (dump_array_domain): Likewise.
11702 (dump_omp_clause): Likewise.
11703 (dump_location): Likewise.
11704 (dump_generic_node): Likewise.
11705 (print_struct_decl): Likewise.
11706 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11708 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11710 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11711 candidate table when replacing a candidate statement.
11712 (replace_rhs_if_not_dup): Likewise.
11713 (replace_one_candidate): Likewise.
11715 2013-08-02 Jan Hubicka <jh@suse.cz>
11716 Martin Liska <marxin.liska@gmail.com>
11718 * cgraphunit.c (add_new_function): Fix logic when adding from
11720 (assemble_thunk): Rename to ...
11721 (expand_thunk); .. this one; export; get it working with
11722 general functions; make produced gimple valid.
11723 * cgraph.h (expand_thunk): Declare.
11725 2013-08-02 Jan Hubicka <jh@suse.cz>
11727 * ipa-cp.c (gather_context_independent_values): Use
11728 ipa_get_param_move_cost.
11729 (get_replacement_map): Remove PARAM; move parameter folding
11731 (create_specialized_node): Update.
11732 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11733 assert that we have gimple body; update move_cost.
11734 (count_formal_params): Assert that we have gimple body.
11735 (ipa_dump_param): New function.
11736 (ipa_alloc_node_params): Break out from ...
11737 (ipa_initialize_node_params): ... here.
11738 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11739 (ipa_write_node_info): Stream move costs.
11740 (ipa_read_node_info): Read move costs.
11741 (ipa_update_after_lto_read): Do not recompute node params.
11742 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11743 (ipa_get_param): Check we are not in WPA.
11744 (ipa_get_param_move_cost): New.
11745 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11746 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11747 parm numbers to be present.
11749 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11751 PR rtl-optimization/58048
11752 * lra-constraints.c (process_alt_operands): Don't check asm
11753 operand on register.
11755 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11757 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11758 the implied StoreLoad barrier for atomic operations if before.
11760 2013-08-02 Jan Hubicka <jh@suse.cz>
11761 Martin Liska <marxin.liska@gmail.com>
11763 * cgraph.c (cgraph_function_body_availability): Do not check
11765 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11766 symtab_node_availability): Declare.
11767 * ipa.c (can_replace_by_local_alias): New.
11768 (function_and_variable_visibility): Use it.
11769 * symtab.c (symtab_for_node_and_aliases,
11770 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11772 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11774 PR rtl-optimization/57963
11775 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11776 (lra_constraints): Use them.
11778 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11780 * config/arm/types.md (define_attr "type"): Add "load_acq"
11782 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11784 (cortex_a53_store1): Likewise.
11786 2013-08-01 Jan Hubicka <jh@suse.cz>
11788 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11789 partitions are not needed.
11791 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11793 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11794 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11795 MAYBE_NON_Q_CLASS_P where appropriate.
11797 2013-08-01 Jan Hubicka <jh@suse.cz>
11799 * cgraph.h (release_function_body): Declare.
11800 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11801 of unused delcarations.
11803 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11805 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11806 RTL form when subtracting a constant.
11808 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11810 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11811 Generate canonical plus rtx with negated immediate instead of minus
11813 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11815 2013-08-01 Jan Hubicka <jh@suse.cz>
11817 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11818 (cgraph_release_function_body): Likewise.
11819 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11820 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11821 to used_as_abstract_origin.
11822 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11823 symbols used as abstract origins.
11824 * cgraphunit.c (analyze_functions): Update.
11825 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11826 used_as_abstract_origin.
11827 * tree-inline.c (tree_function_versioning): Update
11828 used_as_abstract_origin; be ready for DECL_RESULT and
11829 DECL_ARGUMENTS to be NULL.
11831 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11832 for abstract functions.
11833 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11836 2013-08-01 Jan Hubicka <jh@suse.cz>
11838 * profile.c (compute_value_histograms): Fix thinko.
11840 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11842 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11843 aarch-common-protos.h to extra_headers.
11844 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11845 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11846 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11848 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11850 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11851 Include "../arm/types.md". Define "type" attribute for all patterns.
11852 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11855 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11857 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11858 to support power8 load fusion.
11859 (fusion_gpr_mem_load): Likewise.
11861 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11863 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11864 declarations for power8 load fusion.
11865 (emit_fusion_gpr_load): Likewise.
11867 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11868 tuning for power8, turn on fusion mode by default. Turn on sign
11869 extending fusion mode if normal fusion mode is on, and we are at
11871 (fusion_gpr_load_p): New function, return true if we can fuse an
11872 addis instruction with a dependent load to a GPR.
11873 (emit_fusion_gpr_load): Emit the instructions for power8 load
11876 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11877 (VSX load fusion peepholes): New peepholes to fuse together an
11878 addi instruction with a VSX load instruction.
11880 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11881 peepholes to fuse an addis instruction with a load to a GPR base
11882 register. If we are supporting sign extending fusions, convert
11883 sign extending loads to zero extending loads and add an explicit
11886 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11888 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11889 aarch-common-protos.h to extra_headers.
11890 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11891 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11892 (arm_early_store_addr_dep): Likewise.
11893 (arm_no_early_alu_shift_dep): Likewise.
11894 (arm_no_early_alu_shift_value_dep): Likewise.
11895 (arm_no_early_mul_dep): Likewise.
11896 (arm_no_early_store_addr_dep): Likewise.
11897 (arm_mac_accumulator_is_mul_result): Likewise.
11898 (arm_mac_accumulator_is_result): Likewise.
11899 * config/arm/aarch-common.c: ... here. New file.
11900 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11902 (arm_early_store_addr_dep): Likewise.
11903 (arm_no_early_alu_shift_dep): Likewise.
11904 (arm_no_early_alu_shift_value_dep): Likewise.
11905 (arm_no_early_mul_dep): Likewise.
11906 (arm_no_early_store_addr_dep): Likewise.
11907 (arm_mac_accumulator_is_mul_result): Likewise.
11908 (arm_mac_accumulator_is_result): Likewise.
11909 * config/arm/aarch-common-protos.h: ... here. New file.
11910 * config/arm/t-arm (aarch-common.o): Define.
11912 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11914 * config/arm/arm.md: Include new file "types.md".
11915 (define_attr "type"): Move from here to ...
11916 (define_attr "mul32"): Likewise.
11917 (define_attr "mul64"): Likewise.
11918 * config/arm/types.md: ... here. New file.
11920 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11922 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11923 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11925 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11927 * gen-pass-instances.awk: Fix offset of substr().
11929 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11931 * Makefile.in (pass-instances.def): New.
11932 (passes.o): Replace dependency on passes.def with one on
11935 * gen-pass-instances.awk: New.
11937 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11938 rather than passes.def, updating local definition of NEXT_PASS
11939 macro to add an extra NUM parameter (currently unused).
11941 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11943 * Makefile.in (PASS_MANAGER_H): New.
11944 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11945 (passes.o): Likewise.
11946 (statistics.o): Likewise.
11947 (cgraphunit.o): Likewise.
11948 (context.o): Depend on PASS_MANAGER_H.
11950 * pass_manager.h: New.
11952 * cgraphunit.c (cgraph_add_new_function): Update for moves
11953 of globals to fields of pass_manager.
11954 (analyze_function): Likewise.
11955 (expand_function): Likewise.
11956 (ipa_passes): Likewise.
11957 (compile): Likewise.
11959 * context.c (context::context): New.
11960 * context.h (context::context): New.
11961 (context::get_passes): New.
11962 (context::passes_): New.
11964 * lto-cgraph.c (input_node): Update for moves of globals to
11965 fields of pass_manager.
11967 * passes.c (all_passes): Remove, in favor of a field of the
11968 same name within the new class pass_manager.
11969 (all_small_ipa_passes): Likewise.
11970 (all_lowering_passes): Likewise.
11971 (all_regular_ipa_passes): Likewise.
11972 (all_late_ipa_passes): Likewise.
11973 (all_lto_gen_passes): Likewise.
11974 (passes_by_id): Likewise.
11975 (passes_by_id_size): Likewise.
11976 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
11977 the new class pass_manager.
11978 (set_pass_for_id): Convert to...
11979 (pass_manager::set_pass_for_id): ...method.
11980 (get_pass_for_id): Convert to...
11981 (pass_manager::get_pass_for_id): ...method.
11982 (register_one_dump_file): Move body of implementation into...
11983 (pass_manager::register_one_dump_file): ...here.
11984 (register_dump_files_1): Convert to...
11985 (pass_manager::register_dump_files_1): ...method.
11986 (register_dump_files): Convert to...
11987 (pass_manager::register_dump_files): ...method.
11988 (create_pass_tab): Update for moves of globals to fields of
11990 (dump_passes): Move body of implementation into...
11991 (pass_manager::dump_passes): ...here.
11992 (register_pass): Move body of implementation into...
11993 (pass_manager::register_pass): ...here.
11994 (init_optimization_passes): Convert into...
11995 (pass_manager::pass_manager): ...constructor for new
11996 pass_manager class, and initialize the pass_lists array.
11997 (check_profile_consistency): Update for moves of globals to
11998 fields of pass_manager.
11999 (dump_profile_report): Move body of implementation into...
12000 (pass_manager::dump_profile_report): ...here.
12001 (ipa_write_summaries_1): Update for moves of pass lists from
12002 being globals to fields of pass_manager.
12003 (ipa_write_optimization_summaries): Likewise.
12004 (ipa_read_summaries): Likewise.
12005 (ipa_read_optimization_summaries): Likewise.
12006 (execute_all_ipa_stmt_fixups): Likewise.
12008 * statistics.c (statistics_fini): Update for moves of globals to
12009 fields of pass_manager.
12011 * toplev.c (general_init): Replace call to
12012 init_optimization_passes with construction of the pass_manager
12015 * tree-pass.h (all_passes): Remove, in favor of a field of the
12016 same name within the new class pass_manager.
12017 (all_small_ipa_passes): Likewise.
12018 (all_lowering_passes): Likewise.
12019 (all_regular_ipa_passes): Likewise.
12020 (all_lto_gen_passes): Likewise.
12021 (all_late_ipa_passes): Likewise.
12022 (passes_by_id): Likewise.
12023 (passes_by_id_size): Likewise.
12024 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12025 the new class pass_manager.
12026 (get_pass_for_id): Remove.
12028 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12030 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
12033 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12035 * arm.md (mulhi3): New expand pattern.
12037 2013-07-30 Jan Hubicka <jh@suse.cz>
12038 Martin Liska <marxin.liska@gmail.com>
12040 * profile.c (compute_value_histograms): Do not ICE when
12041 there is mismatch only on some counters.
12043 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12045 PR rtl-optimization/57637
12046 * function.c (move_insn_for_shrink_wrap): Also check the
12047 GEN set of the LIVE problem for the liveness analysis
12048 if it exists, otherwise give up.
12050 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12052 PR tree-optimization/57993
12053 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
12054 replaced statement in the candidate table.
12055 (phi_add_costs): Return infinite cost when the hidden basis does
12056 not dominate all phis on which the candidate is dependent.
12057 (replace_one_candidate): Record replaced statement in the
12060 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
12062 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
12063 (ashlv2si3): New expander.
12064 (*ashlv2si3_i): New define_insn_and_split.
12065 * predicates.md (float_operation): Allow patterns with three
12066 basic sub-patterns.
12068 PR rtl-optimization/58021
12069 * mode-switching.c (create_pre_exit): Always split off preceding
12070 insns if we are not at the basic block head.
12072 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
12074 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
12075 (UCLIBC_DYNAMIC_LINKER): New macro.
12076 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
12078 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
12079 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
12081 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
12082 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
12083 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
12084 for SF and DF modes. Use ieee_quad_format for TF mode.
12085 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
12086 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
12087 (mips_option_override): Handle `-mnan=legacy'.
12088 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
12089 `-mabs=2008' and `-mnan=2008'.
12090 (OPTION_DEFAULT_SPECS): Add "nan" default.
12091 (ASM_SPEC): Handle `-mnan='.
12092 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
12093 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
12094 comment accordingly.
12095 (neg<mode>2): Likewise.
12096 * config/mips/mips.opt (mabs, mnan): New options.
12097 * doc/install.texi (Configuration): Document `--with-nan=' option.
12098 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
12100 (MIPS Options): Document them.
12101 * config.gcc <mips*-*-*>: Handle `--with-nan='.
12102 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
12103 * configure: Regenerate.
12104 * config.in: Regenerate.
12106 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12108 * config/i386/i386.md (float post-reload splitters): Do not check
12109 for subregs of SSE registers.
12111 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12112 H.J. Lu <hongjiu.lu@intel.com>
12116 * config/i386/i386.md (post-reload splitter
12117 to avoid partial SSE reg dependency stalls): New pattern.
12119 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12121 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12123 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12125 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12126 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12128 2013-07-26 Cary Coutant <ccoutant@google.com>
12130 * dwarf2out.c (die_checksum_ordered): Don't include template
12131 instantiations in signature.
12132 (is_template_parameter): New function.
12133 (is_template_instantiation): New function.
12134 (generate_skeleton_bottom_up): Don't include template instantiations
12136 (generate_skeleton): Likewise.
12137 (break_out_comdat_types): Move recursive call to break out nested
12139 (prune_unused_types_mark_generic_parms_dies): Call
12140 is_template_parameter.
12142 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12144 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12145 uses vector registers.
12146 * config/aarch64/iterators.md: Add attributes rtn and vas.
12148 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12149 Richard Earnshaw <richard.earnshaw@arm.com>
12151 * combine.c (simplify_comparison): Re-canonicalize operands
12153 * config/arm/arm.md (movcond_addsi): New splitter.
12155 2013-07-25 Sterling Augustine <saugustine@google.com>
12157 * dwarf2out.c (size_of_pubnames): Move code to...
12158 (include_pubname_in_output): ...here. New.
12159 (want_pubnames): Rearrange.
12160 (output_pubnames): Call include_pubname_in_output. Move assertion.
12162 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12164 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12166 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12169 * doc/extend.texi: Remove obsolete builtins. Fix
12170 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12172 2013-07-25 Jan Hubicka <jh@suse.cz>
12174 * cgraph.c (release_function_body): Break out from ...
12175 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12176 and DECL_ARGUMENTS.
12177 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12178 old_tree in the map.
12179 (create_specialized_node): Update.
12180 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12182 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12183 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12184 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12185 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12188 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12190 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12191 addsi3_carryin_alt2_<optab>): Correct output template.
12193 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12195 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12196 Adjust for arm_restrict_it.
12197 Remove trailing whitespace.
12199 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12201 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12202 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12204 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12206 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12208 PR rtl-optimization/57960
12209 * lra-constraints.c (process_alt_operands): Use the right mode
12210 when checking strict_low.
12212 2013-07-25 Jan Hubicka <jh@suse.cz>
12214 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12215 * cgraph.c (cgraph_remove_node): Do not release function body
12216 when in cgraph streaming.
12217 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12218 in other partitions are not considered reachable; fix handling of
12221 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12223 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12225 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12230 * config/arm/arm.md ("*sibcall_insn): Replace use of
12231 Ss with US. Adjust output for v5 and v4t.
12232 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12234 * config/arm/constraints.md ("Ss"): Rename to US.
12236 2013-07-25 Terry Guo <terry.guo@arm.com>
12238 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12239 shift_add/shift_sub0/shift_sub1 RTXs.
12241 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12242 Anton Blanchard <anton@au1.ibm.com>
12244 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12245 (altivec_vpks<VI_char>ss): Likewise.
12246 (altivec_vpks<VI_char>us): Likewise.
12247 (altivec_vpku<VI_char>us): Likewise.
12248 (altivec_vpku<VI_char>um): Likewise.
12250 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12252 Introduce context class.
12254 * Makefile.in (CONTEXT_H): New.
12255 (OBJS): Add context.o.
12256 (toplev.o): Add CONTEXT_H to dependencies.
12259 * toplev.c (general_init): Create the singleton gcc::context instance.
12265 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12267 PR rtl-optimization/57968
12268 * mode-switching.c (create_pre_exit): Allow instructions that
12269 don't set a return register to need a non-exit mode.
12271 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12272 Anton Blanchard <anton@au1.ibm.com>
12274 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12275 operands to vperm for little endian.
12276 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12277 of lvsl to create the control mask for a vperm for little endian.
12279 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12280 Anton Blanchard <anton@au1.ibm.com>
12282 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12283 two operands for little-endian.
12285 2013-07-23 Steve Ellcey <sellcey@mips.com>
12287 * config/mips/mips.c (mips_case_values_threshold): New.
12288 (TARGET_CASE_VALUES_THRESHOLD): Define.
12290 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12291 Anton Blanchard <anton@au1.ibm.com>
12293 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12294 selection of field for vector splat in little endian mode.
12296 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12298 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12299 expanders to rs6000.md.
12300 (ior<mode>3): Likewise.
12301 (and<mode>3): Likewise.
12302 (one_cmpl<mode>2): Likewise.
12303 (nor<mode>3): Likewise.
12304 (andc<mode>3): Likewise.
12305 (eqv<mode>3): Likewise.
12306 (nand<mode>3): Likewise.
12307 (orc<mode>3): Likewise.
12309 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12312 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12313 to split multi-word logical operations.
12314 (rs6000_split_logical_di): Likewise.
12315 (rs6000_split_logical): Likewise.
12317 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12318 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12319 and allow TImode operations in 32-bit.
12320 (vsx_and<mode>3_64bit): Likewise.
12321 (vsx_ior<mode>3_32bit): Likewise.
12322 (vsx_ior<mode>3_64bit): Likewise.
12323 (vsx_xor<mode>3_32bit): Likewise.
12324 (vsx_xor<mode>3_64bit): Likewise.
12325 (vsx_one_cmpl<mode>2_32bit): Likewise.
12326 (vsx_one_cmpl<mode>2_64bit): Likewise.
12327 (vsx_nor<mode>3_32bit): Likewise.
12328 (vsx_nor<mode>3_64bit): Likewise.
12329 (vsx_andc<mode>3_32bit): Likewise.
12330 (vsx_andc<mode>3_64bit): Likewise.
12331 (vsx_eqv<mode>3_32bit): Likewise.
12332 (vsx_eqv<mode>3_64bit): Likewise.
12333 (vsx_nand<mode>3_32bit): Likewise.
12334 (vsx_nand<mode>3_64bit): Likewise.
12335 (vsx_orc<mode>3_32bit): Likewise.
12336 (vsx_orc<mode>3_64bit): Likewise.
12338 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12339 logical types in GPRs.
12341 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12342 logical insns to rs6000.md, and allow TImode operations in
12344 (altivec_ior<mode>3): Likewise.
12345 (altivec_xor<mode>3): Likewise.
12346 (altivec_one_cmpl<mode>2): Likewise.
12347 (altivec_nor<mode>3): Likewise.
12348 (altivec_andc<mode>3): Likewise.
12350 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12351 attributes for moving the 128-bit logical operations into
12353 (BOOL_REGS_OUTPUT): Likewise.
12354 (BOOL_REGS_OP1): Likewise.
12355 (BOOL_REGS_OP2): Likewise.
12356 (BOOL_REGS_UNARY): Likewise.
12357 (BOOL_REGS_AND_CR0): Likewise.
12358 (one_cmpl<mode>2): Add support for DI logical operations on
12359 32-bit, splitting the operations to 32-bit.
12360 (anddi3): Likewise.
12361 (iordi3): Likewise.
12362 (xordi3): Likewise.
12363 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12364 changes to combine the 32/64-bit code, allow logical operations on
12365 TI mode in 32-bit, and to use similar match_operator patterns like
12366 scalar mode uses. Combine the Altivec and VSX code for logical
12367 operations, and move it here.
12368 (ior<mode>3, 128-bit types): Likewise.
12369 (xor<mode>3, 128-bit types): Likewise.
12370 (one_cmpl<mode>3, 128-bit types): Likewise.
12371 (nor<mode>3, 128-bit types): Likewise.
12372 (andc<mode>3, 128-bit types): Likewise.
12373 (eqv<mode>3, 128-bit types): Likewise.
12374 (nand<mode>3, 128-bit types): Likewise.
12375 (orc<mode>3, 128-bit types): Likewise.
12376 (and<mode>3_internal): Likewise.
12377 (bool<mode>3_internal): Likewise.
12378 (boolc<mode>3_internal1): Likewise.
12379 (boolc<mode>3_internal2): Likewise.
12380 (boolcc<mode>3_internal1): Likewise.
12381 (boolcc<mode>3_internal2): Likewise.
12382 (eqv<mode>3_internal1): Likewise.
12383 (eqv<mode>3_internal2): Likewise.
12384 (one_cmpl1<mode>3_internal): Likewise.
12386 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12388 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12389 Rename flag_stack_usage to flag_stack_usage_info.
12391 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12393 * config/microblaze/sync.md: New file.
12394 * config/microblaze/microblaze.md: Include sync.md
12395 * config/microblaze/microblaze.c: Add print_operand 'y'.
12396 * config/microblaze/constraints.md: Add memory_contraint
12397 'Q' which is a single register.
12399 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12401 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12403 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12405 * reload.c (find_reloads): Exit loop once we find this operand
12406 cannot be reloaded somehow for this alternative.
12408 * reload.c (find_reloads): Exit loop once we find a hard register.
12410 * rtlanal.c (computed_jump_p): Exit loop once we find label
12413 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12415 * cfgloopmanip.c (remove_path): Exit loop after setting
12418 * gensupport.c (subst_dup): Avoid loop if code is not
12419 MATCH_DUP nor MATCH_OP_DUP.
12421 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12423 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12425 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12427 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12428 true for SP_REGNUM if mode == ptr_mode.
12429 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12430 with value R0_REGNUM + 31.
12432 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12434 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12435 pad pointer-typed argument downward.
12437 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12439 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12440 and __ILP32__ when the ILP32 model is in use.
12442 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12444 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12445 (aarch64_load_symref_appropriately): In the case of
12446 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12447 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12448 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12449 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12450 if 'mode' doesn't equal to 'ptr_mode'.
12451 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12452 'vcall_offset'; change to call aarch64_emit_move differently depending
12453 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12454 to calculate the upper bound of 'vcall_offset'.
12455 (aarch64_cannot_force_const_mem): Change to also return true if
12457 (aarch64_legitimize_reload_address): In the case of large
12458 displacements, add new local variable 'xmode' and an assertion
12459 based on it; change to use 'xmode' to generate the new rtx and
12461 (aarch64_asm_trampoline_template): Change to generate the template
12462 differently depending on TARGET_ILP32 or not; change to use
12463 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12464 (aarch64_trampoline_size): Removed.
12465 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12466 and replace immediate literals with it. Change to use 'ptr_mode'
12467 instead of 'DImode' and call convert_memory_address if the mode
12468 of 'fnaddr' doesn't equal to 'ptr_mode'.
12469 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12471 (aarch64_elf_asm_destructor): Likewise.
12472 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12473 on TARGET_ILP32 instead of aarch64_trampoline_size.
12474 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12475 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12476 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12477 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12478 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12479 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12480 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12481 depending on the value of 'mode'.
12482 (add_losym_<mode>): New.
12483 (ldr_got_small_<mode>): New, based on ldr_got_small.
12484 (ldr_got_small): Remove.
12485 (ldr_got_small_sidi): New.
12486 * config/aarch64/iterators.md (P): New.
12487 (PTR): Change to 'ptr_mode' in the condition.
12489 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12491 * config.gcc (aarch64*-*-*): Support --with-abi.
12492 (aarch64*-*-elf): Support --with-multilib-list.
12493 (aarch64*-*-linux*): Likewise.
12494 (supported_defaults): Add abi to aarch64*-*-*.
12495 * configure.ac: Mention AArch64 for --with-multilib-list.
12496 * configure: Re-generated.
12497 * config/aarch64/biarchilp32.h: New file.
12498 * config/aarch64/biarchlp64.h: New file.
12499 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12501 (MULTILIB_DEFAULTS): Ditto.
12502 (DRIVER_SELF_SPECS): Ditto.
12503 (ASM_SPEC): Update to also substitute -mabi.
12504 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12505 file whose name depends on -mabi= and -mbig-endian.
12506 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12508 (POINTER_SIZE): New define.
12509 (POINTERS_EXTEND_UNSIGNED): Ditto.
12510 (enum aarch64_abi_type): New enumeration tag.
12511 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12512 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12513 (TARGET_ILP32): New define.
12514 * config/aarch64/aarch64.opt (mabi): New.
12515 (aarch64_abi): New.
12516 (ilp32, lp64): New values for -mabi.
12517 * config/aarch64/t-aarch64 (comma): New define.
12518 (MULTILIB_OPTIONS): Ditto.
12519 (MULTILIB_DIRNAMES): Ditto.
12520 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12521 * doc/invoke.texi: Document -mabi for AArch64.
12523 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12525 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12527 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12528 Anton Blanchard <anton@au1.ibm.com>
12530 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12531 endianness when selecting field to splat.
12533 2013-07-22 Eric Christopher <echristo@gmail.com>
12535 * dwarf2out.c (die_odr_checksum): New function to use
12536 CHECKSUM_ macros and ULEB128 for DIE tag.
12537 (generate_type_signature): Use.
12539 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12541 * config.gcc (sparc*-*-*): Accept leon3 processor.
12542 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12543 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12544 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12545 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12546 (mfix-ut699): Adjust comment.
12547 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12548 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12549 (CPP_CPU_SPEC): Likewise.
12550 (ASM_CPU_SPEC): Likewise.
12551 * config/sparc/sparc.c (leon3_cost): New constant.
12552 (sparc_option_override): Add leon3 support.
12553 (mem_ref): New function.
12554 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12555 (sparc_do_work_around_errata): Look into the instruction in the delay
12556 slot and adjust accordingly. Add fix for the data cache nullify issues
12557 of the UT699. Change insertion position for the NOP.
12558 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12559 (leon3_load): New reservation.
12560 (leon_store): Bump latency to 2.
12561 (grfpu): New automaton.
12562 (grfpu_alu): New unit.
12563 (grfpu_ds): Likewise.
12564 (leon_fp_alu): Adjust.
12565 (leon_fp_mult): Delete.
12566 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12567 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12568 * config/sparc/sparc.md (cpu): Add leon3.
12569 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12570 (swapsi): Likewise.
12571 (atomic_test_and_set): Likewise.
12572 (ldstub): Likewise.
12574 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12576 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12577 default for R5900 targets.
12578 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12579 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12580 * config/mips/mips.c (mips_option_override): Report an error for
12581 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12582 for -march=r5900 -mhard-float.
12584 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12586 * df-problems.c (can_move_insns_across): Exit loop once we
12587 find a non-fixed, non-global register.
12589 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12592 * omega.c (omega_eliminate_red): Break after setting red_found.
12593 (omega_problem_has_red_equations): Similarly after setting found.
12594 (omega_query_variable): Similarly after setting coupled.
12596 2013-07-22 Marek Polacek <polacek@redhat.com>
12598 * gimplify.c: Don't include gimple.h twice.
12600 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12602 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12603 instead of TARGET_THUMB1.
12604 (Pz): New constraint.
12605 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12607 (compare_negsi_si): Likewise.
12608 (compare_addsi2_op0): Likewise.
12609 (compare_addsi2_op1): Likewise.
12610 (addsi3_carryin_<optab>): Likewise.
12611 (addsi3_carryin_alt2_<optab>): Likewise.
12612 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12613 for arm_restrict_it.
12614 (subsi3_carryin): Likewise.
12615 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12616 (minmax_arithsi): Disable for arm_restrict_it.
12617 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12618 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12619 (satsi_<SAT:code>_shift): Likewise.
12620 (arm_shiftsi3): Add alternative for 16-bit encoding.
12621 (arm32_movhf): Disable for arm_restrict_it.
12622 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12623 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12625 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12627 * config/arm/arm.md (attribute "insn"): Delete.
12628 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12629 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12630 (not_shiftsi): Update for attribute change.
12631 (not_shiftsi_compare0): Likewise.
12632 (not_shiftsi_compare0_scratch): Likewise.
12633 (arm_one_cmplsi2): Likewise.
12634 (thumb1_one_cmplsi2): Likewise.
12635 (notsi_compare0): Likewise.
12636 (notsi_compare0_scratch): Likewise.
12637 (thumb1_movdi_insn): Likewise.
12638 (arm_movsi_insn): Likewise.
12639 (movhi_insn_arch4): Likewise.
12640 (movhi_bytes): Likewise.
12641 (arm_movqi_insn): Likewise.
12642 (thumb1_movqi_insn): Likewise.
12643 (arm32_movhf): Likewise.
12644 (thumb1_movhf): Likewise.
12645 (arm_movsf_soft_insn): Likewise.
12646 (thumb1_movsf_insn): Likewise.
12647 (thumb_movdf_insn): Likewise.
12648 (movsicc_insn): Likewise.
12649 (movsfcc_soft_insn): Likewise.
12650 (and_scc): Likewise.
12651 (cond_move): Likewise.
12652 (if_move_not): Likewise.
12653 (if_not_move): Likewise.
12654 (if_shift_move): Likewise.
12655 (if_move_shift): Likewise.
12656 (if_shift_shift): Likewise.
12657 (if_not_arith): Likewise.
12658 (if_arith_not): Likewise.
12659 (cond_move_not): Likewise.
12660 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12661 (neon_mov<mode>): Likewise.
12662 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12663 (thumb2_movsi_vfp): Likewise.
12664 (movsf_vfp): Likewise.
12665 (thumb2_movsf_vfp): Likewise.
12666 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12668 (cortexa7_older_only): Likewise.
12669 (cortexa7_younger): Likewise.
12670 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12671 (1020alu_shift_op): Likewise.
12672 (1020alu_shift_reg_op): Likewise.
12673 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12674 (alu_shift_op): Likewise.
12675 (alu_shift_reg_op): Likewise.
12676 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12677 (11_alu_shift_op): Likewise.
12678 (11_alu_shift_reg_op): Likewise.
12679 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12680 (9_alu_shift_reg_op): Likewise.
12681 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12683 (cortex_a15_alu_shift): Likewise.
12684 (cortex_a15_alu_shift_reg): Likewise.
12685 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12686 (cortex_a5_alu_shift): Likewise.
12687 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12689 (cortex_a53_alu_shift): Likewise.
12690 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12692 (cortex_a7_alu_reg): Likewise.
12693 (cortex_a7_alu_shift): Likewise.
12694 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12695 (cortex_a8_alu_shift): Likewise.
12696 (cortex_a8_alu_shift_reg): Likewise.
12697 (cortex_a8_mov): Likewise.
12698 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12699 (cortex_a9_dp_shift): Likewise.
12700 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12701 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12702 (cortex_r4_mov): Likewise.
12703 (cortex_r4_alu_shift): Likewise.
12704 (cortex_r4_alu_shift_reg): Likewise.
12705 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12706 (526_alu_shift_op): Likewise.
12707 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12708 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12709 (626te_alu_shift_op): Likewise.
12710 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12711 (726te_alu_op): Likewise.
12712 (726te_alu_shift_op): Likewise.
12713 (726te_alu_shift_reg_op): Likewise.
12714 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12715 (mp626_alu_shift_op): Likewise.
12716 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12717 (pj4_alu_e1_conds): Likewise.
12718 (pj4_alu): Likewise.
12719 (pj4_alu_conds): Likewise.
12720 (pj4_shift): Likewise.
12721 (pj4_shift_conds): Likewise.
12722 (pj4_alu_shift): Likewise.
12723 (pj4_alu_shift_conds): Likewise.
12725 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12727 * config/arm/predicates.md (shiftable_operator_strict_it):
12729 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12730 Disable cond_exec version for arm_restrict_it.
12731 (thumb2_smaxsi3): Convert to generate cond_exec.
12732 (thumb2_sminsi3): Likewise.
12733 (thumb32_umaxsi3): Likewise.
12734 (thumb2_uminsi3): Likewise.
12735 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12736 (thumb2_neg_abssi2): Likewise.
12737 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12738 (thumb2_movsicc_insn): Adjust alternatives.
12739 (thumb2_mov_negscc): Disable for arm_restrict_it.
12740 (thumb2_mov_negscc_strict_it): New pattern.
12741 (thumb2_mov_notscc_strict_it): New pattern.
12742 (thumb2_mov_notscc): Disable for arm_restrict_it.
12743 (thumb2_ior_scc): Likewise.
12744 (thumb2_ior_scc_strict_it): New pattern.
12745 (thumb2_cond_move): Adjust for arm_restrict_it.
12746 (thumb2_cond_arith): Disable for arm_restrict_it.
12747 (thumb2_cond_arith_strict_it): New pattern.
12748 (thumb2_cond_sub): Adjust for arm_restrict_it.
12749 (thumb2_movcond): Likewise.
12750 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12751 (thumb2_zero_extendhisi2_v6): Likewise.
12752 (thumb2_zero_extendqisi2_v6): Likewise.
12753 (orsi_notsi_si): Likewise.
12754 (orsi_not_shiftsi_si): Likewise.
12756 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12758 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12759 instruction sequence is 1 byte shorter.
12761 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12763 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12764 it is not needed after split.
12766 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12769 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12770 second label for nonlocal goto receivers. Don't output pic base labels
12771 unless we're producing PIC; mark that action unreachable().
12772 (ix86_save_reg): If the function contains a nonlocal label, save the
12774 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12775 * config/darwin.c (emitted_pic_label_num): New GTY.
12776 (update_pic_label_number_if_needed): New.
12777 (machopic_output_function_base_name): Adjust for nonlocal receiver
12779 (machopic_should_output_picbase_label): New.
12780 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12781 (nonlocal_goto_receiver): New insn and split.
12783 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12785 * config/aarch64/aarch64-builtins.c
12786 (aarch64_fold_builtin): Fold abs in all modes.
12787 * config/aarch64/aarch64-simd-builtins.def
12788 (abs): Enable for all modes.
12789 * config/aarch64/arm_neon.h
12790 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12791 (vabs_f64): Add missing intrinsic.
12793 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12795 * config/aarch64/arm_neon.h (vabs_s64): New function
12797 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12800 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12801 * config/avr/avr.md (adjust_len): Add `round'.
12802 * config/avr/avr-protos.h (avr_out_round): New prototype.
12803 (avr_out_plus): Add `out_label' argument.
12804 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12805 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12806 Handle the case where `insn' is just a pattern.
12807 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12808 (avr_out_round): New function.
12809 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12811 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12813 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12814 Add check for flag_stack_usage to handle -fstack-usage support
12816 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12818 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12819 interaction for new Power8 flags and VSX.
12821 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12823 PR middle-end/57698
12824 * tree-inline.c (expand_call_inline): Emit errors during
12825 early_inlining only if optimization is not turned on.
12827 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12831 * passes.c (init_optimization_passes): Move the construction of
12832 the pass hierarchy into a new passes.def file.
12834 * Makefile.in (passes.o): Add dependency on passes.def.
12836 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12838 * passes.c (init_optimization_passes): Introduce macros for
12839 constructing the tree of passes (INSERT_PASSES_AFTER,
12840 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12841 TERMINATE_PASS_LIST).
12843 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12844 Wei Mi <wmi@google.com>
12846 PR rtl-optimization/57878
12847 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12849 (reload_pseudo_compare_func): Check nregs first for reload
12852 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12854 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12856 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12858 * read-rtl.c (validate_const_int): Once an invalid character is
12859 seen, quit the loop.
12861 * gengtype.c (write_roots): Similarly once we find the "deletable"
12862 or "if_marked" option.
12864 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12866 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12867 "xtab" and "sat". Move value "clz" from here to ...
12868 (attriubte "type"): ... here.
12869 (satsi_<SAT:code>): Delete "insn" attribute.
12870 (satsi_<SAT:code>_shift): Likewise.
12871 (arm_zero_extendqisi2addsi): Likewise.
12872 (arm_extendqisi2addsi): Likewise.
12873 (clzsi2): Update for attribute changes.
12874 (rbitsi2): Likewise.
12875 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12877 (arm_usatsihi): Likewise.
12878 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12880 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12882 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12883 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12884 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12885 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12886 in alphabetical order.
12887 (attribute "core_cycles"): Update for attribute changes.
12888 (arm_addsi3): Likewise.
12889 (addsi3_compare0): Likewise.
12890 (addsi3_compare0_scratch): Likewise.
12891 (addsi3_compare_op1): Likewise.
12892 (addsi3_compare_op2): Likewise.
12893 (compare_addsi2_op0): Likewise.
12894 (compare_addsi2_op1): Likewise.
12895 (addsi3_carryin_shift_<optab>): Likewise.
12896 (subsi3_carryin_shift): Likewise.
12897 (rsbsi3_carryin_shift): Likewise.
12898 (arm_subsi3_insn): Likewise.
12899 (subsi3_compare0): Likewise.
12900 (subsi3_compare): Likewise.
12901 (arm_andsi3_insn): Likewise.
12902 (thumb1_andsi3_insn): Likewise.
12903 (andsi3_compare0): Likewise.
12904 (andsi3_compare0_scratch): Likewise.
12905 (zeroextractsi_compare0_scratch
12906 (andsi_not_shiftsi_si): Likewise.
12907 (iorsi3_insn): Likewise.
12908 (iorsi3_compare0): Likewise.
12909 (iorsi3_compare0_scratch): Likewise.
12910 (arm_xorsi3): Likewise.
12911 (thumb1_xorsi3_insn): Likewise.
12912 (xorsi3_compare0): Likewise.
12913 (xorsi3_compare0_scratch): Likewise.
12914 (satsi_<SAT:code>_shift): Likewise.
12916 (arm_shiftsi3): Likewise.
12917 (shiftsi3_compare0): Likewise.
12918 (not_shiftsi): Likewise.
12919 (not_shiftsi_compare0): Likewise.
12920 (not_shiftsi_compare0_scratch): Likewise.
12921 (arm_one_cmplsi2): Likewise.
12922 (thumb_one_complsi2): Likewise.
12923 (notsi_compare0): Likewise.
12924 (notsi_compare0_scratch): Likewise.
12925 (thumb1_zero_extendhisi2): Likewise.
12926 (arm_zero_extendhisi2): Likewise.
12927 (arm_zero_extendhisi2_v6): Likewise.
12928 (arm_zero_extendhisi2addsi): Likewise.
12929 (thumb1_zero_extendqisi2): Likewise.
12930 (thumb1_zero_extendqisi2_v6): Likewise.
12931 (arm_zero_extendqisi2): Likewise.
12932 (arm_zero_extendqisi2_v6): Likewise.
12933 (arm_zero_extendqisi2addsi): Likewise.
12934 (thumb1_extendhisi2): Likewise.
12935 (arm_extendhisi2): Likewise.
12936 (arm_extendhisi2_v6): Likewise.
12937 (arm_extendqisi): Likewise.
12938 (arm_extendqisi_v6): Likewise.
12939 (arm_extendqisi2addsi): Likewise.
12940 (thumb1_extendqisi2): Likewise.
12941 (thumb1_movdi_insn): Likewise.
12942 (arm_movsi_insn): Likewise.
12943 (movsi_compare0): Likewise.
12944 (movhi_insn_arch4): Likewise.
12945 (movhi_bytes): Likewise.
12946 (arm_movqi_insn): Likewise.
12947 (thumb1_movqi_insn): Likewise.
12948 (arm32_movhf): Likewise.
12949 (thumb1_movhf): Likewise.
12950 (arm_movsf_soft_insn): Likewise.
12951 (thumb1_movsf_insn): Likewise.
12952 (movdf_soft_insn): Likewise.
12953 (thumb_movdf_insn): Likewise.
12954 (arm_cmpsi_insn): Likewise.
12955 (cmpsi_shiftsi): Likewise.
12956 (cmpsi_shiftsi_swp): Likewise.
12957 (arm_cmpsi_negshiftsi_si): Likewise.
12958 (movsicc_insn): Likewise.
12959 (movsfcc_soft_insn): Likewise.
12960 (arith_shiftsi): Likewise.
12961 (arith_shiftsi_compare0
12962 (arith_shiftsi_compare0_scratch
12963 (sub_shiftsi): Likewise.
12964 (sub_shiftsi_compare0
12965 (sub_shiftsi_compare0_scratch
12966 (and_scc): Likewise.
12967 (cond_move): Likewise.
12968 (if_plus_move): Likewise.
12969 (if_move_plus): Likewise.
12970 (if_move_not): Likewise.
12971 (if_not_move): Likewise.
12972 (if_shift_move): Likewise.
12973 (if_move_shift): Likewise.
12974 (if_shift_shift): Likewise.
12975 (if_not_arith): Likewise.
12976 (if_arith_not): Likewise.
12977 (cond_move_not): Likewise.
12978 (thumb1_ashlsi3): Set type attribute.
12979 (thumb1_ashrsi3): Likewise.
12980 (thumb1_lshrsi3): Likewise.
12981 (thumb1_rotrsi3): Likewise.
12982 (shiftsi3_compare0_scratch): Likewise.
12983 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
12984 (neon_mov<mode>): Likewise.
12985 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
12987 (thumb2_movsi_insn): Likewise.
12988 (thumb2_cmpsi_neg_shiftsi): Likewise.
12989 (thumb2_extendqisi_v6): Likewise.
12990 (thumb2_zero_extendhisi2_v6): Likewise.
12991 (thumb2_zero_extendqisi2_v6): Likewise.
12992 (thumb2_shiftsi3_short): Likewise.
12993 (thumb2_addsi3_compare0_scratch): Likewise.
12994 (orsi_not_shiftsi_si): Likewise.
12995 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
12996 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
12998 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
12999 (1020alu_shift_op): Likewise.
13000 (1020alu_shift_reg_op): Likewise.
13001 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13002 (alu_shift_op): Likewise.
13003 (alu_shift_reg_op): Likewise.
13004 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13005 (11_alu_shift_op): Likewise.
13006 (11_alu_shift_reg_op): Likewise.
13007 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13008 (9_alu_shift_reg_op): Likewise.
13009 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13011 (cortex_a15_alu_shift): Likewise.
13012 (cortex_a15_alu_shift_reg): Likewise.
13013 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
13015 (cortex_a5_alu_shift): Likewise.
13016 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
13018 (cortex_a53_alu_shift): Likewise.
13019 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13021 (cortex_a7_alu_reg): Likewise.
13022 (cortex_a7_alu_shift): Likewise.
13023 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
13025 (cortex_a8_alu_shift): Likewise.
13026 (cortex_a8_alu_shift_reg): Likewise.
13027 (cortex_a8_mov): Likewise.
13028 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13029 (cortex_a9_dp_shift): Likewise.
13030 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
13032 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
13034 (cortex_r4_mov): Likewise.
13035 (cortex_r4_alu_shift): Likewise.
13036 (cortex_r4_alu_shift_reg): Likewise.
13037 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13038 (526_alu_shift_op): Likewise.
13039 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13040 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13041 (626te_alu_shift_op): Likewise.
13042 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
13043 (726te_alu_op): Likewise.
13044 (726te_alu_shift_op): Likewise.
13045 (726te_alu_shift_reg_op): Likewise.
13046 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13047 (mp626_alu_shift_op): Likewise.
13048 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
13049 (pj4_alu_e1_conds): Likewise.
13050 (pj4_alu): Likewise.
13051 (pj4_alu_conds): Likewise.
13052 (pj4_shift): Likewise.
13053 (pj4_shift_conds): Likewise.
13054 (pj4_alu_shift): Likewise.
13055 (pj4_alu_shift_conds): Likewise.
13056 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13058 (cortexa7_older_only): Likewise.
13059 (cortexa7_younger): Likewise.
13061 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13063 * ipa-pure-const.c (generate_summary): Rename to...
13064 (pure_const_generate_summary): ... this.
13066 2013-07-17 Iain Sandoe <iain@codesourcery.com>
13068 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
13070 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
13073 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
13076 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13078 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
13079 enabled without -march=zEC12.
13080 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
13083 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13085 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
13086 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
13087 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
13088 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
13089 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
13090 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
13091 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
13093 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
13094 and ISA_HAS_NMADD3_NMSUB3.
13095 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
13096 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
13097 (nmsub4<mode>, nmsub3<mode>): Likewise.
13098 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
13100 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13102 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
13103 TARGET_MIPS5400 checking.
13105 2013-07-16 Jakub Jelinek <jakub@redhat.com>
13106 Peter Bergner <bergner@vnet.ibm.com>
13108 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13109 registers in the comment.
13110 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13111 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13112 rather than FIRST_PSEUDO_REGISTERS.
13114 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13116 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13117 enable extra ISA flags with TARGET_HTM.
13119 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13121 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13124 2013-07-15 Cong Hou <congh@google.com>
13126 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13127 in compare function for sorting.
13129 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13131 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13132 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13133 * config/rs6000/rs6000.opt: Add -mhtm option.
13134 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13135 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13136 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13137 __HTM__ if the HTM instructions are available.
13138 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13139 htm_spr_reg_operand): New define_predicates.
13140 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13141 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13143 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13144 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13145 HTM builtin functions.
13146 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13147 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13148 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13149 (rs6000_builtin_mask_calculate): Likewise.
13150 (rs6000_option_override_internal): Likewise.
13151 (bdesc_htm): Add new HTM builtin support.
13152 (htm_spr_num): New function.
13153 (htm_spr_regno): Likewise.
13154 (rs6000_htm_spr_icode): Likewise.
13155 (htm_expand_builtin): Likewise.
13156 (htm_init_builtins): Likewise.
13157 (rs6000_expand_builtin): Add support for HTM builtin functions.
13158 (rs6000_init_builtins): Likewise.
13159 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13161 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13162 (TARGET_HTM, MASK_HTM): Define macros.
13163 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13164 (FIXED_REGISTERS): Likewise.
13165 (CALL_USED_REGISTERS): Likewise.
13166 (CALL_REALLY_USED_REGISTERS): Likewise.
13167 (REG_ALLOC_ORDER): Likewise.
13168 (enum reg_class): Likewise.
13169 (REG_CLASS_NAMES): Likewise.
13170 (REG_CLASS_CONTENTS): Likewise.
13171 (REGISTER_NAMES): Likewise.
13172 (ADDITIONAL_REGISTER_NAMES): Likewise.
13173 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13174 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13175 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13176 * config/rs6000/htm.md: New file.
13177 * config/rs6000/htmintrin.h: New file.
13178 * config/rs6000/htmxlintrin.h: New file.
13180 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13182 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13183 Define SYMBOL_TINY_GOT, update comment.
13184 * config/aarch64/aarch64.c
13185 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13186 (aarch64_expand_mov_immediate): Likewise.
13187 (aarch64_print_operand): Likewise.
13188 (aarch64_classify_symbol): Likewise.
13189 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13190 (ldr_got_tiny): Define.
13192 2013-07-13 Tobias Grosser <tobias@grosser.es>
13194 PR tree-optimization/54094
13195 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13196 scheduling dimension for the parallelism check from the polyhedral
13197 information in the AST.
13198 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13201 2013-07-13 Jason Merrill <jason@redhat.com>
13203 * print-tree.c (debug_vec_tree): Use debug_raw.
13204 (debug_raw (vec<tree, va_gc> &)): New.
13205 (debug_raw (vec<tree, va_gc> *)): New.
13206 * tree.h: Declare them.
13208 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13210 * ifcvt.c (ifcvt_after_combine): New static variable.
13211 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13213 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13214 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13215 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13217 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13219 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13221 2013-07-12 Michael Matz <matz@suse.de>
13223 PR middle-end/55771
13224 * convert.c (convert_to_real): Reject non-float inner types.
13226 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13228 * config/aarch64/aarch64-protos.h
13229 (aarch64_simd_immediate_valid_for_move): Remove.
13230 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13231 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13233 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13235 2013-07-11 Steve Ellcey <sellcey@mips.com>
13237 * config/mips/mips.c (mips_conditional_register_usage): Do not
13238 use t[0-7] registers in MIPS16 mode when optimizing for size.
13240 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13242 * config/i386/i386.c (dispatch_function_versions): Fix array
13243 indexing of function_version_info to match actual_versions.
13245 2013-07-11 Teresa Johnson <tejohnson@google.com>
13247 * vec.h (struct va_gc): Move release out-of-line.
13248 (va_gc::release): Call ggc_free on released vec.
13250 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13252 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13253 Require GOT register as additional operand in UNSPEC.
13254 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13255 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13256 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13257 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13258 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13259 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13260 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13262 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13265 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13266 name seen by assembler/linker rather if available.
13268 2013-07-11 Andreas Schwab <schwab@suse.de>
13270 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13272 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13274 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13276 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13278 * doc/tm.texi.in: Move hook documentation to ....
13279 * target.def: ... here.
13281 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13282 text on @hook line.
13283 * doc/tm.texi: Regenerate.
13285 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13288 * doc/invoke.texi: Document -Wconditionally-supported.
13290 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13293 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13296 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13299 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13300 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13301 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13302 Remove duplicate devices.
13303 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13304 * config/avr/t-multilib: Regenerate.
13305 * config/avr/avr-tables.opt: Regenerate.
13306 * doc/avr-mmcu.texi: Regenerate.
13308 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13311 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13313 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13315 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13316 the cost of MULT when optimizing for size.
13318 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13320 * config/cr16/cr16-protos.h: Don't include target.h.
13322 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13324 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13325 adjust register size for TDmode and TFmode for VSX registers.
13327 2013-07-08 Kai Tietz <ktietz@redhat.com>
13330 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13331 hook_bool_const_tree_true.
13333 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13335 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13336 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13337 * config/s390/s390.md: Define FPR*_REGNUM constants.
13338 Fix FPR2_REGNUM constant (18 -> 17).
13339 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13340 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13341 ("trunc<BFP:mode><DFP_ALL:mode>2")
13342 ("trunc<DFP_ALL:mode><BFP:mode>2")
13343 ("*extend<BFP:mode><DFP_ALL:mode>2")
13344 ("*extend<DFP_ALL:mode><BFP:mode>2")
13345 ("extend<BFP:mode><DFP_ALL:mode>2")
13346 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13349 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13351 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13353 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13355 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13356 and cfun_fpr_bit_p to cfun_fpr_save_p.
13357 (s390_frame_area, s390_register_info, s390_frame_info)
13358 (s390_emit_prologue, s390_emit_epilogue)
13359 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13361 * config/s390/s390.h: Define *_REGNUM macros for floating point
13364 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13366 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13368 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13370 PR rtl-optimization/57786
13371 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13372 and break out of the loop when it is set to false.
13374 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13377 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13378 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13379 (const_int 63)) 0)).
13380 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13381 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13382 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13384 PR rtl-optimization/57829
13385 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13386 mask bits outside of mode are just sign-extension from mode to HWI.
13388 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13390 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13391 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13392 adjust_address instead of change_address to keep info about alignment.
13393 (emit_strmov): Remove.
13394 (emit_memmov): New function.
13395 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13396 (expand_movmem_epilogue): Likewise and return updated rtx for
13398 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13399 destination and source.
13400 (decide_alignment): Refactor, handle vector_loop.
13401 (ix86_expand_movmem): Likewise.
13402 (ix86_expand_setmem): Likewise.
13403 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13405 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13407 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13408 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13410 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13412 * config/i386/sse.md (sse_movlhps): Change alternative 3
13413 of operand 2 to "m".
13415 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13418 * config/i386/sse.md (iptr): New mode attribute.
13419 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13420 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13421 (<sse>_vmmul<mode>3): Ditto.
13422 (<sse>_vmdiv<mode>3): Ditto.
13423 (sse_vmrcpv4sf2): Ditto.
13424 (<sse>_vmsqrt<mode>2): Ditto.
13425 (sse_vmrsqrtv4sf2): Ditto.
13426 (<sse>_vm<code><mode>3): Ditto.
13427 (avx_vmcmp<mode>3): Ditto.
13428 (<sse>_vmmaskcmp<mode>3): Ditto.
13429 (<sse>_comi): Ditto.
13430 (<sse>_ucomi): Ditto.
13431 (*xop_vmfrcz_<mode>): Ditto.
13432 (*fmai_fmadd_<mode>): Ditto.
13433 (*fmai_fmsub_<mode>): Ditto.
13434 (*fmai_fnmadd_<mode>): Ditto.
13435 (*fmai_fnmsub_<mode>): Ditto.
13436 (*fma4i_vmfmadd_<mode>): Ditto.
13437 (*fma4i_vmfmsub_<mode>): Ditto.
13438 (*fma4i_vmfnmadd_<mode>): Ditto.
13439 (*fma4i_vmfnmsub_<mode>): Ditto.
13440 (*xop_vmfrcz_<mode>): Ditto.
13441 (sse_cvtps2pi): Ditto.
13442 (sse_cvttps2pi): Ditto.
13443 (sse_cvtss2si): Ditto.
13444 (sse_cvtss2si_2): Ditto.
13445 (sse_cvtss2siq_2): Ditto.
13446 (sse_cvttss2si): Ditto.
13447 (sse_cvttss2siq): Ditto.
13448 (sse_cvtsd2si): Ditto.
13449 (sse_cvtsd2si_2): Ditto.
13450 (sse_cvtsd2siq_2): Ditto.
13451 (sse_cvttsd2si): Ditto.
13452 (sse_cvttsd2siq): Ditto.
13453 (sse_cvtsd2ss): Ditto.
13454 (sse_cvtss2sd): Ditto.
13455 (avx2_pbroadcast<mode>): Ditto.
13456 (avx2_pbroadcast<mode>_1): Ditto.
13457 (*avx_vperm_broadcast_v4sf): Ditto.
13459 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13460 (sse_movlhps): Ditto.
13461 (sse_storehps): Ditto.
13462 (sse_loadhps): Ditto.
13463 (sse_storelps): Ditto.
13464 (sse_loadlps): Ditto.
13465 (*vec_concatv4sf): Ditto.
13466 (*vec_interleave_highv2df): Ditto.
13467 (*vec_interleave_lowv2df): Ditto.
13468 (*vec_extractv2df_1_sse): Ditto.
13469 (*vec_extractv2df_0_sse): Ditto.
13470 (sse2_storelpd): Ditto.
13471 (sse2_loadlpd): Ditto.
13472 (sse2_movsd): Ditto.
13473 (*vec_concatv4si): Ditto.
13474 (vec_concatv2di): Ditto.
13476 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13477 for Intel asm dialect.
13478 (mmx_punpcklwd): Ditto.
13479 (mmx_punpckldq): Ditto.
13481 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13482 for intel assembler dialect.
13484 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13487 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13488 for BUILT_IN_C{LZ,LRSB}*.
13489 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13490 * tree-vrp.c (extract_range_basic): Handle
13491 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13492 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13493 fall thru to code calling set_value*.
13494 * builtins.c (expand_builtin): Remove *IMAX cases.
13495 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13496 if width is bigger than 2*HWI.
13498 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13500 PR rtl-optimization/55342
13501 * lra-int.h (lra_subreg_reload_pseudos): New.
13502 * lra.c: Add undoing optional reloads to the block diagram.
13503 (lra_subreg_reload_pseudos): New.
13504 (lra_optional_reload_pseudos): Change comments.
13505 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13506 lra_optional_reload_pseudos after undo transformations.
13507 * lra-assigns.c (pseudo_prefix_title): New.
13508 (lra_setup_reg_renumber): Use it.
13509 (spill_for): Ditto. Check subreg reload pseudos too.
13510 (assign_by_spills): Consider subreg reload pseudos too.
13511 * lra-constraints.c (simplify_operand_subreg): Use
13512 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13513 (curr_insn_transform): Recognize and do optional reloads.
13514 (undo_optional_reloads): New.
13515 (lra_undo_inheritance): Call undo_optional_reloads.
13517 2013-07-05 Thomas Quinot <quinot@adacore.com>
13519 * tree-complex.c (expand_complex_operations_1): Fix typo.
13521 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13523 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13524 (tune_params): New member 'const vec_costs'.
13525 * config/aarch64/aarch64.c (generic_vector_cost): New.
13526 (generic_tunings): New member 'generic_vector_cost'.
13527 (aarch64_builtin_vectorization_cost): New.
13528 (aarch64_add_stmt_cost): New.
13529 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13530 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13532 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13535 * config/i386/predicates.md (vsib_address_operand): Disallow
13536 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13538 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13540 PR middle-end/55030
13541 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13542 expand_builtin_setjmp_receiver.
13543 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13544 with NULL for the label parameter.
13545 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13546 the frame-pointer. Adjust comments.
13547 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13548 only if LABEL is non-NULL.
13550 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13552 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13553 (ARM_ABI_AAPCS64): Ditto.
13555 (ARM_DEFAULT_ABI): Ditto.
13557 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13559 * config/aarch64/aarch64-builtins.c
13560 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13561 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13563 * config/aarch64/aarch64-simd.md
13564 (aarch64_ld1<VALL:mode>): New.
13565 (aarch64_st1<VALL:mode>): Likewise.
13566 * config/aarch64/arm_neon.h
13567 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13569 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13571 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13573 (ix86_option_override_internal):Turn on all -mavx target flags by
13574 default as they are dependent on AVX anyway.
13576 2013-07-02 Cary Coutant <ccoutant@google.com>
13578 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13579 deterministic hash.
13580 (loc_checksum_ordered): Likewise.
13581 (hash_loc_operands): Remove inline keyword.
13583 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13585 PR tree-optimization/57741
13586 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13587 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13588 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13589 Allow REAL_CST step_exprs if flag_associative_math.
13590 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13592 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13594 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13596 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13598 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13600 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13602 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13604 (iorsi3_insn): Likewise.
13605 (arm_xorsi3): Likewise.
13607 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13609 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13611 (attribute "type"): ... here, and prefix with "wmmx_".
13612 (attribute "core_cycles"): Update for attribute changes.
13613 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13614 (tbcstv4hi): Likewise.
13615 (tbcstv2si): Likewise.
13616 (iwmmxt_iordi3): Likewise.
13617 (iwmmxt_xordi3): Likewise.
13618 (iwmmxt_anddi3): Likewise.
13619 (iwmmxt_nanddi3): Likewise.
13620 (iwmmxt_arm_movdi): Likewise.
13621 (iwmmxt_movsi_insn): Likewise.
13622 (mov<mode>_internal): Likewise.
13623 (and<mode>3_iwmmxt): Likewise.
13624 (ior<mode>3_iwmmxt): Likewise.
13625 (xor<mode>3_iwmmxt): Likewise.
13626 (add<mode>3_iwmmxt): Likewise.
13627 (ssaddv8qi3): Likewise.
13628 (ssaddv4hi3): Likewise.
13629 (ssaddv2si3): Likewise.
13630 (usaddv8qi3): Likewise.
13631 (usaddv4hi3): Likewise.
13632 (usaddv2si3): Likewise.
13633 (sub<mode>3_iwmmxt): Likewise.
13634 (sssubv8qi3): Likewise.
13635 (sssubv4hi3): Likewise.
13636 (sssubv2si3): Likewise.
13637 (ussubv8qi3): Likewise.
13638 (ussubv4hi3): Likewise.
13639 (ussubv2si3): Likewise.
13640 (mulv4hi3_iwmmxt): Likewise.
13641 (smulv4hi3_highpart): Likewise.
13642 (umulv4hi3_highpart): Likewise.
13643 (iwmmxt_wmacs): Likewise.
13644 (iwmmxt_wmacsz): Likewise.
13645 (iwmmxt_wmacu): Likewise.
13646 (iwmmxt_wmacuz): Likewise.
13647 (iwmmxt_clrdi): Likewise.
13648 (iwmmxt_clrv8qi): Likewise.
13649 (iwmmxt_clr4hi): Likewise.
13650 (iwmmxt_clr2si): Likewise.
13651 (iwmmxt_uavgrndv8qi3): Likewise.
13652 (iwmmxt_uavgrndv4hi3): Likewise.
13653 (iwmmxt_uavgv8qi3): Likewise.
13654 (iwmmxt_uavgv4hi3): Likewise.
13655 (iwmmxt_tinsrb): Likewise.
13656 (iwmmxt_tinsrh): Likewise.
13657 (iwmmxt_tinsrw): Likewise.
13658 (iwmmxt_textrmub): Likewise.
13659 (iwmmxt_textrmsb): Likewise.
13660 (iwmmxt_textrmuh): Likewise.
13661 (iwmmxt_textrmsh): Likewise.
13662 (iwmmxt_textrmw): Likewise.
13663 (iwmxxt_wshufh): Likewise.
13664 (eqv8qi3): Likewise.
13665 (eqv4hi3): Likewise.
13666 (eqv2si3): Likewise.
13667 (gtuv8qi3): Likewise.
13668 (gtuv4hi3): Likewise.
13669 (gtuv2si3): Likewise.
13670 (gtv8qi3): Likewise.
13671 (gtv4hi3): Likewise.
13672 (gtv2si3): Likewise.
13673 (smax<mode>3_iwmmxt): Likewise.
13674 (umax<mode>3_iwmmxt): Likewise.
13675 (smin<mode>3_iwmmxt): Likewise.
13676 (umin<mode>3_iwmmxt): Likewise.
13677 (iwmmxt_wpackhss): Likewise.
13678 (iwmmxt_wpackwss): Likewise.
13679 (iwmmxt_wpackdss): Likewise.
13680 (iwmmxt_wpackhus): Likewise.
13681 (iwmmxt_wpackwus): Likewise.
13682 (iwmmxt_wpackdus): Likewise.
13683 (iwmmxt_wunpckihb): Likewise.
13684 (iwmmxt_wunpckihh): Likewise.
13685 (iwmmxt_wunpckihw): Likewise.
13686 (iwmmxt_wunpckilb): Likewise.
13687 (iwmmxt_wunpckilh): Likewise.
13688 (iwmmxt_wunpckilw): Likewise.
13689 (iwmmxt_wunpckehub): Likewise.
13690 (iwmmxt_wunpckehuh): Likewise.
13691 (iwmmxt_wunpckehuw): Likewise.
13692 (iwmmxt_wunpckehsb): Likewise.
13693 (iwmmxt_wunpckehsh): Likewise.
13694 (iwmmxt_wunpckehsw): Likewise.
13695 (iwmmxt_wunpckelub): Likewise.
13696 (iwmmxt_wunpckeluh): Likewise.
13697 (iwmmxt_wunpckeluw): Likewise.
13698 (iwmmxt_wunpckelsb): Likewise.
13699 (iwmmxt_wunpckelsh): Likewise.
13700 (iwmmxt_wunpckelsw): Likewise.
13701 (ror<mode>3): Likewise.
13702 (ashr<mode>3_iwmmxt): Likewise.
13703 (lshr<mode>3_iwmmxt): Likewise.
13704 (ashl<mode>3_iwmmxt): Likewise.
13705 (ror<mode>3_di): Likewise.
13706 (ashr<mode>3_di): Likewise.
13707 (lshr<mode>3_di): Likewise.
13708 (ashl<mode>3_di): Likewise.
13709 (iwmmxt_wmadds): Likewise.
13710 (iwmmxt_wmaddu): Likewise.
13711 (iwmmxt_tmia): Likewise.
13712 (iwmmxt_tmiaph): Likewise.
13713 (iwmmxt_tmiabb): Likewise.
13714 (iwmmxt_tmiatb): Likewise.
13715 (iwmmxt_tmiabt): Likewise.
13716 (iwmmxt_tmiatt): Likewise.
13717 (iwmmxt_tmovmskb): Likewise.
13718 (iwmmxt_tmovmskh): Likewise.
13719 (iwmmxt_tmovmskw): Likewise.
13720 (iwmmxt_waccb): Likewise.
13721 (iwmmxt_wacch): Likewise.
13722 (iwmmxt_waccw): Likewise.
13723 (iwmmxt_waligni): Likewise.
13724 (iwmmxt_walignr): Likewise.
13725 (iwmmxt_walignr0): Likewise.
13726 (iwmmxt_walignr1): Likewise.
13727 (iwmmxt_walignr2): Likewise.
13728 (iwmmxt_walignr3): Likewise.
13729 (iwmmxt_wsadb): Likewise.
13730 (iwmmxt_wsadh): Likewise.
13731 (iwmmxt_wsadbz): Likewise.
13732 (iwmmxt_wsadhz): Likewise.
13733 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13734 (iwmmxt_wabsdiffb): Likewise.
13735 (iwmmxt_wabsdiffh): Likewise.
13736 (iwmmxt_wabsdiffw): Likewise.
13737 (iwmmxt_waddsubhx): Likewise
13738 (iwmmxt_wsubaddhx): Likewise.
13739 (addc<mode>3): Likewise.
13740 (iwmmxt_avg4): Likewise.
13741 (iwmmxt_avg4r): Likewise.
13742 (iwmmxt_wmaddsx): Likewise.
13743 (iwmmxt_wmaddux): Likewise.
13744 (iwmmxt_wmaddsn): Likewise.
13745 (iwmmxt_wmaddun): Likewise.
13746 (iwmmxt_wmulwsm): Likewise.
13747 (iwmmxt_wmulwum): Likewise.
13748 (iwmmxt_wmulsmr): Likewise.
13749 (iwmmxt_wmulumr): Likewise.
13750 (iwmmxt_wmulwsmr): Likewise.
13751 (iwmmxt_wmulwumr): Likewise.
13752 (iwmmxt_wmulwl): Likewise.
13753 (iwmmxt_wqmulm): Likewise.
13754 (iwmmxt_wqmulwm): Likewise.
13755 (iwmmxt_wqmulmr): Likewise.
13756 (iwmmxt_wqmulwmr): Likewise.
13757 (iwmmxt_waddbhusm): Likewise.
13758 (iwmmxt_waddbhusl): Likewise.
13759 (iwmmxt_wqmiabb): Likewise.
13760 (iwmmxt_wqmiabt): Likewise.
13761 (iwmmxt_wqmiatb): Likewise.
13762 (iwmmxt_wqmiatt): Likewise.
13763 (iwmmxt_wqmiabbn): Likewise.
13764 (iwmmxt_wqmiabtn): Likewise.
13765 (iwmmxt_wqmiatbn): Likewise.
13766 (iwmmxt_wqmiattn): Likewise.
13767 (iwmmxt_wmiabb): Likewise.
13768 (iwmmxt_wmiabt): Likewise.
13769 (iwmmxt_wmiatb): Likewise.
13770 (iwmmxt_wmiatt): Likewise.
13771 (iwmmxt_wmiabbn): Likewise.
13772 (iwmmxt_wmiabtn): Likewise.
13773 (iwmmxt_wmiatbn): Likewise.
13774 (iwmmxt_wmiattn): Likewise.
13775 (iwmmxt_wmiawbb): Likewise.
13776 (iwmmxt_wmiawbt): Likewise.
13777 (iwmmxt_wmiawtb): Likewise.
13778 (iwmmxt_wmiawtt): Likewise.
13779 (iwmmxt_wmiawbbn): Likewise.
13780 (iwmmxt_wmiawbtn): Likewise.
13781 (iwmmxt_wmiawtbn): Likewise.
13782 (iwmmxt_wmiawttn): Likewise.
13783 (iwmmxt_wmerge): Likewise.
13784 (iwmmxt_tandc<mode>3): Likewise.
13785 (iwmmxt_torc<mode>3): Likewise.
13786 (iwmmxt_torvsc<mode>3): Likewise.
13787 (iwmmxt_textrc<mode>3): Likewise.
13788 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13789 (wmmxt_pack): Likewise.
13790 (wmmxt_mult_c1): Likewise.
13791 (wmmxt_mult_c2): Likewise.
13792 (wmmxt_alu_c1): Likewise.
13793 (wmmxt_alu_c2): Likewise.
13794 (wmmxt_alu_c3): Likewise.
13795 (wmmxt_transfer_c1): Likewise.
13796 (wmmxt_transfer_c2): Likewise.
13797 (wmmxt_transfer_c3): Likewise.
13798 (marvell_f_iwmmxt_wstr): Likewise.
13799 (marvell_f_iwmmxt_wldr): Likewise.
13801 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13803 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13805 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13808 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13809 * lra-constraints.c (need_for_split_p): Check call used hard regs
13810 living through calls.
13812 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13813 call used regs for call insn.
13815 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13818 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13819 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13821 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13823 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13825 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13827 * lra-constraints.c (need_for_split_p): Check call used hard regs
13828 living through calls.
13830 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13833 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13834 to tie with any other modes. Eliminate Altivec vector mode tests,
13835 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13836 code, to return 0 if testing MODE2 for a condition, if we've
13837 already tested MODE1 for the same condition.
13839 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13841 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13844 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13846 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13847 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13849 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13851 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13853 (aarch64_symbolic_constant_p): Remove.
13854 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13855 static. Fix line length and white space.
13856 (aarch64_symbolic_constant_p): Remove.
13857 * config/aarch64/predicates.md (aarch64_valid_symref):
13858 Use aarch64_classify_symbol_expression.
13860 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13862 * config/arm/constraints.md (Ts): New constraint.
13863 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13865 (compare_scc): Use "Ts" constraint for operand 0.
13866 (ior_scc_scc): Likewise.
13867 (and_scc_scc): Likewise.
13868 (and_scc_scc_nodom): Likewise.
13869 (ior_scc_scc_cmp): Likewise for operand 7.
13870 (and_scc_scc_cmp): Likewise.
13871 * config/arm/thumb2.md (thumb2_movsi_insn):
13872 Add alternatives for 16-bit encodings.
13873 (thumb2_movhi_insn): Likewise.
13874 (thumb2_movsicc_insn): Likewise.
13875 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13876 (thumb2_negscc): Use "Ts" constraint.
13877 Move mvn instruction outside cond_exec block.
13878 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13879 for 16-bit encodings.
13881 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13883 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13885 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13886 (mulsi3subsi): Likewise.
13887 (mulsidi3adddi): Likewise.
13888 (mulsidi3_v6): Likewise.
13889 (umulsidi3_v6): Likewise.
13890 (umulsidi3adddi_v6): Likewise.
13891 (smulsi3_highpart_v6): Likewise.
13892 (umulsi3_highpart_v6): Likewise.
13893 (mulhisi3tb): Likewise.
13894 (mulhisi3bt): Likewise.
13895 (mulhisi3tt): Likewise.
13896 (maddhisi4): Likewise.
13897 (maddhisi4tb): Likewise.
13898 (maddhisi4tt): Likewise.
13899 (maddhidi4): Likewise.
13900 (maddhidi4tb): Likewise.
13901 (maddhidi4tt): Likewise.
13902 (zeroextractsi_compare0_scratch): Likewise.
13903 (insv_zero): Likewise.
13904 (insv_t2): Likewise.
13905 (anddi_notzesidi_di): Likewise.
13906 (anddi_notsesidi_di): Likewise.
13907 (andsi_notsi_si): Likewise.
13908 (iordi_zesidi_di): Likewise.
13909 (xordi_zesidi_di): Likewise.
13910 (andsi_iorsi3_notsi): Likewise.
13911 (smax_0): Likewise.
13912 (smax_m1): Likewise.
13913 (smin_0): Likewise.
13914 (not_shiftsi): Likewise.
13915 (unaligned_loadsi): Likewise.
13916 (unaligned_loadhis): Likewise.
13917 (unaligned_loadhiu): Likewise.
13918 (unaligned_storesi): Likewise.
13919 (unaligned_storehi): Likewise.
13920 (extv_reg): Likewise.
13921 (extzv_t2): Likewise.
13922 (divsi3): Likewise.
13923 (udivsi3): Likewise.
13924 (arm_zero_extendhisi2addsi): Likewise.
13925 (arm_zero_extendqisi2addsi): Likewise.
13926 (compareqi_eq0): Likewise.
13927 (arm_extendhisi2_v6): Likewise.
13928 (arm_extendqisi2addsi): Likewise.
13929 (arm_movt): Likewise.
13930 (thumb2_ldrd): Likewise.
13931 (thumb2_ldrd_base): Likewise.
13932 (thumb2_ldrd_base_neg): Likewise.
13933 (thumb2_strd): Likewise.
13934 (thumb2_strd_base): Likewise.
13935 (thumb2_strd_base_neg): Likewise.
13936 (arm_negsi2): Add alternative for 16-bit encoding.
13937 (arm_one_cmplsi2): Likewise.
13939 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13941 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13942 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13943 (movdfcc): Likewise.
13944 * config/arm/vfp.md (*thumb2_movsf_vfp):
13945 Disable predication for arm_restrict_it.
13946 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13947 (*thumb2_movdfcc_vfp): Likewise.
13948 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13949 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13950 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13951 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13952 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13953 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13954 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13955 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13956 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13957 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13958 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13959 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13960 Disable predication for arm_restrict_it.
13962 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13964 * config/i386/bmiintrin.h (_bextr_u32): New.
13965 (_bextr_u64): Ditto.
13967 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13969 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13970 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13971 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13972 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13973 * config/mips/n32-elf.h: ...this new file.
13975 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
13978 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
13979 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
13981 2013-06-27 Catherine Moore <clm@codesourcery.com>
13983 * config/mips/mips-tables.opt: Regenerate.
13984 * config/mips/mips-cpus.def: Add m14ke and m14kec.
13985 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
13986 * doc/invoke.texi: Add -m14kc.
13988 2013-06-27 Jakub Jelinek <jakub@redhat.com>
13991 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
13992 constraints of operand 1 and 2.
13995 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
13996 to match RTL canonicalization. Swap predicates and
13997 constraints of operand 1 and 2.
13999 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
14001 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
14002 Process OP_INOUT regs for splitting too.
14004 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14006 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
14007 decl before the loop, initialize to NULL.
14008 (vectorizable_load): Initialize ptr_incr to NULL.
14010 2013-06-27 Martin Jambor <mjambor@suse.cz>
14013 * ipa-ref.h (ipa_maybe_record_reference): Declare.
14014 * ipa-ref.c (ipa_maybe_record_reference): New function.
14015 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
14016 * ipa-cp.c (create_specialized_node): Record potential references from
14018 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
14020 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14022 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
14023 parameter 'mode' of type 'enum machine_mode mode'; change to pass
14024 'mode' to force_reg.
14025 (aarch64_add_offset): Update calls to aarch64_force_temporary.
14026 (aarch64_expand_mov_immediate): Likewise.
14028 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14030 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
14031 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
14033 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14035 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
14036 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14037 (struct machine_function): Add tbegin_p.
14038 (s390_canonicalize_comparison): Fold CC mode compares to
14039 conditional jump if possible.
14040 (s390_emit_jump): Return the emitted jump.
14041 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
14042 Handle CCRAWmode compares.
14043 (s390_option_override): Default to -mhtm if available.
14044 (s390_reg_clobbered_rtx): Handle floating point regs as well.
14045 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
14046 FPRs instead of df_regs_ever_live_p.
14047 (s390_optimize_nonescaping_tx): New function.
14048 (s390_init_frame_layout): Extend clobbered_regs array to cover
14050 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
14051 (s390_expand_tbegin): New function.
14052 (enum s390_builtin): New enum definition.
14053 (code_for_builtin): New array definition.
14054 (s390_init_builtins): New function.
14055 (s390_expand_builtin): New function.
14056 (TARGET_INIT_BUILTINS): Define.
14057 (TARGET_EXPAND_BUILTIN): Define.
14058 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
14059 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
14060 (s390_alc_comparison): Likewise.
14061 * config/s390/s390-modes.def: Add CCRAWmode.
14062 * config/s390/s390.h (processor_flags): Add PF_TX.
14063 (TARGET_CPU_HTM): Define macro.
14064 (TARGET_HTM): Define macro.
14065 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
14066 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
14067 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14068 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
14069 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
14070 (TBEGIN_MASK, TBEGINC_MASK): New constants.
14071 ("*cc_to_int"): Move up.
14072 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
14073 constants other than 0.
14074 ("*ccraw_to_int"): New insn and splitter definition.
14075 ("tbegin", "tbegin_nofloat", "tbegin_retry")
14076 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
14077 ("tx_assist"): New expander.
14078 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
14079 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
14080 * config/s390/s390.opt: Add -mhtm option.
14081 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
14082 * config/s390/htmxlintrin.h: New file.
14083 * config/s390/htmintrin.h: New file.
14084 * config/s390/s390intrin.h: New file.
14085 * doc/extend.texi: Document htm builtins.
14086 * config.gcc: Add the new header files to extra_headers.
14088 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
14090 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
14091 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
14093 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
14094 Pat Haugen <pthaugen@us.ibm.com>
14095 Peter Bergner <bergner@vnet.ibm.com>
14097 * config/rs6000/power8.md: New.
14098 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
14099 setting for power8 entry.
14100 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
14101 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
14102 test for Power4/Power5 only.
14103 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
14105 (force_new_group): Adjust comment.
14106 * config/rs6000/rs6000.md: Include power8.md.
14108 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14110 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14111 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14113 (tune_params): Update comment.
14114 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14115 (arm_max_conditional_execute): New function.
14116 (thumb2_final_prescan_insn): Use max_insn_skipped and
14117 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14119 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14121 PR tree-optimization/57705
14122 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14123 SSA_NAME step, provided that it is not defined inside the loop.
14124 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14125 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14127 2013-06-25 Martin Jambor <mjambor@suse.cz>
14129 PR middle-end/57670
14130 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14131 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14133 (ipa_note_param_call): Initialize member_ptr flag.
14134 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14135 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14136 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14137 (ipa_read_indirect_edge_info): Likewise.
14139 2013-06-25 Richard Biener <rguenther@suse.de>
14141 PR middle-end/56977
14142 * passes.c (init_optimization_passes): Move pass_fold_builtins
14143 and pass_dce earlier with -Og.
14145 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14147 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14148 <BIT_FIELD_REF>: Remove trailing TAB.
14149 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14152 2013-06-24 Martin Jambor <mjambor@suse.cz>
14154 PR tree-optimization/57358
14155 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14156 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14157 (ipa_analyze_params_uses): Generate pessimistic info when true.
14159 2013-06-24 Martin Jambor <mjambor@suse.cz>
14161 PR tree-optimization/57539
14162 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14163 global.inlined_to of the new node to it. All callers changed.
14164 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14165 inlining_into, pass it to cgraph_clone_node.
14166 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14167 ipa_free_edge_args_substructures.
14168 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14169 rdesc linked list. Do not assert rdesc edges have inlined caller.
14170 Assert we have found an rdesc in the rdesc list.
14172 2013-06-24 Richard Biener <rguenther@suse.de>
14174 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14175 (pointer_set_lookup): Declare.
14176 (class pointer_map): New template class implementing a
14177 generic pointer to T map.
14178 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14179 pointer_map<T>::contains, pointer_map<T>::insert,
14180 pointer_map<T>::traverse): New functions.
14181 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14182 (pointer_set_lookup): New function.
14183 (pointer_set_contains): Use pointer_set_lookup.
14184 (pointer_set_insert): Likewise.
14185 (insert_aux): Remove.
14186 (struct pointer_map_t): Embed a pointer_set_t.
14187 (pointer_map_create): Adjust.
14188 (pointer_map_destroy): Likewise.
14189 (pointer_map_contains): Likewise.
14190 (pointer_map_insert): Likewise.
14191 (pointer_map_traverse): Likewise.
14192 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14193 pointer_map<unsigned> instead of a pointer_map_t.
14194 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14195 (streamer_tree_cache_lookup): Likewise.
14196 (streamer_tree_cache_create): Likewise.
14197 (streamer_tree_cache_delete): Likewise.
14198 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14199 pointer_map<unsigned> instead of a pointer_map_t.
14200 (lto_init_tree_ref_encoder): Adjust.
14201 (lto_destroy_tree_ref_encoder): Likewise.
14202 * lto-section-out.c (lto_output_decl_index): Likewise.
14203 (lto_record_function_out_decl_state): Likewise.
14204 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14206 2013-06-24 Richard Biener <rguenther@suse.de>
14208 PR tree-optimization/57488
14209 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14211 2013-06-24 Alan Modra <amodra@gmail.com>
14213 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14214 (gen_easy_altivec_constant): Likewise.
14215 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14216 easy_vector_constant_msb): Likewise.
14218 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14221 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14222 add missing return true.
14224 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14227 * config/sh/predicates.md (general_extend_operand): Invoke
14228 general_movsrc_operand for memory operands.
14229 (general_movsrc_operand): Allow reg+reg addressing, do not use
14230 general_operand for memory operands.
14232 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14234 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14235 when current target options does not apply.
14236 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14237 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14238 * config/i386/bmiintrin.h: Pass appropriate target
14239 attributes to header.
14240 * config/i386/mmintrin.h: Ditto.
14241 * config/i386/nmmintrin.h: Ditto.
14242 * config/i386/avx2intrin.h: Ditto.
14243 * config/i386/fxsrintrin.h: Ditto.
14244 * config/i386/tbmintrin.h: Ditto.
14245 * config/i386/xsaveintrin.h: Ditto.
14246 * config/i386/f16cintrin.h: Ditto.
14247 * config/i386/xtestintrin.h: Ditto.
14248 * config/i386/xsaveoptintrin.h: Ditto.
14249 * config/i386/bmi2intrin.h: Ditto.
14250 * config/i386/lzcntintrin.h: Ditto.
14251 * config/i386/smmintrin.h: Ditto.
14252 * config/i386/wmmintrin.h: Ditto.
14253 * config/i386/x86intrin.h: Remove all header include guards.
14254 * config/i386/prfchwintrin.h: Ditto.
14255 * config/i386/pmmintrin.h: Ditto.
14256 * config/i386/tmmintrin.h: Ditto.
14257 * config/i386/xmmintrin.h: Ditto.
14258 * config/i386/popcntintrin.h: Ditto.
14259 * config/i386/rdseedintrin.h: Ditto.
14260 * config/i386/ammintrin.h: Ditto.
14261 * config/i386/emmintrin.h: Ditto.
14262 * config/i386/immintrin.h: Remove all header include guards.
14263 * config/i386/fma4intrin.h: Ditto.
14264 * config/i386/lwpintrin.h: Ditto.
14265 * config/i386/xopintrin.h: Ditto.
14266 * config/i386/ia32intrin.h: Ditto.
14267 * config/i386/avxintrin.h: Ditto.
14268 * config/i386/rtmintrin.h: Ditto.
14269 * config/i386/fmaintrin.h: Ditto.
14270 * config/i386/mm3dnow.h: Ditto.
14272 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14274 * common/config/i386/i386-common.c: Handle LZCNT.
14276 2013-06-22 Andi Kleen <ak@linux.intel.com>
14278 * doc/extend.texi: Use __atomic_store_n instead of
14279 __atomic_store in HLE example.
14281 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14283 * config/sh/sh.c: Remove <cstdlib> workaround.
14285 2013-06-21 Andi Kleen <ak@linux.intel.com>
14287 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14289 2013-06-21 Andi Kleen <ak@linux.intel.com>
14291 * doc/extend.texi: Document that __atomic_clear and
14292 __atomic_test_and_set should only be used with bool.
14294 2013-06-20 Jan Hubicka <jh@suse.cz>
14296 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14297 types_same_for_odr.
14298 * tree.c (decls_same_for_odr): New function.
14299 (same_for_edr): New function.
14300 (types_same_for_odr): New function.
14301 (get_binfo_at_offset): Use it.
14302 * tree.h (types_same_for_odr): Declare.
14304 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14305 Jason Merrill <jason@redhat.com>
14307 * system.h: Include <cstdlib> as well as <stdlib.h>.
14309 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14312 * config/i386/i386.c (construct_container): Report error if
14313 long double is used with disabled x87 float returns.
14315 2013-06-20 Jan Hubicka <jh@suse.cz>
14317 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14319 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14321 PR rtl-optimization/57425
14322 PR rtl-optimization/57569
14323 * alias.c (write_dependence_p): Remove parameters mem_mode and
14324 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14325 Changed all callers.
14326 (canon_anti_dependence): Get comments and semantics in sync.
14327 Add parameter mem_canonicalized. Changed all callers.
14328 * rtl.h (canon_anti_dependence): Update prototype.
14330 2013-06-20 Richard Biener <rguenther@suse.de>
14332 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14333 byte case, inline streamer_read_uchar and defer section
14336 2013-06-20 Richard Biener <rguenther@suse.de>
14338 PR tree-optimization/57584
14339 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14340 SSA names into the expanded expression that take part in
14341 abnormal coalescing.
14343 2013-06-19 Sharad Singhai <singhai@google.com>
14345 * gcov.c (print_usage): Handle new option.
14346 (process_args): Ditto.
14347 (get_gcov_intermediate_filename): New function.
14348 (output_intermediate_file): New function.
14349 (output_gcov_file): New function
14350 (generate_results): Handle new option.
14351 (release_function): Relase demangled name.
14352 (read_graph_file): Handle demangled name.
14353 (output_lines): Ditto.
14354 * doc/gcov.texi: Document gcov intermediate format.
14356 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14359 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14360 (lra_emit_add): Use the functions. Add comment about Y as an
14363 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14366 * collect2.c (collect_atexit): New.
14367 (collect_exit): Delete.
14368 (main): Register collect_atexit with atexit.
14369 (collect_wait): Change collect_exit to exit.
14371 * collect2.h (collect_exit): Delete.
14372 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14374 2013-06-19 Wei Mi <wmi@google.com>
14376 PR rtl-optimization/57518
14377 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14378 if regno is used in paradoxical subreg.
14379 (update_equiv_regs): Check pdx_subregs[regno] before
14380 set a reg to be equivalent with a mem.
14382 2013-06-19 Matthias Klose <doko@ubuntu.com>
14385 * file-find.h (find_a_file): Add a mode parameter.
14386 * file-find.c (find_a_file): Likewise.
14387 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14388 with X_OK for the executables.
14389 * collect2.c (main): Call find_a_file with X_OK.
14391 2013-06-19 Steve Ellcey <sellcey@mips.com>
14394 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14395 Use NEXT_INSN instead of next_real_insn.
14397 2013-06-19 Jan Hubicka <jh@suse.cz>
14399 * cgraph.h (const_value_known_p): Replace by ...
14400 (ctor_for_folding): .. this one.
14401 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14402 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14403 * expr.c (expand_expr_real_1): Likewise.
14404 (string_constant): Likewise.
14405 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14406 * ipa.c (process_references): Likewise.
14407 (symtab_remove_unreachable_nodes): Likewise.
14408 * ipa-inline-analysis.c (param_change_prob): Likewise.
14409 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14410 (get_base_constructor): Likwise.
14411 * varpool.c (varpool_remove_node): Likewise.
14412 (varpool_remove_initializer): LIkewise.
14413 (dump_varpool_node): LIkwise.
14414 (const_value_known_p): Rewrite to ...
14415 (ctor_for_folding): ... this one.
14417 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14420 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14421 PERSONALITY in $PATH derived prefixes.
14423 2013-06-19 Jeff Law <law@redhat.com>
14425 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14428 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14429 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14432 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14434 * config/arm/vfp.md: Move VFP instruction classification documentation
14436 * config/arm/arm.md: ... here. Update instruction classification
14439 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14441 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14442 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14443 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14445 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14448 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14449 with NEXT_INSN. Use tablejump_p to check for jump table data
14452 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14455 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14456 that now in C++ the value is correct per the C++ standards.
14458 2013-06-19 Richard Biener <rguenther@suse.de>
14460 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14461 for global context.
14463 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14466 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14469 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14470 with next_active_insn.
14472 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14474 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14475 functions are inlined during failures to flag an error.
14476 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14477 in early inline pass.
14479 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14481 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14484 2013-06-18 Julian Brown <julian@codesourcery.com>
14486 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14487 Permit virtual register pre-reload if !strict.
14488 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14490 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14492 * config/arm/neon.md (movmisalign<mode>): Use
14493 neon_perm_struct_or_reg_operand instead of
14494 neon_struct_or_register_operand.
14495 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14496 neon_permissive_struct_operand instead of neon_struct_operand.
14497 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14498 neon_vector_mem_operand.
14499 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14500 neon_vector_mem_operand.
14501 (neon_permissive_struct_operand): New.
14502 (neon_struct_or_register_operand): Rename to...
14503 (neon_perm_struct_or_reg_operand): This. Adjust call to
14504 neon_vector_mem_operand.
14506 2013-06-18 Richard Biener <rguenther@suse.de>
14508 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14509 * lto-streamer.h: Include pointer-set.h.
14510 (struct lto_decl_slot): Remove.
14511 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14512 Remove next_index entry.
14513 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14514 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14515 (lto_init_tree_ref_encoder): Adjust.
14516 (lto_destroy_tree_ref_encoder): Likewise.
14517 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14518 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14519 (lto_output_decl_index): Adjust.
14520 (lto_new_out_decl_state): Likewise.
14521 (lto_record_function_out_decl_state): Likewise.
14522 * lto-streamer-out.c (copy_function): Likewise.
14524 2013-06-18 Richard Biener <rguenther@suse.de>
14526 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14527 * cgraphunit.c: Include cfgloop.h.
14528 (init_lowered_empty_function): Initialize the loop tree.
14529 (assemble_thunk): Insert new BBs into loops.
14531 2013-06-18 Richard Biener <rguenther@suse.de>
14533 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14534 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14535 the map from cache entry to cache index optional.
14536 (streamer_tree_cache_replace_tree): Adjust accordingly.
14537 (streamer_tree_cache_append): Likewise.
14538 (streamer_tree_cache_delete): Likewise.
14539 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14540 streamer cache map from cache entry to cache index.
14541 * lto-streamer-out.c (create_output_block): Adjust.
14543 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14545 * config/arm/arm.md (attribute "insn"): Move multiplication and
14546 division attributes to...
14547 (attribute "type"): ... here. Remove mult.
14548 (attribute "mul32"): New attribute.
14549 (attribute "mul64"): Add umaal.
14550 (*arm_mulsi3): Update attributes.
14551 (*arm_mulsi3_v6): Likewise.
14552 (*thumb_mulsi3): Likewise.
14553 (*thumb_mulsi3_v6): Likewise.
14554 (*mulsi3_compare0): Likewise.
14555 (*mulsi3_compare0_v6): Likewise.
14556 (*mulsi_compare0_scratch): Likewise.
14557 (*mulsi_compare0_scratch_v6): Likewise.
14558 (*mulsi3addsi): Likewise.
14559 (*mulsi3addsi_v6): Likewise.
14560 (*mulsi3addsi_compare0): Likewise.
14561 (*mulsi3addsi_compare0_v6): Likewise.
14562 (*mulsi3addsi_compare0_scratch): Likewise.
14563 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14564 (*mulsi3subsi): Likewise.
14565 (*mulsidi3adddi): Likewise.
14566 (*mulsi3addsi_v6): Likewise.
14567 (*mulsidi3adddi_v6): Likewise.
14568 (*mulsidi3_nov6): Likewise.
14569 (*mulsidi3_v6): Likewise.
14570 (*umulsidi3_nov6): Likewise.
14571 (*umulsidi3_v6): Likewise.
14572 (*umulsidi3adddi): Likewise.
14573 (*umulsidi3adddi_v6): Likewise.
14574 (*smulsi3_highpart_nov6): Likewise.
14575 (*smulsi3_highpart_v6): Likewise.
14576 (*umulsi3_highpart_nov6): Likewise.
14577 (*umulsi3_highpart_v6): Likewise.
14578 (mulhisi3): Likewise.
14579 (*mulhisi3tb): Likewise.
14580 (*mulhisi3bt): Likewise.
14581 (*mulhisi3tt): Likewise.
14582 (maddhisi4): Likewise.
14583 (*maddhisi4tb): Likewise.
14584 (*maddhisi4tt): Likewise.
14585 (maddhidi4): Likewise.
14586 (*maddhidi4tb): Likewise.
14587 (*maddhidi4tt): Likewise.
14588 (divsi3): Likewise.
14589 (udivsi3): Likewise.
14590 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14591 (thumb2_mulsi_short_compare0): Likewise.
14592 (thumb2_mulsi_short_compare0_scratch): Likewise.
14593 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14594 (1020mult2): Likewise.
14595 (1020mult3): Likewise.
14596 (1020mult4): Likewise.
14597 (1020mult5): Likewise.
14598 (1020mult6): Likewise.
14599 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14601 (cortex_a15_mult64): Likewise.
14602 (cortex_a15_sdiv): Likewise.
14603 (cortex_a15_udiv): Likewise.
14604 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14610 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14611 (pj4_ir_div): Likewise.
14612 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14613 (11_mult2): Likewise.
14614 (11_mult3): Likewise.
14615 (11_mult4): Likewise.
14616 (11_mult5): Likewise.
14617 (11_mult6): Likewise.
14618 (11_mult7): Likewise.
14619 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14620 (cortex_a8_mla): Likewise.
14621 (cortex_a8_mull): Likewise.
14622 (cortex_a8_smulwy): Likewise.
14623 (cortex_a8_smlald): Likewise.
14624 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14625 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14626 (cortex_r4_mul_3): Likewise.
14627 (cortex_r4_mla_4): Likewise.
14628 (cortex_r4_mla_3): Likewise.
14629 (cortex_r4_smlald): Likewise.
14630 (cortex_r4_mull): Likewise.
14631 (cortex_r4_sdiv): Likewise.
14632 (cortex_r4_udiv): Likewise.
14633 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14634 (cortex_a7_idiv): Likewise.
14635 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14636 (9_mult2): Likewise.
14637 (9_mult3): Likewise.
14638 (9_mult4): Likewise.
14639 (9_mult5): Likewise.
14640 (9_mult6): Likewise.
14641 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14642 (cortex_a53_sdiv): Likewise.
14643 (cortex_a53_udiv): Likewise.
14644 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14645 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14646 (mp626_mult2): Likewise.
14647 (mp626_mult3): Likewise.
14648 (mp626_mult4): Likewise.
14649 * config/arm/fa526.md (526_mult1): Update attribute change.
14650 (526_mult2): Likewise.
14651 * config/arm/arm-generic.md (mult): Update attribute change.
14652 (mult_ldsched_strongarm): Likewise.
14653 (mult_ldsched): Likewise.
14654 (multi_cycle): Likewise.
14655 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14656 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14657 (606te_mult2): Likewise.
14658 (606te_mult3): Likewise.
14659 (606te_mult4): Likewise.
14660 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14661 (cortex_a9_mac16): Likewise.
14662 (cortex_a9_multiply): Likewise.
14663 (cortex_a9_mac): Likewise.
14664 (cortex_a9_multiply_long): Likewise.
14665 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14666 (626te_mult2): Likewise.
14667 (626te_mult3): Likewise.
14668 (626te_mult4): Likewise.
14670 2013-06-18 Richard Biener <rguenther@suse.de>
14673 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14675 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14678 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14679 with next_active_insn.
14681 2013-06-18 Alan Modra <amodra@gmail.com>
14683 * config/rs6000/rs6000.h (enum data_align): New.
14684 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14685 (DATA_ABI_ALIGNMENT): Define.
14686 (CONSTANT_ALIGNMENT): Correct comment.
14687 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14688 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14690 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14692 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14693 ATTRIBUTE_UNUSED marking.
14695 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14697 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14698 alternative and update.
14699 (aarch64_dup_lanedi): Delete.
14700 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14701 * config/aarch64/aarch64-simd-builtins.def: Update.
14703 2013-06-17 Richard Biener <rguenther@suse.de>
14705 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14706 (lto_input_scc): Declare.
14707 (lto_input_tree_1): Likewise.
14708 (struct lto_stats_d): Add num_tree_bodies_output and
14709 num_pickle_refs_output.
14710 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14711 (lto_read_tree_1): Split out from ...
14712 (lto_read_tree): ... this.
14713 (lto_input_scc): New function.
14714 (lto_input_tree_1): Split out from ...
14715 (lto_input_tree): ... this. Handle LTO_tree_scc.
14716 (lto_data_in_create): Create the streamer cache without hashes.
14717 * lto-streamer-out.c (create_output_block): Create the streamer
14718 cache with hashes when not doing WPA.
14719 (lto_write_tree_1): Split out from ...
14720 (lto_write_tree): ... this.
14721 (get_symbol_initial_value): New function.
14722 (lto_output_tree_1): Split out from ...
14723 (lto_output_tree): ... this. Write trees as series of SCCs
14724 using a DFS walk via DFS_write_tree.
14725 (struct sccs, struct scc_entry): New types.
14726 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14727 (DFS_write_tree_body): New function.
14728 (DFS_write_tree): Likewise.
14729 (hash_tree): Likewise.
14730 (scc_entry_compare): Likewise.
14731 (hash_scc): Likewise.
14732 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14733 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14734 TREE_CHAIN as regular reference.
14735 (streamer_read_integer_cst): Remove.
14736 (streamer_get_pickled_tree): Adjust.
14737 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14738 of DECL_EXTERNALs in BLOCK_VARS for now.
14739 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14741 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14742 Add hash value argument and record that if hashes are recorded
14744 (streamer_tree_cache_insert_1): Adjust.
14745 (streamer_tree_cache_insert): Likewise.
14746 (streamer_tree_cache_insert_at): Rename to ...
14747 (streamer_tree_cache_replace_tree): ... this and adjust.
14748 (streamer_tree_cache_append): Adjust.
14749 (record_common_node): Likewise.
14750 (streamer_tree_cache_create): Add argument whether to
14751 record hash values together with trees.
14752 (streamer_tree_cache_delete): Adjust.
14753 * tree-streamer.h (struct streamer_tree_cache_d): Add
14755 (streamer_read_integer_cst): Remove.
14756 (streamer_tree_cache_insert): Adjust.
14757 (streamer_tree_cache_append): Likewise.
14758 (streamer_tree_cache_insert_at): Rename to ...
14759 (streamer_tree_cache_replace_tree): ... this and adjust.
14760 (streamer_tree_cache_create): Add argument whether to record hashes.
14761 (streamer_tree_cache_get): Rename to ...
14762 (streamer_tree_cache_get_tree): ... this.
14763 (streamer_tree_cache_get_hash): New function.
14764 * tree.c (cache_integer_cst): New function.
14765 * tree.h (cache_integer_cst): Declare.
14766 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14767 * lto-symtab.c (lto_varpool_replace_node): Only release
14768 DECL_INITIAL of non-prevailing decls.
14769 * varpool.c (varpool_remove_initializer): Do not release
14770 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14772 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14774 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14775 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14776 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14777 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14778 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14779 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14780 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14781 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14782 instead of TARGET_64BIT.
14783 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14784 Require ISA_HAS_<D>DIV.
14786 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14788 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14789 (mips*-*-linux*): Move default with_llsc setting to where other
14791 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14793 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14794 Likewise. Remove default with_tune setting. Move default float
14795 setting to its own block. Handle with_llsc in the same block as above.
14797 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14799 PR rtl-optimization/57425
14800 PR rtl-optimization/57569
14801 * alias.c (write_dependence_p): Add new parameters mem_mode,
14802 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14803 Changed all callers.
14804 (canon_anti_dependence): New function.
14805 * cse.c (check_dependence): Use canon_anti_dependence.
14806 * cselib.c (cselib_invalidate_mem): Likewise.
14807 * rtl.h (canon_anti_dependence): Declare.
14809 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14811 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14812 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14813 ".set mips3" for 64-bit targets.
14815 2013-06-15 Dehao Chen <dehao@google.com>
14817 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14818 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14819 (gimple_check_call_args): Likewise.
14820 * value-prof.c (check_ic_target): Likewise.
14821 * ipa-inline.c (early_inliner): Likewise.
14822 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14823 * cgraph.c (cgraph_create_edge_1): Likewise.
14824 (cgraph_make_edge_direct): Likewise.
14826 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14829 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14830 rs6000_output_move_128bit to handle emitting quad memory
14831 operations. Set attribute length to 8 bytes.
14833 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14835 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14837 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14838 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14839 (aarch64_<su>mlsl<mode>): Likewise.
14841 2013-06-14 Mike Stump <mikestump@comcast.net>
14843 * Makefile.in (TARGET_H): Add insn-codes.h.
14845 2013-06-14 Alan Modra <amodra@gmail.com>
14847 PR middle-end/57134
14848 PR middle-end/57586
14849 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14850 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14851 bitfield expansion when EXPAND_MEMORY.
14852 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14854 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14856 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14857 test for clearing quad memory on 32-bit later.
14859 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14861 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14862 (fold_negate_expr): Likewise.
14863 (fold_real_zero_addition_p): Handle vectors.
14864 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14866 2013-06-14 Alan Modra <amodra@gmail.com>
14868 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14870 2013-06-13 Jan Hubicka <jh@suse.cz>
14872 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14873 Local comdats are not externally visible.
14874 * symtab.c (dump_symtab_base): Dump externally visible.
14875 (verify_symtab_base): Verify back links in the symtab hash.
14877 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14879 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14880 CONVERT_EXPR as equal nodes.
14882 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14884 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14886 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14888 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14889 Generalize to complex and vector.
14890 * tree.c (build_all_ones_cst): New function.
14891 * tree.h (build_all_ones_cst): Declare it.
14893 2013-06-13 Alan Modra <amodra@gmail.com>
14895 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14896 * config/rs6000/rs6000.md (signbittf2): New insn.
14897 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14898 (abstf2_internal, cmptf_internal2): Likewise.
14899 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14901 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14902 Pat Haugen <pthaugen@us.ibm.com>
14903 Peter Bergner <bergner@vnet.ibm.com>
14905 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14906 power8 byte, half-word, and quad-word atomic instructions.
14907 (emit_store_conditional): Likewise.
14908 (rs6000_expand_atomic_compare_and_swap): Likewise.
14909 (rs6000_expand_atomic_op): Likewise.
14911 * config/rs6000/sync.md (larx): Add new modes for power8.
14913 (AINT): New mode iterator to include TImode as well as normal
14914 integer modes on power8.
14915 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14916 that VSX registers are not considered. Use AINT mode iterator
14917 instead of INT1 to allow inclusion of quad word atomic operations
14919 (load_locked<mode>): Likewise.
14920 (store_conditional<mode>): Likewise.
14921 (atomic_compare_and_swap<mode>): Likewise.
14922 (atomic_exchange<mode>): Likewise.
14923 (atomic_nand<mode>): Likewise.
14924 (atomic_fetch_<fetchop_name><mode>): Likewise.
14925 (atomic_nand_fetch<mode>): Likewise.
14926 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14928 (ATOMIC): On power8, add QImode, HImode modes.
14929 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14930 modes that promote to SImode.
14931 (load_lockedti): Convert TImode arguments to PTImode, so that we
14932 get a guaranteed even/odd register pair.
14933 (load_lockedpti): Likewise.
14934 (store_conditionalti): Likewise.
14935 (store_conditionalpti): Likewise.
14937 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14938 atomic load/store instructions.
14941 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14943 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14945 (insn_count): New attribute, with most cases extracted from...
14946 (length): ...here. Redefine most cases in terms of insn_count.
14947 (single_insn): Delete.
14948 (can_delay): Use insn_count to check for single instructions.
14949 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14950 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14951 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14952 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14953 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14954 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14955 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14956 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14957 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14958 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14959 rather than "length".
14960 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14961 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14962 Use "insn_count" rather than "length".
14963 * config/mips/mips-dsp.md
14964 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14965 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14968 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14970 PR tree-optimization/57361
14971 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14973 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
14975 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
14977 (aarch64_simd_combine<mode>): New instruction expansion.
14978 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
14979 function prototype.
14980 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
14981 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
14983 2013-06-12 Jan Hubicka <jh@suse.cz>
14985 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
14986 decl has when in streaming stage.
14987 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
14988 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
14990 2013-06-12 Roland Stigge <stigge@antcom.de>
14993 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
14995 2013-06-12 Jakub Jelinek <jakub@redhat.com>
14997 PR tree-optimization/57537
14998 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
14999 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
15001 2013-06-12 Richard Biener <rguenther@suse.de>
15003 * data-streamer.h (streamer_write_char_stream): CSE
15004 obs->current_pointer.
15005 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
15006 streamer_write_char_stream manually and optimize the resulting loop.
15007 (streamer_write_hwi_stream): Likewise.
15009 2013-06-12 Jan Hubicka <jh@suse.cz>
15011 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
15012 * cgraph.h (varpool_create_empty_node): Declare.
15013 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
15015 * symtab.c (symtab_unregister_node): Be lax about missin entries
15017 (symtab_get_node): Update comment.
15018 * varpool.c (varpool_create_empty_node): Break out from ...
15019 (varpool_node_for_decl): ... here.
15020 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
15022 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
15024 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
15025 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
15026 part. Use straight-line flow at the end.
15027 <COMPONENT_REF>: Remove superfluous else.
15028 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
15030 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15033 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
15034 target hook even for !TREE_PUBLIC decls. If no resolution info
15035 is available, return false for common and external decls.
15037 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15039 * config/rl78/constraints.md (U): New constraint.
15040 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
15043 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15046 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
15047 to allow returning address to AT_PLATFORM name.
15049 2013-06-11 Jan Hubicka <jh@suse.cz>
15051 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
15052 * cgraph.h (symtab_node_base): Add weakref flag.
15053 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
15054 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
15055 (output_weakrefs): Use weakref flag.
15056 * fold-const.c (simple_operand_p): Handle WEAK.
15057 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
15058 * ipa.c (varpool_externally_visible_p): Drop weakref.
15059 (function_and_variable_visibility): Update comment; fix weakref
15060 sanity checks; do not clear DECL_WEAK on them.
15061 * lto-cgraph.c (lto_output_node): update.
15062 (lto_output_varpool_node): Update.
15063 (input_overwrite_node): Update.
15064 (input_node): Update.
15065 (input_varpool_node): Update.
15066 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
15067 (lto_symtab_merge_symbols): Add sanity check.
15068 (lto_symtab_prevailing_decl): Do not special case weakrefs.
15069 * passes.c (rest_of_decl_compilation): Set static flag, too.
15070 * symtab.c (dump_symtab_base): Dump weakref.
15071 (verify_symtab_base): Sanity check weakrefs.
15072 (symtab_make_decl_local): Remove duplicated code.
15073 (symtab_alias_ultimate_target): Simplify.
15074 * varpool.c (varpool_create_variable_alias): Set weakref flag.
15076 2013-06-11 Tom de Vries <tom@codesourcery.com>
15078 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
15079 sequence_vect == NULL.
15081 2013-06-11 DJ Delorie <dj@redhat.com>
15083 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
15084 (rl78_unwind_word_mode): New.
15086 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15088 * final.c (debug_prefix_maps): Make static.
15090 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15092 * function.c (initial_trampoline): Remove stray copy.
15094 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
15096 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
15098 2013-06-11 Martin Jambor <mjambor@suse.cz>
15100 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
15101 within bounds at the beginning of the function.
15103 2013-06-11 Alan Modra <amodra@gmail.com>
15105 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15107 (get_named_section): Don't NULL !DECL_P decl.
15109 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15111 * doc/invoke.texi (core-avx2): Document.
15113 (atom): Updated with MOVBE.
15115 2013-06-11 Richard Biener <rguenther@suse.de>
15117 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15119 2013-06-11 Anton Blanchard <anton@samba.org>
15121 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15122 correct shift value in little-endian mode.
15124 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15127 * varasm.c (get_variable_align): Move #endif to the right place.
15129 2013-06-10 Cary Coutant <ccoutant@google.com>
15131 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15132 for hash so that hash table traversal order is deterministic.
15134 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15135 Pat Haugen <pthaugen@us.ibm.com>
15136 Peter Bergner <bergner@vnet.ibm.com>
15138 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15139 of vectors in GPRS if they are direct moves or quad word load or
15142 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15144 (direct_move_p): Likewise.
15145 (quad_load_store_p): Likewise.
15147 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15148 classes into bins based on the physical register type.
15149 (reg_class_to_reg_type): Likewise.
15150 (IS_STD_REG_TYPE): Likewise.
15151 (IS_FP_VECT_REG_TYPE): Likewise.
15152 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15153 use direct move instructions.
15154 (reload_gpr_vsx): Likewise.
15155 (reload_vsx_gpr): Likewise.
15156 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15157 information that is a simplification of register classes. Also
15158 precalculate direct move reload helpers.
15159 (direct_move_p): New function to return true if the operation can
15160 be done as a direct move instruciton.
15161 (quad_load_store_p): New function to return true if the operation
15162 is a quad memory operation.
15163 (rs6000_legitimize_address): If quad memory, only allow register
15164 indirect for TImode addresses.
15165 (rs6000_legitimate_address_p): Likewise.
15166 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15167 (rs6000_reload_register_type): Likewise.
15168 (register_to_reg_type): Return register type.
15169 (rs6000_secondary_reload_simple_move): New helper function for
15170 secondary reload and secondary memory needed to identify anything
15171 that is a simple move, and does not need reloading.
15172 (rs6000_secondary_reload_direct_move): New helper function for
15173 secondary reload to identify cases that can be done with several
15174 instructions via the direct move instructions.
15175 (rs6000_secondary_reload_move): New helper function for secondary
15176 reload to identify moves between register types that can be done.
15177 (rs6000_secondary_reload): Add support for quad memory operations
15178 and for direct move.
15179 (rs6000_secondary_memory_needed): Likewise.
15180 (rs6000_debug_secondary_memory_needed): Change argument names.
15181 (rs6000_output_move_128bit): New function to return the move to
15182 use for 128-bit moves, including knowing about the various
15183 limitations of quad memory operations.
15185 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15186 memory operations. call rs6000_output_move_128bit for the actual
15187 instruciton(s) to generate.
15188 (vsx_movti_64bit): Likewise.
15190 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15191 (UNSPEC_P8V_MTVSRWZ): Likewise.
15192 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15193 (UNSPEC_P8V_MTVSRD): Likewise.
15194 (UNSPEC_P8V_XXPERMDI): Likewise.
15195 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15196 (UNSPEC_FUSION_GPR): Likewise.
15197 (FMOVE128_GPR): New iterator for direct move.
15198 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15199 (f32_sv): Likewise.
15200 (f32_dm): Likewise.
15201 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15202 loads and direct move instructions.
15203 (zero_extendsidi2_lfiwzx): Likewise.
15204 (extendsidi2_lfiwax): Likewise.
15205 (extendsidi2_nocell): Likewise.
15206 (floatsi<mode>2_lfiwax): Likewise.
15207 (lfiwax): Likewise.
15208 (floatunssi<mode>2_lfiwzx): Likewise.
15209 (lfiwzx): Likewise.
15210 (fix_trunc<mode>_stfiwx): Likewise.
15211 (fixuns_trunc<mode>_stfiwx): Likewise.
15212 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15213 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15214 (parity<mode>2_cmpb): Set length/type attr.
15215 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15216 for 'mr.' to fast_compare.
15217 (bpermd_<mode>): Change type attr to popcnt.
15218 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15219 (p8_mtvsrwz_1): Likewise.
15220 (p8_mtvsrwz_2): Likewise.
15221 (reload_fpr_from_gpr<mode>): Likewise.
15222 (p8_mtvsrd_1): Likewise.
15223 (p8_mtvsrd_2): Likewise.
15224 (p8_xxpermdi_<mode>): Likewise.
15225 (reload_vsx_from_gpr<mode>): Likewise.
15226 (reload_vsx_from_gprsf): Likewise.
15227 (p8_mfvsrd_3_<mode>): LIkewise.
15228 (reload_gpr_from_vsx<mode>): Likewise.
15229 (reload_gpr_from_vsxsf): Likewise.
15230 (p8_mfvsrd_4_disf): Likewise.
15231 (multi-word GPR splits): Do not split direct moves or quad memory
15234 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15236 * tree-into-ssa.c (interesting_blocks): Make static.
15238 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15241 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15242 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15243 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15244 (get_variable_align): New function.
15245 (get_variable_section, emit_bss, emit_common,
15246 assemble_variable_contents, place_block_symbol): Use
15247 get_variable_align instead of DECL_ALIGN.
15248 (assemble_noswitch_variable): Add align argument, use it
15249 instead of DECL_ALIGN.
15250 (assemble_variable): Adjust caller. Use get_variable_align
15251 instead of DECL_ALIGN.
15252 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15254 (DATA_ABI_ALIGNMENT): Define.
15255 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15256 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15257 opt is false, only return the psABI mandated alignment increase.
15258 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15259 (DATA_ABI_ALIGNMENT): ... this.
15260 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15261 (DATA_ABI_ALIGNMENT): ... this.
15262 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15263 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15264 (DATA_ABI_ALIGNMENT): ... this.
15265 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15266 * doc/tm.texi: Regenerated.
15268 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15270 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15271 cmp_code to construct REG_EQUAL note.
15273 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15276 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15277 that operands[2] doesn't overlap with operands[0].
15279 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15280 Jan Hubicka <jh@suse.cz>
15282 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15283 hack to mark symbols as used.
15285 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15287 PR rtl-optimization/57559
15288 * lra-constraints.c (process_alt_operands): Don't discourage
15289 memory with known offset for offsetable memory constraint.
15290 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15292 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15294 * varasm.c (struct oc_local_state): Reorder fields.
15295 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15296 and adjust accordingly.
15297 (output_constructor): Reorder initialization code and adjust call to
15298 output_constructor_bitfield.
15300 2013-06-07 Jan Hubicka <jh@suse.cz>
15302 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15304 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15306 * tree-object-size.c (unknown): Make const.
15308 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15310 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15311 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15312 for last alternative in the cpu_facility attribute.
15314 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15317 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15318 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15319 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15320 * config/arm/constraints.md (Dg): New constraint.
15321 * config/arm/neon.md (xordi3_neon): Remove.
15322 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15323 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15325 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15327 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15328 Clean up alternatives.
15330 2013-06-07 Alan Modra <amodra@gmail.com>
15332 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15335 2013-06-07 Alan Modra <amodra@gmail.com>
15337 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15339 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15341 * config/arm/constraints.md (Df): New constraint.
15342 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15343 Correct length attribute for last two alternatives.
15345 2013-06-07 Alan Modra <amodra@gmail.com>
15347 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15348 override user -mfp-in-toc.
15349 (offsettable_ok_by_alignment): Consider just the current access
15350 rather than the whole object, unless BLKmode. Handle
15351 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15352 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15353 for -mcmodel=medium.
15354 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15355 override user -mfp-in-toc or -msum-in-toc. Default to
15356 -mno-fp-in-toc for -mcmodel=medium.
15358 2013-06-06 DJ Delorie <dj@redhat.com>
15360 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15361 TARGET_VALID_POINTER_MODE.
15363 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15364 Pat Haugen <pthaugen@us.ibm.com>
15365 Peter Bergner <bergner@vnet.ibm.com>
15367 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15368 Document new power8 builtins.
15370 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15371 condition code register, to allow 128-bit logical operations to be
15372 done in the VSX or GPR registers.
15373 (nor<mode>3): Use the canonical form for nor.
15374 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15375 vclz*, and vpopcnt* vector instructions.
15376 (nand<mode>3): Likewise.
15377 (orc<mode>3): Likewise.
15378 (clz<mode>2): LIkewise.
15379 (popcount<mode>2): Likewise.
15381 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15382 that only the GPRs are recognized.
15384 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15385 support for new power8 builtins.
15387 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15389 (xscvdpspn): Likewise.
15395 (vpopcnt): Likewise.
15396 (vpopcntb): Likewise.
15397 (vpopcnth): Likewise.
15398 (vpopcntw): Likewise.
15399 (vpopcntd): Likewise.
15401 (vmrgew): Likewise.
15402 (vmrgow): Likewise.
15404 (eqv_v16qi3): Likewise.
15405 (eqv_v8hi3): Likewise.
15406 (eqv_v4si3): Likewise.
15407 (eqv_v2di3): Likewise.
15408 (eqv_v4sf3): Likewise.
15409 (eqv_v2df3): Likewise.
15411 (nand_v16qi3): Likewise.
15412 (nand_v8hi3): Likewise.
15413 (nand_v4si3): Likewise.
15414 (nand_v2di3): Likewise.
15415 (nand_v4sf3): Likewise.
15416 (nand_v2df3): Likewise.
15418 (orc_v16qi3): Likewise.
15419 (orc_v8hi3): Likewise.
15420 (orc_v4si3): Likewise.
15421 (orc_v2di3): Likewise.
15422 (orc_v4sf3): Likewise.
15423 (orc_v2df3): Likewise.
15425 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15426 allow power8 quad mode in 64-bit.
15427 (rs6000_builtin_vectorized_function): Add support to vectorize
15428 ISA 2.07 count leading zeros, population count builtins.
15429 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15430 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15431 (builtin_function_type): Add vgbbd builtin function which takes an
15433 (altivec_expand_vec_perm_const): Add support for new power8 merge
15436 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15437 that does not include TImdoe for use with 32-bit.
15438 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15440 (UNSPEC_VSX_CVDPSPN): Likewise.
15441 (vsx_xscvdpspn): Likewise.
15442 (vsx_xscvspdpn): Likewise.
15443 (vsx_xscvdpspn_scalar): Likewise.
15444 (vsx_xscvspdpn_directmove): Likewise.
15445 (vsx_and<mode>3): Split logical operations into 32-bit and
15446 64-bit. Add support to do logical operations on TImode as well as
15447 VSX vector types. Allow logical operations to be done in either
15448 VSX registers or in general purpose registers in 64-bit mode. Add
15449 splitters if GPRs were used. For AND, add clobber of CCmode to
15450 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15452 (vsx_and<mode>3_32bit): Likewise.
15453 (vsx_and<mode>3_64bit): Likewise.
15454 (vsx_ior<mode>3): Likewise.
15455 (vsx_ior<mode>3_32bit): Likewise.
15456 (vsx_ior<mode>3_64bit): Likewise.
15457 (vsx_xor<mode>3): Likewise.
15458 (vsx_xor<mode>3_32bit): Likewise.
15459 (vsx_xor<mode>3_64bit): Likewise.
15460 (vsx_one_cmpl<mode>2): Likewise.
15461 (vsx_one_cmpl<mode>2_32bit): Likewise.
15462 (vsx_one_cmpl<mode>2_64bit): Likewise.
15463 (vsx_nor<mode>3): Likewise.
15464 (vsx_nor<mode>3_32bit): Likewise.
15465 (vsx_nor<mode>3_64bit): Likewise.
15466 (vsx_andc<mode>3): Likewise.
15467 (vsx_andc<mode>3_32bit): Likewise.
15468 (vsx_andc<mode>3_64bit): Likewise.
15469 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15470 and xxlorc instructions.
15471 (vsx_eqv<mode>3_64bit): Likewise.
15472 (vsx_nand<mode>3_32bit): Likewise.
15473 (vsx_nand<mode>3_64bit): Likewise.
15474 (vsx_orc<mode>3_32bit): Likewise.
15475 (vsx_orc<mode>3_64bit): Likewise.
15477 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15479 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15481 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15482 (p8_vmrgow): Likewise.
15483 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15484 GPRs to be split under VSX.
15485 (p8v_clz<mode>2): Add power8 count leading zero support.
15486 (p8v_popcount<mode>2): Add power8 population count support.
15487 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15490 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15493 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15495 (vec_nand): Likewise.
15496 (vec_vclz): Likewise.
15497 (vec_vclzb): Likewise.
15498 (vec_vclzd): Likewise.
15499 (vec_vclzh): Likewise.
15500 (vec_vclzw): Likewise.
15501 (vec_vgbbd): Likewise.
15502 (vec_vmrgew): Likewise.
15503 (vec_vmrgow): Likewise.
15504 (vec_vpopcnt): Likewise.
15505 (vec_vpopcntb): Likewise.
15506 (vec_vpopcntd): Likewise.
15507 (vec_vpopcnth): Likewise.
15508 (vec_vpopcntw): Likewise.
15510 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15512 PR rtl-optimization/57468
15513 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15516 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15518 PR rtl-optimization/57459
15519 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15520 type when setting live regs.
15522 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15524 * config/s390/s390.opt (mlra): New option.
15525 * config/s390/s390.c (s390_decompose_address): Check displacement
15526 for all registers for LRA.
15527 (s390_secondary_reload): Don't used secondary reloads for LRA.
15528 (s390_lra_p): New function.
15529 (TARGET_LRA_P): Define.
15530 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15531 of attribute cpu_facility to zarch for the last alternative.
15532 (*cmpmem_short): Ditto.
15534 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15536 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15537 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15538 (arm_expand_prologue): Likewise.
15540 2013-06-06 Teresa Johnson <tejohnson@google.com>
15543 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15544 as this is now done by redirect_edge_and_branch_force.
15545 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15546 barriers, and fix interaction with splitting.
15547 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15548 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15549 reflect changes made in the routine.
15550 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15551 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15552 since this is called in cfglayout mode, and replace partition fixup
15553 with assert as that is now done by force_nonfallthru_and_redirect.
15554 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15555 already be marked with region crossing note.
15556 (insert_section_boundary_note): Make non-static, gate on flag
15557 has_bb_partition, rewrite to also check for multiple partitions.
15558 (rest_of_handle_reorder_blocks): Remove call to
15559 insert_section_boundary_note, now done later during free_cfg.
15560 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15561 * bb-reorder.h (insert_section_boundary_note): Declare.
15562 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15563 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15564 invoke insert_section_boundary_note.
15565 (try_redirect_by_replacing_jump): Remove unnecessary
15566 check for region crossing note.
15567 (fixup_partition_crossing): New function.
15568 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15569 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15570 in non-cfglayout mode.
15571 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15572 remove old code that tried to do this. Emit barrier correctly
15573 when we are in cfglayout mode.
15574 (last_bb_in_partition): New function.
15575 (rtl_split_edge): Correctly fixup partition boundaries.
15576 (commit_one_edge_insertion): Remove old code that tried to
15577 fixup region crossing edge since this is now handled in
15578 split_block, and set up insertion point correctly since
15579 block may now end in a jump.
15580 (verify_hot_cold_block_grouping): Guard against checking when not in
15581 linearized RTL mode.
15582 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15584 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15585 rtl_verify_flow_info, so not called in cfglayout mode.
15586 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15587 (fixup_reorder_chain): Remove old code that attempted to fixup region
15588 crossing note as this is now handled in force_nonfallthru_and_redirect.
15589 (duplicate_insn_chain): Don't duplicate switch section notes.
15590 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15592 * basic-block.h (emit_barrier_after_bb): Declare.
15594 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15596 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15597 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15598 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15600 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15602 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15604 * config/arm/ldmstm.md: Regenerate.
15606 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15608 * config/arm/sync.md (atomic_loaddi_1):
15609 Disable predication for arm_restrict_it.
15610 (arm_load_exclusive<mode>): Likewise.
15611 (arm_load_exclusivesi): Likewise.
15612 (arm_load_exclusivedi): Likewise.
15613 (arm_load_acquire_exclusive<mode>): Likewise.
15614 (arm_load_acquire_exclusivesi): Likewise.
15615 (arm_load_acquire_exclusivedi): Likewise.
15616 (arm_store_exclusive<mode>): Likewise.
15617 (arm_store_exclusive<mode>): Likewise.
15618 (arm_store_release_exclusivedi): Likewise.
15619 (arm_store_release_exclusive<mode>): Likewise.
15621 2013-06-06 Richard Biener <rguenther@suse.de>
15623 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15625 (lto_tag_is_tree_code_p): Adjust.
15626 (lto_tag_is_gimple_code_p): Likewise.
15627 (lto_gimple_code_to_tag): Likewise.
15628 (lto_tag_to_gimple_code): Likewise.
15629 (lto_tree_code_to_tag): Likewise.
15630 (lto_tag_to_tree_code): Likewise.
15631 * data-streamer.h (streamer_write_hwi_in_range): Use
15632 uhwi streaming to stream the normalized range.
15633 (streamer_read_hwi_in_range): Likewise.
15635 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15637 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15638 (predicable_short_it): Likewise.
15639 (predicated): Likewise.
15640 (enabled): Handle above.
15641 (define_cond_exec): Set predicated attribute to yes.
15643 2013-06-05 Mike Stump <mikestump@comcast.net>
15645 * gdbinit.in (__FUNCTION__): Add.
15647 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15649 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15650 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15652 2013-06-05 Jan Hubicka <jh@suse.cz>
15654 * varasm.c (mark_decl_referenced): Revert the removal until targets
15657 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15659 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15660 instead of mark_decl_referenced.
15662 2013-06-05 Jan Hubicka <jh@suse.cz>
15664 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15665 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15666 and symtab_used_from_object_file_p.
15667 (cgraph_make_node_local_1): Clear forced_by_abi.
15668 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15669 * cgraph.h (symtab_node_base): Add forced_by_abi.
15670 (decide_is_variable_needed): Remove.
15671 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15672 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15673 (decide_is_symbol_needed): ... this one; handle symbols in general;
15674 always analyze virtuals; honnor forced_by_abi.
15675 (cgraph_finalize_function): Update.
15676 (varpool_finalize_decl): Update.
15677 (symbol_defined_and_needed): Remove.
15678 (analyze_functions): Update.
15679 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15680 output_refs, input_overwrite_node): Handle forced_by_abi.
15681 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15682 (address_taken_from_non_vtable_p): ... this one.
15683 (comdat_can_be_unshared_p_1): New function.
15684 (cgraph_comdat_can_be_unshared_p): Rename to ...
15685 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15686 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15687 (function_and_variable_visibility): Clear forced_by_abi as needed.
15688 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15689 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15690 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15691 * varpool.c (decide_is_variable_needed): Remove.
15693 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15695 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15696 (arm_option_override): Override arm_restrict_it where appropriate.
15697 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15698 * config/arm/arm.opt (mrestrict-it): New command-line option.
15699 * doc/invoke.texi: Document -mrestrict-it.
15701 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15703 * tsan.c (tsan_atomic_table): Make const.
15705 2013-06-05 Richard Biener <rguenther@suse.de>
15707 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15708 index associated with the tree we are supposed to replace.
15709 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15710 TREE_ASM_WRITTEN as zero for everything but SSA names.
15712 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15714 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15716 2013-06-04 Jan Hubicka <jh@suse.cz>
15718 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15719 (input_node, input_varpool_node): Handle correctly external same
15721 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15722 nodes at ltrans stage.
15724 2013-06-04 Jan Hubicka <jh@suse.cz>
15726 * ipa-inline.c (update_caller_keys): Fix availability test.
15727 (update_callee_keys): Likewise.
15728 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15729 to follow ELF standard.
15731 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15733 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15734 (mips64r5900el-*-elf*): New configurations.
15735 * config/mips/mips-cpus.def (r5900): New processor.
15736 * config/mips/mips-tables.opt: Regenerate.
15737 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15738 (mips_issue_rate): Handle PROCESSOR_R5900.
15739 (mips_reorg_process_insns): Force reorder mode for the R5900.
15740 * config/mips/mips.h (TARGET_MIPS5900): Define.
15741 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15743 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15745 * config/mips/mips.md (processor): Add r5900.
15746 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15748 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15750 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15751 into function to generate MOVI instruction.
15752 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15753 (aarch64_preferred_simd_mode): Turn into wrapper.
15754 (aarch64_output_scalar_simd_mov_immediate): New function.
15755 * config/aarch64/aarch64-protos.h: Add prototype for above.
15757 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15759 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15760 element_char member.
15761 (sizetochar): Return signed char.
15762 (aarch64_simd_valid_immediate): Remove elchar and other
15763 unnecessary variables.
15764 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15765 Calculate element_char as required.
15766 * config/aarch64/aarch64-protos.h: Update and move prototype
15767 for aarch64_output_simd_mov_immediate.
15768 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15771 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15773 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15774 information completed by aarch64_simd_valid_immediate.
15775 (aarch64_legitimate_constant_p): Update arguments.
15776 (aarch64_simd_valid_immediate): Work with struct rather than many
15778 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15779 (aarch64_simd_make_constant): Update arguments.
15780 (aarch64_output_simd_mov_immediate): Work with struct rather than
15781 many pointers. Output immediate directly rather than as operand.
15782 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15784 * config/aarch64/constraints.md (Dn): Update arguments.
15786 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15788 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15790 (aarch64_simd_immediate_valid_for_move): Remove.
15791 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15792 (aarch64_simd_make_constant): Update call.
15793 (aarch64_output_simd_mov_immediate): Update call.
15794 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15796 * config/aarch64/constraints.md (Dn): Update call.
15798 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15800 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15801 return type to bool for prototype.
15802 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15803 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15804 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15806 2013-06-04 Catherine Moore <clm@codesourcery.com>
15808 * config/mips/mips.opt (meva): New.
15809 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15810 (ASM_SPEC): Handle -meva.
15811 * doc/invoke.texi (meva): Document.
15813 2013-06-04 Alan Modra <amodra@gmail.com>
15815 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15818 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15820 * rtl.def: Add extra fourth optional field to define_cond_exec.
15821 * gensupport.c (process_one_cond_exec): Process attributes from
15823 * doc/md.texi: Document fourth field in define_cond_exec.
15825 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15827 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15828 out the processing order as in store_bit_field_1.
15830 2013-06-04 Jan Hubicka <jh@suse.cz>
15832 PR middle-end/57500
15833 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15834 non-VAR_DECL node if it does not exist yet.
15836 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15838 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15839 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15840 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15841 target_cpu_default setting.
15843 2013-06-03 Teresa Johnson <tejohnson@google.com>
15845 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15846 default to -fopt-info=optimized instead of all.
15847 * doc/invoke.texi: Ditto.
15848 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15849 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15850 (execute_vect_slp): Emit BB vectorization success under
15851 MSG_OPTIMIZED_LOCATIONS.
15852 * tree-vect-slp.c (vect_slp_transform_bb): Change
15853 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15854 * tree-vect-loop.c (vect_transform_loop): Ditto.
15856 2013-06-03 Jason Merrill <jason@redhat.com>
15859 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15860 Use TARGET_EXPR for C++.
15862 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15864 PR rtl-optimization/57268
15865 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15866 if DEBUG_INSN_P (insn).
15869 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15871 PR rtl-optimization/57268
15872 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15873 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15875 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15877 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15878 (ix86_avoid_lea_for_addr): Likewise.
15879 (exact_dependency_1): Likewise.
15880 (ix86_adjust_cost): Likewise.
15881 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15883 (do_reorder_for_imul): Fix typo, formatting and
15884 !reload_completed check removed.
15885 (ix86_sched_reorder): Fix typo and formatting.
15886 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15889 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15891 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15893 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15895 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15896 <STRING_CST>: Likewise.
15897 <VECTOR_CST>: Likewise.
15899 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15900 Mikael Morin <mikael@gcc.gnu.org>
15902 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15903 * config.in: Regenerated.
15904 * configure: Regenerated.
15906 2013-06-01 Jan Hubicka <jh@suse.cz>
15908 PR middle-end/57366
15909 * cgraphunit.c (compile): When weakref is not supported,
15910 set up transparent aliases before final output pass.
15911 * varasm.c (assemble_alias): Do not try to do it here.
15913 2013-06-01 Jan Hubicka <jh@suse.cz>
15915 PR middle-end/57467
15916 * passes.c (for_per_function): Skip unanalyzed functions.
15918 2013-06-01 Jan Hubicka <jh@suse.cz>
15920 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15921 (lto_symtab_merge_symbols_1): ... this one.
15922 (lto_symtab_merge_cgraph_nodes): Rename to ...
15923 (lto_symtab_merge_symbols): ... this one; simplify.
15924 * cgraph.c (same_body_aliases_done): Rename to ...
15925 (cpp_implicit_aliases_done): ... this one.
15926 (cgraph_create_function_alias): Update.
15927 (cgraph_same_body_alias): Update.
15928 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15929 (verify_edge_corresponds_to_fndecl): Simplify.
15930 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15931 (cgraph_node): Remove same_body_alias.
15932 (varpool_node): Remove alias_of and extra_name_alias.
15933 (same_body_aliases_done): Rename to ..
15934 (cpp_implicit_aliases_done): ... this one.
15935 (symtab_alias_ultimate_target): Add default parameter.
15936 (symtab_resolve_alias): New function.
15937 (fixup_same_cpp_alias_visibility): Declare.
15938 (cgraph_function_node): Add default parameter.
15939 (cgraph_node_asm_name): Likewise.
15940 (cgraph_function_or_thunk_node): Add default parameter; do
15941 not ICE when it is NULL.
15942 (varpool_variable_node): Likewise.
15943 * tree-emutls.c (create_emultls_var): Update.
15944 (ipa_lower_emutls): Update.
15945 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15946 (cgraph_reset_node): Reset alias info.
15947 (cgraph_finalize_function): Update.
15948 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15949 (analyze_function): Simplify.
15950 (cgraph_process_same_body_aliases): Simplify.
15951 (analyze_functions): Fixup same body aliases.
15952 (handle_alias_pairs): Simplify.
15953 (assemble_thunk): Update.
15954 (assemble_thunks_and_aliases): Update.
15955 (output_weakrefs): Rewrite.
15956 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15957 (lto_output_varpool_node): Likewise.
15958 (compute_ltrans_boundary): Remve assert.
15959 (get_alias_symbol): New functoin.
15960 (input_node): Rewrite alias handling.
15961 (input_varpool_node): Likewise.
15962 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15963 * ipa.c (process_references): Handle weakrefs correctly.
15964 (symtab_remove_unreachable_nodes): Likewise.
15965 * trans-mem.c (get_cg_data): Update.
15966 (ipa_tm_create_version_alias): Update.
15967 (ipa_tm_execute): Update.
15968 * symtab.c (dump_symtab_base): Dump aliases.
15969 (verify_symtab_base): Verify aliases.
15970 (symtab_node_availability): New function.
15971 (symtab_alias_ultimate_target): Simplify.
15972 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15973 handle all the fixup cases.
15974 (symtab_resolve_alias): New function.
15975 * passes.c (ipa_write_summaries): Handle weakrefs.
15976 * varpool.c (varpool_analyze_node): Simplify.
15977 (assemble_aliases): Update.
15978 (varpool_create_variable_alias): Simplify.
15979 (varpool_extra_name_alias): Simplify.
15980 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
15981 (lto_symtab_merge_symbols): ... this one.
15983 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
15986 PR rtl-optimization/57268
15987 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15988 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15990 2013-06-01 Tobias Burnus <burnus@net-b.de>
15992 Partially reverted:
15993 2013-05-31 Tobias Burnus <burnus@net-b.de>
15995 PR middle-end/57073
15996 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
15999 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16001 PR rtl-optimization/57268
16002 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16003 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16005 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16007 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
16008 unordered comparison operators when -fno-trapping-math is in effect
16010 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
16011 and implement unordered comparison operators properly on the e500.
16013 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16015 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
16016 for constant scalar integers.
16017 (simplify_relational_operation_1): Likewise.
16019 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
16021 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
16022 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
16025 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16026 Igor Zamyatin <igor.zamyatin@intel.com>
16028 Silvermont (SLM) architecture performance tuning.
16029 * config/i386/i386.h (enum ix86_tune_indices): Add
16030 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
16031 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
16033 * config/i386/i386.c (initial_ix86_tune_features)
16034 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
16035 (ix86_lea_outperforms): Handle Silvermont tuning.
16036 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
16038 (ix86_use_lea_for_mov): Likewise.
16039 (ix86_avoid_lea_for_addr): Likewise.
16040 (ix86_lea_for_add_ok): Likewise.
16041 (exact_dependency_1): New function.
16042 (exact_store_load_dependency): Likewise.
16043 (ix86_adjust_cost): Handle Silvermont tuning.
16044 (do_reoder_for_imul): Likewise.
16045 (swap_top_of_ready_list): New function.
16046 (ix86_sched_reorder): Changed to handle Silvermont tuning.
16048 * config/i386/i386.md (peepholes that split memory operand in fp
16051 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16053 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
16054 Remove un-necessary braces.
16056 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16058 * config/aarch64/aarch64.c (aarch64_classify_symbol):
16059 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
16061 2013-05-31 Tobias Burnus <burnus@net-b.de>
16063 PR middle-end/57073
16064 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
16066 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16069 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
16070 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
16071 * config/arm/neon.md (iordi3_neon): Remove.
16072 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
16073 * config/arm/predicates.md (imm_for_neon_logic_operand):
16074 Move to earlier in the file.
16075 (neon_logic_op2): Likewise.
16076 (arm_iordi_operand_neon): New predicate.
16078 2013-05-31 Richard Biener <rguenther@suse.de>
16080 PR tree-optimization/57478
16081 PR tree-optimization/57453
16082 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
16085 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16087 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
16088 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
16090 2013-05-30 Tobias Burnus <burnus@net-b.de>
16091 Thomas Koenig <tkoenig@gcc.gnu.org>
16093 PR middle-end/57073
16094 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
16095 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
16097 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
16099 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
16101 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
16103 * target.def (register_usage_leveling_p): New hook.
16104 * targhooks.c (default_register_usage_leveling_p): New.
16105 * targhooks.h (default_register_usage_leveling_p): New prototype.
16106 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16107 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16108 * doc/tm.texi: Update.
16109 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16111 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16113 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16114 (*insv_reg<mode>): New define_insn.
16116 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16118 PR rtl-optimization/57439
16119 * postreload.c (move2add_valid_value_p): Check that we have
16120 a zero subreg_regno_offset when accessing the register in
16121 the requested mode.
16123 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16124 Igor Zamyatin <igor.zamyatin@intel.com>
16126 Silvermont (SLM) architecture pipeline model, tuning and
16128 * config.gcc: Add slm config options and target.
16130 * config/i386/slm.md: New.
16132 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16134 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16136 (ix86_target_macros_internal): Likewise.
16138 * config/i386/i386.c (slm_cost): New cost.
16139 (m_SLM): New macro flag.
16140 (initial_ix86_tune_features): Set m_SLM.
16141 (x86_accumulate_outgoing_args): Likewise.
16142 (x86_arch_always_fancy_math_387): Likewise.
16143 (processor_target_table): Add slm cost.
16144 (cpu_names): Add slm cpu name.
16145 (x86_option_override_internal): Set SLM ISA.
16146 (ix86_issue_rate): New case PROCESSOR_SLM.
16147 (ia32_multipass_dfa_lookahead): Likewise.
16148 (fold_builtin_cpu): Add slm.
16150 * config/i386/i386.h (TARGET_SLM): New target macro.
16151 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16152 (processor_type): Add PROCESSOR_SLM.
16154 * config/i386/i386.md (cpu): Add new value "slm".
16155 (slm.md): Include slm.md.
16157 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16158 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16160 * config/arm/arm-protos.h: Add and update function protos.
16161 * config/arm/arm.c (use_simple_return_p): New added.
16162 (thumb2_expand_return): Check simple_return flag.
16163 * config/arm/arm.md: Add simple_return and conditional simple_return.
16164 * config/arm/iterators.md: Add iterator for return and simple_return.
16166 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16168 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16169 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16170 (arm_emit_vfp_multi_reg_pop): Likewise.
16171 (thumb2_emit_ldrd_pop): Likewise.
16172 (arm_expand_epilogue): Add misc REG_CFA notes.
16173 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16175 2013-05-29 Lawrence Crowl <crowl@google.com>
16177 * config/arm/t-arm: Update for below.
16179 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16180 Change type to hash_table. Update dependent calls and types.
16182 * config/i386/t-cygming: Update for below.
16184 * config/i386/t-interix: Update for below.
16186 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16187 Change type to hash_table. Update dependent calls and types.
16188 (i386_find_on_wrapper_list::wrappers): Likewise.
16190 * config/ia64/t-ia64: Update for below.
16192 * config/ia64/ia64.c (bundle_state_table):
16193 Change type to hash_table. Update dependent calls and types.
16195 * config/mips/mips.c (mips_reorg_process_insns::htab):
16196 Change type to hash_table. Update dependent calls and types.
16198 * config/sol2.c (solaris_comdat_htab):
16199 Change type to hash_table. Update dependent calls and types.
16201 * config/t-sol2: Update for above.
16203 2013-05-29 Teresa Johnson <tejohnson@google.com>
16205 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16206 functions are not yet marked as defined.
16208 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16209 Pat Haugen <pthaugen@us.ibm.com>
16210 Peter Bergner <bergner@vnet.ibm.com>
16212 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16216 (vrotl<mode>3): Likewise.
16217 (vashl<mode>3): Likewise.
16218 (vlshr<mode>3): Likewise.
16219 (vashr<mode>3): Likewise.
16221 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16222 support for power8 V2DI builtins.
16224 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16225 power8 V2DI builtins.
16226 (vupkhsw): Likewise.
16227 (vupklsw): Likewise.
16228 (vaddudm): Likewise.
16229 (vminsd): Likewise.
16230 (vmaxsd): Likewise.
16231 (vminud): Likewise.
16232 (vmaxud): Likewise.
16233 (vpkudum): Likewise.
16234 (vpksdss): Likewise.
16235 (vpkudus): Likewise.
16236 (vpksdus): Likewise.
16241 (vsubudm): Likewise.
16242 (vcmpequd): Likewise.
16243 (vcmpgtsd): Likewise.
16244 (vcmpgtud): Likewise.
16245 (vcmpequd_p): Likewise.
16246 (vcmpgtsd_p): Likewise.
16247 (vcmpgtud_p): Likewise.
16248 (vupkhsw): Likewise.
16249 (vupklsw): Likewise.
16250 (vaddudm): Likewise.
16251 (vmaxsd): Likewise.
16252 (vmaxud): Likewise.
16253 (vminsd): Likewise.
16254 (vminud): Likewise.
16255 (vpksdss): Likewise.
16256 (vpksdus): Likewise.
16257 (vpkudum): Likewise.
16258 (vpkudus): Likewise.
16263 (vsubudm): Likewise.
16265 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16266 support for power8 V2DI instructions.
16268 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16269 power8 V2DI instructions. Combine pack and unpack insns to use an
16270 iterator for each mode. Check whether a particular mode supports
16271 Altivec instructions instead of just checking TARGET_ALTIVEC.
16272 (UNSPEC_VPKUWUM): Likewise.
16273 (UNSPEC_VPKSHSS): Likewise.
16274 (UNSPEC_VPKSWSS): Likewise.
16275 (UNSPEC_VPKUHUS): Likewise.
16276 (UNSPEC_VPKSHUS): Likewise.
16277 (UNSPEC_VPKUWUS): Likewise.
16278 (UNSPEC_VPKSWUS): Likewise.
16279 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16280 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16281 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16282 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16283 (UNSPEC_VUPKHSB): Likewise.
16284 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16285 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16286 (UNSPEC_VUPKHSH): Likewise.
16287 (UNSPEC_VUPKLSB): Likewise.
16288 (UNSPEC_VUPKLSH): Likewise.
16290 (VI_char): Likewise.
16291 (VI_scalar): Likewise.
16292 (VI_unit): Likewise.
16294 (VP_small): Likewise.
16295 (VP_small_lc): Likewise.
16296 (VU_char): Likewise.
16297 (add<mode>3): Likewise.
16298 (altivec_vaddcuw): Likewise.
16299 (altivec_vaddu<VI_char>s): Likewise.
16300 (altivec_vadds<VI_char>s): Likewise.
16301 (sub<mode>3): Likewise.
16302 (altivec_vsubcuw): Likewise.
16303 (altivec_vsubu<VI_char>s): Likewise.
16304 (altivec_vsubs<VI_char>s): Likewise.
16305 (altivec_vavgs<VI_char>): Likewise.
16306 (altivec_vcmpbfp): Likewise.
16307 (altivec_eq<mode>): Likewise.
16308 (altivec_gt<mode>): Likewise.
16309 (altivec_gtu<mode>): Likewise.
16310 (umax<mode>3): Likewise.
16311 (smax<mode>3): Likewise.
16312 (umin<mode>3): Likewise.
16313 (smin<mode>3): Likewise.
16314 (altivec_vpkuhum): Likewise.
16315 (altivec_vpkuwum): Likewise.
16316 (altivec_vpkshss): Likewise.
16317 (altivec_vpkswss): Likewise.
16318 (altivec_vpkuhus): Likewise.
16319 (altivec_vpkshus): Likewise.
16320 (altivec_vpkuwus): Likewise.
16321 (altivec_vpkswus): Likewise.
16322 (altivec_vpks<VI_char>ss): Likewise.
16323 (altivec_vpks<VI_char>us): Likewise.
16324 (altivec_vpku<VI_char>us): Likewise.
16325 (altivec_vpku<VI_char>um): Likewise.
16326 (altivec_vrl<VI_char>): Likewise.
16327 (altivec_vsl<VI_char>): Likewise.
16328 (altivec_vsr<VI_char>): Likewise.
16329 (altivec_vsra<VI_char>): Likewise.
16330 (altivec_vsldoi_<mode>): Likewise.
16331 (altivec_vupkhsb): Likewise.
16332 (altivec_vupkhs<VU_char>): Likewise.
16333 (altivec_vupkls<VU_char>): Likewise.
16334 (altivec_vupkhsh): Likewise.
16335 (altivec_vupklsb): Likewise.
16336 (altivec_vupklsh): Likewise.
16337 (altivec_vcmpequ<VI_char>_p): Likewise.
16338 (altivec_vcmpgts<VI_char>_p): Likewise.
16339 (altivec_vcmpgtu<VI_char>_p): Likewise.
16340 (abs<mode>2): Likewise.
16341 (vec_unpacks_hi_v16qi): Likewise.
16342 (vec_unpacks_hi_v8hi): Likewise.
16343 (vec_unpacks_lo_v16qi): Likewise.
16344 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16345 (vec_unpacks_lo_v8hi): Likewise.
16346 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16347 (vec_pack_trunc_v8h): Likewise.
16348 (vec_pack_trunc_v4si): Likewise.
16349 (vec_pack_trunc_<mode>): Likewise.
16351 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16353 (vec_vmaxsd): Likewise.
16354 (vec_vmaxud): Likewise.
16355 (vec_vminsd): Likewise.
16356 (vec_vminud): Likewise.
16357 (vec_vpksdss): Likewise.
16358 (vec_vpksdus): Likewise.
16359 (vec_vpkudum): Likewise.
16360 (vec_vpkudus): Likewise.
16361 (vec_vrld): Likewise.
16362 (vec_vsld): Likewise.
16363 (vec_vsrad): Likewise.
16364 (vec_vsrd): Likewise.
16365 (vec_vsubudm): Likewise.
16366 (vec_vupkhsw): Likewise.
16367 (vec_vupklsw): Likewise.
16369 2013-05-29 Jan Hubicka <jh@suse.cz>
16371 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16372 flags; reorder rest of fields in more consistent way.
16373 (varpool_node): Remove analyzed, finalized and alias.
16374 (cgraph_ndoe): Likewise.
16375 (symtab_alias_ultimate_target): New function.
16376 (cgraph_function_node): Move offline.
16377 (cgraph_reset_node): Declare.
16378 (cgraph_comdat_can_be_unshared_p): Remove.
16379 (varpool_remove_initializer): Declare.
16380 (varpool_first_defined_variable, varpool_next_defined_variable
16381 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16382 (cgraph_function_with_gimple_body_p): Update.
16383 (varpool_all_refs_explicit_p): Update.
16384 (symtab_alias_target): New function.
16385 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16386 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16387 (cgraph_function_or_thunk_node): Simplify using
16388 symtab_alias_ultimate_target.
16389 (varpool_variable_node): Likewise.
16390 * cgraph.c (cgraph_create_function_alias): Update.
16391 (cgraph_add_thunk): Update.
16392 (cgraph_remove_node): Update.
16393 (dump_cgraph_node): Do not dump removed flags.
16394 (cgraph_function_body_availability): Update.
16395 (cgraph_propagate_frequency): Update.
16396 (verify_cgraph_node): Check sanity of local flag.
16397 (cgraph_function_node): Move here from cgraph.h; revamp for
16398 cgraph_function_or_thunk_node.
16399 * lto-symtab.c (lto_varpool_replace_node): Update.
16400 (lto_symtab_resolve_can_prevail_p): Update.
16401 (lto_symtab_merge_cgraph_nodes): Update.
16402 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16403 propagate_constants_accross_call, devirtualization_time_bonus,
16404 ipcp_propagate_stage): Update.
16405 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16406 * ipa-inline-transform.c (clone_inlined_nodes,
16407 preserve_function_body_p): Update.
16408 * ipa-reference.c (propagate): Update.
16409 (write_node_summary_p): Update.
16410 * toplev.c (wrapup_global_declaration_2): Update.
16411 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16412 (analyze_function) ... this one.
16413 (cgraph_process_new_functions): Update.
16414 (cgraph_reset_node): Export.
16415 (cgraph_finalize_function): Update.
16416 (cgraph_add_new_function): Update.
16417 (process_function_and_variable_attributes): Update.
16418 (varpool_finalize_decl): Update.
16419 (symbol_finalized): Remove.
16420 (symbol_finalized_and_needed): Rename to ...
16421 (symbol_defined_and_needed): ... update.
16422 (cgraph_analyze_functions): Update.
16423 (handle_alias_pairs): Update.
16424 (mark_functions_to_output): Update.
16425 (assemble_thunk): Update.
16426 (output_in_order): Update.
16427 (output_weakrefs): Update.
16428 (finalize_compilation_unit): Update.
16429 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16430 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16431 input_node, input_varpool_node): Update.
16432 * dbxout.c (dbxout_expand_expr): Update.
16433 * cgraphclones.c (cgraph_clone_node): Update.
16434 (cgraph_copy_node_for_versioning): Update.
16435 (cgraph_materialize_clone): Update.
16436 (cgraph_materialize_all_clones): Update.
16437 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16438 propagate_pure_const, propagate_nothrow): Update.
16439 * lto-streamer-out.c (lto_output, write_symbol): Update.
16440 * ipa-utils.c (ipa_reverse_postorder): Update.
16441 * ipa-inline.c (can_inline_edge_p): Update.
16442 (update_caller_keys, ipa_inline): Update.
16443 * dwarf2out.c (reference_to_unused,
16444 premark_types_used_by_global_vars_helper): Update.
16445 * tree-eh.c (tree_could_trap_p): Update.
16446 * ipa-split.c (consider_split, execute_split_functions): Update.
16447 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16448 has_addr_references_p): Update; move ahead in file for better
16450 (process_references): Simplify.
16451 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16452 bodies are removed.
16453 (cgraph_comdat_can_be_unshared_p): Make static.
16454 (cgraph_externally_visible_p): Update.
16455 (varpool_externally_visible_p): Update.
16456 (function_and_variable_visibility): Update.
16457 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16458 ipa_tm_mark_force_output_node): Update.
16459 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16460 estimate_edge_devirt_benefit, inline_generate_summary,
16461 inline_write_summary): Update.
16462 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16463 * ipa-prop.c (ipa_compute_jump_functions): Update.
16464 (ipa_print_node_params, ipa_prop_read_section,
16465 ipa_update_after_lto_read, read_replacements_section): Update.
16466 * varasm.c (mark_decl_referenced): Update.
16467 (assemble_alias, dump_tm_clone_pairs): Update.
16468 * tree-inline.c (copy_bb): Update.
16469 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16471 * symtab.c (dump_symtab_base): Print new flags.
16472 (verify_symtab_base): Verify new flags.
16473 (symtab_alias_ultimate_target): New function.
16474 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16475 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16477 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16479 * i386.c (ix86_get_function_versions_dispatcher,
16480 ix86_generate_version_dispatcher_body): Update.
16481 (fold_builtin_cpu): Use varpool_add_new_variable.
16482 * varpool.c (varpool_remove_initializer): Break out from ...
16483 (varpool_remove_node): ... this one.
16484 (dump_varpool_node, varpool_node_for_asm,
16485 cgraph_variable_initializer_availability, varpool_analyze_node,
16486 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16487 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16490 2013-05-29 Jan Hubicka <jh@suse.cz>
16492 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16494 2013-05-29 Easwaran Raman <eraman@google.com>
16496 PR tree-optimization/57442
16497 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16498 when control exits the main loop.
16500 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16502 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16504 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16505 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16506 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16509 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16511 PR tree-optimization/57441
16512 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16513 Don't limit size of incr_vec to number of candidates.
16515 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16517 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16518 and mips16 directories.
16519 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16520 (MULTILIB_DIRNAMES): Ditto.
16521 (MULTILIB_EXCEPTIONS): Add new exceptions.
16522 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16523 (MULTILIB_DIRNAMES): Ditto.
16524 (MULTILIB_EXCEPTIONS): Add new exceptions.
16526 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16527 Marcus Shawcroft <marcus.shawcroft@arm.com>
16529 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16530 SYMBOL_TINY_ABSOLUTE.
16531 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16532 SYMBOL_TINY_ABSOLUTE.
16533 (aarch64_expand_mov_immediate): Likewise.
16534 (aarch64_classify_symbol): Likewise.
16535 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16536 Permit SYMBOL_TINY_ABSOLUTE.
16537 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16539 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16540 Marcus Shawcroft <marcus.shawcroft@arm.com>
16542 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16543 Refactor if/switch. Replace gcc_assert with if.
16545 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16547 * config/i386/i386.c (initial_ix86_tune_features): Enable
16548 FP Reassociation for AMD bdver1 and bdver2.
16550 2013-05-29 Martin Jambor <mjambor@suse.cz>
16552 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16553 and IMAGPART_EXPR do not occur within other handled_components.
16555 2013-05-29 Richard Biener <rguenther@suse.de>
16557 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16558 access on whether the use is in the BB we currently try to
16560 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16561 vectorize to vect_bb_slp_scalar_cost.
16563 2013-05-29 Richard Biener <rguenther@suse.de>
16565 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16566 computing scalar cost offsetted by stmts that are kept live
16568 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16569 for computation of scalar cost.
16571 2013-05-28 Steve Ellcey <sellcey@mips.com>
16573 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16575 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16577 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16578 array notation built-in reduction functions.
16579 * doc/passes.texi (Passes): Added documentation about changes done
16581 * doc/invoke.texi (C Dialect Options): Added documentation about
16582 the -fcilkplus flag.
16583 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16584 (BUILTINS_DEF): Depend on cilkplus.def.
16585 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16586 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16587 * cilkplus.def: New file.
16589 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16591 PR rtl-optimization/57439
16592 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16594 2013-05-28 Easwaran Raman <eraman@google.com>
16596 PR tree-optimization/57337
16597 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16598 (find_insert_point): Correctly identify the insertion point
16599 when two statements with the same UID is compared.
16601 2013-05-28 Richard Biener <rguenther@suse.de>
16603 PR tree-optimization/56787
16604 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16605 from the list of data references.
16606 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16608 (vect_analyze_loop_operations): Likewise.
16609 (vect_transform_loop): Remove clobbers.
16611 2013-05-28 Martin Jambor <mjambor@suse.cz>
16613 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16614 and REALPART_EXPRs have scalar type.
16616 2013-05-28 Richard Biener <rguenther@suse.de>
16618 PR tree-optimization/57411
16619 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16621 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16622 virtual operand propagation.
16624 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16626 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16627 destination register for bmasksi_vis.
16628 (vector_init_bshuffle): Likewise.
16629 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16631 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16633 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16634 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16635 mode if the instruction isn't available in the original mode.
16636 * config/sparc/sparc.opt (mfix-ut699): New option.
16637 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16638 (divdf3): Turn into expander.
16639 (divdf3_nofix): New insn.
16640 (divdf3_fix): Likewise.
16641 (divsf3): Disable if -mfix-ut699.
16642 (sqrtdf2): Turn into expander.
16643 (sqrtdf2_nofix): New insn.
16644 (sqrtdf2_fix): Likewise.
16645 (sqrtsf2): Disable if -mfix-ut699.
16647 2013-05-27 Richard Biener <rguenther@suse.de>
16649 PR middle-end/57412
16650 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16651 block for the new loop.
16653 2013-05-27 Richard Biener <rguenther@suse.de>
16655 PR tree-optimization/57343
16656 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16657 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16658 (number_of_iterations_cond): Do not build the folded tree.
16660 2013-05-27 Richard Biener <rguenther@suse.de>
16663 PR middle-end/57381
16664 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16665 OEP_CONSTANT_ADDRESS_OF retained.
16667 PR tree-optimization/57417
16668 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16669 for unchanged base.
16670 (set_ssa_val_to): Compare addresses using
16671 get_addr_base_and_unit_offset.
16673 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16675 PR rtl-optimization/56833
16676 * postreload.c (move2add_record_mode): New function.
16677 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16678 (move2add_use_add2_insn): Use move2add_record_sym_value.
16679 (move2add_use_add3_insn): Likewise.
16680 (reload_cse_move2add): Use move2add_valid_value_p and
16681 move2add_record_mode. Invalidate call-clobbered and REG_INC
16682 affected regs by setting reg_mode to VOIDmode.
16683 (move2add_note_store): Don't pretend the inside of a SUBREG is
16684 the actual destination. Invalidate single/leading registers by
16685 setting reg_mode to VOIDmode.
16686 Use move2add_record_sym_value, move2add_valid_value_p and
16687 move2add_record_mode.
16689 2013-05-27 Richard Biener <rguenther@suse.de>
16691 PR tree-optimization/57396
16692 * tree-affine.c (double_int_constant_multiple_p): Properly
16693 return false for val == 0 and div != 0.
16695 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16697 * config/mips/mips.h: Use #elif in preprocessor conditions.
16699 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16702 * config/mips/constraints.md (kl): New constraint.
16703 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16704 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16705 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16706 from LO for MIPS16.
16707 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16709 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16712 * config/mips/mips.c (mips_can_inline_p): New function.
16713 (TARGET_CAN_INLINE_P): Define.
16715 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16717 * sched-int.h (ds_t, dw_t): Make unsigned int.
16718 Fix documentation that describes how all the ds_t bits are used.
16719 Reserve the last bit for delayed-branch scheduling.
16720 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16721 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16722 (gen_dep_weak_1): Remove prototype.
16723 * sched-deps.c (get_dep_weak_1): Make static.
16724 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16725 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16726 * doc/tm.texi: Regenerate.
16727 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16729 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16732 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16734 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16735 Sandra Loosemore <sandra@codesourcery.com>
16737 * config.gcc (powerpc-*): Allow native for with-cpu.
16739 2013-05-24 Jeff Law <law@redhat.com>
16741 PR tree-optimization/57124
16742 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16743 conversion feeding a condition if the range has an overflow
16744 if -fstrict-overflow. Add warnings for when we do make the
16747 2013-05-24 Dehao Chen <dehao@google.com>
16749 * tree-cfg.c (locus_discrim_map): Fix the typo.
16750 (locus_discrim_hasher): Likewise.
16751 (locus_discrim_hasher::hash): Likewise.
16752 (locus_discrim_hasher::equal): Likewise.
16754 2013-05-24 Martin Jambor <mjambor@suse.cz>
16756 PR tree-optimization/57294
16757 * cgraph.h (ipa_record_stmt_references): Declare.
16758 * cgraphbuild.c (ipa_record_stmt_references): New function.
16759 (build_cgraph_edges): Use ipa_record_stmt_references.
16760 (rebuild_cgraph_edges): Likewise.
16761 (cgraph_rebuild_references): Likewise.
16762 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16763 associated with the old statement and build references from the
16764 newly built statements.
16765 * ipa-ref.c (ipa_remove_stmt_references): New function.
16766 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16768 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16770 * lra-constraints.c (emit_spill_move): Use smaller mode for
16772 (check_and_process_move): Consider mem-reg moves for secondary
16774 (curr_insn_transform): Don't lose insns emitted before for
16775 secondary memory moves.
16776 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16777 reg set up in the current insn.
16779 2013-05-24 Dehao Chen <dehao@google.com>
16781 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16783 (locus_descrim_hasher::equal): Likewise.
16784 (build_gimple_cfg): New discriminator assignment algorithm.
16785 (make_edges): Likewise.
16786 (next_discriminator_for_locus): Likewise.
16787 (same_line_p): Likewise.
16788 (assign_discriminators): Likewise.
16789 (make_cond_expr_edges): Likewise.
16790 (make_gimple_switch_edges): Likewise.
16791 (make_goto_expr_edges): Likewise.
16792 (make_gimple_asm_edges): Likewise.
16794 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16796 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16797 X format specifier to only display bottom 16 bits.
16798 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16799 immediate to match for operand 2, since it will be masked.
16801 2013-05-24 Richard Biener <rguenther@suse.de>
16803 PR tree-optimization/57287
16804 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16805 all SSA names that occur in abnormal PHIs.
16807 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16810 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16811 that index is not negative.
16813 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16815 PR rtl-optimization/55177
16816 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16817 (simplify_byte_swapping_operation): New.
16818 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16819 (simplify_relational_operation_1): Deal with BSWAP.
16821 2013-05-23 Richard Henderson <rth@redhat.com>
16824 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16825 (ix86_reorg): Call it.
16827 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16830 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16831 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16832 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16834 2013-05-23 Christian Bruel <christian.bruel@st.com>
16837 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16839 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16840 Marcus Shawcroft <marcus.shawcroft@arm.com>
16842 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16843 * config/aarch64/constraints.md (Usa): Remove.
16844 * doc/md.texi (AArch64 Usa): Remove.
16846 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16847 Marcus Shawcroft <marcus.shawcroft@arm.com>
16849 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16850 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16851 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16852 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16854 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16856 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16857 instruction (AdvSIMD).
16858 * config/aarch64/aarch64-builtins.c
16859 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16860 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16862 2013-05-23 Martin Jambor <mjambor@suse.cz>
16864 PR middle-end/57347
16865 * tree.h (contains_bitfld_component_ref_p): Declare.
16866 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16867 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16869 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16870 not access a bit-field. Assert all final offsets are byte-aligned.
16872 2013-05-23 Richard Biener <rguenther@suse.de>
16874 PR tree-optimization/57380
16875 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16876 least one invariant or re-used load.
16877 * passes.c (init_optimization_passes): Move pass_phiprop before
16880 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16882 * config/aarch64/aarch64-simd.md
16883 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16885 2013-05-23 Richard Biener <rguenther@suse.de>
16887 PR middle-end/57381
16888 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16889 OEP_CONSTANT_ADDRESS_OF retained.
16891 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16893 PR middle-end/57344
16894 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16895 don't lower unit. Handle unit not being always BITS_PER_WORD.
16897 2013-05-23 Richard Biener <rguenther@suse.de>
16899 PR rtl-optimization/57341
16900 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16901 instead of true_dependence.
16903 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16905 * bb-reorder.c (branch_threshold): Make const.
16906 (exec_threshold): Ditto.
16908 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16909 Pat Haugen <pthaugen@us.ibm.com>
16910 Peter Bergner <bergner@vnet.ibm.com>
16912 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16913 documentation for the power8 crypto builtins.
16915 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16917 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16918 macros for defining power8 builtin functions.
16919 (BU_P8V_AV_2): Likewise.
16920 (BU_P8V_AV_P): Likewise.
16921 (BU_P8V_VSX_1): Likewise.
16922 (BU_P8V_OVERLOAD_1): Likewise.
16923 (BU_P8V_OVERLOAD_2): Likewise.
16924 (BU_CRYPTO_1): Likewise.
16925 (BU_CRYPTO_2): Likewise.
16926 (BU_CRYPTO_3): Likewise.
16927 (BU_CRYPTO_OVERLOAD_1): Likewise.
16928 (BU_CRYPTO_OVERLOAD_2): Likewise.
16929 (XSCVSPDP): Fix typo, point to the correct instruction.
16930 (VCIPHER): Add power8 crypto builtins.
16931 (VCIPHERLAST): Likewise.
16932 (VNCIPHER): Likewise.
16933 (VNCIPHERLAST): Likewise.
16934 (VPMSUMB): Likewise.
16935 (VPMSUMH): Likewise.
16936 (VPMSUMW): Likewise.
16937 (VPERMXOR_V2DI): Likewise.
16938 (VPERMXOR_V4SI: Likewise.
16939 (VPERMXOR_V8HI: Likewise.
16940 (VPERMXOR_V16QI: Likewise.
16941 (VSHASIGMAW): Likewise.
16942 (VSHASIGMAD): Likewise.
16943 (VPMSUM): Likewise.
16944 (VPERMXOR): Likewise.
16945 (VSHASIGMA): Likewise.
16947 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16948 __CRYPTO__ if the crypto instructions are available.
16949 (altivec_overloaded_builtins): Add support for overloaded power8
16952 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16953 support for power8 crypto builtins.
16954 (builtin_function_type): Likewise.
16955 (altivec_init_builtins): Add support for builtins that take vector
16956 long long (V2DI) arguments.
16958 * config/rs6000/crypto.md: New file, define power8 crypto
16961 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16962 Pat Haugen <pthaugen@us.ibm.com>
16963 Peter Bergner <bergner@vnet.ibm.com>
16965 * doc/invoke.texi (Option Summary): Add power8 options.
16966 (RS/6000 and PowerPC Options): Likewise.
16968 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16969 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16970 wm, wn, wr documentation.
16972 * config/rs6000/constraints.md (wm): New constraint for VSX
16973 registers if direct move instructions are enabled.
16974 (wn): New constraint for no registers.
16975 (wq): New constraint for quad word even GPR registers.
16976 (wr): New constraint if 64-bit instructions are enabled.
16977 (wv): New constraint if power8 vector instructions are enabled.
16978 (wQ): New constraint for quad word memory locations.
16980 * config/rs6000/predicates.md (const_0_to_15_operand): New
16981 constraint for 0..15 for crypto instructions.
16982 (gpc_reg_operand): If VSX allow registers in VSX registers as well
16983 as GPR and floating point registers.
16984 (int_reg_operand): New predicate to match only GPR registers.
16985 (base_reg_operand): New predicate to match base registers.
16986 (quad_int_reg_operand): New predicate to match even GPR registers
16987 for quad memory operations.
16988 (vsx_reg_or_cint_operand): New predicate to allow vector logical
16989 operations in both GPR and VSX registers.
16990 (quad_memory_operand): New predicate for quad memory operations.
16991 (reg_or_indexed_operand): New predicate for direct move support.
16993 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
16994 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
16995 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
16996 (POWERPC_MASKS): Add power8 options.
16997 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
17000 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
17001 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
17003 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
17004 (-mpower8-fusion): New power8 options.
17005 (-mpower8-fusion-sign): Likewise.
17006 (-mpower8-vector): Likewise.
17007 (-mcrypto): Likewise.
17008 (-mdirect-move): Likewise.
17009 (-mquad-memory): Likewise.
17011 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
17012 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
17014 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
17015 (rs6000_debug_vector_unit): Add p8_vector.
17016 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
17017 definitions. Also print fusion state.
17018 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
17019 (rs6000_builtin_mask_calculate): Add power8 builtin support.
17020 (rs6000_option_override_internal): Add support for power8.
17021 (rs6000_common_init_builtins): Add debugging for skipped builtins
17022 if -mdebug=builtin.
17023 (rs6000_adjust_cost): Add power8 support.
17024 (rs6000_issue_rate): Likewise.
17025 (insn_must_be_first_in_group): Likewise.
17026 (insn_must_be_last_in_group): Likewise.
17027 (force_new_group): Likewise.
17028 (rs6000_register_move_cost): Likewise.
17029 (rs6000_opt_masks): Likewise.
17031 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
17032 power8 capable assembler, default to power7 options.
17033 (TARGET_DIRECT_MOVE): Likewise.
17034 (TARGET_CRYPTO): Likewise.
17035 (TARGET_P8_VECTOR): Likewise.
17036 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
17037 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
17038 (VECTOR_MEM_P8_VECTOR_P): Likewise.
17039 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
17040 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
17041 (TARGET_XSCVDPSPN): Likewise.
17042 (TARGET_XSCVSPDPN): Likewsie.
17043 (TARGET_SYNC_HI_QI): Likewise.
17044 (TARGET_SYNC_TI): Likewise.
17045 (MASK_CRYPTO): Likewise.
17046 (MASK_DIRECT_MOVE): Likewise.
17047 (MASK_P8_FUSION): Likewise.
17048 (MASK_P8_VECTOR): Likewise.
17049 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
17050 temporary used by some of the direct move instructions to get two FP
17051 temporary registers does not force creation of a stack frame.
17052 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
17053 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
17054 that any VSX registers are tieable, even if they are also an
17055 Altivec vector mode.
17056 (r6000_reg_class_enum): Add wm, wr, wv constraints.
17057 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
17058 (RS6000_BTM_CRYPTO): Likewise.
17059 (RS6000_BTM_COMMON): Likewise.
17061 * config/rs6000/rs6000.md (cpu attribute): Add power8.
17062 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
17063 (enum rs6000_vector): Add power8 vector support.
17065 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17069 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
17070 (any_sibcall_could_use_r3): this and handle indirect calls.
17071 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
17073 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17075 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
17077 2013-05-22 Richard Biener <rguenther@suse.de>
17079 PR middle-end/57349
17080 * profile.c (branch_prob): Do not split blocks that are
17081 abnormally receiving from ECF_RETURNS_TWICE functions.
17083 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17085 * recog.c (offsettable_address_addr_space_p): Fix calculation of
17086 address mode. Move pointer mode initialization to the same place.
17088 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
17090 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
17091 while it has any effect.
17093 2013-05-21 Easwaran Raman <eraman@google.com>
17095 PR tree-optimization/57322
17096 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
17097 UID of the statement added to the BB to be 1.
17099 2013-05-21 Jakub Jelinek <jakub@redhat.com>
17101 PR tree-optimization/57331
17102 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
17103 of conversion from pointer type to integral type with integer.
17105 2013-05-21 Martin Jambor <mjambor@suse.cz>
17108 * ipa-prop.c (ipa_read_node_info): Process param_used and
17109 controlled_uses in the same order as when writing.
17111 2013-05-21 Magnus Granberg <baldrick@free.fr>
17114 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17116 2013-05-21 Richard Biener <rguenther@suse.de>
17118 PR tree-optimization/57318
17119 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17120 estimate stmts with side-effects as likely eliminated.
17122 2013-05-21 Richard Biener <rguenther@suse.de>
17124 PR tree-optimization/57330
17125 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17126 preserve the call stmts fntype.
17128 2013-05-21 Richard Biener <rguenther@suse.de>
17130 PR tree-optimization/57303
17131 * tree-ssa-sink.c (statement_sink_location): Improve killing
17132 stmt detection and properly handle self-assignments.
17134 2013-05-21 Christian Bruel <christian.bruel@st.com>
17136 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17137 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17138 registers. Set register size out of the PARALLEL loop.
17140 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17143 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17144 (*fmasf4, *fmasf4_media): New insns.
17146 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17148 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17149 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17150 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17151 (mips_idiv_insns): Update the comments to say that the returned
17152 instruction counts are in units of BASE_INSN_LENGTH.
17153 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17154 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17155 using 2 rather than 4 as the length of indirect MIPS16 and
17156 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17157 length of a NOP. Don't divide MIPS16 lengths by 2.
17158 (mips16_split_long_branches): Assume a branch is long if the
17159 length is greater than 4 rather than 8.
17160 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17161 rather than multiplying them by 2. Multiply instruction counts
17162 by BASE_INSN_LENGTH rather than 4.
17163 (*jump_mips16, tls_get_tp_mips16_<mode>)
17164 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17166 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17168 * config/mips/mips.md (extended_mips16): Remove branch case.
17169 (length): Remove duplicated extended_mips16 test.
17171 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17173 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17175 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17177 * recog.h (Recog_data): Rename to...
17178 (recog_data_d): ...this.
17179 (recog_data): Update accordingly.
17180 * recog.c (recog_data): Likewise.
17181 * reload.c (save_recog_data): Likewise.
17182 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17183 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17185 2013-05-17 Julian Brown <julian@codesourcery.com>
17187 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17188 found in a REG_EQUAL note, invalidate it.
17190 2013-05-17 Easwaran Raman <eraman@google.com>
17192 * tree-ssa-reassoc.c (find_insert_point): New function.
17193 (insert_stmt_after): Likewise.
17194 (get_def_stmt): Likewise.
17195 (ensure_ops_are_available): Likewise.
17196 (not_dominated_by): Likewise.
17197 (rewrite_expr_tree): Do not move statements beyond what is
17198 necessary. Remove call to swap_ops_for_binary_stmt...
17199 (reassociate_bb): ... and move it here.
17200 (build_and_add_sum): Assign UIDs for new statements.
17201 (linearize_expr): Likewise.
17202 (do_reassoc): Renumber gimple statement UIDs.
17204 2013-05-17 Jan Hubicka <jh@suse.cz>
17206 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17208 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17209 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17211 (output_weakrefs): Update.
17213 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17214 Martin Jambor <mjambor@suse.cz>
17216 PR middle-end/57276
17217 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17218 value that corresponds to the given aggval is found in values vector.
17220 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17222 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17223 sse, sse2, sse3, ssse3 and sse4a flags to options.
17225 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17227 * gengtype-state.c: (s_expr_writer): New class, to handle
17228 prettifying of output layout of s-expressions.
17229 (state_writer): New class, to write out gtype.state.
17230 (state_written_type_count): Move this variable into member data of
17232 (s_expr_writer::s_expr_writer): New code: constructor for new class
17233 (state_writer::state_writer(): ditto
17234 (s_expr_writer::write_new_line): New function
17235 (s_expr_writer::write_any_indent): ditto
17236 (s_expr_writer::begin_s_expr): ditto
17237 (s_expr_writer::end_s_expr): ditto
17238 (write_state_fileloc): convert to method of state_writer...
17239 (state_writer:: write_state_fileloc): ...and use methods of
17240 s_expr_writer to write indentation into the gtype.state output file
17241 to visually represent the hierarchical structure of the list
17243 (write_state_fields): ditto, renaming to...
17244 (state_writer::write_state_fields)
17245 (write_state_a_string): ditto, renaming to...
17246 (state_writer::write_state_a_string)
17247 (write_state_string_option): ditto, renaming to...
17248 (state_writer::write_state_string_option)
17249 (write_state_type_option): ditto, renaming to...
17250 (state_writer::write_state_type_option)
17251 (write_state_nested_option): ditto, renaming to...
17252 (state_writer::write_state_nested_option)
17253 (write_state_option): ditto, renaming to...
17254 (state_writer::write_state_option)
17255 (write_state_options): ditto, renaming to...
17256 (state_writer::write_state_options)
17257 (write_state_lang_bitmap): ditto, renaming to...
17258 (state_writer::write_state_lang_bitmap)
17259 (write_state_version): ditto, renaming to...
17260 (state_writer::write_state_version)
17261 (write_state_scalar_type): ditto, renaming to...
17262 (state_writer::write_state_scalar_type)
17263 (write_state_string_type): ditto, renaming to...
17264 (state_writer::write_state_string_type)
17265 (write_state_undefined_type): ditto, renaming to...
17266 (state_writer::write_state_undefined_type)
17267 (write_state_struct_union_type): ditto, renaming to...
17268 (state_writer::write_state_struct_union_type)
17269 (write_state_struct_type): ditto, renaming to...
17270 (state_writer::write_state_struct_type)
17271 (write_state_user_struct_type): ditto, renaming to...
17272 (state_writer::write_state_user_struct_type)
17273 (write_state_lang_struct_type): ditto, renaming to...
17274 (state_writer::write_state_lang_struct_type)
17275 (write_state_param_struct_type): ditto, renaming to...
17276 (state_writer::write_state_param_struct_type)
17277 (write_state_pointer_type): ditto, renaming to...
17278 (state_writer::write_state_pointer_type)
17279 (write_state_array_type): ditto, renaming to...
17280 (state_writer::write_state_array_type)
17281 (write_state_gc_used): ditto, renaming to...
17282 (state_writer::write_state_gc_used)
17283 (write_state_common_type_content): ditto, renaming to...
17284 (state_writer::write_state_common_type_content)
17285 (write_state_type): ditto, renaming to...
17286 (state_writer::write_state_type)
17287 (write_state_pair_list): ditto, renaming to...
17288 (state_writer::write_state_pair_list)
17289 (write_state_pair): ditto, renaming to...
17290 (state_writer::write_state_pair)
17291 (write_state_typedefs): ditto, renaming to...
17292 (state_writer::write_state_typedefs)
17293 (write_state_structures): ditto, renaming to...
17294 (state_writer::write_state_structures)
17295 (write_state_param_structs): ditto, renaming to...
17296 (state_writer::write_state_param_structs)
17297 (write_state_variables): ditto, renaming to...
17298 (state_writer::write_state_variables)
17299 (write_state_srcdir): ditto, renaming to...
17300 (state_writer::write_state_srcdir)
17301 (write_state_files_list): ditto, renaming to...
17302 (state_writer::write_state_files_list)
17303 (write_state_languages): ditto, renaming to...
17304 (state_writer::write_state_languages)
17305 (write_state): create a state_writer instance and use it when
17306 writing out the state file
17308 2013-05-17 Mike Stump <mikestump@comcast.net>
17310 PR rtl-optimization/57304
17311 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17312 accessing DF_REF_REAL_LOC.
17314 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17316 PR rtl-optimization/57281
17317 PR rtl-optimization/57300
17318 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17319 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17320 what the other splitter did if the registers are dead.
17322 2013-05-17 Richard Biener <rguenther@suse.de>
17324 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17327 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17329 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17332 2013-05-17 Marek Polacek <polacek@redhat.com>
17334 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17335 length when doing non-zero store of storing '\0' to '\0'.
17337 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17339 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17340 vect_external_def oprnd1 with loop_vinfo, try to emit
17341 optional cast, negation and and stmts on the loop preheader
17342 edge instead of into the pattern def seq.
17344 PR tree-optimization/57051
17345 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17346 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17348 2013-05-16 Nick Clifton <nickc@redhat.com>
17350 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17351 (rl78_is_naked_func): New function.
17352 (rl78_expand_prologue): Skip prologue generation for naked functions.
17353 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17354 * doc/extend.texi (naked): Add RL78 to the list of processors
17355 that supports this attribute.
17357 2013-05-16 Jeff Law <law@redhat.com>
17359 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17361 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17363 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17364 cache parameters using detect_caches_amd also for CYRIX,
17365 NSC and TM2 signatures.
17367 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17368 Dzianis Kahanovich <mahatma@eu.by>
17372 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17373 VIA/Centaur processors and determine their cache parameters
17374 using detect_caches_amd.
17376 2013-05-16 Teresa Johnson <tejohnson@google.com>
17378 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17379 (rtl_verify_edges): New function.
17380 (rtl_verify_bb_insns): Ditto.
17381 (rtl_verify_bb_pointers): Ditto.
17382 (rtl_verify_bb_insn_chain): Ditto.
17383 (rtl_verify_fallthru): Ditto.
17384 (rtl_verify_bb_layout): Ditto.
17385 (rtl_verify_flow_info_1): Outline checks into new functions.
17386 (rtl_verify_flow_info): Ditto.
17388 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17390 * cfghooks.c (copy_bbs): Add update_dominance argument.
17391 * cfghooks.h (copy_bbs): Update prototype.
17392 * tree-cfg.c (gimple_duplicate_sese_region):
17393 Add update_dominance argument.
17394 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17395 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17396 gimple_duplicate_sese_region call.
17397 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17398 Update copy_bbs call.
17399 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17400 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17402 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17404 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17405 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17406 vect_recog_rotate_pattern.
17407 (vect_recog_rotate_pattern): New function.
17409 2013-05-16 Jason Merrill <jason@redhat.com>
17411 * Makefile.in (LLINKER): New variable.
17412 (mostlyclean): Remove link mutex.
17413 * configure.ac: Handle --enable-link-mutex.
17414 * lock-and-run.sh: New script.
17416 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17419 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17422 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17424 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17426 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17428 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17429 * config/arm/arm.c (next_consecutive_mem): New function.
17430 (gen_movmem_ldrd_strd): Likewise.
17431 * config/arm/arm.md (movmemqi): Update condition and code.
17432 (unaligned_loaddi, unaligned_storedi): New patterns.
17434 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17436 * config.gcc: Obsolete *-*-solaris2.9*.
17437 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17439 2013-05-16 Richard Biener <rguenther@suse.de>
17441 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17442 earlier, after GRAPHITE transforms and IV canonicalization.
17444 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17446 * omp-low.c (extract_omp_for_data): For collapsed loops,
17447 if at least one of the loops is known at compile time to
17448 iterate zero times, set count to 0.
17449 (expand_omp_regimplify_p): New function.
17450 (expand_omp_for_generic): For collapsed loops, if at least
17451 one of the loops isn't known to iterate at least once,
17452 add runtime check with setting count to 0.
17453 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17454 For unsigned types if it isn't known at compile time that
17455 the loop will iterate at least once, add runtime check to bypass
17456 the whole loop if initial condition isn't true.
17458 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17460 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17462 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17464 PR middle-end/57286
17465 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17466 transformations to avoid an infinite loop.
17468 2013-05-16 Marek Polacek <polacek@redhat.com>
17470 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17472 2013-05-15 Leif Ekblad <leif@rdos.net>
17474 * config/i386/i386.c (ix86_decompose_address): Use
17475 DEFAULT_TLS_SEG_REG to access TLS segment register.
17476 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17477 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17478 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17480 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17483 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17484 sibling calls to functions that would normally be lazily bound,
17485 unless $gp is call-clobbered.
17487 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17489 * config/i386/i386.c (ix86_option_override_internal): Update
17490 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17491 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17492 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17493 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17495 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17497 2013-05-15 Andreas Schwab <schwab@suse.de>
17499 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17500 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17503 2013-05-15 Teresa Johnson <tejohnson@google.com>
17505 * loop-unroll.c (report_unroll_peel): Check decision before
17506 emitting unroll/peel message.
17508 2013-05-15 Teresa Johnson <tejohnson@google.com>
17510 * function.h (has_bb_partition): New rtl_data flag.
17511 (bb_reorder_complete): Ditto.
17512 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17513 instead of flag_reorder_blocks_and_partition.
17514 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17515 with some enhancements.
17516 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17517 * bb-reorder.c (connect_traces): Check for has_bb_partition
17518 instead of flag_reorder_blocks_and_partition.
17519 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17520 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17521 verify_hot_cold_block_grouping.
17522 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17524 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17527 * config/arm/predicates.md (call_insn_operand): New predicate.
17528 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17529 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17530 if insn is not a tail call.
17531 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17533 * config/arm/arm.h (enum reg_class): New caller save register class.
17534 (REG_CLASS_NAMES): Likewise.
17535 (REG_CLASS_CONTENTS): Likewise.
17536 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17539 2013-05-15 Richard Biener <rguenther@suse.de>
17541 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17542 of MSG_OPTIMIZED_LOCATIONS.
17543 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17544 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17546 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17547 of MSG_OPTIMIZED_LOCATIONS.
17548 (execute_vect_slp): Likewise.
17549 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17550 (vect_create_cond_for_alias_checks): Likewise.
17551 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17552 (vect_recog_widen_mult_pattern): Likewise.
17553 (vect_recog_widen_sum_pattern): Likewise.
17554 (vect_recog_over_widening_pattern): Likewise.
17555 (vect_recog_widen_shift_pattern): Likewise.
17556 (vect_recog_vector_vector_shift_pattern): Likewise.
17557 (vect_recog_divmod_pattern): Likewise.
17558 (vect_recog_mixed_size_cond_pattern): Likewise.
17559 (vect_recog_bool_pattern): Likewise.
17560 (vect_pattern_recog_1): Likewise.
17562 2013-05-15 Martin Jambor <mjambor@suse.cz>
17564 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17565 non-functions to builtin_unreachable.
17566 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17567 correct when new direct edges were discovered.
17569 2013-05-15 Martin Jambor <mjambor@suse.cz>
17571 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17572 header, print symbol order instead of node uid, print more information
17573 about indirect edge targets.
17574 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17576 (ipa_make_edge_direct_to_target): Likewise.
17577 (remove_described_reference): Likewise.
17578 (propagate_controlled_uses): Likewise.
17579 (ipa_print_node_params): Also print symbol order.
17580 (ipcp_transform_function): Print symbol order instead of node uids.
17581 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17582 (cgraph_get_create_real_symbol_node): Likewise.
17583 * ipa-cp.c (print_lattice): Likewise.
17584 (print_all_lattices): Likewise.
17585 (determine_versionability): Likewise.
17586 (initialize_node_lattices): Likewise.
17587 (estimate_local_effects): Likewise.
17588 (update_profiling_info): Likewise.
17589 (create_specialized_node): Likewise.
17590 (perhaps_add_new_callers): Likewise.
17591 (decide_about_value): Likewise.
17592 (decide_whether_version_node): Likewise.
17593 (identify_dead_nodes): Likewise.
17594 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17595 (dump_inline_summary): Likewise.
17596 (estimate_node_size_and_time): Likewise.
17597 (inline_analyze_function): Likewise.
17598 * ipa-inline.c (report_inline_failed_reason): Likewise.
17599 (want_early_inline_function_p): Likewise.
17600 (edge_badness): Likewise.
17601 (update_edge_key): Likewise.
17602 (inline_small_functions): Likewise. Add dumping of order to two other
17604 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17605 instead of node uids.
17606 (propagate_pure_const): Likewise.
17607 (propagate_pure_const): Likewise.
17608 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17609 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17610 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17612 * tree-pretty-print.c (dump_function_header): Likewise.
17613 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17614 Print symbol order instead of node uids.
17616 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17618 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17619 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17621 2013-05-15 Richard Biener <rguenther@suse.de>
17623 PR tree-optimization/57275
17624 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17625 return value for fail to do runtime alias checks for gather loads.
17627 2013-05-15 Jan Hubicka <jh@suse.cz>
17631 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17632 weakrefs are not external.
17633 (lto_symtab_merge_decls): Fix thinko when dealing with
17634 non-lto_symtab decls.
17635 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17636 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17637 * varpool.c (dump_varpool_node): Dump more flags.
17639 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17641 * config/i386/i386.c (processor_alias_table): Add instruction
17642 FSGSBASE for AMD bdver3 architecture.
17644 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17646 * tree.c (warn_deprecated_use): Print file:line using locus color.
17647 * diagnostic.c (diagnostic_report_current_module): Print file:line
17648 and file:line:column using locus color.
17650 2013-05-14 Mike Stump <mikestump@comcast.net>
17652 * gdbinit.in: Add __null.
17654 2013-05-14 Mike Stump <mikestump@comcast.net>
17656 * recog.h: Rename struct recog_data to Recog_data.
17657 * recog.c: Likewise.
17658 * reload.c (can_reload_into): Likewise.
17659 * config/picochip/picochip.c: Likewise.
17661 2013-05-14 Mike Stump <mikestump@comcast.net>
17663 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17665 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17667 * resource.h (struct resources): Remove unch_memory member.
17668 (CLEAR_RESOURCE): Don't clear unch_memory.
17669 * resource.c (mark_referenced_resources): Don't set it.
17670 (mark_set_resources): Likewise.
17671 (mark_target_live_regs): Don't clear it.
17672 (init_resource_info): Likewise.
17673 * reorg.c (resource_conflicts_p): Don't compare it.
17674 (redundant_insn): Don't set it.
17676 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17678 * emit-rtl.c (next_label): Remove unused function.
17679 (skip_consecutive_labels, link_cc0_insns): Move to ...
17680 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17681 only place where these functions are used, and make them static.
17683 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17685 * fold-const.c (fold_negate_expr): Handle vectors.
17686 (fold_truth_not_expr): Make it static.
17687 (fold_invert_truthvalue): New static function.
17688 (invert_truthvalue_loc): Handle vectors. Do not call
17689 fold_truth_not_expr directly.
17690 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17691 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17692 (fold_comparison): Handle vector constants.
17693 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17694 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17695 * tree.h (fold_truth_not_expr): Remove declaration.
17697 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17699 * config/aarch64/aarch64-simd.md
17700 (aarch64_vcond_internal<mode>): Rename to...
17701 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17702 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17703 float modes. Clarify all iterator modes.
17704 (vcond<mode><mode>): Use new name for vcond expanders.
17705 (vcond<v_cmp_result><mode>): Likewise.
17706 (vcondu<mode><mode>: Likewise.
17707 * config/aarch64/iterators.md (VDQF_COND): New.
17709 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17712 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17713 variable for the shift amount. Check that we shift by non-negative
17716 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17719 * config/arm/arm.h (EPILOGUE_USES): Only return true
17720 for LR_REGNUM after epilogue_completed.
17722 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17724 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17725 is error_mark_node.
17727 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17730 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17731 and Solaris 11+/x86 with gld.
17732 * configure: Regenerate.
17734 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17736 * expmed.c (expand_shift_1): Canonicalize rotates by
17737 constant bitsize / 2 to bitsize - 1.
17738 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17739 case ROTATERT>: Likewise.
17742 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17744 * config/i386/i386.md (rotateinv): New code attr.
17745 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17746 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17747 roll $31, %eax, etc.
17749 2013-05-14 Richard Biener <rguenther@suse.de>
17751 PR middle-end/57235
17752 * tree-eh.c (sink_clobbers): Give up for successors with
17753 multiple predecessors and no virtual uses.
17755 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17757 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17758 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17760 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17762 PR middle-end/57251
17763 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17764 the case when both op0 and op1 have VOIDmode.
17766 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17768 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17769 in multiply-accumulate mode.
17771 2013-05-13 Guozhi Wei <carrot@google.com>
17773 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17775 2013-05-13 Kai Tietz <ktietz@redhat.com>
17778 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17779 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17780 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17781 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17782 (ix86_expand_prologue): Likewise.
17783 (ix86_expand_split_stack_prologue): Likewise.
17784 (legitimate_pic_address_disp_p): Likewise.
17785 (legitimize_pic_address): Likewise.
17786 (legitimize_tls_address): Likewise.
17787 (legitimize_pe_coff_symbol): Likewise.
17788 (output_pic_addr_const): Likewise.
17789 (construct_plt_address): Likewise.
17790 (ix86_expand_call): Likewise.
17791 (x86_output_mi_thunk): Likewise.
17792 (x86_function_profiler): Likewise.
17794 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17796 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17797 similar switch cases.
17798 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17799 (aarch64_simd_mov_to_<mode>low): Delete.
17800 (aarch64_simd_mov_to_<mode>high): Delete.
17801 (move_lo_quad_<mode>): Add w<-r alternative.
17802 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17803 (aarch64_simd_mov_from_*): Update type attribute.
17804 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17807 2013-05-13 Jan Hubicka <jh@suse.cz>
17809 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17810 * config/i386/i386.c (ix86_compute_frame_layout,
17811 ix86_expand_epilogue, emit_i387_cw_initialization,
17812 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17813 ix86_local_alignment): Fix use of size/speed predicates.
17815 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17817 PR tree-optimization/45216
17818 PR tree-optimization/57157
17819 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17820 the (-Y) & (B - 1) variant if OP is |.
17821 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17822 return shifted. Use (-op1) & (prec - 1) as other_amount
17823 instead of prec - op1.
17825 2013-05-13 Martin Jambor <mjambor@suse.cz>
17827 PR middle-end/42371
17828 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17829 (ipa_constant_data): New type.
17830 (ipa_jump_func): Use ipa_constant_data to hold information about
17831 constant jump functions.
17832 (ipa_get_jf_constant): Adjust to jump function type changes.
17833 (ipa_get_jf_constant_rdesc): New function.
17834 (ipa_param_descriptor): New field controlled_uses.
17835 (ipa_get_controlled_uses): New function.
17836 (ipa_set_controlled_uses): Likewise.
17837 * ipa-ref.h (ipa_find_reference): Declare.
17838 * ipa-prop.c (ipa_cst_ref_desc): New type.
17839 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17841 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17842 New parameter cs. Adjust all callers.
17843 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17844 (remove_described_reference): New function.
17845 (jfunc_rdesc_usable): Likewise.
17846 (try_make_edge_direct_simple_call): Decrement controlled use count,
17847 attempt to remove reference if it hits zero.
17848 (combine_controlled_uses_counters): New function.
17849 (propagate_controlled_uses): Likewise.
17850 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17851 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17852 (ipa_print_node_params): Print described use counter.
17853 (ipa_write_jump_function): Adjust to jump function type changes.
17854 (ipa_read_jump_function): New parameter CS, pass it to
17855 ipa_set_jf_constant. Adjust caller.
17856 (ipa_write_node_info): Stream controlled use count
17857 (ipa_read_node_info): Likewise.
17858 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17860 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17861 count. Remove cloning-added reference if it reaches zero.
17862 * ipa-ref.c (ipa_find_reference): New function.
17864 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17866 * config/i386/i386.c (processor_target_table): Modified default
17867 alignment values for AMD BD and BT architectures.
17869 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17871 * tree-vect-generic.c (uniform_vector_p): Move ...
17872 * tree.c (uniform_vector_p): ... here.
17873 * tree.h (uniform_vector_p): Declare it.
17874 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17877 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17879 PR tree-optimization/57230
17880 * tree-ssa-strlen.c (handle_char_store): Record length for
17881 array store from STRING_CST.
17883 PR tree-optimization/57230
17884 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17887 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17889 * config/epiphany/epiphany.c (epiphany_init): Check size of
17890 NUM_MODES_FOR_MODE_SWITCHING.
17891 (epiphany_expand_prologue):
17892 Remove CONFIG_REGNUM initial value handling code.
17893 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17894 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17895 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17896 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17897 Don't return 1 for FP_MODE_NONE.
17898 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17899 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17900 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17901 * config/epiphany/epiphany.md (save_config): New pattern.
17903 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17905 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17907 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17909 * config/i386/i386.md (memory): Handle sseishft1.
17910 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17911 (*vec_extractv2di_1): Ditto.
17913 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17915 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17918 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17920 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17921 Add mthumb/march=armv7-a multilib.
17922 Add mthumb/march=armv7-r multilib.
17923 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17925 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17927 * config/v850/t-rtems: Add more multilibs.
17929 2013-05-10 Richard Biener <rguenther@suse.de>
17931 PR tree-optimization/57214
17932 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17933 not propagate from SSA names that occur in abnormal PHI nodes.
17935 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17937 * stor-layout.c (element_precision): New function.
17938 * machmode.h (element_precision): Declare it.
17939 * tree.c (build_minus_one_cst): New function.
17940 (element_precision): Likewise.
17941 * tree.h (build_minus_one_cst): Declare new function.
17942 (element_precision): Likewise.
17943 * fold-const.c (operand_equal_p): Use element_precision.
17944 (fold_binary_loc): Handle vector types.
17945 * convert.c (convert_to_integer): Use element_precision.
17946 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17949 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17951 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17952 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17953 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17954 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17955 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17956 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17957 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17958 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17959 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17960 (Uuw8): New constraints.
17961 (Usb4): Move into alphabetical order.
17962 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17963 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17964 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17965 previously unnamed patterns.
17966 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17967 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17968 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17969 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17970 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17971 of specific lengths.
17973 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17975 * config/i386/i386.md (rotateinv): New code attr.
17976 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17977 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17978 roll $31, %eax, etc.
17980 PR tree-optimization/45216
17981 PR tree-optimization/57157
17982 * tree-ssa-forwprop.c (simplify_rotate): New function.
17983 (ssa_forward_propagate_and_combine): Call it.
17985 2013-05-10 Richard Biener <rguenther@suse.de>
17987 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
17988 disable peeling when we version for aliasing.
17989 (vector_alignment_reachable_p): Honor explicit user alignment.
17990 (vect_supportable_dr_alignment): Likewise.
17991 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
17992 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
17993 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
17994 then peeling to arrange for the cost-model check to come first.
17996 2013-05-10 Alan Modra <amodra@gmail.com>
17998 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
17999 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
18000 * configure: Regenerate.
18002 2013-05-10 Alan Modra <amodra@gmail.com>
18005 * varasm.c (default_elf_select_section): Move !DECL_P check..
18006 (get_named_section): ..to here before calling get_section_name.
18008 (default_section_type_flags): Add DECL_P check.
18009 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
18010 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
18012 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
18014 * config/epiphany/epiphany.c (epiphany_expand_prologue):
18015 When using gen_stack_adjust_str with a register offset, add a
18016 REG_FRAME_RELATED_EXPR note.
18018 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
18020 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
18021 (*vec_extractv4si_zext_mem): Ditto.
18022 (*vec_extractv2di): Add 0->x and x->x alternatives.
18023 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
18024 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
18026 2013-05-09 Jason Merrill <jason@redhat.com>
18028 N3639 C++1y VLA support
18029 * gimplify.c (gimplify_vla_decl): Don't touch an existing
18032 * tree.c (build_constructor_va): New.
18033 * tree.h: Declare it.
18035 2013-05-09 Martin Jambor <mjambor@suse.cz>
18038 * gimple-fold.c (canonicalize_constructor_val): Call
18039 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
18041 2013-05-09 Jan Hubicka <jh@suse.cz>
18042 Richard Biener <rguenther@suse.de>
18045 * symtab.c (symtab_make_decl_local): Do not add private names.
18047 2013-05-09 Jan Hubicka <jh@suse.cz>
18050 * symtab.c (insert_to_assembler_name_hash): Handle clones.
18051 (unlink_from_assembler_name_hash): Likewise.
18052 (symtab_prevail_in_asm_name_hash, symtab_register_node,
18053 symtab_unregister_node, symtab_initialize_asm_name_hash,
18054 change_decl_assembler_name): Update.
18056 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
18058 * config/aarch64/aarch64.md: New movtf split.
18059 (*movtf_aarch64): Update.
18060 (aarch64_movdi_tilow): Handle TF modes and rename to
18061 aarch64_movdi_<mode>low.
18062 (aarch64_movdi_tihigh): Handle TF modes and rename to
18063 aarch64_movdi_<mode>high
18064 (aarch64_movtihigh_di): Handle TF modes and rename to
18065 aarch64_mov<mode>high_di
18066 (aarch64_movtilow_di): Handle TF modes and rename to
18067 aarch64_mov<mode>low_di
18068 (aarch64_movtilow_tilow): Remove spurious whitespace.
18069 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
18071 (aarch64_print_operand): Update.
18073 2013-05-09 Alan Modra <amodra@gmail.com>
18075 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
18077 * configure: Regenerate.
18079 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18081 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
18082 splitter preparation statements.
18083 * config/i386/sse.md (*vec_extract* splitters): Ditto.
18084 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
18087 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18089 * gimple-ssa-strength-reduction.c (count_candidates): Change
18090 return value to int.
18091 (analyze_candidates_and_replace): Change type of length to int.
18093 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18095 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
18096 (*vec_extract<mode>): Use VI12_128 mode iterator.
18097 (*vec_extract<mode>_mem): Ditto.
18098 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
18101 2013-05-08 Diego Novillo <dnovillo@google.com>
18106 2012-08-17 Diego Novillo <dnovillo@google.com>
18109 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18110 * config.in: Regenerate.
18111 * configure: Regenerate.
18112 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18114 2013-05-08 Jan Hubicka <jh@suse.cz>
18117 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18118 * cgraph.h (symtab_node_base): Add unique_name.
18119 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18120 input_overwrite_node, input_varpool_node): Stream unique_name.
18121 * cgraphclones.c (cgraph_create_virtual_clone,
18122 cgraph_function_versioning): Set unique_name.
18123 * ipa.c (function_and_variable_visibility): Set unique_name.
18125 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18127 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18128 (alloc_cand_and_find_basis): Restrict conditional candidate
18129 processing to CAND_MULTs.
18131 2013-05-08 Jan Hubicka <jh@suse.cz>
18134 lto-symtab.c (lto_symtab_symbol_p): New function.
18135 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18136 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18137 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18138 Skip static symbols.
18140 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18142 PR tree-optimization/57200
18143 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18144 Only call inform if the preceding warning_at returns true.
18146 2013-05-07 Han Shen <shenhan@google.com>
18148 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18149 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18150 * common.opt (fstack-protector-strong): New option.
18151 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18152 * doc/invoke.texi (Optimization Options): Document
18153 "-fstack-protector-strong".
18154 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18156 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18158 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18160 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18162 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18163 it is not lhs of the insn.
18164 (setup_reg_equiv): Remove insn having equiv memory which it is not
18166 * lra-constraints.c (process_address): Try to improve generation
18167 code for address base + disp.
18168 (lra_constraints): Make correct the code for checking insn setting
18169 up backward equivalence. Remove insn only if it is in the init
18171 * lra-eliminations.c (update_reg_eliminate): Change return value.
18172 (lra_eliminate): Use the result.
18174 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18176 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18177 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18178 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18179 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18180 register target operands.
18181 (*vec_extractv8hi_sse2): New pattern.
18182 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18183 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18184 (*vec_extract<mode>_mem): New insn and split pattern.
18186 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18188 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18189 (TARGET_ASAN_SHADOW_OFFSET): Define.
18190 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18191 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18193 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18195 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18196 (incr_vec_index): Return -1 if increment not found.
18197 (create_add_on_incoming_edge): Assert if increment not found.
18198 (record_increment): Limit number of increments recorded.
18199 (all_phi_incrs_profitable): Return false if an increment not found.
18200 (replace_profitable_candidates): Don't process increments that were
18202 (analyze_candidates_and_replace): Limit size of incr_vec.
18204 2013-05-07 Richard Biener <rguenther@suse.de>
18206 * calls.c (special_function_p): setjmp-like functions are leaf.
18207 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18208 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18210 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18212 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18213 (aarch64_simd_mov<mode>): New expander.
18214 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18215 (aarch64_simd_mov_to_<mode>high): Likewise.
18216 (aarch64_simd_mov_from_<mode>low): Likewise.
18217 (aarch64_simd_mov_from_<mode>high): Likewise.
18218 (aarch64_dup_lane<mode>): Update.
18219 (aarch64_dup_lanedi): New instruction pattern.
18220 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18221 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18223 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18225 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18226 (replace_mult_candidate): Remove unnecessary argument; remove
18227 unnecessary parameter from call to introduce_cast_before_cand.
18228 (replace_unconditional_candidate): Remove unnecessary parameter
18229 from call to replace_mult_candidate.
18230 (replace_conditional_candidate): Likewise.
18231 (insert_initializers): Use make_temp_ssa_name.
18232 (introduce_cast_before_cand): Remove unnecessary argument; use
18233 make_temp_ssa_name.
18234 (replace_one_candidate): Remove unnecessary argument; remove
18235 unnecessary parameter from calls to introduce_cast_before_cand.
18236 (replace_profitable_candidates): Remove unnecessary parameters
18237 from calls to replace_one_candidate.
18239 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18241 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18242 phi def as possibly hiding a basis for a CAND_ADD whose operands
18243 have been commuted in the analysis.
18244 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18246 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18248 * config/aarch64/aarch64.md
18249 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18250 shift value between 0-4.
18252 2013-05-07 Richard Biener <rguenther@suse.de>
18254 * double-int.h (rshift): New overload.
18255 * double-int.c (rshift): New function.
18256 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18257 (create_reference_ops_from_ref): Remove.
18258 (vn_reference_insert): Use shared ops for constructing the
18259 reference and copy it.
18261 2013-05-07 Richard Biener <rguenther@suse.de>
18263 PR middle-end/57190
18264 * tree-eh.c (sink_clobbers): Properly propagate
18265 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18267 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18269 PR tree-optimization/57149
18270 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18271 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18272 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18273 uninit_undefined_value_p instead of ssa_undefined_value_p.
18276 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18277 for modifier == EXPAND_INITIALIZER.
18279 2013-05-07 Anton Blanchard <anton@samba.org>
18281 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18282 for powerpc64 little endian.
18283 * configure: Regenerate.
18285 2013-05-06 Graham Stott <grahams@btinternet.com>
18287 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18288 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18289 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18290 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18292 2013-05-06 Graham Stott <grahams@btinternet.com>
18294 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18295 codes which allow non-lvalues.
18297 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18299 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18300 components are all 1s.
18301 (integer_minus_onep): New function.
18302 * tree.h (integer_minus_onep): Declare it.
18303 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18304 integer_minus_onep instead of integer_all_onesp.
18306 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18309 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18310 variations of these patterns.
18312 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18314 * config/i386/i386.md (isa): Add x64_sse4 member.
18315 (enabled): Handle x64_sse4.
18316 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18317 instruction for 64bit SSE4_1 targets. Update insn attributes.
18318 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18319 instruction for SSE4_1 targets. Update insn attributes.
18320 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18321 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18322 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18324 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18325 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18326 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18328 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18331 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18333 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18335 * final.c (do_assembler_dialects): Don't handle curly braces and
18336 vertical bar escaped by % as dialect delimiters.
18337 (output_asm_insn): Print curly braces and vertical bar if escaped
18338 by % and ASSEMBLER_DIALECT defined.
18339 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18340 * doc/tm.texi: Regenerated.
18342 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18344 * config/mips/mips.c: Include tree-pass.h.
18345 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18346 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18347 (pass_mips_machine_reorg2): New machine specific pass.
18348 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18349 (mips_option_override): Register the new pass.
18350 * rtl.h (cleanup_barriers): Remove prototype.
18351 (dbr_schedule): Likewise.
18352 * jump.c (cleanup_barriers): Make static.
18353 * reorg.c (dbr_schedule): Likewise.
18355 2013-05-06 Richard Biener <rguenther@suse.de>
18357 PR tree-optimization/57185
18358 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18359 SSA names properly.
18361 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18364 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18365 between "first_arg" and "insn", not "last" and "insn".
18367 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18369 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18370 (find_candidates_in_block): Re-enable slsr_process_phi.
18371 (create_phi_basis): Fix double counting of candidate adjustment.
18373 2013-05-06 Richard Biener <rguenther@suse.de>
18375 PR middle-end/57147
18376 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18377 the edge is also fallthru, preserve it and just clear the
18379 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18380 also complex, preserve that and just clear the fallthru flag.
18381 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18382 update virtual operands.
18384 2013-05-06 Alan Modra <amodra@gmail.com>
18386 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18387 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18388 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18389 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18390 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18391 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18393 2013-05-06 Alan Modra <amodra@gmail.com>
18395 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18396 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18397 (DEFAULT_ASM_ENDIAN): Define.
18398 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18399 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18400 Update -K PIC clause from sysv4.h.
18401 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18402 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18404 2013-05-06 Alan Modra <amodra@gmail.com>
18406 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18407 twice for little-endian.
18408 (ashrdi3_no_power, ashrdi3): Support little-endian.
18410 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18413 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18414 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18416 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18417 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18418 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18421 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18422 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18424 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18425 * config/pa/pa.opt: Make mbig-switch a no-op.
18426 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18427 (CASE_VECTOR_MODE): Always return SImode.
18428 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18429 for the !TARGET_BIG_SWITCH case.
18430 * config/pa/pa-linux.h: Likewise.
18431 * config/pa/pa-openbsd.h: Likewise.
18432 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18433 * config/pa/pa.md (short_jump): Remove define_insn.
18434 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18435 (casesi0): Remove define_insn.
18436 (type): Remove btable_branch.
18437 (pa_combine_type): Likewise.
18438 (in_nullified_branch_delay): Likewise.
18439 (in_call_delay): Likewise.
18440 (define_delay): Likewise.
18441 (define_insn_reservation "Z3"): Likewise.
18442 (define_insn_reservation "Z4"): Likewise.
18443 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18444 (pa_adjust_insn_length): Remove adjustment for btable branches.
18445 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18448 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18450 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18451 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18452 Add m->r,x alternatives.
18453 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18454 splitters using SWI48x mode iterator.
18455 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18456 TARGET_64BIT. Add m->x alternative.
18457 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18458 Add o->x alternative. Enable for TARGET_SSE.
18459 (sse_storeq): Remove expander.
18460 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18461 with memory input operand.
18462 (*vec_extractv2di_1 splitter): New.
18463 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18464 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18466 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18468 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18469 (INT_LOWPART): Delete.
18470 (extract_MB): Adjust.
18471 (extract_ME): Adjust.
18472 (print_operand): Adjust.
18474 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18476 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18477 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18478 (reg_or_logical_cint_operand, easy_fp_constant,
18479 logical_const_operand): Delete "CONST_DOUBLE" case.
18480 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18481 "HOST_BITS_PER_WIDE_INT == 64" test.
18482 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18483 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18484 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18485 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18487 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18488 CONST_DOUBLE DImode/VOIDmode case.
18489 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18490 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18491 CONST_DOUBLE VOIDmode case.
18492 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18493 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18494 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18495 Delete CONST_DOUBLE case.
18496 (splitters for mov FMOVE64 const_double): Delete
18497 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18498 "HOST_BITS_PER_WIDE_INT >= 64" test.
18499 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18501 (mov DI const_double): Delete.
18503 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18505 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18506 on op shows all bits zero in mode of a lowpart subreg, return zero.
18508 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18511 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18512 to save TFmode registers and DImode to save TImode registers for
18513 caller save operations.
18514 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18515 mark being partially clobbered since they only use the first
18518 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18519 and TDmode only use the upper 64-bits of each VSX register.
18521 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18523 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18524 (find_candidates_in_block): Disable slsr_process_phi.
18526 2013-05-03 Guozhi Wei <carrot@google.com>
18528 * coverage.c (coverage_obj_init): Move the construction of gcov
18530 (build_init_ctor): ... here.
18532 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18534 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18535 (slsr_cand_d): Redefine def_phi.
18536 (stride_status, phi_adjust_status, count_phis_status): New enums.
18537 (find_phi_def): New.
18538 (find_basis_for_base_expr): New.
18539 (find_basis_for_candidate): Handle hidden bases.
18540 (alloc_cand_and_find_basis): Handle phi candidates.
18541 (slsr_process_phi): New.
18542 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18543 (create_mul_imm_cand): Likewise.
18544 (create_add_ssa_cand): Exclude phi base candidates.
18545 (create_add_imm_cand): Likewise.
18546 (slsr_process_cast): Likewise.
18547 (slsr_process_copy): Likewise.
18548 (find_candidates_in_block): Handle phi candidates.
18549 (dump_candidate): Likewise.
18550 (unconditional_cands): Delete.
18551 (unconditional_cands_with_known_stride_p): Delete.
18552 (phi_dependent_cand_p): New.
18553 (cand_increment): Handle phi-dependent candidates.
18554 (replace_dependent): Delete.
18555 (replace_mult_candidate): New.
18556 (replace_unconditional_candidate): New.
18557 (incr_vec_index): Move to avoid forward reference.
18558 (create_add_on_incoming_edge): New.
18559 (create_phi_basis): New.
18560 (replace_dependents): Delete.
18561 (replace_conditional_candidate): New.
18562 (phi_add_costs): New.
18563 (replace_uncond_cands_and_profitable_phis): New.
18564 (record_increment): Handle phi adjustments.
18565 (record_phi_increments): New.
18566 (record_increments): Handle phi adjustments.
18567 (phi_incr_cost): New.
18568 (lowest_cost_path): Handle phis.
18569 (total_savings): Likewise.
18570 (analyze_increments): Likewise.
18571 (ncd_with_phi): New.
18572 (ncd_of_cand_and_phis): New.
18573 (nearest_common_dominator_for_cands): Handle phi increments.
18574 (all_phi_incrs_profitable): New.
18575 (replace_profitable_candidates): Handle phi-dependent candidates.
18576 (analyze_candidates_and_replace): Likewise.
18578 2013-05-03 Teresa Johnson <tejohnson@google.com>
18581 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18582 do not exceed REG_BR_PROB_BASE.
18584 2013-05-03 Jeff Law <law@redhat.com>
18586 PR tree-optimization/57144
18587 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18588 operand of the condition will bit into the new type when eliminating
18589 a cast feeding a condition.
18591 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18593 PR rtl-optimization/57130
18594 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18595 of COMPARE as in_code to the recursive call if needed.
18597 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18599 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18600 (enabled): Handle new members.
18601 * config/i386/sse.md (*vec_concatv2si): Merge from
18602 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18603 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18605 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18607 PR tree-optimization/57027
18608 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18609 for fnms opportunity, check we got the prerequisite kind
18610 of tree / gimple before using accessor functions.
18612 2013-05-03 Richard Biener <rguenther@suse.de>
18614 * double-int.h (lshift): New overload without precision
18615 and arith argument.
18616 (operator *=, operator +=, operator -=): Move ...
18617 * double-int.c (operator *=, operator +=, operator -=): ... here
18618 and implement more efficiently.
18619 (mul_double_with_sign): Remove.
18620 (lshift_double): Adjust to take unsinged shift argument, push
18621 dispatching code to callers.
18622 (mul_double_wide_with_sign): Add early out for callers that
18623 are not interested in high parts or overflow.
18624 (lshift): New function.
18625 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18626 dispatch code here.
18627 (lrotate, rrotate): Use logical shifts.
18628 * expr.c (get_inner_reference): Use lshift.
18629 * fixed-value.c (do_fixed_divide): Likewise.
18630 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18631 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18632 (indirect_refs_may_alias_p): Likewise.
18633 (stmt_kills_ref_p_1): Likewise.
18635 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18637 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18639 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18641 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18642 scalar form of FABD instruction.
18644 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18646 * lra-constraints.c (process_alt_operands): Add checking alt
18647 number to choose the best alternative.
18649 2013-05-02 Richard Biener <rguenther@suse.de>
18651 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18652 bitmap and its handling.
18653 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18655 2013-05-02 Richard Biener <rguenther@suse.de>
18657 PR middle-end/57140
18658 * tree-inline.c (copy_loops): Properly handle removed loops.
18659 (copy_cfg_body): Mark destination loops for fixup if source
18660 loops needed fixup.
18662 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18665 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18666 generating simple_return for naked functions.
18668 2013-05-02 Martin Jambor <mjambor@suse.cz>
18670 PR middle-end/56988
18671 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18672 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18674 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18675 ipa_agg_replacement_value structures.
18676 (known_aggs_to_agg_replacement_list): Likewise.
18677 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18678 (read_agg_replacement_chain): Likewise.
18679 (ipcp_transform_function): Also check that by_ref flags match.
18681 2013-05-02 Richard Biener <rguenther@suse.de>
18683 * graphds.h (struct graph): Add obstack member.
18684 * graphds.c (new_graph): Initialize obstack and allocate
18686 (add_edge): Allocate edge from the obstack.
18687 (free_graph): Free the obstack instead of all edges and vertices.
18689 2013-05-02 Teresa Johnson <tejohnson@google.com>
18691 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18693 * cfg.c (update_bb_profile_for_threading): Ditto.
18694 * tree-inline.c (copy_bb): Ditto.
18695 (copy_edges_for_bb): Ditto.
18696 (initialize_cfun): Ditto.
18697 (copy_cfg_body): Ditto.
18698 (expand_call_inline): Ditto.
18699 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18700 (estimate_node_size_and_time): Ditto.
18701 (inline_merge_summary): Ditto.
18702 * cgraphclones.c (cgraph_clone_edge): Ditto.
18703 (cgraph_clone_node): Ditto.
18704 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18705 (compute_trg_info): Ditto.
18707 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18709 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18710 S reg when fp attribute set.
18711 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18713 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18715 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18717 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18718 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18719 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18721 2013-05-02 Richard Biener <rguenther@suse.de>
18723 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18724 (struct instantiate_cache_entry): New type.
18725 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18726 (struct instantiate_cache_type): New type.
18727 (set_instantiated_value, get_instantiated_value): Remove.
18728 (get_instantiated_value_entry): New function.
18729 (instantiate_scev_name): Use the new cache and adjust.
18730 (instantiate_scev_poly): Adjust.
18731 (instantiate_scev_binary): Likewise.
18732 (instantiate_array_ref): Likewise.
18733 (instantiate_scev_convert): Likewise.
18734 (instantiate_scev_not): Likewise.
18735 (instantiate_scev_3): Likewise.
18736 (instantiate_scev_2): Likewise.
18737 (instantiate_scev_r): Likewise.
18738 (instantiate_scev): Likewise.
18739 (resolve_mixers): Likewise.
18741 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18744 * lra-constraints.c (best_small_class_operands_num): Remove.
18745 (process_alt_operands): Remove small_class_operands_num. Take
18746 small classes operands into losers and only if the operand is not
18747 matched. Modify debugging output.
18748 (curr_insn_transform): Remove best_small_class_operands_num.
18751 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18753 * config/aarch64/aarch64-builtins.c
18754 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18755 * config/aarch64/aarch64-simd-builtins.def
18756 (reduc_splus_): Add new modes.
18757 (reduc_uplus_): New.
18758 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18759 (reduc_uplus_v4sf): Likewise.
18760 (reduc_splus_v4sf): Likewise.
18761 (aarch64_addv<mode>): Likewise.
18762 (reduc_uplus_<mode>): Likewise.
18763 (reduc_splus_<mode>): Likewise.
18764 (aarch64_addvv2di): Likewise.
18765 (reduc_uplus_v2di): Likewise.
18766 (reduc_splus_v2di): Likewise.
18767 (aarch64_addvv2si): Likewise.
18768 (reduc_uplus_v2si): Likewise.
18769 (reduc_splus_v2si): Likewise.
18770 (reduc_<sur>plus_<mode>): New.
18771 (reduc_<sur>plus_v2di): Likewise.
18772 (reduc_<sur>plus_v2si): Likewise.
18773 (reduc_<sur>plus_v4sf): Likewise.
18774 (aarch64_addpv4sf): Likewise.
18775 * config/aarch64/arm_neon.h
18776 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18777 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18778 add UNSPEC_SADDV, UNSPEC_UADDV.
18780 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18782 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18784 * config/aarch64/arm_neon.h
18785 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18787 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18789 * config/aarch64/aarch64-builtins
18790 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18792 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18794 * config/aarch64/aarch64-simd-builtins.def
18795 (reduc_smax_): New.
18796 (reduc_smin_): Likewise.
18797 (reduc_umax_): Likewise.
18798 (reduc_umin_): Likewise.
18799 (reduc_smax_nan_): Likewise.
18800 (reduc_smin_nan_): Likewise.
18803 (smax): Update for V2SF, V4SF and V2DF modes.
18806 (smin_nan): Likewise.
18807 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18808 (<su><maxmin><mode>3): ...This, refactor.
18809 (s<maxmin><mode>3): New.
18810 (<maxmin_uns><mode>3): Likewise.
18811 (reduc_<maxmin_uns>_<mode>): Refactor.
18812 (reduc_<maxmin_uns>_v4sf): Likewise.
18813 (reduc_<maxmin_uns>_v2si): Likewise.
18814 (aarch64_<fmaxmin><mode>: Remove.
18815 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18817 (vmin<q>_f<32,64>): Likewise.
18818 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18820 (su): Add mappings for smax, smin, umax, umin.
18822 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18823 (FMAXMIN): Rename as...
18824 (FMAXMIN_UNS): ...This.
18826 (fmaxminv): Likewise.
18827 (fmaxmin): Likewise.
18829 (maxmin_uns_op): Likewise.
18831 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18833 * config/aarch64/arm_neon.h
18834 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18835 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18836 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18838 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18840 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18841 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18843 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18845 * config/aarch64/aarch64-simd.md
18846 (vcond<mode>_internal): Handle special cases for constant masks.
18847 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18848 (vcondu<mode><mode>): Likewise.
18849 (vcond<v_cmp_result><mode>): New.
18851 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18853 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18854 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18855 * config/aarch64/aarch64-simd-builtins.def
18856 (cmeq): Update to BUILTIN_VALLDI.
18861 * config/aarch64/arm_neon.h
18862 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18863 to builtins or C as appropriate.
18865 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18867 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18869 (cmhi): Rename to...
18871 * config/aarch64/aarch64-simd.md
18872 (simd_mode): Add SF.
18873 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18874 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18875 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18876 (cstore<mode>_neg): ...This.
18877 * config/aarch64/iterators.md
18879 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18880 (COMPARISONS): New.
18881 (UCOMPARISONS): Likewise.
18882 (optab): Add missing comparisons.
18888 (VCMP_S): Likewise.
18889 (VCMP_U): Likewise.
18890 (V_cmp_result): Add DF, SF modes.
18891 (v_cmp_result): Likewise.
18893 (vmtype): Likewise.
18894 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18896 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18898 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18899 define_insn to define_insn_and_split.
18900 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18901 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18902 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18903 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18904 (thumb2_negscc): Likewise.
18906 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18908 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18910 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18912 * config/arm/thumb2.md: Remove trailing whitespaces.
18914 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18916 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18917 Use gen_int_mode rather than GEN_INT.
18919 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18921 * value-prof.c (stream_in_histogram_value): Remove the strayed
18924 2013-04-30 Richard Biener <rguenther@suse.de>
18926 PR middle-end/57122
18927 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18929 2013-04-30 Richard Biener <rguenther@suse.de>
18931 PR middle-end/57107
18932 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18934 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18936 PR rtl-optimization/56957
18937 PR rtl-optimization/57105
18938 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18939 variable. Use just INSN_UID for determining whether an insn
18940 should be only disconnected from the insn stream.
18941 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18943 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18945 PR tree-optimization/57104
18946 * tsan.c (instrument_expr): Don't instrument accesses to
18947 DECL_HARD_REGISTER VAR_DECLs.
18949 2013-04-30 Richard Biener <rguenther@suse.de>
18951 * function.h (loops_for_fn): New inline function.
18952 (set_loops_for_fn): Likewise.
18953 * cfgloop.h (place_new_loop): Add struct function parameter.
18954 (get_loop): Likewise.
18955 (get_loops): Likewise.
18956 (number_of_loops): Likewise.
18957 (fel_next): Adjust.
18958 (fel_init): Likewise.
18959 * cfg.c (get_loop_copy): Adjust.
18960 * cfgloop.c (flow_loops_dump): Likewise.
18961 (record_loop_exits): Likewise.
18962 (verify_loop_structure): Likewise.
18963 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18964 (estimate_reg_pressure_cost): Likewise.
18965 (mark_loop_exit_edges): Likewise.
18966 * cfgloopmanip.c (place_new_loop): Likewise.
18967 (add_loop): Likewise.
18968 (duplicate_loop): Likewise.
18969 * graph.c (draw_cfg_nodes): Likewise.
18970 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18971 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18972 (extract_affine_chrec): Likewise.
18973 (build_scop_iteration_domain): Likewise.
18974 * graphite.c (graphite_initialize): Likewise.
18975 * ira-build.c (create_loop_tree_nodes): Likewise.
18976 (more_one_region_p): Likewise.
18977 (rebuild_regno_allocno_maps): Likewise.
18978 (mark_loops_for_removal): Likewise.
18979 (mark_all_loops_for_removal): Likewise.
18980 (remove_unnecessary_regions): Likewise.
18981 (ira_build): Likewise.
18982 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
18983 * loop-init.c (fix_loop_structure): Likewise.
18984 (gate_rtl_move_loop_invariants): Likewise.
18985 (gate_rtl_unswitch): Likewise.
18986 (gate_rtl_unroll_and_peel_loops): Likewise.
18987 (rtl_doloop): Likewise.
18988 * lto-streamer-in.c (input_cfg): Likewise.
18989 * lto-streamer-out.c (output_cfg): Likewise.
18990 * modulo-sched.c (sms_schedule): Likewise.
18991 * predict.c (tree_estimate_probability): Likewise.
18992 (tree_estimate_probability_driver): Likewise.
18993 (estimate_loops): Likewise.
18994 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
18995 (move_sese_region_to_fn): Likewise.
18996 (debug_loop_num): Likewise.
18997 * tree-chrec.c (chrec_evaluate): Likewise.
18998 (hide_evolution_in_other_loops_than_loop): Likewise.
18999 (chrec_component_in_loop_num): Likewise.
19000 (reset_evolution_in_loop): Likewise.
19001 (evolution_function_is_invariant_rec_p): Likewise.
19002 * tree-if-conv.c (main_tree_if_conversion): Likewise.
19003 * tree-inline.c (copy_loops): Likewise.
19004 (copy_cfg_body): Likewise.
19005 (tree_function_versioning): Likewise.
19006 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
19007 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
19009 (add_to_evolution_1): Likewise.
19010 (scev_const_prop): Likewise.
19011 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
19012 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
19013 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
19014 (tree_ssa_lim_initialize): Likewise.
19015 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
19016 (verify_loop_closed_ssa): Likewise.
19017 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
19018 (tree_ssa_loop_im): Likewise.
19019 (tree_ssa_loop_unswitch): Likewise.
19020 (tree_vectorize): Likewise.
19021 (check_data_deps): Likewise.
19022 (tree_ssa_loop_ivcanon): Likewise.
19023 (tree_ssa_loop_bounds): Likewise.
19024 (tree_complete_unroll): Likewise.
19025 (tree_complete_unroll_inner): Likewise.
19026 (tree_parallelize_loops): Likewise.
19027 (tree_ssa_loop_prefetch): Likewise.
19028 (tree_ssa_loop_ivopts): Likewise.
19029 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
19030 * tree-vectorizer.c (vectorize_loops): Likewise.
19032 2013-04-29 Mike Frysinger <vapier@gentoo.org>
19034 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
19035 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
19036 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
19037 with EABI_LINK_SPEC.
19039 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19042 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
19045 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
19048 * lra-constraints.c (process_alt_operands): Discourage a bit more
19049 using memory for pseudos. Print cost dump for alternatives.
19050 Modify cost values for conflicts with early clobbers.
19051 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
19053 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19056 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
19058 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19060 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
19061 from/to S register.
19062 (movdi_aarch64): Support LDR/STR from/to D register.
19064 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19066 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
19067 or higher by default.
19069 2013-04-29 Richard Biener <rguenther@suse.de>
19071 PR middle-end/57075
19072 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
19073 even if not adding abnormal edges for calls that can make
19076 2013-04-29 Richard Biener <rguenther@suse.de>
19078 PR middle-end/57103
19079 * tree-cfg.c (move_stmt_op): Fix condition under which to update
19081 (move_stmt_r): Remove redundant checking.
19083 2013-04-29 Teresa Johnson <tejohnson@google.com>
19086 * basic-block.h (apply_scale): New function.
19087 (apply_probability): Use apply_scale.
19088 * gimple-streamer-in.c (input_bb): Ditto.
19089 * lto-streamer-in.c (input_cfg): Ditto.
19090 * lto-cgraph.c (merge_profile_summaries): Ditto.
19091 * tree-optimize.c (execute_fixup_cfg): Ditto.
19092 * tree-inline.c (copy_bb): Update comment to use apply_scale.
19093 (copy_edges_for_bb): Ditto.
19094 (copy_cfg_body): Ditto.
19096 2013-04-29 Tom de Vries <tom@codesourcery.com>
19098 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
19099 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
19100 (tail_merge_optimize): Handle current_loops == NULL.
19102 2013-04-26 Jeff Law <law@redhat.com>
19104 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
19105 (simplify_cond_using_ranges): Generalize code to simplify
19106 COND_EXPRs where one argument is a constant and the other
19107 is an SSA_NAME created by an integral type conversion.
19109 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19111 * config/arm/arm.md (store_minmaxsi): Use only when
19112 optimize_insn_for_size_p.
19114 2013-04-29 Christian Bruel <christian.bruel@st.com>
19117 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19119 2013-04-29 Richard Biener <rguenther@suse.de>
19121 PR middle-end/57089
19122 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19123 loop tree make sure to schedule a fixup for the child as well.
19124 (expand_omp_for_generic): Properly add loops.
19125 (expand_omp_for_static_nochunk): Likewise.
19126 (expand_omp_for_static_chunk): Likewise.
19127 (expand_omp_for): For the degenerate case fixup loops.
19128 (expand_omp_sections): Fix default bb placement in loops.
19129 (expand_omp_atomic_pipeline): Properly add loops.
19131 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19133 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19135 2013-04-29 Tom de Vries <tom@codesourcery.com>
19137 * tree-ssa-tail-merge.c: Update header comment.
19139 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19141 * config/aarch64/arm_neon.h
19142 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19143 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19144 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19145 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19146 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19147 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19148 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19149 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19151 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19153 * config/aarch64/aarch64-simd.md
19154 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19155 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19156 fix_trunc, fixuns_trunc.
19157 (ftrunc<VDQF:mode>2): New.
19158 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19159 (fix_trunc_optab): New.
19161 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19163 * config/aarch64/aarch64-builtins.c
19164 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19165 iceilf, lround, iroundf.
19167 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19170 * config/i386/i386.h (enum ix86_tune_indices)
19171 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19172 New, split from X86_TUNE_INTER_UNIT_MOVES.
19173 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19174 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19175 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19176 (TARGET_INTER_UNIT_MOVES): Remove.
19177 * config/i386/i386.c (initial_ix86_tune_features): Update.
19178 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19179 (ix86_expand_convert_uns_didf_sse): Use
19180 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19181 (ix86_expand_vector_init_one_nonzero): Ditto.
19182 (ix86_expand_vector_init_interleave): Ditto.
19183 (inline_secondary_memory_needed): Return true for moves from SSE class
19184 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19185 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19186 * config/i386/constraints.md (Yi, Ym): Depend on
19187 TARGET_INTER_UNIT_MOVES_TO_VEC.
19188 (Yj, Yn): New constraints.
19189 * config/i386/i386.md (*movdi_internal): Change constraints of
19190 operand 1 from Yi to Yj and from Ym to Yn.
19191 (*movsi_internal): Ditto.
19192 (*movdf_internal): Ditto.
19193 (*movsf_internal): Ditto.
19194 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19195 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19196 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19197 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19198 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19199 * config/i386/sse.md (movdi_to_sse): Ditto.
19200 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19201 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19202 TARGET_INTER_UNIT_MOVES.
19203 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19204 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19205 instead of TARGET_INTER_UNIT_MOVES.
19206 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19207 operand 1 from Yi to Yj and from Ym to Yn.
19209 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19211 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19212 (float_truncate_hi_): Likewise.
19213 (float_extend_lo_): Likewise.
19214 (float_truncate_lo_): Likewise.
19215 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19216 (aarch64_float_extend_lo_v2df): Likewise.
19217 (vec_unpacks_hi_v4sf): Likewise.
19218 (aarch64_float_truncate_lo_v2sf): Likewise.
19219 (aarch64_float_truncate_hi_v4sf): Likewise.
19220 (vec_pack_trunc_v2df): Likewise.
19221 (vec_pack_trunc_df): Likewise.
19223 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19225 * config/aarch64/aarch64-builtins.c
19226 (aarch64_fold_builtin): Fold float conversions.
19227 * config/aarch64/aarch64-simd-builtins.def
19228 (floatv2si, floatv4si, floatv2di): New.
19229 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19230 * config/aarch64/aarch64-simd.md
19231 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19232 * config/aarch64/iterators.md (FLOATUORS): New.
19233 (optab): Add float, floatuns.
19234 (su_optab): Likewise.
19236 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19238 * config/aarch64/aarch64-builtins.c
19239 (aarch64_builtin_vectorized_function): Use new names for
19241 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19242 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19243 (fcvtzu): Split as...
19244 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19245 (fcvtas): Split as...
19246 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19247 (fcvtau): Split as...
19248 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19249 (fcvtps): Split as...
19250 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19251 (fcvtpu): Split as...
19252 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19253 (fcvtms): Split as...
19254 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19255 (fcvtmu): Split as...
19256 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19257 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19258 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19259 (lfrintnusf, lfrintnudf): Likewise.
19260 * config/aarch64/aarch64-simd.md
19261 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19263 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19264 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19265 (fcvt_pattern): Likewise.
19267 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19269 * config/aarch64/aarch64-simd.md
19270 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19271 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19273 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19275 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19276 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19277 (vrnd<a,m,n,p>_f32): Implement using builtins.
19278 (vrnd<i,x><q>_f<32, 64>): New.
19280 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19282 * config/aarch64/aarch64-builtins.c
19283 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19284 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19285 (frintz): Rename to...
19287 (frintp): Rename to...
19289 (frintm): Rename to...
19291 (frinti): Rename to...
19292 (nearbyint): ...this.
19293 (frintx): Rename to...
19295 (frinta): Rename to...
19297 * config/aarch64/aarch64-simd.md
19298 (aarch64_frint<frint_suffix><mode>): Delete.
19299 (<frint_pattern><mode>2): Convert to insn.
19300 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19301 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19302 (frint_pattern): Likewise.
19303 (frint_suffix): Likewise.
19305 2013-04-29 Richard Biener <rguenther@suse.de>
19307 PR tree-optimization/57081
19308 * loop-init.c: Include tree-flow.h.
19309 (loop_optimizer_finalize): Free number of iteration estimates.
19310 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19312 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19314 PR tree-optimization/57083
19315 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19316 non-singleton shift count range, zero extend low_bound for uns case.
19318 * config/i386/predicates.md (general_vector_operand): New predicate.
19319 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19320 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19321 if they aren't nonimmediate operands. If their original values
19322 satisfy const_vector_equal_evenodd_p, don't shift them.
19323 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19324 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19325 if not nonimmediate_operand.
19326 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19327 instead of register_operand.
19328 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19330 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19332 * stor-layout.c (finalize_size_functions): Allocate a structure and
19333 reset cfun before dumping the functions.
19335 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19337 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19340 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19341 use xop_pmacsdqh if uns_p.
19342 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19343 the immediate rotate count.
19345 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19347 * rtl.h (struct rtx_def): Add comment for field jump.
19348 (LRA_SUBREG_P): New macro.
19349 * recog.c (register_operand): Check LRA_SUBREG_P.
19350 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19352 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19353 (lra_final_code_change): Skip subreg change for operators.
19354 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19355 if there are no operand changes.
19356 * lra-constraints.c (curr_insn_set): New.
19357 (match_reload): Set LRA_SUBREG_P.
19358 (emit_spill_move): Ditto.
19359 (check_and_process_move): Use curr_insn_set. Process only single
19360 set insns. Don't initialize sec_mem_p and change_p.
19361 (simplify_operand_subreg): Use LRA_SUBREG_P.
19362 (reg_in_class_p): New function.
19363 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19364 of #ifdef. Add code to remove cycling.
19365 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19366 non-null disp. Reload inner instead of disp when base and index
19367 are null. Try to put lo_sum into register.
19368 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19369 (check_and_process_move): Move code for move cost check to
19370 simple_move_p. Remove equiv_substitution.
19371 (simple_move_p): New function.
19372 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19373 curr_insn_set. Call check_and_process_move only for single set
19374 insns. Use the new function. Move call of check_and_process_move
19375 after operand equiv substitution and address process.
19377 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19380 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19381 with nonlocal goto receivers or returns twice calls, ignore
19382 unininitialized values from abnormal edges to nl goto receiver
19383 or returns twice call.
19385 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19387 PR tree-optimization/57051
19388 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19389 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19392 2013-04-26 Richard Biener <rguenther@suse.de>
19394 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19395 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19396 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19397 (fixup_loop_arrays_after_move): New function.
19398 (move_sese_region_to_fn): Properly outline the loop tree parts
19399 of the SESE region.
19401 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19403 * config/i386/i386.md (type, unit): Fix long lines.
19405 2013-04-26 Richard Biener <rguenther@suse.de>
19407 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19408 (lto-streamer-out.o): Likewise.
19409 * cfgloop.c (init_loops_structure): Export, add struct function
19410 argument and adjust.
19411 (flow_loops_find): Adjust.
19412 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19413 (init_loops_structure): Declare.
19414 * lto-streamer-in.c: Include cfgloop.h.
19415 (input_cfg): Input the loop tree.
19416 * lto-streamer-out.c: Include cfgloop.h.
19417 (output_cfg): Output the loop tree.
19418 (output_struct_function_base): Do not drop PROP_loops.
19420 2013-03-26 Richard Biener <rguenther@suse.de>
19422 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19423 (pass_build_cfg): Provide PROP_loops.
19424 (move_sese_region_to_fn): Remove loops that are outlined into fn
19426 * tree-inline.c: Include cfgloop.h.
19427 (initialize_cfun): Do not drop PROP_loops.
19428 (copy_loops): New function.
19429 (copy_cfg_body): Copy loop structure.
19430 (tree_function_versioning): Initialize destination loop tree.
19431 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19432 (pass_parallelize_loops): Do IL verification.
19433 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19434 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19435 the CFG make sure we fixup loops as well.
19436 * tree-ssa-tail-merge.c: Include cfgloop.h.
19437 (replace_block_by): When merging loop latches mark loops for fixup.
19438 * lto-streamer-out.c (output_struct_function_base): Drop
19439 PROP_loops for now.
19440 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19441 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19442 * ipa-split.c: Include cfgloop.h.
19443 (split_function): Add the new return block to the loop tree root.
19444 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19445 whether we have removed the forwarder block.
19446 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19447 * cfgloop.h (place_new_loop): Declare.
19448 * cfgloopmanip.c (place_new_loop): Export.
19449 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19450 (tree-switch-conversion.o): Likewise.
19451 (tree-complex.o): Likewise.
19452 (tree-inline.o): Likewise.
19453 (tree-ssa-tailmerge.o): Likewise.
19454 (ipa-split.o): Likewise.
19455 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19456 (tree-ssa-copy.o): Likewise.
19457 * tree-switch-conversion.c: Include cfgloop.h
19458 (process_switch): If we emit a bit-test cascade, schedule loops
19460 * tree-complex.c: Include cfgloop.h.
19461 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19462 * asan.c: Include cfgloop.h.
19463 (create_cond_insert_point): Properly add new basic-blocks to
19464 loops, schedule loop fixup.
19465 * cfgloop.c (verify_loop_structure): Check that looks are not
19467 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19469 (expand_omp_for_generic): Likewise.
19470 (expand_omp_sections): Likewise.
19471 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19472 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19473 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19474 is initialized, not when loops are present.
19475 * tree-parloops.c (parallelize_loops): Remove checking here.
19476 * passes.c (init_optimization_passes): Schedule a copy-propagation
19477 pass before complete unrolling of inner loops.
19479 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19481 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19482 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19483 (colorize_init): Add argument to _WIN32 version.
19484 * toplev.c: Include diagnostic-color.h.
19485 (process_options): Default to -fdiagnostics-color=auto if
19486 GCC_COLORS env var is in the environment.
19487 * common.opt (fdiagnostics-color=): Add Var and Init.
19488 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19489 env var is in the environment, the default is auto rather than never.
19491 * diagnostic.h (file_name_as_prefix): Add context argument.
19492 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19493 the string as locus.
19494 * langhooks.c (lhd_print_error_function): Adjust caller.
19496 2013-04-25 Lawrence Crowl <crowl@google.com>
19498 * var-tracking.c (shared_hash_def::htab):
19499 Change type to hash_table. Update dependent calls and types.
19501 2013-04-25 Lawrence Crowl <crowl@google.com>
19503 * Makefile.in: Update as needed below.
19505 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19506 Move declaration to after the type's method definitons.
19508 * attribs.c (htab_t scoped_attributes::attribute_hash):
19509 Change type to hash_table. Update dependent calls and types.
19511 * bitmap.c (htab_t bitmap_desc_hash):
19512 Change type to hash_table. Update dependent calls and types.
19514 * cselib.c (htab_t cselib_hash_table):
19515 Change type to hash_table. Update dependent calls and types.
19517 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19518 (hash_string_slot_node): Move implementation into lto-streamer.h
19519 struct string_slot_hasher.
19520 (eq_string_slot_node): Likewise.
19522 * data-streamer-out.c: Update output_block::string_hash_table
19523 dependent calls and types.
19525 * dwarf2cfi.c (htab_t trace_index):
19526 Change type to hash_table. Update dependent calls and types.
19528 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19529 Change type to hash_table. Update dependent calls and types.
19530 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19531 (htab_t optimize_external_refs::map): Likewise.
19532 (htab_t output_comp_unit::extern_map): Likewise.
19533 (htab_t output_comdat_type_unit::extern_map): Likewise.
19534 (htab_t output_macinfo::macinfo_htab): Likewise.
19535 (htab_t optimize_location_lists::htab): Likewise.
19536 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19538 * except.c (htab_t ehspec_hash_type):
19539 Change type to hash_table. Update dependent calls and types.
19540 (assign_filter_values::ttypes): Likewise.
19541 (assign_filter_values::ehspec): Likewise.
19542 (sjlj_assign_call_site_values::ar_hash): Likewise.
19543 (convert_to_eh_region_ranges::ar_hash): Likewise.
19545 * gcse.c (htab_t pre_ldst_table):
19546 Change type to hash_table. Update dependent calls and types.
19548 * ggc-common.c (htab_t saving_htab):
19549 Change type to hash_table. Update dependent calls and types.
19550 (htab_t loc_hash): Likewise.
19551 (htab_t ptr_hash): Likewise.
19552 (call_count): Rename ggc_call_count.
19553 (call_alloc): Rename ggc_call_alloc.
19554 (loc_descriptor): Rename make_loc_descriptor.
19555 (add_statistics): Rename ggc_add_statistics.
19557 * ggc-common.c (saving_htab):
19558 Change type to hash_table. Update dependent calls and types.
19560 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19561 (push_gimplify_context): Likewise.
19562 (pop_gimplify_context): Likewise.
19563 (struct gimple_temp_hash_elt): Added.
19564 (struct gimplify_hasher): Likewise.
19565 (struct gimplify_ctx.temp_htab):
19566 Change type to hash_table. Update dependent calls and types.
19568 * gimple-fold.c: Include gimplify-ctx.h.
19570 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19571 Change type to hash_table. Update dependent calls and types.
19572 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19573 avoid potential global name collision.
19575 * gimplify.c: Include gimplify-ctx.h.
19576 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19577 (htab_t gimplify_ctx::temp_htab):
19578 Update dependent calls and types for new type hash_table.
19579 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19580 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19582 * gimplify-ctx.h: New.
19583 (struct gimple_temp_hash_elt): Move from gimplify.c.
19584 (class gimplify_hasher): New.
19585 (struct gimplify_ctx): Move from gimple.h.
19586 (htab_t gimplify_ctx::temp_htab):
19587 Change type to hash_table. Update dependent calls and types.
19589 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19590 (htab_t ivs_params::newivs_index):
19591 Change type to hash_table. Update dependent calls and types.
19592 (htab_t ivs_params::params_index): Likewise.
19593 (htab_t print_generated_program::params_index): Likewise.
19594 (htab_t gloog::newivs_index): Likewise.
19595 (htab_t gloog::params_index): Likewise.
19597 * graphite.c: Include graphite-htab.h.
19598 4htab_t graphite_transform_loops::bb_pbb_mapping):
19599 Change type to hash_table. Update dependent calls and types.
19601 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19602 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19603 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19605 * graphite-dependences.c: Include graphite-htab.h.
19606 (loop_is_parallel_p): Change hash table type of parameter.
19608 * graphite-htab.h: New.
19609 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19610 (extern find_pbb_via_hash): Move from graphite-poly.h.
19611 (extern loop_is_parallel_p): Move from graphite-poly.h.
19612 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19614 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19615 (extern loop_is_parallel_p): Move to graphite-htab.h.
19616 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19618 * haifa-sched.c (htab_t delay_htab):
19619 Change type to hash_table. Update dependent calls and types.
19620 (htab_t delay_htab_i2): Likewise.
19622 * ira-color.c (htab_t allocno_hard_regs_htab):
19623 Change type to hash_table. Update dependent calls and types.
19625 * ira-costs.c (htab_t cost_classes_htab):
19626 Change type to hash_table. Update dependent calls and types.
19628 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19629 Change type to hash_table. Update dependent calls and types.
19631 * loop-iv.c (htab_t bivs):
19632 Change type to hash_table. Update dependent calls and types.
19634 * loop-unroll.c (htab_t opt_info::insns_to_split):
19635 Change type to hash_table. Update dependent calls and types.
19636 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19638 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19639 (struct string_slot_hasher): New.
19640 (htab_t output_block::string_hash_table):
19641 Change type to hash_table. Update dependent calls and types.
19643 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19644 (htab_t file_name_hash_table):
19645 Change type to hash_table. Update dependent calls and types.
19647 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19650 * lto-streamer.c (htab_t tree_htab):
19651 Change type to hash_table. Update dependent calls and types.
19653 * omp-low.c: Include gimplify-ctx.h.
19655 * passes.c (htab_t name_to_pass_map):
19656 Change type to hash_table. Update dependent calls and types.
19657 (pass_traverse): Rename to passes_pass_traverse.
19659 * plugin.c (htab_t event_tab):
19660 Change type to hash_table. Update dependent calls and types.
19662 * postreload-gcse.c (htab_t expr_table):
19663 Change type to hash_table. Update dependent calls and types.
19664 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19666 * sese.c (debug_rename_map_1): Make extern.
19667 (htab_t copy_bb_and_scalar_dependences::rename_map):
19668 Change type to hash_table. Update dependent calls and types.
19670 * sese.h (extern debug_rename_map): Move to .c file.
19672 * store-motion.c (htab_t store_motion_mems_table):
19673 Change type to hash_table. Update dependent calls and types.
19675 * trans-mem.c (htab_t tm_new_mem_hash):
19676 Change type to hash_table. Update dependent calls and types.
19678 * tree-browser.c (htab_t TB_up_ht):
19679 Change type to hash_table. Update dependent calls and types.
19681 * tree-cfg.c (htab_t discriminator_per_locus):
19682 Change type to hash_table. Update dependent calls and types.
19684 * tree-complex.c: Include tree-hasher.h
19685 (htab_t complex_variable_components):
19686 Change type to hash_table. Update dependent calls and types.
19688 * tree-eh.c (htab_t finally_tree):
19689 Change type to hash_table. Update dependent calls and types.
19691 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19692 struct int_tree_hasher.
19693 (extern int_tree_map_eq): Likewise.
19694 (uid_decl_map_hash): Removed.
19695 (extern decl_tree_map_eq): Likewise.
19697 * tree-hasher.h: New.
19698 (struct int_tree_hasher): New.
19699 (typedef int_tree_htab_type): New.
19701 * tree-inline.c: Include gimplify-ctx.h.
19703 * tree-mudflap.c: Include gimplify-ctx.h.
19705 * tree-parloops.c: Include tree-hasher.h.
19706 (htab_t eliminate_local_variables_stmt::decl_address):
19707 Change type to hash_table. Update dependent calls and types.
19708 (htab_t separate_decls_in_region::decl_copies): Likewise.
19710 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19711 Change type to hash_table. Update dependent calls and types.
19713 * tree-sra.c (candidates):
19714 Change type to hash_table. Update dependent calls and types.
19716 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19718 (int_tree_map_hash): Likewise.
19720 * tree-ssa-dom.c (htab_t avail_exprs):
19721 Change type to hash_table. Update dependent calls and types.
19723 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19724 Change type to hash_table. Update dependent calls and types.
19726 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19727 Change type to hash_table. Update dependent calls and types.
19729 * tree-ssa-phiopt.c (seen_ssa_names):
19730 Change type to hash_table. Update dependent calls and types.
19732 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19733 Change type to hash_table. Update dependent calls and types.
19735 * tree-ssa-uncprop.c (equiv):
19736 Change type to hash_table. Update dependent calls and types.
19738 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19740 PR rtl-optimization/57003
19741 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19742 call note_stores with kill_clobbered_value callback again after
19743 killing regs_invalidated_by_call.
19745 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19747 * config/aarch64/aarch64-simd.md
19748 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19749 (aarch64_simd_bsl<mode>): Likewise.
19750 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19752 2013-04-25 Marek Polacek <polacek@redhat.com>
19754 PR tree-optimization/57066
19755 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19757 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19759 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19761 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19763 * config/aarch64/aarch64-builtins.c
19764 (aarch64_fold_builtin): New.
19765 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19766 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19767 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19768 * config/aarch64/arm_neon.h
19769 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19771 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19772 Tejas Belagod <tejas.belagod@arm.com>
19774 * config/aarch64/aarch64-builtins.c
19775 (aarch64_gimple_fold_builtin): New.
19776 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19777 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19778 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19779 (addvv4sf): Update.
19780 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19782 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19784 * config/aarch64/aarch64.md
19785 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19787 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19789 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19790 (*ngcsi_uxtw): New pattern.
19792 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19793 Julian Brown <julian@codesourcery.com>
19795 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19796 (TB_DREG): Add T_V4HF.
19797 (v4hf_UP): New macro.
19798 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19799 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19800 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19802 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19803 (arm_vector_mode_supported_p): Handle V4HF.
19804 (arm_mangle_map): Handle V4HFmode.
19805 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19806 * config/arm/arm_neon_builtins.def: Add entries for
19807 vcvtv4hfv4sf, vcvtv4sfv4hf.
19808 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19809 (neon_vcvtv4hfv4sf): Likewise.
19810 * config/arm/neon-gen.ml: Handle half-precision floating point
19812 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19813 * config/arm/arm_neon.h: Regenerate.
19814 * config/arm/neon.ml (type elts): Add F16.
19815 (type vectype): Add T_float16x4, T_floatHF.
19816 (type vecmode): Add V4HF.
19817 (type features): Add Requires_FP_bit feature.
19818 (elt_width): Handle F16.
19819 (elt_class): Likewise.
19820 (elt_of_class_width): Likewise.
19821 (mode_of_elt): Refactor.
19822 (type_for_elt): Handle F16, fix error messages.
19823 (vectype_size): Handle T_float16x4.
19824 (vcvt_sh): New function.
19825 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19826 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19827 (string_of_mode): Handle V4HF.
19828 * doc/arm-neon-intrinsics.texi: Regenerate.
19830 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19832 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19833 format specifier in 'X' case.
19835 2013-04-25 Alan Modra <amodra@gmail.com>
19838 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19839 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19840 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19841 Repeat for many other rotate/shift and mask patterns using subregs.
19842 Name lshiftrt insns.
19843 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19844 on WORDS_BIG_ENDIAN.
19846 2013-04-25 Alan Modra <amodra@gmail.com>
19848 * config.gcc: Support little-endian powerpc-linux targets.
19849 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19850 (LINK_OS_LINUX_SPEC): Define.
19851 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19852 Preserve MASK_LITTLE_ENDIAN.
19853 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19854 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19855 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19856 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19857 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19858 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19859 Correct fp word order for little-endian. Don't shift toc entries
19860 smaller than a word for little-endian.
19861 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19862 (bswapdi2 splits): Correct low-part subreg for little-endian.
19863 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19864 low/high where such is correct only for be.
19865 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19866 little-endian for -mcall-aixdesc.
19868 2013-04-25 Alan Modra <amodra@gmail.com>
19870 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19871 replace_equiv_address_nv.
19873 2013-04-25 Alan Modra <amodra@gmail.com>
19875 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19877 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19880 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19881 * rtl.h (struct rtx_def): ...
19883 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19885 PR rtl-optimizations/57046
19886 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19887 for multi-reg splits.
19889 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19891 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19893 2013-04-24 Sterling Augustine <saugustine@google.com>
19895 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19896 (comp_dir_string, debug_str_dwo_section): New.
19897 (DEBUG_STR_DWO_SECTION): Rename to ...
19898 (DEBUG_DWO_STR_SECTION): ... this.
19899 (DEBUG_NORM_STR_SECTION): Delete.
19900 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19901 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19902 (find_AT_string): Move most logic to ...
19903 (find_AT_string_in_table): ... here. New.
19904 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19905 add_skeleton_AT_string. Delete logic.
19906 (output_skeleton_debug_sections): Remove call to
19907 add_top_level_skeleton_die_attrs.
19908 (add_comp_dir_attribute): Move logic to comp_dir_string.
19909 (dwarf2out_init): Initialize debug_str_dwo_section.
19910 (output_indirect_string): Call find_string_form.
19911 (output_indirect_strings): Rewrite.
19912 (prune_unused_types): Empty skeleton_debug_str_hash.
19913 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19914 (dwarf2out_finish): Call output_indirect_strings.
19916 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19918 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19920 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19922 * rtl.h (struct rtx_def): Add comment for field jump.
19923 (LRA_SUBREG_P): New macro.
19924 * recog.c (register_operand): Check LRA_SUBREG_P.
19925 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19927 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19928 (lra_final_code_change): Skip subreg change for operators.
19929 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19930 if there are no operand changes.
19931 * lra-constraints.c (curr_insn_set): New.
19932 (match_reload): Set LRA_SUBREG_P.
19933 (emit_spill_move): Ditto.
19934 (check_and_process_move): Use curr_insn_set. Process only single
19935 set insns. Don't initialize sec_mem_p and change_p.
19936 (simplify_operand_subreg): Use LRA_SUBREG_P.
19937 (reg_in_class_p): New function.
19938 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19939 of #ifdef. Add code to remove cycling.
19940 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19941 non-null disp. Reload inner instead of disp when base and index
19942 are null. Try to put lo_sum into register.
19943 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19944 (check_and_process_move): Move code for move cost check to
19945 simple_move_p. Remove equiv_substitution.
19946 (simple_move_p): New function.
19947 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19948 curr_insn_set. Call check_and_process_move only for single set
19949 insns. Use the new function. Move call of check_and_process_move
19950 after operand equiv substitution and address process.
19952 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19954 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19955 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19956 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19958 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19960 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19962 2013-04-24 Marek Polacek <polacek@redhat.com>
19964 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19965 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19966 (select_loops_exit_conditions): Likewise.
19967 (number_of_iterations_for_all_loops): Likewise.
19968 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19969 (scev_analysis): Likewise.
19971 2013-04-02 Catherine Moore <clm@codesourcery.com>
19972 Chao-ying Fu <fu@mips.com>
19974 * config/mips/micromips.md (jraddiusp): New pattern.
19975 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
19976 instruction if possible.
19978 2013-04-24 Alan Modra <amodra@gmail.com>
19980 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
19982 2013-04-24 Julian Brown <julian@codesourcery.com>
19983 Chung-Lin Tang <cltang@codesourcery.com>
19985 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
19986 dependency behavior in enumeration type DIE generation. Add TODO note
19987 to comments about future DW_FORM_sdata/udata re-work of related code.
19989 2013-04-23 Lawrence Crowl <crowl@google.com>
19991 * Makefile.in: Update as needed below.
19993 * hash-table.h (class hash_table):
19994 Correct many methods with parameter types compare_type to the correct
19995 value_type. (Correct code was unlikely to notice the change.)
19996 (hash_table::elements_with_deleted) New.
19997 (class hashtable::iterator): New.
19998 (hashtable::begin()): New.
19999 (hashtable::end()): New.
20000 (FOR_EACH_HASH_TABLE_ELEMENT): New.
20002 * statistics.c (statistics_hashes):
20003 Change type to hash_table. Update dependent calls and types.
20005 * tree-into-ssa.c (var_infos):
20006 Change type to hash_table. Update dependent calls and types.
20008 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
20009 Change type to hash_table. Update dependent calls and types.
20011 * tree-ssa-loop-im.c (struct mem_ref.refs):
20012 Change type to hash_table. Update dependent calls and types.
20014 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
20015 Change type to hash_table. Update dependent calls and types.
20017 * tree-ssa-sccvn.c (vn_tables_s::nary):
20018 Change type to hash_table. Update dependent calls and types.
20019 (vn_tables_s::phis): Likewise.
20020 (vn_tables_s::references): Likewise.
20022 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
20023 (vn_reference_eq): Update parameter and return types.
20025 * tree-ssa-structalias.c (pointer_equiv_class_table):
20026 Change type to hash_table. Update dependent calls and types.
20027 (location_equiv_class_table): Likewise.
20029 * tree-vect-data-refs.c: Consequential changes for making
20030 peeling a hash_table.
20032 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
20033 (destroy_loop_vec_info): Dependent hash_table update.
20035 * tree-vectorizer.h (peeling_htab):
20036 Change type to hash_table. Update dependent calls and types.
20038 2013-04-23 Shiva Chen <shiva0217@gmail.com>
20040 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
20041 to check the register content is equal or not.
20042 * lra-constraints.c (match_reload): Use lra_assign_reg_val
20043 to assign register content record.
20044 * lra-eliminations.c (update_reg_eliminate): Use
20045 lra_update_reg_val_offset to update register content offset.
20046 * lra-int.h (struct lra_reg): Add offset member.
20047 (lra_reg_val_equal_p): New static inline function.
20048 (lra_update_reg_val_offset): New static inline function.
20049 (lra_assign_reg_val): New static inline function.
20050 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
20051 to assign register content record.
20052 (initialize_lra_reg_info_element): Initial offset to zero.
20054 2013-04-23 Catherine Moore <clm@codesourcery.com>
20056 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
20057 operands. Record compression.
20059 2013-04-23 Xinliang David Li <davidxl@google.com>
20061 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
20063 2013-04-23 Richard Biener <rguenther@suse.de>
20065 PR middle-end/57036
20066 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
20067 parameter, only add abnormal goto edges from the copied body
20068 if the call could perform abnormal gotos.
20069 (copy_cfg_body): Adjust.
20071 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
20073 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
20075 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
20077 * coretypes.h (gimple_stmt_iterator): Add struct to make
20080 2013-04-23 Richard Biener <rguenther@suse.de>
20082 PR tree-optimization/57026
20083 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
20084 from SSA names occuring in abnormal PHI nodes.
20086 2013-04-22 Andi Kleen <ak@linux.intel.com>
20088 * lto/lto.c (print_lto_report_1): Fix LTO report names.
20090 2013-04-22 Andi Kleen <ak@linux.intel.com>
20092 * lto/lto.c (print_lto_report_1): Declare early.
20093 (read_cgraph_and_symbols): Call print_lto_report_1 early.
20095 2013-04-22 Andi Kleen <ak@linux.intel.com>
20097 * common.opt (-flto-report-wpa): Add.
20098 * doc/invoke.texi (-flto-report-wpa): Add.
20099 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
20102 2013-04-22 Xinliang David Li <davidxl@google.com>
20104 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
20105 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
20106 * Makefile.in: New dependency
20108 David Daney <ddaney.cavm@gmail.com>
20110 * configure.ac (gcc_cv_as_micromips_support): Use the
20111 --fatal-warnings option.
20112 * configure: Regenerate.
20114 2013-04-22 Marek Polacek <polacek@redhat.com>
20117 * tsan.c (instrument_expr): Don't instrument expression
20118 in case its size is zero.
20120 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20124 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20126 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20128 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20130 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20131 (gimple_stmt_iterator): New typedef.
20132 * gimple.h (gimple_stmt_iterator): Rename to...
20133 (gimple_stmt_iterator_d): ... This.
20134 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20135 trees be valid for GIMPLE and GENERIC.
20136 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20137 * gimple-fold.c (gimple_fold_call): Call target hook
20138 gimple_fold_builtin.
20139 * hooks.c (hook_bool_gsiptr_false): New.
20140 * hooks.h (hook_bool_gsiptr_false): New.
20141 * target.def (fold_stmt): New.
20142 * doc/tm.texi: Regenerate.
20144 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20147 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20148 a set sp if no stack realignment.
20150 2013-04-22 Nick Clifton <nickc@redhat.com>
20152 * config.gcc (tilegx-linux): Extend extra_objs rather than
20154 (tilepro-linux): Likewise.
20156 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20158 * config/aarch64/aarch64-builtins.c
20160 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20161 (VAR<1-12>): Add MAP parameter.
20162 (BUILTIN_*): Likewise.
20163 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20164 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20165 (aarch64_ushl_n<mode>): Likewise.
20166 (aarch64_sshr_n<mode>): Likewise.
20167 (aarch64_ushr_n<mode>): Likewise.
20168 (aarch64_<maxmin><mode>): Likewise.
20169 (aarch64_sqrt<mode>): Likewise.
20170 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20171 (vshr<q>_n_*): Likewise.
20173 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20175 * config/aarch64/aarch64-builtins.c
20176 (aarch64_simd_builtin_type_mode): Handle SF types.
20178 (BUILTIN_GPF): Define.
20179 (aarch64_init_simd_builtins): Handle SF types.
20180 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20181 (frecps): Likewise.
20182 (frecpx): Likewise.
20183 * config/aarch64/aarch64-simd.md
20184 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20185 (aarch64_frecpe<mode>): New.
20186 (aarch64_frecps<mode>): Likewise.
20187 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20188 (v8type): Add frecp<esx>.
20189 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20190 (aarch64_frecps<mode>): Likewise.
20191 * config/aarch64/iterators.md (FRECP): New.
20192 (frecp_suffix): Likewise.
20193 * config/aarch64/arm_neon.h
20194 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20196 2013-04-22 Christian Bruel <christian.bruel@st.com>
20199 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20200 (REG_CLASS_NAMES): Idem.
20201 (REG_CLASS_CONTENTS): Idem.
20202 (REGCLASS_HAS_FP_REG): Idem.
20203 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20204 (sh_conditional_register_usage): Idem.
20206 2013-04-21 Jeff Law <law@redhat.com>
20208 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20209 (ssa_forward_propagate_and_combine): Use it.
20211 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20213 * lra.c: Update the flow chart diagram.
20215 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20217 PR rtl-optimization/56847
20218 * lra-constraints.c (process_alt_operands): Discourage alternative
20219 with non-matche doffsettable memory constraint fro memory with
20222 2013-04-19 Richard Biener <rguenther@suse.de>
20224 PR tree-optimization/56982
20225 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20227 * gimplify.c (gimplify_call_expr): Notice special calls.
20228 (gimplify_modify_expr): Likewise.
20229 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20230 abnormal control flow receivers.
20231 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20232 in the same way as cfun->has_nonlocal_labels.
20233 (gimple_purge_dead_abnormal_call_edges): Likewise.
20234 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20235 receivers start a basic-block.
20237 2013-04-19 Richard Biener <rguenther@suse.de>
20239 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20241 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20242 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20243 (SLP_TREE_LOAD_PERMUTATION): Add.
20244 (vect_transform_slp_perm_load): Adjust prototype.
20245 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20246 (vect_free_slp_instance): Likewise.
20247 (vect_create_new_slp_node): Likewise.
20248 (vect_supported_slp_permutation_p): Remove.
20249 (vect_slp_rearrange_stmts): Adjust.
20250 (vect_supported_load_permutation_p): Likewise. Inline
20251 vect_supported_slp_permutation_p here.
20252 (vect_analyze_slp_instance): Compute load permutations per
20253 slp node instead of per instance.
20254 (vect_get_slp_defs): Adjust.
20255 (vect_transform_slp_perm_load): Likewise.
20256 (vect_schedule_slp_instance): Remove redundant code.
20257 (vect_schedule_slp): Remove hack for PR56270, add it ...
20258 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20259 CSE loads for SLP. Adjust.
20261 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20263 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20264 spelling in two comments.
20266 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20269 * config/arm/arm.c (load_multiple_sequence): Require SP
20270 as base register for loads if SP is in the register list.
20272 2013-04-19 Martin Jambor <mjambor@suse.cz>
20274 PR tree-optimization/56718
20275 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20276 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20277 and made public. Adjusted all callers.
20278 (ipa_intraprocedural_devirtualization): New function.
20279 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20280 (ipa_intraprocedural_devirtualization): Likewise.
20281 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20283 2013-04-19 Richard Biener <rguenther@suse.de>
20285 PR tree-optimization/57000
20286 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20288 2013-04-19 Terry Guo <terry.guo@arm.com>
20290 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20292 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20293 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20294 (cortex_m4_fmacs): Use new reservations.
20295 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20297 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20299 PR rtl-optimization/56999
20300 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20302 (lra_coalesce): Remove split_origin_bitmap and related code.
20303 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20304 ranges if necessary.
20306 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20308 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20310 (ix86_expand_call): Remove clobbered_registers array and use
20311 x86_64_ms_sysv_extra_clobbered_registers instead.
20312 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20314 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20316 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20317 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20318 (*call_value_rex64_ms_sysv): Ditto.
20320 2013-04-18 Cary Coutant <ccoutant@google.com>
20322 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20323 parent instead of die_mark.
20325 2013-04-18 Diego Novillo <dnovillo@google.com>
20327 * gimple.c (create_gimple_tmp): New.
20328 (get_expr_type): New.
20329 (build_assign): New.
20330 (build_type_cast): New.
20331 * gimple.h (enum ssa_mode): Define.
20332 (gimple_seq_set_location): New.
20333 * asan.c (build_check_stmt): Change some gimple_build_* calls
20334 to use build_assign and build_type_cast.
20336 2013-04-18 Richard Biener <rguenther@suse.de>
20338 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20339 handle negative step. Remove redundant checks.
20340 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20341 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20342 for negative step and grouped loads fail to vectorize.
20344 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20346 * emit-rtl.c (reset_insn_used_flags): New function.
20347 (reset_all_used_flags): Use it.
20348 (verify_insn_sharing): New function.
20349 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20351 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20353 PR tree-optimization/56984
20354 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20355 and (x >> M) >= N don't register any assertion if N << M is the
20358 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20360 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20361 do not delete_insn insns that have not yet been emitted, only
20362 unlink them with remove_insn.
20363 * df-scan.c (df_insn_delete): Revert r197492.
20365 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20367 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20368 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20370 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20372 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20373 define_insn_and_split.
20374 (and_scc,ior_scc,negscc): Likewise.
20375 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20377 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20379 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20380 can benefit from using a sequence of LDRD instructions in epilogue
20381 instead of a single LDM instruction.
20383 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20386 * doc/extend.texi: Fix typo.
20388 2013-04-17 Richard Biener <rguenther@suse.de>
20390 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20391 (vect_build_slp_tree): ... here.
20392 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20393 match. Remove special-casing of mismatched complex loads.
20394 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20395 re-try the match with swapped commutative operands.
20396 (vect_supported_load_permutation_p): Remove special-casing of
20397 mismatched complex loads.
20398 (vect_analyze_slp_instance): Adjust.
20400 2013-04-17 Richard Biener <rguenther@suse.de>
20402 PR rtl-optimization/56921
20403 * cfgloop.h (struct loop): Add simple_loop_desc member.
20404 (struct niter_desc): Mark with GTY(()).
20405 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20406 * loop-iv.c (get_simple_loop_desc): Likewise.
20407 (free_simple_loop_desc): Likewise.
20410 2013-04-16 Richard Biener <rguenther@suse.de>
20412 PR rtl-optimization/56921
20413 * loop-init.c (pass_rtl_move_loop_invariants): Add
20414 TODO_do_not_ggc_collect to todo_flags_finish.
20415 (pass_rtl_unswitch): Same.
20416 (pass_rtl_unroll_and_peel_loops): Same.
20417 (pass_rtl_doloop): Same.
20419 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20421 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20422 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20423 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20425 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20426 * tree-streamer.c (record_common_node): Adjust reference in comment.
20428 2013-04-17 Terry Guo <terry.guo@arm.com>
20430 * config/arm/cortex-m4.md: Add a new bypass.
20432 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20434 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20436 (*subs_<optab><mode>_multp2): New pattern.
20437 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20438 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20440 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20442 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20443 (*subs_mul_imm_<mode>): New pattern.
20445 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20448 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20449 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20450 (vsx_movti_32bit): Change j->wa to O->wa.
20452 2013-04-16 Richard Biener <rguenther@suse.de>
20454 PR rtl-optimization/56921
20455 * loop-init.c (pass_rtl_move_loop_invariants): Add
20456 TODO_do_not_ggc_collect to todo_flags_finish.
20457 (pass_rtl_unswitch): Same.
20458 (pass_rtl_unroll_and_peel_loops): Same.
20459 (pass_rtl_doloop): Same.
20461 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20463 * config/arm/arm.c (emit_multi_reg_push): New declaration
20464 for an existing function.
20465 (arm_emit_strd_push): New function.
20466 (arm_expand_prologue): Used here.
20467 (arm_emit_ldrd_pop): New function.
20468 (arm_expand_epilogue): Used here.
20469 (arm_get_frame_offsets): Update condition.
20470 (arm_emit_multi_reg_pop): Add a special case for load of a single
20471 register with writeback.
20473 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20475 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20478 2013-04-16 Richard Biener <rguenther@suse.de>
20480 PR tree-optimization/56756
20481 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20482 (first_mem_ref_loc): New.
20483 (execute_sm): Place the load temporarily before a previous
20484 access instead of in the latch edge to ensure its SSA dependencies
20485 are defined at points dominating the load.
20487 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20489 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20490 correct fix by moving header and footer insn to the footer of
20491 the merged basic block. Clear BB_END of the merged-away block.
20493 PR middle-end/43631
20494 * emit-rtl.c (make_note_raw): New function.
20495 (link_insn_into_chain): New static inline function.
20496 (add_insn): Use it.
20497 (add_insn_before, add_insn_after): Factor insn chain linking code...
20498 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20499 using link_insn_into_chain.
20500 (note_outside_basic_block_p): New helper function for emit_note_after
20501 and emit_note_before.
20502 (emit_note_after): Use nobb variant of add_insn_after if the note
20503 should not be contained in a basic block.
20504 (emit_note_before): Use nobb variant of add_insn_before if the note
20505 should not be contained in a basic block.
20506 (emit_note_copy): Use make_note_raw.
20507 (emit_note): Likewise.
20508 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20509 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20510 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20511 the moved barrier the tail of the basic block it follows.
20512 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20514 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20516 PR tree-optimization/56962
20517 * gimple-ssa-strength-reduction.c (record_increment): Only set
20518 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20519 either rhs1 or rhs2 is equal to c->base_expr.
20521 2013-04-15 Richard Biener <rguenther@suse.de>
20523 PR tree-optimization/56933
20524 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20526 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20527 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20528 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20529 dependence check ...
20530 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20532 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20533 GROUP_READ_WRITE_DEPENDENCE.
20535 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20537 * emit-rtl.c (reset_all_used_flags): New function.
20538 (verify_rtl_sharing): Call reset_all_used_flags before and after
20539 performing the checks.
20541 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20543 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20544 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20545 * config/arm/constraints.md (De): New constraint.
20546 * config/arm/neon.md (anddi3_neon): Delete.
20547 (neon_vand<mode>): Expand to standard anddi3 pattern.
20548 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20549 Move earlier in the file.
20550 (neon_inv_logic_op2): Likewise.
20551 (arm_anddi_operand_neon): New predicate.
20553 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20555 * configure.ac (gcc_cv_ld_as_needed): Set
20556 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20557 Use -z ignore, -z record on *-*-solaris2*.
20558 (HAVE_LD_AS_NEEDED): Update comment.
20559 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20560 * configure: Regenerate.
20561 * config.in: Regenerate.
20562 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20563 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20564 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20565 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20566 equivalents. Fix markup.
20567 * doc/tm.texi: Regenerate.
20569 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20571 * config/i386/i386.opt: New option mstack-protector-guard=.
20572 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20573 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20574 TARGET_SSP_TLS_GUARD.
20575 * config/i386/i386.c (ix86_option_override_internal): Set
20576 ix86_stack_protector_guard.
20577 * config/i386/i386.md (stack_protect_set): Enable for
20578 TARGET_SSP_TLS_GUARD only.
20579 (stack_protect_set_<mode>): Ditto.
20580 (stack_protect_test): Ditto.
20581 (stack_protect_test_<mode>): Ditto.
20582 * doc/invoke.texi (i386 Option): Document.
20584 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20587 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20588 (S_MODES): Set H_MODE bit.
20589 (SF_MODES): Set only S_MODE and SF_MODE bits.
20590 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20591 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20592 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20593 <MODE_FLOAT>: Likewise.
20595 2013-04-15 Joey Ye <joey.ye@arm.com>
20597 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20599 2013-04-15 Joey Ye <joey.ye@arm.com>
20601 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20603 (thumb_far_jump_used_p): Count instruction size and set
20606 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20608 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20609 * resource.c (mark_target_live_regs): Reformat conditional block.
20611 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20613 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20614 notes, they are emitted only just before final.
20615 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20617 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20619 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20620 * cfgrtl.c (delete_insn): Call it here instead.
20621 * lra-spills.c (lra_final_code_change): Use delete_insn.
20622 * haifa-sched.c (sched_remove_insn): Likewise.
20623 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20624 returning to the nop pool.
20625 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20626 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20628 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20630 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20631 * doc/tm.texi: Regenerated.
20633 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20635 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20638 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20640 * df-core.c (df_find_def): Compare register numbers.
20641 (df_find_use): Likewise.
20643 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20646 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20647 lra_in_progress for return.
20649 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20651 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20652 define_insn into define_insn_and_split and emit movsicc patterns.
20654 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20656 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20658 2013-04-12 Richard Biener <rguenther@suse.de>
20660 * tree-pass.h (TODO_do_not_ggc_collect): New.
20661 * passes.c (execute_one_ipa_transform_pass): Honor
20662 TODO_do_not_ggc_collect.
20663 (execute_one_pass): Likewise.
20666 2013-04-10 Richard Biener <rguenther@suse.de>
20668 * passes.c (init_optimization_passes): Remove reload pass.
20669 * ira.c (do_reload): Merge into ...
20671 (rest_of_handle_reload): Remove.
20672 (pass_reload): Likewise.
20673 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20674 of reload for vzeroupper pass placement.
20676 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20678 PR tree-optimization/56918
20679 PR tree-optimization/56920
20680 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20681 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20682 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20683 use wide_mul_with_sign method.
20685 2013-04-12 Richard Biener <rguenther@suse.de>
20687 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20688 not be considered a gimple constant.
20690 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20692 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20693 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20696 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20697 Jakub Jelinek <jakub@redhat.com>
20699 * opts.c: Include diagnostic-color.h.
20700 (common_handle_option): Handle OPT_fdiagnostics_color_.
20701 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20702 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20703 (diagnostic-color.o): New.
20704 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20705 (diagnostic_color_rule): New enum.
20706 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20707 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20708 the location string.
20709 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20710 either NULL, or color kind.
20711 * diagnostic-color.c: New file.
20712 * diagnostic-color.h: New file.
20713 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20715 * doc/invoke.texi (-fdiagnostics-color): Document.
20716 * pretty-print.h (pp_show_color): Define.
20717 (struct pretty_print_info): Add show_color field.
20718 * diagnostic.c: Include diagnostic-color.h.
20719 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20720 macros. Colorize error:, warning: etc. strings and also the location
20722 (diagnostic_show_locus): Colorize the caret line.
20723 * pretty-print.c: Include diagnostic-color.h.
20724 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20725 inside of %< %> quotes or quoted through q format modifier.
20727 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20729 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20731 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20733 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20734 code in CC_NZ mode.
20735 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20738 2013-04-11 Marek Polacek <polacek@redhat.com>
20740 PR tree-optimization/48184
20741 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20743 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20745 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20746 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20747 (skip_simple_arithmetic): Tidy up.
20748 * tree.h (skip_simple_constant_arithmetic): Declare.
20750 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20752 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20754 2013-04-11 Richard Biener <rguenther@suse.de>
20756 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20757 generate vector constants.
20759 2013-04-11 Richard Biener <rguenther@suse.de>
20761 PR tree-optimization/56878
20762 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20763 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20765 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20766 Prefer to align the DR with the most invariant base address.
20768 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20770 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20773 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20775 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20776 floating-point vector comparisons against 0.
20778 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20780 PR tree-optimization/56899
20781 * fold-const.c (extract_muldiv_1): Apply distributive law
20782 only if TYPE_OVERFLOW_WRAPS (ctype).
20784 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20787 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20788 loading has side effect.
20790 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20792 * config/sparc/sparc.c: Include tree-pass.h.
20793 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20794 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20795 head of file. Change return type. Split off gate function.
20796 (sparc_gate_work_around_errata): New function.
20797 (pass_work_around_errata): New pass definition.
20798 (insert_pass_work_around_errata) New pass insert definition to
20799 insert pass_work_around_errata just after delayed-branch scheduling.
20800 (sparc_option_override): Insert the pass.
20801 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20803 2013-04-10 David S. Miller <davem@davemloft.net>
20805 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20806 or -mcpu=hypersparc.
20808 * target.def (cstore_mode): New hook.
20809 * target.h: Include insn-codes.h
20810 * targhooks.c: Likewise.
20811 (default_cstore_mode): New function.
20812 * targhooks.h: Declare it.
20813 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20814 * doc/tm.texi: Rebuild.
20815 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20816 target hook, rather than inspecting the insn_data.
20817 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20818 (TARGET_CSTORE_MODE): Redefine.
20819 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20821 * config/sparc/predicates.md (cstore_result_operand): New special
20823 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20824 Use it for operand 0.
20825 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20826 (*snesi_special): Likewise.
20827 (*snesi_zero): Likewise.
20828 (*seqsi_zero): Likewise.
20829 (*sltu_insn): Likewise.
20830 (*sgeu_insn): Likewise.
20831 (*seqdi_special): Make operand 0 and comparison operation be of
20833 (*snedi_special): Likewise.
20834 (*snedi_special_vis3): Likewise.
20835 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20836 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20837 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20838 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20839 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20840 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20841 (*sltu_extend_sp64): Likewise.
20842 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20843 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20844 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20845 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20846 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20848 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20850 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20851 (aarch64_start_file): Use the new function.
20853 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20854 Jason Merrill <jason@redhat.com>
20856 * common.opt: Add -gdwarf.
20857 * opts.c (common_handle_option): Handle it.
20858 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20860 2013-04-10 Richard Biener <rguenther@suse.de>
20862 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20863 (execute_one_ipa_transform_pass): But unconditionally here.
20864 (execute_one_pass): And here.
20865 (init_optimization_passes): Remove reload pass.
20866 * tree-pass.h (TODO_ggc_collect): Remove.
20867 (pass_reload): Likewise.
20868 * ira.c (do_reload): Merge into ...
20870 (rest_of_handle_reload): Remove.
20871 (pass_reload): Likewise.
20872 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20873 of reload for vzeroupper pass placement.
20874 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20875 and todo_flags_finish of all passes.
20877 2013-04-10 Richard Biener <rguenther@suse.de>
20879 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20880 first_const_oprnd field, rename first_def_type to first_op_type.
20881 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20882 (vect_get_and_check_slp_defs): Always use the type of the
20883 operand. Allow mixed vect_external_def, vect_constant_def types.
20884 (vect_get_constant_vectors): Handle mixed vect_external_def,
20885 vect_constant_def types.
20887 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20889 PR tree-optimization/55524
20890 * tree-ssa-math-opts.c
20891 (convert_mult_to_fma): Don't use an fms construct
20892 when we don't have an fms operation, but fnma, and it looks
20893 likely that we'll be able to use the latter.
20895 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20897 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20899 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20900 inline fail caused by overwritable functions.
20902 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20904 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20905 unnecessary bits in the constant power of two case.
20907 2013-04-10 Richard Biener <rguenther@suse.de>
20909 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20910 broken code swapping operands.
20911 (vect_build_slp_tree): Do not compute load permutations here.
20912 (vect_analyze_slp_instance): Compute load permutations here,
20913 after building the SLP tree.
20915 2013-04-09 Christian Bruel <christian.bruel@st.com>
20917 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20918 of next/prev_real_insn.
20920 2013-04-09 Jan Hubicka <jh@suse.cz>
20922 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20923 Drop aliased parameter.
20924 (function_and_variable_visibility): Do not handle alias pairs.
20925 * cgraph.c (varpool_externally_visible_p): Update prototype.
20926 * varpool.c (varpool_add_new_variable): Update.
20928 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20930 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20932 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20934 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20936 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20938 2013-04-09 Marek Polacek <polacek@redhat.com>
20940 PR tree-optimization/48762
20941 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20943 2013-04-09 Richard Biener <rguenther@suse.de>
20945 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20947 (vect_build_slp_tree): Likewise.
20948 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20949 calculating the cost of a SLP instance.
20950 (vect_analyze_slp_instance): Use it from here, after building
20953 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20955 PR middle-end/56883
20956 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20957 expand_omp_for_static_chunk): Use simple_p = true in
20958 force_gimple_operand_gsi calls when assigning to addressable decls.
20960 2013-04-09 Jeff Law <law@redhat.com>
20962 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20963 when the boolean was created by converting a wider object which
20964 had a boolean range.
20966 2013-04-09 Richard Biener <rguenther@suse.de>
20968 * tree-vectorizer.h (slp_void_p): Remove.
20969 (slp_tree): Typedef before _slp_tree declaration.
20970 (struct _slp_tree): Use a vector of slp_tree as children.
20971 (vect_get_place_in_interleaving_chain): Remove.
20972 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
20974 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
20976 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
20977 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
20978 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
20979 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
20980 Use slp_node instead of slp_void_p and adjust.
20982 2013-04-09 Richard Biener <rguenther@suse.de>
20984 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
20985 work that is not necessary.
20987 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20989 PR tree-optimization/56854
20990 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
20991 forward into clobber stmts if it would change MEM_REF lhs into
20994 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
20996 * tree.c (type_hash_lookup, type_hash_add): Make static.
20997 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
20999 2013-04-09 Richard Biener <rguenther@suse.de>
21001 * tree.h (unsave_expr_now): Remove.
21002 * tree-inline.c (mark_local_for_remap_r): Remove.
21003 (unsave_expr_1): Likewise.
21004 (unsave_r): Likewise.
21005 (unsave_expr_now): Likewise.
21006 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
21007 (propagate_tree_value): Likewise.
21009 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
21011 * doc/rtl.texi (sequence): Rewrite documentation to match the
21012 current use of SEQUENCE rtl objects.
21013 * rtl.def (SEQUENCE): Likewise.
21015 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
21016 Update documentation.
21017 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
21018 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
21020 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
21022 2013-04-08 Teresa Johnson <tejohnson@google.com>
21024 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
21025 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
21027 (estimate_edge_size_and_time): Add comment to suggest using rounding
21029 (estimate_node_size_and_time): Ditto.
21030 (remap_edge_change_prob): Use helper rounding divide methods.
21031 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
21032 (gimple_mod_pow2_value_transform): Ditto.
21033 (gimple_mod_subtract_transform): Ditto.
21034 (gimple_ic_transform): Ditto.
21035 (gimple_stringops_transform): Ditto.
21036 * stmt.c (conditional_probability): Ditto.
21037 (emit_case_dispatch_table): Ditto.
21038 * lto-cgraph.c (merge_profile_summaries): Ditto.
21039 * tree-optimize.c (execute_fixup_cfg): Ditto.
21040 * cfgcleanup.c (try_forward_edges): Ditto.
21041 * cfgloopmanip.c (scale_loop_profile): Ditto.
21043 (duplicate_loop_to_header_edge): Ditto.
21044 (lv_adjust_loop_entry_edge): Ditto.
21045 * tree-vect-loop.c (vect_transform_loop): Ditto.
21046 * profile.c (compute_branch_probabilities): Ditto.
21047 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
21048 * lto-streamer-in.c (input_cfg): Ditto.
21049 * gimple-streamer-in.c (input_bb): Ditto.
21050 * ipa-cp.c (update_profiling_info): Ditto.
21051 (update_specialized_profile): Ditto.
21052 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
21053 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
21055 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21056 (compute_trg_info): Ditto.
21057 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
21058 (purge_dead_edges): Ditto.
21059 * loop-unswitch.c (unswitch_loop): Ditto.
21060 * cgraphclones.c (cgraph_clone_edge): Ditto.
21061 (cgraph_clone_node): Ditto.
21062 * tree-inline.c (copy_bb): Ditto.
21063 (copy_edges_for_bb): Ditto.
21064 (initialize_cfun): Ditto.
21065 (copy_cfg_body): Ditto.
21066 (expand_call_inline): Ditto.
21068 2013-04-08 Kai Tietz <ktietz@redhat.com>
21070 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
21071 TARGET_CYGWIN64 by TARGET_64BIT.
21073 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21075 * config/epiphany/epiphany.md (GPR_1): New constant.
21076 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
21077 * config/epiphany/epiphany.c (gen_compare_reg):
21078 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
21079 is already in place.
21080 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
21081 Don't require being called during rtl expansion; If y operlaps r0,
21083 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
21084 (epiphany_expand_epilogue): Likewise.
21086 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
21087 Don't use CC_FPmode for ORDERED / UNORDERED.
21088 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
21090 * config/epiphany/constraints.md (CnL): New constraint.
21091 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
21092 * config/epiphany/predicates.md (add_operand): Allow 1024.
21094 * config/epiphany/epiphany.md (logical_op): New code iterator.
21095 (op_mnc): New code attribute.
21096 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
21097 (mov_f+1, mov_f+2): New peephole2 patterns.
21099 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
21100 (cstoresi4): Also allow re-use of zero result when doing a NE
21101 comparison to a non-zero operand.
21102 Use (clobber (scratch)) for first insn if the gpr output is not needed.
21104 * config/epiphany/epiphany.md (<insn_opname>v2si3):
21105 Use gen_addsi3_i / gen_subsi3_i.
21107 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21111 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21112 contain anything but clobbers, at most one __builtin_stack_restore,
21113 optionally debug stmts and final resx, and if it has at least one
21114 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21115 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21116 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21117 which isn't defaut definition, remove them.
21118 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21119 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21120 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21121 with MEM_REF LHS with SSA_NAME address.
21123 2013-04-08 Jeff Law <law@redhat.com>
21125 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21127 2013-04-08 Richard Biener <rguenther@suse.de>
21129 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21131 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21132 determined vector type.
21133 (vect_analyze_data_refs): Likewise.
21134 (vect_get_new_vect_var): Adjust.
21135 (vect_create_destination_var): Preserve SSA name versions.
21136 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21137 not dump anything here.
21139 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21141 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21142 Add member lr_slot_known.
21143 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21145 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21146 Remove code that sets lr_slot_offset according to what a previous
21147 version of epiphany_emit_save_restore used to do.
21148 (epiphany_emit_save_restore): When doing an lr save or restore,
21149 set/verify lr_slot_known and lr_slot_offset.
21151 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21154 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21157 2013-04-08 Richard Biener <rguenther@suse.de>
21159 * alias.c (find_base_term): Fix thinko in previous change.
21161 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21163 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21164 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21165 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21166 if possible to compute val.
21167 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21168 For QImode integers don't require anything about precision. Use
21169 const_with_all_bytes_same to find out if the constant doesn't have
21170 repeated bytes in it.
21172 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21174 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21177 2013-04-08 Marek Polacek <polacek@redhat.com>
21179 PR rtl-optimization/48182
21180 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21183 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21186 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21187 nuses, make sure we have a label.
21189 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21192 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21193 (rs6000_emit_swdiv_low_precision): Remove.
21194 (rs6000_emit_swdiv): Rewrite to handle between one and four
21195 iterations of Newton-Raphson generally; modify required number of
21196 iterations for some cases.
21197 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21199 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21201 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21202 set-but-unused variable.
21204 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21205 basic blocks of released function bodies garbage-collectable.
21207 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21208 (struct rtl_opt_pass): Add TODO_df_finish.
21210 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21212 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21214 * config/arm/constraints.md (q): New constraint.
21215 * config/arm/ldrdstrd.md: New file.
21216 * config/arm/arm.md (ldrdstrd.md) New include.
21217 (arm_movdi): Use "q" instead of "r" constraint
21218 for double-word memory access.
21219 (movdf_soft_insn): Likewise.
21220 * config/arm/vfp.md (movdi_vfp): Likewise.
21221 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21222 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21223 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21224 (mem_ok_for_ldrd_strd): Likewise.
21225 (output_move_double): Update assertion.
21227 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21229 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21231 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21233 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21234 define_insn_and_split.
21235 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21237 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21239 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21240 define_insn_and_split.
21241 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21242 (shiftsi3_compare): New pattern.
21243 (rrx): New pattern.
21244 * config/arm/unspecs.md (UNSPEC_RRX): New.
21246 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21248 * config/arm/arm.md (negdi_extendsidi): New pattern.
21249 (negdi_zero_extendsidi): Likewise.
21251 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21253 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21254 define_insn_and_split.
21255 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21256 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21258 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21260 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21261 define_insn_and_split.
21262 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21263 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21265 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21267 * config/arm/arm.md (subsi3_carryin): New pattern.
21268 (subsi3_carryin_const): Likewise.
21269 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21270 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21272 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21274 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21276 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21278 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21279 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21281 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21283 * config/arm/arm.c (arm_expand_builtin): Change fcode
21284 type to unsigned int.
21286 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21288 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21290 2013-04-04 Ian Lance Taylor <iant@google.com>
21292 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21295 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21297 PR middle-end/56729
21298 * df-scan.c (df_insn_delete): Disable failing assert.
21300 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21302 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21303 New function prototype.
21304 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21305 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21306 (arm_builtin_vectorized_function): New function.
21308 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21310 * config/arm/arm_neon_builtins.def: New file.
21311 * config/arm/arm.c (neon_builtin_data): Move contents to
21312 arm_neon_builtins.def.
21313 (enum arm_builtins): Include neon builtin definitions.
21314 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21315 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21317 2013-04-04 Marek Polacek <polacek@redhat.com>
21319 PR tree-optimization/48186
21320 * predict.c (maybe_hot_frequency_p): Return false if
21321 HOT_BB_FREQUENCY_FRACTION is 0.
21322 (cgraph_maybe_hot_edge_p): Likewise.
21324 2013-04-04 Richard Biener <rguenther@suse.de>
21326 PR tree-optimization/56826
21327 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21330 2013-04-04 Richard Biener <rguenther@suse.de>
21332 PR tree-optimization/56213
21333 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21334 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21336 2013-04-04 Richard Biener <rguenther@suse.de>
21338 PR tree-optimization/56837
21339 * tree-loop-distribution.c (classify_partition): For non-zero
21340 values require that the value has the same precision as its
21341 mode to be useful as memset value.
21343 2013-04-03 Nick Clifton <nickc@redhat.com>
21345 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21346 (fmssf4): Use fmsf.s on E3V5 architectures.
21347 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21348 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21350 2013-04-03 Jeff Law <law@redhat.com>
21352 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21353 (lra-eliminations.o): Likewise.
21355 2013-04-03 Teresa Johnson <tejohnson@google.com>
21357 * gcov-io.c (compute_working_sets): Moved most of body of old
21358 compute_working_sets here from profile.c.
21359 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21360 (gcov_working_set_t): Moved typedef here from basic-block.h
21361 (compute_working_set): Declare.
21362 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21363 (get_working_sets): Renamed from compute_working_set,
21364 replace most of body with call to new compute_working_sets.
21365 (get_exec_counts): Replace call to compute_working_sets
21366 to get_working_sets.
21367 * profile.h (get_working_sets): Renamed from compute_working_set.
21368 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21369 to get_working_sets.
21370 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21371 * gcov-dump.c (dump_working_sets): New function.
21373 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21375 * hwint.c (sext_hwi, zext_hwi): New functions.
21376 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21377 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21378 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21379 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21380 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21381 (sext_hwi, zext_hwi): New functions.
21383 2013-04-03 Jeff Law <law@redhat.com>
21385 PR tree-optimization/56799
21386 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21387 back test for widening conversion erroneously dropped in prior change.
21389 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21392 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21393 instead of next_real_insn.
21395 2013-04-03 Marek Polacek <polacek@redhat.com>
21398 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21400 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21403 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21405 (thumb1_output_casesi): Likewise.
21406 (thumb2_output_casesi): Likewise.
21408 2013-04-03 Richard Biener <rguenther@suse.de>
21410 PR tree-optimization/56817
21411 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21413 (tree_unroll_loops_completely_1): ... new function to manually
21414 walk the loop tree, properly defering outer loops of unrolled
21415 loops to later iterations.
21417 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21419 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21420 (vectorizable_load): Likewise.
21421 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21422 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21424 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21426 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21429 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21431 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21433 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21435 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21437 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21439 PR tree-optimization/56790
21440 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21443 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21445 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21447 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21448 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21451 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21454 * tree.h (force_folding_builtin_constant_p): New decl.
21455 * builtins.c (force_folding_builtin_constant_p): New variable.
21456 (fold_builtin_constant_p): Fold immediately also if
21457 force_folding_builtin_constant_p.
21459 2013-04-03 Richard Biener <rguenther@suse.de>
21461 PR tree-optimization/56812
21462 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21463 DRs of the same interleaving chain are independent.
21465 2013-04-02 Jason Merrill <jason@redhat.com>
21467 * gdbinit.in (pbb): Use debug fn.
21469 2013-04-02 Lawrence Crowl <crowl@google.com>
21471 * sese.h (struct ivtype_map_elt_s): Remove unused.
21472 (extern debug_ivtype_map): Remove unused.
21473 (extern eq_ivtype_map_elts): Remove unused.
21474 * sese.c (debug_ivtype_map): Removed unused.
21475 (debug_ivtype_map_1): Removed unused.
21476 (debug_ivtype_elt): Remove unused.
21477 (eq_ivtype_map_elts): Remove unused.
21480 2013-04-02 Kai Tietz <ktietz@redhat.com>
21483 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21484 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21485 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21487 (legitimize_pe_coff_symbol): Likewise.
21488 (is_imported_p): New helper-function.
21489 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21490 for Windows x64 targets.
21491 (ix86_expand_prologue): Optimize for pe-coff targets.
21492 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21493 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21494 medium/large code-model.
21495 (legitimize_pic_address): Likewise.
21496 (legitimize_tls_address): Likewise.
21497 (ix86_expand_call): Likewise.
21498 (x86_output_mi_thunk): Likewise.
21499 (get_dllimport_decl): Add new beimport argument.
21500 (construct_plt_address): Don't assert for x64 pe-coff targets.
21501 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21503 (SYMBOL_FLAG_STUBVAR): New macro.
21504 (SYMBOL_REF_STUBVAR_P): Likewise.
21505 * config/i386/winnt.c (stub_list): New structure.
21506 (stub_head): New local variable.
21507 (i386_pe_record_stub): New function.
21508 (i386_pe_file_end): Emit refptr-stubs.
21510 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21512 PR rtl-optimization/56745
21513 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21514 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21517 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21518 and both of them are MEM_REFs, just compare first argument for
21519 equality and attempt to deal even with differing offsets.
21522 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21523 of gimple_clobber_p to be MEM_REF.
21524 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21525 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21526 after gimplification.
21527 * asan.c (get_mem_ref_of_assignment): Don't instrument
21528 gimple_clobber_p stmts.
21529 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21530 gimple_clobber_p stmt if they have MEM_REF lhs and
21531 are dead because of another gimple_clobber_p stmt.
21532 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21533 gimple_clobber_p stmts like debug stmts.
21534 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21535 that refer to unused VAR_DECLs or uninitialized values.
21536 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21537 gimple_clobber_p stmts if they refer to removed parameters.
21538 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21541 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21543 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21544 using SWI48 mode attribute.
21546 2013-04-02 Wei Mi <wmi@google.com>
21548 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21549 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21550 *<rotate_insn><mode>3_mask in i386.md.
21552 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21554 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21556 2013-04-02 Richard Biener <rguenther@suse.de>
21558 PR tree-optimization/56778
21559 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21560 Runtime alias tests are not supported for gather loads.
21561 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21562 stmts referenced from SSA operands before updating SSA form.
21564 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21565 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21567 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21568 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21569 * config/arm/cortex-a53.md: New file.
21570 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21571 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21572 * config/arm/arm.c (arm_issue_rate): Likewise.
21573 * config/arm/arm-tune.md: Regenerate
21574 * config/arm/arm-tables.opt: Regenerate.
21575 * config/arm/arm-cores.def: Add cortex-a53.
21577 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21579 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21582 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21584 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21585 scalar load/store operations using B/H registers.
21586 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21588 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21590 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21592 * config/aarch64/aarch64.c
21593 (aarch64_simd_scalar_immediate_valid_for_move): New.
21594 * config/aarch64/aarch64-protos.h
21595 (aarch64_simd_scalar_immediate_valid_for_move): New.
21596 * config/aarch64/constraints.md (Dh, Dq): New.
21597 * config/aarch64/iterators.md (hq): New.
21599 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21601 * reorg.c (get_branch_condition): Deal with conditional returns.
21602 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21604 2013-04-01 Wei Mi <wmi@google.com>
21606 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21607 Truncate operand 2 using %b asm operand modifier.
21608 (*<shift_insn><mode>3_mask): Ditto.
21609 (*<rotate_insn><mode>3_mask): Ditto.
21611 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21613 PR middle-end/56798
21614 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21616 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21618 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21620 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21622 2013-03-30 Lawrence Crowl <crowl@google.com>
21624 * dse.c (clear_alias_sets): Remove never set.
21625 (disqualified_clear_alias_sets): Remove never set.
21626 (clear_alias_mode_pool): Remove never set.
21627 (dse_step0): Remove condition that is never true.
21628 (canon_address): Remove condition that is never true.
21629 (dse_step7): Remove condition that is never true.
21630 (rest_of_handle_dse): Remove condition that is never true.
21631 (rest_of_handle_dse::did_global): Remove never read from above.
21632 (dse_step2_spill): Remove never called from above.
21633 (dse_step5_spill): Remove never called from above.
21635 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21637 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21638 JUMP_TABLE_DATA changes.
21639 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21640 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21641 (Insns) <jump_table_data>: New entry.
21642 * doc/tm.texi: Regenerate.
21644 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21646 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21647 for table jump at the end of a basic block using tablejump_p.
21648 * targhooks.c (default_invalid_within_doloop): Likewise.
21649 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21650 target hook implementation that is identical to the default hook.
21651 (rs6000_invalid_within_doloop): Remove.
21653 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21654 unused variable from tablejump_p call.
21656 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21657 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21658 (INSN_DELETED_P): Likewise.
21659 (emit_jump_table_data): New prototype.
21660 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21661 after 4th as unused.
21662 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21663 * sched-vis.c (print_insn): Likewise.
21664 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21665 insn for compatibility with back ends that use next_active_insn to
21666 identify jump table data.
21667 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21668 (remove_insn): Likewise.
21669 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21671 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21672 (emit_jump_table_data): New function.
21674 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21675 basic block, a JUMP_TABLE_DATA never is.
21676 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21677 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21678 off from code handling real insns.
21679 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21680 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21681 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21682 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21683 is not a NONDEBUG_INSN_P.
21684 * ira-costs.c (scan_one_insn): Likewise.
21685 * jump.c (mark_all_labels): Likewise.
21686 (mark_jump_label_1): Likewise.
21687 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21688 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21690 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21691 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21692 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21693 JUMP_TABLE_DATA_P insns.
21694 (calculate_elim_costs_all_insns): Likewise.
21695 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21696 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21697 (delete_output_reload): Code style fixups.
21698 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21699 insn flags on this non-insn.
21700 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21701 as scheduling barriers, for pre-change compatibility.
21702 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21703 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21705 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21706 redundant JUMP_TABLE_DATA_P test.
21707 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21708 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21709 (frv_for_each_packet): Likewise.
21710 * config/i386/i386.c (min_insn_size): Likewise.
21711 (ix86_avoid_jump_mispredicts): Likewise.
21712 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21713 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21714 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21715 (mips16_insn_length): Robustify.
21716 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21717 (mips16_split_long_branches): Likewise.
21718 * config/pa/pa.c (pa_combine_instructions): Likewise.
21719 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21720 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21721 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21722 as contributing to pool range lengths.
21723 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21724 Remove redundant JUMP_TABLE_DATA_P test.
21725 (sh_loop_align): Likewise.
21726 (split_branches): Likewise.
21727 (sh_insn_length_adjustment): Likewise.
21728 * config/spu/spu.c (get_branch_target): Likewise.
21730 2013-03-29 Jan Hubicka <jh@suse.cz>
21732 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21733 gcov streaming; stream hot bb threshold to ltrans.
21734 * predict.c (get_hot_bb_threshold): Break out from ....
21735 (maybe_hot_count_p): ... here.
21736 (set_hot_bb_threshold): New function.
21737 * lto-section-in.c (lto_section_name): Add profile.
21738 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21739 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21740 and data-streamer.h
21741 (histogram_entry): New structure.
21742 (histogram, histogram_pool): New global vars.
21743 (histogram_hash): New structure.
21744 (histogram_hash::hash): New method.
21745 (histogram_hash::equal): Likewise.
21746 (account_time_size): New function.
21747 (cmp_counts): New function.
21748 (dump_histogram): New function.
21749 (ipa_profile_generate_summary): New function.
21750 (ipa_profile_write_summary): New function.
21751 (ipa_profile_read_summary): New function.
21752 (ipa_profile): Decide on threshold.
21753 (pass_ipa_profile): Add ipa_profile_write_summary and
21754 ipa_profile_read_summary.
21755 * Makefile.in (ipa.o): Update dependencies.
21756 * lto-streamer.h (LTO_section_ipa_profile): New section.
21758 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21760 * tree.h (VAR_P): New.
21762 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21765 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21767 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21769 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21770 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21771 (control_flow_insn_p): Likewise.
21772 * cfgrtl.c (duplicate_insn_chain): Likewise.
21773 * final.c (get_attr_length_1): Likewise.
21774 (shorten_branches): Likewise.
21775 (final_scan_insn): Likewise.
21776 * function.c (instantiate_virtual_regs): Likewise.
21777 * gcse.c (insert_insn_end_basic_block): Likewise.
21778 * ira-costs.c (scan_one_insn): Likewise.
21779 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21780 * lra.c (check_rtl): Likewise.
21781 * reload1.c (elimination_costs_in_insn): Likewise.
21782 * reorg.c (follow_jumps): Likewise.
21784 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21785 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21786 (thumb_far_jump_used_p): Likewise.
21787 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21788 (workaround_speculation): Likewise.
21789 (add_sched_insns_for_speculation): Likewise.
21790 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21791 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21792 (frv_for_each_packet): Likewise.
21793 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21794 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21795 (final_emit_insn_group_barriers): Likewise.
21796 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21797 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21798 (mips16_insn_length): Likewise.
21799 * config/pa/pa.c (pa_reorg): Likewise.
21800 (pa_combine_instructions): Likewise.
21801 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21802 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21803 (sh_reorg): Likewise.
21804 (split_branches): Likewise.
21805 * config/spu/spu.c (get_branch_target): Likewise.
21807 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21810 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21812 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21813 Fix declaration name.
21815 2013-03-28 Lawrence Crowl <crowl@google.com>
21817 * graphds.h (struct graph.indicies): Remove unused.
21818 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21819 (SCOP_ORIGINAL_PDDRS): Remove unused.
21820 * sese.h (extern insert_loop_close_phis): Removed unused.
21821 (extern insert_guard_phis): Removed unused.
21822 (extern ivtype_map_elt_info): Removed unused.
21823 (new_ivtype_map_elt): Removed unused.
21824 * sese.c (ivtype_map_elt_info): Removed unused.
21826 2013-03-28 Lawrence Crowl <crowl@google.com>
21828 * Makefile.in: Add several missing include dependences.
21830 (test-dump.o): New. This object is not added to any executable,
21831 but is present for ad-hoc testing.
21833 (debug (const bitmap_head_def &)): New.
21834 (debug (const bitmap_head_def *)): New.
21836 (extern debug (const bitmap_head_def &)): New.
21837 (extern debug (const bitmap_head_def *)): New.
21839 (debug (edge_def &)): New.
21840 (debug (edge_def *)): New.
21842 (debug (basic_block_def &)): New.
21843 (debug (basic_block_def *)): New.
21845 (dump_node (const_tree, int, FILE *)): Correct source file.
21847 (debug (die_struct &)): New.
21848 (debug (die_struct *)): New.
21850 (extern debug (die_struct &)): New.
21851 (extern debug (die_struct *)): New.
21852 * gimple-pretty-print.c
21853 (debug (gimple_statement_d &)): New.
21854 (debug (gimple_statement_d *)): New.
21855 * gimple-pretty-print.h
21856 (extern debug (gimple_statement_d &)): New.
21857 (extern debug (gimple_statement_d *)): New.
21859 (debug (ira_allocno_copy &)): New.
21860 (debug (ira_allocno_copy *)): New.
21861 (debug (ira_allocno &)): New.
21862 (debug (ira_allocno *)): New.
21864 (extern debug (ira_allocno_copy &)): New.
21865 (extern debug (ira_allocno_copy *)): New.
21866 (extern debug (ira_allocno &)): New.
21867 (extern debug (ira_allocno *)): New.
21869 (debug (live_range &)): New.
21870 (debug (live_range *)): New.
21872 (debug (lra_live_range &)): New.
21873 (debug (lra_live_range *)): New.
21875 (debug (lra_live_range &)): New.
21876 (debug (lra_live_range *)): New.
21878 (debug (omega_pb_d &)): New.
21879 (debug (omega_pb_d *)): New.
21881 (extern debug (omega_pb_d &)): New.
21882 (extern debug (omega_pb_d *)): New.
21884 (debug (const rtx_def &)): New.
21885 (debug (const rtx_def *)): New.
21887 (debug_tree (tree): Move within file.
21888 (debug_raw (const tree_node &)): New.
21889 (debug_raw (const tree_node *)): New.
21890 (dump_tree_via_hooks (const tree_node *, int)): New.
21891 (debug (const tree_node &)): New.
21892 (debug (const tree_node *)): New.
21893 (debug_verbose (const tree_node &)): New.
21894 (debug_verbose (const tree_node *)): New.
21895 (debug_head (const tree_node &)): New.
21896 (debug_head (const tree_node *)): New.
21897 (debug_body (const tree_node &)): New.
21898 (debug_body (const tree_node *)): New.
21899 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21900 (debug (vec<tree, va_gc> &)): New.
21901 (debug (vec<tree, va_gc> *)): New.
21903 (extern debug (const rtx_def &)): New.
21904 (extern debug (const rtx_def *)): New.
21906 (debug_raw (simple_bitmap_def &)): New.
21907 (debug_raw (simple_bitmap_def *)): New.
21908 (debug (simple_bitmap_def &)): New.
21909 (debug (simple_bitmap_def *)): New.
21911 (extern debug (simple_bitmap_def &)): New.
21912 (extern debug (simple_bitmap_def *)): New.
21913 (extern debug_raw (simple_bitmap_def &)): New.
21914 (extern debug_raw (simple_bitmap_def *)): New.
21916 (debug (vinsn_def &)): New.
21917 (debug (vinsn_def *)): New.
21918 (debug_verbose (vinsn_def &)): New.
21919 (debug_verbose (vinsn_def *)): New.
21920 (debug (expr_def &)): New.
21921 (debug (expr_def *)): New.
21922 (debug_verbose (expr_def &)): New.
21923 (debug_verbose (expr_def *)): New.
21924 (debug (vec<rtx> &)): New.
21925 (debug (vec<rtx> *)): New.
21927 (extern debug (vinsn_def &)): New.
21928 (extern debug (vinsn_def *)): New.
21929 (extern debug_verbose (vinsn_def &)): New.
21930 (extern debug_verbose (vinsn_def *)): New.
21931 (extern debug (expr_def &)): New.
21932 (extern debug (expr_def *)): New.
21933 (extern debug_verbose (expr_def &)): New.
21934 (extern debug_verbose (expr_def *)): New.
21935 (extern debug (vec<rtx> &)): New.
21936 (extern debug (vec<rtx> *)): New.
21938 (_list_iter_cond_expr): Make inline instead of static.
21940 (debug (sreal &)): New.
21941 (debug (sreal *)): New.
21943 (extern debug (sreal &)): New.
21944 (extern debug (sreal *)): New.
21946 (extern debug_raw (const tree_node &)): New.
21947 (extern debug_raw (const tree_node *)): New.
21948 (extern debug (const tree_node &)): New.
21949 (extern debug (const tree_node *)): New.
21950 (extern debug_verbose (const tree_node &)): New.
21951 (extern debug_verbose (const tree_node *)): New.
21952 (extern debug_head (const tree_node &)): New.
21953 (extern debug_head (const tree_node *)): New.
21954 (extern debug_body (const tree_node &)): New.
21955 (extern debug_body (const tree_node *)): New.
21956 (extern debug (vec<tree, va_gc> &)): New.
21957 (extern debug (vec<tree, va_gc> *)): New.
21959 (debug (struct loop &)): New.
21960 (debug (struct loop *)): New.
21961 (debug_verbose (struct loop &)): New.
21962 (debug_verbose (struct loop *)): New.
21963 * tree-dump.c: Add header dependence.
21965 (extern debug (struct loop &)): New.
21966 (extern debug (struct loop *)): New.
21967 (extern debug_verbose (struct loop &)): New.
21968 (extern debug_verbose (struct loop *)): New.
21970 (debug (data_reference &)): New.
21971 (debug (data_reference *)): New.
21972 (debug (vec<data_reference_p> &)): New.
21973 (debug (vec<data_reference_p> *)): New.
21974 (debug (vec<ddr_p> &)): New.
21975 (debug (vec<ddr_p> *)): New.
21977 (extern debug (data_reference &)): New.
21978 (extern debug (data_reference *)): New.
21979 (extern debug (vec<data_reference_p> &)): New.
21980 (extern debug (vec<data_reference_p> *)): New.
21981 (extern debug (vec<ddr_p> &)): New.
21982 (extern debug (vec<ddr_p> *)): New.
21984 (debug (pt_solution &)): New.
21985 (debug (pt_solution *)): New.
21987 (extern debug (pt_solution &)): New.
21988 (extern debug (pt_solution *)): New.
21990 (debug (_var_map &)): New.
21991 (debug (_var_map *)): New.
21992 (debug (tree_live_info_d &)): New.
21993 (debug (tree_live_info_d *)): New.
21995 (extern debug (_var_map &)): New.
21996 (extern debug (_var_map *)): New.
21997 (extern debug (tree_live_info_d &)): New.
21998 (extern debug (tree_live_info_d *)): New.
22000 2013-03-28 Jan Hubicka <jh@suse.cz>
22002 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
22004 2013-03-28 Ian Bolton <ian.bolton@arm.com>
22006 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
22007 record only when desired or required.
22009 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
22011 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
22012 *vec_extractv2di_1_rex64. Use x64 isa attribute.
22014 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22016 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
22017 (*andsi3_compare0_uxtw): New pattern.
22018 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
22019 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
22021 2013-03-28 Jan Hubicka <jh@suse.cz>
22023 * data-streamer-in.c (streamer_read_gcov_count): New function.
22024 * gimple-streamer-out.c: Include value-prof.h.
22025 (output_gimple_stmt): Output histogram.
22026 (output_bb): Use streamer_write_gcov_count.
22027 * value-prof.c: Include data-streamer.h
22028 (dump_histogram_value): Add HIST_TYPE_MAX.
22029 (stream_out_histogram_value): New function.
22030 (stream_in_histogram_value): New function.
22031 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
22032 (stream_out_histogram_value, stream_in_histogram_value): Declare.
22033 * data-streamer-out.c (streamer_write_gcov_count): New function.
22034 (streamer_write_gcov_count_stream): New function.
22035 * lto-cgraph.c (lto_output_edge): Update counter streaming.
22036 (lto_output_node): Likewise.
22037 (input_node, input_edge): Likewise.
22038 * lto-streamer-out.c (output_cfg): Update streaming.
22039 * lto-streamer-in.c (input_cfg): Likewise.
22040 * data-streamer.h (streamer_write_gcov_count,
22041 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
22042 * gimple-streamer-in.c: Include value-prof.h
22043 (input_gimple_stmt): Input histograms.
22044 (input_bb): Update profile streaming.
22046 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22048 * genmodes.c (emit_max_int): New function.
22049 (emit_insn_modes_h): Added call to emit_max_function.
22050 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
22052 * machmode.def: Fixed comment.
22054 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22056 * combine.c (try_combine): Removed useless assert.
22057 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
22059 2013-03-28 Marek Polacek <polacek@redhat.com>
22060 Richard Biener <rguenther@suse.de>
22062 PR tree-optimization/56695
22063 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
22064 build signed result of a vector comparison.
22065 * tree-cfg.c (verify_gimple_comparison): Check that a result
22066 of a vector comparison has signed type.
22068 2013-03-28 Richard Biener <rguenther@suse.de>
22070 PR tree-optimization/37021
22071 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
22072 do not restrict gaps between groups.
22073 * tree-vect-stmts.c (vectorizable_load): Properly account for
22074 a gap between groups.
22076 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
22078 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
22079 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
22082 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
22084 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
22085 * doc/extend.texi (Named Address Spaces): Ditto.
22086 (Variable Attributes): Ditto.
22088 2013-03-27 Kai Tietz <ktietz@redhat.com>
22090 * config.build: Add support for cygwin x64 target.
22091 * config.gcc: Likewise.
22092 * config.host: Likewise.
22093 * configure.ac: Likewise
22094 * configure: Regenerated.
22096 2013-03-27 Kai Tietz <ktietz@redhat.com>
22098 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
22099 * config/i386/t-cygwin-w64: New file.
22100 * config/i386/cygwin-w64.h: New file.
22101 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
22102 and add support for x64-cygwin target.
22103 (CPP_SPEC): Likewise.
22104 (CXX_WRAP_SPEC_LIST): Undefine before define.
22105 (LIBGCJ_SONAME): Use 15 as version.
22107 2013-03-27 Richard Biener <rguenther@suse.de>
22109 PR tree-optimization/56716
22110 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22111 dumping for ref nodes.
22113 2013-03-27 Martin Jambor <mjambor@suse.cz>
22115 PR tree-optimization/55334
22116 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22117 restricted pointers to arrays.
22119 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22121 * Makefile.in (.SUFFIXES): Add .cc.
22122 (.c.o): Apply same recipe for implicit rule .cc.o.
22124 2013-03-27 Richard Biener <rguenther@suse.de>
22126 PR tree-optimization/37021
22127 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22128 REALPART/IMAGPART_EXPRs around the supported refs.
22129 * tree-ssa-structalias.c (find_func_aliases): Assume that
22130 floating-point values are not used to transfer pointers.
22132 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22134 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22135 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22136 * doc/tm.texi: Regenerate.
22137 * targhooks.h (default_has_ifunc_p): New.
22138 * targhooks.c (default_has_ifunc_p): Ditto.
22139 * config/linux-protos.h: New file.
22140 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22141 hook for linux which disables support of indirect functions in android.
22142 * config/linux-android.c: New file.
22143 * config/t-linux-android.c: Ditto.
22144 * config.gcc: Added new object file linux-android.o.
22145 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22146 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22147 * varasm.c (do_assemble_alias): Likewise.
22148 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22149 doesn't support indirect functions.
22150 * configure: Regenerate.
22152 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22155 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22156 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22159 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22161 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22163 2013-03-27 Terry Guo <terry.guo@arm.com>
22165 * config/arm/arm-cores.def: Added core cortex-r7.
22166 * config/arm/arm-tune.md: Regenerated.
22167 * config/arm/arm-tables.opt: Regenerated.
22168 * doc/invoke.texi: Added entry for core cortex-r7.
22170 2013-03-27 Walter Lee <walt@tilera.com>
22172 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22173 double-decrement of next_scratch_regno.
22175 2013-03-27 Walter Lee <walt@tilera.com>
22177 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22179 (insn_v1mulus): Ditto.
22180 (insn_v2muls): Ditto.
22182 2013-03-27 Walter Lee <walt@tilera.com>
22184 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22185 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22187 2013-03-27 Walter Lee <walt@tilera.com>
22189 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22190 (*sibcall_value): Ditto.
22192 2013-03-27 Walter Lee <walt@tilera.com>
22194 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22195 (insn_mnz_v8qi): ... this ...
22196 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22198 (insn_v<n>mnz): Replaced by ...
22199 (insn_v1mnz): ... this ...
22200 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22202 (insn_mz_<mode>): Replaced by ...
22203 (insn_mz_v8qi): ... this ...
22204 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22206 (insn_v<n>mz): Replaced by ...
22207 (insn_v1mz): ... this ...
22208 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22211 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22213 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22215 2013-03-26 Roland McGrath <mcgrathr@google.com>
22217 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22218 than fprintf with a non-constant, non-format string.
22220 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22222 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22223 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22224 operand 0 predicate.
22225 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22226 attribute. Use general_x64nomem_operand as operand 1 predicate.
22227 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22228 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22229 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22230 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22231 (mov<mode>_insv_1): Remove expander. Merge insn with
22232 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22233 Use general_x64nomem_operand as operand 1 predicate.
22234 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22235 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22236 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22237 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22238 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22239 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22240 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22241 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22242 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22243 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22244 (general_x64nomem_operand): Ditto.
22246 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22248 * config/rtems.opt: Add -pthread option.
22250 2013-03-26 Richard Biener <rguenther@suse.de>
22252 * alias.c (find_base_term): Avoid redundant and not used recursion.
22253 (base_alias_check): Get the initial base term from the caller.
22254 (true_dependence_1): Compute and pass base terms to base_alias_check.
22255 (write_dependence_p): Likewise.
22256 (may_alias_p): Likewise.
22258 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22260 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22261 PC-relative load in SI modes and above only.
22263 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22265 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22266 * config/arm/iwmmxt.md (WCGR0): Update.
22267 (WCGR1, WCGR2, WCGR3): Likewise.
22269 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22271 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22272 Use x64 and nox64 isa attributes.
22274 2013-03-26 Richard Biener <rguenther@suse.de>
22276 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22277 alignment computations and rely on get_object_alignment_1
22278 for the !TYPE_P case.
22279 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22281 2013-03-26 Walter Lee <walt@tilera.com>
22283 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22284 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22286 2013-03-25 Jeff Law <law@redhat.com>
22288 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22289 check for INTEGRAL_TYPE_P that was missing due to checking in
22290 wrong version of prior patch.
22292 2013-03-25 Walter Lee <walt@tilera.com>
22294 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22295 TILEGX_INSN_SHUFFLEBYTES1.
22296 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22298 (tilegx_builtins): Ditto.
22299 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22301 2013-03-25 Walter Lee <walt@tilera.com>
22303 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22304 (floatunssisf2): New pattern.
22305 (floatsidf2): New pattern.
22306 (floatunssidf2): New pattern.
22308 2013-03-25 Walter Lee <walt@tilera.com>
22310 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22311 tests for constraint J, K, N, P.
22313 2013-03-25 Walter Lee <walt@tilera.com>
22315 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22316 Use indirect/pcrel encoding.
22317 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22320 2013-03-25 Steve Ellcey <sellcey@mips.com>
22322 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22323 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22324 * config/mips/mips.c (mips_option_override): Set IMADD default.
22325 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22326 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22327 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22328 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22329 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22331 2013-03-25 Jeff Law <law@redhat.com>
22333 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22334 slightly to avoid creating and folding useless trees. Simplify
22335 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22337 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22339 * config/i386/i386.md (*zero_extendsidi2): Merge with
22340 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22341 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22342 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22343 targets. Clarify comment.
22345 2013-03-25 Martin Jambor <mjambor@suse.cz>
22347 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22348 pass-through jump functions differently.
22349 (ipa_read_jump_function): Likewise. Also use setter functions to set
22352 2013-03-25 Martin Jambor <mjambor@suse.cz>
22354 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22355 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22357 (ipa_get_indirect_edge_target): New function.
22358 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22359 ipa_get_indirect_edge_target. Update all callers.
22360 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22361 ipa_get_indirect_edge_target_1 instead of calling
22362 ipa_get_indirect_edge_target.
22363 (create_specialized_node): Pass aggvlas to
22364 ipcp_discover_new_direct_edges.
22366 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22368 * config/arm/arm.md (f_sels, f_seld): New types.
22369 (*cmov<mode>): New pattern.
22370 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22373 2013-03-25 Kai Tietz <ktietz@redhat.com>
22375 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22376 POSIX-printf for mingw-hosted builds.
22378 2013-03-25 Richard Biener <rguenther@suse.de>
22380 PR middle-end/56694
22381 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22382 must-not-throw stmt location.
22384 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22386 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22387 Emit load-acquire versions when acq is true.
22388 (arm_emit_store_exclusive): Add rel parameter.
22389 Emit store-release versions when rel is true.
22390 (arm_split_compare_and_swap): Use acquire-release instructions
22392 of barriers when appropriate.
22393 (arm_split_atomic_op): Likewise.
22394 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22395 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22396 (VUNSPEC_SLX): Likewise.
22397 (VUNSPEC_LDA): Likewise.
22398 (VUNSPEC_STL): Likewise.
22399 * config/arm/sync.md (atomic_load<mode>): New pattern.
22400 (atomic_store<mode>): Likewise.
22401 (arm_load_acquire_exclusive<mode>): Likewise.
22402 (arm_load_acquire_exclusivesi): Likewise.
22403 (arm_load_acquire_exclusivedi): Likewise.
22404 (arm_store_release_exclusive<mode>): Likewise.
22406 2013-03-25 Catherine Moore <clm@codesourcery.com>
22408 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22409 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22410 * config/mip/predicates.md (lwsp_swsp_operand,
22411 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22412 sb16_operand, db4_operand, db7_operand, ib3_operand,
22413 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22414 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22415 andi16_operand): New predicates.
22416 * config/mips/mips.md (compression): New attribute.
22417 (enabled): New attribute.
22418 (length): Consider compression in computing length.
22419 (shift_compression): New code attribute.
22420 (*add<mode>3): New operands. Record compression.
22421 (sub<mode>3): Likewise.
22422 (one_cmpl<mode>2): Likewise.
22423 (*and<mode>3): Likewise.
22424 (*ior<mode>3): Likewise.
22425 (unnamed pattern for xor): Likewise.
22426 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22427 (*<optab><mode>3): Likewise.
22428 (*mov<mode>_internal: Likewise.
22429 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22430 (mips_unsigned_immediate_p): New.
22431 (umips_lwsp_swsp_address_p): New.
22432 (m16_based_address_p): New.
22433 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22434 (mips_unsigned_immediate_p): New prototype.
22435 (lwsp_swsp_address_p): New prototype.
22436 (m16_based_address_p): New prototype.
22437 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22438 (mips_signed_immediate_p): New function.
22439 (m16_based_address_p): New function.
22440 (lwsp_swsp_address_p): New function.
22441 (mips_print_operand_punctuation): Recognize short delay slot insns
22442 for microMIPS.add<mode>3"
22444 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22447 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22448 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22450 2013-03-25 Richard Biener <rguenther@suse.de>
22452 PR tree-optimization/56689
22453 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22456 2013-03-25 Richard Biener <rguenther@suse.de>
22458 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22460 (memory_references): Likewise.
22461 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22462 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22463 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22464 (gather_mem_refs_in_loops): Fold into ...
22465 (analyze_memory_references): ... this. Move initialization
22466 to tree_ssa_lim_initialize.
22467 (fill_always_executed_in): Rename to ...
22468 (fill_always_executed_in_1): ... this.
22469 (fill_always_executed_in): Move contains_call computation to
22470 this new function from ...
22471 (tree_ssa_lim_initialize): ... here.
22472 (tree_ssa_lim): Call fill_always_executed_in.
22474 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22476 * postreload.c (reload_combine): Fix code detecting returns.
22478 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22480 * function.c (emit_use_return_register_into_block): On cc0 targets,
22481 do not emit the sequence between cc0 setter and user.
22483 2013-03-25 Kai Tietz <ktietz@redhat.com>
22485 * config/i386/predicates.md (local_symbolic_operand): Interpret
22486 dll-imported symbols as none-local.
22488 2013-03-25 Richard Biener <rguenther@suse.de>
22490 * tree-ssa-loop-im.c (struct depend): Remove.
22491 (struct lim_aux_data): Make depends a vec of gimples.
22492 (free_lim_aux_data): Adjust.
22493 (add_dependency): Likewise.
22494 (set_level): Likewise.
22496 2013-03-25 Richard Biener <rguenther@suse.de>
22498 PR middle-end/56434
22499 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22500 the pointer returned by calls with ECF_MALLOC set.
22502 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22504 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22506 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22508 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22509 using MMXMODE mode iterator.
22510 (*move<mode>_internal): Merge with *movv2sf_internal and
22511 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22513 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22515 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22516 (record_last_mem_set_info): Likewise.
22518 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22519 of XNEWVEC followed by memset.
22520 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22522 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22524 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22525 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22526 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22527 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22528 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22529 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22530 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22531 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22532 BARRIER_P instead of GET_CODE.
22534 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22536 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22537 inaccuracy in the probing code.
22539 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22540 (ctrapdi4): Likewise.
22542 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22544 * calls.c (expand_call): Add missing guard to code handling return
22545 of non-BLKmode structures in MSB.
22546 * function.c (expand_function_end): Likewise.
22548 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22550 * combine.c (try_combine): Adjust comment. Do not add the set of
22551 insn #0 if the destination indirectly is set or dies in insn #2.
22552 Tidy up code to distribute a new note.
22554 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22556 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22557 also for alternatives 16 and 17.
22559 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22561 * config/i386/sse.md (*mov<mode>_internal): Merge with
22562 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22563 Emit insn template depending on type attribute. Use
22564 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22565 movd instead of movq mnemonic for interunit moves. Rewrite mode
22566 attribute calculation. Remove unit attribute calculation.
22567 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22568 Set prefix_data16 attribute for DImode ssemov types.
22569 Use Ym instead of y for SSE-MMX conversion alternatives.
22570 Reorder operand constraints.
22572 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22574 * df.h (df_insn_delete): Adjust prototype.
22575 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22576 and let it decide whether mark the basic block dirty.
22577 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22578 * df-scan.c (df_insn_info_delete): New helper function, split
22579 off from df_insn_delete.
22580 (df_scan_free_bb_info): Use it.
22581 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22583 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22584 that the insn is actually an insn and it has a non-NULL basic block.
22585 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22587 2013-03-22 Richard Biener <rguenther@suse.de>
22589 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22591 (mem_ref_alloc): Do not allocate them.
22592 (refs_independent_p): Do not query or maintain a cache.
22594 2013-03-22 Richard Biener <rguenther@suse.de>
22596 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22597 (gather_mem_refs_in_loops): Do not compute it.
22598 (analyze_memory_references): Do not allocate it.
22599 (tree_ssa_lim_finalize): Do not free it.
22600 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22602 2013-03-22 Richard Biener <rguenther@suse.de>
22604 * is-a.h (as_a): Use gcc_checking_assert.
22606 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22608 * config/aarch64/aarch64.c (aarch64_print_operand): New
22609 format specifier for printing a constant in hex.
22610 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22611 format specifier for printing second operand.
22613 2013-03-22 Richard Biener <rguenther@suse.de>
22615 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22617 (gather_mem_refs_in_loops): Perform store accumulation here.
22618 (create_vop_ref_mapping_loop): Remove.
22619 (create_vop_ref_mapping): Likewise.
22620 (analyze_memory_references): Initialize refs_stored_in_loop.
22621 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22622 (record_indep_loop): Remove.
22623 (record_dep_loop): New function.
22624 (ref_indep_loop_p_1): Adjust to only walk over references
22625 in the loop, not its subloops.
22626 (ref_indep_loop_p): Rename to ...
22627 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22628 maintaining a more fine-grained cache.
22629 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22630 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22632 2013-03-22 Richard Biener <rguenther@suse.de>
22634 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22635 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22636 aggregate mem_ref_loc.
22637 (free_mem_ref_locs): Inline into ...
22638 (memref_free): ... this and adjust.
22639 (mem_ref_alloc): Adjust.
22640 (mem_ref_locs_alloc): Remove.
22641 (record_mem_ref_loc): Adjust.
22642 (get_all_locs_in_loop): Rewrite into ...
22643 (for_all_locs_in_loop): ... this iterator.
22644 (rewrite_mem_ref_loc): New functor.
22645 (rewrite_mem_refs): Use for_all_locs_in_loop.
22646 (sm_set_flag_if_changed): New functor.
22647 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22648 (ref_always_accessed): New functor.
22649 (ref_always_accessed_p): Use for_all_locs_in_loop.
22651 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22653 * tree-pass.h (PROP_gimple_lvec): New.
22654 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22655 (init_optimization_passes): Move pass_lower_vector.
22656 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22658 (pass_lower_vector): Provide PROP_gimple_lvec.
22659 (pass_lower_vector_ssa): Likewise.
22660 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22662 2013-03-21 Mark Wielaard <mjw@redhat.com>
22664 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22666 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22668 * config/i386/i386.md (*movdi_internal): Disparage slightly
22669 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22670 conversion alternatives.
22672 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22674 PR middle-end/48087
22675 * diagnostic.def (DK_WERROR): New kind.
22676 * diagnostic.h (werrorcount): Define.
22677 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22678 promoted to DK_ERROR, increment DK_WERROR counter instead of
22680 * toplev.c (toplev_main): Call print_ignored_options even if
22681 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22682 even if just werrorcount is non-zero.
22685 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22687 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22688 (string_cst_pool_decl): New function.
22689 (optimize_one_addr_into_implicit_ptr): New function.
22690 (resolve_addr_in_expr): Optimize DWARF location expression
22691 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22692 which doesn't live in memory, but has DW_AT_location or
22693 DW_AT_const_value, or refers to a string literal, into
22694 DW_OP_GNU_implicit_pointer.
22695 (optimize_location_into_implicit_ptr): New function.
22696 (resolve_addr): If removing DW_AT_location of a variable because
22697 it was DW_OP_addr of address of the variable, but the variable doesn't
22698 live in memory, try to emit const value attribute for the initializer.
22700 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22702 * tree.h (VECTOR_TYPE_P): New macro.
22703 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22704 TYPE_MODE): Use it.
22705 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22706 VEC_COND_EXPR cannot be lvalues.
22707 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22709 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22711 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22712 Restrict the transformation to equal modes.
22714 2013-03-21 Richard Biener <rguenther@suse.de>
22716 PR tree-optimization/39326
22717 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22718 (MEM_ANALYZABLE): Adjust.
22719 (record_mem_ref_loc): Move bitmap ops ...
22720 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22721 unanalyzable refs, do not record locations for it.
22722 (analyze_memory_references): Allocate ref zero as shared
22724 (refs_independent_p): Do not test for unanalyzed mems here.
22725 (ref_indep_loop_p_1): Special-case disambiguation against
22726 the unanalyzed ref.
22727 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22729 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22731 * config/arm/arm-protos.h (tune_params): Add
22732 prefer_neon_for_64bits field.
22733 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22734 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22735 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22736 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22737 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22738 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22739 (arm_option_override): Handle -mneon-for-64bits new option.
22740 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22741 (prefer_neon_for_64bits): Declare new variable.
22742 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22743 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22744 (arch_enabled): Handle new arch types. Remove support for onlya8
22746 (one_cmpldi2): Use new arch names.
22747 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22748 * config/arm/arm.opt (mneon-for-64bits): Add option.
22749 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22750 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22751 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22753 * doc/invoke.texi (-mneon-for-64bits): Document.
22755 2013-03-21 Richard Biener <rguenther@suse.de>
22757 PR tree-optimization/39326
22758 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22759 (sort_bbs_in_loop_postorder_cmp): New function.
22760 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22762 2013-03-21 Richard Biener <rguenther@suse.de>
22764 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22765 (vect_insert_into_interleaving_chain): Likewise.
22766 (vect_drs_dependent_in_basic_block): Inline ...
22767 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22769 (vect_analyze_data_ref_dependence): ... here. Simplify.
22770 (vect_check_interleaving): Simplify.
22771 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22772 (vect_slp_analyze_data_ref_dependences): ... this new function.
22773 (dr_group_sort_cmp): New function.
22774 (vect_analyze_data_ref_accesses): Compute data-reference groups
22775 here instead of in vect_analyze_data_ref_dependence. Use
22776 a more efficient algorithm.
22777 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22778 vect_slp_analyze_data_ref_dependences. Call
22779 vect_analyze_data_ref_accesses earlier.
22780 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22781 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22782 (vect_slp_analyze_data_ref_dependences): New prototype.
22784 2013-03-21 Richard Biener <rguenther@suse.de>
22786 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22787 ref is stored in the loop.
22788 (find_refs_for_sm): Walk only over all stores.
22789 (store_motion_loop): Allocate from lim_bitmap_obstack.
22790 (store_motion): Likewise.
22792 2013-03-21 Richard Biener <rguenther@suse.de>
22794 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22795 Update virtual SSA form.
22797 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22799 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22800 * configure: Regenerate.
22801 * config.in: Regenerate.
22802 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22803 if !HAVE_LD_EH_FRAME_CIEV3.
22805 2013-03-21 Richard Biener <rguenther@suse.de>
22807 * tree-cfg.c (verify_expr_no_block): New function.
22808 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22809 nor DECL_VALUE_EXPR have locations with associated blocks.
22810 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22811 (clear_unused_block_pointer): Remove code dealing with
22812 blocks in DECL_DEBUG_EXPR locations.
22814 2013-03-21 Richard Biener <rguenther@suse.de>
22816 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22817 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22818 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22819 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22820 instead of DECL_DEBUG_EXPR_IS_FROM.
22821 * gimplify.c (gimplify_modify_expr): Likewise.
22822 * tree-cfg.c (verify_expr_location_1): Likewise.
22823 * tree-complex.c (create_one_component_var): Likewise.
22824 * tree-sra.c (create_access_replacement): Likewise.
22825 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22826 (clear_unused_block_pointer): Likewise.
22827 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22828 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22829 * var-tracking.c (var_debug_decl): Likewise.
22830 (track_expr_p): Likewise.
22831 * tree-inline.c (add_local_variables): Likewise. Set
22832 DECL_HAS_DEBUG_EXPR_P after copying it.
22833 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22834 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22836 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22839 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22840 * configure: Regenerate.
22841 * config.in: Regenerate.
22842 * config/i386/i386.md (*movdf_internal): Use
22843 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22844 movd instead of movq mnemonic for interunit moves.
22845 (*movdi_internal): Ditto.
22847 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22849 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22850 (abd<mode>_3): New pattern.
22851 (aba<mode>_3): New pattern.
22852 (fabd<mode>_3): New pattern.
22854 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22856 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22857 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22858 occurrence of REGISTER_PREFIX as its empty string.
22860 2013-03-20 Jeff Law <law@redhat.com>
22862 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22863 addititional equivalences for equality comparisons between an SSA_NAME
22864 and a constant where the SSA_NAME was set from a widening conversion.
22866 2013-03-20 Walter Lee <walt@tilera.com>
22868 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22870 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22872 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22873 depending on type attribute.
22874 (*movti_internal): Ditto.
22875 (*movtf_internal): Ditto.
22876 (*movxf_internal): Ditto.
22877 (*movdf_internal): Ditto.
22878 (*movsf_internal): Ditto.
22880 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22882 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22883 maybe_vex for sselog1 and ssemov types.
22884 (*movdi_internal): Reorder operand constraints.
22885 (*movsi_internal): Ditto. Set prefix attribute to
22886 maybe_vex for sselog1 and ssemov types.
22887 (*movtf_internal): Set prefix attribute to maybe_vex
22888 for sselog1 and ssemov types.
22889 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22890 DImode ssemov types. Reorder operand constraints.
22891 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22892 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22893 attribute for SImode ssemov types. Reorder operand constraints.
22895 2013-03-20 Martin Jambor <mjambor@suse.cz>
22897 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22898 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22900 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22902 * config/rs6000/predicates.md (indexed_address, update_address_mem
22903 update_indexed_address_mem): New predicates.
22904 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22905 attribute for load/store instructions.
22906 * config/rs6000/dfp.md (movsd_store): Likewise.
22907 (movsd_load): Likewise.
22908 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22909 (unnamed HI->DI extend define_insn): Likewise.
22910 (unnamed SI->DI extend define_insn): Likewise.
22911 (unnamed QI->SI extend define_insn): Likewise.
22912 (unnamed QI->HI extend define_insn): Likewise.
22913 (unnamed HI->SI extend define_insn): Likewise.
22914 (unnamed HI->SI extend define_insn): Likewise.
22915 (extendsfdf2_fpr): Likewise.
22916 (movsi_internal1): Likewise.
22917 (movsi_internal1_single): Likewise.
22918 (movhi_internal): Likewise.
22919 (movqi_internal): Likewise.
22920 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22921 attribute for load/store instructions.
22922 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22924 (mov<mode>_softfloat): Likewise.
22925 (mov<mode>_hardfloat32): Likewise.
22926 (mov<mode>_hardfloat64): Likewise.
22927 (mov<mode>_softfloat64): Likewise.
22928 (movdi_internal32): Likewise.
22929 (movdi_internal64): Likewise.
22930 (probe_stack_<mode>): Likewise.
22932 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22934 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22935 floating point, and decimal floating point to reload iterator.
22937 * config/rs6000/constraints.md (wl constraint): New constraints to
22938 return FLOAT_REGS if certain options are used to reduce the number
22939 of separate patterns that exist in the file.
22940 (wx constraint): Likewise.
22941 (wz constraint): Likewise.
22943 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22944 -mdebug=reg, print wg, wl, wx, and wz constraints.
22945 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22946 Initialize the reload functions for 64-bit binary/decimal floating
22948 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22949 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22950 create the buffer on the stack to overcome not having a 32-bit
22952 (rs6000_emit_move): Likewise.
22953 (rs6000_secondary_memory_needed_rtx): Likewise.
22954 (rs6000_alloc_sdmode_stack_slot): Likewise.
22955 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22956 via xxlxor, just like DFmode 0.0.
22958 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22959 define as 1 if we are running on a power7 or newer.
22960 (enum r6000_reg_class_enum): Add new constraints.
22962 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22963 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22964 with other moves by using conditional constraits (wg). Use LFIWZX
22965 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22966 (movsd splitter): Likewise.
22967 (movsd_hardfloat): Likewise.
22968 (movsd_softfloat): Likewise.
22970 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22971 binary and decimal floating point moves.
22972 (fmove_ok): New attributes to combine binary and decimal floating
22973 point moves, and to combine power6x (mfpgpr) moves along normal
22975 (real_value_to_target): Likewise.
22976 (f32_lr): Likewise.
22977 (f32_lm): Likewise.
22978 (f32_li): Likewise.
22979 (f32_sr): Likewise.
22980 (f32_sm): Likewise.
22981 (f32_si): Likewise.
22982 (movsf): Combine binary and decimal floating point moves. Combine
22983 power6x (mfpgpr) moves with other moves by using conditional
22984 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
22985 (mov<mode> for SFmode/SDmode); Likewise.
22986 (SFmode/SDmode splitters): Likewise.
22987 (movsf_hardfloat): Likewise.
22988 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
22989 (movsf_softfloat): Likewise.
22990 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
22992 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
22993 wx and wz constraints.
22995 * config/rs6000/constraints.md (wg constraint): New constraint to
22996 return FLOAT_REGS if -mmfpgpr (power6x) was used.
22998 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
23001 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23002 -mdebug=reg, print wg, wl, wx, and wz constraints.
23003 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23004 Initialize the reload functions for 64-bit binary/decimal floating
23006 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23007 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23008 create the buffer on the stack to overcome not having a 32-bit
23010 (rs6000_emit_move): Likewise.
23011 (rs6000_secondary_memory_needed_rtx): Likewise.
23012 (rs6000_alloc_sdmode_stack_slot): Likewise.
23013 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23014 via xxlxor, just like DFmode 0.0.
23016 * config/rs6000/dfp.md (movdd): Delete, combine with binary
23017 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23018 with other moves by using conditional constraits (wg). Use LFIWZX
23019 and STFIWX for loading SDmode on power7.
23020 (movdd splitters): Likewise.
23021 (movdd_hardfloat32): Likewise.
23022 (movdd_softfloat32): Likewise.
23023 (movdd_hardfloat64_mfpgpr): Likewise.
23024 (movdd_hardfloat64): Likewise.
23025 (movdd_softfloat64): Likewise.
23027 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
23028 64-bit binary and decimal floating point moves.
23029 (FMOVE64X): Likewise.
23030 (movdf): Combine 64-bit binary and decimal floating point moves.
23031 Combine power6x (mfpgpr) moves with other moves by using
23032 conditional constraits (wg).
23033 (mov<mode> for DFmode/DDmode): Likewise.
23034 (DFmode/DDmode splitters): Likewise.
23035 (movdf_hardfloat32): Likewise.
23036 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
23037 (movdf_softfloat32): Likewise.
23038 (movdf_hardfloat64_mfpgpr): Likewise.
23039 (movdf_hardfloat64): Likewise.
23040 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
23041 (movdf_softfloat64): Likewise.
23042 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
23043 (reload_<mode>_load): Move to later in the file so they aren't in
23044 the middle of the floating point move insns.
23045 (reload_<mode>_store): Likewise.
23047 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
23050 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
23051 constraint if -mdebug=reg.
23052 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
23053 Enable using dd reload support if needed.
23055 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
23056 binary and decimal floating point moves in rs6000.md.
23057 (movtd_internal): Likewise.
23059 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
23060 decimal floating point moves.
23062 (movtf_internal): Likewise.
23063 (mov<mode>_internal, TDmode/TFmode): Likewise.
23064 (movtf_softfloat): Likewise.
23065 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
23067 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
23068 movdi_internal64, using wg constraint for move direct operations.
23069 (movdi_internal64): Likewise.
23071 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
23072 MODES_TIEABLE_P for selected modes. Print the numerical value of
23073 the various virtual registers. Use GPR/FPR first/last values,
23074 instead of hard coding the register numbers. Print which modes
23075 have reload functions registered.
23076 (rs6000_option_override_internal): If -mdebug=reg, trace the options
23077 settings before/after setting cpu, target and subtarget settings.
23078 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
23079 and for secondary reload failures in rs6000_secondary_reload_inner.
23080 (rs6000_secondary_reload_fail): Likewise.
23081 (rs6000_secondary_reload_inner): Likewise.
23083 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
23084 macros for first/last GPR and FPR registers.
23085 (LAST_GPR_REGNO): Likewise.
23086 (FIRST_FPR_REGNO): Likewise.
23087 (LAST_FPR_REGNO): Likewise.
23089 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
23090 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
23091 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
23092 (vcond<mode><mode>): Likewise.
23093 (vcondu<mode><mode>): Likewise.
23094 (vector_gtu<mode>): Likewise.
23095 (vector_gte<mode>): Likewise.
23096 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
23097 to prevent the compiler from converting DImode operations to TImode.
23098 (ior<mode>3): Likewise.
23099 (and<mode>3): Likewise.
23100 (one_cmpl<mode>2): Likewise.
23101 (nor<mode>3): Likewise.
23102 (andc<mode>3): Likewise.
23104 * config/rs6000/constraints.md (wt constraint): New constraint
23105 that returns VSX_REGS if TImode is allowed in VSX registers.
23107 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23108 constant under VSX.
23110 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23111 similar to TImode, but it is restricted to being in the GPRs.
23113 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23114 TImode to occupy a single VSX register.
23116 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23117 -mvsx-timode for power7/power8.
23118 (power7 cpu): Likewise.
23119 (power8 cpu): Likewise.
23121 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23122 sure that TFmode/TDmode take up two registers if they are ever
23123 allowed in the upper VSX registers.
23124 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23126 (rs6000_init_hard_regno_mode_ok): Likewise.
23127 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23128 constraint. Print if LRA is turned on.
23129 (rs6000_option_override_internal): Give an error if -mvsx-timode
23130 and VSX is not enabled.
23131 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23132 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23133 to reg+offset addressing. Use PTImode when checking offset
23134 addresses for validity.
23135 (reg_offset_addressing_ok_p): Likewise.
23136 (rs6000_legitimate_offset_address_p): Likewise.
23137 (rs6000_legitimize_address): Likewise.
23138 (rs6000_legitimize_reload_address): Likewise.
23139 (rs6000_legitimate_address_p): Likewise.
23140 (rs6000_eliminate_indexed_memrefs): Likewise.
23141 (rs6000_emit_move): Likewise.
23142 (rs6000_secondary_reload): Likewise.
23143 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23144 reloads to fpr registers to continue to use reg+offset addressing,
23145 but 64-bit reloads to altivec registers need reg+reg addressing.
23146 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23147 it. Treat LO_SUM like a PLUS operation.
23148 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23149 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23150 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23151 registers to share a register with a smaller sized type, since VSX
23152 puts scalars in the upper 64-bits.
23153 (print_operand): Add support for PTImode.
23154 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23155 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23156 registers, but don't have arithmetic support.
23157 (rs6000_memory_move_cost): Add test for VSX.
23158 (rs6000_opt_masks): Add -mvsx-timode.
23160 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23163 (VSr): Use wt constraint for TImode.
23164 (VSv): Drop TImode support.
23165 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23166 (vsx_movti_64bit): Likewise.
23167 (vsx_movti_32bit): Likewise.
23168 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23169 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23170 one '?' on the appropriate output constraint. Do not allow TImode
23171 logical operations on 32-bit systems.
23172 (vsx_ior<mode>3): Likewise.
23173 (vsx_xor<mode>3): Likewise.
23174 (vsx_one_cmpl<mode>2): Likewise.
23175 (vsx_nor<mode>3): Likewise.
23176 (vsx_andc<mode>3): Likewise.
23177 (vsx_concat_<mode>): Likewise.
23178 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23180 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23181 OPTION_MASK_VSX_TIMODE.
23182 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23183 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23185 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23186 (TI2 iterator): New iterator for TImode, PTImode.
23187 (wd mode attribute): Add values for vector types.
23188 (movti_string): Replace TI move operations with operations for TImode
23189 and PTImode. Add support for TImode being allowed in VSX registers.
23190 (mov<mode>_string, TImode/PTImode): Likewise.
23191 (movti_ppc64): Likewise.
23192 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23193 (TI mode splitters): Likewise.
23195 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23198 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23200 PR tree-optimization/56355
23201 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23202 Also handle integers with undefined overflow.
23204 2013-03-20 Catherine Moore <clm@codesourcery.com>
23205 Maciej W. Rozycki <macro@codesourcery.com>
23206 Tom de Vries <tom@codesourcery.com>
23207 Nathan Sidwell <nathan@codesourcery.com>
23208 Iain Sandoe <iain@codesourcery.com>
23209 Nathan Froyd <froydnj@codesourcery.com>
23210 Chao-ying Fu <fu@mips.com>
23212 * doc/extend.texi: (micromips, nomicromips, nocompression):
23213 Document new function attributes.
23214 * doc/invoke.texi (minterlink-compressed, mmicromips,
23215 m14k, m14ke, m14kec): Document new options.
23216 (minterlink-mips16): Update documentation.
23217 * doc/md.texi (ZC, ZD): Document new constraints.
23218 * configure.ac (gcc_cv_as_micromips): Check if linker
23219 supports the .set micromips directive.
23220 * configure: Regenerate.
23221 * config.in: Regenerate.
23222 * config/mips/mips-tables.opt: Regenerate.
23223 * config/mips/micromips.md: New file.
23224 * constraints.md (ZC, ZD): New constraints.
23225 * config/mips/predicates.md (movep_src_register): New predicate.
23226 (movep_src_operand): New predicate.
23227 (non_volatile_mem_operand): New predicate.
23228 * config/mips/mips.md (multimem): New type.
23229 (length): Differentiate between 17-bit and 18-bit branch offsets.
23230 (MOVEP1, MOVEP2): New mode iterator.
23231 (mov_<load>l): Use ZC constraint.
23232 (mov_<load>r): Likewise.
23233 (mov_<store>l): Likewise.
23234 (mov_<store>r): Likewise.
23235 (*branch_equality<mode>_inverted): Add microMIPS support.
23236 (*branch_equality<mode>): Likewise.
23237 (*jump_absolute): Likewise.
23238 (indirect_jump_<mode>): Likewise.
23239 (tablejump_<mode>): Likewise.
23240 (<optab>_internal): Likewise.
23241 (sibcall_internal): Likewise.
23242 (sibcall_value_internal): Likewise.
23243 (prefetch): Use constraint ZD.
23244 * config/mips/mips.opt (minterlink-compressed): New option.
23245 (minterlink-mips16): Now an alias for minterlink-compressed.
23246 (mmicromips): New option.
23247 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23248 (compare_and_swap_12): Likewise.
23249 (sync_add<mode>): Likewise.
23250 (sync_<optab>_12): Likewise.
23251 (sync_old_<optab>_12): Likewise.
23252 (sync_new_<optab>_12): Likewise.
23253 (sync_nand_12): Likewise.
23254 (sync_old_nand_12): Likewise.
23255 (sync_new_nand_12): Likewise.
23256 (sync_sub<mode>): Likewise.
23257 (sync_old_add<mode>): Likewise.
23258 (sync_old_sub<mode>): Likewise.
23259 (sync_new_add<mode>): Likewise.
23260 (sync_new_sub<mode>): Likewise.
23261 (sync_<optab><mode>): Likewise.
23262 (sync_old_<optab><mode>): Likewise.
23263 (sync_new_<optab><mode>): Likewise.
23264 (sync_nand<mode>): Likewise.
23265 (sync_old_nand<mode>): Likewise.
23266 (sync_new_nand<mode>): Likewise.
23267 (sync_lock_test_and_set<mode>): Likewise.
23268 (test_and_set_12): Likewise.
23269 (atomic_compare_and_swap<mode>): Likewise.
23270 (atomic_exchange<mode>_llsc): Likewise.
23271 (atomic_fetch_add<mode>_llsc): Likewise.
23272 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23273 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23274 (umips_save_restore_pattern_p): Likewise.
23275 (umips_load_store_pair_p): Likewise.
23276 (umips_output_load_store_pair): Likewise.
23277 (umips_movep_target_p): Likewise.
23278 (umips_12bit_offset_address_p): Likewise.
23279 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23280 (mips_base_mips16): Rename this...
23281 (mips_base_compression_flags): ...to this. Update all uses.
23282 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23283 (mips_mips16_decl_p): Delete.
23284 (mips_nomips16_decl_p): Delete.
23285 (mips_get_compress_on_flags): New function.
23286 (mips_get_compress_off_flags): New function.
23287 (mips_get_compress_mode): New function.
23288 (mips_get_compress_on_name): New function.
23289 (mips_get_compress_off_name): New function.
23290 (mips_insert_attributes): Support multiple compression types.
23291 (mips_merge_decl_attributes): Likewise.
23292 (umips_12bit_offset_address_p): New function.
23293 (mips_start_function_definition): Emit .set micromips directive.
23294 (mips_call_may_need_jalx_p): New function.
23295 (mips_function_ok_for_sibcall): Add microMIPS support.
23296 (mips_print_operand_punctuation): Support short delay slots and
23298 (umips_swm_mask, umips_swm_encoding): New.
23299 (umips_build_save_restore): New function.
23300 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23301 (was_mips16_p): Remove.
23302 (old_compression_mode): New.
23303 (mips_set_compression_mode): New function.
23304 (mips_set_current_function): Add microMIPS support.
23305 (mips_option_override): Likewise.
23306 (umips_save_restore_pattern_p): New function.
23307 (umips_output_save_restore): New function.
23308 (umips_load_store_pair_p_1): New function.
23309 (umips_load_store_pair_p): New function.
23310 (umips_output_load_store_pair_1): New function.
23311 (umips_output_load_store_pair): New function.
23312 (umips_movep_target_p) New function.
23313 (mips_prepare_pch_save): Add microMIPS support.
23314 * config/mips/mips.h (TARGET_COMPRESSION): New.
23315 (TARGET_CPU_CPP_BUILTINS): Update macro
23316 to use new compression flags and to support microMIPS.
23317 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23318 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23319 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23320 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23321 (ASM_SPEC): Support mmicromips and mno-micromips.
23322 (M16STORE_REG_P): New macro.
23323 (MIPS_CALL): Support TARGET_MICROMIPS.
23324 (MICROMIPS_J): New macro.
23325 (mips_base_mips16): Rename this...
23326 (mips_base_compression_flags): ...to this.
23327 (UMIPS_12BIT_OFFSET_P): New macro.
23328 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23329 (MULTILIB_DIRNAMES): Likewise.
23330 2013-03-20 Richard Biener <rguenther@suse.de>
23332 PR tree-optimization/56661
23333 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23334 the result does not have to be distinct.
23336 2013-03-20 Richard Biener <rguenther@suse.de>
23338 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23341 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23342 Steven Bosscher <steven@gcc.gnu.org>
23344 PR rtl-optimization/56605
23345 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23347 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23350 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23351 that require movd instead of movq.
23353 2013-03-20 Richard Biener <rguenther@suse.de>
23355 * tree-ssa-structalias.c (struct variable_info): Add pointer
23356 to the first field of an aggregate with sub-vars. Make
23357 this and the pointer to the next subfield its ID.
23358 (vi_next): New function.
23359 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23360 storedanything_id, integer_id): Increment by one.
23361 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23362 get_call_clobber_vi): Adjust.
23363 (solution_set_expand): Simplify and speedup.
23364 (solution_set_add): Inline into ...
23365 (set_union_with_increment): ... this. Adjust accordingly.
23366 (do_sd_constraint): Likewise.
23367 (do_ds_constraint): Likewise.
23368 (do_complex_constraint): Simplify.
23369 (build_pred_graph): Adjust.
23370 (solve_graph): Likewise. Simplify and speedup.
23371 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23372 get_constraint_for_component_ref, get_constraint_for_1,
23373 first_vi_for_offset, first_or_preceding_vi_for_offset,
23374 create_function_info_for, create_variable_info_for_1,
23375 create_variable_info_for, intra_create_variable_infos): Adjust.
23376 (init_base_vars): Push NULL for ID zero.
23377 (compute_points_to_sets): Adjust.
23379 2013-03-20 Richard Biener <rguenther@suse.de>
23381 * cfgloop.c (verify_loop_structure): Streamline and avoid
23382 ICEing on corrupt loop tree.
23383 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23386 2013-03-20 Richard Biener <rguenther@suse.de>
23388 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23389 check whether an SSA update is needed.
23391 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23393 * config/mips/constraints.md (T): Rename to...
23397 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23398 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23400 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23402 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23403 (*subsi3_carryin_uxtw): Likewise.
23405 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23407 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23408 (*rorsi3_insn_uxtw): Likewise.
23410 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23412 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23413 (*extrsi5_insn_uxtw): Likewise.
23415 2013-03-19 Richard Biener <rguenther@suse.de>
23417 PR tree-optimization/56273
23418 * passes.c (init_optimization_passes): Move second VRP after DOM.
23420 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23422 * config/i386/i386.md (*movti_internal): Merge from
23423 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23424 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23425 nox64 isa attributes.
23427 2013-03-18 Richard Biener <rguenther@suse.de>
23429 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23431 (merge_node_constraints): Likewise.
23432 (build_succ_graph): Likewise.
23433 (valid_graph_edge): Inline into single caller.
23434 (unify_nodes): Likewise. Use bitmap_set_bit return value
23436 (scc_visit): Fix formatting and variable use.
23437 (do_sd_constraint): Use gcc_checking_assert.
23438 (do_ds_constraint): Likewise.
23439 (do_complex_constraint): Likewise.
23440 (condense_visit): Likewise. Cleanup.
23441 (dump_pred_graph): New function.
23442 (perform_var_substitution): Dump the pred-graph before
23443 variable substitution.
23444 (find_equivalent_node): Use gcc_checking_assert.
23445 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23447 2013-03-18 Richard Biener <rguenther@suse.de>
23449 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23450 Remove cond_expr_stmt_list argument and do not gimplify the
23452 (vect_loop_versioning): Adjust.
23453 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23454 Cleanup to use less temporaries.
23455 (vect_create_data_ref_ptr): Cleanup.
23457 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23459 PR tree-optimization/56635
23460 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23461 require types_compatible_p types.
23463 2013-03-18 Nick Clifton <nickc@redhat.com>
23465 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23466 spurious backslash.
23468 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23469 Add missing line to comment describing function.
23471 2013-03-18 Richard Biener <rguenther@suse.de>
23473 PR tree-optimization/56210
23474 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23475 Handle string / character search functions.
23476 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23478 2013-03-18 Richard Biener <rguenther@suse.de>
23480 PR middle-end/56483
23481 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23482 and implement properly.
23483 * gimple.h (gimple_cond_single_var_p): Remove.
23485 2013-03-18 Richard Biener <rguenther@suse.de>
23487 * tree-data-ref.h (find_data_references_in_loop): Declare.
23488 * tree-data-ref.c (get_references_in_stmt): Use a stack
23489 vector pre-allocated in the callers.
23490 (find_data_references_in_stmt): Adjust.
23491 (graphite_find_data_references_in_stmt): Likewise.
23492 (create_rdg_vertices): Likewise.
23493 (find_data_references_in_loop): Export.
23494 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23495 Compute dependences here...
23496 (vect_analyze_data_refs): ...not here. When we encounter
23497 a non-vectorizable data reference in basic-block vectorization
23498 truncate the data reference vector. Do not bother to
23499 fixup data-dependence information for gather loads.
23500 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23501 of data references, as reported.
23503 2013-03-18 Richard Biener <rguenther@suse.de>
23505 PR tree-optimization/3713
23506 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23507 has_constants and expr.
23508 (stmt_has_constants): Properly valueize SSA names when deciding
23509 whether the stmt has constants.
23511 2013-03-18 Richard Biener <rguenther@suse.de>
23513 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23514 whole function when there is nothing to do.
23515 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23516 * tree-vectorizer.c (vectorize_loops): Update virtual and
23517 loop-closed SSA once.
23518 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23520 2013-03-18 Richard Biener <rguenther@suse.de>
23522 PR middle-end/56113
23523 * domwalk.c (bb_postorder): New global static.
23524 (cmp_bb_postorder): New function.
23525 (walk_dominator_tree): Replace scheme imposing an order for
23526 visiting dominator sons by one sorting them at the time they
23527 are pushed on the stack.
23529 2013-03-18 Richard Biener <rguenther@suse.de>
23531 PR tree-optimization/39326
23532 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23533 (struct mem_ref): Replace mem member with ao_ref typed member.
23534 (MEM_ANALYZABLE): Adjust.
23535 (memref_eq): Likewise.
23536 (mem_ref_alloc): Likewise.
23537 (gather_mem_refs_stmt): Likewise.
23538 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23539 (execute_sm_if_changed_flag_set): Adjust.
23540 (execute_sm): Likewise.
23541 (ref_always_accessed_p): Likewise.
23542 (refs_independent_p): Likewise.
23543 (can_sm_ref_p): Likewise.
23545 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23548 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23549 return 1 even for !unsignedp.
23551 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23553 * config/i386/i386.md (isa): Add x64 and nox64.
23554 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23555 (*pushtf): Enable *roF alternative for x64 isa only.
23556 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23557 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23558 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23559 (*movtf_internal): Merge from *movtf_internal_rex64 and
23560 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23561 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23562 nox64 isa attributes.
23563 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23564 nox64 isa attributes.
23565 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23567 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23569 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23571 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23574 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23575 class names. Remove trailing comma after "ALL_REGS".
23577 2013-03-16 Jan Hubicka <jh@suse.cz>
23579 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23580 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23581 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23582 of cgraph_get_create_node.
23583 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23585 2013-03-16 Jason Merrill <jason@redhat.com>
23588 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23589 with DW_AT_default_value.
23591 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23593 * BASE-VER: Set to 4.9.0.
23595 2013-03-14 Andi Kleen <ak@linux.intel.com>
23598 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23599 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23600 Document _x* TSX intrinsics.
23602 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23603 David Holsgrove <david.holsgrove@xilinx.com>
23605 * configure.ac: Add MicroBlaze TLS support detection.
23606 * configure: Regenerate.
23607 * config/microblaze/microblaze-protos.h
23608 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23609 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23610 * config/microblaze/microblaze.c (microblaze_address_type): Add
23611 ADDRESS_TLS and tls_reloc address types.
23612 (microblaze_address_info): Add tls_reloc.
23613 (TARGET_HAVE_TLS): Define.
23614 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23615 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23616 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23617 load_tls_operand, microblaze_call_tls_get_addr,
23618 microblaze_legitimize_tls_address): New functions.
23619 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23620 (get_base_reg): Use microblaze_tls_symbol_p.
23621 (microblaze_classify_address): Handle TLS.
23622 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23623 label_mentioned_p and microblaze_tls_referenced_p.
23624 (microblaze_legitimize_address): Handle TLS.
23625 (microblaze_address_insns): Handle ADDRESS_TLS.
23626 (pic_address_needs_scratch): Handle TLS.
23627 (print_operand_address): Handle TLS.
23628 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23629 (microblaze_expand_move): Handle TLS.
23630 (microblaze_legitimate_constant_p): Check
23631 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23632 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23633 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23634 (PIC_OFFSET_TABLE_REGNUM): Set.
23635 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23636 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23637 (addsi3, movsi_internal2, movdf_internal): Update constraints
23638 * config/microblaze/predicates.md (arith_plus_operand): Define
23639 (move_operand): Redefine as move_src_operand,
23640 check microblaze_tls_referenced_p.
23642 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23644 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23645 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23647 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23649 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23652 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23654 PR tree-optimization/53265
23655 * common.opt (Waggressive-loop-optimizations): New option.
23656 * tree-ssa-loop-niter.c: Include tree-pass.h.
23657 (do_warn_aggressive_loop_optimizations): New function.
23658 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23659 if number_of_latch_executions returned constant.
23660 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23661 early. If number_of_latch_executions returned constant, set
23662 nb_iterations_upper_bound back to it.
23663 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23665 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23666 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23668 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23669 (MULTILIB_OSDIRNAMES): Set.
23670 * genmultilib: If defaultosdirname doesn't start with :: , set
23671 defaultosdirname2 instead, clear it and emit two . multilib_raw
23672 entries instead of just one.
23674 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23676 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23677 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23678 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23679 (SUBTARGET_OVERRIDE_OPTIONS): New.
23681 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23684 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23685 (musermode): Convert to Var(TARGET_USERMODE).
23686 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23687 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23688 * config/sh/sh.c (sh_option_override): Use
23689 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23690 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23692 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23694 (udivsi3_i4_single, divsi3_i4_single): Use
23695 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23697 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
23699 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23702 2013-03-13 Richard Biener <rguenther@suse.de>
23704 PR tree-optimization/56608
23705 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23706 calls when vectorizing basic-blocks.
23708 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23711 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23714 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23716 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23718 2013-03-11 Jan Hubicka <jh@suse.cz>
23721 * lto-streamer-out.c (output_symbol_p): Skip references from
23722 constructors of external variables.
23724 2013-03-11 Jan Hubicka <jh@suse.cz>
23726 PR middle-end/56571
23727 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23729 * emit-rtl.c (verify_rtx_sharing): Likewise.
23730 (copy_insn_1): Likewise.
23731 * rtl.c (copy_rtx): Likewise.
23733 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23736 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23737 output_operand_lossage message.
23739 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23742 * arm.c (shift_op): Validate RTL pattern on the fly.
23743 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23746 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23749 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23750 new call patterns that clobber %fr12.
23751 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23752 split and postreload patterns.
23753 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23754 registers %fr12 and %fr12R as call used.
23756 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23758 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23759 (canon_address, record_store, replace_read, check_mem_read_rtx,
23760 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23761 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23762 rest_of_handle_dse): Likewise.
23764 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23766 PR middle-end/56524
23767 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23769 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23770 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23771 (save_optabs_if_changed): Replace with...
23772 (init_tree_optimization_optabs): ...this.
23773 * optabs.c (save_optabs_if_changed): Rename to...
23774 (init_tree_optimization_optabs): ...this. Take the optimization node
23775 as argument. Do nothing if the base optabs are already correct.
23776 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23777 to recompute optabs.
23778 * function.h (function): Remove optabs field.
23779 * function.c (invoke_set_current_function_hook): Call
23780 init_tree_optimization_optabs. Use the result to initialize
23783 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23785 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23786 if GTMA_HAS_NO_INSTRUMENTATION.
23787 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23788 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23789 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23790 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23791 GTMA_HAS_NO_INSTRUMENTATION.
23793 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23795 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23798 2013-03-08 Marek Polacek <polacek@redhat.com>
23799 Jakub Jelinek <jakub@redhat.com>
23801 PR tree-optimization/56478
23802 * predict.c (is_comparison_with_loop_invariant_p): Change the
23803 type of loop_step to tree.
23804 (predict_loops): Adjust.
23805 (predict_iv_comparison): Perform the computations on double_ints.
23807 2013-03-08 Richard Biener <rguenther@suse.de>
23809 PR tree-optimization/56570
23810 * tree-cfg.c (verify_expr_location_1): Verify locations for
23812 * tree-sra.c (create_access_replacement): Strip locations
23813 from DECL_DEBUG_EXPRs.
23815 2013-03-08 Richard Biener <rguenther@suse.de>
23817 * tree-inline.c (expand_call_inline): Do not associate
23818 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23819 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23821 2013-03-08 Richard Biener <rguenther@suse.de>
23823 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23824 or block changes with -Og. Fix for location / block encoding
23825 changes and PHI arguments with locations.
23827 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23829 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23831 (struct output_info): Likewise.
23832 (register_overhead): Remove bad gcc_assert.
23833 (bitmap_find_bit): If there is only a single bitmap element, do not
23834 count a miss as a search.
23835 (print_statistics): Update for counter type changes.
23836 (dump_bitmap_statistics): Likewise. Print headers such that they
23837 are properly lined up with the printed counters.
23839 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23841 PR tree-optimization/56559
23842 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23843 check that it has only a single use.
23845 2013-03-07 Richard Biener <rguenther@suse.de>
23847 * doc/invoke.texi (fwhole-program): Discourage use in combination
23850 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23852 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23854 PR tree-optimization/56539
23855 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23856 instead of GSI_CONTINUE_LINKING as last argument to
23857 force_gimple_operand_gsi. Adjust function comment.
23859 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23862 PR middle-end/56548
23863 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23864 promoted mode, convert the result back to the original mode.
23866 2013-03-06 Richard Biener <rguenther@suse.de>
23868 PR middle-end/56294
23869 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23870 (insert_updated_phi_nodes_compare_uids): New function.
23871 (update_ssa): Sort symbols_to_rename after UID before
23872 traversing it to insert PHI nodes.
23874 2013-03-06 Richard Biener <rguenther@suse.de>
23876 PR middle-end/50494
23877 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23878 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23881 2013-02-13 Richard Biener <rguenther@suse.de>
23884 * varasm.c (output_constant_def_1): Get the decl representing
23885 the constant as argument.
23886 (output_constant_def): Wrap output_constant_def_1.
23887 (make_decl_rtl): Use output_constant_def_1 with the decl
23888 representing the constant.
23889 (build_constant_desc): Optionally re-use a decl already
23890 representing the constant.
23891 (tree_output_constant_def): Adjust.
23893 2013-03-06 Joey Ye <joey.ye@arm.com>
23896 * gcc.c (convert_white_space): New function.
23897 (main): Handles white space in function name.
23899 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23902 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23903 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23904 to SH_DIV_CALL_TABLE for TARGET_SH2.
23905 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23907 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23908 call-table options.
23910 2013-03-05 Sterling Augustine <saugustine@google.com>
23911 Cary Coutant <ccoutant@google.com>
23914 * dwarf2out.c (resolve_addr): Don't call
23915 remove_loc_list_addr_table_entries a second time for the same
23918 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23921 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23922 (avoid_complex_debug_insns): New function.
23923 (expand_debug_locations): Call it.
23925 PR rtl-optimization/56484
23926 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23927 lifetimes of hard registers on small register class machines.
23929 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23931 * config/microblaze/microblaze-protos.h: Rename
23932 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23933 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23935 (microblaze_fast_interrupt_function_p): New function.
23936 (microblaze_is_interrupt_handler): Rename to
23937 microblaze_is_interrupt_variant and add fast_interrupt check.
23938 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23939 (save_restore_insns): Likewise.
23940 (compute_frame_size): Likewise.
23941 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23942 (microblaze_globalize_label): Likewise.
23943 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23944 * config/microblaze/microblaze.md: Use wrapper
23945 microblaze_is_interrupt_variant.
23947 2013-03-05 Kai Tietz <ktietz@redhat.com>
23949 * sdbout.c (sdbout_one_type): Switch to current function's section
23950 supporting cold/hot.
23952 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23954 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23957 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23959 PR middle-end/56461
23960 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23961 if VALGRIND_GET_VBITS is defined, temporarily make object
23962 memory all defined, and restore previous valgrind addressability
23963 and definability afterwards. Free this_object at the end.
23965 PR middle-end/56461
23966 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23967 right before calling lra_create_live_ranges, also call it
23968 when clearing live_p. Only call lra_clear_live_ranges
23969 at the end if live_p.
23971 PR middle-end/56461
23972 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
23974 2013-03-05 Richard Biener <rguenther@suse.de>
23976 PR tree-optimization/56521
23977 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
23980 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
23983 * except.h (remove_unreachable_eh_regions): New prototype.
23984 * except.c (remove_eh_handler_splicer): New function, split out
23985 of remove_eh_handler.
23986 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
23987 warning about running it on many EH regions one at a time.
23988 (remove_unreachable_eh_regions_worker): New function, walk the
23989 EH tree in depth-first order and remove non-marked regions.
23990 (remove_unreachable_eh_regions): New function.
23991 * tree-eh.c (mark_reachable_handlers): New function, split out
23992 from remove_unreachable_handlers.
23993 (remove_unreachable_handlers): Use mark_reachable_handlers and
23994 remove_unreachable_eh_regions.
23995 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
23996 and remove_unreachable_eh_regions.
23998 2013-03-05 Richard Biener <rguenther@suse.de>
24000 PR middle-end/56525
24001 * loop-init.c (fix_loop_structure): Remove loops in two stages,
24002 not freeing them until the end.
24004 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24006 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
24008 2013-03-05 Richard Biener <rguenther@suse.de>
24010 PR tree-optimization/56270
24011 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
24012 of loads after scheduling an SLP instance.
24014 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24016 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
24018 (check_gcc_parallelize): Run guality.exp as a separate job from
24019 vect.exp with unsorted.exp and $(dg_target_exps) separately from
24020 struct-layout-1.exp with stackalign.exp.
24022 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
24024 PR middle-end/56461
24025 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
24026 load_index sbitmap even if some bit in it isn't set.
24028 PR middle-end/56461
24029 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
24030 (discover_iteration_bound_by_body_walk): Change queues to
24031 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
24032 spelling in comment. Call safe_push on queues[bound_index] directly.
24033 Release queues[queue_index] in every iteration unconditionally.
24034 Release bounds vector.
24036 PR middle-end/56461
24037 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
24038 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
24039 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
24042 2013-03-05 Richard Biener <rguenther@suse.de>
24045 * tree-inline.c (remap_blocks_to_null): New function.
24046 (expand_call_inline): When expanding a call stmt without
24047 an associated block inline remap all callee blocks to NULL.
24049 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24051 PR rtl-optimization/56494
24052 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
24053 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
24054 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
24056 PR middle-end/56461
24057 * sel-sched-ir.c (free_sched_pools): Release
24058 succs_info_pool.stack[succs_info_pool.max_top] vectors too
24059 if succs_info_pool.max_top isn't -1.
24062 * opts.c (opts_obstack, opts_concat): Moved to...
24063 * opts-common.c (opts_obstack, opts_concat): ... here.
24065 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24067 PR middle-end/56461
24068 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
24070 2013-03-04 Martin Jambor <mjambor@suse.cz>
24072 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
24073 all appropriate places.
24075 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
24077 PR tree-optimization/56424
24078 * ipa-split.c (split_function): Do not set the RSO flag if result is
24079 not by reference and its type is a register type.
24081 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
24083 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
24084 (microblaze_legitimate_pic_operand): Likewise
24085 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
24086 new function microblaze_legitimate_pic_operand
24087 * config/microblaze/microblaze-protos.h
24088 (microblaze_legitimate_pic_operand): Declare.
24090 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24092 * config/microblaze/predicates.md (call_insn_simple_operand):
24093 New predicate for supported rtx code types.
24094 * config/microblaze/microblaze.md (call_internal1): Use
24095 call_insn_simple_operand predicate.
24097 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24099 PR middle-end/56461
24100 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
24101 partitions.ordered_remove.
24103 PR middle-end/56461
24104 * tree-vect-stmts.c (vectorizable_conversion): Don't call
24105 vec_oprnds0.create (1) for modifier == NONE.
24107 PR middle-end/56461
24108 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24109 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24110 vec_oprnds1 right before pushing anything to it for
24113 PR middle-end/56461
24114 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24115 set nbbs to 0 instead of having separate code path.
24116 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24117 instead of false as last argument if returning NULL.
24119 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24121 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24122 the attribute is now called "target" instead of "option".
24123 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24124 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24125 attribute/pragma name for TARGET_OPTION_VALID_P and
24126 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24127 * doc/tm.texi: Regenerated.
24129 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24131 * config/microblaze/microblaze.c:
24132 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24133 * config/microblaze/microblaze.h: Add -mxl-reorder to
24135 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24136 instructions emitted if TARGET_REORDER.
24137 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24138 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24141 2013-03-01 Xinliang David Li <davidxl@google.com>
24143 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24146 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24148 PR middle-end/56461
24149 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24150 vector even when returning true. Fix up function comment formatting.
24152 PR middle-end/56461
24153 * ira-build.c (ira_loop_nodes_count): New variable.
24154 (create_loop_tree_nodes): Initialize it.
24155 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24157 PR middle-end/56461
24158 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24159 method on dr_chain and result_chain.
24160 * tree-vect-stmts.c (vectorizable_store): Only call
24161 result_chain.create if j == 0.
24163 PR middle-end/56461
24164 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24165 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24166 before overwriting it.
24168 2013-03-01 Tobias Burnus <burnus@net-b.de>
24170 * doc/extended.texi (C Extensions): Change order in @menu
24172 (Other MIPS Built-in Functions): Move last MIPS entry before
24173 "picoChip Built-in Functions".
24174 (SH Built-in Functions): Move after RX Built-in Functions.
24175 * doc/gcc.texi (Introduction): Change order in @menu
24177 * doc/md.texi (Constraints): Ditto.
24178 * gty.texi (Type Information): Ditto.
24179 (User-provided marking routines for template types): Make
24181 * doc/invoke.texi (AArch64 Options): Move before
24182 "Adapteva Epiphany Options".
24184 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24185 Jakub Jelinek <jakub@redhat.com>
24188 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24189 no_address_safety_analysis attribute.
24190 * doc/extend.texi (no_address_safety_attribute): Rename to
24191 no_sanitize_address attribute, mention no_address_safety_analysis
24192 attribute as deprecated alias.
24194 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24196 PR middle-end/56461
24197 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24198 type to vec<vec<tree> > *.
24199 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24200 to be vec<tree> instead of vec<tree> *, set vec_defs
24201 to vNULL and call vec_defs.create (number_of_vects), adjust other
24203 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24204 vectorizable_condition): Adjust vect_get_slp_defs callers.
24206 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24208 * config/aarch64/aarch64.c
24209 (aarch64_float_const_representable): Remove unused variable.
24211 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24213 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24215 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24217 * config/aarch64/aarch64-builtins.c
24218 (aarch64_init_simd_builtins): Make static.
24220 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24222 * config/aarch64/aarch64.c
24223 (aarch64_simd_make_constant): Make static.
24225 2013-02-28 Martin Jambor <mjambor@suse.cz>
24227 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24228 with no initialization to the RHS of debug statements.
24230 2013-02-28 Martin Jambor <mjambor@suse.cz>
24232 PR tree-optimization/56294
24233 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24235 (get_access_replacement): Do not call create_access_replacement.
24236 Assert a replacement exists.
24237 (get_repl_default_def_ssa_name): Create the replacement declaration
24240 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24242 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24243 final_end_function.
24245 2013-02-28 Marek Polacek <polacek@redhat.com>
24247 PR rtl-optimization/56466
24248 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24249 if we're changing a loop.
24250 (peel_loops_completely): Likewise.
24252 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24255 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24257 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24260 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24261 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24262 INTX_FTYPE_FX, FX_FTYPE_INTX.
24263 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24265 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24267 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24268 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24269 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24270 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24271 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24272 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24273 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24274 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24275 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24276 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24277 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24278 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24279 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24280 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24281 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24282 (avrxmega6): Increase max flash segments from 5 to 6.
24283 * config/avr/t-multilib: Regenerate.
24284 * config/avr/avr-tables.opt: Regenerate.
24285 * doc/avr-mmcu.texi: Regenerate.
24287 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24289 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24290 (avr_device_to_arch): Rename to avr_device_to_ld.
24291 (avr_device_to_as): New prototype.
24292 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24293 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24294 * config/avr/driver-avr.c (avr_device_to_as): New.
24295 (avr_device_to_arch): Rename to avr_device_to_ld.
24297 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24299 PR middle-end/56461
24300 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24301 method on dr_chain and result_chain.
24303 PR middle-end/56461
24304 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24305 pointer_set_destroy on not_executed_last_iteration.
24307 PR middle-end/56461
24308 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24310 PR middle-end/56461
24311 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24312 FOR_EACH_DEFINED_FUNCTION when freeing state.
24314 PR middle-end/56461
24315 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24317 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24320 PR middle-end/56461
24321 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24322 known_aggs[i].items and release known_aggs vector.
24324 PR middle-end/56461
24325 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24327 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24329 * config/microblaze/microblaze.c (microblaze_emit_compare):
24330 Use xor for EQ/NE comparisions.
24331 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24332 (cbranchsf4): Adjust operator to comparison_operator.
24334 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24336 PR middle-end/56461
24337 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24339 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24340 vec_safe_push, always update *slot.
24341 (redirect_edge_var_map_clear): Use vec_free.
24342 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24343 (free_var_map_entry): Use vec_free.
24344 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24345 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24347 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24349 PR middle-end/45472
24350 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24351 when the may_trap_p bit of the exprs being merged differs.
24352 Reorder tests for speculativeness in the logical and operator.
24354 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24356 * incpath.c (add_standard_paths): Use reconcat instead of concat
24357 where appropriate and avoid leaking memory.
24359 * opts.h: Include obstack.h.
24360 (opts_concat): New prototype.
24361 (opts_obstack): New declaration.
24362 * opts.c (opts_concat): New function.
24363 (opts_obstack): New variable.
24364 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24365 (finish_options): Use opts_concat instead of concat
24366 and XOBNEWVEC instead of XNEWVEC.
24367 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24368 generate_option): Likewise.
24369 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24370 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24373 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24374 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24376 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24378 PR middle-end/56461
24379 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24381 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24383 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24384 (arm_block_move_unaligned_straight): Likewise.
24385 (arm_adjust_block_mem): Likewise.
24387 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24390 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24391 temp, cond and label.
24392 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24395 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24396 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24399 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24400 of prologue/epilogue functions.
24403 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24404 Remove unused variable cfa_offset.
24405 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24408 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24409 type promotion to unsigned.
24412 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24413 for HOST_WIDE_INT of 32 bit / same size as int.
24414 (arm_block_move_unaligned_straight): Likewise.
24415 (arm_adjust_block_mem): Likewise.
24418 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24421 2013-02-26 Marek Polacek <polacek@redhat.com>
24423 PR tree-optimization/56426
24424 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24426 2013-02-26 Richard Biener <rguenther@suse.de>
24429 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24430 unused variable loops.
24432 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24434 PR tree-optimization/56448
24435 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24436 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24437 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24438 later operands of the references, or even first operand for
24439 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24441 PR tree-optimization/56443
24442 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24443 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24444 to type_for_mode langhook.
24446 2013-02-25 Matt Turner <mattst88@gmail.com>
24448 * doc/invoke.texi: Document r4700.
24450 2013-02-25 Richard Biener <rguenther@suse.de>
24452 PR tree-optimization/56175
24453 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24455 (simplify_bitwise_binary): ... here. Also guard the conversion
24456 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24458 2013-02-25 Catherine Moore <clm@codesourcery.com>
24461 2013-02-24 Catherine Moore <clm@codesourcery.com>
24462 Maciej W. Rozycki <macro@codesourcery.com>
24463 Tom de Vries <tom@codesourcery.com>
24464 Nathan Sidwell <nathan@codesourcery.com>
24465 Iain Sandoe <iain@codesourcery.com>
24466 Nathan Froyd <froydnj@codesourcery.com>
24467 Chao-ying Fu <fu@mips.com>
24469 * doc/extend.texi: (micromips, nomicromips, nocompression):
24470 Document new function attributes.
24471 * doc/invoke.texi (minterlink-compressed, mmicromips,
24472 m14k, m14ke, m14kec): Document new options.
24473 (minterlink-mips16): Update documentation.
24474 * doc/md.texi (ZC, ZD): Document new constraints.
24475 * configure.ac (gcc_cv_as_micromips): Check if linker
24476 supports the .set micromips directive.
24477 * configure: Regenerate.
24478 * config.in: Regenerate.
24479 * config/mips/mips-tables.opt: Regenerate.
24480 * config/mips/micromips.md: New file.
24481 * constraints.md (ZC, AD): New constraints.
24482 * config/mips/predicates.md (movep_src_register): New predicate.
24483 (movep_src_operand): New predicate.
24484 (non_volatile_mem_operand): New predicate.
24485 * config/mips/mips.md (multimem): New type.
24486 (length): Differentiate between 17-bit and 18-bit branch offsets.
24487 (MOVEP1, MOVEP2): New mode iterator.
24488 (mov_<load>l): Use ZC constraint.
24489 (mov_<load>r): Likewise.
24490 (mov_<store>l): Likewise.
24491 (mov_<store>r): Likewise.
24492 (*branch_equality<mode>_inverted): Add microMIPS support.
24493 (*branch_equality<mode>): Likewise.
24494 (*jump_absolute): Likewise.
24495 (indirect_jump_<mode>): Likewise.
24496 (tablejump_<mode>): Likewise.
24497 (<optab>_internal): Likewise.
24498 (sibcall_internal): Likewise.
24499 (sibcall_value_internal): Likewise.
24500 (prefetch): Use constraint ZD.
24501 * config/mips/mips.opt (minterlink-compressed): New option.
24502 (minterlink-mips16): Now an alias for minterlink-compressed.
24503 (mmicromips): New option.
24504 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24505 (compare_and_swap_12): Likewise.
24506 (sync_add<mode>): Likewise.
24507 (sync_<optab>_12): Likewise.
24508 (sync_old_<optab>_12): Likewise.
24509 (sync_new_<optab>_12): Likewise.
24510 (sync_nand_12): Likewise.
24511 (sync_old_nand_12): Likewise.
24512 (sync_new_nand_12): Likewise.
24513 (sync_sub<mode>): Likewise.
24514 (sync_old_add<mode>): Likewise.
24515 (sync_old_sub<mode>): Likewise.
24516 (sync_new_add<mode>): Likewise.
24517 (sync_new_sub<mode>): Likewise.
24518 (sync_<optab><mode>): Likewise.
24519 (sync_old_<optab><mode>): Likewise.
24520 (sync_new_<optab><mode>): Likewise.
24521 (sync_nand<mode>): Likewise.
24522 (sync_old_nand<mode>): Likewise.
24523 (sync_new_nand<mode>): Likewise.
24524 (sync_lock_test_and_set<mode>): Likewise.
24525 (test_and_set_12): Likewise.
24526 (atomic_compare_and_swap<mode>): Likewise.
24527 (atomic_exchange<mode>_llsc): Likewise.
24528 (atomic_fetch_add<mode>_llsc): Likewise.
24529 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24530 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24531 (umips_save_restore_pattern_p): Likewise.
24532 (umips_load_store_pair_p): Likewise.
24533 (umips_output_load_store_pair): Likewise.
24534 (umips_movep_target_p): Likewise.
24535 (umips_12bit_offset_address_p): Likewise.
24536 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24537 (mips_base_mips16): Rename this...
24538 (mips_base_compression_flags): ...to this. Update all uses.
24539 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24540 (mips_mips16_decl_p): Delete.
24541 (mips_nomips16_decl_p): Delete.
24542 (mips_get_compress_on_flags): New function.
24543 (mips_get_compress_off_flags): New function.
24544 (mips_get_compress_mode): New function.
24545 (mips_get_compress_on_name): New function.
24546 (mips_get_compress_off_name): New function.
24547 (mips_insert_attributes): Support multiple compression types.
24548 (mips_merge_decl_attributes): Likewise.
24549 (umips_12bit_offset_address_p): New function.
24550 (mips_start_function_definition): Emit .set micromips directive.
24551 (mips_call_may_need_jalx_p): New function.
24552 (mips_function_ok_for_sibcall): Add microMIPS support.
24553 (mips_print_operand_punctuation): Support short delay slots and
24555 (umips_swm_mask, umips_swm_encoding): New.
24556 (umips_build_save_restore): New function.
24557 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24558 (was_mips16_p): Remove.
24559 (old_compression_mode): New.
24560 (mips_set_compression_mode): New function.
24561 (mips_set_current_function): Add microMIPS support.
24562 (mips_option_override): Likewise.
24563 (umips_save_restore_pattern_p): New function.
24564 (umips_output_save_restore): New function.
24565 (umips_load_store_pair_p_1): New function.
24566 (umips_load_store_pair_p): New function.
24567 (umips_output_load_store_pair_1): New function.
24568 (umips_output_load_store_pair): New function.
24569 (umips_movep_target_p) New function.
24570 (mips_prepare_pch_save): Add microMIPS support.
24571 * config/mips/mips.h (TARGET_COMPRESSION): New.
24572 (TARGET_CPU_CPP_BUILTINS): Update macro
24573 to use new compression flags and to support microMIPS.
24574 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24575 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24576 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24577 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24578 (ASM_SPEC): Support mmicromips and mno-micromips.
24579 (M16STORE_REG_P): New macro.
24580 (MIPS_CALL): Support TARGET_MICROMIPS.
24581 (MICROMIPS_J): New macro.
24582 (mips_base_mips16): Rename this...
24583 (mips_base_compression_flags): ...to this.
24584 (UMIPS_12BIT_OFFSET_P): New macro.
24585 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24586 (MULTILIB_DIRNAMES): Likewise.
24588 2013-02-25 Tom de Vries <tom@codesourcery.com>
24590 PR rtl-optimization/56131
24591 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24592 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24593 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24595 2013-02-25 Tobias Burnus <burnus@net-b.de>
24597 * doc/invoke.texi (-fsanitize=): Move from optimization
24598 to debugging options.
24600 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24602 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24604 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24605 Alexander Monakov <amonakov@ispras.ru>
24607 PR middle-end/56077
24608 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24609 flush pending lists also on non-jumps. Adjust comment.
24611 2013-02-24 Catherine Moore <clm@codesourcery.com>
24612 Maciej W. Rozycki <macro@codesourcery.com>
24613 Tom de Vries <tom@codesourcery.com>
24614 Nathan Sidwell <nathan@codesourcery.com>
24615 Iain Sandoe <iain@codesourcery.com>
24616 Nathan Froyd <froydnj@codesourcery.com>
24617 Chao-ying Fu <fu@mips.com>
24619 * doc/extend.texi: (micromips, nomicromips, nocompression):
24620 Document new function attributes.
24621 * doc/invoke.texi (minterlink-compressed, mmicromips,
24622 m14k, m14ke, m14kec): Document new options.
24623 (minterlink-mips16): Update documentation.
24624 * doc/md.texi (ZC, ZD): Document new constraints.
24625 * configure.ac (gcc_cv_as_micromips): Check if linker
24626 supports the .set micromips directive.
24627 * configure: Regenerate.
24628 * config.in: Regenerate.
24629 * config/mips/mips-tables.opt: Regenerate.
24630 * config/mips/micromips.md: New file.
24631 * constraints.md (ZC, AD): New constraints.
24632 * config/mips/predicates.md (movep_src_register): New predicate.
24633 (movep_src_operand): New predicate.
24634 (non_volatile_mem_operand): New predicate.
24635 * config/mips/mips.md (multimem): New type.
24636 (length): Differentiate between 17-bit and 18-bit branch offsets.
24637 (MOVEP1, MOVEP2): New mode iterator.
24638 (mov_<load>l): Use ZC constraint.
24639 (mov_<load>r): Likewise.
24640 (mov_<store>l): Likewise.
24641 (mov_<store>r): Likewise.
24642 (*branch_equality<mode>_inverted): Add microMIPS support.
24643 (*branch_equality<mode>): Likewise.
24644 (*jump_absolute): Likewise.
24645 (indirect_jump_<mode>): Likewise.
24646 (tablejump_<mode>): Likewise.
24647 (<optab>_internal): Likewise.
24648 (sibcall_internal): Likewise.
24649 (sibcall_value_internal): Likewise.
24650 (prefetch): Use constraint ZD.
24651 * config/mips/mips.opt (minterlink-compressed): New option.
24652 (minterlink-mips16): Now an alias for minterlink-compressed.
24653 (mmicromips): New option.
24654 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24655 (compare_and_swap_12): Likewise.
24656 (sync_add<mode>): Likewise.
24657 (sync_<optab>_12): Likewise.
24658 (sync_old_<optab>_12): Likewise.
24659 (sync_new_<optab>_12): Likewise.
24660 (sync_nand_12): Likewise.
24661 (sync_old_nand_12): Likewise.
24662 (sync_new_nand_12): Likewise.
24663 (sync_sub<mode>): Likewise.
24664 (sync_old_add<mode>): Likewise.
24665 (sync_old_sub<mode>): Likewise.
24666 (sync_new_add<mode>): Likewise.
24667 (sync_new_sub<mode>): Likewise.
24668 (sync_<optab><mode>): Likewise.
24669 (sync_old_<optab><mode>): Likewise.
24670 (sync_new_<optab><mode>): Likewise.
24671 (sync_nand<mode>): Likewise.
24672 (sync_old_nand<mode>): Likewise.
24673 (sync_new_nand<mode>): Likewise.
24674 (sync_lock_test_and_set<mode>): Likewise.
24675 (test_and_set_12): Likewise.
24676 (atomic_compare_and_swap<mode>): Likewise.
24677 (atomic_exchange<mode>_llsc): Likewise.
24678 (atomic_fetch_add<mode>_llsc): Likewise.
24679 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24680 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24681 (umips_save_restore_pattern_p): Likewise.
24682 (umips_load_store_pair_p): Likewise.
24683 (umips_output_load_store_pair): Likewise.
24684 (umips_movep_target_p): Likewise.
24685 (umips_12bit_offset_address_p): Likewise.
24686 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24687 (mips_base_mips16): Rename this...
24688 (mips_base_compression_flags): ...to this. Update all uses.
24689 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24690 (mips_mips16_decl_p): Delete.
24691 (mips_nomips16_decl_p): Delete.
24692 (mips_get_compress_on_flags): New function.
24693 (mips_get_compress_off_flags): New function.
24694 (mips_get_compress_mode): New function.
24695 (mips_get_compress_on_name): New function.
24696 (mips_get_compress_off_name): New function.
24697 (mips_insert_attributes): Support multiple compression types.
24698 (mips_merge_decl_attributes): Likewise.
24699 (umips_12bit_offset_address_p): New function.
24700 (mips_start_function_definition): Emit .set micromips directive.
24701 (mips_call_may_need_jalx_p): New function.
24702 (mips_function_ok_for_sibcall): Add microMIPS support.
24703 (mips_print_operand_punctuation): Support short delay slots and
24705 (umips_swm_mask, umips_swm_encoding): New.
24706 (umips_build_save_restore): New function.
24707 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24708 (was_mips16_p): Remove.
24709 (old_compression_mode): New.
24710 (mips_set_compression_mode): New function.
24711 (mips_set_current_function): Add microMIPS support.
24712 (mips_option_override): Likewise.
24713 (umips_save_restore_pattern_p): New function.
24714 (umips_output_save_restore): New function.
24715 (umips_load_store_pair_p_1): New function.
24716 (umips_load_store_pair_p): New function.
24717 (umips_output_load_store_pair_1): New function.
24718 (umips_output_load_store_pair): New function.
24719 (umips_movep_target_p) New function.
24720 (mips_prepare_pch_save): Add microMIPS support.
24721 * config/mips/mips.h (TARGET_COMPRESSION): New.
24722 (TARGET_CPU_CPP_BUILTINS): Update macro
24723 to use new compression flags and to support microMIPS.
24724 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24725 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24726 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24727 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24728 (ASM_SPEC): Support mmicromips and mno-micromips.
24729 (M16STORE_REG_P): New macro.
24730 (MIPS_CALL): Support TARGET_MICROMIPS.
24731 (MICROMIPS_J): New macro.
24732 (mips_base_mips16): Rename this...
24733 (mips_base_compression_flags): ...to this.
24734 (UMIPS_12BIT_OFFSET_P): New macro.
24735 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24736 (MULTILIB_DIRNAMES): Likewise.
24738 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24741 * target-globals.c (save_target_globals): For init_reg_sets and
24742 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24744 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
24746 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24747 * config/aarch64/t-aarch64
24748 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24750 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24752 PR inline-asm/56148
24753 * lra-constraints.c (process_alt_operands): Reload operand
24754 conflicting with earlier clobber only if no more other conflicting
24757 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24760 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24761 if not linking a shared library.
24763 2013-02-22 Seth LaForge <sethml@google.com>
24765 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24767 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24769 * config/arm/arm.md (split for extendsidi): Update condition.
24770 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24771 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24772 (qhs_zextenddi_cstr): Likewise.
24774 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24776 PR middle-end/56420
24777 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24778 avoid signed wrapping.
24779 (expand_mult): Handle properly multiplication by
24780 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24781 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24782 in the compiler if coeff is HOST_WIDE_INT_MIN.
24783 (expand_divmod): Don't make ext_op1 static, change it's type to
24784 uhwi. Avoid undefined behavior in -INTVAL (op1).
24786 PR rtl-optimization/50339
24787 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24789 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24790 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24791 into splitting_ashiftrt field.
24792 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24794 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24797 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24799 PR middle-end/56108
24800 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24801 are sure to go irrevocable.
24803 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24805 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24806 scalars are valid operands.
24808 2013-02-21 Martin Jambor <mjambor@suse.cz>
24810 PR tree-optimization/56310
24811 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24812 only matching indices and non-negative final offsets.
24813 (intersect_aggregates_with_edge): Pass src_idx to
24814 agg_replacements_to_vector. Pass src_idx insstead of index to
24815 intersect_with_agg_replacements.
24817 2013-02-21 Martin Jambor <mjambor@suse.cz>
24819 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24820 instead of hard-wired defaults.
24822 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24824 * doc/invoke.texi (MIPS Options): Update documentation of the
24825 floating-point multiply-accumulate instruction restrictions.
24827 2013-02-21 Kostya Serebryany <kcc@google.com>
24829 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24830 asan_shadow_offset on x86_64 linux.
24832 2013-02-21 Richard Biener <rguenther@suse.de>
24834 PR tree-optimization/56415
24836 2013-02-11 Richard Biener <rguenther@suse.de>
24838 PR tree-optimization/56273
24839 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24842 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24845 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24848 PR inline-asm/56405
24849 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24850 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24852 2013-02-20 Jan Hubicka <jh@suse.cz>
24854 PR tree-optimization/56265
24855 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24856 when target is referenced for first time.
24858 2013-02-20 Richard Biener <rguenther@suse.de>
24860 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24861 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24862 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24863 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24864 not return anything.
24865 (rename_ssa_copies): Do not remove unused locals.
24866 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24867 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24868 * passes.c (execute_function_todo): Do not schedule unused locals
24869 removal if cleanup_tree_cfg did something.
24870 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24871 about the number of removed locals.
24873 2013-02-20 Richard Biener <rguenther@suse.de>
24875 PR tree-optimization/56398
24876 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24878 2013-02-20 Martin Jambor <mjambor@suse.cz>
24880 PR tree-optimization/55334
24881 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24882 restricted pointers to arrays.
24884 2013-02-20 Richard Biener <rguenther@suse.de>
24885 Jakub Jelinek <jakub@redhat.com>
24887 PR tree-optimization/56396
24888 * tree-ssa-ccp.c (n_const_val): New static variable.
24889 (get_value): Return NULL for SSA names we don't have a lattice
24891 (ccp_initialize): Initialize n_const_val.
24892 * tree-ssa-copy.c (n_copy_of): New static variable.
24893 (init_copy_prop): Initialize n_copy_of.
24894 (get_value): Return NULL_TREE for SSA names we don't have a
24897 2013-02-20 Martin Jambor <mjambor@suse.cz>
24899 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24901 2013-02-20 Richard Biener <rguenther@suse.de>
24903 * genpreds.c (write_lookup_constraint): Do not compare first
24904 letter of the constraint again.
24906 2013-02-20 Richard Biener <rguenther@suse.de>
24908 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24910 (get_use_iv_cost): Terminate hashtable walk when coming across
24913 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24915 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24916 reassociation for avx2 targets.
24918 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24920 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24921 Add version check for v8.10.a to enable microblaze_has_clz
24922 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24923 version and TARGET_PATTERN_COMPARE check
24924 * config/microblaze/microblaze.md: New clzsi2 instruction
24926 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24928 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24929 function before branching.
24931 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24933 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24935 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24937 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24939 PR middle-end/55889
24940 * sel-sched.c: Include ira.h.
24941 (implicit_clobber_conflict_p): New function.
24942 (moveup_expr): Use it.
24943 * Makefile.in (sel-sched.o): Depend on ira.h.
24945 2013-02-19 Richard Biener <rguenther@suse.de>
24947 PR tree-optimization/56384
24948 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24949 (vn_hash_type): Split out from ...
24950 (vn_hash_constant_with_type): ... here.
24951 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24952 (vn_phi_eq): Compare types from vn_phi_s structure.
24953 (vn_phi_lookup): Populate vn_phi_s type.
24954 (vn_phi_insert): Likewise.
24956 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24958 PR tree-optimization/56350
24959 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24960 if haven't found reduction or nested cycle operand, rather than
24961 asserting we must find it.
24963 PR tree-optimization/56381
24964 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24967 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24968 Jakub Jelinek <jakub@redhat.com>
24971 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24972 (swap_optab_enable): Same.
24973 (init_all_optabs): Use argument instead of global.
24974 * tree.h (struct tree_optimization_option): New field target_optabs.
24975 * expr.h (init_all_optabs): Add argument to prototype.
24976 (TREE_OPTIMIZATION_OPTABS): New.
24977 (save_optabs_if_changed): Protoize.
24978 * optabs.h: Declare this_fn_optabs.
24979 * optabs.c (save_optabs_if_changed): New.
24980 Declare this_fn_optabs.
24981 (init_optabs): Add argument to init_all_optabs() call.
24982 * function.c (invoke_set_current_function_hook): Handle per
24984 * function.h (struct function): New field optabs.
24985 * config/mips/mips.c (mips_set_mips16_mode): Handle when
24986 optimization_current_node has changed.
24987 * target-globals.h (save_target_globals_default_opts): Protoize.
24988 * target-globals.c (save_target_globals_default_opts): New.
24990 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24993 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
24994 registers %fr12 and %fr12R as call used.
24997 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
24998 and HImode, require all displacements to be an integer multiple of
25000 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
25001 only allow QImode and HImode when reload is in progress and strict is
25002 true. Likewise for symbolic addresses. Use base14_operand to check
25003 displacements in REG+BASE addresses.
25005 2013-02-18 Richard Biener <rguenther@suse.de>
25007 PR tree-optimization/56366
25008 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25009 handle sign-conversion of outer-loop initial induction value.
25011 2013-02-18 Richard Biener <rguenther@suse.de>
25013 PR middle-end/56349
25014 * cfghooks.c (merge_blocks): If we merge a latch into another
25015 block adjust references to it.
25016 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
25017 (verify_loop_structure): Verify that a recorded latch is in fact
25020 2013-02-18 Richard Biener <rguenther@suse.de>
25022 PR tree-optimization/56321
25023 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
25024 order SSA name release and virtual operand unlinking.
25026 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25028 * config/microblaze/microblaze.md (save_stack_block): Define.
25029 (restore_stack_block): Likewise.
25031 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25033 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
25034 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
25035 * config/microblaze/microblaze.c (microblaze_option_override):
25036 Bail out early for PIC modes when target does not support PIC.
25038 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25040 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
25041 Replace with a microblaze version.
25042 (microblaze_trampoline_init): Adapt for microblaze.
25043 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
25046 2013-02-16 Jakub Jelinek <jakub@redhat.com>
25047 Dodji Seketeli <dodji@redhat.com>
25050 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
25051 (instrument_mem_region_access): Do not forget to always put
25052 instrumentation of the of 'base' and 'base + len' in a "if (len !=
25053 0) statement, even for cases where either 'base' or 'base + len'
25054 are not instrumented -- because they have been previously
25055 instrumented. Simplify the logic by putting all the statements
25056 instrument 'base + len' inside a sequence, and then insert that
25057 sequence right before the current insertion point. Then, to
25058 instrument 'base + len', just get an iterator on that statement.
25059 And do not forget to update the pointer to iterator the function
25060 received as argument.
25062 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
25064 PR rtl-optimization/56348
25065 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
25067 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
25069 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
25070 (clean_graph_dump_file): Pass base to start_graph_dump.
25072 2013-02-14 Richard Henderson <rth@redhat.com>
25075 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
25077 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
25079 * collect2-aix.h: Define F_LOADONLY.
25081 2013-02-14 Richard Biener <rguenther@suse.de>
25084 * varasm.c (output_constant_def_1): Get the decl representing
25085 the constant as argument.
25086 (output_constant_def): Wrap output_constant_def_1.
25087 (make_decl_rtl): Use output_constant_def_1 with the decl
25088 representing the constant.
25089 (build_constant_desc): Optionally re-use a decl already
25090 representing the constant.
25091 (tree_output_constant_def): Adjust.
25093 2013-02-14 Dodji Seketeli <dodji@redhat.com>
25096 * asan.c (instrument_builtin_call): Really put the length of the
25097 second source argument into src1_len.
25099 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25101 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
25102 argument. If it is false, don't create edge from then_bb to
25104 (insert_if_then_before_iter): Pass true to it.
25105 (build_check_stmt): Pass false to it.
25106 (transform_statements): Flush hash table only on extended basic
25107 block boundaries, rather than at the beginning of every bb.
25108 Don't flush hash table on nonfreeing_call_p calls.
25109 * tree-flow.h (nonfreeing_call_p): New prototype.
25110 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25112 2013-02-13 David S. Miller <davem@davemloft.net>
25114 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25116 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25119 * ira.c (max_regno_before_ira): Move from ...
25121 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25124 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25126 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25128 2013-02-13 Richard Biener <rguenther@suse.de>
25131 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25132 globals in MEM_REFs.
25134 2013-02-13 Richard Biener <rguenther@suse.de>
25136 * loop-init.c (loop_optimizer_init): Clear loop state when
25137 re-initializing preserved loops.
25138 * loop-unswitch.c (unswitch_single_loop): Return whether
25139 we unswitched the loop. Do not verify loop state here.
25140 (unswitch_loops): When we unswitched a loop discover new loops.
25142 2013-02-13 Kostya Serebryany <kcc@google.com>
25144 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25146 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25148 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25150 Avoid instrumenting duplicated memory access in the same basic block
25151 * Makefile.in (asan.o): Add new dependency on hash-table.h
25152 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25153 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25154 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25155 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25156 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25157 (get_mem_ref_of_assignment): New functions.
25158 (get_mem_refs_of_builtin_call): Extract from
25159 instrument_builtin_call and tweak a little bit to make it fit with
25161 (instrument_builtin_call): Use the new
25162 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25163 of is_gimple_builtin_call.
25164 (instrument_derefs, instrument_mem_region_access): Insert the
25165 instrumented memory reference into the hash table.
25166 (maybe_instrument_assignment): Renamed instrument_assignment into
25167 this, and change it to advance the iterator when instrumentation
25168 actually happened and return true in that case. This makes it
25169 homogeneous with maybe_instrument_assignment, and thus give a
25170 chance to callers to be more 'regular'.
25171 (transform_statements): Clear the memory reference hash table
25172 whenever we enter a new BB, when we cross a function call, or when
25173 we are done transforming statements. Use
25174 maybe_instrument_assignment instead of instrumentation. No more
25175 need to special case maybe_instrument_assignment and advance the
25176 iterator after calling it; it's now handled just like
25177 maybe_instrument_call. Update comment.
25179 2013-02-13 Richard Biener <rguenther@suse.de>
25181 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25182 Fix loop discovery code.
25184 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25186 PR inline-asm/56148
25187 * lra-constraints.c (process_alt_operands): Match early clobber
25188 operand with itself. Check conflicts with earlyclobber only if
25189 the operand is not reloaded. Prefer to reload conflicting operand
25190 if earlyclobber and matching operands are the same.
25192 2013-02-12 Richard Biener <rguenther@suse.de>
25195 * lto-streamer-out.c (write_symbol): Do not output symbols
25196 for hard register variables.
25198 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25201 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25202 (umulsidi3_insn, mulsidi3_insn): New insns.
25204 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25206 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25207 (struct tune_params): Add vec_costs field.
25208 * config/arm/arm.c (arm_builtin_vectorization_cost)
25209 (arm_add_stmt_cost): New functions.
25210 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25211 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25212 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25213 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25214 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25215 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25216 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25218 2013-02-12 Richard Biener <rguenther@suse.de>
25221 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25222 decls again if possible.
25224 2013-02-12 Richard Biener <rguenther@suse.de>
25226 PR middle-end/56288
25227 * tree-ssa.c (verify_ssa_name): Fix check, move
25228 SSA_NAME_IN_FREE_LIST check up.
25230 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25231 Steven Bosscher <steven@gcc.gnu.org>
25233 PR rtl-optimization/56151
25234 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25235 equal to op0 or op1, and last_insn pattern is CODE operation
25236 with MEM dest and one of the operands matches that MEM.
25238 2013-02-11 Sriraman Tallam <tmsriram@google.com>
25240 * doc/extend.texi: Document Function Multiversioning and "default"
25241 parameter string to target attribute.
25242 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25243 target attribute parameter is "default".
25244 (ix86_compare_version_priority): Remove checks for target attribute.
25245 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25246 Remove check for target attribute equal to NULL. Add assert.
25247 (ix86_generate_version_dispatcher_body): Change error to sorry.
25249 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25250 Jack Howarth <howarth@bromo.med.uc.edu>
25251 Patrick Marlier <patrick.marlier@gmail.com>
25254 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25255 define ENDFILE_SPEC as TM_DESTRUCTOR.
25256 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25258 2013-02-11 Alexander Potapenko <glider@google.com>
25259 Jack Howarth <howarth@bromo.med.uc.edu>
25260 Jakub Jelinek <jakub@redhat.com>
25263 * config/darwin.c (cdtor_record): Rename ctor_record.
25264 (sort_cdtor_records): Rename sort_ctor_records.
25265 (finalize_dtors): New routine to sort destructors by
25266 priority before use in assemble_integer.
25267 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25269 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25271 PR rtl-optimization/56275
25272 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25273 offset is non-negative and less than cmode size before
25274 calling simplify_subreg.
25276 2013-02-11 Richard Biener <rguenther@suse.de>
25278 PR tree-optimization/56264
25279 * cfgloop.h (fix_loop_structure): Adjust prototype.
25280 * loop-init.c (fix_loop_structure): Return the number of
25281 newly discovered loops.
25282 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25283 are discovered, do a full loop-closed SSA rewrite.
25285 2013-02-11 Richard Biener <rguenther@suse.de>
25287 PR tree-optimization/56273
25288 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25290 (check_array_ref): Fix missing newline in dumps.
25291 (search_for_addr_array): Likewise.
25293 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25295 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25297 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25300 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25302 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25304 PR rtl-optimization/56246
25305 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25307 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25308 constraints are satisfied.
25310 2013-02-08 Jeff Law <law@redhat.com>
25313 * emit-rtl.c (reg_is_parm_p): New function.
25314 * regs.h (reg_is_parm_p): New prototype.
25315 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25316 callee-clobbered registers.
25318 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25321 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25322 If there is no implicit builtin declaration, just return NULL.
25324 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25326 * config/i386/sse.md (FMAMODEM): New mode iterator.
25327 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25328 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25330 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25332 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25333 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25334 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25336 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25338 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25339 (microblaze*-*-elf): Likewise.
25340 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25342 * config/microblaze/microblaze-c.c: Add builtin defines for
25343 _LITTLE_ENDIAN and _BIG_ENDIAN.
25344 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25345 add to TARGET_DEFAULT flags.
25346 Expand ASM_SPEC and LINK_SPEC.
25347 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25348 * config/microblaze/microblaze.md: Update extendsidi2 and
25349 movdi_internal instructions to use low-order / high-order reg
25351 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25352 options and inversemask / mask of LITTLE_ENDIAN.
25353 * config/microblaze/t-microblaze: Expand multilib options to
25354 include mlittle-endian (le) and update exceptions patterns.
25356 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25358 PR rtl-optimization/56195
25359 * lra-constraints.c (get_reload_reg): Don't reuse regs
25360 if they have smaller mode than requested, if they have
25361 wider mode than requested, try to return a SUBREG.
25363 PR tree-optimization/56250
25364 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25365 if type is unsigned and code isn't MULT_EXPR.
25367 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25369 PR tree-optimization/56064
25370 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25371 bits according to mode.
25372 * fixed-value.h (fixed_from_double_int)
25373 (const_fixed_from_double_int): Adjust comments.
25375 2013-02-08 Richard Biener <rguenther@suse.de>
25378 * lto-streamer.h (struct data_in): Remove current_file, current_line
25379 and current_col members.
25380 * lto-streamer-out.c (lto_output_location): Stream changed bits
25381 en-block for efficiency.
25382 * lto-streamer-in.c (clear_line_info): Remove.
25383 (lto_input_location): Cache current file, line and column
25384 globally via local statics. Read changed bits en-block.
25385 (input_function): Do not call clear_line_info.
25386 (lto_read_body): Likewise.
25387 (lto_input_toplevel_asms): Likewise.
25389 2013-02-08 Michael Matz <matz@suse.de>
25391 PR tree-optimization/52448
25392 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25393 (nt_call_phase): New static.
25394 (add_or_mark_expr): Only mark accesses with newer phase than any
25396 (nonfreeing_call_p): New.
25397 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25398 (nt_fini_block): Keep blocks marked as visited.
25399 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25401 2013-02-08 Richard Biener <rguenther@suse.de>
25403 * ira.c (ira): Free broken dominator information.
25405 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25407 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25409 2013-02-08 Marek Polacek <polacek@redhat.com>
25411 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25413 2013-02-08 Richard Biener <rguenther@suse.de>
25415 PR middle-end/56181
25416 * cfgloop.h (flow_loops_find): Adjust.
25417 (bb_loop_header_p): Declare.
25418 * cfgloop.c (bb_loop_header_p): New function split out from ...
25419 (flow_loops_find): ... here. Adjust function signature,
25420 support incremental loop structure update.
25421 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25422 * cfgloopmanip.c (fix_loop_structure): Move ...
25423 * loop-init.c (fix_loop_structure): ... here.
25424 (apply_loop_flags): Split out from ...
25425 (loop_optimizer_init): ... here.
25426 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25427 in incremental mode, only remove dead loops here.
25429 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25432 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25433 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25434 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25435 (*round<mode>3.libgcc): New insns for fixed-modes.
25436 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25437 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25438 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25439 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25440 implementations. Define to __builtin_avr_absFX,
25441 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25442 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25443 __builtin_avr_countlsFX, respectively.
25444 * config/avr/avr-c.c (target.h): Include it.
25445 (enum avr_builtin_id): New enum.
25446 (avr_resolve_overloaded_builtin): New static function.
25447 (avr_register_target_pragmas): Use it to set
25448 targetm.resolve_overloaded_builtin.
25449 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25450 tree nodes used by DEF_BUILTIN.
25451 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25452 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25453 <AVR_BUILTIN_xxBITS>: Same.
25455 2013-02-08 Richard Biener <rguenther@suse.de>
25457 * cfgloop.c (verify_loop_structure): Properly handle
25458 a loop exiting to another loop header.
25459 * ira-int.h (ira_loops): Remove.
25460 * ira.c (ira_loops): Remove.
25461 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25462 (do_reload): Use loop_optimizer_finalize.
25463 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25464 number_of_loops to access the loop tree.
25465 (more_one_region_p): Likewise.
25466 (finish_loop_tree_nodes): Likewise.
25467 (rebuild_regno_allocno_maps): Likewise.
25468 (mark_loops_for_removal): Likewise.
25469 (mark_all_loops_for_removal): Likewise.
25470 (remove_unnecessary_regions): Likewise.
25471 (ira_build): Likewise.
25472 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25474 2013-02-08 Richard Biener <rguenther@suse.de>
25476 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25477 * ipa-pure-const.c (analyze_function): Avoid calling
25478 mark_irreducible_loops twice.
25479 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25481 2013-02-07 David S. Miller <davem@davemloft.net>
25483 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25485 * var-tracking.c (vt_add_function_parameter): Test the presence of
25486 HAVE_window_save properly and do not remap argument registers when
25487 we have a leaf function.
25489 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25492 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25494 * config/i386/i386.c (ix86_print_operand): Ditto.
25496 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25498 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25500 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25502 PR rtl-optimization/56225
25503 * lra-constraints.c (process_alt_operands): Check that reload hard
25504 reg can hold value for strict_low_part.
25506 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25509 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25510 dwarf2out_end_function.
25511 (in_first_function_p, maybe_at_text_label_p,
25512 first_loclabel_num_not_at_text_label): New variables.
25513 (dwarf2out_var_location): In the first function find out
25514 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25515 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25518 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25520 PR rtl-optimization/56178
25521 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25522 SUBREG of a register. Tidy up related block of code.
25523 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25524 note if the source is a register or a SUBREG of a register.
25526 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25529 * config/rs6000/rs6000.md (ptrm): New mode attr.
25530 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25531 call_value_indirect_aix<pttrsize>,
25532 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25535 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25537 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25538 if -bnortl. Convert to strcmp and strncmp.
25540 2013-02-07 Alan Modra <amodra@gmail.com>
25543 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25544 addresses won't wrap when offsetting.
25545 (rs6000_secondary_reload): Provide secondary reloads needed for
25546 wrapping LO_SUM addresses.
25548 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25550 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25551 MACH, just __MACH__.
25553 2013-02-06 Richard Biener <rguenther@suse.de>
25555 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25556 instead of calling fix_loop_structure.
25558 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25560 PR middle-end/56217
25561 * omp-low.c (use_pointer_for_field): Return false if
25562 lower_send_shared_vars doesn't generate any copy-out code.
25564 2013-02-06 Tom de Vries <tom@codesourcery.com>
25566 PR rtl-optimization/56131
25567 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25568 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25569 of the label is NULL. Add comment.
25571 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25573 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25576 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25577 (STATIC_LIBTSAN_LIBS): Likewise.
25578 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25579 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25580 is defined, don't add anything else beyond that.
25581 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25582 (LINK_COMMAND_SPEC): Use them.
25584 PR tree-optimization/56205
25585 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25586 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25587 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25589 2013-02-05 Richard Biener <rguenther@suse.de>
25591 PR tree-optimization/53342
25592 PR tree-optimization/53185
25593 * tree-vectorizer.h (vect_check_strided_load): Remove.
25594 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25595 not disallow peeling for vectorized strided loads.
25596 (vect_check_strided_load): Make static and simplify.
25597 (vect_analyze_data_refs): Adjust.
25598 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25599 correctly when vectorizing strided loads.
25601 2013-02-05 Richard Biener <rguenther@suse.de>
25603 * doc/install.texi: Refer to ISL, not PPL.
25605 2013-02-05 Jan Hubicka <jh@suse.cz>
25607 PR tree-optimization/55789
25608 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25610 2013-02-05 Jan Hubicka <jh@suse.cz>
25612 PR tree-optimization/55789
25613 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25614 the dead call anyway.
25616 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25619 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25621 2013-02-04 Alexander Potapenko <glider@google.com>
25622 Jack Howarth <howarth@bromo.med.uc.edu>
25623 Jakub Jelinek <jakub@redhat.com>
25626 * config/darwin.c (sort_ctor_records): Stabilized qsort
25627 on constructor priority by using original position.
25628 (finalize_ctors): New routine to sort constructors by
25629 priority before use in assemble_integer.
25630 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25632 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25635 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25636 about visibility on artificial decls.
25637 * config/sol2.c (solaris_assemble_visibility): Likewise.
25639 2013-02-04 Kai Tietz <ktietz@redhat.com>
25642 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25643 argument and improve checking of return-argument types for 16-byte
25645 (ix86_function_value_1): Add additional valtype argument on call
25646 of function_value_64.
25647 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25648 handling infunction_value_64 function.
25650 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25652 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25654 2013-02-04 Richard Biener <rguenther@suse.de>
25656 PR tree-optimization/56188
25657 * tree-ssa-structalias.c (label_visit): Consider case with
25658 initially non-empty points-to set.
25659 (perform_var_substitution): Dump node mapping and clean up.
25661 2013-02-04 Richard Guenther <rguenther@suse.de>
25664 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25665 node prevail as last resort.
25666 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25667 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25669 2013-02-04 Richard Biener <rguenther@suse.de>
25671 PR tree-optimization/56113
25672 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25674 (equiv_class_lookup_or_add): ... this.
25675 (label_visit): Adjust and fix error in previous patch.
25676 (perform_var_substitution): Adjust.
25678 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25680 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25681 * config/sh/sh4-300.md: Likewise.
25682 * config/sh/sh4a.md: Likewise.
25683 * config/sh/constraints.md: Likewise.
25684 * config/sh/sh.md: Likewise.
25685 * config/sh/netbsd-elf.h: Likewise.
25686 * config/sh/predicates.md: Likewise.
25687 * config/sh/sh-protos.h: Likewise.
25688 * config/sh/ushmedia.h: Likewise.
25689 * config/sh/linux.h: Likewise.
25690 * config/sh/sh.c: Likewise.
25691 * config/sh/superh.h: Likewise.
25692 * config/sh/elf.h: Likewise.
25693 * config/sh/sh4.md: Likewise.
25694 * config/sh/sh.h: Likewise.
25696 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25698 * config/pa/constraints.md: Adjust unused letters. Change "T"
25699 constraint to match_test floating_point_store_memory_operand().
25700 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25701 (base14_operand): New.
25702 (floating_point_store_memory_operand): New.
25703 (integer_store_memory_operand): Revise to use base14_operand and
25704 reg_plus_base_memory_operand.
25705 (move_dest_operand): Allow symbolic_memory_operands.
25706 (symbolic_memory_operand): Check for LO_SOM.
25707 (symbolic_operand): Change default case to break.
25708 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25709 CONST_DOUBLE values to be reloaded by putting them into memory when
25710 the destination is a floating point register.
25711 (movdf): Remove code to handle CONST_DOUBLE.
25713 (reload_indf_r1): New.
25714 (reload_insf_r1): New.
25715 Consistently use "Q" and "T" constraints with integer and floating
25716 point move instructions, respectively.
25717 (movdi): Remove FAIL.
25718 Change predicate for source operand unamed DImode move from
25719 general_operand to move_src_operand.
25720 (umulsidi3): Change predicate for destination operand to
25722 Likewise for similar unamed patterns.
25723 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25724 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25725 (hppa_legitimize_address): Simplify mask calculation.
25726 (pa_emit_move_sequence): Revised handling of secondary reloads from
25727 REG+D addresses for floating point loads and stores. Directly handle
25728 loading CONST0_RTX (mode) to a floating point register.
25729 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25730 to floating point registers. Don't restrict secondary reloads to
25731 floating point registers to integer modes. Revise some comments and
25733 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25734 (pa_legitimate_address_p): New.
25735 (pa_legitimize_reload_address): New.
25736 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25737 (STRICT_REG_OK_FOR_BASE_P): New.
25738 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25739 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25741 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25742 Andrew Dixie <andrewd@gentrack.com>
25744 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25747 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25749 * expmed.c (extract_bit_field_1): Pass the full width of the
25750 structure to get_best_reg_extraction_insn.
25752 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25755 * configure.ac (use_cxa_atexit): Add AIX.
25756 * configure: Regenerate.
25758 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25760 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25763 * final.c (need_profile_function): New variable.
25764 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25765 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25766 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25767 notes, targetm.asm_out.function_prologue doesn't emit anything,
25768 HAVE_prologue and profiler should be emitted before prologue,
25769 set need_profile_function instead of emitting it.
25770 (final_scan_insn): If need_profile_function, emit
25771 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25772 NOTE_INSN_FUNCTION_BEG note.
25774 2013-02-01 Richard Henderson <rth@redhat.com>
25776 * config/rs6000/rs6000.md (smulditi3): New.
25779 * config/alpha/alpha.md (umulditi3): New.
25781 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25783 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25784 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25786 2013-02-01 Richard Biener <rguenther@suse.de>
25788 PR tree-optimization/56113
25789 * tree-ssa-structalias.c (label_visit): Reduce work for
25790 single-predecessor nodes.
25792 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25794 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25795 range isn't testing for zero.
25797 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25799 PR middle-end/56113
25800 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25802 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25803 Nick Clifton <nickc@redhat.com>
25805 * config/v850/constraints.md (Q): Define as a memory constraint.
25806 * config/v850/predicates.md (label_ref_operand): New predicate.
25807 (e3v5_shift_operand): New predicate.
25808 (ior_operator): New predicate.
25809 * config/v850/t-v850: Add e3v5 multilib.
25810 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25811 (v850_gen_movdi): Prototype.
25812 * config/v850/v850.c: Add support for e3v5 architecture.
25813 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25815 (construct_save_jarl): Add e3v5 long JARL support.
25816 (v850_adjust_insn_length): New function. Adjust length of call
25817 insns when using e3v5 instructions.
25818 (v850_gen_movdi): New function: Generate instructions to move a
25820 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25821 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25822 (TARGET_USE_FPU): Enable for e3v5.
25823 (CONST_OK_FOR_W): New macro.
25824 (ADJUST_INSN_LENGTH): Define.
25825 * config/v850/v850.md (UNSPEC_LOOP): Define.
25826 (attr cpu): Add v850e3v5.
25827 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25828 (movdi): New pattern.
25829 (movdi_internal): New pattern.
25830 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25831 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25832 (cstoresf4): Likewise.
25833 (cstoredf4): Likewise.
25834 (insv): New pattern.
25835 (rotlso3_a): New pattern.
25836 (rotlsi3_b): New pattern
25837 (rotlsi3_v850e3v5): New pattern.
25838 (doloop_begin): New pattern.
25839 (fix_loop_counter): New pattern.
25840 (doloop_end): New pattern.
25841 (branch_normal): Add e3v5 long branch support.
25842 (branch_invert): Likewise.
25843 (branch_z_normal): Likewise.
25844 (branch_z_invert): Likewise.
25845 (branch_nz_normal): Likewise.
25846 (branch_nz_invert): Likewise.
25847 (call_internal_short): Add e3v5 register-indirect JARL support.
25848 (call_internal_long): Likewise.
25849 (call_value_internal_short): Likewise.
25850 (call_value_internal_long): Likewise.
25851 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25852 (mloop): New option.
25853 * config.gcc: Add support for configuring v840e3v5 target.
25854 * doc/invoke.texi: Document new v850 specific command line options.
25856 2013-01-31 Paul Koning <ni1d@arrl.net>
25860 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25861 children if parent is a class.
25862 (prune_unused_types_prune): Don't add DW_AT_declaration.
25864 2013-01-31 Richard Biener <rguenther@suse.de>
25866 PR tree-optimization/56157
25867 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25868 match up operand with SLP child.
25870 2013-01-31 Jason Merrill <jason@redhat.com>
25873 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25874 parameters the first time.
25875 (gen_scheduled_generic_parms_dies): Check completeness here.
25877 2013-01-31 Richard Biener <rguenther@suse.de>
25879 PR middle-end/53073
25880 * common.opt (faggressive-loop-optimizations): New flag,
25881 enabled by default.
25882 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25883 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25884 infer_loop_bounds_from_undefined by it.
25886 2013-01-31 Richard Biener <rguenther@suse.de>
25888 PR tree-optimization/56150
25889 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25890 visit virtual operands.
25891 (find_uses_to_rename_bb): Likewise.
25893 2013-01-31 Richard Biener <rguenther@suse.de>
25895 PR tree-optimization/56150
25896 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25897 mixed store non-store stmts.
25899 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25902 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25903 LIBASAN_EARLY_SPEC is defined.
25904 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25905 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25907 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25910 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25911 invalid args instead of ICEing on it.
25912 (ix86_valid_target_attribute_tree): Return error_mark_node if
25913 ix86_valid_target_attribute_inner_p failed.
25914 (ix86_valid_target_attribute_p): Return false only if
25915 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25916 target("default") attribute.
25917 (sorted_attr_string): Change argument from const char * to tree,
25918 merge in all target attribute arguments rather than just one.
25919 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25920 instead of free. Avoid using strcat.
25921 (ix86_mangle_function_version_assembler_name): Mangle
25922 target("default") as if no target attribute is present. Adjust
25923 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25924 instead of xmalloc and XDELETEVEC instead of free.
25925 (ix86_function_versions): Don't return true if one of the decls
25926 doesn't have target attribute. If they don't and one of the decls
25927 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25928 sorted_attr_string caller. Use XDELETEVEC instead of free.
25929 (ix86_supports_function_versions): Remove.
25930 (make_name): Fix up formatting.
25931 (make_dispatcher_decl): Remove resolver_name and its initialization.
25932 Avoid leaking memory.
25933 (is_function_default_version): Return true if there is
25934 target("default") attribute rather than no target attribute at all.
25935 (make_resolver_func): Avoid leaking memory.
25936 (ix86_generate_version_dispatcher_body): Likewise.
25937 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25938 * target.def (supports_function_versions): Remove.
25939 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25940 * doc/tm.texi: Regenerated.
25942 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25944 PR rtl-optimization/56144
25945 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25946 for values with side effects.
25948 2013-01-30 Richard Biener <rguenther@suse.de>
25950 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25951 (sparseset_pop): Likewise.
25952 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25953 to be able to use quick_push in the worker loop.
25955 2013-01-30 Marek Polacek <polacek@redhat.com>
25957 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25959 2013-01-30 Richard Biener <rguenther@suse.de>
25962 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25964 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25966 PR tree-optimization/56064
25967 * fixed-value.c (fixed_from_double_int): New function.
25968 * fixed-value.h (fixed_from_double_int): New prototype.
25969 (const_fixed_from_double_int): New static inline function.
25970 * fold-const.c (native_interpret_fixed): New static function.
25971 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25972 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25973 (native_encode_fixed): New static function.
25974 (native_encode_expr) <FIXED_CST>: Use it.
25975 (native_interpret_int): Move double_int worker code to...
25976 * double-int.c (double_int::from_buffer): ...this new static method.
25977 * double-int.h (double_int::from_buffer): Prototype it.
25979 2013-01-30 Richard Biener <rguenther@suse.de>
25981 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
25982 New pointer-map and obstack.
25983 (init_alias_vars): Allocate pointer-map and obstack.
25984 (delete_points_to_sets): Free them.
25985 (find_what_var_points_to): Cache result.
25986 (find_what_p_points_to): Adjust for changed interface of
25987 find_what_var_points_to.
25988 (compute_points_to_sets): Likewise.
25989 (ipa_pta_execute): Likewise.
25991 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
25993 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
25994 * configure: Regenerate.
25995 * config.in: Regenerate.
25996 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
25997 #nobits/#progbits if supported.
25999 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
26002 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
26003 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
26004 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
26006 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26008 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
26009 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
26011 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26013 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
26014 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
26016 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26018 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
26020 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
26021 * config/arm/cortex-a7.md: New bypasses using
26022 arm_mac_accumulator_is_result.
26024 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26026 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
26027 (cortex_a7_neon_mla): Likewise.
26028 (cortex_a7_fpfmad): New reservation.
26029 (cortex_a7_fpmacs): Use ffmas and update required units.
26030 (cortex_a7_fpmuld): Update required units and latency.
26031 (cortex_a7_fpmacd): Likewise.
26032 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
26033 (cortex_a7_neon). Likewise.
26034 (bypass) Update participating units.
26036 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26038 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
26039 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
26041 * config/arm/vfp11.md (vfp_farith): Use ffmas.
26042 (vfp_fmul): Use ffmad.
26043 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
26044 (cortex_r4_fmacd): Use ffmad.
26045 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
26046 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
26047 (cortex_a9_fmacd): Use ffmad.
26048 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
26049 (cortex_a8_vfp_macd): Use ffmad.
26050 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
26051 (cortex_a5_fpmacd): Use ffmad.
26052 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
26053 (cortex_a15_vfp_macd): Use ffmad.
26054 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
26056 2013-01-29 Jason Merrill <jason@redhat.com>
26059 * varasm.c (default_assemble_visibility): Don't warn about
26060 visibility on artificial decls.
26062 2013-01-29 Richard Biener <rguenther@suse.de>
26064 PR tree-optimization/56113
26065 * tree-ssa-structalias.c (equiv_class_lookup): Also return
26067 (label_visit): Free duplicate bitmaps and record the leader instead.
26068 (perform_var_substitution): Adjust.
26070 2013-01-29 Richard Biener <rguenther@suse.de>
26072 PR tree-optimization/55270
26073 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
26074 the CFG, schedule loops for fixup.
26076 2013-01-29 Nick Clifton <nickc@redhat.com>
26078 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
26081 2013-01-28 Leif Ekblad <leif@rdos.net>
26083 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
26084 * config/i386/i386.h (TARGET_RDOS): New macro.
26085 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
26086 * config/i386/i386.c (ix86_option_override_internal): For 64bit
26087 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
26088 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
26089 DEFAULT_LARGE_SECTION_THRESHOLD.
26090 * config/i386/i386.md (R14_REG, R15_REG): New constants.
26091 * config/i386/rdos.h: New file.
26092 * config/i386/rdos64.h: New file.
26094 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
26097 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
26100 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26102 PR rtl-optimization/56117
26103 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
26104 call cselib_lookup_from_insn on the MEM before calling
26105 add_insn_mem_dependence.
26107 2013-01-28 Richard Biener <rguenther@suse.de>
26109 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26110 to a stmt that didn't have one.
26111 (copy_phis_for_bb): Likewise for PHI arguments.
26112 (copy_debug_stmt): Likewise for debug stmts.
26114 2013-01-28 Richard Biener <rguenther@suse.de>
26116 PR tree-optimization/56034
26117 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26118 (partition_builtin_p): Adjust.
26119 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26120 it is the last partition.
26121 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26122 up the vertex for the definition.
26123 (classify_partition): Classify whether a partition is a
26124 PKIND_REDUCTION, thus has uses outside of the loop.
26125 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26126 Merge all PKIND_REDUCTION partitions into the last partition.
26127 (tree_loop_distribution): Seed partitions from reductions as well.
26129 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26131 PR tree-optimization/56125
26132 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26133 pow(x,c) into sqrt(x) * powi(x, n/2) or
26134 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26135 optimizing for size.
26136 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26137 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26140 PR tree-optimization/56094
26141 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26142 to UNKNOWN_LOCATION while gimplifying expr.
26144 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26147 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26148 operand 0 in movabs insn template for -masm=intel asm alternative.
26149 (*movabs<mode>_2): Ditto for operand 1.
26151 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26154 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26157 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26159 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26162 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26164 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26165 Undef to avoid warning.
26167 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26169 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26170 * configure: Regenerate.
26172 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26174 PR tree-optimization/56098
26175 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26176 for stmts with volatile ops.
26177 (cond_store_replacement): Don't optimize if assign has volatile ops.
26178 (cond_if_else_store_replacement_1): Don't optimize if either
26179 then_assign or else_assign have volatile ops.
26180 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26183 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26185 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26187 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26189 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26190 missing ':' in asm example.
26192 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26194 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26195 entries into lane and laneq entries.
26196 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26197 Remove AdvSIMD scalar modes.
26198 (aarch64_sq<r>dmulh_laneq<mode>): New.
26199 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26201 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26202 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26203 * config/aarch64/iterators.md (VCOND): New.
26206 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26209 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26210 Add NULL LIBNAME argument to existing definitions.
26211 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26212 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26213 * config/avr/avr.c (DEF_BUILTIN): Same.
26214 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26215 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26216 implementation is available (DECL_ASSEMBLER_NAME is set).
26217 (avr_fold_absfx): New static function.
26218 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26219 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26220 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26221 AVR_BUILTIN_ABSLLK.
26222 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26223 (abshk, absk, abslk, absllk): Provide as static inline functions.
26225 2013-01-25 Marek Polacek <polacek@redhat.com>
26227 PR tree-optimization/56035
26228 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26230 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26232 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26233 (*movtf_internal_rex64): Add (!o,C) alternative
26234 (*movxf_internal_rex64): Ditto.
26235 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26237 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26239 * doc/invoke.texi: fix typo.
26240 * doc/objc.texi: fix typo.
26242 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26244 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26245 for the first two alternatives.
26247 2013-01-24 Diego Novillo <dnovillo@google.com>
26249 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26250 (ggc-zone.o): Remove.
26251 * configure.ac: Remove option --with-gc.
26252 * configure: Re-generate.
26253 * doc/install.texi: Remove documentation for --with-gc.
26254 * gengtype.c (write_enum_defn): Remove. Update all users.
26255 (write_Types_process_field): Remove generation of gt_e_* argument.
26256 (output_type_enum): Remove. Update all users.
26257 (write_enum_defn): Remove. Update all users.
26258 (enum alloc_zone): Remove. Update all users.
26259 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26260 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26261 Update all callers.
26262 (struct ptr_data): Remove field TYPE. Update all users.
26263 (gt_pch_note_object): Remove argument TYPE. Update all users.
26264 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26266 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26267 (struct alloc_zone): Remove.
26268 (ggc_internal_alloc_zone_stat): Remove.
26269 (ggc_internal_cleared_alloc_zone_stat): Remove.
26270 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26271 (ggc_pch_count_object): Remove last argument. Update all users.
26272 (ggc_pch_alloc_object): Remove last argument. Update all users.
26273 (struct alloc_zone): Remove.
26274 * ggc-zone.c: Remove.
26275 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26276 (struct alloc_zone): Remove.
26277 (ggc_alloc_typed_stat): Remove.
26278 (ggc_alloc_typed): Remove.
26279 (ggc_splay_alloc): Remove first argument.
26280 (rtl_zone): Remove. Update all users.
26281 (tree_zone): Remove. Update all users.
26282 (tree_id_zone): Remove. Update all users.
26283 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26284 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26285 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26286 * tree-ssanames.c: Remove references to zone allocator in comments.
26288 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26290 * config/avr/avr.c (avr_out_fract): Make register numbers that
26291 might be outside of source operand signed.
26293 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26295 * config/i386/constraints.md (Yf): New constraint.
26296 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26297 of f constraint to conditionaly disable x87 register preferences.
26298 (*movdf_internal): Ditto.
26299 (*movsf_internal): Ditto.
26301 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26303 PR inline-asm/55934
26304 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26305 that have operands with impossible constraints.
26306 Add a FIXME for a speed-up opportunity.
26307 * lra-constraints.c (process_alt_operands): Verify that a class
26308 selected from constraints on asms is valid for the operand mode.
26309 (curr_insn_transform): Remove incorrect comment.
26311 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26313 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26314 TOC operand is a valid symbol ref in the constant pool.
26316 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26318 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26320 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26323 * config/avr/stdfix.h: New file.
26324 * t-avr (stdfix-gcc.h): New rule to build it.
26325 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26327 2013-01-23 Kostya Serebryany <kcc@google.com>
26329 * config/darwin.h: remove dependency on
26330 CoreFoundation (asan on Mac OS).
26332 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26335 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26336 instead of cmpdi_operand for first comparison operand.
26337 Don't assert that comparison operands aren't both constants.
26339 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26341 * doc/install.texi (Downloading the Source): Update references to
26342 downloading separate components.
26344 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26346 * doc/extend.texi (__int128): Improve grammar.
26348 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26351 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26352 alternative to (o,r).
26353 (*movdi_internal_rex64): Remove (!o,n) alternative.
26354 (DImode immediate->memory splitter): Remove.
26355 (DImode immediate->memory peephole2): Remove.
26356 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26357 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26358 alternative to (!o,*r).
26359 (*movtf_internal_sse): New pattern.
26360 (*movxf_internal_rex64): New pattern.
26361 (*movxf_internal): Disable for TARGET_64BIT.
26362 (*movdf_internal_rex64): Remove (!o,F) alternative.
26364 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26366 PR middle-end/56074
26367 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26368 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26369 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26370 stmt locations where LOCATION_LOCUS of the stmt location is
26371 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26374 * config/i386/i386.md (UNSPEC_STOS): New.
26375 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26376 *strsetqi_1): Add UNSPEC_STOS.
26378 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26381 * doc/invoke.texi: Remove left over -Wsynth example.
26383 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26385 PR tree-optimization/56051
26386 * fold-const.c (fold_binary_loc): Don't fold
26387 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26388 a narrowing conversion, or widening conversion from signed
26391 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26393 PR rtl-optimization/56023
26394 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26395 dependent on debug instruction.
26397 2013-01-21 Martin Jambor <mjambor@suse.cz>
26399 PR middle-end/56022
26400 * function.c (allocate_struct_function): Call
26401 invoke_set_current_function_hook earlier.
26403 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26405 * reload1.c (init_reload): Only initialize reload_obstack
26406 during the first call.
26408 2013-01-21 Marek Polacek <polacek@redhat.com>
26410 * cfgloop.c (verify_loop_structure): Fix up grammar.
26412 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26414 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26415 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26417 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26420 * config/arm/marvell-pj4.md: Update copyright year.
26421 Fix up use of alu to alu_reg and simple_alu_imm.
26423 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26425 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26427 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26430 * lra-constraints.c (curr_insn_transform): Don't reuse original
26431 insn for secondary memory move when memory mode should be different.
26433 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26435 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26436 atomic_storedi_1): New patterns.
26438 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26440 btver2 pipeline descriptions.
26441 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26443 * config/i386/i386.md (btver2_decode): New type attributes.
26444 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26446 * config/i386/btver2.md: New file describing btver2 pipelines.
26448 2013-01-19 Andrew Pinski <apinski@cavium.com>
26450 PR tree-optimization/52631
26451 * tree-ssa-sccvn (visit_use): Before looking up the original
26452 statement, try looking up the simplified expression.
26454 2013-01-19 Anthony Green <green@moxielogic.com>
26456 * config/moxie/moxie.c (moxie_expand_prologue): Set
26457 current_function_static_stack_size.
26459 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26461 PR tree-optimization/56029
26462 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26463 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26465 2013-01-18 Sharad Singhai <singhai@google.com>
26467 PR tree-optimization/55995
26468 * dumpfile.c (dump_loc): Print location only if available.
26469 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26471 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26474 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26475 secondary memory move.
26476 (inherit_reload_reg): Use rclass instead of cl for
26477 check_secondary_memory_needed_p.
26479 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26481 PR middle-end/56015
26482 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26483 the case where writing real complex part of target modifies op1.
26485 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26487 * config/aarch64/aarch64-simd.md
26488 (aarch64_vcond_internal<mode>): Handle unordered cases.
26489 * config/aarch64/iterators.md (v_cmp_result): New.
26491 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26492 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26494 * config/arm/marvell-pj4.md: New file.
26495 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26496 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26497 (generic_vfp): Likewise.
26498 * config/arm/arm-cores.def: Add marvell-pj4.
26499 * config/arm/arm-tune.md: Regenerate.
26500 * config/arm/arm-tables.opt: Regenerate.
26501 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26502 * doc/invoke.texi: Document marvell-pj4.
26504 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26506 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26508 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26513 * var-tracking.c (negative_power_of_two_p): New.
26514 (global_get_addr_cache, local_get_addr_cache): New.
26515 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26516 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26518 (vt_stack_offset_p): Remove.
26519 (vt_canon_true_dep): Always canonicalize loc's address.
26520 (clobber_overlapping_mems): Make sure we have a MEM.
26521 (local_get_addr_clear_given_value): New.
26522 (val_reset): Clear local cached entries.
26523 (compute_bb_dataflow): Create and release the local cache.
26524 Disable duplicate MEMs clobbering.
26525 (emit_notes_in_bb): Clobber MEMs likewise.
26526 (vt_emit_notes): Create and release the local cache.
26527 (vt_initialize, vt_finalize): Create and release the global
26528 cache, respectively.
26529 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26531 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26533 PR libmudflap/53359
26534 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26535 not found in the symtab.
26537 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26540 PR rtl-optimization/55547
26541 PR rtl-optimization/53827
26544 * alias.c (offset_overlap_p): New, factored out of...
26545 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26546 the conservative special case for symbolic constants. Don't
26547 adjust zero sizes on alignment.
26549 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26551 PR rtl-optimization/52573
26552 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26553 REG_UNUSED for the same register.
26555 2013-01-17 Richard Biener <rguenther@suse.de>
26556 Marek Polacek <polacek@redhat.com>
26558 PR rtl-optimization/55833
26559 * loop-unswitch.c (unswitch_loops): Move loop verification...
26560 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26561 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26562 Set it to true when we're removing a loop from hierarchy tree in
26563 an irreducible region.
26564 (fix_bb_placements): Adjust caller.
26565 (fix_loop_placements): Likewise.
26567 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26569 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26570 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26571 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26572 Remove ID. Adjust comments.
26573 * config/avr/avr-c.c (avr_builtin_name): Remove.
26574 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26575 * config/avr/avr.c (avr_tolower): New static function.
26576 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26577 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26578 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26581 2013-01-17 Jan Hubicka <jh@suse.cz>
26583 PR tree-optimization/55273
26584 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26586 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26589 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26590 store through atomic_store<mode>_1.
26591 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26593 2013-01-17 Martin Jambor <mjambor@suse.cz>
26595 PR tree-optimizations/55264
26596 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26597 for virtual methods.
26598 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26599 virtual methods before inlining is over.
26600 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26602 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26605 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26607 PR rtl-optimization/56005
26608 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26609 pending reads for prefetch.
26611 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26613 * config/aarch64/aarch64.md
26614 (*cstoresi_neg_uxtw): New pattern.
26615 (*cmovsi_insn_uxtw): New pattern.
26616 (*<optab>si3_uxtw): New pattern.
26617 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26618 (*<optab>si3_insn_uxtw): New pattern.
26619 (*bswapsi2_uxtw): New pattern.
26621 2013-01-16 Richard Biener <rguenther@suse.de>
26623 * tree-inline.c (tree_function_versioning): Remove set but
26624 never used variable.
26626 2013-01-16 Richard Biener <rguenther@suse.de>
26628 PR tree-optimization/55964
26629 * tree-flow.h (rename_variables_in_loop): Remove.
26630 (rename_variables_in_bb): Likewise.
26631 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26632 (copy_loop_before): Adjust and delete update-ssa status.
26633 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26634 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26635 (rename_variables_in_loop): Remove.
26636 (slpeel_update_phis_for_duplicate_loop): Likewise.
26637 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26638 use available cfg machinery instead of duplicating it.
26639 Update PHI nodes and perform poor-mans SSA update here.
26640 (slpeel_tree_peel_loop_to_edge): Adjust.
26642 2013-01-16 Richard Biener <rguenther@suse.de>
26644 PR tree-optimization/54767
26645 PR tree-optimization/53465
26646 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26647 (vrp_visit_phi_node): For PHI arguments coming via backedges
26648 drop all symbolical range information.
26649 (execute_vrp): Compute backedges.
26651 2013-01-16 Richard Biener <rguenther@suse.de>
26653 * doc/install.texi: Update CLooG and ISL requirements to
26656 2013-01-16 Christian Bruel <christian.bruel@st.com>
26659 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26660 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26661 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26663 2013-01-16 DJ Delorie <dj@redhat.com>
26665 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26666 (UNSPECV_SP_SWITCH_E): New.
26667 (sp_switch_1): Change to an unspec.
26668 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26671 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26673 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26674 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26675 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26676 (expand_mem_thread_fence): Ditto.
26677 (expand_mem_signal_fence): Ditto.
26678 (expand_atomic_load): Ditto.
26679 (expand_atomic_store): Ditto.
26681 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26683 PR rtl-optimization/55547
26684 PR rtl-optimization/53827
26687 * alias.c (memrefs_conflict_p): Set sizes to negative after
26690 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26693 * function.c (thread_prologue_and_epilogue_insns): Always
26694 add crtl->drap_reg to set_up_by_prologue.set, even if
26695 stack_realign_drap is false.
26697 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26699 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26700 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26701 *call): Fix indention.
26703 2013-01-15 Tom de Vries <tom@codesourcery.com>
26706 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26709 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26711 PR rtl-optimization/55153
26712 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26714 2013-01-15 Martin Jambor <mjambor@suse.cz>
26716 PR tree-optimization/55920
26717 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26718 accesses as grp_to_be_debug_replaced.
26720 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26722 PR tree-optimization/55920
26723 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26724 there is non-useless type conversion needed from debug rhs to lhs,
26725 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26727 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26728 Mikael Pettersson <mikpe@it.uu.se>
26731 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26733 (ASM_OUTPUT_CASE_LABEL): Remove.
26734 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26735 * final.c (shorten_branches): Update alignment of labels before
26736 jump tables if CASE_VECTOR_SHORTEN_MODE.
26738 2013-01-15 Richard Biener <rguenther@suse.de>
26741 * system.h: Do not include gmp.h for building host tools.
26743 2013-01-15 Richard Biener <rguenther@suse.de>
26745 PR middle-end/55882
26746 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26747 account for bitpos when computing alignment.
26749 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26751 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26752 (ix86_target_macros_internal): Likewise.
26754 * config/i386/i386.c (m_CORE2I7): Removed.
26755 (m_CORE_HASWELL): New macro.
26756 (m_CORE_ALL): Likewise.
26757 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26758 (initial_ix86_arch_features): Likewise.
26759 (processor_target_table): Initializations for Core avx2.
26760 (cpu_names): New names "core-avx2".
26761 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26762 PROCESSOR_CORE_HASWELL.
26763 (ix86_issue_rate): New case.
26764 (ia32_multipass_dfa_lookahead): Likewise.
26765 (ix86_sched_init_global): Likewise.
26767 * config/i386/i386.h (TARGET_HASWELL): New macro.
26768 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26769 (processor_type): New PROCESSOR_HASWELL.
26771 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26773 PR tree-optimization/55955
26774 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26775 *SHIFT_EXPR and *ROTATE_EXPR codes.
26777 PR tree-optimization/48766
26778 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26779 -ftrapv disable -fwrapv.
26781 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26784 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26785 etc. to 1 and not to __flash.
26786 Use LL suffix for __INT24_MAX__ with -mint8.
26787 Use ULL suffix for __UINT24_MAX__ with -mint8.
26789 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26791 * config/avr/avr-arch.h
26792 (struct base_arch_s): Use typedef avr_arch_t instead.
26793 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26794 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26795 * config/avr/avr.c: Same.
26796 * config/avr/avr-devices.c: Same.
26797 * config/avr/driver-avr.c: Same.
26798 * config/avr/gen-avr-mmcu-texi.c: Same.
26799 * config/avr/avr-mcus.def: Adjust comment.
26801 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26803 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26804 * config/aarch64/iterators.md (VALLDI): New.
26806 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26807 Andi Kleen <ak@linux.intel.com>
26810 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26811 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26814 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26816 * config/avr/avr-stdint.h: Remove trailing blanks.
26817 * config/avr/avr-log.h: Same.
26818 * config/avr/avr-arch.h: Same.
26819 * config/avr/avr-devices.c: Same.
26820 * config/avr/avr-dimode.md: Same.
26821 * config/avr/predicates.md: Same.
26822 * config/avr/avr-c.c: Same. And fix typo.
26824 * config/avr/avr-protos.h: Same. And:
26825 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26826 (init_cumulative_args): Rename to avr_init_cumulative_args.
26827 (expand_prologue): Rename to avr_expand_prologue.
26828 (expand_epilogue): Rename to avr_expand_epilogue.
26829 (adjust_insn_length): Rename to avr_adjust_insn_length.
26830 (notice_update_cc): Rename to avr_notice_update_cc.
26831 (final_prescan_insn): Rename to avr_final_prescan_insn.
26832 * config/avr/avr.c: Same.
26833 * config/avr/avr.h: Same.
26834 * config/avr/avr.md: Remove trailing blanks.
26835 (prologue): Use avr_expand_prologue.
26836 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26838 2013-01-14 Richard Biener <rguenther@suse.de>
26840 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26841 verify_location, collect_subblocks): New functions.
26842 (verify_gimple_in_cfg): Verify that locations only reference
26843 BLOCKs in the functions BLOCK tree.
26845 2013-01-14 Richard Biener <rguenther@suse.de>
26847 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26849 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26851 (insert_out_of_ssa_copy_on_edge): Likewise.
26852 (rewrite_close_phi_out_of_ssa): Likewise.
26853 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26855 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26856 propagated constants.
26857 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26860 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26862 * config/avr/avr-modes.def: Add GPL copyright notice.
26864 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26866 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26867 MEMMODEL_MASK to determine memory model.
26868 (atomic_store<mode>): Ditto from operands[2].
26869 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26871 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26874 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26875 (fold_gimple_assign): Don't call unshare_expr here.
26876 (fold_ctor_reference): Call unshare_expr.
26878 2013-01-13 Terry Guo <terry.guo@arm.com>
26880 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26881 * doc/fragments.texi: Document MULTILIB_REUSE.
26882 * gcc.c (multilib_reuse): New internal spec.
26883 (set_multilib_dir): Also search multilib from multilib_reuse.
26884 * genmultilib (tmpmultilib3): Refactor code.
26885 (tmpmultilib4): Ditto.
26886 (multilib_reuse): New multilib argument.
26888 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26890 * Makefile.in: Update copyright.
26892 2013-01-12 Tom de Vries <tom@codesourcery.com>
26894 PR middle-end/55890
26895 * calls.c (expand_call): Check if arg_nr is valid.
26897 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26899 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26900 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26901 documentation. Add missing '__' in front of
26902 __builtin_ia32_packssdw256.
26904 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26907 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26908 NO_REGS for larl operands.
26909 (s390_reload_larl_operand): Use s390_load_address instead of
26912 2013-01-11 Richard Biener <rguenther@suse.de>
26914 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26915 (verify_node_sharing): ... here.
26916 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26918 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26920 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26921 Substitute TREECHECKING.
26922 * configure: Regenerate.
26923 * Makefile.in (TREECHECKING): New.
26925 2013-01-11 Richard Guenther <rguenther@suse.de>
26927 PR tree-optimization/44061
26928 * tree-vrp.c (extract_range_basic): Compute zero as
26929 value-range for __builtin_constant_p of function parameters.
26931 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26933 Update copyright years.
26935 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26937 PR rtl-optimization/55672
26938 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26939 const to be eliminable.
26941 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26943 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26944 * configure: Regenerate.
26946 2013-01-10 Richard Biener <rguenther@suse.de>
26948 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26950 2013-01-10 Richard Biener <rguenther@suse.de>
26953 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26954 locations for virtual PHI arguments.
26955 (rewrite_update_phi_arguments): Likewise.
26957 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26959 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26962 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26964 PR tree-optimization/55921
26965 * tree-complex.c (expand_complex_asm): New function.
26966 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26968 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26971 * config/s390/s390.c (s390_symref_operand_p)
26972 (s390_loadrelative_operand_p): Merge the two functions.
26973 (s390_check_qrst_address, print_operand_address): Add parameters
26974 to s390_loadrelative_operand_p invokation.
26975 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
26976 (s390_reload_larl_operand, s390_secondary_reload): Use
26977 s390_loadrelative_operand_p instead of s390_symref_operand_p.
26978 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
26980 2013-01-09 Mike Stump <mikestump@comcast.net>
26982 * dse.c (record_store): Remove unnecessary assert.
26984 2013-01-09 Jan Hubicka <jh@suse.cz>
26986 PR tree-optimization/55569
26987 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
26988 * cfgloop.h (scale_loop_profile): Likewise.
26990 2013-01-09 Jan Hubicka <jh@suse.cz>
26993 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
26995 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
26997 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
26999 PR middle-end/55114
27000 * expr.h (maybe_emit_group_store): Declare.
27001 * expr.c (maybe_emit_group_store): New function.
27002 * builtins.c (expand_builtin_int_roundingfn): Call it.
27003 (expand_builtin_int_roundingfn_2): Likewise.
27005 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
27007 PR rtl-optimization/55829
27008 * lra-constraints.c (match_reload): Add code for absent output.
27009 (curr_insn_transform): Add code for reloads of matched inputs
27012 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
27014 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
27015 attribute of movddup insn to DF.
27016 (*vec_interleave_lowv2df): Ditto.
27017 (vec_dupv2df): Ditto.
27019 2013-01-09 Jan Hubicka <jh@suse.cz>
27021 PR tree-optimiation/55875
27022 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
27023 EVERY_ITERATION parameter.
27024 (number_of_iterations_exit): Check if exit is executed every iteration.
27025 (idx_infer_loop_bounds): Similarly here.
27026 (n_of_executions_at_most): Simplify
27027 to only test for cases where statement is dominated by the
27028 particular bound; handle correctly the "postdominance" test.
27029 (scev_probably_wraps_p): Use max loop iterations info
27030 as a global bound first.
27032 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
27033 Nick Clifton <nickc@redhat.com>
27035 * config/v850/v850.md (cbranchsf4): New pattern.
27036 (cstoresf4): New pattern.
27037 (cbranchdf4): New pattern.
27038 (cstoredf4): New pattern.
27039 (movsicc): Disallow floating point comparisons.
27040 (cmpsf_le_insn): Fix order of operators.
27041 (cmpsf_lt_insn): Likewise.
27042 (cmpsf_eq_insn): Likewise.
27043 (cmpdf_le_insn): Likewise.
27044 (cmpdf_lt_insn): Likewise.
27045 (cmpdf_eq_insn): Likewise.
27046 (cmpsf_ge_insn): Use LE comparison.
27047 (cmpdf_ge_insn): Likewise.
27048 (cmpsf_gt_insn): Use LT comparison.
27049 (cmpdf_gt_insn): Likewise.
27050 (cmpsf_ne_insn): Delete pattern.
27051 (cmpdf_ne_insn): Delete pattern.
27052 * config/v850/v850.c (v850_gen_float_compare): Use
27053 gen_cmpdf_eq_insn for NE comparison.
27054 (v850_float_z_comparison_operator)
27055 (v850_float_nz_comparison_operator): Move from here ...
27056 * config/v850/predicates.md: ... to here. Move GT and GE
27057 comparisons into v850_float_z_comparison_operator.
27058 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
27060 (v850_float_nz_comparison_operator): Likewise.
27062 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27064 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
27065 with calls to gen_insvsi/gen_insvdi.
27067 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27069 * config/i386/i386.c (initial_ix86_tune_features): Set up
27070 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
27072 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
27073 Jakub Jelinek <jakub@redhat.com>
27075 PR tree-optimization/48189
27076 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
27077 If nitercst is 0, don't predict the exit edge.
27079 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27081 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
27082 in asm_fprintf with reg_names.
27083 (aarch64_print_operand_address): Likewise.
27084 (aarch64_return_addr): Likewise.
27085 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
27087 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27089 * config/pa/pa.h (VAL_U6_BITS_P): Define.
27090 (INT_U6_BITS): Likewise.
27091 * config/pa/predicates.md (uint6_operand): New predicate.
27092 (shift5_operand, shift6_operand): Likewise.
27093 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
27095 (lshrdi3): Use shift6_operand.
27096 (shrpsi4, shrpdi4): New insn patterns.
27097 (extzv): Delete expander.
27098 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
27099 predicates in unamed zero extract patterns. Tighten common constraint.
27100 (extv): Delete expander.
27101 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
27102 predicates in unamed sign extract patterns. Tighten common constraint.
27103 (insv): Delete expander.
27104 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
27105 predicates in unamed insert patterns. Tighten common constraint.
27106 Change uint32_operand predicate to uint6_operand predicate in unamed
27107 DImode pattern to insert constant values of type 1...1xxxx.
27109 2013-01-04 Jan Hubicka <jh@suse.cz>
27111 PR tree-optimization/55823
27112 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27115 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27116 Uros Bizjak <ubizjak@gmail.com>
27118 PR rtl-optimization/55845
27119 * df-problems.c (can_move_insns_across): Stop scanning at
27120 volatile_insn_p source instruction or give up if
27121 across_from .. across_to range contains any volatile_insn_p
27124 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27126 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27127 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27129 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27130 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27132 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27135 * asan.c (asan_clear_shadow): New function.
27136 (asan_emit_stack_protection): Use it.
27138 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27140 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27141 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27142 with tab instead of space.
27144 2013-01-08 Nick Clifton <nickc@redhat.com>
27146 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27147 register bank 0 at the start of an interrupt handler.
27148 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27151 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27153 * config/aarch64/aarch64-simd.md
27154 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27155 (aarch64_simd_bsl): Likewise.
27156 (aarch64_vcond_internal<mode>): Likewise.
27157 (vcond<mode><mode>): Likewise.
27158 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27159 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27161 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27163 * config/aarch64/aarch64-builtins.c
27164 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27166 2013-01-08 Martin Jambor <mjambor@suse.cz>
27169 * tree-sra.c (analyze_access_subtree): Return true also after
27170 potentially creating a debug-only replacement.
27172 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27174 PR middle-end/55890
27175 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27177 PR tree-optimization/54120
27178 * tree-vrp.c (range_fits_type_p): Don't allow
27179 src_precision < precision from signed vr to unsigned_p
27180 if vr->min or vr->max is negative.
27181 (simplify_float_conversion_using_ranges): Test can_float_p
27182 against CODE_FOR_nothing.
27184 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27185 Richard Biener <rguenther@suse.de>
27187 PR middle-end/55851
27188 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27189 types instead of just INTEGER_TYPE types.
27191 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27193 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27196 2013-01-07 Steve Ellcey <sellcey@mips.com>
27199 * config/mips/mips.opt: Change mad to mmad to match documentation.
27201 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27204 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27205 .progmemx.data now.
27207 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27210 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27211 (avr_addrspace_t): Add .section_name field.
27212 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27214 (avr_addrspace): Same. Initialize .section_name. Remove last
27215 NULL entry. Put __memx into .progmemx.data.
27216 (progmem_section_prefix): Remove.
27217 (avr_asm_init_sections): No need to initialize progmem_section.
27218 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27219 section name prefix.
27220 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27221 retrieve the progmem section.
27222 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27223 boundary to run over avr_addrspace[].
27224 (avr_register_target_pragmas): Ditto.
27226 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27228 * varasm.c (output_constant_def_contents): For asan_protect_global
27229 protected strings, adjust DECL_ALIGN if needed, before testing for
27231 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27232 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27234 (output_object_block): For asan protected decls, emit asan padding
27235 after their contents.
27236 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27237 (asan_finish_file): Test it here instead.
27239 2013-01-07 Nick Clifton <nickc@redhat.com>
27240 Matthias Klose <doko@debian.org>
27241 Doug Kwan <dougkwan@google.com>
27242 H.J. Lu <hongjiu.lu@intel.com>
27245 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27247 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27249 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27251 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27253 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27255 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27258 * doc/install.texi (Cross-Compiler-Specific Options): Document
27261 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27263 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27264 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27265 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27266 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27267 vqmovun_high_s64): Fix source operand number and update copyright.
27269 2013-01-07 Richard Biener <rguenther@suse.de>
27271 PR middle-end/55890
27272 * gimple.h (gimple_call_builtin_p): New overload.
27273 * gimple.c (validate_call): New function.
27274 (gimple_call_builtin_p): Likewise.
27275 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27276 Use gimple_call_builtin_p.
27277 (find_func_clobbers): Likewise.
27278 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27279 (strlen_optimize_stmt): Likewise.
27281 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27283 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27284 (vld1q_dup_*): Likewise.
27285 (vld1_*): Likewise.
27286 (vld1q_*): Likewise.
27287 (vld1_lane_*): Likewise.
27288 (vld1q_lane_*): Likewise.
27290 2013-01-07 Richard Biener <rguenther@suse.de>
27292 * lto-streamer.h (LTO_minor_version): Bump to 2.
27294 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27296 * config/aarch64/aarch64-protos.h
27297 (aarch64_const_double_zero_rtx_p): Rename to...
27298 (aarch64_float_const_zero_rtx_p): ...this.
27299 (aarch64_float_const_representable_p): New.
27300 (aarch64_output_simd_mov_immediate): Likewise.
27301 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27302 move immediate case.
27303 * config/aarch64/aarch64.c
27304 (aarch64_const_double_zero_rtx_p): Rename to...
27305 (aarch64_float_const_zero_rtx_p): ...this.
27306 (aarch64_print_operand): Allow printing of new constants.
27307 (aarch64_valid_floating_const): New.
27308 (aarch64_legitimate_constant_p): Check for valid floating-point
27310 (aarch64_simd_valid_immediate): Likewise.
27311 (aarch64_vect_float_const_representable_p): New.
27312 (aarch64_float_const_representable_p): Likewise.
27313 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27314 (aarch64_output_simd_mov_immediate): New.
27315 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27316 (*movdf_aarch64): Likewise.
27317 * config/aarch64/constraints.md (Ufc): New.
27318 (Y): call aarch64_float_const_zero_rtx.
27319 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27321 2013-01-07 Richard Biener <rguenther@suse.de>
27323 PR tree-optimization/55888
27324 PR tree-optimization/55862
27325 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27326 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27327 not if it is contained therein.
27329 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27331 * config/avr/t-avr: Typo.
27333 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27336 * config/avr/t-avr: Don't automatically rebuild
27337 $(srcdir)/config/avr/t-multilib
27338 $(srcdir)/config/avr/avr-tables.opt
27339 $(srcdir)/doc/avr-mmcu.texi
27340 (avr-mcus): New phony target to build them on request.
27341 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27342 * avr/avr-mcus.def: Adjust comments.
27344 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27346 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27348 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27350 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27352 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27354 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27356 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27358 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27359 to generate profiling.
27360 * config/rs6000/aix64.h (LIB_SPEC): Same.
27362 2013-01-04 Andrew Pinski <apinski@cavium.com>
27364 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27366 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27368 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27370 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27372 (ix86_expand_move): Ditto.
27373 (ix86_zero_extend_to_Pmode): Ditto.
27374 (ix86_expand_call): Ditto.
27375 (ix86_expand_special_args_builtin): Ditto.
27376 (ix86_expand_builtin): Ditto.
27378 2013-01-04 Richard Biener <rguenther@suse.de>
27380 PR tree-optimization/55862
27381 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27382 translating them through PHI nodes.
27384 2013-01-04 Martin Jambor <mjambor@suse.cz>
27386 PR tree-optimization/55755
27387 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27388 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27390 2013-01-04 Marek Polacek <polacek@redhat.com>
27392 PR middle-end/55859
27393 * opts.c (default_options_optimization): Clarify error message.
27395 2013-01-04 Richard Biener <rguenther@suse.de>
27397 PR middle-end/55863
27398 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27401 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27404 * config/pa/pa.md (movsi): Revert previous change.
27405 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27408 2013-01-03 Richard Henderson <rth@redhat.com>
27410 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27411 after eliminating TLS symbols.
27413 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27416 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27417 * graphite-poly.c (debug_gmp_value): Likewise.
27419 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27422 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27423 selected code model, define __code_mode_small__, __code_model_medium__,
27424 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27425 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27426 xchg temporary register with %k. Declare temporary register as
27428 [__x86_64__]: For medium and large code models, preserve %rbx register.
27430 2013-01-03 Richard Biener <rguenther@suse.de>
27432 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27433 (dump_subscript): Adjust.
27434 (finalize_ddr_dependent): Do not dump redundant info.
27435 (analyze_siv_subscript): Adjust.
27436 (subscript_dependence_tester): Likewise.
27437 (compute_affine_dependence): Likewise.
27439 2013-01-03 Richard Biener <rguenther@suse.de>
27442 2013-01-03 Richard Biener <rguenther@suse.de>
27444 PR tree-optimization/55857
27445 * tree-vect-stmts.c (vectorizable_load): Do not setup
27446 re-alignment for invariant loads.
27448 2013-01-02 Richard Biener <rguenther@suse.de>
27450 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27451 invariant load do not generate a vector load from the scalar location.
27453 2013-01-03 Richard Biener <rguenther@suse.de>
27455 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27456 for not vectorizing.
27457 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27458 not build INDIRECT_REFs, call get_name once only.
27459 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27460 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27462 2013-01-03 Richard Biener <rguenther@suse.de>
27464 PR tree-optimization/55857
27465 * tree-vect-stmts.c (vectorizable_load): Do not setup
27466 re-alignment for invariant loads.
27468 2013-01-03 Richard Biener <rguenther@suse.de>
27471 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27472 prefer a built-in decl.
27474 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27476 * gcc.c (process_command): Update copyright notice dates.
27477 * gcov.c (print_version): Likewise.
27478 * gcov-dump.c (print_version): Likewise.
27480 PR rtl-optimization/55838
27481 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27482 iv0.step, iv1.step and step.
27484 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27485 Marc Glisse <marc.glisse@inria.fr>
27487 PR tree-optimization/55832
27488 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27489 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27490 integer_{one,zero}_node.
27492 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27495 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27496 * var-tracking.c (reverse_op): Don't add reverse ops to
27497 VALUEs that have already
27498 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27500 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27502 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27504 2013-01-02 Teresa Johnson <tejohnson@google.com>
27506 * dumpfile.c (dump_loc): Print filename with location.
27507 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27508 new location_t parameter to emit complete unroll message with
27509 new dump framework.
27510 (canonicalize_loop_induction_variables): Compute loops location
27511 and pass to try_unroll_loop_completely.
27512 * loop-unroll.c (report_unroll_peel): New function.
27513 (peel_loops_completely): Use new dump format with location
27514 for main dumpfile message, and invoke report_unroll_peel on success.
27515 (decide_unrolling_and_peeling): Ditto.
27516 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27517 by report_unroll_peel.
27518 (decide_peel_completely): Ditto.
27519 (decide_unroll_constant_iterations): Ditto.
27520 (decide_unroll_runtime_iterations): Ditto.
27521 (decide_peel_simple): Ditto.
27522 (decide_unroll_stupid): Ditto.
27523 * cfgloop.c (get_loop_location): New function.
27524 * cfgloop.h (get_loop_location): Declare.
27526 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27528 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27531 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27533 PR middle-end/55198
27534 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27535 BLKmode objects when EXPAND_MEMORY is specified.
27537 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27539 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27541 (fold_builtin_cpu): Do not share cpu model decls across statements.
27543 2013-01-02 Jason Merrill <jason@redhat.com>
27546 * tree.c (build_array_type_1): Revert earlier change.
27548 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27550 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27552 * config/aarch64/aarch64-tune.md: Re-generate.
27554 2013-01-02 Richard Biener <rguenther@suse.de>
27556 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27557 invariant load do not generate a vector load from the scalar location.
27559 2013-01-02 Richard Biener <rguenther@suse.de>
27562 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27563 * configure: Regenerate.
27565 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27567 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27568 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27569 (expand_builtin_int_roundingfn_2): Keep the original target around
27570 for the fallback case.
27572 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27574 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27575 to be clear for sign changes.
27577 2013-01-01 Jan Hubicka <jh@suse.cz>
27579 * ipa-inline-analysis.c: Fix formatting.
27581 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27583 PR tree-optimization/55831
27584 * tree-vect-loop.c (get_initial_def_for_induction): Use
27585 gsi_after_labels instead of gsi_start_bb.
27587 Copyright (C) 2013 Free Software Foundation, Inc.
27589 Copying and distribution of this file, with or without modification,
27590 are permitted in any medium without royalty provided the copyright
27591 notice and this notice are preserved.