1 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
4 * Makefile.in (CXXDEPMODE): Assign and change users.
6 * configure: Regenerate.
8 2013-10-23 David Malcolm <dmalcolm@redhat.com>
10 * gengtype-parse.c (require_without_advance): New.
11 (type): For GTY-marked types that are not GTY((user)), parse any
12 base classes, requiring them to be single-inheritance, and not
13 be templates. For non-GTY-marked types and GTY((user)),
14 continue to skip over any C++ inheritance specification.
15 * gengtype-state.c (state_writer::write_state_struct_type):
16 Write base class of type (if any).
17 (read_state_struct_type): Read base class of type (if any).
18 * gengtype.c (new_structure): Add a "base_class" parameter.
19 (create_optional_field_): Update for new parameter to
21 (adjust_field_rtx_def): Likewise.
22 (adjust_field_tree_exp): Likewise.
23 * gengtype.h (struct type): Add "base_class" field to the s
25 (new_structure): Add "base" parameter.
27 2013-10-23 Sriraman Tallam <tmsriram@google.com>
30 * config/i386/i386.c (ix86_option_override_internal):
31 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
32 (ix86_valid_target_attribute_tree):
33 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
34 Change TARGET_SSE to TARGET_SSE_P (opts->...)
36 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
38 * tree-ssa-loop.h: Remove include files.
39 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
40 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
41 * graphite-clast-to-gimple.c: Likewise.
42 * graphite-scop-detection.c: Likewise.
43 * graphite-sese-to-poly.c: Likewise.
44 * ipa-inline-analysis.c: Likewise.
45 * ipa-pure-const.c: Likewise.
46 * loop-init.c: Likewise.
48 * predict.c: Likewise.
49 * tree-cfg.c: Likewise.
50 * tree-cfgcleanup.c: Likewise.
51 * tree-chrec.c: Likewise.
52 * tree-data-ref.c: Likewise.
53 * tree-loop-distribution.c: Likewise.
54 * tree-parloops.c: Likewise.
55 * tree-predcom.c: Likewise.
56 * tree-scalar-evolution.c: Likewise.
57 * tree-ssa-address.c: Likewise.
58 * tree-ssa.c: Likewise.
59 * tree-ssa-dce.c: Likewise.
60 * tree-ssa-loop.c: Likewise.
61 * tree-ssa-loop-im.c: Likewise.
62 * tree-ssa-loop-ivcanon.c: Likewise.
63 * tree-ssa-loop-ivopts.c: Likewise.
64 * tree-ssa-loop-manip.c: Likewise.
65 * tree-ssa-loop-niter.c: Likewise.
66 * tree-ssa-loop-prefetch.c: Likewise.
67 * tree-ssa-loop-unswitch.c: Likewise.
68 * tree-ssa-reassoc.c: Likewise.
69 * tree-vect-data-refs.c: Likewise.
70 * tree-vect-loop.c: Likewise.
71 * tree-vect-loop-manip.c: Likewise.
72 * tree-vectorizer.c: Likewise.
73 * tree-vect-stmts.c: Likewise.
74 * tree-vrp.c: Likewise.
76 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
78 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
80 2013-10-23 Jakub Jelinek <jakub@redhat.com>
82 PR tree-optimization/58775
83 PR tree-optimization/58791
84 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
85 (insert_stmt_after): Rewritten, don't move the stmt, but really
87 (get_stmt_uid_with_default): Remove.
88 (build_and_add_sum): Use insert_stmt_after and
89 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only
91 (update_range_test): Set uid on stmts added by
92 force_gimple_operand_gsi. Don't immediately modify statements
93 in inter-bb optimization, just update oe->op values.
94 (optimize_range_tests): Return bool whether any changed have
96 (update_ops): New function.
97 (struct inter_bb_range_test_entry): New type.
98 (maybe_optimize_range_tests): Perform statement changes here.
99 (not_dominated_by, appears_later_in_bb, get_def_stmt,
100 ensure_ops_are_available): Remove.
101 (find_insert_point): Rewritten.
102 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
103 return LHS of the (new resp. old) stmt. Don't call
104 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
105 instead of last, move new stmt at the right place.
106 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
107 (negate_value): Likewise. Set uids.
108 (break_up_subtract_bb): Initialize uids.
109 (reassociate_bb): Adjust rewrite_expr_tree caller.
110 (do_reassoc): Don't call renumber_gimple_stmt_uids.
112 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
115 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
116 TARGET_32BIT final condition.
117 (mulsi3_internal2 and splitter): Same.
119 2013-10-23 Jeff Law <law@redhat.com>
121 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
122 through joiner blocks with abnormal outgoing edges.
124 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
125 Add parameter JOINERS, to allow/disallow threading through joiner
127 (thread_block): New. Call thread_block_1.
128 (mark_threaded_blocks): Remove code to filter out certain cases
129 of threading through joiner blocks.
130 (thread_through_all_blocks): Document how we can have a dangling
131 edge AUX field and clear it.
133 2013-10-23 Ian Lance Taylor <iant@google.com>
135 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
136 (C++ Dialect Options): Likewise.
137 (Optimize Options): Likewise.
139 2013-10-23 Tom de Vries <tom@codesourcery.com>
141 PR tree-optimization/58805
142 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
144 2013-10-23 Jakub Jelinek <jakub@redhat.com>
146 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
147 sequence based on get_range_info returned range.
149 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
151 * tree-ssa.h: Remove all #include's
152 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
153 * alias.c: Move required includes from tree-ssa.h.
155 * builtins.c: Likewise.
157 * cfgexpand.c: Likewise.
158 * cfghooks.c: Likewise.
159 * cfgloop.c: Likewise.
160 * cfgloopmanip.c: Likewise.
161 * cgraph.c: Likewise.
162 * cgraphbuild.c: Likewise.
163 * cgraphclones.c: Likewise.
164 * cgraphunit.c: Likewise.
166 * except.c: Likewise.
169 * fold-const.c: Likewise.
170 * ggc-page.c: Likewise.
171 * gimple-builder.c: Likewise.
172 * gimple-fold.c: Likewise.
173 * gimple-iterator.c: Likewise.
174 * gimple-low.c: Likewise.
175 * gimple-pretty-print.c: Likewise.
176 * gimple-ssa-strength-reduction.c: Likewise.
177 * gimple-streamer-in.c: Likewise.
178 * gimple-streamer-out.c: Likewise.
179 * gimplify.c: Likewise.
180 * graphite-blocking.c: Likewise.
181 * graphite-clast-to-gimple.c: Likewise.
182 * graphite-dependences.c: Likewise.
183 * graphite-interchange.c: Likewise.
184 * graphite-optimize-isl.c: Likewise.
185 * graphite-poly.c: Likewise.
186 * graphite-scop-detection.c: Likewise.
187 * graphite-sese-to-poly.c: Likewise.
188 * graphite.c: Likewise.
189 * ipa-cp.c: Likewise.
190 * ipa-inline-analysis.c: Likewise.
191 * ipa-inline-transform.c: Likewise.
192 * ipa-inline.c: Likewise.
193 * ipa-prop.c: Likewise.
194 * ipa-pure-const.c: Likewise.
195 * ipa-reference.c: Likewise.
196 * ipa-split.c: Likewise.
197 * ipa-utils.c: Likewise.
198 * loop-init.c: Likewise.
199 * lto-cgraph.c: Likewise.
200 * lto-section-in.c: Likewise.
201 * lto-section-out.c: Likewise.
202 * lto-streamer-in.c: Likewise.
203 * lto-streamer-out.c: Likewise.
204 * lto-streamer.c: Likewise.
205 * omp-low.c: Likewise.
206 * passes.c: Likewise.
207 * predict.c: Likewise.
208 * print-tree.c: Likewise.
209 * profile.c: Likewise.
211 * targhooks.c: Likewise.
212 * tracer.c: Likewise.
213 * trans-mem.c: Likewise.
214 * tree-call-cdce.c: Likewise.
215 * tree-cfg.c: Likewise.
216 * tree-cfgcleanup.c: Likewise.
217 * tree-chrec.c: Likewise.
218 * tree-complex.c: Likewise.
219 * tree-data-ref.c: Likewise.
220 * tree-dfa.c: Likewise.
221 * tree-eh.c: Likewise.
222 * tree-emutls.c: Likewise.
223 * tree-if-conv.c: Likewise.
224 * tree-inline.c: Likewise.
225 * tree-into-ssa.c: Likewise.
226 * tree-loop-distribution.c: Likewise.
227 * tree-mudflap.c: Likewise.
228 * tree-nested.c: Likewise.
229 * tree-nrv.c: Likewise.
230 * tree-object-size.c: Likewise.
231 * tree-outof-ssa.c: Likewise.
232 * tree-parloops.c: Likewise.
233 * tree-phinodes.c: Likewise.
234 * tree-predcom.c: Likewise.
235 * tree-pretty-print.c: Likewise.
236 * tree-profile.c: Likewise.
237 * tree-scalar-evolution.c: Likewise.
238 * tree-sra.c: Likewise.
239 * tree-ssa-address.c: Likewise.
240 * tree-ssa-alias.c: Likewise.
241 * tree-ssa-ccp.c: Likewise.
242 * tree-ssa-coalesce.c: Likewise.
243 * tree-ssa-copy.c: Likewise.
244 * tree-ssa-copyrename.c: Likewise.
245 * tree-ssa-dce.c: Likewise.
246 * tree-ssa-dom.c: Likewise.
247 * tree-ssa-dse.c: Likewise.
248 * tree-ssa-forwprop.c: Likewise.
249 * tree-ssa-ifcombine.c: Likewise.
250 * tree-ssa-live.c: Likewise.
251 * tree-ssa-loop-ch.c: Likewise.
252 * tree-ssa-loop-im.c: Likewise.
253 * tree-ssa-loop-ivcanon.c: Likewise.
254 * tree-ssa-loop-ivopts.c: Likewise.
255 * tree-ssa-loop-manip.c: Likewise.
256 * tree-ssa-loop-niter.c: Likewise.
257 * tree-ssa-loop-prefetch.c: Likewise.
258 * tree-ssa-loop-unswitch.c: Likewise.
259 * tree-ssa-loop.c: Likewise.
260 * tree-ssa-math-opts.c: Likewise.
261 * tree-ssa-operands.c: Likewise.
262 * tree-ssa-phiopt.c: Likewise.
263 * tree-ssa-phiprop.c: Likewise.
264 * tree-ssa-pre.c: Likewise.
265 * tree-ssa-propagate.c: Likewise.
266 * tree-ssa-reassoc.c: Likewise.
267 * tree-ssa-sccvn.c: Likewise.
268 * tree-ssa-sink.c: Likewise.
269 * tree-ssa-strlen.c: Likewise.
270 * tree-ssa-structalias.c: Likewise.
271 * tree-ssa-tail-merge.c: Likewise.
272 * tree-ssa-ter.c: Likewise.
273 * tree-ssa-threadedge.c: Likewise.
274 * tree-ssa-threadupdate.c: Likewise.
275 * tree-ssa-uncprop.c: Likewise.
276 * tree-ssa-uninit.c: Likewise.
277 * tree-ssa.c: Likewise.
278 * tree-ssanames.c: Likewise.
279 * tree-stdarg.c: Likewise.
280 * tree-streamer-in.c: Likewise.
281 * tree-switch-conversion.c: Likewise.
282 * tree-tailcall.c: Likewise.
283 * tree-vect-data-refs.c: Likewise.
284 * tree-vect-generic.c: Likewise.
285 * tree-vect-loop-manip.c: Likewise.
286 * tree-vect-loop.c: Likewise.
287 * tree-vect-patterns.c: Likewise.
288 * tree-vect-slp.c: Likewise.
289 * tree-vect-stmts.c: Likewise.
290 * tree-vectorizer.c: Likewise.
291 * tree-vrp.c: Likewise.
294 * value-prof.c: Likewise.
295 * var-tracking.c: Likewise.
296 * varpool.c: Likewise.
297 * vtable-verify.c: Likewise.
299 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
301 * config/tilegx/tilegx.c: Include "tree.h".
303 2013-10-23 Jakub Jelinek <jakub@redhat.com>
305 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
306 the info, not after it.
307 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
308 dump_ssaname_info call.
309 (pp_gimple_stmt_1): Adjust caller.
310 (dump_phi_nodes): Likewise. Don't print "# " here.
312 2013-10-22 Jan Hubicka <jh@suse.cz>
314 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
315 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
316 * i386.c (expand_small_movmem_or_setmem): New function.
317 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
318 (alg_usable_p): Add support for value ranges; cleanup.
319 (ix86_expand_set_or_movmem): Add support for misaligned moves.
321 2013-10-22 Sterling Augustine <saugustine@google.com>
323 * doc/invoke.texi: Document -ggnu-pubnames.
324 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
326 * dwarf2out.c: Include gdb/gdb-index.h.
327 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
328 debug_generate_pub_sections.
329 (is_java, output_pubtables, output_pubname): New functions.
330 (include_pubname_in_output): Handle debug_generate_pub_sections at
332 (size_of_pubnames): Use new local space_for_flags based on
333 debug_generate_pub_sections.
334 (output_pubnames): Unify pubnames and pubtypes output logic.
335 Genericize comments. Call output_pubname.
336 (dwarf2out_finish): Move logic to output_pubnames and call it.
338 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
341 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
342 Remove CCCmode handling.
343 <case LTU>: Return 'c' suffix for CCCmode.
344 <case GEU>: Return 'nc' suffix for CCCmode.
345 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
346 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
347 (*sub<mode>3_cc_overflow): Ditto.
348 (*subsi3_zext_cc_overflow): Ditto.
350 2013-10-22 Steve Ellcey <sellcey@mips.com>
352 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
353 (AND (NOT OP1) (NOT OP2)).
355 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
357 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
358 meaning of merge-high and merge-low masks for little endian; avoid
359 use of vector-pack masks for little endian for mismatched modes.
361 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
363 * config/tilepro/tilepro.c: Include "tree.h".
365 2013-10-22 Andreas Schwab <schwab@suse.de>
367 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
370 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
372 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
373 [Ed Smith-Rowland]): New entries.
374 ([Stephen M. Webb]): Update.
376 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
378 * tree-ssa-ter.h: Remove duplicate copy of file contents.
380 2013-10-21 Marek Polacek <polacek@redhat.com>
383 * fold-const.c (fold_range_test): Return 0 if the type is not
386 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
388 * system.h: Move hwint.h include further down.
389 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
390 gcc_checking_asserts.
391 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
393 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
395 Fix volatile issues in optimize_bit_field_compare.
396 * fold-const.c (optimize_bit_field_compare): Bail out if
397 lvolatilep or rvolatilep.
399 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
401 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
402 and get_inner_reference returning different pmode for non-volatile
403 bit-field members dependent on flag_strict_volatile_bitfields.
404 * stor-layout.c (layout_decl): Remove special handling of
405 flag_strict_volatile_bitfields.
406 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
407 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
409 2013-10-21 Paulo Matos <pmatos@broadcom.com>
411 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
414 2013-10-21 Jeff Law <law@redhat.com>
416 * tree-ssa-threadedge.c (thread_through_normal_block): New
417 argument VISITED. Remove VISISTED as a local variable. When we
418 have a threadable jump, verify the destination of the jump has not
420 (thread_across_edge): Allocate VISITED bitmap once at function
421 scope and use it throughout. Make sure to set appropriate bits in
422 VISITED for E (start of jump thread path).
423 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
424 through a joiner if any edge on the path has a recorded jump thread.
426 2013-10-21 Ian Lance Taylor <iant@google.com>
428 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
429 don't imply that attributes can solve all problems.
430 (Directory Options): Fix typo.
432 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
434 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
435 extend and extend_add.
437 2013-10-21 Richard Biener <rguenther@suse.de>
439 PR tree-optimization/58794
440 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
441 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
443 2013-10-21 Richard Biener <rguenther@suse.de>
446 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
447 to (T) X for sign-changing conversions (or no conversion).
449 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
451 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
453 2013-10-20 Jan Hubicka <jh@suse.cz>
455 * config/i386/i386-tune.def: Add comment; organize into categories
457 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
459 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
460 argument. Update function comment.
461 (expand_set_or_movmem_via_rep): New function combining
462 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
463 (expand_movmem_via_rep_mov): Remove.
464 expand_setmem_via_rep_stos): Remove.
465 (expand_movmem_epilogue): Update calls correspondingly.
466 (expand_setmem_epilogue_via_loop): Likewise.
468 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
469 (expand_set_or_movmem_prologue): New function combining
470 expand_movmem_prologue and expand_setmem_prologue.
471 (expand_movmem_prologue): Remove.
472 (expand_setmem_prologue): Remove.
473 (expand_set_or_movmem_constant_prologue): New function combining
474 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
475 (expand_constant_movmem_prologue): Remove.
476 (expand_constant_setmem_prologue): Remove.
477 (promote_duplicated_reg): Allow vector-const0 value.
478 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
479 and ix86_expand_setmem.
480 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
481 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
483 2013-10-21 Diego Novillo <dnovillo@google.com>
485 * asan.c: Include tree.h
486 * bb-reorder.c: Likewise.
487 * cfgcleanup.c: Likewise.
488 * cfgloopmanip.c: Likewise.
489 * data-streamer-in.c: Likewise.
490 * data-streamer-out.c: Likewise.
491 * data-streamer.c: Likewise.
492 * dwarf2cfi.c: Likewise.
493 * graphite-blocking.c: Likewise.
494 * graphite-clast-to-gimple.c: Likewise.
495 * graphite-dependences.c: Likewise.
496 * graphite-interchange.c: Likewise.
497 * graphite-optimize-isl.c: Likewise.
498 * graphite-poly.c: Likewise.
499 * graphite-scop-detection.c: Likewise.
500 * graphite-sese-to-poly.c: Likewise.
501 * graphite.c: Likewise.
502 * ipa-devirt.c: Likewise.
503 * ipa-profile.c: Likewise.
506 * loop-init.c: Likewise.
507 * loop-unroll.c: Likewise.
508 * lower-subreg.c: Likewise.
509 * lto/lto-object.c: Likewise.
511 * reginfo.c: Likewise.
512 * tree-loop-distribution.c: Likewise.
513 * tree-parloops.c: Likewise.
514 * tree-ssa-strlen.c: Likewise.
515 * tree-streamer.c: Likewise.
516 * value-prof.c: Likewise.
517 * target-globals.c: Likewise.
518 * expr.h: Include tree-core.h instead of tree.h.
519 * gimple.h: Likewise.
520 * ipa-prop.h: Likewise.
521 * ipa-utils.h: Likewise.
522 * lto-streamer.h: Likewise.
523 * streamer-hooks.h: Likewise.
524 * ipa-reference.h: Include cgraph.h instead of tree.h.
525 * cgraph.h: Include basic-block.h instead of tree.h.
526 * tree-streamer.h: Do not include tree.h.
527 * genattrtab.c (write_header): Generate inclusion of
529 * genautomata.c (main): Likewise.
530 * genemit.c: Likewise.
531 * genopinit.c: Likewise.
532 * genoutput.c (output_prologue): Likewise.
533 * genpeep.c: Likewise.
535 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
537 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
539 (vec_unpacku_hi_v8hi): Likewise.
540 (vec_unpacku_lo_v16qi): Likewise.
541 (vec_unpacku_lo_v8hi): Likewise.
543 2013-10-20 Jan Hubicka <jh@suse.cz>
545 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
546 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
547 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
549 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
551 * config/mips/mips.h (ISA_HAS_WSBH): Define.
552 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
554 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
556 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
559 * system.h: Undef m_slot.
561 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
563 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
564 all elements for both endian flavors.
566 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
569 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
570 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
571 and SI_REG for 64bit SYSV ABI targets.
573 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
575 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
576 to multi_reg_return. Clarify that we are skipping USEs of multiple
577 return registers. Use bool type where appropriate.
579 2013-10-18 Jan Hubicka <jh@suse.cz>
581 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
583 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
584 (X86_TUNE_PUSH_MEMORY): Likewise.
585 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
586 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
587 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
589 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
590 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
592 (ix86_option_override_internal): Update to use tune features instead
595 2013-10-18 Cong Hou <congh@google.com>
597 PR tree-optimization/58508
598 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
599 statement that contains data refs with zero-step.
601 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
603 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
604 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
605 * gimple-low.c (gimple_check_call_arg,
606 gimple_check_call_matching_types): Move to cgraph.c.
607 * gimple-low.h: Remove prototype.
608 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
609 Relocate from gimple-low.c.
610 * cgraph.h: Add prototype. Don't include basic-block.h.
611 * gimplify.c: Add gimple-low to include list.
612 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
613 * tree-eh.c: Add gimple-low to include list.
614 * tree-nested.c: Likewise.
615 * cfgexpand.c: Add tree-ssa-address.h to include list.
617 * gimple-fold.c: Likewise.
618 * gimple-ssa-strength-reduction.c: Likewise.
619 * trans-mem.c: Likewise.
620 * tree-mudflap.c: Likewise.
621 * tree-ssa-loop-ivopts.c: Likewise.
622 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
623 (degenerate_phi_result): Move to tree-phinodes.c.
624 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
625 * tree-ssa-threadedge.c: Likewise.
626 * tree-vrp.c: Likewise.
627 * tree-phinodes.c (degenerate_phi_result): Relocate here.
628 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
629 * tree-phinodes.h (degenerate_phi_result): Add prototype.
630 * tree-ssa-copy.c: Include tree-ssa-dom.h.
631 * tree-ssa-forwprop.c: Likewise.
632 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
633 pass_data_cleanup_cfg_post_optimizing,
634 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
635 * tree-optimize.c: Delete File.
636 * graphite.c: Include tree-cfgcleanup.h.
637 * passes.c: Likewise.
638 * tree-cfg.c: Likewise.
639 * tree-profile.c: Likewise.
640 * tree-ssa-dse.c: Likewise.
641 * tree-ssa-loop-ivcanon.c: Likewise.
642 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
643 * tree-outof-ssa.c: Include sbitmap.h.
644 * tree-ssa-live.c: Likewise.
645 * tree-ssa-propagate.c: Likewise.
646 * tree-ssa-structalias.c: Likewise.
647 * tree-stdarg.c: Likewise.
648 * Makefile.in (OBJS): Delete tree-optimize.o.
649 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
650 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
651 * varasm.c: Include basic-block.h.
652 * cfgloop.h: Include function.h instead of basic-block.h
653 (bb_loop_depth): Move to cfgloop.c.
654 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
656 2013-10-18 Teresa Johnson <tejohnson@google.com>
658 * predict.c (probably_never_executed): Compare frequency-based
659 count to number of training runs.
660 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
662 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
664 * config/arm/arm.c (cortexa9_extra_costs): New table.
665 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
667 2013-10-18 Jeff Law <law@redhat.com>
669 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
671 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
672 (register_jump_thread): Add "registered_jump_thread" debug
674 * dbgcnt.def (registered_jump_thread): New debug counter.
676 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
678 * config/rs6000/rs6000.c: Include cgraph.h.
680 2013-10-18 Teresa Johnson <tejohnson@google.com>
682 * tree-ssa-tail-merge.c (replace_block_by): Update edge
683 weights during merging.
685 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
687 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
688 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
689 * tree-ssa-operands.h: Remove prototype.
690 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
691 * gimple.c (virtual_operand_p): Relocate from gimple.c.
692 * gimple.h: Add prototype.
693 * gimple-ssa.h: Include tree-ssa-operands.h.
694 * tree-dump.c: Add tree-cfg.h to include list.
695 * tree-ssa-alias.c: Add ipa-reference.h to include list.
696 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
697 * config/i386/i386.c: Don't include tree-flow.h.
698 * config/rs6000/rs6000.c: Likewise.
700 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
702 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
704 2013-10-18 Richard Biener <rguenther@suse.de>
706 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
707 or TYPE_UNSIGNED of integral types.
708 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
709 NULL_TREE for zero-precision integral types.
711 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
713 * config/aarch64/arm_neon.h
714 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
716 2013-10-17 Sriraman Tallam <tmsriram@google.com>
719 * opth-gen.awk: Define target_flags_explicit.
721 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
723 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
724 fields to the reg_addr array that describes the valid addressing
725 mode for any register, general purpose registers, floating point
726 registers, and Altivec registers.
727 (FIRST_RELOAD_REG_CLASS): Likewise.
728 (LAST_RELOAD_REG_CLASS): Likewise.
729 (struct reload_reg_map_type): Likewise.
730 (reload_reg_map_type): Likewise.
731 (RELOAD_REG_VALID): Likewise.
732 (RELOAD_REG_MULTIPLE): Likewise.
733 (RELOAD_REG_INDEXED): Likewise.
734 (RELOAD_REG_OFFSET): Likewise.
735 (RELOAD_REG_PRE_INCDEC): Likewise.
736 (RELOAD_REG_PRE_MODIFY): Likewise.
737 (reg_addr): Likewise.
738 (mode_supports_pre_incdec_p): New helper functions to say whether
739 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
740 (mode_supports_pre_modify_p): Likewise.
741 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
742 print the valid address mode bits for each mode.
743 (rs6000_debug_print_mode): Likewise.
744 (rs6000_debug_reg_global): Likewise.
745 (rs6000_setup_reg_addr_masks): New function to set up the address
746 mask bits for each type.
747 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
748 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
749 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
750 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
751 PRE_MODIFY are supported.
752 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
753 registers, instead of {src,dest}_av_p.
754 (rs6000_print_options_internal): Tweak the debug output slightly.
756 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
758 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
761 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
763 * tree-flow.h (struct omp_region): Move to omp-low.c.
764 Remove omp_ prototypes and variables.
765 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
766 (copy_var_decl): Relocate prototype from tree-flow.h.
767 * gimple.c (copy_var_decl): Relocate from omp-low.c.
768 * tree.h: Move prototype to omp-low.h.
769 * omp-low.h: New File. Relocate prototypes here.
770 * omp-low.c (struct omp_region): Make local here.
771 (root_omp_region): Make static.
772 (copy_var_decl) Move to gimple.c.
773 (new_omp_region): Make static.
774 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
775 * tree-cfg.c: Include omp-low.h.
776 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
777 * gimplify.c: Include omp-low.h.
778 * tree-parloops.c: Likewise.
780 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
782 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
783 operands are in memory, prefer to force non-matched operand 1 to
786 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
789 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
790 restrict TImode addresses to single indirect registers if both
791 -mquad-memory and -mvsx-timode are used.
792 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
793 we should emit load/store quad. Remove using %y for quad memory
796 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
797 constraints to allow load/store quad on machines where TImode is
798 not allowed in VSX registers. Use 'n' instead of 'F' constraint
799 for TImode to load integer constants.
801 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
803 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
805 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
807 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
808 handling of STACK_REG.
810 2013-10-17 Richard Biener <rguenther@suse.de>
812 PR tree-optimization/58143
813 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
815 (rewrite_to_defined_overflow): Likewise.
816 (move_computations_dom_walker::before_dom): Rewrite stmts
817 with undefined signed overflow that are not always executed
818 into unsigned arithmetic.
820 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
823 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
824 explicit isa flag to be an options variable, instead of using
825 global_options_set. Remove define from rs6000.h.
826 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
828 * config/rs6000/rs6000.c (rs6000_option_override_internal):
829 Initialize rs6000_isa_flags_explicit.
830 (rs6000_function_specific_save): Add gcc_options* parameter, so
831 that the powerpc builds after the 2013-10-15 changes.
832 (rs6000_function_specific_restore): Likewise.
834 2013-10-16 DJ Delorie <dj@redhat.com>
836 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
837 op is a REG before checking REGNO.
838 (rl78_alloc_physical_registers): Verify pattern is a SET before
841 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
843 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
845 (vec_unpacks_lo_v4sf): Likewise.
846 (vec_unpacks_float_hi_v4si): Likewise.
847 (vec_unpacks_float_lo_v4si): Likewise.
848 (vec_unpacku_float_hi_v4si): Likewise.
849 (vec_unpacku_float_lo_v4si): Likewise.
851 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
853 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
854 (vsx_concat_v2sf): Likewise.
856 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
858 * config/aarch64/aarch64.md
859 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
861 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
863 PR tree-optimization/58697
864 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
865 estimated_loop_iterations_int.
866 (max_stmt_executions_int): Call get_max_loop_iterations_int.
867 (get_max_loop_iterations_int): New. HWINT version of
868 get_max_loop_iterations.
869 * cfgloop.h: Add prototypes.
870 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
871 * loop-unroll.c (decide_peel_once_rolling): Call
872 get_estimated_loop_iterations_int.
873 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
874 * tree-ssa-loop-niter.h: Tweak prototypes.
876 2013-10-16 David Malcolm <dmalcolm@redhat.com>
878 * gengtype-parse.c (struct_field_seq): Ignore access-control
879 keywords ("public:" etc).
881 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
883 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
884 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
886 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
888 * config/arm/arm.opt (mlra): New option.
889 * config/arm/arm.c (arm_lra_p): New function.
890 (TARGET_LRA_P): Define.
892 2013-10-16 Paulo Matos <pmatos@broadcom.com>
894 * tree-core.h (tree_code_name): Remove.
895 * tree.h (get_tree_code_name): New prototype.
896 * tree.c (tree_code_name): Make static.
897 (get_tree_code_name): New function.
898 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
899 tree_class_check_failed, tree_range_check_failed,
900 tree_not_class_check_failed, omp_clause_check_failed,
901 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
902 wrapper get_tree_code_name instead of calling tree_code_name directly.
903 * tree-vrp.c (dump_asserts_for): Likewise.
904 * tree-dump.c (dequeue_and_dump): Likewise.
905 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
906 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
907 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
908 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
909 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
910 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
911 dump_gimple_omp_for): Likewise.
912 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
913 * tree-ssa-pre.c (print_pre_expr): Likewise.
914 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
915 * print-tree.c (print_node_brief, print_node): Likewise.
916 * gimple.c (gimple_check_failed): Likewise.
917 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
918 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
919 * config/mep/mep.c (mep_validate_vliw): Likewise.
920 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
921 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
923 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
925 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
928 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
930 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
931 (MULTILIB_MATCHES): Add multilib for -march=v8.
933 2013-10-15 Sriraman Tallam <tmsriram@google.com>
936 * optc-save-gen.awk: Add extra parameter to the save and restore
938 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
939 * tree.c (build_optimization_node): New parameter. Add extra parameter
940 to call to cl_optimization_save.
941 (build_target_option_node): New parameter. Add extra parameter
942 to call to cl_target_option_save.
943 * tree.h (build_optimization_node): New parameter.
944 (build_target_option_node): New parameter.
945 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
946 build_optimization_node and build_target_option_node.
947 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
948 (handle_pragma_push_options): Ditto.
949 * toplev.c (process_options): Ditto.
950 * opts.c (init_options_struct): Check for opts_set non-null.
951 * target.def (target_option.save): New parameter.
952 (target_option.restore): New parameter.
954 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
955 (ix86_pragma_target_parse): Ditto.
956 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
958 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
959 to build_optimization_node and build_target_option_node.
960 (rs6000_valid_attribute_p): Ditto.
961 (rs6000_pragma_target_parse): Ditto.
962 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
964 * config/i386/i386.h:
965 TARGET_64BIT_P: New Macro
966 TARGET_MMX_P: New Macro.
967 TARGET_3DNOW_P: New Macro.
968 TARGET_3DNOW_A_P: New Macro.
969 TARGET_SSE_P: New Macro.
970 TARGET_SSE2_P: New Macro.
971 TARGET_SSE3_P: New Macro.
972 TARGET_SSSE3_P: New Macro.
973 TARGET_SSE4_1_P: New Macro.
974 TARGET_SSE4_2_P: New Macro.
975 TARGET_AVX_P: New Macro.
976 TARGET_AVX2_P: New Macro.
977 TARGET_AVX512F_P: New Macro.
978 TARGET_AVX512PF_P: New Macro.
979 TARGET_AVX512ER_P: New Macro.
980 TARGET_AVX512CD_P: New Macro.
981 TARGET_FMA_P: New Macro.
982 TARGET_SSE4A_P: New Macro.
983 TARGET_FMA4_P: New Macro.
984 TARGET_XOP_P: New Macro.
985 TARGET_LWP_P: New Macro.
986 TARGET_ABM_P: New Macro.
987 TARGET_BMI_P: New Macro.
988 TARGET_BMI2_P: New Macro.
989 TARGET_LZCNT_P: New Macro.
990 TARGET_TBM_P: New Macro.
991 TARGET_POPCNT_P: New Macro.
992 TARGET_SAHF_P: New Macro.
993 TARGET_MOVBE_P: New Macro.
994 TARGET_CRC32_P: New Macro.
995 TARGET_AES_P: New Macro.
996 TARGET_PCLMUL_P: New Macro.
997 TARGET_CMPXCHG16B_P: New Macro.
998 TARGET_FSGSBASE_P: New Macro.
999 TARGET_RDRND_P: New Macro.
1000 TARGET_F16C_P: New Macro.
1001 TARGET_RTM_P: New Macro.
1002 TARGET_HLE_P: New Macro.
1003 TARGET_RDSEED_P: New Macro.
1004 TARGET_PRFCHW_P: New Macro.
1005 TARGET_ADX_P: New Macro.
1006 TARGET_FXSR_P: New Macro.
1007 TARGET_XSAVE_P: New Macro.
1008 TARGET_XSAVEOPT_P: New Macro.
1009 TARGET_LP64_P: New Macro.
1010 TARGET_X32_P: New Macro.
1011 TARGET_FPMATH_DEFAULT_P: New Macro.
1012 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
1013 * config/i386/i386.c (ix86_option_override_internal): New parameters.
1015 Change ix86_tune_string to access opts->x_ix86_tune_string.
1016 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
1017 Change ix86_arch_string to access opts->x_ix86_arch_string.
1018 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
1019 Change ix86_pmode to access opts->x_ix86_pmode.
1020 Change ix86_abi to access opts->x_ix86_abi.
1021 Change ix86_cmodel to access opts->x_ix86_cmodel.
1022 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
1023 Change ix86_isa_flags_explicit to access
1024 opts->x_ix86_isa_flags_explicit.
1025 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
1026 Change ix86_regparm to access opts->x_ix86_regparm.
1027 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
1028 Change ix86_preferred_stack_boundary_arg to access
1029 opts->x_ix86_preferred_stack_boundary_arg.
1030 Change ix86_force_align_arg_pointer to access
1031 opts->x_ix86_force_align_arg_pointer.
1032 Change ix86_incoming_stack_boundar_arg to access
1033 opts->x_ix86_incoming_stack_boundar_arg.
1034 Change ix86_fpmath to access opts->x_ix86_fpmath.
1035 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
1036 Change ix86_recip_name to access opts->x_ix86_recip_name.
1037 Change ix86_stack_protector_guard to access
1038 opts->x_ix86_stack_protector_guard.
1039 Change ix86_tune_memcpy_strategy to access
1040 opts->x_ix86_tune_memcpy_strategy.
1041 Change ix86_tune_memset_strategy to access
1042 opts->x_ix86_tune_memset_strategy.
1043 Change global_options to access opts.
1044 Change global_options_set to access opts_set.
1045 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
1046 Change TARGET_MMX to TARGET_MMX_P (opts->...).
1047 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
1048 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
1049 Change TARGET_SSE to TARGET_SSE_P (opts->...).
1050 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
1051 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
1052 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
1053 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
1054 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
1055 Change TARGET_AVX to TARGET_AVX_P (opts->...).
1056 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
1057 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
1058 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
1059 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
1060 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
1061 Change TARGET_FMA to TARGET_FMA_P (opts->...).
1062 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
1063 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
1064 Change TARGET_XOP to TARGET_XOP_P (opts->...).
1065 Change TARGET_LWP to TARGET_LWP_P (opts->...).
1066 Change TARGET_ABM to TARGET_ABM_P (opts->...).
1067 Change TARGET_BMI to TARGET_BMI_P (opts->...).
1068 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
1069 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
1070 Change TARGET_TBM to TARGET_TBM_P (opts->...).
1071 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
1072 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
1073 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
1074 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
1075 Change TARGET_AES to TARGET_AES_P (opts->...).
1076 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
1077 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
1078 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
1079 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
1080 Change TARGET_F16C to TARGET_F16C_P (opts->...).
1081 Change TARGET_RTM to TARGET_RTM_P (opts->...).
1082 Change TARGET_HLE to TARGET_HLE_P (opts->...).
1083 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
1084 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
1085 Change TARGET_ADX to TARGET_ADX_P (opts->...).
1086 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
1087 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
1088 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
1089 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
1090 Change TARGET_X32 to TARGET_X32_P (opts->...).
1091 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
1092 Change TARGET_FLOAT_RETURNS_IN_80387 to
1093 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
1094 (ix86_function_specific_save): New parameter. Use opts-> fields
1095 to replace global fields.
1096 (ix86_function_specific_restore): Ditto.
1097 (ix86_valid_target_attribute_inner_p): New parameters.
1099 Fix call to ix86_handle_option and set_option.
1100 (ix86_valid_target_attribute_tree): New parameters.
1101 Change global_options to access opts.
1102 Change global_options_set to access opts_set.
1103 Fix call to ix86_valid_target_attribute_inner_p.
1104 Change ix86_tune_string to access opts->x_ix86_tune_string.
1105 Change ix86_arch_string to access opts->x_ix86_arch_string.
1106 Change ix86_fpmath to access opts->x_ix86_fpmath
1107 Fix call to ix86_option_override_internal.
1108 Fix call to ix86_add_new_builtins.
1109 Fix calls to build_optimization_node and build_target_option_node.
1110 (ix86_valid_target_attribute_p): Remove access to global_options.
1111 Use new gcc_options structure func_options.
1112 Fix call to ix86_valid_target_attribute_tree.
1113 Fix call to build_optimization_node.
1114 (get_builtin_code_for_version): Fix call to
1115 ix86_valid_target_attribute_tree.
1117 2013-10-15 David Malcolm <dmalcolm@redhat.com>
1119 * Makefile.in (PICFLAG): New.
1120 (enable_host_shared): New.
1121 (INTERNAL_CFLAGS): Use PICFLAG.
1122 (LIBIBERTY): Use pic build of libiberty.a if configured with
1123 --enable-host-shared.
1124 * configure.ac: Add --enable-host-shared, setting up new
1126 * configure: Regenerate.
1127 * doc/install.texi (--enable-shared): Add note contrasting it with ...
1128 (--enable-host-shared): New option.
1130 2013-10-15 Richard Biener <rguenther@suse.de>
1132 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1133 for built-in functions.
1135 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1137 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1138 (optimize_range_tests_1): New function,
1139 extracted from optimize_range_tests.
1140 (optimize_range_tests_xor): Similarly.
1141 (optimize_range_tests_diff): New function.
1142 (optimize_range_tests): Use optimize_range_tests_1.
1144 2013-10-15 Cong Hou <congh@google.com>
1146 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1147 requirement of the reduction pattern so that one operand of the
1148 reduction operation can come from outside of the loop.
1150 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1152 * config/arm/neon-schedgen.ml: Remove.
1153 * config/arm/cortex-a9-neon.md: Remove comment regarding
1156 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1158 * config/arm/types: Remove old neon types.
1160 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1162 * config/arm/cortex-a7.md
1163 (cortex_a7_neon_type): New.
1164 (cortex_a7_neon_mul): Update for new types.
1165 (cortex_a7_neon_mla): Likewise.
1166 (cortex_a7_neon): Likewise.
1168 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1170 * config/arm/cortex-a15-neon.md
1171 (cortex_a15_neon_type): New,
1173 (cortex_a15_neon_int_1): Remove.
1174 (cortex_a15_neon_int_2): Likewise.
1175 (cortex_a15_neon_int_3): Likewise.
1176 (cortex_a15_neon_int_4): Likewise.
1177 (cortex_a15_neon_int_5): Likewise.
1178 (cortex_a15_neon_vqneg_vqabs): Likewise.
1179 (cortex_a15_neon_vmov): Likewise.
1180 (cortex_a15_neon_vaba): Likewise.
1181 (cortex_a15_neon_vaba_qqq): Likewise.
1182 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1183 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1184 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1186 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1187 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1188 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1189 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1190 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1191 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1192 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1193 (cortex_a15_neon_shift_1): Likewise.
1194 (cortex_a15_neon_shift_2): Likewise.
1195 (cortex_a15_neon_shift_3): Likewise.
1196 (cortex_a15_neon_vshl_ddd): Likewise.
1197 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1198 (cortex_a15_neon_vsra_vrsra): Likewise.
1199 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1200 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1201 (cortex_a15_neon_bp_3cycle): Likewise.
1202 (cortex_a15_neon_ldm_2): Likewise.
1203 (cortex_a15_neon_stm_2): Likewise.
1204 (cortex_a15_neon_mcr): Likewise.
1205 (cortex_a15_neon_mrc): Likewise.
1206 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1207 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1208 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1209 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1210 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1211 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1212 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1213 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1214 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1215 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1216 (cortex_a15_neon_bp_simple): Likewise.
1217 (cortex_a15_neon_bp_2cycle): Likewise.
1218 (cortex_a15_neon_bp_3cycle): Likewise.
1219 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1220 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1221 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1222 (cortex_a15_neon_vld2_4_regs): Likewise.
1223 (cortex_a15_neon_vld3_vld4): Likewise.
1224 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1225 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1226 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1227 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1228 (cortex_a15_neon_vst3_vst4): Rename to...
1229 (cortex_a15_neon_vst4): This, update for new attributes.
1230 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1231 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1232 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1233 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1234 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1235 (cortex_a15_neon_ldm_2): Likewise.
1236 (cortex_a15_neon_stm_2): Likewise.
1237 (cortex_a15_neon_mcr): Likewise.
1238 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1239 (cortex_a15_neon_mrc): Likewise.
1240 (cortex_a15_neon_mrrc): Likewise.
1242 (cortex_a15_neon_abd): New.
1243 (cortex_a15_neon_abd_q): Likewise.
1244 (cortex_a15_neon_aba): Likewise.
1245 (cortex_a15_neon_aba_q): Likewise.
1246 (cortex_a15_neon_acc): Likewise.
1247 (cortex_a15_neon_acc_q): Likewise.
1248 (cortex_a15_neon_arith_basic): Likewise.
1249 (cortex_a15_neon_arith_complex): Likewise.
1250 (cortex_a15_neon_multiply): Likewise.
1251 (cortex_a15_neon_multiply_q): Likewise.
1252 (cortex_a15_neon_mla): Likewise.
1253 (cortex_a15_neon_mla_q): Likewise.
1254 (cortex_a15_neon_sat_mla_long): Likewise.
1255 (cortex_a15_neon_shift_acc): Likewise.
1256 (cortex_a15_neon_shift_imm_basic): Likewise.
1257 (cortex_a15_neon_shift_imm_complex): Likewise.
1258 (cortex_a15_neon_shift_reg_basic): Likewise.
1259 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1260 (cortex_a15_neon_shift_reg_complex): Likewise.
1261 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1262 (cortex_a15_neon_fp_negabs): Likewise
1263 (cortex_a15_neon_fp_arith): Likewise
1264 (cortex_a15_neon_fp_arith_q): Likewise
1265 (cortex_a15_neon_fp_cvt_int): Likewise
1266 (cortex_a15_neon_fp_cvt_int_q): Likewise
1267 (cortex_a15_neon_fp_cvt_16): Likewise
1268 (cortex_a15_neon_fp_mul): Likewise
1269 (cortex_a15_neon_fp_mul_q): Likewise
1270 (cortex_a15_neon_fp_mla): Likewise
1271 (cortex_a15_neon_fp_mla_q): Likewise
1272 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1273 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1274 (cortex_a15_neon_bitops): Likewise.
1275 (cortex_a15_neon_bitops_q): Likewise.
1276 (cortex_a15_neon_from_gp): Likewise.
1277 (cortex_a15_neon_from_gp_q): Likewise.
1278 (cortex_a15_neon_tbl3_tbl4): Likewise.
1279 (cortex_a15_neon_zip_q): Likewise.
1280 (cortex_a15_neon_to_gp): Likewise.
1281 (cortex_a15_neon_load_a): Likewise.
1282 (cortex_a15_neon_load_b): Likewise.
1283 (cortex_a15_neon_load_c): Likewise.
1284 (cortex_a15_neon_load_d): Likewise.
1285 (cortex_a15_neon_load_e): Likewise.
1286 (cortex_a15_neon_load_f): Likewise.
1287 (cortex_a15_neon_store_a): Likewise.
1288 (cortex_a15_neon_store_b): Likewise.
1289 (cortex_a15_neon_store_c): Likewise.
1290 (cortex_a15_neon_store_d): Likewise.
1291 (cortex_a15_neon_store_e): Likewise.
1292 (cortex_a15_neon_store_f): Likewise.
1293 (cortex_a15_neon_store_g): Likewise.
1294 (cortex_a15_neon_store_h): Likewise.
1295 (cortex_a15_vfp_to_from_gp): Likewise.
1297 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1299 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1301 (cortex_a9_neon_vshl_ddd): Remove.
1302 (cortex_a9_neon_vst3_vst4): Likewise.
1303 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1305 (cortex_a9_neon_bit_ops_q): New.
1307 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1308 (cortex_a9_neon_int_2): Likewise.
1309 (cortex_a9_neon_int_3): Likewise.
1310 (cortex_a9_neon_int_4): Likewise.
1311 (cortex_a9_neon_int_5): Likewise.
1312 (cortex_a9_neon_vqneg_vqabs): Likewise.
1313 (cortex_a9_neon_vmov): Likewise.
1314 (cortex_a9_neon_vaba): Likewise.
1315 (cortex_a9_neon_vaba_qqq): Likewise.
1316 (cortex_a9_neon_shift_1): Likewise.
1317 (cortex_a9_neon_shift_2): Likewise.
1318 (cortex_a9_neon_shift_3): Likewise.
1319 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1320 (cortex_a9_neon_vsra_vrsra): Likewise.
1321 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1322 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1323 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1325 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1326 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1327 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1329 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1330 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1331 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1332 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1333 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1334 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1335 (cortex_a9_neon_fp_vsum): Likewise.
1336 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1337 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1338 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1339 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1340 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1341 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1342 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1343 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1344 (cortex_a9_neon_bp_simple): Likewise.
1345 (cortex_a9_neon_bp_2cycle): Likewise.
1346 (cortex_a9_neon_bp_3cycle): Likewise.
1347 (cortex_a9_neon_ldr): Likewise.
1348 (cortex_a9_neon_str): Likewise.
1349 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1350 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1351 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1352 (cortex_a9_neon_vld2_4_regs): Likewise.
1353 (cortex_a9_neon_vld3_vld4): Likewise.
1354 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1355 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1356 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1357 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1358 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1359 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1360 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1361 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1362 (cortex_a9_neon_mcr): Likewise.
1363 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1364 (cortex_a9_neon_mrc): Likewise.
1365 (cortex_a9_neon_mrrc): Likewise.
1367 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1369 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1371 (cortex_a8_neon_vshl_ddd): Remove.
1372 (cortex_a8_neon_vst3_vst4): Likewise.
1373 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1375 (cortex_a8_neon_bit_ops_q): New.
1377 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1378 (cortex_a8_neon_int_2): Likewise..
1379 (cortex_a8_neon_int_3): Likewise.
1380 (cortex_a8_neon_int_5): Likewise.
1381 (cortex_a8_neon_vqneg_vqabs): Likewise.
1382 (cortex_a8_neon_int_4): Likewise.
1383 (cortex_a8_neon_vaba): Likewise.
1384 (cortex_a8_neon_vaba_qqq): Likewise.
1385 (cortex_a8_neon_shift_1): Likewise.
1386 (cortex_a8_neon_shift_2): Likewise.
1387 (cortex_a8_neon_shift_3): Likewise.
1388 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1389 (cortex_a8_neon_vsra_vrsra): Likewise.
1390 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1391 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1392 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1394 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1395 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1396 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1398 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1399 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1400 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1401 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1402 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1403 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1404 (cortex_a8_neon_fp_vsum): Likewise.
1405 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1406 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1407 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1408 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1409 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1410 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1411 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1412 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1413 (cortex_a8_neon_bp_simple): Likewise.
1414 (cortex_a8_neon_bp_2cycle): Likewise.
1415 (cortex_a8_neon_bp_3cycle): Likewise.
1416 (cortex_a8_neon_ldr): Likewise.
1417 (cortex_a8_neon_str): Likewise.
1418 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1419 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1420 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1421 (cortex_a8_neon_vld2_4_regs): Likewise.
1422 (cortex_a8_neon_vld3_vld4): Likewise.
1423 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1424 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1425 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1426 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1427 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1428 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1429 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1430 (cortex_a8_neon_mcr): Likewise.
1431 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1432 (cortex_a8_neon_mrc): Likewise.
1433 (cortex_a8_neon_mrrc): Likewise.
1435 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1437 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1439 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1440 (aarch64_simd_dup<mode>): Add "type" attribute.
1441 (aarch64_dup_lane<mode>): Likewise.
1442 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1443 (*aarch64_simd_mov<mode>): Likewise.
1444 (aarch64_simd_mov_from_<mode>low): Likewise.
1445 (aarch64_simd_mov_from_<mode>high): Likewise.
1446 (orn<mode>3): Likewise.
1447 (bic<mode>3): Likewise.
1448 (add<mode>3): Likewise.
1449 (sub<mode>3): Likewise.
1450 (mul<mode>3): Likewise.
1451 (*aarch64_mul3_elt<mode>): Likewise.
1452 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1453 (*aarch64_mul3_elt_to_128df): Likewise.
1454 (*aarch64_mul3_elt_to_64v2df): Likewise.
1455 (neg<mode>2): Likewise.
1456 (abs<mode>2): Likewise.
1457 (abd<mode>_3): Likewise.
1458 (aba<mode>_3): Likewise.
1459 (fabd<mode>_3): Likewise.
1460 (*fabd_scalar<mode>3): Likewise.
1461 (and<mode>3): Likewise.
1462 (ior<mode>3): Likewise.
1463 (xor<mode>3): Likewise.
1464 (one_cmpl<mode>2): Likewise.
1465 (aarch64_simd_vec_set<mode>): Likewise.
1466 (aarch64_simd_lshr<mode>): Likewise.
1467 (aarch64_simd_ashr<mode>): Likewise.
1468 (aarch64_simd_imm_shl<mode>): Likewise.
1469 (aarch64_simd_reg_sshl<mode): Likewise.
1470 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1471 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1472 (aarch64_simd_vec_setv2di): Likewise.
1473 (aarch64_simd_vec_set<mode>): Likewise.
1474 (aarch64_mla<mode>): Likewise.
1475 (*aarch64_mla_elt<mode>): Likewise.
1476 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1477 (aarch64_mls<mode>): Likewise.
1478 (*aarch64_mls_elt<mode>): Likewise.
1479 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1480 (<su><maxmin><mode>3): Likewise.
1481 (move_lo_quad_<mode>): Likewise.
1482 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1483 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1484 (vec_pack_trunc_<mode>): Likewise.
1485 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1486 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1487 (*aarch64_<su>mlal_lo<mode>): Likewise.
1488 (*aarch64_<su>mlal_hi<mode>): Likewise.
1489 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1490 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1491 (*aarch64_<su>mlal<mode>): Likewise.
1492 (*aarch64_<su>mlsl<mode>): Likewise.
1493 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1494 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1495 (add<mode>3): Likewise.
1496 (sub<mode>3): Likewise.
1497 (mul<mode>3): Likewise.
1498 (div<mode>3): Likewise.
1499 (neg<mode>2): Likewise.
1500 (abs<mode>2): Likewise.
1501 (fma<mode>4): Likewise.
1502 (*aarch64_fma4_elt<mode>): Likewise.
1503 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1504 (*aarch64_fma4_elt_to_128df): Likewise.
1505 (*aarch64_fma4_elt_to_64v2df): Likewise.
1506 (fnma<mode>4): Likewise.
1507 (*aarch64_fnma4_elt<mode>): Likewise.
1508 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1509 (*aarch64_fnma4_elt_to_128df): Likewise.
1510 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1511 (<frint_pattern><mode>2): Likewise.
1512 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1513 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1514 (vec_unpacks_lo_v4sf): Likewise.
1515 (aarch64_float_extend_lo_v2df): Likewise.
1516 (vec_unpacks_hi_v4sf): Likewise.
1517 (aarch64_float_truncate_lo_v2sf): Likewise.
1518 (aarch64_float_truncate_hi_v4sf): Likewise.
1519 (aarch64_vmls<mode>): Likewise.
1520 (<su><maxmin><mode>3): Likewise.
1521 (<maxmin_uns><mode>3): Likewise.
1522 (reduc_<sur>plus_<mode>): Likewise.
1523 (reduc_<sur>plus_v2di): Likewise.
1524 (reduc_<sur>plus_v2si): Likewise.
1525 (reduc_<sur>plus_<mode>): Likewise.
1526 (aarch64_addpv4sf): Likewise.
1527 (clz<mode>2): Likewise.
1528 (reduc_<maxmin_uns>_<mode>): Likewise.
1529 (reduc_<maxmin_uns>_v2di): Likewise.
1530 (reduc_<maxmin_uns>_v2si): Likewise.
1531 (reduc_<maxmin_uns>_<mode>): Likewise.
1532 (reduc_<maxmin_uns>_v4sf): Likewise.
1533 (aarch64_simd_bsl<mode>_internal): Likewise.
1534 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1535 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1536 (aarch64_get_lane<mode>): Likewise.
1537 (*aarch64_combinez<mode>): Likewise.
1538 (aarch64_combine<mode>): Likewise.
1539 (aarch64_simd_combine<mode>): Likewise.
1540 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1541 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1542 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1543 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1544 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1545 (aarch64_<sur>h<addsub><mode>): Likewise.
1546 (aarch64_<sur><addsub>hn<mode>): Likewise.
1547 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1548 (aarch64_pmul<mode>): Likewise.
1549 (aarch64_<su_optab><optab><mode>): Likewise.
1550 (aarch64_<sur>qadd<mode>): Likewise.
1551 (aarch64_sqmovun<mode>): Likewise.
1552 (aarch64_<sur>qmovn<mode>): Likewise.
1553 (aarch64_s<optab><mode>): Likewise.
1554 (aarch64_sq<r>dmulh<mode>): Likewise.
1555 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1556 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1557 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1558 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1559 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1560 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1561 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1562 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1563 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1564 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1565 (aarch64_sqdmull<mode>): Likewise.
1566 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1567 (aarch64_sqdmull_n<mode>): Likewise.
1568 (aarch64_sqdmull2<mode>_internal): Likewise.
1569 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1570 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1571 (aarch64_<sur>shl<mode>): Likewise.
1572 (aarch64_<sur>q<r>shl<mode>
1573 (aarch64_<sur>shll_n<mode>): Likewise.
1574 (aarch64_<sur>shll2_n<mode>): Likewise.
1575 (aarch64_<sur>shr_n<mode>): Likewise.
1576 (aarch64_<sur>sra_n<mode>): Likewise.
1577 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1578 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1579 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1580 (aarch64_cm<optab><mode>): Likewise.
1581 (aarch64_cm<optab>di): Likewise.
1582 (aarch64_cm<optab><mode>): Likewise.
1583 (aarch64_cm<optab>di): Likewise.
1584 (aarch64_cmtst<mode>): Likewise.
1585 (aarch64_cmtstdi): Likewise.
1586 (aarch64_cm<optab><mode>): Likewise.
1587 (*aarch64_fac<optab><mode>): Likewise.
1588 (aarch64_addp<mode>): Likewise.
1589 (aarch64_addpdi): Likewise.
1590 (sqrt<mode>2): Likewise.
1591 (vec_load_lanesoi<mode>): Likewise.
1592 (vec_store_lanesoi<mode>): Likewise.
1593 (vec_load_lanesci<mode>): Likewise.
1594 (vec_store_lanesci<mode>): Likewise.
1595 (vec_load_lanesxi<mode>): Likewise.
1596 (vec_store_lanesxi<mode>): Likewise.
1597 (*aarch64_mov<mode>): Likewise.
1598 (aarch64_ld2<mode>_dreg): Likewise.
1599 (aarch64_ld2<mode>_dreg): Likewise.
1600 (aarch64_ld3<mode>_dreg): Likewise.
1601 (aarch64_ld3<mode>_dreg): Likewise.
1602 (aarch64_ld4<mode>_dreg): Likewise.
1603 (aarch64_ld4<mode>_dreg): Likewise.
1604 (aarch64_tbl1<mode>): Likewise.
1605 (aarch64_tbl2v16qi): Likewise.
1606 (aarch64_combinev16qi): Likewise.
1607 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1608 (aarch64_st2<mode>_dreg): Likewise.
1609 (aarch64_st2<mode>_dreg): Likewise.
1610 (aarch64_st3<mode>_dreg): Likewise.
1611 (aarch64_st3<mode>_dreg): Likewise.
1612 (aarch64_st4<mode>_dreg): Likewise.
1613 (aarch64_st4<mode>_dreg): Likewise.
1614 (*aarch64_simd_ld1r<mode>): Likewise.
1615 (aarch64_frecpe<mode>): Likewise.
1616 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1617 (aarch64_frecps<mode>): Likewise.
1619 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1621 * config/arm/iterators.md (V_elem_ch): New.
1623 (VQH_type): Likewise.
1624 * config/arm/arm.md (is_neon_type): New.
1625 (conds): Use is_neon_type.
1626 (anddi3_insn): Update type attribute.
1627 (xordi3_insn): Likewise.
1628 (one_cmpldi2): Likewise.
1629 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1630 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1631 (*movmisalign<mode>_neon_store): Likewise.
1632 (*movmisalign<mode>_neon_load): Likewise.
1633 (vec_set<mode>_internal): Likewise.
1634 (vec_set<mode>_internal): Likewise.
1635 (vec_setv2di_internal): Likewise.
1636 (vec_extract<mode>): Likewise.
1637 (vec_extract<mode>): Likewise.
1638 (vec_extractv2di): Likewise.
1639 (*add<mode>3_neon): Likewise.
1640 (adddi3_neon): Likewise.
1641 (*sub<mode>3_neon): Likewise.
1642 (subdi3_neon): Likewise.
1643 (fma<VCVTF:mode>4): Likewise.
1644 (fma<VCVTF:mode>4_intrinsic): Likewise.
1645 (*fmsub<VCVTF:mode>4): Likewise.
1646 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1647 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1648 (ior<mode>3): Likewise.
1649 (and<mode>3): Likewise.
1650 (orn<mode>3_neon): Likewise.
1651 (orndi3_neon): Likewise.
1652 (bic<mode>3_neon): Likewise.
1653 (bicdi3_neon): Likewise.
1654 (xor<mode>3): Likewise.
1655 (one_cmpl<mode>2): Likewise.
1656 (abs<mode>2): Likewise.
1657 (neg<mode>2): Likewise.
1658 (negdi2_neon): Likewise.
1659 (*umin<mode>3_neon): Likewise.
1660 (*umax<mode>3_neon): Likewise.
1661 (*smin<mode>3_neon): Likewise.
1662 (*smax<mode>3_neon): Likewise.
1663 (vashl<mode>3): Likewise.
1664 (vashr<mode>3_imm): Likewise.
1665 (vlshr<mode>3_imm): Likewise.
1666 (ashl<mode>3_signed): Likewise.
1667 (ashl<mode>3_unsigned): Likewise.
1668 (neon_load_count): Likewise.
1669 (ashldi3_neon_noclobber): Likewise.
1670 (ashldi3_neon): Likewise.
1671 (signed_shift_di3_neon): Likewise.
1672 (unsigned_shift_di3_neon): Likewise.
1673 (ashrdi3_neon_imm_noclobber): Likewise.
1674 (lshrdi3_neon_imm_noclobber): Likewise.
1675 (<shift>di3_neon): Likewise.
1676 (widen_ssum<mode>3): Likewise.
1677 (widen_usum<mode>3): Likewise.
1678 (quad_halves_<code>v4si): Likewise.
1679 (quad_halves_<code>v4sf): Likewise.
1680 (quad_halves_<code>v8hi): Likewise.
1681 (quad_halves_<code>v16qi): Likewise.
1682 (reduc_splus_v2di): Likewise.
1683 (neon_vpadd_internal<mode>): Likewise.
1684 (neon_vpsmin<mode>): Likewise.
1685 (neon_vpsmax<mode>): Likewise.
1686 (neon_vpumin<mode>): Likewise.
1687 (neon_vpumax<mode>): Likewise.
1688 (*ss_add<mode>_neon): Likewise.
1689 (*us_add<mode>_neon): Likewise.
1690 (*ss_sub<mode>_neon): Likewise.
1691 (*us_sub<mode>_neon): Likewise.
1692 (neon_vadd<mode>_unspec): Likewise.
1693 (neon_vaddl<mode>): Likewise.
1694 (neon_vaddw<mode>): Likewise.
1695 (neon_vhadd<mode>): Likewise.
1696 (neon_vqadd<mode>): Likewise.
1697 (neon_vaddhn<mode>): Likewise.
1698 (neon_vmul<mode>): Likewise.
1699 (neon_vfms<VCVTF:mode>): Likewise.
1700 (neon_vmlal<mode>): Likewise.
1701 (neon_vmls<mode>): Likewise.
1702 (neon_vmlsl<mode>): Likewise.
1703 (neon_vqdmulh<mode>): Likewise.
1704 (neon_vqdmlal<mode>): Likewise.
1705 (neon_vqdmlsl<mode>): Likewise.
1706 (neon_vmull<mode>): Likewise.
1707 (neon_vqdmull<mode>): Likewise.
1708 (neon_vsub<mode>_unspec): Likewise.
1709 (neon_vsubl<mode>): Likewise.
1710 (neon_vsubw<mode>): Likewise.
1711 (neon_vqsub<mode>): Likewise.
1712 (neon_vhsub<mode>): Likewise.
1713 (neon_vsubhn<mode>): Likewise.
1714 (neon_vceq<mode>): Likewise.
1715 (neon_vcge<mode>): Likewise.
1716 (neon_vcgeu<mode>): Likewise.
1717 (neon_vcgt<mode>): Likewise.
1718 (neon_vcgtu<mode>): Likewise.
1719 (neon_vcle<mode>): Likewise.
1720 (neon_vclt<mode>): Likewise.
1721 (neon_vcage<mode>): Likewise.
1722 (neon_vcagt<mode>): Likewise.
1723 (neon_vtst<mode>): Likewise.
1724 (neon_vabd<mode>): Likewise.
1725 (neon_vabdl<mode>): Likewise.
1726 (neon_vaba<mode>): Likewise.
1727 (neon_vabal<mode>): Likewise.
1728 (neon_vmax<mode>): Likewise.
1729 (neon_vmin<mode>): Likewise.
1730 (neon_vpaddl<mode>): Likewise.
1731 (neon_vpadal<mode>): Likewise.
1732 (neon_vpmax<mode>): Likewise.
1733 (neon_vpmin<mode>): Likewise.
1734 (neon_vrecps<mode>): Likewise.
1735 (neon_vrsqrts<mode>): Likewise.
1736 (neon_vqabs<mode>): Likewise.
1737 (neon_vqneg<mode>): Likewise.
1738 (neon_vcls<mode>): Likewise.
1739 (clz<mode>2): Likewise.
1740 (popcount<mode>2): Likewise.
1741 (neon_vrecpe<mode>): Likewise.
1742 (neon_vrsqrte<mode>): Likewise.
1743 (neon_vget_lane<mode>_sext_internal): Likewise.
1744 (neon_vget_lane<mode>_zext_internal): Likewise.
1745 (neon_vdup_n<mode>): Likewise.
1746 (neon_vdup_n<mode>): Likewise.
1747 (neon_vdup_nv2di): Likewise.
1748 (neon_vdup_lane<mode>_interal): Likewise.
1749 (*neon_vswp<mode>): Likewise.
1750 (neon_vcombine<mode>): Likewise.
1751 (float<mode><V_cvtto>2): Likewise.
1752 (floatuns<mode><V_cvtto>2): Likewise.
1753 (fix_trunc<mode><V_cvtto>2): Likewise.
1754 (fixuns_trunc<mode><V_cvtto>2
1755 (neon_vcvt<mode>): Likewise.
1756 (neon_vcvt<mode>): Likewise.
1757 (neon_vcvtv4sfv4hf): Likewise.
1758 (neon_vcvtv4hfv4sf): Likewise.
1759 (neon_vcvt_n<mode>): Likewise.
1760 (neon_vcvt_n<mode>): Likewise.
1761 (neon_vmovn<mode>): Likewise.
1762 (neon_vqmovn<mode>): Likewise.
1763 (neon_vqmovun<mode>): Likewise.
1764 (neon_vmovl<mode>): Likewise.
1765 (neon_vmul_lane<mode>): Likewise.
1766 (neon_vmul_lane<mode>): Likewise.
1767 (neon_vmull_lane<mode>): Likewise.
1768 (neon_vqdmull_lane<mode>): Likewise.
1769 (neon_vqdmulh_lane<mode>): Likewise.
1770 (neon_vqdmulh_lane<mode>): Likewise.
1771 (neon_vmla_lane<mode>): Likewise.
1772 (neon_vmla_lane<mode>): Likewise.
1773 (neon_vmlal_lane<mode>): Likewise.
1774 (neon_vqdmlal_lane<mode>): Likewise.
1775 (neon_vmls_lane<mode>): Likewise.
1776 (neon_vmls_lane<mode>): Likewise.
1777 (neon_vmlsl_lane<mode>): Likewise.
1778 (neon_vqdmlsl_lane<mode>): Likewise.
1779 (neon_vext<mode>): Likewise.
1780 (neon_vrev64<mode>): Likewise.
1781 (neon_vrev32<mode>): Likewise.
1782 (neon_vrev16<mode>): Likewise.
1783 (neon_vbsl<mode>_internal): Likewise.
1784 (neon_vshl<mode>): Likewise.
1785 (neon_vqshl<mode>): Likewise.
1786 (neon_vshr_n<mode>): Likewise.
1787 (neon_vshrn_n<mode>): Likewise.
1788 (neon_vqshrn_n<mode>): Likewise.
1789 (neon_vqshrun_n<mode>): Likewise.
1790 (neon_vshl_n<mode>): Likewise.
1791 (neon_vqshl_n<mode>): Likewise.
1792 (neon_vqshlu_n<mode>): Likewise.
1793 (neon_vshll_n<mode>): Likewise.
1794 (neon_vsra_n<mode>): Likewise.
1795 (neon_vsri_n<mode>): Likewise.
1796 (neon_vsli_n<mode>): Likewise.
1797 (neon_vtbl1v8qi): Likewise.
1798 (neon_vtbl2v8qi): Likewise.
1799 (neon_vtbl3v8qi): Likewise.
1800 (neon_vtbl4v8qi): Likewise.
1801 (neon_vtbl1v16qi): Likewise.
1802 (neon_vtbl2v16qi): Likewise.
1803 (neon_vcombinev16qi): Likewise.
1804 (neon_vtbx1v8qi): Likewise.
1805 (neon_vtbx2v8qi): Likewise.
1806 (neon_vtbx3v8qi): Likewise.
1807 (neon_vtbx4v8qi): Likewise.
1808 (*neon_vtrn<mode>_insn): Likewise.
1809 (*neon_vzip<mode>_insn): Likewise.
1810 (*neon_vuzp<mode>_insn): Likewise.
1811 (neon_vld1<mode>): Likewise.
1812 (neon_vld1_lane<mode>): Likewise.
1813 (neon_vld1_lane<mode>): Likewise.
1814 (neon_vld1_dup<mode>): Likewise.
1815 (neon_vld1_dup<mode>): Likewise.
1816 (neon_vld1_dupv2di): Likewise.
1817 (neon_vst1<mode>): Likewise.
1818 (neon_vst1_lane<mode>): Likewise.
1819 (neon_vst1_lane<mode>): Likewise.
1820 (neon_vld2<mode>): Likewise.
1821 (neon_vld2<mode>): Likewise.
1822 (neon_vld2_lane<mode>): Likewise.
1823 (neon_vld2_lane<mode>): Likewise.
1824 (neon_vld2_dup<mode>): Likewise.
1825 (neon_vst2<mode>): Likewise.
1826 (neon_vst2<mode>): Likewise.
1827 (neon_vst2_lane<mode>): Likewise.
1828 (neon_vst2_lane<mode>): Likewise.
1829 (neon_vld3<mode>): Likewise.
1830 (neon_vld3qa<mode>): Likewise.
1831 (neon_vld3qb<mode>): Likewise.
1832 (neon_vld3_lane<mode>): Likewise.
1833 (neon_vld3_lane<mode>): Likewise.
1834 (neon_vld3_dup<mode>): Likewise.
1835 (neon_vst3<mode>): Likewise.
1836 (neon_vst3qa<mode>): Likewise.
1837 (neon_vst3qb<mode>): Likewise.
1838 (neon_vst3_lane<mode>): Likewise.
1839 (neon_vst3_lane<mode>): Likewise.
1840 (neon_vld4<mode>): Likewise.
1841 (neon_vld4qa<mode>): Likewise.
1842 (neon_vld4qb<mode>): Likewise.
1843 (neon_vld4_lane<mode>): Likewise.
1844 (neon_vld4_lane<mode>): Likewise.
1845 (neon_vld4_dup<mode>): Likewise.
1846 (neon_vst4<mode>): Likewise.
1847 (neon_vst4qa<mode>): Likewise.
1848 (neon_vst4qb<mode>): Likewise.
1849 (neon_vst4_lane<mode>): Likewise.
1850 (neon_vst4_lane<mode>): Likewise.
1851 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1852 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1853 (neon_vec_<US>mult_lo_<mode>): Likewise.
1854 (neon_vec_<US>mult_hi_<mode>): Likewise.
1855 (neon_vec_<US>shiftl_<mode>): Likewise.
1856 (neon_unpack<US>_<mode>): Likewise.
1857 (neon_vec_<US>mult_<mode>): Likewise.
1858 (vec_pack_trunc_<mode>): Likewise.
1859 (neon_vec_pack_trunc_<mode>): Likewise.
1860 (neon_vabd<mode>_2): Likewise.
1861 (neon_vabd<mode>_3): Likewise.
1863 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1865 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1866 (load_pair): Update type attribute.
1867 (store_pair): Update type attribute.
1868 * config/aarch64/iterators.md (q): New.
1870 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1872 * config/arm/types.md: Add new types for Neon insns.
1874 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1875 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1876 Sergey Lega <sergey.s.lega@intel.com>
1877 Anna Tikhonova <anna.tikhonova@intel.com>
1878 Ilya Tocar <ilya.tocar@intel.com>
1879 Andrey Turetskiy <andrey.turetskiy@intel.com>
1880 Ilya Verbin <ilya.verbin@intel.com>
1881 Kirill Yukhin <kirill.yukhin@intel.com>
1882 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1884 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1885 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1886 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1888 (srcp14<mode>): Ditto.
1889 (rsqrt14<mode>): Ditto.
1890 (rsqrt14<mode>): Ditto.
1891 (avx512f_vmscalef<mode>): Ditto.
1892 (avx512f_scalef<mode>): Ditto.
1893 (avx512f_getexp<mode>): Ditto.
1894 (avx512f_sgetexp<mode>): Ditto.
1895 (avx512f_fixupimm<mode>): Ditto.
1896 (avx512f_sfixupimm<mode>): Ditto.
1897 (avx512f_rndscale<mode>): Ditto.
1898 (*avx512er_exp2<mode>): Ditto.
1899 (*avx512er_rcp28<mode>): Ditto.
1900 (avx512er_rsqrt28<mode>): Ditto.
1901 (avx512f_getmant<mode>): Ditto.
1902 (avx512f_getmant<mode>): Ditto.
1903 (avx512f_rndscale<mode>): Fix formatting.
1905 2013-10-15 Martin Jambor <mjambor@suse.cz>
1907 * ipa-utils.h (ipa_edge_within_scc): Declare.
1908 * ipa-cp.c (edge_within_scc): Moved...
1909 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1911 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1912 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1913 Sergey Lega <sergey.s.lega@intel.com>
1914 Anna Tikhonova <anna.tikhonova@intel.com>
1915 Ilya Tocar <ilya.tocar@intel.com>
1916 Andrey Turetskiy <andrey.turetskiy@intel.com>
1917 Ilya Verbin <ilya.verbin@intel.com>
1918 Kirill Yukhin <kirill.yukhin@intel.com>
1919 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1921 * config/i386/predicates.md (const_8_to_15_operand): New.
1922 (const_16_to_31_operand): Ditto.
1923 * config/i386/sse.md (V8FI): New.
1925 (reduc_splus_v8df): Ditto.
1926 (reduc_splus_v16sf): Ditto.
1927 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1928 (vec_extract_hi_<mode>): Ditto.
1929 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1930 (vec_set_lo_<mode>): Ditto.
1931 (vec_set_hi_<mode>): Ditto.
1932 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1933 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1934 (avx512f_pshufd_1): Ditto.
1935 (avx512f_broadcast<mode>): Ditto.
1936 (avx512f_broadcast<mode>): Ditto.
1937 (define_split): Split vec_extract_lo into move.
1938 (ssequartermode): Ditto.
1939 (ssedoublemode): Extened with wider modes.
1940 (vec_extract_lo_<mode>): Ditto.
1942 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1943 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1944 Sergey Lega <sergey.s.lega@intel.com>
1945 Anna Tikhonova <anna.tikhonova@intel.com>
1946 Ilya Tocar <ilya.tocar@intel.com>
1947 Andrey Turetskiy <andrey.turetskiy@intel.com>
1948 Ilya Verbin <ilya.verbin@intel.com>
1949 Kirill Yukhin <kirill.yukhin@intel.com>
1950 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1952 * config/i386/predicates.md (register_or_constm1_operand): New.
1953 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1954 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1955 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1956 UNSPEC_SCATTER_PREFETCH
1958 (avx512f_ucmp<mode>3): Ditto.
1959 (avx512f_vternlog<mode>): Ditto.
1960 (avx512f_align<mode>): Ditto.
1961 (<shift_insn><mode>3): Ditto.
1962 (avx512f_<rotate>v<mode>): Ditto.
1963 (avx512f_<rotate><mode>): Ditto.
1964 (avx512f_eq<mode>3): Ditto.
1965 (avx512f_eq<mode>3_1): Ditto.
1966 (avx512f_gt<mode>3): Ditto.
1967 (avx512f_testm<mode>3): Ditto.
1968 (avx512f_testnm<mode>3): Ditto.
1969 (avx512pf_gatherpf<mode>): Ditto.
1970 (*avx512pf_gatherpf<mode>_mask): Ditto.
1971 (*avx512pf_gatherpf<mode>): Ditto.
1972 (avx512pf_scatterpf<mode>): Ditto.
1973 (*avx512pf_scatterpf<mode>_mask): Ditto.
1974 (*avx512pf_scatterpf<mode>): Ditto.
1975 (avx512f_vec_dup_gpr<mode>): Ditto.
1976 (clz<mode>2): Ditto.
1977 (conflict<mode>): Ditto.
1978 (REDUC_SMINMAX_MODE): Extened with wider modes.
1979 (reduc_<code>_<mode>): Ditto.
1980 (vlshr<mode>3): Ditto.
1981 (vashl<mode>3): Ditto.
1983 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1984 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1985 Sergey Lega <sergey.s.lega@intel.com>
1986 Anna Tikhonova <anna.tikhonova@intel.com>
1987 Ilya Tocar <ilya.tocar@intel.com>
1988 Andrey Turetskiy <andrey.turetskiy@intel.com>
1989 Ilya Verbin <ilya.verbin@intel.com>
1990 Kirill Yukhin <kirill.yukhin@intel.com>
1991 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1993 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
1996 (avx512fmaskmode): Ditto.
1997 (bcstscalarsuff): Ditto.
1998 (avx512f_blendm<mode>): Ditto.
1999 (cmp_imm_predicate): Ditto.
2000 (avx512f_cmp<mode>3): Ditto.
2001 (avx512f_vec_dup<mode>): Ditto.
2002 (avx512f_vec_dup_mem<mode>): Ditto.
2003 (avx512f_vpermi2var<mode>3): Ditto.
2004 (avx512f_vpermt2var<mode>3): Ditto.
2005 (vec_init<mode>): Ditto.
2006 (avx512f_gathersi<mode>): Ditto.
2007 (*avx512f_gathersi<mode>): Ditto.
2008 (*avx512f_gathersi<mode>_2): Ditto.
2009 (avx512f_gatherdi<mode>): Ditto.
2010 (*avx512f_gatherdi<mode>): Ditto.
2011 (*avx512f_gatherdi<mode>_2): Ditto.
2012 (avx512f_scattersi<mode>): Ditto.
2013 (*avx512f_scattersi<mode>): Ditto.
2014 (avx512f_scatterdi<mode>): Ditto.
2015 (*avx512f_scatterdi<mode>): Ditto.
2016 (sseintprefix): Extened with wider modes.
2017 (VEC_GATHER_IDXSI): Ditto.
2018 (VEC_GATHER_IDXDI): Ditto.
2019 (VEC_GATHER_SRCDI): Ditto.
2021 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
2022 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2024 * config/arm/t-aprofile: New file.
2025 * config.gcc: Handle --with-multilib-list option.
2027 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
2029 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
2032 2013-10-15 Richard Biener <rguenther@suse.de>
2034 * tree-loop-distribution.c (build_empty_rdg): Inline into
2036 (rdg_flag_vertex): Inline into single user.
2037 (rdg_flag_vertex_and_dependent): Likewise.
2038 (build_rdg_partition_for_vertex): Remove processed bitmap.
2039 (rdg_build_partitions): Simplify.
2041 2013-10-15 Richard Biener <rguenther@suse.de>
2043 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
2044 Restructure forwarding through conversions and copies to
2045 avoid performing copy-propagation the wrong way. Adjust
2046 recursion invocations.
2047 (forward_propagate_addr_expr): Add argument stating if we
2048 are recursing from a single-use.
2049 (ssa_forward_propagate_and_combine): Adjust.
2051 2013-10-14 David Malcolm <dmalcolm@redhat.com>
2053 * dumpfile.h (gcc::dump_manager): New class, to hold state
2054 relating to dumpfile management.
2055 (get_dump_file_name): Remove in favor of method of dump_manager.
2056 (dump_initialized_p): Likewise.
2057 (dump_start): Likewise.
2058 (dump_finish): Likewise.
2059 (dump_switch_p): Likewise.
2060 (dump_register): Likewise.
2061 (get_dump_file_info): Likewise.
2062 * context.c (gcc::context::context): Construct the dump_manager
2064 * context.h (gcc::context::get_dumps): New.
2065 (gcc::context::m_dumps): New.
2066 * coverage.c (coverage_init): Port to dump_manager API.
2067 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
2068 (extra_dump_files_in_use): Likewise.
2069 (extra_dump_files_alloced): Likewise.
2070 (gcc::dump_manager::dump_manager): New.
2071 (dump_register): Convert to...
2072 (gcc::dump_manager::dump_register): ...method, replacing
2073 function-static next_dump with m_next_dump field.
2074 (get_dump_file_info): Convert to...
2075 (gcc::dump_manager::get_dump_file_info): ...method.
2076 (get_dump_file_name): Convert to...
2077 (gcc::dump_manager::get_dump_file_name): ...method.
2078 (dump_start): Convert to...
2079 (gcc::dump_manager::dump_start): ...method.
2080 (dump_finish): Convert to...
2081 (gcc::dump_manager::dump_finish): ...method.
2082 (dump_begin): Replace body with...
2083 (gcc::dump_manager::dump_begin): ...new method.
2084 (dump_phase_enabled_p): Convert to...
2085 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2086 (dump_phase_enabled_p): Convert to...
2087 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2088 (dump_initialized_p): Convert to...
2089 (gcc::dump_manager::dump_initialized_p): ...method.
2090 (dump_flag_name): Replace body with...
2091 (gcc::dump_manager::dump_flag_name): ...new method.
2092 (dump_enable_all): Convert to...
2093 (gcc::dump_manager::dump_enable_all): ...new method.
2094 (opt_info_enable_passes): Convert to...
2095 (gcc::dump_manager::opt_info_enable_passes): ...new method.
2096 (dump_switch_p_1): Convert to...
2097 (gcc::dump_manager::dump_switch_p_1): ...new method.
2098 (dump_switch_p): Convert to...
2099 (gcc::dump_manager::dump_switch_p): ...new method.
2100 (opt_info_switch_p): Port to dump_manager API.
2101 (enable_rtl_dump_file): Likewise.
2102 * opts-global.c (handle_common_deferred_options): Port to new
2104 * passes.c (pass_manager::finish_optimization_passes): Likewise.
2105 (pass_manager::register_one_dump_file): Likewise.
2106 (pass_manager::register_pass): Likewise.
2107 (pass_init_dump_file): Likewise.
2108 (pass_fini_dump_file): Likewise.
2109 * statistics.c (statistics_early_init): Likewise.
2111 2013-10-14 Richard Biener <rguenther@suse.de>
2113 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
2114 iterative_hash_canonical_type, gimple_canonical_type_hash,
2115 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
2116 gimple_register_canonical_type, print_gimple_types_stats,
2117 free_gimple_type_tables): Move to lto/lto.c
2118 (gt-gimple.h): Do not include.
2119 * gimple.h (gimple_register_canonical_type,
2120 print_gimple_types_stats, free_gimple_type_tables): Remove.
2121 * Makefile.in (GTFILES): Remove gimple.c.
2123 2013-10-14 Travis Snoozy <quandary@remstate.com>
2126 * config/msp430/msp430.c (msp430_option_override): Correct thinko
2127 scanning for msp430x targets.
2129 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
2132 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2133 (registers_ok_for_ldd_peep): Move around.
2134 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2135 * config/sparc/sparc.md (widening peepholes): Use it.
2137 2013-10-14 Richard Biener <rguenther@suse.de>
2141 * gimple.c (iterative_hash_canonical_type): Make sure to
2142 record the hash into the correct hashtable slot.
2144 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2146 PR rtl-optimization/58662
2147 * combine.c (try_combine): Take into account death nodes on I2 when
2148 splitting a PARALLEL of two independent SETs. Fix dump message.
2150 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2153 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2154 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2155 * config/sh/t-sh (sh_treg_combine.o): New entry.
2156 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2157 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2158 (register_sh_passes): New function. Register sh_treg_combine pass.
2159 (sh_option_override): Invoke it.
2160 (sh_canonicalize_comparison): Handle op0_preserve_value.
2161 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2162 opportunities. Canonicalize branch condition.
2163 (nott): Allow only if pseudos can be created for non-SH2A.
2165 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2168 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2169 (ix86_expand_movmem): Replace copy_addr_to_reg with
2170 ix86_copy_addr_to_reg.
2171 (ix86_expand_setmem): Likewise.
2173 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2175 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2176 provided by ix86_fp_compare_mode instead of CCFPUmode.
2178 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2180 * config/aarch64/arm_neon.h
2181 (vtbx<1,3>_<psu>8): Fix register constriants.
2183 2013-10-11 Jeff Law <law@redhat.com>
2185 PR tree-optimization/58640
2186 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2187 threading paths that cross over two loop entry points.
2189 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2191 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2192 handle vector float as well.
2193 (*vsx_le_perm_load_v4si): Likewise.
2194 (*vsx_le_perm_store_v2di): Likewise.
2195 (*vsx_le_perm_store_v4si): Likewise.
2197 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2199 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2200 directly to circumvent subtract from splat{31} workaround.
2201 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2203 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2204 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2205 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2206 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2207 separate big and little endian logic.
2208 (*altivec_vperm_<mode>_internal): New define_insn.
2209 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2210 separate big and little endian logic.
2211 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2212 (vec_permv16qi): Add little endian logic.
2214 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2216 * doc/extend.texi (returns_nonnull): Remove arguments.
2218 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2219 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2220 Sergey Lega <sergey.s.lega@intel.com>
2221 Anna Tikhonova <anna.tikhonova@intel.com>
2222 Ilya Tocar <ilya.tocar@intel.com>
2223 Andrey Turetskiy <andrey.turetskiy@intel.com>
2224 Ilya Verbin <ilya.verbin@intel.com>
2225 Kirill Yukhin <kirill.yukhin@intel.com>
2226 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2228 * config/i386/sse.md (VI48F_256_512): New.
2229 (avx2_permvar<mode>): Change to ...
2230 (<avx2_avx512f>_permvar<mode>): This.
2232 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2233 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2234 Sergey Lega <sergey.s.lega@intel.com>
2235 Anna Tikhonova <anna.tikhonova@intel.com>
2236 Ilya Tocar <ilya.tocar@intel.com>
2237 Andrey Turetskiy <andrey.turetskiy@intel.com>
2238 Ilya Verbin <ilya.verbin@intel.com>
2239 Kirill Yukhin <kirill.yukhin@intel.com>
2240 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2242 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2243 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2244 * config/i386/sse.md (VI4_AVX): New.
2245 (sf2simodelower): Ditto.
2246 (sse2_cvtps2dq): Change to ...
2247 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2249 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2250 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2251 Sergey Lega <sergey.s.lega@intel.com>
2252 Anna Tikhonova <anna.tikhonova@intel.com>
2253 Ilya Tocar <ilya.tocar@intel.com>
2254 Andrey Turetskiy <andrey.turetskiy@intel.com>
2255 Ilya Verbin <ilya.verbin@intel.com>
2256 Kirill Yukhin <kirill.yukhin@intel.com>
2257 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2259 * config/i386/sse.md (V_512): New.
2261 (vcond<V_512:mode><VF_512:mode>): Ditto.
2262 (vcond<V_512:mode><VI_512:mode>): Ditto.
2263 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2265 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2266 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2267 Sergey Lega <sergey.s.lega@intel.com>
2268 Anna Tikhonova <anna.tikhonova@intel.com>
2269 Ilya Tocar <ilya.tocar@intel.com>
2270 Andrey Turetskiy <andrey.turetskiy@intel.com>
2271 Ilya Verbin <ilya.verbin@intel.com>
2272 Kirill Yukhin <kirill.yukhin@intel.com>
2273 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2275 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2276 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2278 (fma<mode>4): Removed condition.
2279 (fms<mode>4): Ditto.
2280 (fnma<mode>4): Ditto.
2281 (fnms<mode>4): Ditto.
2282 (fma4i_fmadd_<mode>): Ditto.
2283 (*fma_fmadd_<mode>): Ditto.
2284 (*fma_fmsub_<mode>): Ditto.
2285 (*fma_fnmadd_<mode>): Ditto.
2286 (*fma_fnmsub_<mode>): Ditto.
2287 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2288 (*fma_fmaddsub_<mode>): Ditto.
2289 (*fma_fmsubadd_<mode>): Ditto.
2290 (*fmai_fmadd_<mode>): Ditto.
2291 (*fmai_fmsub_<mode>): Ditto.
2292 (*fmai_fnmadd_<mode>): Ditto.
2293 (*fmai_fnmsub_<mode>): Ditto.
2295 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2296 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2297 Sergey Lega <sergey.s.lega@intel.com>
2298 Anna Tikhonova <anna.tikhonova@intel.com>
2299 Ilya Tocar <ilya.tocar@intel.com>
2300 Andrey Turetskiy <andrey.turetskiy@intel.com>
2301 Ilya Verbin <ilya.verbin@intel.com>
2302 Kirill Yukhin <kirill.yukhin@intel.com>
2303 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2305 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2306 (VI124_256): Changed to ...
2307 (VI124_256_48_512): This.
2308 (ssepackmode): Extended with wider modes.
2309 (<code><mode>3): Changed iterator.
2310 (*avx2_<code><mode>3): Ditto.
2311 (vec_pack_trunc_<mode>): Ditto.
2313 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2314 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2315 Sergey Lega <sergey.s.lega@intel.com>
2316 Anna Tikhonova <anna.tikhonova@intel.com>
2317 Ilya Tocar <ilya.tocar@intel.com>
2318 Andrey Turetskiy <andrey.turetskiy@intel.com>
2319 Ilya Verbin <ilya.verbin@intel.com>
2320 Kirill Yukhin <kirill.yukhin@intel.com>
2321 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2323 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2324 (VI8F_256_512): Ditto.
2325 (abs<mode>2): Changed iterator.
2326 (avx2_perm<mode>): Changed to ...
2327 (<avx2_avx512f>_perm<mode>): This.
2328 (avx2_perm<mode>_1): Changed to ...
2329 (<avx2_avx512f>_perm<mode>_1): This.
2331 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2332 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2333 Sergey Lega <sergey.s.lega@intel.com>
2334 Anna Tikhonova <anna.tikhonova@intel.com>
2335 Ilya Tocar <ilya.tocar@intel.com>
2336 Andrey Turetskiy <andrey.turetskiy@intel.com>
2337 Ilya Verbin <ilya.verbin@intel.com>
2338 Kirill Yukhin <kirill.yukhin@intel.com>
2339 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2341 * config/i386/sse.md (VI48_AVX512F): New.
2342 (VI48_AVX2): Changed to ...
2343 (VI48_AVX2_48_AVX512F): This.
2344 (avx2_ashrv<mode>): Changed to ...
2345 (<avx2_avx512f>_ashrv<mode>): This.
2346 (avx2_<shift_insn>v<mode>): Changed to ...
2347 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2349 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2350 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2351 Sergey Lega <sergey.s.lega@intel.com>
2352 Anna Tikhonova <anna.tikhonova@intel.com>
2353 Ilya Tocar <ilya.tocar@intel.com>
2354 Andrey Turetskiy <andrey.turetskiy@intel.com>
2355 Ilya Verbin <ilya.verbin@intel.com>
2356 Kirill Yukhin <kirill.yukhin@intel.com>
2357 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2359 * config/i386/sse.md (VI4_AVX512F): New.
2360 (VI8_AVX2_AVX512F): Ditto.
2361 (mul<mode>3): Extended with wider modes.
2362 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2363 (mul<mode>3): Ditto.
2364 (vec_widen_<s>mult_odd_<mode>): Ditto.
2366 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2367 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2368 Sergey Lega <sergey.s.lega@intel.com>
2369 Anna Tikhonova <anna.tikhonova@intel.com>
2370 Ilya Tocar <ilya.tocar@intel.com>
2371 Andrey Turetskiy <andrey.turetskiy@intel.com>
2372 Ilya Verbin <ilya.verbin@intel.com>
2373 Kirill Yukhin <kirill.yukhin@intel.com>
2374 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2376 * config/i386/sse.md (VI2_AVX512F): New.
2377 (VI124_AVX512F): Ditto.
2378 (sseunpackmode): Extended with wider modes.
2379 (sseunpackfltmode): Ditto.
2380 (vec_unpacks_float_hi_<mode>): Ditto.
2381 (vec_unpacks_float_lo_<mode>): Ditto.
2382 (vec_unpacku_float_hi_<mode>): Ditto.
2383 (vec_unpacku_float_lo_<mode>): Ditto.
2384 (vec_unpacks_lo_<mode>): Ditto.
2385 (vec_unpacks_hi_<mode>): Ditto.
2386 (vec_unpacku_lo_<mode>): Ditto.
2387 (vec_unpacku_hi_<mode>): Ditto.
2389 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2390 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2391 Sergey Lega <sergey.s.lega@intel.com>
2392 Anna Tikhonova <anna.tikhonova@intel.com>
2393 Ilya Tocar <ilya.tocar@intel.com>
2394 Andrey Turetskiy <andrey.turetskiy@intel.com>
2395 Ilya Verbin <ilya.verbin@intel.com>
2396 Kirill Yukhin <kirill.yukhin@intel.com>
2397 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2399 * config/i386/i386.md (multdiv): New.
2400 (multdiv_mnemonic): Ditto.
2401 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2402 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2403 (<sse>_vmdiv<mode>3): Removed.
2405 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2406 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2407 Sergey Lega <sergey.s.lega@intel.com>
2408 Anna Tikhonova <anna.tikhonova@intel.com>
2409 Ilya Tocar <ilya.tocar@intel.com>
2410 Andrey Turetskiy <andrey.turetskiy@intel.com>
2411 Ilya Verbin <ilya.verbin@intel.com>
2412 Kirill Yukhin <kirill.yukhin@intel.com>
2413 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2415 * config/i386/sse.md (V): Extended with wider modes.
2417 (ssehalfvecmode): Ditto.
2419 (ssepackfltmode): Ditto.
2420 (avx_vec_concat<mode>): Ditto.
2421 (V_256_512): New iterator.
2422 (VF2_512_256): Ditto.
2423 (si2dfmode): New attribute.
2424 (si2dfmodelower): Ditto.
2426 (concat_tg_mode): Ditto.
2427 (floatv4siv4df2): Changed to ...
2428 (float<si2dfmodelower><mode>2): This.
2429 (avx_cvtps2pd256): Changed to ...
2430 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2431 (vec_pack_trunc_v4df): Changed to ...
2432 (vec_pack_trunc_<mode>): This.
2433 (avx_vpermil<mode>): Changed to ...
2434 (<sse2_avx_avx512f>_vpermil<mode>): This.
2435 (<fixsuffix>fix_truncv8dfv8si2): New.
2436 (vec_pack_sfix_trunc_v8df): Ditto.
2437 (avx512f_rndscale<mode>): Ditto.
2438 (avx512f_roundpd512): Ditto.
2439 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2441 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2442 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2443 Sergey Lega <sergey.s.lega@intel.com>
2444 Anna Tikhonova <anna.tikhonova@intel.com>
2445 Ilya Tocar <ilya.tocar@intel.com>
2446 Andrey Turetskiy <andrey.turetskiy@intel.com>
2447 Ilya Verbin <ilya.verbin@intel.com>
2448 Kirill Yukhin <kirill.yukhin@intel.com>
2449 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2451 * config/i386/i386.md (any_fix): New iterator.
2452 (fixsuffix): New attribute.
2453 * config/i386/sse.md (VF1): Extened with wider modes.
2457 (sseintvecmodelower): Ditto.
2458 (ssescalarmode): Ditto.
2459 (ssescalarnum): Ditto.
2461 (ssexmmmode): Ditto.
2462 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2463 (<sse>_rcp<mode>2): Change iterator.
2464 (rsqrt<mode>2): Ditto.
2465 (<sse>_rsqrt<mode>2): Ditto.
2466 (avx2_vec_dup<mode>): Ditto.
2467 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2468 (round<mode>2_sfix): Ditto.
2469 (avx2_pbroadcast<mode>): Ditto.
2470 (*andnot<mode>3): Handle XI mode.
2471 (*<code><mode>3): Ditto.
2472 (AVXTOSSEMODE): Removed.
2473 (avx_vpermil<mode>): Changed to ...
2474 (<sse2_avx_avx512f>_vpermil<mode>): This.
2476 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2477 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2478 Sergey Lega <sergey.s.lega@intel.com>
2479 Anna Tikhonova <anna.tikhonova@intel.com>
2480 Ilya Tocar <ilya.tocar@intel.com>
2481 Andrey Turetskiy <andrey.turetskiy@intel.com>
2482 Ilya Verbin <ilya.verbin@intel.com>
2483 Kirill Yukhin <kirill.yukhin@intel.com>
2484 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2486 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2487 (<sse>_comi): Ditto.
2488 (<sse>_ucomi): Ditto.
2489 (sse_cvtss2siq_2): Ditto.
2490 (sse2_cvtsd2si): Ditto.
2491 (sse2_cvtsd2siq): Ditto.
2492 (sse2_cvttsd2si): Ditto.
2493 (sse2_cvttsd2siq): Ditto.
2494 (<shift_insn><mode>3): Ditto.
2495 (sse2_cvtsi2sdq): Update constraint and prefix.
2496 (sse_cvtsi2ss): Update prefix.
2497 (sse_cvtsi2ssq): Ditto.
2499 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2501 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2502 ignore internal calls.
2504 2013-10-11 Richard Biener <rguenther@suse.de>
2506 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2507 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2508 TREE_OVERFLOW is set.
2510 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2512 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2514 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2515 * gimple.h: Likewise.
2517 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2519 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2522 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2524 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2525 * configure: Regenerate.
2527 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2529 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2530 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2531 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2532 similarly pp_right_paren.
2533 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2534 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2535 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2536 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2538 (omp_declare_simd_clauses_equal,
2539 omp_remove_redundant_declare_simd_attrs): New functions.
2540 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2541 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2542 * tree.h (OMP_LOOP_CHECK): Define.
2543 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2544 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2545 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2546 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2547 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2548 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2549 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2550 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2551 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2552 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2553 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2554 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2555 (omp_remove_redundant_declare_simd_attrs): New prototype.
2556 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2557 GIMPLE_OMP_TEAMS): New codes.
2558 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2559 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2561 (target_nesting_level): New variable.
2562 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2563 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2564 for collapse > 1 static schedule unless ordered.
2565 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2566 (determine_parallel_type): Adjust get_ws_args_for caller.
2567 (install_var_field): Handle mask & 4 for double indirection.
2568 (scan_sharing_clauses): Ignore shared clause on teams construct.
2569 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2570 (create_omp_child_function): If inside target or declare target
2571 constructs, set "omp declare target" attribute on the child function.
2572 (find_combined_for): New function.
2573 (scan_omp_parallel): Handle combined loops.
2574 (scan_omp_target, scan_omp_teams): New functions.
2575 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2576 restrictions and set ctx->cancellable for cancellable constructs.
2577 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2578 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2579 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2580 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2581 (omp_clause_aligned_alignment): New function.
2582 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2583 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2584 on teams constructs. Handle user defined reductions and new
2586 (lower_reduction_clauses): Don't set placeholder to address of ref
2587 if it has already the right type.
2588 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2589 (expand_parallel_call): Use the new non-_start suffixed builtins,
2590 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2591 and GOMP_parallel_end after the call.
2592 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2593 (expand_omp_for_init_counts): Handle combined loops.
2594 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2596 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2597 end of cancellable loops.
2598 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2599 Likewise. Handle collapse > 1 loops.
2600 (expand_omp_simd): Handle combined loops.
2601 (expand_omp_for): Add inner_stmt argument, adjust callers of
2602 expand_omp_for* functions, use expand_omp_for_static*chunk even
2603 for collapse > 1 unless ordered.
2604 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2605 of cancellable sections.
2606 (expand_omp_single): Remove need_barrier variable, just rely on
2607 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2608 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2609 (expand_omp_atomic_load, expand_omp_atomic_store,
2610 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2611 (expand_omp_target): New function.
2612 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2613 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2614 (build_omp_regions_1): Immediately close region for
2615 GF_OMP_TARGET_KIND_UPDATE.
2616 (maybe_add_implicit_barrier_cancel): New function.
2617 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2619 (lower_omp_single): Likewise. Add clobber after the barrier.
2620 (lower_omp_taskgroup): New function.
2621 (lower_omp_for): Handle combined loops. Adjust
2622 lower_rec_input_clauses caller. Handle cancellation.
2623 (lower_depend_clauses): New function.
2624 (lower_omp_taskreg): Lower depend clauses. Adjust
2625 lower_rec_input_clauses caller. Add clobber after the call. Handle
2627 (lower_omp_target, lower_omp_teams): New functions.
2628 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2629 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2630 and GOMP_cancellation_point calls.
2631 (lower_omp): Fold stmts inside of target region.
2632 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2633 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2634 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2635 (BT_FN_VOID_OMPFN_PTR_UINT,
2636 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2637 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2638 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2639 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2640 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2641 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2642 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2643 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2644 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2645 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2646 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2647 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2648 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2649 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2650 BUILT_IN_GOMP_PARALLEL_END.
2651 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2652 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2653 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2654 GF_OMP_FOR_KIND_DISTRIBUTE.
2655 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2656 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2657 (dump_gimple_omp_return): Print lhs if it has any.
2658 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2659 gimple_omp_atomic_seq_cst_p.
2660 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2661 and GIMPLE_OMP_TEAMS.
2662 * langhooks.c (lhd_omp_mappable_type): New function.
2663 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2664 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2666 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2667 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2668 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2669 (struct gimplify_omp_ctx): Add combined_loop field.
2670 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2671 on stmts inside of target region.
2672 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2673 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2674 ORT_TARGET and ORT_TARGET_DATA.
2675 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2677 (omp_notice_threadprivate_variable): Complain about threadprivate
2678 variables in target region.
2679 (omp_notice_variable): Complain about vars with non-mappable type
2680 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2681 (omp_check_private): Ignore ORT_TARGET* regions.
2682 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2683 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2684 (find_combined_omp_for): New function.
2685 (gimplify_omp_for): Handle gimplification of combined loops.
2686 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2688 (gimplify_omp_target_update): New function.
2689 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2690 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2691 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2692 (gimplify_body): If fndecl has "omp declare target" attribute, add
2693 implicit ORT_TARGET context around it.
2694 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2695 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2696 * tree-nested.c (convert_nonlocal_reference_stmt,
2697 convert_local_reference_stmt, convert_gimple_call): Handle
2698 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2699 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2700 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2701 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2702 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2703 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2704 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2705 BUILT_IN_GOMP_LOOP_END_CANCEL,
2706 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2707 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2708 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2709 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2710 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2711 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2712 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2713 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2714 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2715 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2716 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2717 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2718 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2719 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2720 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2721 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2722 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2723 gimple_build_omp_teams): New functions.
2724 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2725 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2726 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2727 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2728 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2729 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2730 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2731 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2732 GF_OMP_ATOMIC_SEQ_CST): New.
2733 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2734 gimple_build_omp_teams): New prototypes.
2735 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2736 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2737 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2738 GIMPLE_OMP_SINGLE as end of range.
2739 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2740 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2741 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2742 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2743 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2744 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2745 gimple_omp_target_kind, gimple_omp_target_set_kind,
2746 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2747 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2748 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2749 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2750 gimple_omp_teams_set_clauses): New inlines.
2751 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2752 and GIMPLE_OMP_TASKGROUP.
2753 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2754 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2755 enum omp_clause_proc_bind_kind): New.
2756 (union omp_clause_subcode): Add depend_kind, map_kind and
2757 proc_bind_kind fields.
2758 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2759 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2760 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2761 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2762 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2764 2013-10-10 Teresa Johnson <tejohnson@google.com>
2766 * predict.c (tree_estimate_probability): Add new parameter
2767 for estimate_bb_frequencies.
2768 (estimate_bb_frequencies): Add new parameter to force estimation.
2769 (rebuild_frequencies): When max frequency in function is small,
2770 recompute counts from frequencies.
2771 * predict.h (estimate_bb_frequencies): New parameter.
2773 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2775 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2776 optimizations are disabled.
2778 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2780 * coverage.c (coverage_finish): Fix leak of da_file_name.
2782 2013-10-10 Jan Hubicka <jh@suse.cz>
2784 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2785 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2788 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2791 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2792 if any labels are in FALLTHRU_BB, use a special label emitted
2793 immediately after the asm goto insn rather than label_rtx
2795 (expand_asm_stmt): Adjust caller.
2796 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2797 edge if the last insn in predecessor is a jump with single successor,
2798 but it isn't simplejump_p.
2800 2013-10-10 Richard Biener <rguenther@suse.de>
2802 PR tree-optimization/58656
2803 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2805 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2807 * gimplify.c: Include expr.h and tm_p.h for targets with special
2808 va-arg padding requirements.
2810 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2812 * tree-flow.h: Move some prototypes to gimple.h.
2813 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2814 * gimple.h: Relocate some prototypes from tree-flow.h
2815 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2817 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2818 (build_va_arg_indirect_ref): Relocate and make static.
2819 (std_gimplify_va_arg_expr): Relocate here.
2820 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2821 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2823 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2825 * doc/md.texi: Document the mnemonic attribute.
2827 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2830 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2831 (set_attr_alternative x ...) when searching for user defined
2834 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2836 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2838 2013-10-09 Easwaran Raman <eraman@google.com>
2840 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2841 * cfgexpand.c (defer_stack_allocation): ...use here
2842 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2844 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2846 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2847 (operand_equal_for_value_replacement): New function, extracted from
2848 value_replacement and enhanced to catch more cases.
2849 (value_replacement): Use operand_equal_for_value_replacement.
2851 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2853 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2854 get_max_loop_iterations.
2856 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2858 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2859 Place comment above function.
2861 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2863 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2864 are not related to tree-cfg.c.
2865 * tree-ssa-address.h: New file. Relocate prototypes.
2866 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2867 (addr_for_mem_ref): New. Combine call to get_address_description and
2868 return addr_for_mem_ref.
2869 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2870 * tree-ssa-live.h: Adjust prototypes.
2871 * passes.c: Include tree-ssa-live.h.
2872 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2873 * graphite.c (graphite_transform_loops): Make static.
2874 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2875 make_pass_graphite, pass_data_graphite_transforms,
2876 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2877 * ipa-pure-const.c (warn_function_noreturn): Make static.
2878 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2879 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2880 Relocate from tree-cfg.c
2881 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2883 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2884 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2885 Move to ipa-pure-const.c.
2886 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2887 Relocate from tree-optimize.c.
2888 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2889 make_pass_fixup_cfg): Move to tree-cfg.c.
2890 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2891 Relocate some prototypes.
2892 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2893 * tree-dump.c (dump_function_to_file): Remove prototype.
2894 Add tree-flow.h to the include file.
2895 * tree-dump.h: Remove prototype.
2896 * tree-parloops.h: New File. Add prototypes.
2897 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2898 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2899 from tree-ssa-loop.c.
2900 * tree-predcom.c (run_tree_predictive_commoning,
2901 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2902 Relocate here from tree-ssa-loop.c.
2903 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2904 ssa_name_values.release ().
2905 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2906 (ssa_name_values): Relocate from tree-flow.h.
2907 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2908 * tree-ssa-loop.c (run_tree_predictive_commoning,
2909 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2910 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2911 make_pass_graphite, pass_data_graphite_transforms,
2912 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2913 tree_parallelize_loops, pass_data_parallelize_loops,
2914 make_pass_parallelize_loops): Move to other files.
2915 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2917 * tree.h: Remove prototypes from tree-address.c.
2919 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2921 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2922 (struct int_tree_map): Move to tree-hasher.h
2923 (SCALE, LABEL, PERCENT): Move to gimple.h
2924 * tree-flow-inline.h: Delete. Move functions to other files.
2925 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2926 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2927 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2928 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2929 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2930 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2931 * tree-hasher.h: Don't include tree-flow.h.
2932 (struct int_tree_map): Relocate from tree-flow.h.
2933 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2934 tree-flow-inline.h and make static.
2935 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2937 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2938 tree-flow-inline.h and make static.
2939 * tree.h (is_global_var, may_be_aliased): Relocate from
2941 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2942 * value-prof.c: No longer include tree-flow-inline.h.
2943 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2945 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2947 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2948 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2949 (enum move_pos): Move to tree-ssa-loop-im.h
2950 * cfgloop.h: Move some prototypes.
2951 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2952 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2953 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2954 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2955 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2956 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
2957 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
2958 (enum move_pos): Relocate here.
2959 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
2961 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
2962 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
2963 make_pass_lim): Relocate here from tree-ssa-loop.c.
2964 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
2966 (loop_edge_to_cancel, unloop_loops): Make static.
2967 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2968 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
2969 (tree_complete_unroll, gate_tree_complete_unroll,
2970 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
2971 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
2972 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
2973 * tree-ssa-loop-ivopts.c: Remove local prototypes.
2974 (stmt_invariant_in_loop_p): Remove unused function.
2975 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
2976 * tree-ssa-loop-manip.h: New file. Add prototypes.
2977 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
2978 (gcov_type_to_double_int): Move to cfgloop.h.
2979 (double_int_cmp, bound_index,
2980 estimate_numbers_of_iterations_loop): Make static.
2981 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
2982 (max_loop_iterations): Factor out get_max_loop_iterations.
2983 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
2985 * tree-ssa-loop-niter.h: New file. Add prototypes.
2986 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
2987 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
2988 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
2989 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
2990 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
2991 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
2992 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
2993 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
2994 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
2995 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
2996 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
2997 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
2998 pass_data_complete_unroll, make_pass_complete_unroll,
2999 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3000 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
3001 tree-ssa-loop-ivcanon.c.
3002 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
3003 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
3004 tree-ssa-loop-prefetch.c.
3005 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
3007 (get_lsm_tmp_name): Relocate and add suffix parameter.
3008 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
3009 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
3010 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
3011 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
3012 (get_estimated_loop_iterations): Factor out accessor from
3013 estimated_loop_iterations in tree-ssa-loop-niter.c.
3014 (get_max_loop_iterations): Factor out accessor from
3015 _max_loop_iterations in tree-ssa-niter.c.
3016 * loop-unroll.c (decide_unroll_constant_iterations,
3017 decide_unroll_runtime_iterations, decide_peel_simple,
3018 decide_unroll_stupid): Use new get_* accessors.
3020 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
3022 PR tree-optimization/20318
3023 * doc/extend.texi (returns_nonnull): New function attribute.
3024 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
3026 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
3027 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
3029 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
3032 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
3033 false if both components are bitfields.
3035 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3037 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
3038 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
3039 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
3040 * config/aarch64/aarch64.h
3041 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
3042 * config/aarch64/aarch64-simd-builtins.def
3043 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
3045 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3047 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
3049 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3051 * config/aarch64/arm_neon.h (vdiv_f64): Added.
3053 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3055 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
3056 (vneg_f64): New intrinsic.
3057 (vneg_s8): Asm replaced with C.
3058 (vneg_s16): Likewise.
3059 (vneg_s32): Likewise.
3060 (vneg_s64): New intrinsic.
3061 (vnegq_f32): Asm replaced with C.
3062 (vnegq_f64): Likewise.
3063 (vnegq_s8): Likewise.
3064 (vnegq_s16): Likewise.
3065 (vnegq_s32): Likewise.
3066 (vnegq_s64): Likewise.
3068 2013-10-09 Renlin Li <Renlin.Li@arm.com>
3070 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
3072 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3074 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
3075 packed stack special handling.
3076 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
3077 back to fixed stack slots for FPRs saved due to stdarg.
3079 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3081 * config/s390/s390.c (s390_frame_info): Restructure function.
3083 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3085 * config/s390/s390.c (struct s390_frame_layout): New field
3087 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
3088 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
3089 regs_ever_clobbered to char*.
3090 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
3091 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
3092 (s390_register_info): Enable FPR save slots. Move/Copy some
3093 functionality into ...
3094 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
3095 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
3097 (s390_frame_info): Do gpr slot allocation here now. stdarg does
3098 not imply a stack frame.
3099 (s390_init_frame_layout): Remove variable clobbered_regs.
3100 (s390_update_register_info): Remove function.
3101 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
3102 cannot be used for register renaming.
3103 (s390_hard_regno_scratch_ok): New function.
3104 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
3105 (s390_initial_elimination_offset): Change offset calculation of
3106 the return address pointer.
3107 (save_gprs): Deal with only r6 being saved from the call-saved regs.
3108 (restore_gprs): Set frame related flag.
3109 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
3110 (s390_emit_prologue): Call s390_register_info instead of
3111 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
3112 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
3113 (s390_optimize_prologue): Call s390_optimize_register_info.
3114 Try to remove also FPR slot save/restore INSNs. Remove frame
3115 related flags from restore INSNs.
3117 2013-10-08 DJ Delorie <dj@redhat.com>
3119 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
3122 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
3123 avoid conflict with the MI use of %c.
3124 * config/rl78/rl78-real.md: change %c to %C throughout.
3125 * config/rl78/rl78-virt.md: Likewise.
3127 2013-10-08 Jan Hubicka <jh@suse.cz>
3129 * config/i386/i386.c (ix86_option_override_internal): Switch
3130 to SSE math for -ffast-math when target ISA supports SSE2.
3132 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3134 * tree-flow.h: Remove some prototypes.
3135 * tree.h: Remove some protypes, add a couple.
3136 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3137 using_eh_for_cleanups_p): Add interface routines for front ends.
3138 * tree-eh.h: New file. Add protoptyes.
3139 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3140 (add_stmt_to_eh_lp_fn): Make static.
3141 (lower_try_finally): Use new using_eh_for_cleanups_p.
3142 * emit-rtl.c: Include tree-eh.h.
3143 * gimple.h: Include tree-eh.h.
3145 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3147 PR tree-optimization/58480
3148 * tree-vrp.c (infer_nonnull_range): New function.
3149 (infer_value_range): Call infer_nonnull_range.
3151 2013-10-08 Dehao Chen <dehao@google.com>
3153 PR tree-optimization/58619
3154 * tree-inline.c (copy_phis_for_bb): Combine location data
3157 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3160 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3161 RTX_FRAME_RELATED_P on INSN.
3163 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3165 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3166 (altivec_expand_vec_perm_const): Call it.
3168 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3170 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3171 sequences for LE VSX loads and stores at expand time.
3172 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3174 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3175 (rs6000_gen_le_vsx_permute): New.
3176 (rs6000_gen_le_vsx_load): New.
3177 (rs6000_gen_le_vsx_store): New.
3178 (rs6000_gen_le_vsx_move): New.
3179 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3180 (*vsx_le_perm_load_v4si): New.
3181 (*vsx_le_perm_load_v8hi): New.
3182 (*vsx_le_perm_load_v16qi): New.
3183 (*vsx_le_perm_store_v2di): New.
3184 (*vsx_le_perm_store_v4si): New.
3185 (*vsx_le_perm_store_v8hi): New.
3186 (*vsx_le_perm_store_v16qi): New.
3187 (*vsx_xxpermdi2_le_<mode>): New.
3188 (*vsx_xxpermdi4_le_<mode>): New.
3189 (*vsx_xxpermdi8_le_V8HI): New.
3190 (*vsx_xxpermdi16_le_V16QI): New.
3191 (*vsx_lxvd2x2_le_<mode>): New.
3192 (*vsx_lxvd2x4_le_<mode>): New.
3193 (*vsx_lxvd2x8_le_V8HI): New.
3194 (*vsx_lxvd2x16_le_V16QI): New.
3195 (*vsx_stxvd2x2_le_<mode>): New.
3196 (*vsx_stxvd2x4_le_<mode>): New.
3197 (*vsx_stxvd2x8_le_V8HI): New.
3198 (*vsx_stxvd2x16_le_V16QI): New.
3200 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3202 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3204 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3206 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3207 loop to work also for 31bit ABI.
3208 Save the stack pointer for frame_size > 0.
3210 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3212 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3213 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3214 constraint letters from expanders.
3215 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3216 retry count to general_operand.
3217 ("tabort"): Give operand 0 a mode.
3218 ("tabort_1"): Add mode and constraint letter for operand 0.
3219 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3221 2013-10-04 Jeff Law <law@redhat.com>
3223 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3225 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3227 (thread_across_edge): Here. Call it.
3229 2013-10-04 Cary Coutant <ccoutant@google.com>
3231 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3232 discarding a location list expression (or a piece of one).
3234 2013-10-03 Jan Hubicka <jh@suse.cz>
3236 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3237 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3238 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3240 2013-10-03 Jan Hubicka <jh@suse.cz>
3242 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3243 accumulate-outgoing-args when producing unwind info.
3245 2013-10-03 Wei Mi <wmi@google.com>
3247 * lra-constraints.c (insert_move_for_subreg): New function
3248 extracted from simplify_operand_subreg.
3249 (simplify_operand_subreg): Add reload for paradoxical subreg.
3251 2013-10-03 Rong Xu <xur@google.com>
3253 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3254 the candidate of builtin_expect such that we should fix the
3255 size/time estimation.
3256 (estimate_function_body_sizes): Do the acutally size/time fix-up
3259 2013-10-03 Rong Xu <xur@google.com>
3261 * predict.c (tree_predict_by_opcode): Get the probability
3262 for builtin_expect from param builtin_expect_probability.
3263 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3264 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3265 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3267 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3270 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3271 * calls.c (alloca_call_p): Use get_callee_fndecl.
3272 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3273 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3275 (vrp_visit_stmt): Remove duplicated code.
3277 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3279 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3280 ceildf2, btruncdf2, instead of vsx_* name.
3282 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3283 iterators to only do V2DF and V4SF here. Move the DF code to
3284 rs6000.md where it is combined with SF mode. Replace <VSv> with
3285 just 'v' since only vector operations are handled with these insns
3286 after moving the DF support to rs6000.md.
3287 (vsx_sub<mode>3): Likewise.
3288 (vsx_mul<mode>3): Likewise.
3289 (vsx_div<mode>3): Likewise.
3290 (vsx_fre<mode>2): Likewise.
3291 (vsx_neg<mode>2): Likewise.
3292 (vsx_abs<mode>2): Likewise.
3293 (vsx_nabs<mode>2): Likewise.
3294 (vsx_smax<mode>3): Likewise.
3295 (vsx_smin<mode>3): Likewise.
3296 (vsx_sqrt<mode>2): Likewise.
3297 (vsx_rsqrte<mode>2): Likewise.
3298 (vsx_fms<mode>4): Likewise.
3299 (vsx_nfma<mode>4): Likewise.
3300 (vsx_copysign<mode>3): Likewise.
3301 (vsx_btrunc<mode>2): Likewise.
3302 (vsx_floor<mode>2): Likewise.
3303 (vsx_ceil<mode>2): Likewise.
3304 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3305 (vsx_sminsf3): Likewise.
3306 (vsx_fmadf4): Likewise.
3307 (vsx_fmsdf4): Likewise.
3308 (vsx_nfmadf4): Likewise.
3309 (vsx_nfmsdf4): Likewise.
3310 (vsx_cmpdf_internal1): Likewise.
3312 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3313 simpler to select whether a target has SPE or traditional floating
3314 point support in iterators.
3315 (TARGET_DF_SPE): Likewise.
3316 (TARGET_SF_FPR): Likewise.
3317 (TARGET_DF_FPR): Likewise.
3318 (TARGET_SF_INSN): Macros to say whether floating point support
3319 exists for a given operation for expanders.
3320 (TARGET_DF_INSN): Likewise.
3322 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3323 combining of SF/DF mode operations, using both traditional and VSX
3331 (abs<mode>2): Combine SF/DF modes using traditional floating point
3332 instructions. Add support for using the upper DF registers with
3333 VSX support, and SF registers with power8-vector support. Update
3334 expanders for operations supported by both the SPE and traditional
3335 floating point units.
3336 (abs<mode>2_fpr): Likewise.
3337 (nabs<mode>2): Likewise.
3338 (nabs<mode>2_fpr): Likewise.
3339 (neg<mode>2): Likewise.
3340 (neg<mode>2_fpr): Likewise.
3341 (add<mode>3): Likewise.
3342 (add<mode>3_fpr): Likewise.
3343 (sub<mode>3): Likewise.
3344 (sub<mode>3_fpr): Likewise.
3345 (mul<mode>3): Likewise.
3346 (mul<mode>3_fpr): Likewise.
3347 (div<mode>3): Likewise.
3348 (div<mode>3_fpr): Likewise.
3349 (sqrt<mode>3): Likewise.
3350 (sqrt<mode>3_fpr): Likewise.
3351 (fre<Fs>): Likewise.
3352 (rsqrt<mode>2): Likewise.
3353 (cmp<mode>_fpr): Likewise.
3354 (smax<mode>3): Likewise.
3355 (smin<mode>3): Likewise.
3356 (smax<mode>3_vsx): Likewise.
3357 (smin<mode>3_vsx): Likewise.
3358 (negsf2): Delete SF operations that are merged with DF.
3365 (fmasf4_fpr): Likewise.
3366 (fmssf4_fpr): Likewise.
3367 (nfmasf4_fpr): Likewise.
3368 (nfmssf4_fpr): Likewise.
3369 (sqrtsf2): Likewise.
3370 (rsqrtsf_internal1): Likewise.
3371 (smaxsf3): Likewise.
3372 (sminsf3): Likewise.
3373 (cmpsf_internal1): Likewise.
3374 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3375 (negdf2): Delete DF operations that are merged with SF.
3377 (nabsdf2): Likewise.
3383 (rsqrtdf_internal1): Likewise.
3384 (fmadf4_fpr): Likewise.
3385 (fmsdf4_fpr): Likewise.
3386 (nfmadf4_fpr): Likewise.
3387 (nfmsdf4_fpr): Likewise.
3388 (sqrtdf2): Likewise.
3389 (smaxdf3): Likewise.
3390 (smindf3): Likewise.
3391 (cmpdf_internal1): Likewise.
3392 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3393 (btrunc<mode>2): Delete separate expander, and combine with the
3394 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3395 (btrunc<mode>2_fpr): Likewise.
3396 (ceil<mode>2): Likewise.
3397 (ceil<mode>2_fpr): Likewise.
3398 (floor<mode>2): Likewise.
3399 (floor<mode>2_fpr): Likewise.
3400 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3401 Add support for using the upper registers with VSX and
3402 power8-vector. Move insns to be closer to the define_expands. On
3403 VSX systems, prefer the traditional form of FMA over the VSX
3404 version, since the traditional form allows the target not to
3405 overlap with the inputs.
3406 (fms<mode>4_fpr): Likewise.
3407 (nfma<mode>4_fpr): Likewise.
3408 (nfms<mode>4_fpr): Likewise.
3410 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3411 Richard Earnshaw <richard.earnshaw@arm.com>
3413 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3414 (struct mult_cost_table): Likewise.
3415 (struct mem_cost_table): Likewise.
3416 (struct fp_cost_table): Likewise.
3417 (struct vector_cost_table): Likewise.
3418 (cpu_cost_table): Likewise.
3419 * config/arm/arm.opt (mold-rts-costs): New option.
3420 (mnew-generic-costs): Likewise.
3421 * config/arm/arm.c (generic_extra_costs): New table.
3422 (cortexa15_extra_costs): Likewise.
3423 (arm_slowmul_tune): Use NULL as new costs.
3424 (arm_fastmul_tune): Likewise.
3425 (arm_strongarm_tune): Likewise.
3426 (arm_xscale_tune): Likewise.
3427 (arm_9e_tune): Likewise.
3428 (arm_v6t2_tune): Likewise.
3429 (arm_cortex_a5_tune): Likewise.
3430 (arm_cortex_a9_tune): Likewise.
3431 (arm_v6m_tune): Likewise.
3432 (arm_fa726te_tune): Likewise.
3433 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3434 (arm_cortex_tune): Use generict_extra_costs.
3435 (shifter_op_p): New function.
3436 (arm_unspec_cost): Likewise.
3437 (LIBCALL_COST): Define.
3438 (arm_new_rtx_costs): New function.
3439 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3440 table is available. Use old costs otherwise unless mnew-generic-costs
3442 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3443 (cpu_cost_table): Declare.
3445 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3448 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3449 (*subs_mul_imm_<mode>)
3450 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3451 (*sub_<shift>_<mode>)
3452 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3453 Remove k constraint.
3455 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3457 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3459 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3461 2013-10-02 Teresa Johnson <tejohnson@google.com>
3463 * predict.c (probably_never_executed): New function.
3464 (probably_never_executed_bb_p): Invoke probably_never_executed.
3465 (probably_never_executed_edge_p): Ditto.
3466 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3467 Treat profile insanities conservatively.
3469 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3471 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3473 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3475 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3476 use it. Use smaller increase for scratch. Don't increase reject
3477 for early clobber scratch.
3478 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3479 setting eliminated regs except setting fp from hfp.
3480 (lra_eliminate): Check lra_insn_recog_data on NULL.
3482 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3485 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3486 setting -mvsx-timode by default until the underlying problem is fixed.
3487 (RS6000_CPU, power7 defaults): Likewise.
3489 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3491 * config/x-linux (host-linux.o): Remove header dependencies.
3492 Use $(COMPILE) and $(POSTCOMPILE).
3493 * config/t-linux-android (linux-android.o): Ditto.
3495 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3497 * Makefile.in (expmed.o-warn): Remove.
3499 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3501 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3502 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3504 2013-10-02 Teresa Johnson <tejohnson@google.com>
3506 * dojump.c (do_jump_1): Divide probability between
3507 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3509 2013-10-02 Tom Tromey <tromey@redhat.com>
3511 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3513 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3515 * tree-flow.h: Remove some prototypes.
3516 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3517 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3518 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3519 mark_virtual_phi_result_for_renaming): Relocate here.
3520 * tree-into-ssa.h: Add prototypes.
3521 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3522 single_pred_before_succ_order.
3523 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3524 (nonfreeing_call_p) Move to gimple.c.
3525 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3527 * basic-block.h (single_pred_before_succ_order): Add prototype.
3528 * gimple.c (nonfreeing_call_p): Relocate here.
3529 * gimple.h: Add prototype.
3530 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3531 * tree-ssa-dom.h: New file. Relocate prototypes here.
3532 * tree-ssa.h: Include tree-ssa-dom.h.
3534 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3536 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3537 Use $(COMPILE) and $(POSTCOMPILE).
3539 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3541 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3543 * tree-flow.h: Remove some prototypes.
3544 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3545 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3546 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3547 propagate_tree_value*): Move from here to...
3548 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3549 propagate_tree_value*): Relocate here.
3550 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3551 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3552 * gimple-fold.c: Remove gimple-fold.h from include list.
3553 * tree-vrp.c: Remove gimple-fold.h from include list.
3554 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3555 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3556 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3557 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3558 * sese.c: Add tree-ssa-propagate.h to include list.
3560 2013-10-02 Richard Biener <rguenther@suse.de>
3562 * tree-loop-distribution.c: Include tree-vectorizer.h for
3564 (enum partition_kind): Remove PKIND_REDUCTION.
3565 (struct partition_s): Remove has_writes member, add reduction_p member.
3566 (partition_alloc): Adjust.
3567 (partition_builtin_p): Likewise.
3568 (partition_has_writes): Remove.
3569 (partition_reduction_p): New function.
3570 (partition_merge_into): Likewise.
3571 (generate_code_for_partition): Commonize builtin partition
3573 (rdg_cannot_recompute_vertex_p): Remove.
3574 (already_processed_vertex_p): Likewise.
3575 (rdg_flag_vertex): Do not set has_writes.
3576 (classify_partition): Adjust.
3577 (rdg_build_partitions): Do not set has_writes, treat all
3578 partitions as useful.
3579 (distribute_loop): Record number of library calls generated. Adjust.
3580 (tree_loop_distribution): Report number of loops and library
3581 calls generated as opt-info.
3583 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3585 * tree-flow.h: Include new .h files. Move prototypes.
3586 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3587 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3588 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3589 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3590 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3591 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3592 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3593 * tree.h (get_ref_base_and_extent): Move prototype out.
3594 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3596 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3597 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3598 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3599 * tree-scalar-evolution.c: Include tree.h.
3600 * sese.c: Include tree.h.
3601 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3602 * dwarf2out.c: Include tree-dfa.h.
3603 * tree-chrec.c: Include tree.h.
3604 * tree-data-ref.c: Include tree.h.
3606 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3608 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3611 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3613 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3614 Use $(COMPILE) and $(POSTCOMPILE).
3615 (sol2-cxx.o): Likewise.
3616 (sol2-stubs.o): Likewise.
3618 * config/x-solaris (host-solaris.o): Likewise.
3620 * config/sparc/t-sparc (sparc.o): Remove.
3621 (sparc-c.o): Remove header dependencies.
3622 Use $(COMPILE) and $(POSTCOMPILE).
3623 * config/sparc/x-sparc: Likewise.
3625 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3627 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3628 * config/arc/arc700.md: Likewise.
3629 * config/arc/arc-modes.def: Likewise.
3630 * config/arc/arc-simd.h: Likewise.
3631 * config/arc/t-arc-uClibc: Likewise.
3632 * config/arc/t-arc-newlib: Likewise.
3634 2013-10-02 Renlin Li <renlin.li@arm.com>
3636 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3638 (aarch64_expand_epilogue): Likewise.
3640 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3641 Yufeng Zhang <yufeng.zhang@arm.com>
3643 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3645 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3646 'base_in' represent a conversion and legal_cast_p_1 holds; set
3647 'base_in' with the returned value from get_unwidened.
3649 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3651 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3652 plus_constant is not used.
3654 2013-10-01 Wei Mi <wmi@google.com>
3656 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3657 * config/i386/i386.md: Add define_peephole2 to
3658 break partial reg stall for cvtss2sd/cvtsd2ss.
3660 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3662 * config/arc/arc.c (pass_arc_ifcvt::clone):
3663 Update for ctxt_ -> m_ctxt change.
3665 2013-10-01 Jeff Law <law@redhat.com>
3667 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3668 outgoing_edge and intermediate_edge fields. Instead store the path.
3669 (redirection_data::hash): Hash on the last edge's destination index.
3670 (redirection_data::equal): Check the entire thread path.
3671 (lookup_redirectio_data): Corresponding changes.
3672 (create_edge_and_update_destination_phis): Likewise.
3673 (thread_single_edge): Likewise.
3675 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3676 Diego Novillo <dnovillo@google.com>
3678 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3679 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3680 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3681 versions of these patterns.
3683 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3684 (__builtin_arc_aligned): Likewise.
3686 * config/arc/arc.md: Expand adc_0 comment stating the intended
3687 purpose and why it isn't ready.
3688 Replace commented out call_value_via_label_mixed with a
3689 plain comment about bl_s.
3691 * config/arc/arc.c (stdio.h): Don't include directly.
3692 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3693 Remove if (1) condition.
3694 (arc_encode_section_info): Fix comment.
3696 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3698 * config/arc/arc.c (arc_conditional_register_usage):
3699 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3700 Also set reg_alloc_order for DMA config regs.
3702 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3703 Jeremy Bennett <jeremy.bennett@embecosm.com>
3705 * doc/install.texi (--with-cpu): Mention ARC.
3706 (arc-*-elf32): New paragraph.
3707 (arc-linux-uclibc): Likewise.
3708 * doc/md.texi (Machine Constraints): Add ARC part.
3709 * doc/invoke.texi: (menu): Add ARC Options.
3710 (Machine Dependent Options) <ARC Options>: Add synopsis.
3711 (node ARC Options): Add.
3712 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3713 (ARC Built-in Functions): New section defining
3714 generic ARC built-in functions.
3715 (ARC SIMD Built-in Functions): New section defining SIMD specific
3717 (Declaring Attributes of Functions): Extended
3718 description of short_call and long_call attributes for ARC and
3719 added index entries.
3721 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3722 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3723 Joern Rennecke <joern.rennecke@embecosm.com>
3724 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3725 Brendan Kehoe <brendan@zen.org>
3726 Michael Eager <eager@eagercon.com>
3727 Simon Cook <simon.cook@embecosm.com>
3728 Jeremy Bennett <jeremy.bennett@embecosm.com>
3730 * config/arc, common/config/arc: New directories.
3732 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3733 Brendan Kehoe <brendan@zen.org>
3734 Simon Cook <simon.cook@embecosm.com>
3736 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3738 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3740 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3741 * tree-ssa-coalesce.h: New. Move prototype to here.
3742 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3743 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3744 (gimple_can_coalesce_p): Move to...
3745 * gimple.c (gimple_can_coalesce_p): Here.
3747 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3749 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3750 (dump_decl_set): Move to gimple.c.
3751 * gimple.h: Don't include tree-ssa-operands.h.
3752 (dump_decl_set): Add prototype.
3753 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3754 Move to gimple-ssa.h.
3755 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3756 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3757 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3758 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3759 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3761 (dump_decl_set): Relocate here.
3762 * gimple-ssa.h: New file.
3763 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3764 Relocate from gimple.h.
3765 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3766 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3767 swap_tree_operands and remove non-ssa path.
3768 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3769 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3771 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3772 vect_is_simple_reduction_1): Use swap_ssa_operands.
3773 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3774 (enum need_phi_state): Move to tree-into-ssa.c.
3775 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3776 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3777 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3778 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3779 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3780 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3781 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3782 num_imm_uses): Move to ssa-iterators.h.
3783 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3784 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3786 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3787 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3788 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3789 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3790 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3791 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3792 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3793 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3794 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3795 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3796 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3797 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3798 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3799 (set_phi_nodes): Move to tree-phinodes.h.
3800 * tree-ssa-operands.h (enum ssa_op_iter_type,
3801 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3802 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3803 (dump_decl_set): Remove prototype.
3804 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3805 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3806 (set_phi_nodes): Relocate from tree-flow-inline.h.
3807 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3809 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3810 include list. Temporarily add gimple.h to include list.
3811 * ssa-iterators.h: New file.
3812 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3813 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3814 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3815 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3816 Relocate from tree-ssa-operands.h.
3817 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3818 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3819 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3820 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3821 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3822 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3823 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3824 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3825 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3826 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3827 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3828 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3829 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3830 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3831 Relocate from tree-flow-inline.h.
3832 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3834 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3837 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3838 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3839 Insert '\t' to output template.
3840 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3841 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3842 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3843 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3845 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3847 * doc/install.texi (Host/target specific installation notes for GCC):
3848 Put @anchor before @heading.
3849 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3850 Use @email for email addresses.
3852 2013-10-01 Jeff Law <law@redhat.com>
3854 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3855 a vec. Only delete the path if we create one without successfully
3856 registering a jump thread.
3857 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3859 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3860 (paths): New vector of jump threading paths.
3861 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3862 (THREAD_PATH): New accessor macro for the entire thread path.
3863 (lookup_redirection_data): Get intermediate and final outgoing edge
3864 from the thread path.
3865 (create_edge_and_update_destination_phis): Copy the threading path.
3866 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3867 jump threading path.
3868 (ssa_redirect_edges): Get edges and block types from the jump threading
3869 path. Free the path vector.
3870 (thread_block): Get edges from the jump threading path. Look at the
3871 entire path to see if we thread to a loop exit. If we cancel a jump
3872 thread request, then free the path vector.
3873 (thread_single_edge): Get edges and block types from the jump threading
3874 path. Free the path vector.
3875 (thread_through_loop_header): Get edges and block types from the jump
3876 threading path. Free the path vector.
3877 (mark_threaded_blocks): Iterate over the vector of paths and store
3878 the path on the appropriate edge. Get edges and block types from the
3879 jump threading path.
3880 (mark_threaded_blocks): Get edges and block types from the jump
3881 threading path. Free the path vector.
3882 (thread_through_all_blocks): Use the vector of paths rather than
3883 a vector of 3-edge sets.
3884 (register_jump_thread): Accept pointer to a path vector rather
3885 than the path vector itself. Store the path vector for later use.
3888 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3889 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3892 * config/s390/s390.c (s390_split_branches): Modify check for table
3894 (s390_chunkify_start): Rearrange table jump insn check in order to
3895 deal with compare and branch insns correctly.
3897 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3901 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3902 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3903 define_insn_and_split.
3904 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3905 (shiftsi3_compare): New pattern.
3907 * config/arm/unspecs.md (UNSPEC_RRX): New.
3909 2013-10-01 Alan Modra <amodra@gmail.com>
3911 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3912 casing inout operands.
3914 2013-10-01 Richard Biener <rguenther@suse.de>
3916 PR tree-optimization/58553
3917 * tree-loop-distribution.c (struct partition_s): Add niter member.
3918 (classify_partition): Populate niter member for the partition
3919 and properly identify whether the relevant store happens before
3920 or after the loop exit.
3921 (generate_memset_builtin): Use niter member from the partition.
3922 (generate_memcpy_builtin): Likewise.
3924 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3926 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3927 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3929 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3931 * basic-block.h (edge_list): Prefix member names with "m_".
3932 * context.h (context): Likewise.
3933 * domwalk.h (dom_walker): Likewise.
3934 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3935 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3936 * hash-table.h (hash_table): Likewise.
3937 * machmode.h (bit_field_mode_iterator): Likewise.
3938 * pass_manager.h (pass_list): Likewise.
3939 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3940 * tree-pass.h (pass_data): Likewise.
3941 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3942 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3943 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3944 * asan.c (pass_data_asan): Update accordingly.
3945 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3946 (control_dependences::control_dependences): Likewise.
3947 (control_dependences::~control_dependences): Likewise.
3948 (control_dependences::~control_dependences): Likewise.
3949 (control_dependences::get_edges_dependent_on): Likewise.
3950 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3951 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3952 * context.c (gcc::context::context): Likewise.
3953 * cprop.c (pass_rtl_cprop::clone): Likewise.
3954 * domwalk.c (dom_walker::walk): Likewise.
3955 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3956 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
3957 * mode-switching.c (pass_mode_switching::clone): Likewise.
3958 * passes.c (opt_pass::opt_pass): Likewise.
3959 (pass_manager::pass_manager): Likewise.
3960 * predict.c (pass_strip_predict_hints::clone): Likewise.
3961 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
3962 (pass_split_all_insns::clone): Likewise.
3963 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
3965 (bit_field_mode_iterator::next_mode): Likewise.
3966 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
3967 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
3968 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
3969 * tree-complex.c (pass_lower_complex::clone): Likewise.
3970 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
3971 * tree-object-size.c (pass_object_sizes::clone): Likewise.
3972 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
3973 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
3974 (pass_fold_builtins::clone): Likewise.
3975 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
3976 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
3977 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
3978 (pass_cd_dce::clone): Likewise.
3979 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
3980 (pass_phi_only_cprop::clone): Likewise.
3981 * tree-ssa-dse.c (pass_dse::clone): Likewise.
3982 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
3983 * tree-ssa-loop.c (pass_lim::clone): Likewise.
3984 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
3985 * tree-ssa-pre.c (pass_fre::clone): Likewise.
3986 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
3987 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
3988 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
3989 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
3990 * tree-vrp.c (pass_vrp::clone): Likewise.
3991 * tsan.c (pass_tsan::clone): Likewise.
3993 2013-09-30 Jakub Jelinek <jakub@redhat.com>
3996 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
3997 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
4000 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
4001 optimization, punt if sign_bit_p looked through any zero extension.
4003 2013-09-30 Teresa Johnson <tejohnson@google.com>
4005 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
4006 Update redirected out edge count in joiner case.
4007 (ssa_redirect_edges): Common the joiner and non-joiner cases
4008 so that joiner case gets profile updates.
4010 2013-09-30 Richard Biener <rguenther@suse.de>
4012 PR tree-optimization/58554
4013 * tree-loop-distribution.c (classify_partition): Require
4014 unconditionally executed stores for memcpy and memset recognition.
4015 (tree_loop_distribution): Calculate dominance info.
4017 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4019 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
4020 (NO_PROFILE_COUNTERS): Likewise.
4021 (PROFILE_HOOK): Likewise.
4022 (FUNCTION_PROFILER): Likewise.
4023 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
4025 2013-09-30 Iain Sandoe <iain@codesourcery.com>
4027 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
4028 calls and defines in TARGET_MACHO conditional.
4029 (load_macho_picbase_di): Likewise.
4030 (reload_macho_picbase): Likewise.
4031 (reload_macho_picbase_si): Likewise.
4032 (reload_macho_picbase_di): Likewise.
4033 (nonlocal_goto_receiver): Likewise.
4035 2013-09-30 Nick Clifton <nickc@redhat.com>
4037 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
4038 that use the MSP430X ISA.
4039 (msp430_option_override): Scan -mmcu command line option for any
4040 MCU name that supports the MSP430X ISA.
4041 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
4042 -mmcu options which enable the MSP430X ISA.
4044 2013-09-30 Richard Biener <rguenther@suse.de>
4047 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
4048 before looking for setjmp-like calls.
4050 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4053 * config/darwin-protos.h (machopic_get_function_picbase): New.
4054 * config/darwin.c (machopic_get_function_picbase): New.
4055 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
4056 label for a new func. (load_macho_picbase_di): Likewise.
4057 (reload_macho_picbase): New expand.
4058 (reload_macho_picbase_si): New insn.
4059 (reload_macho_picbase_di): New insn.
4060 (nonlocal_goto_receiver): New define and split.
4061 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
4062 (unspecv enum): Add UNSPECV_NLGR.
4064 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4066 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
4067 that altivec registers are correctly sized on Darwin.
4069 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4071 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
4072 darwin-driver.o): Use COMPILE and POSTCOMPILE.
4073 * config/x-darwin (host-darwin.o): Likewise.
4074 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
4075 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
4076 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
4078 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
4080 * doc/invoke.texi: Fix usage of @tie{} command.
4082 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
4084 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
4086 2013-09-28 Jan Hubicka <jh@suse.cz>
4088 * config/i386/x86-tune.def: Add documentation for each of the options;
4091 2013-09-28 Jan Hubicka <jh@suse.cz>
4093 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
4095 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
4096 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
4097 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
4098 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
4099 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
4101 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
4103 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
4104 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
4105 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
4106 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
4107 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
4108 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
4109 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
4110 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
4111 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
4112 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
4113 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
4114 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
4115 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
4116 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
4117 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
4118 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
4119 gimple.h, godump.c, graphite-clast-to-gimple.c,
4120 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
4121 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
4122 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
4123 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4124 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
4125 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
4126 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
4127 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
4128 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
4129 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
4130 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
4131 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
4132 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4133 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4134 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4135 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4136 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4137 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4138 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4139 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4140 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4141 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4142 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4143 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4144 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4145 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4146 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4147 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4148 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4149 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4150 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4151 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4152 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4153 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4154 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4155 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4156 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4157 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4158 whitespace before "(".
4160 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4162 * expr.h (extract_bit_field): Remove packedp parameter.
4163 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4164 from forward declaration.
4165 (store_split_bit_field): Remove packedp arg from calls to
4166 extract_fixed_bit_field.
4167 (extract_bit_field_1): Remove packedp parameter and packedp
4168 argument from recursive calls and calls to extract_fixed_bit_field.
4169 (extract_bit_field): Remove packedp parameter and corresponding
4170 arg to extract_bit_field_1.
4171 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4173 (extract_split_bit_field): Remove packedp arg from call to
4174 extract_fixed_bit_field.
4175 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4176 (copy_blkmode_from_reg): Likewise.
4177 (copy_blkmode_to_reg): Likewise.
4178 (read_complex_part): Likewise.
4179 (store_field): Likewise.
4180 (expand_expr_real_1): Likewise.
4181 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4182 to extract_bit_field.
4183 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4184 call to extract_bit_field.
4185 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4186 call to extract_bit_field.
4187 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4188 and special packedp behavior from -fstrict-volatile-bitfields
4191 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4193 * lra-eliminations.c (init_elim_table): Guard value_p.
4195 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4197 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4198 DFmode, DImode, and SFmode in the upper VSX registers based on the
4199 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4200 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4201 of the VSX register handling.
4203 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4205 (zero_extendsidi2_lfiwzx): Add support for loading into the
4206 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4207 only do VSX memory options on Altivec registers.
4208 (extendsidi2_lfiwax): Likewise.
4209 (extendsfdf2_fpr): Likewise.
4210 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4211 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4212 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4213 (movdi_internal64): Likewise.
4215 2013-09-27 Xinliang David Li <davidxl@google.com>
4217 * opts.c (finish_options): Adjust parameters
4218 according to vect cost model.
4219 (common_handle_option): Set dynamic vect cost
4221 targhooks.c (default_add_stmt_cost): Compute stmt cost
4223 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4224 Use helper function.
4225 * tree-vectorizer.h (unlimited_cost_model): New function.
4226 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4227 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4229 (vect_enhance_data_refs_alignment): Ditto.
4230 * flag-types.h: New enum.
4231 * common/config/i386/i386-common.c (ix86_option_init_struct):
4232 No need to initialize vect_cost_model flag.
4233 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4236 2013-09-27 Diego Novillo <dnovillo@google.com>
4238 * gimple.h (enum ssa_mode): Remove.
4240 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4242 * cfgloop.h (number_of_loops): Fix typo in check for null.
4244 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4247 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4248 are children of outermost saved_cfun's loop, and set it up to
4249 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4250 if loop != loop0 and SESE region contains bbs that belong to loop0.
4252 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4254 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4255 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4256 (set_address_segment, set_address_base, set_address_index)
4257 (set_address_disp): Accept the argument unconditionally.
4258 (baseness): Remove must_be_base_p and must_be_index_p checks.
4259 (decompose_normal_address): Classify as much as possible in the
4262 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4264 * cse.c (count_reg_usage): Handle INT_LIST.
4265 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4266 * reginfo.c (reg_scan_mark_refs): Likewise.
4267 * reload1.c (eliminate_regs_1): Likewise.
4269 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4272 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4275 2013-09-27 Richard Biener <rguenther@suse.de>
4277 PR tree-optimization/58459
4278 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4279 restriction not propagating into loops.
4281 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4283 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4284 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4285 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4287 2013-09-26 Richard Biener <rguenther@suse.de>
4289 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4291 2013-09-26 Richard Biener <rguenther@suse.de>
4293 * alias.h (component_uses_parent_alias_set): Rename to ...
4294 (component_uses_parent_alias_set_from): ... this.
4295 * alias.c (component_uses_parent_alias_set): Rename to ...
4296 (component_uses_parent_alias_set_from): ... this and return
4298 (reference_alias_ptr_type_1): Use the result from
4299 component_uses_parent_alias_set_from instead of stripping
4300 components one at a time.
4301 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4303 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4305 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4306 Move prototypes to...
4307 * tree-ssa-ter.h: New File. Move prototypes here.
4308 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4309 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4310 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4311 from is_replaceable_p.
4312 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4313 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4314 (process_replaceable): Use ter_is_replaceable_p.
4315 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4316 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4317 newly refactored ssa_is_replaceable_p.
4318 * cfgexpand.c: Include tree-outof-ssa.h.
4319 * ssaexpand.h: Delete.
4321 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4323 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4324 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4326 (create_gimple_tmp): Delete.
4327 (get_expr_type, build_assign, build_type_cast): Move to...
4328 * gimple-builder.c: New File.
4329 (get_expr_type): Relocate from gimple.c.
4330 (build_assign, build_type_cast): Change to only create ssanames.
4331 * gimple.h: Move prototypes to...
4332 * gimple-builder.h: New File. Here.
4333 * tree-ssa.h: And here.
4334 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4335 count_uses_and_derefs): Relocate from gimple.c.
4336 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4337 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4338 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4339 gimple_replace_ssa_lhs.
4340 * asan.c: Include gimple-builder.h.
4341 * Makefile.in: Add gimple-builder.o.
4343 2013-09-26 Richard Biener <rguenther@suse.de>
4345 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4346 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4347 (loe_visit_block): Use gcc_checking_assert.
4348 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4349 gimple_assign_ssa_name_copy_p.
4350 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4353 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4355 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4356 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4358 2013-09-26 Richard Biener <rguenther@suse.de>
4360 PR tree-optimization/58539
4361 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4362 the fact that debug statements are not taking part in loop-closed
4365 2013-09-26 Nick Clifton <nickc@redhat.com>
4367 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4368 time warning message.
4369 (msp430_print_operand_raw): Delete unused letter parameter.
4370 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4371 (msp430_print_operand_address): New function.
4372 (msp430_print_operand): Move address printing code from here to
4374 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4376 (zero_extendpsisi2): Likewise.
4377 (extendpsisi2): New pattern.
4378 (andneghi3): New pattern.
4380 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4382 * config/aarch64/aarch64.opt (mlra): New option.
4383 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4384 (TARGET_LRA_P): Define.
4386 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4388 * expr.c (expand_assignment): Remove obsolete comment.
4390 2013-09-25 Jeff Law <law@redhat.com>
4392 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4393 tree-ssa-threadupdate.h.
4394 (register_jump_thread): Similarly.
4395 * tree-ssa-threadupdate.h: New header file.
4396 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4397 * tree-vrp.c: Likewise.
4398 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4399 (thread_around_empty_blocks): Change type of path vector argument to
4400 an edge,type pair from just an edge. Initialize both elements when
4401 appending to a jump threading path. Tweak references to elements
4403 (thread_across_edge): Similarly. Release memory for the elements
4405 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4406 (dump_jump_thread_path): New function broken out from
4407 register_jump_thread.
4408 (register_jump_thread): Use dump_jump_thread_path. Change type of
4409 path vector entries. Search the path for NULL edges and dump
4410 the path if one is found. Tweak the conversion of path to 3-edge
4411 form to use the block copy type information embedded in the path.
4413 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4415 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4417 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4418 Vladimir Makarov <vmakarov@redhat.com>
4420 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4421 from the least significant bit.
4422 (strip_address_mutations): Add bitfield operations handling.
4423 (must_be_index_p): Add shifting and rotate operations handling.
4424 (set_address_base): Use must_be_base_p predicate.
4425 (set_address_index): Use must_be_index_p predicate.
4427 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4428 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4429 Sergey Lega <sergey.s.lega@intel.com>
4430 Anna Tikhonova <anna.tikhonova@intel.com>
4431 Ilya Tocar <ilya.tocar@intel.com>
4432 Andrey Turetskiy <andrey.turetskiy@intel.com>
4433 Ilya Verbin <ilya.verbin@intel.com>
4434 Kirill Yukhin <kirill.yukhin@intel.com>
4435 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4437 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4439 (ix86_expand_vector_move_misalign): Support new unaligned load and
4440 stores and use new names.
4441 (CODE_FOR_sse2_storedqu): Rename to ...
4442 (CODE_FOR_sse2_storedquv16qi): ... this.
4443 (CODE_FOR_sse2_loaddqu): Rename to ...
4444 (CODE_FOR_sse2_loaddquv16qi): ... this.
4445 (CODE_FOR_avx_loaddqu256): Rename to ...
4446 (CODE_FOR_avx_loaddquv32qi): ... this.
4447 (CODE_FOR_avx_storedqu256): Rename to ...
4448 (CODE_FOR_avx_storedquv32qi): ... this.
4449 * config/i386/i386.md (fpint_logic): New.
4450 * config/i386/sse.md (VMOVE): Extend for AVX512.
4454 (VI_UNALIGNED_LOADSTORE): Ditto.
4455 (sse2_avx_avx512f): Ditto.
4456 (sse2_avx2): Extend for AVX512.
4457 (sse4_1_avx2): Ditto.
4458 (avx2_avx512f): New.
4459 (sse): Extend for AVX512.
4462 (avxsizesuffix): Ditto.
4463 (sseintvecmode): Ditto.
4465 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4466 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4467 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4468 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4469 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4470 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4471 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4472 (STORENT_MODE): Extend for AVX512.
4473 (*absneg<mode>2): Replace constraint "x" with "v".
4474 (*mul<mode>3): Ditto.
4475 (*ieee_smin<mode>3): Ditto.
4476 (*ieee_smax<mode>3): Ditto.
4477 (avx_cmp<mode>3): Replace VF with VF_128_256.
4478 (*<sse>_maskcmp<mode>3_comm): Ditto.
4479 (<sse>_maskcmp<mode>3): Ditto.
4480 (<sse>_andnot<mode>3): Extend for AVX512.
4481 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4482 (<code><mode>3, fpint_logic): New.
4483 (*<code><mode>3): Extend for AVX512.
4484 (avx512flogicsuff): New.
4485 (avx512f_<logic><mode>): Ditto.
4486 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4488 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4489 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4490 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4491 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4492 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4493 (xop_vpermil2<mode>3): Ditto.
4494 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4495 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4496 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4497 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4499 2013-09-25 Tom Tromey <tromey@redhat.com>
4501 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4502 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4503 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4504 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4505 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4506 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4507 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4508 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4509 (GRAPHITE_HTAB_H): Remove.
4511 2013-09-25 Tom Tromey <tromey@redhat.com>
4513 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4515 2013-09-25 Tom Tromey <tromey@redhat.com>
4517 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4519 2013-09-25 Tom Tromey <tromey@redhat.com>
4521 * config/i386/t-i386 (i386.o): Remove.
4522 (i386-c.o): Use COMPILE and POSTCOMPILE.
4524 2013-09-25 Tom Tromey <tromey@redhat.com>
4526 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4528 2013-09-25 Tom Tromey <tromey@redhat.com>
4530 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4531 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4532 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4533 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4534 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4535 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4536 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4537 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4538 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4539 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4540 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4541 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4542 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4543 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4544 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4545 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4546 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4547 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4548 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4549 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4550 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4551 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4552 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4553 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4554 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4555 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4556 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4557 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4558 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4559 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4560 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4561 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4562 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4563 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4564 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4565 (graphite-clast-to-gimple.o, graphite-dependences.o)
4566 (graphite-interchange.o, graphite-poly.o)
4567 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4568 (graphite-optimize-isl.o, tree-vect-loop.o)
4569 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4570 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4571 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4572 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4573 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4574 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4575 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4576 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4577 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4578 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4579 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4580 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4581 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4582 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4583 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4584 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4585 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4586 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4587 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4588 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4589 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4590 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4591 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4592 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4593 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4594 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4595 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4596 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4597 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4598 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4599 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4600 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4601 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4602 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4603 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4604 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4605 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4606 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4607 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4608 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4609 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4610 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4611 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4612 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4613 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4614 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4615 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4616 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4617 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4618 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4619 (gcov-dump.o): Remove.
4620 (default-c.o): Use COMPILE and POSTCOMPILE.
4621 (CFLAGS-gcc.o): New variable.
4622 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4624 2013-09-25 Tom Tromey <tromey@redhat.com>
4626 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4627 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4628 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4629 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4630 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4631 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4632 (c-family/array-notation-common.o, c-family/stub-objc.o)
4633 (c-family/c-ubsan.o): Remove.
4635 2013-09-25 Tom Tromey <tromey@redhat.com>
4637 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4639 2013-09-25 Tom Tromey <tromey@redhat.com>
4641 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4642 to add -DENABLE_SHARED_LIBGCC.
4643 (gcc.o): Don't use subshell.
4645 2013-09-25 Tom Tromey <tromey@redhat.com>
4647 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4648 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4649 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4650 * configure, config.in: Rebuild.
4652 2013-09-25 Tom Tromey <tromey@redhat.com>
4654 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4655 (COMPILE, POSTCOMPILE): New variables.
4656 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4657 (DEPFILES): New variable.
4658 Include ".Po" files.
4659 * configure.ac: Add checks for dependency checking.
4660 * configure, aclocal.m4: Regenerate.
4662 2013-09-25 Tom Tromey <tromey@redhat.com>
4664 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4665 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4667 2013-09-25 Tom Tromey <tromey@redhat.com>
4669 * Makefile.in (generated_files): Add options.h,
4670 target-hooks-def.h, insn-opinit.h,
4671 common/common-target-hooks-def.h, pass-instances.def,
4672 c-family/c-target-hooks-def.h.
4674 2013-09-25 Jeff Law <law@redhat.com>
4676 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4677 than foo[foo.length () - 1] to access last member in a vec.
4678 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4680 2013-09-25 Richard Biener <rguenther@suse.de>
4683 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4685 2013-09-25 Jan Hubicka <jh@suse.cz>
4687 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4690 2013-09-25 Marek Polacek <polacek@redhat.com>
4693 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4694 when determining the type name.
4696 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4698 * config/sh/sh.md: Fix formatting.
4700 2013-09-24 Xinliang David Li <davidxl@google.com>
4702 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4703 max peel iterations parameter.
4704 * param.def: New parameter.
4705 * doc/invoke.texi: Document New parameter.
4707 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4709 * gimple-pretty-print.c: Various whitespace tweaks.
4710 * tree-core.h: Likewise.
4711 * tree-pretty-print.c: Likewise.
4712 * tree-ssa-alias.c: Likewise.
4713 * tree-ssa-copy.c: Likewise.
4714 * tree-ssanames.c: Likewise.
4715 * tree-ssanames.h: Likewise.
4716 * tree-vrp.c: Likewise.
4718 2013-09-24 Alan Modra <amodra@gmail.com>
4722 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4723 for output operands that disallow regs. Don't use EXPAND_WRITE on
4726 2013-09-24 Richard Biener <rguenther@suse.de>
4729 * tree.c (reference_alias_ptr_type): Move ...
4730 * alias.c (reference_alias_ptr_type): ... here and implement
4731 in terms of the new reference_alias_ptr_type_1.
4732 (ref_all_alias_ptr_type_p): New helper.
4733 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4734 use ref_all_alias_ptr_type_p.
4735 (get_deref_alias_set): Add flag_strict_aliasing check here.
4736 (reference_alias_ptr_type_1): New function, split out from ...
4737 (get_alias_set): ... here.
4738 (alias_ptr_types_compatible_p): New function.
4739 * alias.h (reference_alias_ptr_type): Declare.
4740 (alias_ptr_types_compatible_p): Likewise.
4741 * tree.h (reference_alias_ptr_type): Remove.
4742 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4743 to compare MEM_REF alias types.
4745 2013-09-24 Richard Biener <rguenther@suse.de>
4747 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4749 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4751 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4752 reload helper function arrays into a single array reg_addr.
4753 (reload_fpr_gpr): Likewise.
4754 (reload_gpr_vsx): Likewise.
4755 (reload_vsx_gpr): Likewise.
4756 (struct rs6000_reg_addr): Likewise.
4757 (reg_addr): Likewise.
4758 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4759 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4760 (rs6000_init_hard_regno_mode_ok): Likewise.
4761 (rs6000_secondary_reload_direct_move): Likewise.
4762 (rs6000_secondary_reload): Likewise.
4764 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4765 constraints: wu, ww, and wy. Repurpose wv constraint added during
4766 power8 changes. Put wg constraint in alphabetical order.
4768 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4769 for future work to add ISA 2.07 VSX single precision support.
4770 (-mvsx-scalar-double): Change default from -1 to 1, update
4771 documentation comment.
4772 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4773 (-mupper-regs-df): New debug switch to control whether DF values
4774 can go in the traditional Altivec registers.
4775 (-mupper-regs-sf): New debug switch to control whether SF values
4776 can go in the traditional Altivec registers.
4778 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4780 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4781 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4782 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4783 (rs6000_debug_legitimate_address_p): Print if we are running
4784 before, during, or after reload.
4785 (rs6000_secondary_reload): Add a comment.
4786 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4788 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4789 constraints. Update documentation string.
4790 (wd constraint): Likewise.
4791 (wf constraint): Likewise.
4792 (wg constraint): Likewise.
4793 (wn constraint): Likewise.
4794 (ws constraint): Likewise.
4795 (wt constraint): Likewise.
4796 (wx constraint): Likewise.
4797 (wz constraint): Likewise.
4798 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4800 (ww constraint): Likewise.
4801 (wy constraint): Likewise.
4802 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4803 in the previous submissions.
4804 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4806 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4808 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4809 (REG_BR_PROB): Say that the probability is stored in an int_list.
4810 * reg-notes.def: Update commentary to mention INT_LIST.
4811 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4812 (INT_LIST): New rtx.
4813 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4814 * rtlanal.c (int_reg_note_p): New function.
4815 (alloc_reg_note): Assert that the note does not have an int argument.
4816 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4817 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4818 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4819 rather than an INSN_LIST. Handle INT_LIST.
4820 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4821 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4822 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4823 Manipulate them as ints rather than rtxes.
4824 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4825 * regmove.c (copy_src_to_dest): Likewise.
4826 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4828 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4829 into the cases that need it.
4830 * config/arm/arm.c (arm_unwind_emit): Likewise.
4832 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4833 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4834 * loop-doloop.c (add_test, doloop_modify): Likewise.
4835 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4836 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4837 * predict.c (combine_predictions_for_insn): Likewise.
4838 * print-rtl.c (print_rtx): Handle INT_LIST.
4839 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4840 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4841 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4842 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4843 (ix86_split_fp_branch, predict_jump): Likewise.
4844 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4845 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4846 * config/spu/spu.c (ea_load_store_inline): Likewise.
4848 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4849 value of a REG_BR_PROB note.
4850 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4851 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4852 * emit-rtl.c (try_split): Likewise.
4853 * predict.c (br_prob_note_reliable_p): Likewise.
4854 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4855 * reorg.c (mostly_true_jump): Likewise.
4856 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4857 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4858 * config/i386/i386.c (ix86_print_operand): Likewise.
4859 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4860 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4861 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4862 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4863 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4864 * config/sparc/sparc.c (output_cbranch): Likewise.
4865 * config/spu/spu.c (get_branch_target): Likewise.
4866 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4867 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4869 2013-09-23 Jan Hubicka <jh@suse.cz>
4871 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4873 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4874 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4876 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4877 (gimple_fold_call): Dump inconsistent devirtualizations; add
4878 sanity check for type based devirtualizations.
4879 * ipa-prop.c: Include ipa-utils.h
4880 (ipa_intraprocedural_devirtualization): Add sanity check.
4881 (try_make_edge_direct_virtual_call): Likewise.
4883 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4885 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4886 assignment statements.
4888 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4890 * gimple-pretty-print.c (dump_ssaname_info): New function.
4891 (dump_gimple_phi): Call it.
4892 (pp_gimple_stmt_1): Likewise.
4893 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4894 (range_info_def): Declare.
4895 * tree-pretty-print.c (pp_double_int): New function.
4896 (dump_generic_node): Call it.
4897 * tree-pretty-print.h (pp_double_int): Declare.
4898 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4899 * tree-ssanames.h (range_info_def): New structure.
4900 (value_range_type): Move definition here.
4901 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4903 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4905 (set_range_info): New function.
4906 (get_range_info): Likewise.
4907 (duplicate_ssa_name_range_info): Likewise.
4908 (duplicate_ssa_name_fn): Check pointer type and call
4909 duplicate_ssa_name_range_info.
4910 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4911 * tree-vrp.c (value_range_type): Remove definition, now in
4913 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4914 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4915 (SSA_NAME_RANGE_INFO): New macro.
4917 2013-09-23 Richard Biener <rguenther@suse.de>
4919 PR tree-optimization/58464
4920 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4921 (phi_trans_add): Change to add conditionally on being not present.
4922 (phi_translate_1): Remove recursion detection here.
4923 (phi_translate): Pre-seed the cache with NULL to catch
4924 recursion here in a more generic way.
4925 (bitmap_find_leader): Adjust comment.
4926 (get_representative_for): Dump value-numbers.
4927 (create_expression_by_pieces): Likewise.
4928 (insert_into_preds_of_block): Likewise.
4930 2013-09-23 Christian Bruel <christian.bruel@st.com>
4933 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4934 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4936 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4939 2013-09-20 Renlin Li <renlin.li@arm.com>
4941 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4943 (aarch64_expand_epilogue): Likewise.
4944 (aarch64_legitimize_reload_address): Likewise.
4946 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4948 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4949 NULL_TREE before pushing them onto the vector. Likewise for labels.
4951 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4953 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4954 comparison operators when -fno-trapping-math is in effect.
4955 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4956 comparison operators in TFmode and assert that unsupported operators
4958 (ia64_print_operand): Likewise.
4960 2013-09-21 Jan Hubicka <jh@suse.cz>
4962 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
4963 (sse_typeless_stores): Enable for core
4964 (sse_load0_by_pxor): Likewise.
4965 (four_jump_limit): Disable for core.
4966 (pad_returns): Likewise.
4967 (avoid_vector_decode): Likewise.
4968 (fuse_cmp_and_branch): Enable for cores.
4969 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
4971 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
4974 * config/pa/pa.c (pa_option_override): Disable auto increment and
4975 decrement instructions until reload is completed.
4977 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
4978 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
4979 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
4981 2013-09-20 DJ Delorie <dj@redhat.com>
4982 Nick Clifton <nickc@redhat.com>
4984 * config/rl78/rl78.c: Various whitespace and comment tweaks.
4985 (need_to_save): Save bank 0 on interrupts.
4986 (characterize_address): Strip far address wrappers.
4987 (rl78_as_legitimate_address): Likewise.
4988 (transcode_memory_rtx): Likewise.
4989 (rl78_peep_movhi_p): Disable this peephole after devirt.
4990 (rl78_propogate_register_origins): Forget all origins when a
4992 * config/rl78/rl78-virt.md: Various whitespace tweaks.
4993 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
4995 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
4996 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
4997 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
4999 * config/rl78/constraints.md: Various whitespace and paren tweaks.
5001 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5003 * config/pa/pa.md: In "scc" insn patterns, change output template to
5004 handle const0_rtx in reg_or_0_operand operands.
5006 2013-09-20 Martin Husemann <martin@NetBSD.org>
5009 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
5010 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
5012 2013-09-20 Richard Biener <rguenther@suse.de>
5015 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
5016 remembering SSA name version and block index.
5017 (new_scev_info_str): Adjust.
5018 (hash_scev_info): Likewise. Also hash the block index.
5019 (eq_scev_info): Adjust.
5020 (find_var_scev_info): Likewise.
5021 (struct instantiate_cache_entry): Remove.
5022 (struct instantiate_cache_type): Use a htab to map name, block
5024 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5025 (get_instantiated_value_entry): Likewise.
5026 (hash_idx_scev_info, eq_idx_scev_info): New functions.
5027 (instantiate_scev_name): Adjust.
5029 2013-09-20 Jeff Law <law@redhat.com>
5031 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
5033 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
5035 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
5036 Call aarch64_simd_expand_args to update op[argc].
5038 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5040 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
5043 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5045 * gengtype.c (file_rules): Added rule for *.cc files.
5046 (get_output_file_with_visibility): Give fatal message when no
5049 2013-09-20 Renlin Li <renlin.li@arm.com>
5051 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
5052 (aarch64_expand_epilogue): Likewise.
5053 (aarch64_legitimize_reload_address): Likewise.
5055 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
5058 * expr.c (expand_assignment): Remove misalignp code path.
5060 2013-09-20 Marek Polacek <polacek@redhat.com>
5063 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
5064 TYPE_PRECISION. Add asserts.
5066 2013-09-20 Richard Biener <rguenther@suse.de>
5068 PR tree-optimization/58453
5069 * tree-loop-distribution.c (distribute_loop): Apply the cost
5070 model for -ftree-loop-distribute-patterns, too.
5072 2013-09-20 Richard Biener <rguenther@suse.de>
5075 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
5076 make type comparison less strict.
5078 2013-09-20 Alan Modra <amodra@gmail.com>
5080 * configure: Regenerate.
5081 * aclocal.m4: Regenerate.
5083 2013-09-20 Marek Polacek <polacek@redhat.com>
5086 * doc/extend.texi: Document that attribute used is meant to be used
5087 on variables with static storage duration.
5089 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5091 PR tree-optimization/58472
5092 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
5093 simd_lane_access set inv_p = false.
5094 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
5095 the simduid magic VAR_DECL.
5097 2013-09-19 Jan Hubicka <jh@suse.cz>
5099 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
5101 2013-09-17 Jeff Law <law@redhat.com>
5103 * tree-ssa-dom.c (record_temporary_equivalences): New function
5104 split out of dom_opt_dom_walker::after_dom_children.
5105 (dom_opt_dom_walker::thread_across_edge): Move common code
5106 in here from dom_opt_dom_walker::after_dom_children.
5107 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
5109 2013-09-19 Jan Hubicka <jh@suse.cz>
5111 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
5112 (TARGET_GENERIC): Use PROCESOR_GENERIC
5113 (enum processor_type): Unify generic32 and 64.
5114 * i386.md (cpu): Likewise.
5115 * x86-tune.def (use_leave): Enable for generic32.
5116 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
5117 * athlon.md: Change generic64 to generic in all occurences.
5118 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
5119 (ix86_target_macros_internal): Likewise.
5120 * driver-i386.c (host_detect_local_cpu): Likewise.
5121 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
5123 (generic_memcpy, generic_memset, generic_cost): This one.
5124 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
5125 (m_GENERIC32, m_GENERIC64): Remove.
5126 (m_GENERIC): Turn into one flag.
5127 (processor_target): Unify generic tunnings.
5128 (ix86_option_override_internal): Replace generic32/64 by generic.
5129 (ix86_issue_rate): Likewise.
5130 (ix86_adjust_cost): Likewise.
5132 2013-09-19 Jan Hubicka <jh@suse.cz>
5134 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5135 of speculative flag.
5137 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5139 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5140 GOMP_sections_start, even if !exit_reachable.
5142 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5144 * lra-constraints.c (need_for_all_save_p): Use macro
5145 HARD_REGNO_CALL_PART_CLOBBERED.
5146 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5147 set up pseudo conflict hard regs.
5149 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5152 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5155 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5157 PR rtl-optimization/58438
5158 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5159 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5160 from previous subpasses.
5162 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5164 * arm.c (arm_get_frame_offsets): Validate architecture supports
5165 LDRD/STRD before accepting the tuning preference.
5166 (arm_expand_prologue): Likewise.
5167 (arm_expand_epilogue): Likewise.
5169 2013-09-18 Richard Biener <rguenther@suse.de>
5171 PR tree-optimization/58417
5172 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5173 have chrecs with symbols defined in the loop as operands.
5174 (chrec_fold_multiply): Likewise.
5175 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5176 parameters before folding binary operations.
5177 (struct instantiate_cache_entry_hasher): Remove.
5178 (struct instantiate_cache_type): Use a pointer-map.
5179 (instantiate_cache_type::instantiate_cache_type): New function.
5180 (instantiate_cache_type::get): Likewise.
5181 (instantiate_cache_type::set): Likewise.
5182 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5183 (get_instantiated_value_entry): Likewise.
5184 (global_cache): New global.
5185 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5186 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5187 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5188 (instantiate_scev_name): Adjust.
5189 (instantiate_scev): Construct global instead of local cache.
5190 (resolve_mixers): Likewise.
5192 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5193 Paolo Carlini <paolo.carlini@oracle.com>
5196 * doc/implement-cxx.texi: Fix references to the C++ standards.
5198 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5200 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5201 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5202 simd_lane_access drs, update also DR_ALIGNED_TO.
5204 2013-09-18 Marek Polacek <polacek@redhat.com>
5207 * doc/extend.texi: Document no_sanitize_undefined attribute.
5208 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5209 no_sanitize_undefined attribute.
5211 2013-09-18 Nick Clifton <nickc@redhat.com>
5213 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5214 (ASM_DECLARE_FUNCTION_NAME): Define.
5216 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5218 * compare-elim.c (find_comparison_dom_walker): New class
5219 (find_comparisons_in_bb): Rename to
5220 find_comparison_dom_walker::before_dom_children
5221 (find_comparisons): Adjust
5222 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5224 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5225 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5226 (init_walk_dominator_tree): Remove declaration.
5227 (fini_walk_dominator_tree): Remove declaration.
5228 * fwprop.c (single_def_use_dom_walker): New class
5229 (single_def_use_enter_block): Convert to
5230 single_def_use_dom_walker::before_dom_children.
5231 (single_def_use_leave_block): Convert to
5232 single_def_use_dom_walker::after_dom_children.
5233 (build_single_def_use_links): Adjust.
5234 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5236 (find_candidates_in_block): Convert to
5237 find_candidates_dom_walker::before_dom_children.
5238 (execute_strength_reduction): Adjust.
5239 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5240 (sese_dom_walker): New class.
5241 (sese_dom_walker::sese_dom_walker): New constructor.
5242 (sese_dom_walker::~sese_dom_walker): New destructor.
5243 (build_sese_conditions_before): Convert to
5244 sese_dom_walker::before_dom_children.
5245 (build_sese_conditions_after): Convert to
5246 sese_dom_walker::after_dom_children.
5247 (build_poly_scop): Adjust
5248 * tree-into-ssa.c (rewrite_dom_walker): New class
5249 (rewrite_enter_block): Convert to
5250 rewrite_dom_walker::before_dom_children.
5251 (rewrite_leave_block): Convert to
5252 rewrite_dom_walker::after_dom_children.
5253 (rewrite_update_dom_walker): New class.
5254 (rewrite_update_enter_block): Convert to
5255 rewrite_update_dom_walker::before_dom_children.
5256 (rewrite_update_leave_block): Convert to
5257 rewrite_update_dom_walker::after_dom_children.
5258 (rewrite_blocks, rewrite_into_ssa): Adjust.
5259 (mark_def_dom_walker): New class.
5260 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5261 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5262 (mark_def_sites_blocks): Convert to
5263 mark_def_dom_walker::before_dom_children.
5264 (mark_def_site_blocks): Remove.
5265 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5266 (tree_ssa_dominator_optimize): Adjust.
5267 (dom_thread_across_edge): Convert to method
5268 dom_opt_dom_walker::thread_across_edge.
5269 (dom_opt_enter_block): Convert to member function
5270 dom_opt_dom_walker::before_dom_children.
5271 (dom_opt_leave_block): Convert to member function
5272 dom_opt_dom_walker::after_dom_children.
5273 * tree-ssa-dse.c (dse_dom_walker): New class.
5274 (dse_enter_block): Convert to member function
5275 dse_dom_walker::before_dom_children.
5276 (tree_ssa_dse): Adjust.
5277 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5278 (determine_invariantness_stmt): Convert to method
5279 invariantness_dom_walker::before_dom_children.
5280 (determine_invariantness): Remove
5281 (move_computations_dom_walker): New class.
5282 (move_computations_stmt): Convert to method
5283 move_computations_dom_walker::before_dom_children.
5284 (move_computations, tree_ssa_lim): Adjust.
5285 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
5286 (nt_init_block): Make method
5287 notrappping_dom_walker::before_dom_children.
5288 (nt_fini_block): Make
5289 method nontrapping_dom_walker::after_dom_children.
5290 (get_non_trapping): Adjust.
5291 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5292 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
5293 (eliminate_leave_block): Make method.
5294 eliminate_dom_walker::after_dom_children.
5296 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5297 (strlen_enter_block): Make method
5298 strlen_dom_walker::before_dom_children.
5299 (strlen_leave_block): Make
5300 method strlen_dom_walker::after_dom_children.
5301 (tree_ssa_strlen): Adjust.
5302 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5303 (tree_ssa_uncprop): Adjust.
5304 (uncprop_leave_block): Make method
5305 uncprop_dom_walker::after_dom_children.
5306 (uncprop_leave_block): Make method
5307 uncprop_dom_walker::before_dom_children.
5309 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5311 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5312 branch in same basic block. Check both src and dest of the move insn.
5314 2013-09-17 Nick Clifton <nickc@redhat.com>
5316 * config/rl78/rl78-real.md (bf): New pattern.
5318 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5319 (rl78_print_operand): Do not put a # before a %B.
5320 * config/rl78/rl78.opt: Tweak doc strings.
5322 2013-09-17 DJ Delorie <dj@redhat.com>
5324 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5325 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5326 (rl78_option_override): Likewise, if -mallregs.
5327 (is_virtual_register): Likewise.
5328 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5329 (REGNO_OK_FOR_BASE_P): Likewise.
5330 * config/rl78/rl78.opt (-mallregs): New.
5332 2013-09-17 Nick Clifton <nickc@redhat.com>
5334 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5335 For interrupt functions: save all call clobbered registers if the
5336 interrupt handler is not a leaf function.
5337 (rl78_expand_prologue): Always recompute the frame information.
5338 For interrupt functions: only select bank 0 if one of the bank 0
5339 registers is going to be psuhed.
5341 2013-09-17 DJ Delorie <dj@redhat.com>
5343 * config/rl78/constraints.md: For each W* constraint, rename to C*
5344 and create a W* constraint that checks for an optional ES: prefix
5346 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5347 (es_addr): New. Used to wrap far addresses.
5348 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5349 (rl78_es_base): New.
5350 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5351 wrapped far addresses.
5352 (rl78_print_operand_1): Unwrap far addresses before processing.
5353 (rl78_lo16): Wrap far addresses in unspecs.
5354 (rl78_es_addr): New.
5355 (rl78_es_base): New.
5356 (insn_ok_now): Check for not-yet-wrapped far addresses.
5357 (transcode_memory_rtx): Properly re-wrap far addresses.
5359 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5361 * config/sparc/t-rtems: Add leon3 multilibs.
5363 2013-09-17 Cong Hou <congh@google.com>
5365 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5366 when checking the dot production pattern. The type of rhs operand
5367 of multiply is now checked correctly.
5369 2013-09-17 Jeff Law <law@redhat.com>
5371 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5372 edge implied equivalences into successor phis.
5373 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5374 here from tree-ssa-threadedge.c.
5375 (mark_threaded_blocks): When threading through a joiner, if both
5376 successors of the joiner's clone reach the same block, verify the
5377 PHI arguments are equal. If not, cancel the jump threading request.
5378 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5379 tree-ssa-threadupdate.c
5380 (thread_across_edge): Don't check PHI argument equality when
5381 threading through joiner block here.
5383 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5385 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5386 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5387 (warn_uninit, warn_uninitialized_vars,
5388 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5389 Move to tree-ssa-uninit.c.
5390 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5391 (has_undefined_value_p): New. Pass dependant parts of
5392 ssa_undefined_value_p.
5393 (uninit_undefined_value_p): Use has_undefined_value_p.
5394 (warn_uninit, warn_uninitialized_vars,
5395 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5396 Move from tree-ssa.c.
5397 * tree-ssa.h: Adjust prototypes.
5399 2013-09-17 Jan Hubicka <jh@suse.cz>
5402 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5403 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5404 FUNCTION_NOT_OPTIMIZED.
5405 * ipa-inline-analysis.c (compute_inline_parameters): Function
5406 not optimized is not inlinable unless it is alwaysinline.
5407 (inline_analyze_function): Force calls in not optimized
5408 function not inlinable.
5410 2013-09-17 Jan Hubicka <jh@suse.cz>
5413 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5415 * ipa.c (function_and_variable_visibility): Likewise.
5416 * ipa-profile.c (ipa_profile): Likewise.
5418 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5421 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5422 attribute, and don't inline the resolver in this case.
5424 2013-09-17 Teresa Johnson <tejohnson@google.com>
5426 * coverage.c (get_coverage_counts): Add missing newline.
5428 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5430 PR tree-optimization/58088
5431 * fold-const.c (mask_with_trailing_zeros): New function.
5432 (fold_binary_loc): Make sure we don't recurse infinitely
5433 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5434 Use mask_with_trailing_zeros where appropriate.
5436 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5438 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5439 of 'prev' var to get better distance estimation.
5441 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5443 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5444 * tree-inline.c (is_parameter_of): New predicate.
5445 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5446 a parameter has been remapped.
5447 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5448 (optimize_inline_calls): Initialize transform_parameter.
5449 (copy_gimple_seq_and_replace_locals): Likewise.
5450 (tree_function_versioning): Likewise.
5451 (maybe_inline_call_in_expr): Likewise.
5453 2013-09-17 Nick Clifton <nickc@redhat.com>
5455 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5456 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5458 (is_attr_func): New function.
5459 (msp430_is_interrupt_func): New function.
5460 (is_naked_func): New function.
5461 (is_reentrant_func): New function.
5462 (is_critical_func): New function.
5463 (msp430_start_function): Add annotations for function attributes.
5464 (msp430_attr): New function.
5465 (msp430_attribute_table): New.
5466 (msp430_function_section): New function.
5467 (TARGET_ASM_FUNCTION_SECTION): Define.
5468 (msp430_builtin): New enum.
5469 (msp430_init_builtins): New function.
5470 (msp430_builtin_devl): New function.
5471 (msp430_expand_builtin): New function.
5472 (TARGET_INIT_BUILTINS): Define.
5473 (TARGET_EXPAND_BUILTINS): Define.
5474 (TARGET_BUILTIN_DECL): Define.
5475 (msp430_expand_prologue): Add support for naked, interrupt,
5476 critical and reentrant functions.
5477 (msp430_expand_epilogue): Likewise.
5478 (msp430_print_operand): Handle 'O' character.
5479 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5481 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5482 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5483 (pushm): Use a 'n' rather than an 'i' constraint.
5484 (msp_return): Add generation of the interrupt return instruction.
5485 (disable_interrupts): New pattern.
5486 (enable_interrupts): New pattern.
5487 (push_intr_state): New pattern.
5488 (pop_intr_state): New pattern.
5489 (bic_SR): New pattern.
5490 (bis_SR): New pattern.
5491 * doc/extend.texi: Document MSP430 function attributes and builtin
5494 2013-09-17 Richard Biener <rguenther@suse.de>
5496 PR tree-optimization/58432
5497 * tree-loop-distribution.c (tree_loop_distribution): Also
5498 scan PHIs for outside loop uses and seed a partition from them.
5500 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5502 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5503 (restructure_reference): Call backtrace_base_for_ref.
5505 2013-09-17 Alan Modra <amodra@gmail.com>
5508 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5511 2013-09-16 DJ Delorie <dj@redhat.com>
5513 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5514 vregs location for RL78/G10.
5515 (rl78_expand_prologue): Avoid SEL on G10.
5516 (rl78_expand_epilogue): Likewise.
5517 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5518 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5519 __RL78_G10__ when appropriate.
5520 (ASM_SPEC): Pass -mg10 along to the assembler.
5521 * config/rl78/rl78.md (sel_rb): Disable for G10.
5522 * config/rl78/rl78.opt: Add -mg10 option.
5523 * config/rl78/t-rl78: Add -mg10 multilib.
5525 2013-09-16 Xinliang David Li <davidxl@google.com>
5527 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5528 * omp-low.c (omp_max_vf): Ditto.
5529 (expand_omp_simd): Ditto.
5530 * tree-vectorizer.c (vectorize_loops): Ditto.
5531 (gate_vect_slp): Ditto.
5532 (gate_increase_alignment): Ditto.
5533 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5534 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5535 (gate_tree_vectorize): Name change.
5536 (tree_vectorize): Ditto.
5537 (pass_vectorize::gate): Call new function.
5538 (pass_vectorize::execute): Ditto.
5539 * opts.c: O3 default setting change.
5540 (finish_options): Check new flag.
5541 * doc/invoke.texi: Document new flags.
5542 * common.opt: New flags.
5544 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5546 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5547 (Condition Code): Don't reference it.
5548 * doc/tm.texi: Regenerate.
5550 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5553 * lra-constraints.c (undo_optional_reloads): Consider all optional
5554 reload even if it did not get a hard reg.
5556 2013-09-16 Teresa Johnson <tejohnson@google.com>
5558 * dumpfile.c (dump_loc): Remove newline emission.
5559 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5560 emission to dump_printf_loc calls where missing.
5561 (vect_mark_for_runtime_alias_test): Ditto.
5562 (vect_analyze_data_ref_dependence): Ditto.
5563 (vect_analyze_data_ref_dependences): Ditto.
5564 (vect_slp_analyze_data_ref_dependence): Ditto.
5565 (vect_slp_analyze_data_ref_dependences): Ditto.
5566 (vect_compute_data_ref_alignment): Ditto.
5567 (vect_update_misalignment_for_peel): Ditto.
5568 (vect_verify_datarefs_alignment): Ditto.
5569 (vector_alignment_reachable_p): Ditto.
5570 (vect_get_data_access_cost): Ditto.
5571 (vect_enhance_data_refs_alignment): Ditto.
5572 (vect_find_same_alignment_drs): Ditto.
5573 (vect_analyze_data_refs_alignment): Ditto.
5574 (vect_analyze_group_access): Ditto.
5575 (vect_analyze_data_ref_access): Ditto.
5576 (vect_analyze_data_ref_accesses): Ditto.
5577 (vect_prune_runtime_alias_test_list): Ditto.
5578 (vect_analyze_data_refs): Ditto.
5579 (vect_create_addr_base_for_vector_ref): Ditto.
5580 (vect_create_data_ref_ptr): Ditto.
5581 (vect_grouped_store_supported): Ditto.
5582 (vect_grouped_load_supported): Ditto.
5583 * value-prof.c (check_counter): Ditto.
5584 (check_ic_target): Ditto.
5585 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5586 (vect_recog_widen_mult_pattern): Ditto.
5587 (vect_recog_widen_sum_pattern): Ditto.
5588 (vect_recog_over_widening_pattern): Ditto.
5589 (vect_recog_widen_shift_pattern): Ditto.
5590 (vect_recog_rotate_pattern): Ditto.
5591 (vect_recog_vector_vector_shift_pattern): Ditto.
5592 (vect_recog_divmod_pattern): Ditto.
5593 (vect_recog_mixed_size_cond_pattern): Ditto.
5594 (vect_recog_bool_pattern): Ditto.
5595 (vect_pattern_recog_1): Ditto.
5596 (vect_pattern_recog): Ditto.
5597 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5598 (vect_is_simple_iv_evolution): Ditto.
5599 (vect_analyze_scalar_cycles_1): Ditto.
5600 (vect_get_loop_niters): Ditto.
5601 (vect_analyze_loop_1): Ditto.
5602 (vect_analyze_loop_form): Ditto.
5603 (vect_analyze_loop_operations): Ditto.
5604 (vect_analyze_loop_2): Ditto.
5605 (vect_analyze_loop): Ditto.
5606 (report_vect_op): Ditto.
5607 (vect_is_slp_reduction): Ditto.
5608 (vect_is_simple_reduction_1): Ditto.
5609 (vect_get_known_peeling_cost): Ditto.
5610 (vect_estimate_min_profitable_iters): Ditto.
5611 (vect_model_reduction_cost): Ditto.
5612 (vect_model_induction_cost): Ditto.
5613 (get_initial_def_for_induction): Ditto.
5614 (vect_create_epilog_for_reduction): Ditto.
5615 (vectorizable_reduction): Ditto.
5616 (vectorizable_induction): Ditto.
5617 (vectorizable_live_operation): Ditto.
5618 (vect_loop_kill_debug_uses): Ditto.
5619 (vect_transform_loop): Ditto.
5620 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5621 (vect_stmt_relevant_p): Ditto.
5622 (process_use): Ditto.
5623 (vect_mark_stmts_to_be_vectorized): Ditto.
5624 (vect_model_simple_cost): Ditto.
5625 (vect_model_promotion_demotion_cost): Ditto.
5626 (vect_model_store_cost): Ditto.
5627 (vect_get_store_cost): Ditto.
5628 (vect_model_load_cost): Ditto.
5629 (vect_get_load_cost): Ditto.
5630 (vect_init_vector_1): Ditto.
5631 (vect_get_vec_def_for_operand): Ditto.
5632 (vect_finish_stmt_generation): Ditto.
5633 (vectorizable_call): Ditto.
5634 (vectorizable_conversion): Ditto.
5635 (vectorizable_assignment): Ditto.
5636 (vectorizable_shift): Ditto.
5637 (vectorizable_operation): Ditto.
5638 (vectorizable_store): Ditto.
5639 (vectorizable_load): Ditto.
5640 (vectorizable_condition): Ditto.
5641 (vect_analyze_stmt): Ditto.
5642 (vect_transform_stmt): Ditto.
5643 (vect_is_simple_use): Ditto.
5644 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5645 (vect_can_advance_ivs_p): Ditto.
5646 (vect_update_ivs_after_vectorizer): Ditto.
5647 (vect_do_peeling_for_loop_bound): Ditto.
5648 (vect_gen_niters_for_prolog_loop): Ditto.
5649 (vect_update_inits_of_drs): Ditto.
5650 (vect_create_cond_for_alias_checks): Ditto.
5651 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5652 (vect_build_slp_tree_1): Ditto.
5653 (vect_supported_load_permutation_p): Ditto.
5654 (vect_analyze_slp_instance): Ditto.
5655 (vect_analyze_slp): Ditto.
5656 (vect_make_slp_decision): Ditto.
5657 (vect_detect_hybrid_slp): Ditto.
5658 (vect_bb_vectorization_profitable_p): Ditto.
5659 (vect_slp_analyze_bb_1): Ditto.
5660 (vect_update_slp_costs_according_to_vf): Ditto.
5661 (vect_get_mask_element): Ditto.
5662 (vect_transform_slp_perm_load): Ditto.
5663 (vect_schedule_slp_instance): Ditto.
5664 (vect_schedule_slp): Ditto.
5665 (vect_slp_transform_bb): Ditto.
5666 * profile.c (read_profile_edge_counts): Ditto.
5667 (compute_branch_probabilities): Ditto.
5668 * coverage.c (get_coverage_counts): Ditto.
5670 2013-09-16 Diego Novillo <dnovillo@google.com>
5672 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5674 2013-09-16 Jan Hubicka <jh@suse.cz>
5676 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5677 abstract functions; for static functions check the presence of body.
5679 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5681 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5682 * config/aarch64/aarch64-simd.md
5683 (aarch64_mla_elt<mode>): New.
5684 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5685 (aarch64_mls_elt<mode>): Likewise.
5686 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5687 (aarch64_fma4_elt<mode>): Likewise.
5688 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5689 (aarch64_fma4_elt_to_128v2df): Likewise.
5690 (aarch64_fma4_elt_to_64df): Likewise.
5691 (fnma<mode>4): Likewise.
5692 (aarch64_fnma4_elt<mode>): Likewise.
5693 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5694 (aarch64_fnma4_elt_to_128v2df): Likewise.
5695 (aarch64_fnma4_elt_to_64df): Likewise.
5696 * config/aarch64/iterators.md (VDQSF): New.
5697 * config/aarch64/arm_neon.h
5698 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5699 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5701 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5703 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5704 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5705 (aarch64_mul3_elt_to_128df): Likewise.
5706 (aarch64_mul3_elt_to_64v2df): Likewise.
5707 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5709 (VMUL_CHANGE_NLANES) Likewise.
5712 * config/aarch64/arm_neon.h
5713 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5715 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5717 * config/aarch64/arm_neon.h
5718 (vcvtx_high_f32_f64): Fix parameters.
5720 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5721 Uros Bizjak <ubizjak@gmail.com>
5723 * config/alpha.c: Include tree-ssanames.h.
5725 2013-09-16 Richard Biener <rguenther@suse.de>
5727 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5728 (dot_rdg_1): Handle control_dd.
5729 (create_edge_for_control_dependence): New function.
5730 (create_rdg_edges): Add control dependences if asked for.
5731 (build_rdg): Likewise.
5732 (generate_loops_for_partition): If there are not necessary
5733 control stmts remove all their dependencies.
5734 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5735 (distribute_loop): Pass on control dependences.
5736 (tree_loop_distribution): Compute control dependences and remove
5737 restriction on number of loop nodes.
5739 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5741 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5744 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5746 * cse.c (try_const_anchors): Punt on CC modes.
5748 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5750 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5752 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5755 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5758 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5761 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5763 (darwin_objc1_section): Likewise.
5764 (darwin_file_end): Emit Image Info section when required.
5766 2013-09-14 Jan Hubicka <jh@suse.cz>
5768 * tree-into-ssa.c (gate_into_ssa): New.
5769 (pass_data_build_ssa): Use it.
5770 * cgraph.h (expand_thunk): Update prototype.
5771 * cgraphunit.c (analyze_function): Expand thunks early.
5772 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5773 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5774 set lowered flag; do not add new function.
5775 (assemble_thunks_and_aliases): Update.
5776 * tree-ssa.c (gate_init_datastructures): New gate.
5777 (pass_data_init_datastructures): Use it.
5779 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5782 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5783 xmm register set described in the psABI.
5785 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5787 * dwarf2out.c (should_emit_struct_debug): Add check
5788 for type_decl variable is not NULL.
5790 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5792 * config.gcc: Use new winnt-c.c target hooks
5793 * config/t-winnt: New file
5794 * config/winnt-c.c: New file
5795 * doc/tm.texi.in: Document new hook
5796 * doc/tm.texi: Regenerated
5798 2013-09-13 Jan Hubicka <jh@suse.cz>
5801 * ipa-inline.c (check_callers): New function.
5802 (check_caller_edge): Remove.
5803 (want_inline_function_to_all_callers_p): Also permit alises that are
5805 (inline_to_all_callers): Terminate the walk when devirtualization
5806 introduce new calls.
5808 2013-09-13 Jan Hubicka <jh@suse.cz>
5810 * ipa-inline-analysis.c (struct growth_data): Add node.
5811 (do_estimate_growth_1): Fix detection of recursion.
5813 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5815 PR tree-optimization/58392
5816 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5817 to avoid shadowing of outer loop variable. If
5818 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5819 replace_by_duplicate_decl simduid of loops that have it set and
5820 set dest_cfun->has_simduid_loops and/or
5821 dest_cfun->has_force_vect_loops.
5822 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5823 instead of maybe_lookup_decl.
5824 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5825 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5826 call. Copy over force_vect and copy and remap simduid. Set
5827 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5828 (copy_cfg_body): Remove blocks_to_copy argument. Use
5829 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5830 caller. Don't set cfun->has_simduid_loops and/or
5831 cfun->has_force_vect_loops here.
5832 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5834 (expand_call_inline, tree_function_versioning): Adjust copy_body
5837 2013-09-13 Martin Jambor <mjambor@suse.cz>
5840 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5841 the assert if the edge was a speculative one.
5843 2013-09-13 Richard Biener <rguenther@suse.de>
5845 * tree-data-ref.h (known_dependences_p): Move here ...
5846 * tree-loop-distribution.c (known_dependences_p): ... from here.
5847 (dump_rdg_component, debug_rdg_component): Remove.
5849 (generate_loops_for_partition): Use gimple_uid instead of
5850 relying on matching stmt visit order.
5851 (rdg_build_partitions): Take starting stmt vector.
5852 (ldist_gen): Merge into ...
5853 (distribute_loop): ... this function. Do not compute starting
5855 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5857 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5859 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5860 Set type attribute correctly. Set predicable_short_it attribute.
5861 (cmpsi_shiftsi): Remove %? from output template.
5863 2013-09-13 Richard Biener <rguenther@suse.de>
5865 * tree-loop-distribution.c (struct rdg_component,
5866 rdg_defs_used_in_other_loops_p, free_rdg_components,
5867 rdg_build_components): Remove.
5868 (stmts_from_loop): Do not record virtual PHIs.
5869 (generate_loops_for_partition): Skip virtual PHIs.
5870 (build_rdg_partition_for_component): Rename to ...
5871 (build_rdg_partition_for_vertex): ... this and adjust.
5872 (rdg_build_partitions): Take a vector of starting vertices
5873 instead of components. Remove unnecessary leftover handling.
5874 (ldist_gen): Do not build components or record other stores.
5875 (distribute_loop): Do not distribute loops containing stmts
5878 2013-09-13 Christian Bruel <christian.bruel@st.com>
5881 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5883 2013-09-13 Kai Tietz <ktietz@redhat.com>
5885 * config.gcc: Separate cases for mingw and cygwin targets,
5886 and add 64-bit cygwin target case.
5888 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5889 dll-export inline-functions.
5890 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5893 2013-09-13 Jeff Law <law@redhat.com>
5897 2013-09-06 Jeff Law <law@redhat.com>
5899 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5900 edge implied equivalences into successor phis.
5902 2013-09-12 DJ Delorie <dj@redhat.com>
5904 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5907 2013-09-12 Brooks Moses <bmoses@google.com>
5910 * Makefile.in: Do not install driver binaries in $(target)/bin.
5912 2013-09-12 DJ Delorie <dj@redhat.com>
5914 * config/rl78/rl78.opt (mrelax): New.
5915 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5916 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5918 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5920 (rl78_expand_epilogue): Likewise.
5922 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5925 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5926 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5927 pointer elimination with using remove_reg_equal_offset_note.
5929 2013-09-12 DJ Delorie <dj@redhat.com>
5931 * config/msp430/: New port.
5932 * config.gcc (msp430): Added.
5933 * doc/invoke.texi: Document MSP430 options.
5934 * doc/install.texi: Document msp430-elf
5935 * doc/md.texi: Document msp430-elf
5936 * doc/contrib.texi: Document msp430-elf
5938 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5941 2013-09-12 Martin Jambor <mjambor@suse.cz>
5944 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5945 reference descriptor is NULL.
5946 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5949 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5951 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5952 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5953 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5954 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5955 make_temp_ssa_name): move to tree-ssanames.h
5956 * tree-ssa-alias.h: Move prototype.
5957 * tree-ssa.h: Include tree-ssanames.h.
5958 * tree-ssanames.c (FREE_SSANAMES): Move to here.
5959 * tree-ssanames.h: New. Move items from tree-flow*.h
5960 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
5962 2013-09-12 Richard Biener <rguenther@suse.de>
5964 PR tree-optimization/58404
5965 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
5966 propagate non-invariant addresses into dereferences wrapped
5967 in component references.
5969 2013-09-12 Richard Biener <rguenther@suse.de>
5971 PR tree-optimization/58402
5972 * passes.def: Move pass_late_warn_uninitialized later.
5974 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5976 * tree-ssa.h: New. Move content from tree-flow.h and
5978 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
5979 Move prototypes belonging to tree-ssa.c.
5980 * tree-flow-inline.h (redirect_edge_var_map_def,
5981 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
5983 * gimple.h: Adjust prototypes.
5984 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
5986 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
5987 * tree.h: Move prototype to tree-ssa.h.
5988 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
5989 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
5990 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
5991 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
5992 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
5993 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
5994 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
5995 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
5996 graphite-blocking.c, graphite-clast-to-gimple.c,
5997 graphite-dependences.c, graphite-interchange.c,
5998 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
5999 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
6000 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
6001 ipa-reference.c, ipa-split.c, ipa-utils.c,
6002 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
6003 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
6004 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
6005 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6006 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
6007 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
6008 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
6009 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
6010 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
6011 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
6012 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
6013 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6014 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
6015 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
6016 value-prof.c, var-tracking.c,
6017 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
6019 2013-09-12 Richard Biener <rguenther@suse.de>
6021 PR tree-optimization/58396
6022 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
6023 (build_rdg): Take a loop-nest parameter, fix memleaks.
6024 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
6026 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6028 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
6031 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
6033 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
6034 The FMA instruction names should have a 'v' prefix.
6036 2013-09-12 Richard Biener <rguenther@suse.de>
6038 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
6039 (dot_rdg): Use popen instead of system in optional code.
6040 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
6041 (already_processed_vertex_p): Adjust.
6042 (has_anti_or_output_dependence, predecessor_has_mem_write,
6043 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
6044 rdg_flag_uses): Remove.
6045 (rdg_flag_vertex): Simplify.
6046 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
6048 (build_rdg_partition_for_component): Process the first vertex
6049 of a component only.
6050 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
6052 2013-09-12 Alan Modra <amodra@gmail.com>
6054 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
6056 2013-09-11 DJ Delorie <dj@redhat.com>
6057 Nick Clifton <nickc@redhat.com>
6059 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
6060 (rl78_stack_based_mem): New.
6061 * config/rl78/constraints.md (Iv08): New.
6070 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
6072 (movsi): Change from expand to insn-and-split.
6073 (ashrsi3): Clobber AX.
6077 * config/rl78/rl78.md (CC_REG): Fix.
6078 (addsi3): Allow memory and immediate operands.
6079 (addsi3_internal): Split into...
6080 (addsi3_internal_virt): ...new, and ...
6081 (addsi3_internal_real): ...new.
6083 (subsi3_internal_virt): New.
6084 (subsi3_internal_real): New.
6085 (mulsi3): Add memory operand.
6086 (mulsi3_rl78): Likewise.
6087 (mulsi3_g13): Likewise.
6088 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
6089 (cbranchqi4_real): Add more constraint options.
6090 (cbranchhi4_real): Expand pattern.
6091 (cbranchhi4_real_signed): New.
6092 (cbranchhi4_real_inverted): New.
6093 (cbranchsi4_real_lt): New.
6094 (cbranchsi4_real_ge): New.
6095 (cbranchsi4_real_signed): New.
6096 (cbranchsi4_real): New.
6098 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
6100 (lshrsi3_virt): Likewise.
6101 (ashlsi3_virt): Likewise.
6102 (cbranchqi4_virt_signed): New.
6103 (cbranchhi4_virt_signed): New.
6104 (cbranchsi4_virt): New.
6105 * config/rl78/rl78.c: Whitespace fixes throughout.
6106 (move_elim_pass): New.
6107 (pass_data_rl78_move_elim): New.
6108 (pass_rl78_move_elim): New.
6109 (make_pass_rl78_move_elim): New.
6110 (rl78_devirt_info): Run devirt earlier.
6111 (rl78_move_elim_info): New.
6112 (rl78_asm_file_start): Register it.
6113 (rl78_split_movsi): New.
6114 (rl78_as_legitimate_address): Allow virtual base registers when
6116 (rl78_addr_space_convert): Remove spurious debug stuff.
6117 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
6118 (rl78_print_operand): More cases for not printing '#'.
6119 (rl78_expand_compare): Remove most of the logic.
6120 (content_memory): New.
6121 (clear_content_memory): New.
6122 (get_content_index): New.
6123 (get_content_name): New.
6124 (display_content_memory): New.
6125 (update_content): New.
6126 (record_content): New.
6127 (already_contains): New.
6128 (insn_ok_now): Re-recog insns with virtual registers.
6129 (add_postponed_content_update): New.
6130 (process_postponed_content_update): New.
6131 (gen_and_emit_move): New.
6132 (transcode_memory_rtx): Record new location content.
6133 Use gen_and_emit_move.
6134 (force_into_acc): New.
6135 (move_to_acc): Use gen_and_emit_move.
6136 (move_from_acc): Likewise.
6137 (move_acc_to_reg): Likewise.
6138 (move_to_x): Likewise.
6139 (move_to_hl): Likewise.
6140 (move_to_de): Likewise.
6141 (rl78_alloc_physical_registers_op1): Record location content.
6142 (has_constraint): New.
6143 (rl78_alloc_physical_registers_op2): Record location content.
6145 (rl78_alloc_physical_registers_ro1): Likewise.
6146 (rl78_alloc_physical_registers_cmp): Likewise.
6147 (rl78_alloc_physical_registers_umul): Likewise.
6148 (rl78_alloc_address_registers_macax): New.
6149 (rl78_alloc_physical_registers): Initialize and set location
6150 content memory as needed.
6151 (rl78_reorg): Make sure split2 is called.
6152 (rl78_rtx_costs): New.
6154 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6156 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6157 for (not (neg ...)) and (neg (not ...)) cases.
6159 2013-09-11 Richard Biener <rguenther@suse.de>
6162 * passes.def: Split critical edges before late uninit warning passes.
6163 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6165 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6167 PR tree-optimization/58385
6168 * fold-const.c (build_range_check): If both low and high are NULL,
6169 use omit_one_operand_loc to preserve exp side-effects.
6171 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6173 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6175 2013-09-11 Richard Biener <rguenther@suse.de>
6177 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6178 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6179 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6180 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6181 stmts_from_loop, known_dependences_p, build_empty_rdg,
6182 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6183 * tree-loop-distribution.c: ... here.
6184 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6185 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6186 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6187 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6188 * tree-loop-distribution.c: ... here.
6189 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6190 (struct partition_s): Add loops member.
6191 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6192 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6193 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6195 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6196 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6197 Sergey Lega <sergey.s.lega@intel.com>
6198 Anna Tikhonova <anna.tikhonova@intel.com>
6199 Ilya Tocar <ilya.tocar@intel.com>
6200 Andrey Turetskiy <andrey.turetskiy@intel.com>
6201 Ilya Verbin <ilya.verbin@intel.com>
6202 Kirill Yukhin <kirill.yukhin@intel.com>
6203 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6205 * config/i386/constraints.md (k): New.
6207 * config/i386/i386.c (const regclass_map): Add new mask registers.
6208 (dbx_register_map): Ditto.
6209 (dbx64_register_map): Ditto.
6210 (svr4_dbx_register_map): Ditto.
6211 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6213 (ix86_preferred_reload_class): Disable constants for mask registers.
6214 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6215 (ix86_hard_regno_mode_ok): Support new mask registers.
6216 (x86_order_regs_for_local_alloc): Ditto.
6217 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6218 (FIXED_REGISTERS): Add new mask registers.
6219 (CALL_USED_REGISTERS): Ditto.
6220 (REG_ALLOC_ORDER): Ditto.
6221 (VALID_MASK_REG_MODE): New.
6222 (FIRST_MASK_REG): Ditto.
6223 (LAST_MASK_REG): Ditto.
6224 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6225 (MAYBE_MASK_CLASS_P): New.
6226 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6227 (REG_CLASS_CONTENTS): Ditto.
6228 (MASK_REGNO_P): New.
6229 (ANY_MASK_REG_P): Ditto.
6230 (HI_REGISTER_NAMES): Add new mask registers.
6231 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6232 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6234 (attribute "type"): Add mskmov, msklog.
6235 (attribute "length_immediate"): Support them.
6236 (attribute "memory"): Ditto.
6237 (attribute "prefix_0f"): Ditto.
6238 (*movhi_internal): Support new mask registers.
6239 (*movqi_internal): Ditto.
6240 (define_split): Split out clobber pattern is a logic
6241 insn on mask registers.
6242 (*k<logic><mode>): New.
6243 (*andhi_1): Extend to support mask regs.
6244 (*andqi_1): Extend to support mask regs.
6246 (define_split): Split and-not to and and not if operands
6248 (*<code><mode>_1): Separate HI mode to new pattern...
6249 (*<code>hi_1): This.
6250 (*<code>qi_1): Extend to support mask regs.
6252 (kortestzhi): Ditto.
6253 (kortestchi): Ditto.
6255 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6256 (*one_cmplhi2_1): ...Here, now with mask registers support.
6257 (*one_cmplqi2_1): Support new mask registers.
6258 (HI/QImode arithmetics splitter): Don't split if mask registers
6260 (HI/QImode not splitter): Ditto.
6261 * config/i386/predicated.md (mask_reg_operand): New.
6262 (general_reg_operand): Ditto.
6264 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6266 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6267 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6269 2013-09-10 Jeff Law <law@redhat.com>
6271 PR tree-optimization/58380
6272 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6273 of threading through a buried loop header.
6275 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6276 return value for single successor case.
6278 2013-09-10 Jan Hubicka <jh@suse.cz>
6280 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6282 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6285 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6286 support conditional execution.
6287 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6289 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6291 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6293 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6294 Check destination too to check move insn.
6295 (undo_optional_reloads): Add check that the original peudo did not
6296 changed its allocation and the optional reload was inherited on last
6297 inheritance pass. Break loop after deciding to keep optional reload.
6298 (lra_undo_inheritance): Add check that inherited pseudo still in
6301 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6303 * config/aarch64/aarch64.md (generic_sched): New.
6304 * config/aarch64/aarch64-generic.md (load): Make conditional
6305 on generic_sched attribute.
6306 (nonload): Likewise.
6308 2013-09-10 Jan Hubicka <jh@suse.cz>
6310 * lto-cgraph.c: Include ipa-utils.h.
6311 (compute_ltrans_boundary): Also add possible targets into the boundary.
6313 2013-09-10 Jan Hubicka <jh@suse.cz>
6315 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6316 VAR_DECL of vtable rather than full expression.
6318 2013-09-10 Jan Hubicka <jh@suse.cz>
6319 Paolo Carlini <paolo.carlini@oracle.com>
6321 * cgraphunit.c (analyze_functions): Save input_location, set it
6322 to UNKNOWN_LOCATION and restore it at the end.
6324 2013-09-10 Martin Jambor <mjambor@suse.cz>
6326 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6327 represented by a thunk.
6329 2013-09-10 Jeff Law <law@redhat.com>
6331 PR tree-optimization/58343
6332 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6333 jump threading requests through loop headers buried in the middle
6334 of a jump threading path.
6336 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6337 in return value/type.
6339 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6341 PR rtl-optimization/58365
6342 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6343 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6346 2013-09-10 Richard Biener <rguenther@suse.de>
6348 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6349 * tree-data-ref.c (create_rdg_vertices): Collect all data
6350 references, signal failure to the caller, use data-ref API.
6351 (build_rdg): Compute data references only once. Maintain lifetime
6352 of data references and data dependences from within RDG.
6353 (free_rdg): Free dependence relations.
6354 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6355 inventing extra dependences.
6356 (distribute_loop): Update for RDG API changes.
6358 2013-09-10 Kai Tietz <ktietz@redhat.com>
6360 * doc/invoke.texi (fms-extensions): Document changed
6361 behavior for ms-abi targets.
6362 * config/i386/i386.c (ix86_option_override_internal):
6363 Set default value of option -fms-extension for ms-abi targets.
6365 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6367 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6369 2013-09-10 Alan Modra <amodra@gmail.com>
6372 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6374 2013-09-10 Alan Modra <amodra@gmail.com>
6376 * config/rs6000/predicates.md (add_cint_operand): New.
6377 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6378 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6379 using add_cint_operand.
6380 (largetoc_high_plus_aix): Likewise.
6382 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6384 PR tree-optimization/58364
6385 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6386 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6387 the current range can't be an unconditional true or false.
6389 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6391 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6393 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6395 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6397 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6400 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6402 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6404 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6405 NO_REGS for immediate that can't be moved directly into FP_REGS.
6407 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6409 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6410 comparison with negated operand.
6411 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6414 2013-09-09 Richard Biener <rguenther@suse.de>
6417 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6418 of a subloop record all its block as affecting loop-closed SSA form.
6420 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6422 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6423 of an rtx/bitpos pair.
6424 (store_fixed_bit_field): Update accordingly.
6426 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6428 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6430 * builtins.c (expand_errno_check): Likewise.
6431 * dwarf2cfi.c (init_return_column_size): Likewise.
6432 * except.c (sjlj_mark_call_sites): Likewise.
6433 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6434 * lra-constraints.c (emit_inc): Likewise.
6435 * ree.c (combine_set_extension): Likewise.
6436 * regmove.c (fixup_match_2): Likewise.
6437 * reload1.c (inc_for_reload): Likewise.
6439 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6441 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6442 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6443 Use gen_int_mode with the mode of the associated simplify_* call.
6444 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6445 * expmed.c (expand_shift_1): Likewise.
6446 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6447 * loop-iv.c (iv_number_of_iterations): Likewise.
6448 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6449 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6451 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6453 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6454 of the associated expand_* call.
6455 (asan_emit_stack_protection): Likewise.
6456 * builtins.c (round_trampoline_addr): Likewise.
6457 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6458 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6459 (emit_store_flag): Likewise.
6460 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6461 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6463 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6464 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6465 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6466 * modulo-sched.c (generate_prolog_epilog): Likewise.
6467 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6468 (expand_ctz, expand_ffs, expand_unop): Likewise.
6470 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6472 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6473 of the associated gen_rtx_* call.
6474 * caller-save.c (init_caller_save): Likewise.
6475 * combine.c (find_split_point, make_extraction): Likewise.
6476 (make_compound_operation): Likewise.
6477 * dwarf2out.c (mem_loc_descriptor): Likewise.
6478 * explow.c (plus_constant, probe_stack_range): Likewise.
6479 * expmed.c (expand_mult_const): Likewise.
6480 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6481 * reload1.c (init_reload): Likewise.
6482 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6483 * var-tracking.c (adjust_mems): Likewise.
6484 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6485 rather than gen_rtx_fmt_ee.
6487 2013-09-09 Jan Hubicka <jh@suse.cz>
6490 * value-prof.c (gimple_ic): Copy also abnormal edges.
6492 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6494 * asan.c (asan_shadow_cst): Use gen_int_mode.
6496 2013-09-08 Jan Hubicka <jh@suse.cz>
6498 * ipa-profile.c: Add toplevel comment.
6499 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6500 (contains_hot_call_p): New function.
6501 (ipa_propagate_frequency): Set frequencies based on counts when
6503 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6504 profile; do not tamper with profile after inlining if it is read.
6506 2013-09-08 Jan Hubicka <jh@suse.cz>
6508 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6511 2013-09-08 Jan Hubicka <jh@suse.cz>
6513 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6516 2013-09-08 Jeff Law <law@redhat.com>
6519 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6522 2013-09-08 Andi Kleen <ak@linux.intel.com>
6524 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6526 2013-09-08 Jan Hubicka <jh@suse.cz>
6528 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6530 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6532 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6533 for non-debug insns.
6534 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6535 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6536 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6539 2013-09-08 Jan Hubicka <jh@suse.cz>
6541 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6542 targets and devirtualize to BUILT_IN_UNREACHABLE.
6543 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6544 * ipa.c (walk_polymorphic_call_targets): New function.
6545 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6546 functions; use the new timevar.
6547 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6548 was removed from the program.
6549 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6550 not consider it in the walk when its vtable is dead.
6551 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6553 (devirt_variable_node_removal_hook): New function.
6554 (possible_polymorphic_call_targets): Also register
6555 devirt_variable_node_removal_hook.
6556 (ipa_devirt): Do not do non-speculative devirtualization.
6557 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6559 2013-09-08 Jan Hubicka <jh@suse.cz>
6561 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6562 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6563 varpool_remove_variable_insertion_hook): Declare.
6564 * varpool.c (varpool_node_hook_list): New structure.
6565 (first_varpool_node_removal_hook,
6566 first_varpool_variable_insertion_hook): New variables.
6567 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6568 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6569 varpool_remove_variable_insertion_hook,
6570 varpool_call_variable_insertion_hooks): New functions.
6571 (varpool_remove_node): Use it.
6573 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6576 * diagnostic.c (diagnostic_build_prefix): When s.file is
6577 "<built-in>" don't output line and column numbers.
6579 2013-09-06 Jan Hubicka <jh@suse.cz>
6581 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6582 * lto-streamer-out.c: Stream thunks, too.
6583 * lto-streamer-in.c (input_function): Pop cfun here
6584 (lto_read_body): Instead of here.
6586 2013-09-06 Caroline Tice <cmtice@google.com>
6588 * doc/install.texi: Add documentation for the --enable-vtable-verify
6589 and the --disable-libvtv configure options.
6591 2013-09-06 Jeff Law <law@redhat.com>
6593 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6594 edge implied equivalences into successor phis.
6596 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6598 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6600 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6602 * resource.c (mark_target_live_regs): Compute resources taking
6603 into account if a call is predicated or not.
6605 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6607 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6608 the compiler in the function names.
6610 2013-09-06 Jan Hubicka <jh@suse.cz>
6613 * ipa-inline.c (has_caller_p): New function.
6614 (want_inline_function_to_all_callers_p): Use it.
6615 (sum_callers, inline_to_all_callers): Break out from ...
6616 (ipa_inline): ... here.
6618 2013-09-06 Jan Hubicka <jh@suse.cz>
6620 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6621 only when AVX is enabled.
6623 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6625 * config/aarch64/aarch64.md
6626 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6627 is fpsimd_<load/store>2.
6628 (load_pair<mode>): Likewise.
6629 (store_pair<mode>): Likewise.
6631 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6633 * config/arm/types.md (type): Add "mrs" type.
6634 * config/aarch64/aarch64.md
6635 (aarch64_load_tp_hard): Make type "mrs".
6637 (load_tp_hard): Make type "mrs".
6638 * config/arm/cortex-a15.md: Update with new attributes.
6639 * config/arm/cortex-a5.md: Update with new attributes.
6640 * config/arm/cortex-a53.md: Update with new attributes.
6641 * config/arm/cortex-a7.md: Update with new attributes.
6642 * config/arm/cortex-a8.md: Update with new attributes.
6643 * config/arm/cortex-a9.md: Update with new attributes.
6644 * config/arm/cortex-m4.md: Update with new attributes.
6645 * config/arm/cortex-r4.md: Update with new attributes.
6646 * config/arm/fa526.md: Update with new attributes.
6647 * config/arm/fa606te.md: Update with new attributes.
6648 * config/arm/fa626te.md: Update with new attributes.
6649 * config/arm/fa726te.md: Update with new attributes.
6651 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6653 * config/aarch64/aarch64.md
6654 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6655 (*movtf_aarch64): Likewise.
6657 (thumb1_movdi_insn): Use "multiple" for type where more than one
6658 instruction is used for a move.
6659 (*arm32_movhf): Likewise.
6660 (*thumb_movdf_insn): Likewise.
6662 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6664 * config/arm/types.md (type): Rename fcpys to fmov.
6666 (*arm_movsi_vfp): Rename type fcpys as fmov.
6667 (*thumb2_movsi_vfp): Likewise
6668 (*movhf_vfp_neon): Likewise
6669 (*movhf_vfp): Likewise
6670 (*movsf_vfp): Likewise
6671 (*thumb2_movsf_vfp): Likewise
6672 (*movsfcc_vfp): Likewise
6673 (*thumb2_movsfcc_vfp): Likewise
6674 * config/aarch64/aarch64-simd.md
6675 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6676 * config/aarch64/aarch64.md
6677 (*movsi_aarch64): Replace type mov_reg with fmovs.
6678 (*movdi_aarch64): Likewise
6679 (*movsf_aarch64): Likewise
6680 (*movdf_aarch64): Likewise
6682 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6683 * config/arm/iwmmxt.md
6684 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6685 * config/arm/arm1020e.md: Update with new attributes.
6686 * config/arm/cortex-a15-neon.md: Update with new attributes.
6687 * config/arm/cortex-a5.md: Update with new attributes.
6688 * config/arm/cortex-a53.md: Update with new attributes.
6689 * config/arm/cortex-a7.md: Update with new attributes.
6690 * config/arm/cortex-a8-neon.md: Update with new attributes.
6691 * config/arm/cortex-a9.md: Update with new attributes.
6692 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6693 * config/arm/cortex-r4f.md: Update with new attributes.
6694 * config/arm/marvell-pj4.md: Update with new attributes.
6695 * config/arm/vfp11.md: Update with new attributes.
6697 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6699 * config/aarch64/aarch64.md
6700 (*madd<mode>): Fix type attribute.
6701 (*maddsi_uxtw): Likewise.
6702 (*msub<mode>): Likewise.
6703 (*msubsi_uxtw): Likewise.
6704 (<su_optab>maddsidi4): Likewise.
6705 (<su_optab>msubsidi4): Likewise.
6707 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6709 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6710 * config/arm/arm.md (core_cycles): Remove fdiv.
6711 * config/arm/vfp.md:
6712 (*sqrtsf2_vfp): Update for attribute changes.
6713 (*sqrtdf2_vfp): Likewise.
6714 * config/aarch64/aarch64.md:
6715 (sqrt<mode>2): Update for attribute changes.
6716 * config/arm/arm1020e.md: Update with new attributes.
6717 * config/arm/cortex-a15-neon.md: Update with new attributes.
6718 * config/arm/cortex-a5.md: Update with new attributes.
6719 * config/arm/cortex-a53.md: Update with new attributes.
6720 * config/arm/cortex-a7.md: Update with new attributes.
6721 * config/arm/cortex-a8-neon.md: Update with new attributes.
6722 * config/arm/cortex-a9.md: Update with new attributes.
6723 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6724 * config/arm/cortex-r4f.md: Update with new attributes.
6725 * config/arm/marvell-pj4.md: Update with new attributes.
6726 * config/arm/vfp11.md: Update with new attributes.
6728 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6730 * config/arm/types.md
6731 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6732 * config/aarch64/aarch64.md
6733 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6735 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6736 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6737 (float<GPI:mode><GPF:mode>2): Likewise.
6739 (*truncsisf2_vfp): Update with new attributes.
6740 (*truncsidf2_vfp): Likewise.
6741 (fixuns_truncsfsi2): Likewise.
6742 (fixuns_truncdfsi2): Likewise.
6743 (*floatsisf2_vfp): Likewise.
6744 (*floatsidf2_vfp): Likewise.
6745 (floatunssisf2): Likewise.
6746 (floatunssidf2): Likewise.
6747 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6748 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6749 * config/arm/arm1020e.md: Update with new attributes.
6750 * config/arm/cortex-a15-neon.md: Update with new attributes.
6751 * config/arm/cortex-a5.md: Update with new attributes.
6752 * config/arm/cortex-a53.md: Update with new attributes.
6753 * config/arm/cortex-a7.md: Update with new attributes.
6754 * config/arm/cortex-a8-neon.md: Update with new attributes.
6755 * config/arm/cortex-a9.md: Update with new attributes.
6756 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6757 * config/arm/cortex-r4f.md: Update with new attributes.
6758 * config/arm/marvell-pj4.md: Update with new attributes.
6759 * config/arm/vfp11.md: Update with new attributes.
6761 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6763 * config/aarch64/arm_neon.h
6764 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6765 (vqtbx<1,2,3,4><q>_s8): Likewise.
6767 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6769 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6770 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6771 (add<mode>3): Add type attribute.
6772 (add<mode>3): Likewise.
6773 (usadd<mode>3): Likewise.
6774 (ssadd<mode>3): Likewise.
6775 (sub<mode>3): Likewise.
6776 (sub<mode>3): Likewise.
6777 (ussub<mode>3): Likewise.
6778 (sssub<mode>3): Likewise.
6779 (ssmulsa3): Likewise.
6780 (usmulusa3): Likewise.
6781 (arm_usatsihi): Likewise.
6783 (*movdi_vfp): Add types for all instructions.
6784 (*movdi_vfp_cortexa8): Likewise.
6785 (*movhf_vfp_neon): Likewise.
6786 (*movhf_vfp): Likewise.
6787 (*movdf_vfp): Likewise.
6788 (*thumb2_movdf_vfp): Likewise.
6789 (*thumb2_movdfcc_vfp): Likewise.
6790 * config/arm/arm.md: Add type attribute to all insn patterns.
6791 (*thumb1_adddi3): Add type attribute.
6792 (*arm_adddi3): Likewise.
6793 (*adddi_sesidi_di): Likewise.
6794 (*adddi_zesidi_di): Likewise.
6795 (*thumb1_addsi3): Likewise.
6796 (addsi3_compare0): Likewise.
6797 (*addsi3_compare0_scratch): Likewise.
6798 (*compare_negsi_si): Likewise.
6799 (cmpsi2_addneg): Likewise.
6800 (*addsi3_carryin_<optab>): Likewise.
6801 (*addsi3_carryin_alt2_<optab>): Likewise.
6802 (*addsi3_carryin_clobercc_<optab>): Likewise.
6803 (*subsi3_carryin): Likewise.
6804 (*subsi3_carryin_const): Likewise.
6805 (*subsi3_carryin_compare): Likewise.
6806 (*subsi3_carryin_compare_const): Likewise.
6807 (*arm_subdi3): Likewise.
6808 (*thumb_subdi3): Likewise.
6809 (*subdi_di_zesidi): Likewise.
6810 (*subdi_di_sesidi): Likewise.
6811 (*subdi_zesidi_di): Likewise.
6812 (*subdi_sesidi_di): Likewise.
6813 (*subdi_zesidi_ze): Likewise.
6814 (thumb1_subsi3_insn): Likewise.
6815 (*arm_subsi3_insn): Likewise.
6816 (*anddi3_insn): Likewise.
6817 (*anddi_zesidi_di): Likewise.
6818 (*anddi_sesdi_di): Likewise.
6819 (*ne_zeroextracts): Likewise.
6820 (*ne_zeroextracts): Likewise.
6821 (*ite_ne_zeroextr): Likewise.
6822 (*ite_ne_zeroextr): Likewise.
6823 (*anddi_notdi_di): Likewise.
6824 (*anddi_notzesidi): Likewise.
6825 (*anddi_notsesidi): Likewise.
6826 (andsi_notsi_si): Likewise.
6827 (thumb1_bicsi3): Likewise.
6828 (*iordi3_insn): Likewise.
6829 (*iordi_zesidi_di): Likewise.
6830 (*iordi_sesidi_di): Likewise.
6831 (*thumb1_iorsi3_insn): Likewise.
6832 (*xordi3_insn): Likewise.
6833 (*xordi_zesidi_di): Likewise.
6834 (*xordi_sesidi_di): Likewise.
6835 (*arm_xorsi3): Likewise.
6836 (*andsi_iorsi3_no): Likewise.
6837 (*smax_0): Likewise.
6838 (*smax_m1): Likewise.
6839 (*arm_smax_insn): Likewise.
6840 (*smin_0): Likewise.
6841 (*arm_smin_insn): Likewise.
6842 (*arm_umaxsi3): Likewise.
6843 (*arm_uminsi3): Likewise.
6844 (*minmax_arithsi): Likewise.
6845 (*minmax_arithsi_): Likewise.
6846 (*satsi_<SAT:code>): Likewise.
6847 (arm_ashldi3_1bit): Likewise.
6848 (arm_ashrdi3_1bit): Likewise.
6849 (arm_lshrdi3_1bit): Likewise.
6850 (*arm_negdi2): Likewise.
6851 (*thumb1_negdi2): Likewise.
6852 (*arm_negsi2): Likewise.
6853 (*thumb1_negsi2): Likewise.
6854 (*negdi_extendsid): Likewise.
6855 (*negdi_zero_extend): Likewise.
6856 (*arm_abssi2): Likewise.
6857 (*thumb1_abssi2): Likewise.
6858 (*arm_neg_abssi2): Likewise.
6859 (*thumb1_neg_abss): Likewise.
6860 (one_cmpldi2): Likewise.
6861 (extend<mode>di2): Likewise.
6862 (*compareqi_eq0): Likewise.
6863 (*arm_extendhisi2addsi): Likewise.
6864 (*arm_movdi): Likewise.
6865 (*thumb1_movdi_insn): Likewise.
6866 (*arm_movt): Likewise.
6867 (*thumb1_movsi_insn): Likewise.
6868 (pic_add_dot_plus_four): Likewise.
6869 (pic_add_dot_plus_eight): Likewise.
6870 (tls_load_dot_plus_eight): Likewise.
6871 (*thumb1_movhi_insn): Likewise.
6872 (*thumb1_movsf_insn): Likewise.
6873 (*movdf_soft_insn): Likewise.
6874 (*thumb_movdf_insn): Likewise.
6875 (cbranchsi4_insn): Likewise.
6876 (cbranchsi4_scratch): Likewise.
6877 (*negated_cbranchsi4): Likewise.
6878 (*tbit_cbranch): Likewise.
6879 (*tlobits_cbranch): Likewise.
6880 (*tstsi3_cbranch): Likewise.
6881 (*cbranchne_decr1): Likewise.
6882 (*addsi3_cbranch): Likewise.
6883 (*addsi3_cbranch_scratch): Likewise.
6884 (*arm_cmpdi_insn): Likewise.
6885 (*arm_cmpdi_unsig): Likewise.
6886 (*arm_cmpdi_zero): Likewise.
6887 (*thumb_cmpdi_zero): Likewise.
6888 (*deleted_compare): Likewise.
6889 (*mov_scc): Likewise.
6890 (*mov_negscc): Likewise.
6891 (*mov_notscc): Likewise.
6892 (*cstoresi_eq0_thumb1_insn): Likewise.
6893 (cstoresi_nltu_thumb1): Likewise.
6894 (cstoresi_ltu_thu): Likewise.
6895 (thumb1_addsi3_addgeu): Likewise.
6896 (*arm_jump): Likewise.
6897 (*thumb_jump): Likewise.
6898 (*check_arch2): Likewise.
6899 (arm_casesi_internal): Likewise.
6900 (thumb1_casesi_dispatch): Likewise.
6901 (*arm_indirect_jump): Likewise.
6902 (*thumb1_indirect_jump): Likewise.
6904 (*and_scc): Likewise.
6905 (*ior_scc): Likewise.
6906 (*compare_scc): Likewise.
6907 (*cond_move): Likewise.
6908 (*cond_arith): Likewise.
6909 (*cond_sub): Likewise.
6910 (*cmp_ite0): Likewise.
6911 (*cmp_ite1): Likewise.
6912 (*cmp_and): Likewise.
6913 (*cmp_ior): Likewise.
6914 (*ior_scc_scc): Likewise.
6915 (*ior_scc_scc_cmp): Likewise.
6916 (*and_scc_scc): Likewise.
6917 (*and_scc_scc_cmp): Likewise.
6918 (*and_scc_scc_nod): Likewise.
6919 (*negscc): Likewise.
6920 (movcond_addsi): Likewise.
6921 (movcond): Likewise.
6922 (*ifcompare_plus_move): Likewise.
6923 (*if_plus_move): Likewise.
6924 (*ifcompare_move_plus): Likewise.
6925 (*if_move_plus): Likewise.
6926 (*ifcompare_arith_arith): Likewise.
6927 (*if_arith_arith): Likewise.
6928 (*ifcompare_arith_move): Likewise.
6929 (*if_arith_move): Likewise.
6930 (*ifcompare_move_arith): Likewise.
6931 (*if_move_arith): Likewise.
6932 (*ifcompare_move_not): Likewise.
6933 (*if_move_not): Likewise.
6934 (*ifcompare_not_move): Likewise.
6935 (*if_not_move): Likewise.
6936 (*ifcompare_shift_move): Likewise.
6937 (*if_shift_move): Likewise.
6938 (*ifcompare_move_shift): Likewise.
6939 (*if_move_shift): Likewise.
6940 (*ifcompare_shift_shift): Likewise.
6941 (*ifcompare_not_arith): Likewise.
6942 (*ifcompare_arith_not): Likewise.
6943 (*if_arith_not): Likewise.
6944 (*ifcompare_neg_move): Likewise.
6945 (*if_neg_move): Likewise.
6946 (*ifcompare_move_neg): Likewise.
6947 (*if_move_neg): Likewise.
6948 (prologue_thumb1_interwork): Likewise.
6949 (*cond_move_not): Likewise.
6950 (*sign_extract_onebit): Likewise.
6951 (*not_signextract_onebit): Likewise.
6952 (stack_tie): Likewise.
6953 (align_4): Likewise.
6954 (align_8): Likewise.
6955 (consttable_end): Likewise.
6956 (consttable_1): Likewise.
6957 (consttable_2): Likewise.
6958 (consttable_4): Likewise.
6959 (consttable_8): Likewise.
6960 (consttable_16): Likewise.
6961 (*thumb1_tablejump): Likewise.
6962 (prefetch): Likewise.
6963 (force_register_use): Likewise.
6964 (thumb_eh_return): Likewise.
6965 (load_tp_hard): Likewise.
6966 (load_tp_soft): Likewise.
6967 (tlscall): Likewise.
6968 (*arm_movtas_ze): Likewise.
6969 (*arm_rev): Likewise.
6970 (*arm_revsh): Likewise.
6971 (*arm_rev16): Likewise.
6972 * config/arm/thumb2.md
6973 (*thumb2_smaxsi3): Likewise.
6974 (*thumb2_sminsi3): Likewise.
6975 (*thumb32_umaxsi3): Likewise.
6976 (*thumb2_uminsi3): Likewise.
6977 (*thumb2_negdi2): Likewise.
6978 (*thumb2_abssi2): Likewise.
6979 (*thumb2_neg_abss): Likewise.
6980 (*thumb2_movsi_insn): Likewise.
6981 (tls_load_dot_plus_four): Likewise.
6982 (*thumb2_movhi_insn): Likewise.
6983 (*thumb2_mov_scc): Likewise.
6984 (*thumb2_mov_negs): Likewise.
6985 (*thumb2_mov_negs): Likewise.
6986 (*thumb2_mov_nots): Likewise.
6987 (*thumb2_mov_nots): Likewise.
6988 (*thumb2_movsicc_): Likewise.
6989 (*thumb2_movsfcc_soft_insn): Likewise.
6990 (*thumb2_indirect_jump): Likewise.
6991 (*thumb2_and_scc): Likewise.
6992 (*thumb2_ior_scc): Likewise.
6993 (*thumb2_ior_scc_strict_it): Likewise.
6994 (*thumb2_cond_move): Likewise.
6995 (*thumb2_cond_arith): Likewise.
6996 (*thumb2_cond_ari): Likewise.
6997 (*thumb2_cond_sub): Likewise.
6998 (*thumb2_negscc): Likewise.
6999 (*thumb2_movcond): Likewise.
7000 (thumb2_casesi_internal): Likewise.
7001 (thumb2_casesi_internal_pic): Likewise.
7002 (*thumb2_alusi3_short): Likewise.
7003 (*thumb2_mov<mode>_shortim): Likewise.
7004 (*thumb2_addsi_short): Likewise.
7005 (*thumb2_subsi_short): Likewise.
7006 (thumb2_addsi3_compare0): Likewise.
7007 (*thumb2_cbz): Likewise.
7008 (*thumb2_cbnz): Likewise.
7009 (*thumb2_one_cmplsi2_short): Likewise.
7010 (*thumb2_negsi2_short): Likewise.
7011 (*orsi_notsi_si): Likewise.
7012 * config/arm/arm1020e.md: Update with new attributes.
7013 * config/arm/arm1026ejs.md: Update with new attributes.
7014 * config/arm/arm1136jfs.md: Update with new attributes.
7015 * config/arm/arm926ejs.md: Update with new attributes.
7016 * config/arm/cortex-a15.md: Update with new attributes.
7017 * config/arm/cortex-a5.md: Update with new attributes.
7018 * config/arm/cortex-a53.md: Update with new attributes.
7019 * config/arm/cortex-a7.md: Update with new attributes.
7020 * config/arm/cortex-a8.md: Update with new attributes.
7021 * config/arm/cortex-a9.md: Update with new attributes.
7022 * config/arm/cortex-m4.md: Update with new attributes.
7023 * config/arm/cortex-r4.md: Update with new attributes.
7024 * config/arm/fa526.md: Update with new attributes.
7025 * config/arm/fa606te.md: Update with new attributes.
7026 * config/arm/fa626te.md: Update with new attributes.
7027 * config/arm/fa726te.md: Update with new attributes.
7029 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7031 * config/aarch64/aarch64-simd.md
7032 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
7033 <vwx> iterator to ensure correct register choice.
7034 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7035 (aarch64_sqdmull_n<mode>): Likewise.
7036 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7037 * config/aarch64/arm_neon.h
7038 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
7039 (vml<as><q>_n_<su>16): Likewise.
7040 (vml<as>l_high_lane<q>_<su>16): Likewise.
7041 (vml<as>l_high_n_<su>16): Likewise.
7042 (vml<as>l_lane<q>_<su>16): Likewise.
7043 (vml<as>l_n_<su>16): Likewise.
7044 (vmul<q>_lane<q>_<su>16): Likewise.
7045 (vmul<q>_n_<su>16): Likewise.
7046 (vmull_lane<q>_<su>16): Likewise.
7047 (vmull_n_<su>16): Likewise.
7048 (vmull_high_lane<q>_<su>16): Likewise.
7049 (vmull_high_n_<su>16): Likewise.
7050 (vqrdmulh<q>_n_s16): Likewise.
7052 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
7054 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
7055 have the correct lane parameter.
7057 2013-09-06 Richard Biener <rguenther@suse.de>
7059 * cfganal.c (control_dependences::~control_dependences):
7060 Properly free all of the vector.
7062 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
7065 * config/i386/i386.c (ix86_conditional_register_usage):
7066 Proper initialize extended SSE registers.
7068 2013-09-06 Jan Hubicka <jh@suse.cz>
7070 PR tree-optimization/58311
7071 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
7073 2013-09-06 Jan Hubicka <jh@suse.cz>
7075 * Makefile.in (tree-sra.o): Update dependencies.
7076 * tree-sra.c: Include ipa-utils.h
7077 (scan_function): Use recursive_call_p.
7078 (has_caller_p): New function.
7079 (cgraph_for_node_and_aliases): Count also callers of aliases.
7081 2013-09-06 Jan Hubicka <jh@suse.cz>
7084 * cgraph.h (symtab_semantically_equivalent_p): Declare.
7085 * tree-tailcall.c: Include ipa-utils.h.
7086 (find_tail_calls): Use it.
7087 * ipa-pure-const.c (check_call): Likewise.
7088 * ipa-utils.c (recursive_call_p): New function.
7089 * ipa-utils.h (recursive_call_p): Dclare.
7090 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
7091 (symtab_semantically_equivalent_p): New function.
7092 * Makefile.in (tree-tailcall.o): Update dependencies.
7094 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7096 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
7099 2013-09-06 Richard Biener <rguenther@suse.de>
7101 * lto-streamer.h (lto_global_var_decls): Remove.
7102 * Makefile.in (OBJS): Remove lto-symtab.o.
7103 (lto-symtab.o): Remove.
7104 (GTFILES): Remove lto-symtab.c
7105 * lto-symtab.c: Move to lto/
7107 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7109 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
7110 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
7111 (UNSPEC_FPINT_RINT): New constant definitions.
7112 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
7113 definition with 2 attributes.
7114 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
7115 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
7118 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7120 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
7121 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
7122 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
7123 attribute to "z196_alone".
7124 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
7127 2013-09-06 Richard Biener <rguenther@suse.de>
7129 * basic-block.h (class control_dependences): New.
7130 * tree-ssa-dce.c (control_dependence_map): Remove.
7132 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7133 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7134 find_pdom, find_control_dependence, find_all_control_dependences):
7136 (mark_control_dependent_edges_necessary,
7137 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7138 tree_dce_done, perform_tree_ssa_dce): Adjust.
7139 * cfganal.c (set_control_dependence_map_bit,
7140 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7141 find_all_control_dependences): Move from tree-ssa-dce.c and
7142 implement as methods of control_dependences class.
7143 (control_dependences::control_dependences): New.
7144 (control_dependences::~control_dependences): Likewise.
7145 (control_dependences::get_edges_dependent_on): Likewise.
7146 (control_dependences::get_edge): Likewise.
7148 2013-09-04 Jan Hubicka <jh@suse.cz>
7150 * tree.c (types_same_for_odr): Drop overactive check.
7151 * ipa-devirt.c (hash_type_name): Likewise.
7153 2013-09-04 Jan Hubicka <jh@suse.cz>
7155 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7156 (analyze_functions): ... here.
7158 2013-09-04 Jan Hubicka <jh@suse.cz>
7161 * cgraphunit.c (analyze_functions): Clear AUX fields
7162 after processing; initialize assembler name has.
7164 2013-09-05 Jeff Law <law@redhat.com>
7166 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7167 from thread_around_empty_block. Record threading path into PATH.
7168 Recurse if threading through the initial block is successful.
7169 (thread_across_edge): Corresponding changes to slightly simplify.
7171 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7173 * config/aarch64/aarch64.md
7174 (type): Remove frecpe, frecps, frecpx.
7175 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7176 fix to be a TARGET_SIMD instruction.
7177 (aarch64_frecps): Remove.
7178 * config/aarch64/aarch64-simd.md
7179 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7180 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7182 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7183 Sofiane Naci <sofiane.naci@arm.com>
7185 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7186 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7187 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7188 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7189 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7190 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7191 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7192 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7193 into "clz, "rbit". Rename "shift" to "shift_imm".
7194 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7195 changes. Update for attribute changes all occurrences of arlo_* and
7197 * config/arm/arm-fixed.md: Update for attribute changes
7198 all occurrences of arlo_* types.
7199 * config/arm/thumb2.md: Update for attribute changes all occurrences
7201 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7202 (cortexa7_older_only): Likewise.
7203 (cortexa7_younger): Likewise.
7204 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7205 (1020alu_shift_op): Likewise.
7206 (1020alu_shift_reg_op): Likewise.
7207 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7208 (alu_shift_op): Likewise.
7209 (alu_shift_reg_op): Likewise.
7210 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7211 (11_alu_shift_op): Likewise.
7212 (11_alu_shift_reg_op): Likewise.
7213 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7214 (9_alu_shift_reg_op): Likewise.
7215 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7217 (cortex_a15_alu_shift): Likewise.
7218 (cortex_a15_alu_shift_reg): Likewise.
7219 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7221 (cortex_a5_alu_shift): Likewise.
7222 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7224 (cortex_a53_alu_shift): Likewise.
7225 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7227 (cortex_a7_alu_reg): Likewise.
7228 (cortex_a7_alu_shift): Likewise.
7229 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7231 (cortex_a8_alu_shift): Likewise.
7232 (cortex_a8_alu_shift_reg): Likewise.
7233 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7234 (cortex_a9_dp_shift): Likewise.
7235 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7237 * config/arm/cortex-r4.md
7238 (cortex_r4_alu): Update for attribute changes.
7239 (cortex_r4_mov): Likewise.
7240 (cortex_r4_alu_shift_reg): Likewise.
7241 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7242 (526_alu_shift_op): Likewise.
7243 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7244 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7245 (626te_alu_shift_op): Likewise.
7246 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7247 (726te_alu_shift_op): Likewise.
7248 (726te_alu_shift_reg_op): Likewise.
7249 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7250 (mp626_alu_shift_op): Likewise.
7251 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7252 (pj4_alu_conds): Likewise.
7253 (pj4_shift): Likewise.
7254 (pj4_shift_conds): Likewise.
7255 (pj4_alu_shift): Likewise.
7256 (pj4_alu_shift_conds): Likewise.
7257 * config/aarch64/aarch64.md: Update for attribute change
7258 all occurrences of arlo_* and shift* types.
7260 2013-09-05 Mike Stump <mikestump@comcast.net>
7262 * tree.h: Move documentation for tree_function_decl to tree-core.h
7263 with the declaration.
7265 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7268 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7269 looking for widest mode.
7271 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7273 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7275 2013-09-05 Richard Biener <rguenther@suse.de>
7277 PR tree-optimization/58137
7278 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7279 Do not create vectors of pointers.
7280 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7281 types for the components of the vector initializer.
7282 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7283 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7285 2013-09-05 Martin Jambor <mjambor@suse.cz>
7287 * ipa-prop.c (remove_described_reference): Accept missing references,
7288 return false if that hppens, otherwise return true.
7289 (cgraph_node_for_jfunc): New function.
7290 (try_decrement_rdesc_refcount): Likewise.
7291 (try_make_edge_direct_simple_call): Use them.
7292 (ipa_edge_removal_hook): Remove references from rdescs.
7293 (ipa_edge_duplication_hook): Clone rdescs and their references
7294 when the new edge has the same caller as the old one.
7295 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7296 reference before removing any edges.
7298 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7300 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7302 * thumb2.md (thumb2_storewb_parisi): New pattern.
7304 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7306 * config/aarch64/aarch64-option-extensions.def: Add
7307 AARCH64_OPT_EXTENSION of 'crc'.
7308 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7309 (AARCH64_ISA_CRC): Ditto.
7310 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7311 description of the CRC extension.
7313 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7315 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7316 * config/rs6000/linux.h: Ditto.
7317 * alpha/linux.h: Ditto.
7318 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7319 no_c99_libc_has_function.
7320 * config/c6x/uclinux-elf.h: Ditto.
7321 * config/lm32/uclinux-elf.h: Ditto.
7322 * config/m68k/uclinux.h: Ditto.
7323 * config/moxie/uclinux.h: Ditto.
7324 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7325 (crisv32-*-linux*, cris-*-linux*): Ditto.
7326 * config/bfin/bfin.c: Include "tm_p.h".
7328 2013-09-05 Richard Biener <rguenther@suse.de>
7330 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7331 check for a definition without a basic-block.
7333 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7334 Sofiane Naci <sofiane.naci@arm.com>
7336 * config/aarch64/aarch64.md
7337 (*movti_aarch64): Rename r_2_f and f_2_r.
7338 (*movsf_aarch64): Likewise.
7339 (*movdf_aarch64): Likewise.
7340 (*movtf_aarch64): Likewise.
7341 (aarch64_movdi_<mode>low): Likewise.
7342 (aarch64_movdi_<mode>high): Likewise.
7343 (aarch64_mov<mode>high_di): Likewise.
7344 (aarch64_mov<mode>low_di): Likewise.
7345 (aarch64_movtilow_tilow): Likewise.
7346 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7347 values to config/arm/types.md
7348 (attribute "conds"): Update for attribute change.
7349 (anddi3_insn): Likewise.
7350 (iordi3_insn): Likewise.
7351 (xordi3_insn): Likewise.
7352 (one_cmpldi2): Likewise.
7353 * config/arm/types.md (type): Add Neon types.
7354 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7355 use "type" attribute.
7356 (movmisalign<mode>_neon_store): Likewise.
7357 (movmisalign<mode>_neon_load): Likewise.
7358 (vec_set<mode>_internal): Likewise.
7359 (vec_setv2di_internal): Likewise.
7360 (vec_extract<mode>): Likewise.
7361 (vec_extractv2di): Likewise.
7362 (add<mode>3_neon): Likewise.
7363 (adddi3_neon): Likewise.
7364 (sub<mode>3_neon): Likewise.
7365 (subdi3_neon): Likewise.
7366 (mul<mode>3_neon): Likewise.
7367 (mul<mode>3add<mode>_neon): Likewise.
7368 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7369 (fma<VCVTF:mode>4)): Likewise.
7370 (fma<VCVTF:mode>4_intrinsic): Likewise.
7371 (fmsub<VCVTF:mode>4)): Likewise.
7372 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7373 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7374 (ior<mode>3): Likewise.
7375 (and<mode>3): Likewise.
7376 (anddi3_neon): Likewise.
7377 (orn<mode>3_neon): Likewise.
7378 (orndi3_neon): Likewise.
7379 (bic<mode>3_neon): Likewise.
7380 (bicdi3_neon): Likewise.
7381 (xor<mode>3): Likewise.
7382 (one_cmpl<mode>2): Likewise.
7383 (abs<mode>2): Likewise.
7384 (neg<mode>2): Likewise.
7385 (umin<mode>3_neon): Likewise.
7386 (umax<mode>3_neon): Likewise.
7387 (smin<mode>3_neon): Likewise.
7388 (smax<mode>3_neon): Likewise.
7389 (vashl<mode>3): Likewise.
7390 (vashr<mode>3_imm): Likewise.
7391 (vlshr<mode>3_imm): Likewise.
7392 (ashl<mode>3_signed): Likewise.
7393 (ashl<mode>3_unsigned): Likewise.
7394 (neon_load_count): Likewise.
7395 (ashldi3_neon_noclobber): Likewise.
7396 (signed_shift_di3_neon): Likewise.
7397 (unsigned_shift_di3_neon): Likewise.
7398 (ashrdi3_neon_imm_noclobber): Likewise.
7399 (lshrdi3_neon_imm_noclobber): Likewise.
7400 (widen_ssum<mode>3): Likewise.
7401 (widen_usum<mode>3): Likewise.
7402 (quad_halves_<code>v4si): Likewise.
7403 (quad_halves_<code>v4sf): Likewise.
7404 (quad_halves_<code>v8hi): Likewise.
7405 (quad_halves_<code>v16qi): Likewise.
7406 (reduc_splus_v2di): Likewise.
7407 (neon_vpadd_internal<mode>): Likewise.
7408 (neon_vpsmin<mode>): Likewise.
7409 (neon_vpsmax<mode>): Likewise.
7410 (neon_vpumin<mode>): Likewise.
7411 (neon_vpumax<mode>): Likewise.
7412 (ss_add<mode>_neon): Likewise.
7413 (us_add<mode>_neon): Likewise.
7414 (ss_sub<mode>_neon): Likewise.
7415 (us_sub<mode>_neon): Likewise.
7416 (neon_vadd<mode>_unspec): Likewise.
7417 (neon_vaddl<mode>): Likewise.
7418 (neon_vaddw<mode>): Likewise.
7419 (neon_vhadd<mode>): Likewise.
7420 (neon_vqadd<mode>): Likewise.
7421 (neon_vaddhn<mode>): Likewise.
7422 (neon_vmul<mode>): Likewise.
7423 (neon_vmla<mode>): Likewise.
7424 (neon_vmlal<mode>): Likewise.
7425 (neon_vmls<mode>): Likewise.
7426 (neon_vmlsl<mode>): Likewise.
7427 (neon_vqdmulh<mode>): Likewise.
7428 (neon_vqdmlal<mode>): Likewise.
7429 (neon_vqdmlsl<mode>): Likewise.
7430 (neon_vmull<mode>): Likewise.
7431 (neon_vqdmull<mode>): Likewise.
7432 (neon_vsub<mode>_unspec): Likewise.
7433 (neon_vsubl<mode>): Likewise.
7434 (neon_vsubw<mode>): Likewise.
7435 (neon_vqsub<mode>): Likewise.
7436 (neon_vhsub<mode>): Likewise.
7437 (neon_vsubhn<mode>): Likewise.
7438 (neon_vceq<mode>): Likewise.
7439 (neon_vcge<mode>): Likewise.
7440 (neon_vcgeu<mode>): Likewise.
7441 (neon_vcgt<mode>): Likewise.
7442 (neon_vcgtu<mode>): Likewise.
7443 (neon_vcle<mode>): Likewise.
7444 (neon_vclt<mode>): Likewise.
7445 (neon_vcage<mode>): Likewise.
7446 (neon_vcagt<mode>): Likewise.
7447 (neon_vtst<mode>): Likewise.
7448 (neon_vabd<mode>): Likewise.
7449 (neon_vabdl<mode>): Likewise.
7450 (neon_vaba<mode>): Likewise.
7451 (neon_vabal<mode>): Likewise.
7452 (neon_vmax<mode>): Likewise.
7453 (neon_vmin<mode>): Likewise.
7454 (neon_vpaddl<mode>): Likewise.
7455 (neon_vpadal<mode>): Likewise.
7456 (neon_vpmax<mode>): Likewise.
7457 (neon_vpmin<mode>): Likewise.
7458 (neon_vrecps<mode>): Likewise.
7459 (neon_vrsqrts<mode>): Likewise.
7460 (neon_vqabs<mode>): Likewise.
7461 (neon_vqneg<mode>): Likewise.
7462 (neon_vcls<mode>): Likewise.
7463 (clz<mode>2): Likewise.
7464 (popcount<mode>2): Likewise.
7465 (neon_vrecpe): Likewise.
7466 (neon_vrsqrte): Likewise.
7467 (neon_vget_lane<mode>_sext_internal): Likewise.
7468 (neon_vget_lane<mode>_zext_internal): Likewise.
7469 (neon_vdup_n<mode>): Likewise.
7470 (neon_vdup_nv2di): Likewise.
7471 (neon_vdpu_lane<mode>_internal): Likewise.
7472 (neon_vswp<mode>): Likewise.
7473 (float<mode><V_cvtto>2): Likewise.
7474 (floatuns<mode><V_cvtto>2): Likewise.
7475 (fix_trunc<mode><V_cvtto>)2): Likewise
7476 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7477 (neon_vcvt<mode>): Likewise.
7478 (neon_vcvtv4sfv4hf): Likewise.
7479 (neon_vcvtv4hfv4sf): Likewise.
7480 (neon_vcvt_n<mode>): Likewise.
7481 (neon_vmovn<mode>): Likewise.
7482 (neon_vqmovn<mode>): Likewise.
7483 (neon_vqmovun<mode>): Likewise.
7484 (neon_vmovl<mode>): Likewise.
7485 (neon_vmul_lane<mode>): Likewise.
7486 (neon_vmull_lane<mode>): Likewise.
7487 (neon_vqdmull_lane<mode>): Likewise.
7488 (neon_vqdmulh_lane<mode>): Likewise.
7489 (neon_vmla_lane<mode>): Likewise.
7490 (neon_vmlal_lane<mode>): Likewise.
7491 (neon_vqdmlal_lane<mode>): Likewise.
7492 (neon_vmls_lane<mode>): Likewise.
7493 (neon_vmlsl_lane<mode>): Likewise.
7494 (neon_vqdmlsl_lane<mode>): Likewise.
7495 (neon_vext<mode>): Likewise.
7496 (neon_vrev64<mode>): Likewise.
7497 (neon_vrev32<mode>): Likewise.
7498 (neon_vrev16<mode>): Likewise.
7499 (neon_vbsl<mode>_internal): Likewise.
7500 (neon_vshl<mode>): Likewise.
7501 (neon_vqshl<mode>): Likewise.
7502 (neon_vshr_n<mode>): Likewise.
7503 (neon_vshrn_n<mode>): Likewise.
7504 (neon_vqshrn_n<mode>): Likewise.
7505 (neon_vqshrun_n<mode>): Likewise.
7506 (neon_vshl_n<mode>): Likewise.
7507 (neon_vqshl_n<mode>): Likewise.
7508 (neon_vqshlu_n<mode>): Likewise.
7509 (neon_vshll_n<mode>): Likewise.
7510 (neon_vsra_n<mode>): Likewise.
7511 (neon_vsri_n<mode>): Likewise.
7512 (neon_vsli_n<mode>): Likewise.
7513 (neon_vtbl1v8qi): Likewise.
7514 (neon_vtbl2v8qi): Likewise.
7515 (neon_vtbl3v8qi): Likewise.
7516 (neon_vtbl4v8qi): Likewise.
7517 (neon_vtbx1v8qi): Likewise.
7518 (neon_vtbx2v8qi): Likewise.
7519 (neon_vtbx3v8qi): Likewise.
7520 (neon_vtbx4v8qi): Likewise.
7521 (neon_vtrn<mode>_internal): Likewise.
7522 (neon_vzip<mode>_internal): Likewise.
7523 (neon_vuzp<mode>_internal): Likewise.
7524 (neon_vld1<mode>): Likewise.
7525 (neon_vld1_lane<mode>): Likewise.
7526 (neon_vld1_dup<mode>): Likewise.
7527 (neon_vld1_dupv2di): Likewise.
7528 (neon_vst1<mode>): Likewise.
7529 (neon_vst1_lane<mode>): Likewise.
7530 (neon_vld2<mode>): Likewise.
7531 (neon_vld2_lane<mode>): Likewise.
7532 (neon_vld2_dup<mode>): Likewise.
7533 (neon_vst2<mode>): Likewise.
7534 (neon_vst2_lane<mode>): Likewise.
7535 (neon_vld3<mode>): Likewise.
7536 (neon_vld3qa<mode>): Likewise.
7537 (neon_vld3qb<mode>): Likewise.
7538 (neon_vld3_lane<mode>): Likewise.
7539 (neon_vld3_dup<mode>): Likewise.
7540 (neon_vst3<mode>): Likewise.
7541 (neon_vst3qa<mode>): Likewise.
7542 (neon_vst3qb<mode>): Likewise.
7543 (neon_vst3_lane<mode>): Likewise.
7544 (neon_vld4<mode>): Likewise.
7545 (neon_vld4qa<mode>): Likewise.
7546 (neon_vld4qb<mode>): Likewise.
7547 (neon_vld4_lane<mode>): Likewise.
7548 (neon_vld4_dup<mode>): Likewise.
7549 (neon_vst4<mode>): Likewise.
7550 (neon_vst4qa<mode>): Likewise.
7551 (neon_vst4qb<mode>): Likewise.
7552 (neon_vst4_lane<mode>): Likewise.
7553 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7554 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7555 (neon_vec_<US>mult_lo_<mode>): Likewise.
7556 (neon_vec_<US>mult_hi_<mode>): Likewise.
7557 (neon_vec_<US>shiftl_<mode>): Likewise.
7558 (neon_unpack<US>_<mode>): Likewise.
7559 (neon_vec_<US>mult_<mode>): Likewise.
7560 (vec_pack_trunc_<mode>): Likewise.
7561 (neon_vec_pack_trunk_<mode>): Likewise.
7562 (neon_vabd<mode>_2): Likewise.
7563 (neon_vabd<mode>_3): Likewise.
7564 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7565 (thumb2_movsi_vfp): Likewise.
7566 (movdi_vfp): Likewise.
7567 (movdi_vfp_cortexa8): Likewise.
7568 (movhf_vfp_neon): Likewise.
7569 (movhf_vfp): Likewiwse.
7570 (movsf_vfp): Likewiwse.
7571 (thumb2_movsf_vfp): Likewiwse.
7572 (movdf_vfp): Likewise.
7573 (thumb2_movdf_vfp): Likewise.
7574 (movsfcc_vfp): Likewise.
7575 (thumb2_movsfcc_vfp): Likewise.
7576 (movdfcc_vfp): Likewise.
7577 (thumb2_movdfcc_vfp): Likewise.
7578 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7579 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7580 (v10_v2c): Likewise.
7581 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7583 (cortex_a15_neon_int_2): Likewise.
7584 (cortex_a15_neon_int_3): Likewise.
7585 (cortex_a15_neon_int_4): Likewise.
7586 (cortex_a15_neon_int_5): Likewise.
7587 (cortex_a15_neon_vqneg_vqabs): Likewise.
7588 (cortex_a15_neon_vmov): Likewise.
7589 (cortex_a15_neon_vaba): Likewise.
7590 (cortex_a15_neon_vaba_qqq): Likewise.
7591 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7592 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7593 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7594 scalar_64_32_long_scalar): Likewise.
7595 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7596 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7597 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7598 lotype_qdd_64_32_long): Likewise.
7599 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7600 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7601 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7602 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7603 (cortex_a15_neon_shift_1): Likewise.
7604 (cortex_a15_neon_shift_2): Likewise.
7605 (cortex_a15_neon_shift_3): Likewise.
7606 (cortex_a15_neon_vshl_ddd): Likewise.
7607 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7608 (cortex_a15_neon_vsra_vrsra): Likewise.
7609 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7610 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7611 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7612 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7613 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7614 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7615 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7616 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7617 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7618 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7619 (cortex_a15_neon_bp_simple): Likewise.
7620 (cortex_a15_neon_bp_2cycle): Likewise.
7621 (cortex_a15_neon_bp_3cycle): Likewise.
7622 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7623 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7624 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7625 (cortex_a15_neon_vld2_4_regs): Likewise.
7626 (cortex_a15_neon_vld3_vld4): Likewise.
7627 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7628 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7629 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7630 (cortex_a15_neon_vst3_vst4): Likewise.
7631 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7632 (cortex_a15_neon_vld3_vld4_lane" 10
7633 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7634 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7635 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7636 (cortex_a15_neon_ldm_2): Likewise.0
7637 (cortex_a15_neon_stm_2): Likewise.
7638 (cortex_a15_neon_mcr): Likewise.
7639 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7640 (cortex_a15_neon_mrc): Likewise.
7641 (cortex_a15_neon_mrrc): Likewise.
7642 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7644 (cortex_a15_alu_shift): Likewise.
7645 (cortex_a15_alu_shift_reg): Likewise.
7646 (cortex_a15_mult32): Likewise.
7647 (cortex_a15_mult64): Likewise.
7648 (cortex_a15_block): Likewise.
7649 (cortex_a15_branch): Likewise.
7650 (cortex_a15_load1): Likewise.
7651 (cortex_a15_load3): Likewise.
7652 (cortex_a15_store1): Likewise.
7653 (cortex_a15_store3): Likewise.
7654 (cortex_a15_call): Likewise.
7655 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7656 (cortex_a5_f2r): Likewise.
7657 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7659 (cortex_a53_f2r): Likewise.
7660 * config/arm/cortex-a7.md
7661 (cortex_a7_branch): Update for attribute change.
7662 (cortex_a7_call): Likewise.
7663 (cortex_a7_alu_imm): Likewise.
7664 (cortex_a7_alu_reg): Likewise.
7665 (cortex_a7_alu_shift): Likewise.
7666 (cortex_a7_mul): Likewise.
7667 (cortex_a7_load1): Likewise.
7668 (cortex_a7_store1): Likewise.
7669 (cortex_a7_load2): Likewise.
7670 (cortex_a7_store2): Likewise.
7671 (cortex_a7_load3): Likewise.
7672 (cortex_a7_store3): Likewise.
7673 (cortex_a7_load4): Likewise.
7674 (cortex_a7_store4): Likewise.
7675 (cortex_a7_fpalu): Likewise.
7676 (cortex_a7_fconst): Likewise.
7677 (cortex_a7_fpmuls): Likewise.
7678 (cortex_a7_neon_mul): Likewise.
7679 (cortex_a7_fpmacs): Likewise.
7680 (cortex_a7_neon_mla: Likewise.
7681 (cortex_a7_fpmuld: Likewise.
7682 (cortex_a7_fpmacd: Likewise.
7683 (cortex_a7_fpfmad: Likewise.
7684 (cortex_a7_fdivs: Likewise.
7685 (cortex_a7_fdivd: Likewise.
7686 (cortex_a7_r2f: Likewise.
7687 (cortex_a7_f2r: Likewise.
7688 (cortex_a7_f_flags: Likewise.
7689 (cortex_a7_f_loads: Likewise.
7690 (cortex_a7_f_loadd: Likewise.
7691 (cortex_a7_f_stores: Likewise.
7692 (cortex_a7_f_stored: Likewise.
7693 (cortex_a7_neon): Likewise.
7694 * config/arm/cortex-a8-neon.md
7695 (cortex_a8_neon_mrc): Update for attribute change.
7696 (cortex_a8_neon_mrrc): Likewise.
7697 (cortex_a8_neon_int_1): Likewise.
7698 (cortex_a8_neon_int_2): Likewise.
7699 (cortex_a8_neon_int_3): Likewise.
7700 (cortex_a8_neon_int_4): Likewise.
7701 (cortex_a8_neon_int_5): Likewise.
7702 (cortex_a8_neon_vqneg_vqabs): Likewise.
7703 (cortex_a8_neon_vmov): Likewise.
7704 (cortex_a8_neon_vaba): Likewise.
7705 (cortex_a8_neon_vaba_qqq): Likewise.
7706 (cortex_a8_neon_vsma): Likewise.
7707 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7708 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7709 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7710 long_scalar): Likewise.
7711 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7712 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7713 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7714 long_scalar_qdd_64_32_long): Likewise.
7715 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7716 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7717 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7718 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7719 (cortex_a8_neon_shift_1): Likewise.
7720 (cortex_a8_neon_shift_2): Likewise.
7721 (cortex_a8_neon_shift_3): Likewise.
7722 (cortex_a8_neon_vshl_ddd): Likewise.
7723 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7724 (cortex_a8_neon_vsra_vrsra): Likewise.
7725 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7726 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7727 (cortex_a8_neon_fp_vsum): Likewise.
7728 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7729 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7730 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7731 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7732 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7733 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7734 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7735 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7736 (cortex_a8_neon_bp_simple): Likewise.
7737 (cortex_a8_neon_bp_2cycle): Likewise.
7738 (cortex_a8_neon_bp_3cycle): Likewise.
7739 (cortex_a8_neon_ldr): Likewise.
7740 (cortex_a8_neon_str): Likewise.
7741 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7742 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7743 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7744 (cortex_a8_neon_vld2_4_regs): Likewise.
7745 (cortex_a8_neon_vld3_vld4): Likewise.
7746 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7747 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7748 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7749 (cortex_a8_neon_vst3_vst4): Likewise.
7750 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7751 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7752 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7753 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7754 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7755 (cortex_a8_neon_mcr): Likewise.
7756 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7757 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7758 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7760 (ca9_neon_mrrc): Likewise.
7761 (cortex_a9_neon_int_1): Likewise.
7762 (cortex_a9_neon_int_2): Likewise.
7763 (cortex_a9_neon_int_3): Likewise.
7764 (cortex_a9_neon_int_4): Likewise.
7765 (cortex_a9_neon_int_5): Likewise.
7766 (cortex_a9_neon_vqneg_vqabs): Likewise.
7767 (cortex_a9_neon_vmov): Likewise.
7768 (cortex_a9_neon_vaba): Likewise.
7769 (cortex_a9_neon_vaba_qqq): Likewise.
7770 (cortex_a9_neon_vsma): Likewise.
7771 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7772 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7773 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7774 long_scalar): Likewise.
7775 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7776 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7777 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7778 long_scalar_qdd_64_32_long): Likewise.
7779 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7780 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7781 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7782 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7783 (cortex_a9_neon_shift_1): Likewise.
7784 (cortex_a9_neon_shift_2): Likewise.
7785 (cortex_a9_neon_shift_3): Likewise.
7786 (cortex_a9_neon_vshl_ddd): Likewise.
7787 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7788 (cortex_a9_neon_vsra_vrsra): Likewise.
7789 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7790 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7791 (cortex_a9_neon_fp_vsum): Likewise.
7792 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7793 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7794 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7795 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7796 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7797 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7798 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7799 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7800 (cortex_a9_neon_bp_simple): Likewise.
7801 (cortex_a9_neon_bp_2cycle): Likewise.
7802 (cortex_a9_neon_bp_3cycle): Likewise.
7803 (cortex_a9_neon_ldr): Likewise.
7804 (cortex_a9_neon_str): Likewise.
7805 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7806 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7807 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7808 (cortex_a9_neon_vld2_4_regs): Likewise.
7809 (cortex_a9_neon_vld3_vld4): Likewise.
7810 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7811 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7812 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7813 (cortex_a9_neon_vst3_vst4): Likewise.
7814 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7815 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7816 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7817 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7818 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7819 (cortex_a9_neon_mcr): Likewise.
7820 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7821 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7822 (cortex_a9_fps): Likewise.
7823 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7825 (cortex_m4_fmuls): Likewise.
7826 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7828 (cortex_r4_mrc): Likewise.
7829 * config/arm/iterators.md: Update comment referring to neon_type.
7830 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7831 (iwmmxt_movsi_insn): Likewise.
7832 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7834 (pj4_core_to_vfp): Likewise.
7835 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7837 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7838 (vfp_fstore): Likewise.
7839 * doc/md.texi: Change references to neon_type to refer to type.
7841 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7843 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7845 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7847 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7850 2013-09-04 Easwaran Raman <eraman@google.com>
7853 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7854 (build_and_add_sum): Use it.
7855 (appears_later_in_bb): Simplify code.
7857 2013-09-04 Teresa Johnson <tejohnson@google.com>
7859 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7861 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7863 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7865 2013-09-04 Jan Hubicka <jh@suse.cz>
7867 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7868 * ipa-devirt.c: Include diganostic.h
7869 (odr_type_d): Add types and types_set.
7870 (hash_type_name): Work for types with vtables during LTO.
7871 (odr_hasher::remove): Fix comment; destroy types_set.
7872 (add_type_duplicate): New function,
7873 (get_odr_type): Use it.
7874 (dump_type_inheritance_graph): Dump type duplicates.
7875 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7877 * tree.c (types_same_for_odr): Give exact answers on types with
7880 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7882 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7883 explaining their differences.
7885 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7887 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7889 2013-09-03 Jeff Law <law@redhat.com>
7891 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7892 when not threading through a joiner block. Pass joiner/no joiner
7893 state to register_jump_thread.
7894 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7895 state from argument rather than implying on path length.
7896 Dump the entire jump thread path into debugging dump.
7897 * tree-flow.h (register_jump_thread): Update prototype.
7899 2013-08-29 Xinliang David Li <davidxl@google.com>
7901 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7902 Remove a trivial gcc_assert.
7904 2013-08-29 Xinliang David Li <davidxl@google.com>
7906 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7907 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7908 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7909 Delay base decl alignment adjustment.
7910 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7911 * tree-vectorizer.h: New data structure.
7912 (set_dr_misalignment): New function.
7913 (dr_misalignment): Ditto.
7914 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7915 (vectorizable_load): Ditto.
7916 (ensure_base_align): New function.
7917 (vectorize_loops): Add dbg_cnt support.
7918 (execute_vect_slp): Ditto.
7919 * dbgcnt.def: New debug counter.
7920 * Makefile: New dependency.
7922 2013-09-03 Meador Inge <meadori@codesourcery.com>
7926 2013-08-30 Meador Inge <meadori@codesourcery.com>
7928 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7930 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7932 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7933 function descriptor.
7935 2013-09-03 Richard Biener <rguenther@suse.de>
7937 * tree-affine.c (add_elt_to_tree): Fix association issue,
7938 avoid useless converts and make sure to always return a
7939 properly typed result.
7941 2013-09-03 Richard Biener <rguenther@suse.de>
7944 * fold-const.c (negate_expr_p): Fix division case.
7945 (negate_expr): Likewise.
7947 2013-09-03 Richard Biener <rguenther@suse.de>
7950 * tree-streamer-out.c: Include tm.h.
7951 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7953 2013-09-03 Jan Hubicka <jh@suse.cz>
7955 * tree-profile.c (tree_profiling): Cleanup CFG when done.
7957 2013-09-03 Alan Modra <amodra@gmail.com>
7959 * config.gcc (powerpc*-*-linux*): Add support for little-endian
7960 multilibs to big-endian target and vice versa.
7961 * config/rs6000/t-linux64: Use := assignment on all vars.
7962 (MULTILIB_EXTRA_OPTS): Remove fPIC.
7963 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
7964 * config/rs6000/t-linux64le: New file.
7965 * config/rs6000/t-linux64bele: New file.
7966 * config/rs6000/t-linux64lebe: New file.
7968 2013-09-02 Jan Hubicka <jh@suse.cz>
7970 * ipa-inline-transform.c (inline_transform): Do not
7971 optimize_inline_calls when not optimizing.
7973 2013-09-02 Jan Hubicka <jh@suse.cz>
7975 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
7976 duplicated nodes for assembler names.
7977 * symtab.c (symtab_unregister_node): Do not attempt to unlink
7978 hard registers from assembler name hash.
7980 2013-09-02 Jan Hubicka <jh@suse.cz>
7982 * ipa-split.c (execute_split_functions): Split externally visible
7983 functions called once.
7985 2013-09-02 Martin Jambor <mjambor@suse.cz>
7988 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
7989 linked list. When finding the correct duplicate, also consider also
7990 the caller in additon to its inlined_to node.
7992 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
7994 * config/aarch64/aarch64-simd-builtins.def
7995 (dup_lane_scalar): Remove.
7996 * config/aarch64/aarch64-simd.md
7997 (aarch64_simd_dup): Add 'w->w' alternative.
7998 (aarch64_dup_lane<mode>): Allow for VALL.
7999 (aarch64_dup_lane_scalar<mode>): Remove.
8000 (aarch64_dup_lane_<vswap_width_name><mode>): New.
8001 (aarch64_get_lane_signed<mode>): Add w->w altenative.
8002 (aarch64_get_lane_unsigned<mode>): Likewise.
8003 (aarch64_get_lane<mode>): Likewise.
8004 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
8005 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
8006 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
8007 (VCON): Change container of V2SF.
8008 (vswap_width_name): Likewise.
8009 * config/aarch64/arm_neon.h
8010 (__aarch64_vdup_lane_any): New.
8011 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8012 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
8013 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8015 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
8018 * expr.c (emit_move_complex): Do not move complex FP values as parts if
8019 the source or the destination is a single hard register.
8021 2013-09-02 Richard Biener <rguenther@suse.de>
8024 * tree-scalar-evolution.c (instantiate_scev_name): Allow
8027 2013-09-02 Richard Biener <rguenther@suse.de>
8029 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
8030 arithmetic to sizetype.
8032 2013-09-02 Bin Cheng <bin.cheng@arm.com>
8034 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
8035 Find auto-increment use both before and after candidate.
8037 2013-09-02 Marek Polacek <polacek@redhat.com>
8039 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
8041 2013-09-01 Jan Hubicka <jh@suse.cz>
8043 * Makefile.in: Add ipa-profile.o
8044 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
8045 * cgraph.c (struct cgraph_propagate_frequency_data,
8046 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
8047 ipa-profile.c; replace cgraph_ by ipa_ prefix.
8048 * cgraph.h (cgraph_propagate_frequency): Remove.
8049 * ipa-inline-analysis.c: Include ipa-utils.h;
8050 drop duplicated cfgloop.h.
8051 (inline_update_callee_summaries): Update.
8052 * ipa-profile.c: New file.
8053 * ipa-utils.h (ipa_propagate_frequency): Declare.
8054 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
8055 data-streamer.h, value-prof.h.
8056 (symtab_remove_unreachable_nodes): Update profile.
8057 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
8058 account_time_size, cmp_counts, dump_histogram,
8059 ipa_profile_generate_summary, ipa_profile_write_summary,
8060 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
8061 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
8062 Move to ipa-profile.c.
8064 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
8066 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
8068 2013-09-01 Jan Hubicka <jh@suse.cz>
8070 * common.opt (fdevirtualize-speculatively): New function.
8071 * invoke.texi (fdevirtualize-speculatively): Document.
8072 * ipa-devirt.c: Include ipa-inline.h
8073 (likely_target_p): New function.
8074 (ipa_devirt): New function.
8075 (gate_ipa_devirt): New function.
8076 (pass_data_ipa_devirt): New static var.
8077 (pass_ipa_devirt): Likewise.
8078 (make_pass_ipa_devirt): New function.
8079 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
8080 (common_handle_option): Disable devirtualization when
8081 value range profiling is available.
8082 * passes.def (pass_ipa_devirt): Add.
8083 * timever.def (TV_IPA_DEVIRT): New timevar.
8084 * tree-pass.h (make_pass_ipa_devirt):
8086 2013-09-01 Iain Sandoe <iain@codesourcery.com>
8088 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
8089 include sanitize(undefined).
8091 2013-08-31 Diego Novillo <dnovillo@google.com>
8093 * Makefile.in (TREE_CORE_H): Define.
8095 (GTFILES): Add tree-core.h.
8096 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
8098 * tree-core.h: New file.
8099 Move all data structures, enum, typedefs, global
8100 declarations and constants from ...
8103 2013-08-31 Jan Hubicka <jh@suse.cz>
8105 * bulitins.c (expand_builtin): Do not early exit for gcov
8106 instrumented functions.
8108 2013-08-31 Marek Polacek <polacek@redhat.com>
8110 * ubsan.c: Include tm_p.h.
8112 2013-08-31 Jan Hubicka <jh@suse.cz>
8114 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
8117 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
8118 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
8119 * tree-cfg.c (verify_gimple_label): ... here.
8120 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
8121 (ipa_merge_profiles): New function.
8122 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
8123 (lto_input_function_body): Likewise.
8124 * ipa-utils.h (ipa_merge_profiles): Declare.
8125 * lto-streamer.h (lto_input_function_body): Update prototype.
8126 (emit_label_in_global_context_p): Remove.
8127 * lto-symtab.c: Include ipa-utils.h
8128 (lto_cgraph_replace_node): Use ipa_merge_profiles.
8130 2013-08-31 Jan Hubicka <jh@suse.cz>
8132 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8134 2013-08-31 Jan Hubicka <jh@suse.cz>
8136 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8138 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8140 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8141 "cmp" RTX before signed_comparison_operator check to account
8144 2013-08-30 Jan Hubicka <jh@suse.cz>
8146 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8147 (detect_type_change_1): Rename to ...
8148 (detect_type_change): ... this one; early return on non-polymorphic
8150 (detect_type_change_ssa): Add comp_type parameter; update
8151 use of detect_type_change.
8152 (compute_complex_assign_jump_func): Add param_type parameter;
8153 update use of detect_type_change_ssa.
8154 (compute_complex_ancestor_jump_func): Likewise.
8155 (ipa_get_callee_param_type): New function.
8156 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8157 update calls to the jump function computation functions.
8159 2013-08-30 Teresa Johnson <tejohnson@google.com>
8160 Steven Bosscher <steven@gcc.gnu.org>
8162 * cfgrtl.c (fixup_new_cold_bb): New routine.
8163 (commit_edge_insertions): Invoke fixup_partitions.
8164 (find_partition_fixes): New routine.
8165 (fixup_partitions): Ditto.
8166 (verify_hot_cold_block_grouping): Update comments.
8167 (rtl_verify_edges): Invoke find_partition_fixes.
8168 (rtl_verify_bb_pointers): Update comments.
8169 (rtl_verify_bb_layout): Ditto.
8170 * basic-block.h (probably_never_executed_edge_p): Declare.
8171 (fixup_partitions): Ditto.
8172 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8173 * bb-reorder.c (sanitize_hot_paths): New function.
8174 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8176 * predict.c (probably_never_executed_edge_p): New routine.
8177 * cfg.c (check_bb_profile): Add partition insanity warnings.
8179 2013-08-30 Meador Inge <meadori@codesourcery.com>
8181 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8183 2013-08-30 Marek Polacek <polacek@redhat.com>
8185 * Makefile.in (ubsan.o): Add.
8186 (c-family/c-ubsan.o): Add.
8187 (builtins.o): Add ubsan.h dependency.
8188 * ubsan.h: New file.
8189 * ubsan.c: New file.
8190 * common.opt: Add -fsanitize=undefined option.
8191 (flag_sanitize): Add variable.
8192 (fsanitize=): Add option. Add Driver.
8193 (fsanitize=thread): Remove option.
8194 (fsanitize=address): Likewise.
8195 (static-libubsan): New option.
8196 * doc/invoke.texi: Document the new flag and -static-libubsan.
8197 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8198 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8199 * builtin-attrs.def (ATTR_COLD): Define.
8200 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8201 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8202 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8203 * flag-types.h (sanitize_code): New enum.
8204 * opts.c (common_handle_option): Parse command line arguments
8205 of -fsanitize=. Add -fsanitize=unreachable option.
8206 * varasm.c (get_variable_section): Adjust.
8207 (assemble_noswitch_variable): Likewise.
8208 (assemble_variable): Likewise.
8209 (output_constant_def_contents): Likewise.
8210 (categorize_decl_for_section): Likewise.
8211 (place_block_symbol): Likewise.
8212 (output_object_block): Likewise.
8213 * builtins.def: Likewise.
8214 * toplev.c (compile_file): Likewise.
8215 (process_options): Likewise.
8216 * cppbuiltin.c: Likewise.
8217 * tsan.c (tsan_pass): Likewise.
8218 (tsan_gate): Likewise.
8219 (tsan_gate_O0): Likewise.
8220 * cfgexpand.c (partition_stack_vars): Likewise.
8221 (expand_stack_vars): Likewise.
8222 (defer_stack_allocation): Likewise.
8223 (expand_used_vars): Likewise.
8224 * cfgcleanup.c (old_insns_match_p): Likewise.
8225 * asan.c (asan_finish_file): Likewise.
8226 (asan_instrument): Likewise.
8227 (gate_asan): Likewise.
8228 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8229 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8230 (asan_global_struct): Use pointer_sized_int_node instead
8231 calling build_nonstandard_integer_type.
8232 (initialize_sanitizer_builtins): Likewise.
8233 (asan_finish_file): Likewise.
8234 * gcc.c: Document %{%:function(args):X}.
8235 (static_spec_functions): Add sanitize.
8236 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8237 store funcval != NULL there.
8238 (do_spec_1): Adjust handle_spec_function caller.
8239 (handle_braces): Allow %:function(args) as condition.
8240 (sanitize_spec_function): New function.
8241 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8242 (LIBUBSAN_SPEC): Likewise.
8243 (LIBUBSAN_EARLY_SPEC): Likewise.
8244 (SANITIZER_SPEC): Handle libubsan.
8245 (SANITIZER_EARLY_SPEC): Likewise.
8246 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8247 instead of fsanitize=address.
8248 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8249 instead of fsanitize=address*.
8250 * builtins.c: Include ubsan.h.
8251 (fold_builtin_0): Instrument __builtin_unreachable.
8252 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8253 instead of flag_asan.
8254 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8255 (pointer_sized_int_node): Define.
8256 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8258 2013-08-30 Mike Stump <mikestump@comcast.net>
8260 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8263 2013-08-29 Jan Hubicka <jh@suse.cz>
8265 * cgraph.c (cgraph_function_body_availability): Handle weakref
8267 * passes.def: Remove pass_fixup_cfg.
8268 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8269 track when we need to remove functions.
8270 (gate_ipa_inline): Execute inlining always; add comment why.
8271 (pass_data_ipa_inline): Remove TODO_remove_functions.
8272 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8273 do not produce summaries.
8274 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8275 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8276 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8277 constant pool and vtable.
8279 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8281 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8282 New arm_neon.h's internal macros to specify 64-bit constants.
8283 Avoid using stdint.h's macros.
8285 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8287 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8288 before applying REG_P.
8290 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8292 PR tree-optimization/58277
8293 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8294 after seeing too many stmts with vdef in between dombb and current
8295 bb, invalidate everything.
8297 2013-08-30 Richard Biener <rguenther@suse.de>
8299 * fold-const.c (fold_single_bit_test): Fix overflow test.
8301 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8303 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8304 and which can live in a register, always retrieve the value on entry.
8305 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8306 passed by invisible reference specially.
8307 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8308 (vt_add_function_parameter): Correctly deal with a parameter passed by
8309 invisible reference.
8311 2013-08-30 Jan Hubicka <jh@suse.cz>
8313 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8315 2013-08-30 Richard Biener <rguenther@suse.de>
8317 PR tree-optimization/58228
8318 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8319 allow invariant loads in nested loop vectorization.
8321 2013-08-30 Richard Biener <rguenther@suse.de>
8323 PR tree-optimization/58223
8324 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8325 (has_anti_or_output_dependence): ... this and adjust to also
8326 look for output dependences.
8327 (mark_nodes_having_upstream_mem_writes): Adjust.
8328 (rdg_flag_uses): Likewise.
8330 2013-08-30 Richard Biener <rguenther@suse.de>
8332 PR tree-optimization/58010
8333 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8334 assert that we have a loop-closed PHI.
8336 2013-08-29 Jan Hubicka <jh@suse.cz>
8338 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8339 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8340 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8341 (lto_free_function_in_decl_state_for_node): New function.
8343 2013-08-29 Xinliang David Li <davidxl@google.com>
8345 * loop-unroll.c (report_unroll_peel): Minor message change.
8346 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8347 Emit alignment peeling message with default -fopt-info.
8348 (vect_loop_versioning): Emit loop version info message.
8349 * tree-vectorizer.c (vectorize_loops): Minor message change.
8350 (execute_vect_slp): Ditto.
8352 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8354 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8355 of the clone from the DECL_NAME of the original function.
8357 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8359 * passes.c (register_pass): Add overload.
8360 * tree-pass.h (register_pass): Forward declare it. Add comment.
8362 2013-08-29 Jan Hubicka <jh@suse.cz>
8364 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8365 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8366 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8367 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8368 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8369 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8370 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8371 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8373 2013-08-29 Teresa Johnson <tejohnson@google.com>
8375 * dumpfile.c (dump_loc): Output column number.
8376 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8377 * doc/invoke.texi: Document optall -fopt-info flag.
8378 * profile.c (read_profile_edge_counts): Use new dump framework.
8379 (compute_branch_probabilities): Ditto.
8380 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8381 when pass not in any opt group.
8382 * pass_manager.h (pass_manager::get_pass_profile): New method.
8383 * value-prof.c (check_counter): Use new dump framework.
8384 (check_ic_target): Ditto.
8385 * coverage.c (get_coverage_counts): Ditto.
8386 (coverage_init): Setup new dump framework.
8388 2013-08-29 Richard Biener <rguenther@suse.de>
8390 PR tree-optimization/58246
8391 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8392 handle the dominance check inside a basic-block.
8394 2013-08-29 Richard Biener <rguenther@suse.de>
8397 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8398 of default defs that appear in abnormal PHI nodes.
8400 2013-08-29 Richard Biener <rguenther@suse.de>
8402 PR tree-optimization/57685
8403 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8404 single-use operands to avoid exponential complexity.
8406 2013-08-28 Dehao Chen <dehao@google.com>
8408 * ipa-inline.c (edge_badness): Fix integer underflow.
8410 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8412 * gtm-builtins.def (_ITM_free): Declare leaf.
8414 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8417 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8418 (*tls_local_dynamic_base_64_largepic): Likewise.
8419 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8420 Remove predicate from call operand.
8421 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8422 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8424 2013-08-28 Jeff Law <law@redhat.com>
8426 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8427 checks for the number of predecessors and successors allowed.
8428 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8429 which require copying a joiner block if there is a request which
8430 is a subpath that requires no joiner block copying.
8432 2013-08-28 Jan Hubicka <jh@suse.cz>
8434 * lto-streamer-out.c (DFS_write_tree_body): Drop
8435 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8436 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8437 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8438 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8439 Do not read DECL_ERROR_ISSUED.
8440 (unpack_ts_decl_with_vis_value_fields): Do not read
8442 (lto_input_ts_binfo_tree_pointers): Do not read
8443 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8444 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8445 write DECL_ERROR_ISSUED..
8446 (pack_ts_decl_with_vis_value_fields): Do not write
8448 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8449 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8450 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8451 * tree.h (tree_decl_common): Update comment.
8452 (DECL_ERROR_ISSUED): Remove.
8454 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8457 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8459 2013-08-28 Jan Hubicka <jh@suse.cz>
8461 * builtins.def (free): Declare leaf.
8463 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8466 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8467 * configure: Regenerate.
8469 2013-08-27 Martin Jambor <mjambor@suse.cz>
8471 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8472 (ipa_ancestor_jf_data): Likewise.
8473 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8474 (ipa_get_jf_pass_through_type_preserved): New function.
8475 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8476 (ipa_get_jf_ancestor_type_preserved): New function.
8477 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8478 (ipa_get_jf_ancestor_result): Likewise.
8479 (propagate_vals_accross_pass_through): Use
8480 ipa_get_jf_pass_through_result to do all the value mappings.
8481 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8482 type_preserved flag.
8483 (ipa_set_jf_cst_copy): New function.
8484 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8485 (ipa_set_jf_arith_pass_through): Likewise.
8486 (ipa_set_ancestor_jf): Likewise.
8487 (compute_complex_assign_jump_func): Set type_preserved instead of
8489 (ipa_compute_jump_functions_for_edge): Likewise.
8490 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8491 (update_jump_functions_after_inlining): Update type_preserved.
8492 Explicitely create jump functions when combining one with pass_through.
8493 (ipa_write_jump_function): Stream the type_preserved flags.
8494 (ipa_read_jump_function): Likewise.
8496 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8497 Aldy Hernandez <aldyh@redhat.com>
8499 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8500 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8501 * function.h (struct function): Add has_force_vect_loops and
8503 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8504 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8506 * gimple.def: Update CLAUSES comments.
8507 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8508 (gimple_build_omp_for): Add argument to prototype.
8509 (gimple_omp_for_kind): New.
8510 (gimple_omp_for_set_kind): New.
8511 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8512 GOVD_DATA_SHARE_CLASS.
8513 (enum omp_region_type): Add ORT_SIMD.
8514 (gimple_add_tmp_var): Handle ORT_SIMD.
8515 (gimplify_var_or_parm_decl): Same.
8516 (is_gimple_stmt): Same.
8517 (omp_firstprivatize_variable): Same.
8518 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8519 (omp_notice_variable): Handle ORT_SIMD.
8520 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8521 (omp_check_private): Handle ORT_SIMD.
8522 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8524 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8525 Handle OMP_CLAUSE_LASTPRIVATE.
8526 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8528 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8529 (gimplify_expr): Handle OMP_SIMD.
8530 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8531 (expand_GOMP_SIMD_VF): New.
8532 (expand_GOMP_SIMD_LAST_LANE): New.
8533 * internal-fn.def (GOMP_SIMD_LANE): New.
8534 (GOMP_SIMD_VF): New.
8535 (GOMP_SIMD_LAST_LANE): New.
8536 * omp-low.c: Include target.h.
8537 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8539 (check_omp_nesting_restrictions): Same.
8541 (lower_rec_simd_input_clauses): New.
8542 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8544 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8545 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8546 (expand_omp_build_assign): New.
8547 (expand_omp_for_init_counts): New.
8548 (expand_omp_for_init_vars): New.
8549 (extract_omp_for_update_vars): New.
8550 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8551 and rewrite accordingly.
8552 (expand_omp_simd): New.
8553 (expand_omp_for): Use expand_omp_simd.
8554 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8555 (lower_omp_for): Do not lower the body.
8556 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8558 * tree-flow.h (find_omp_clause): Remove prototype.
8559 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8560 forcing vectorization of the loop, or if flag_tree_vectorize.
8561 (gate_tree_if_conversion): Similarly.
8562 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8563 gimple_build_omp_for.
8564 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8565 * tree-parloops (create_parallel_loop): Pass kind argument to
8566 gimple_build_omp_for.
8567 * tree-pretty-print.c (dump_omp_clause): Add cases for
8568 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8569 OMP_CLAUSE__SIMDUID_.
8570 (dump_generic_node): Handle OMP_SIMD.
8571 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8572 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8573 unroll OMP_SIMD loops here.
8574 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8575 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8577 (vect_analyze_data_refs): Handle simd loops.
8578 * tree-vect-loop.c (vectorizable_live_operation): Handle
8580 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8581 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8582 (vectorizable_load): Same.
8583 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8584 (struct simduid_to_vf): New.
8585 (simduid_to_vf::hash): New.
8586 (simduid_to-vf::equal): New.
8587 (struct simd_array_to_simduid): New.
8588 (simd_array_to_simduid::hash): New.
8589 (simd_array_to_simduid::equal): New.
8590 (adjust_simduid_builtins): New.
8591 (struct note_simd_array_uses_struct): New.
8592 (note_simd_array_uses_cb): New.
8593 (note_simd_array_uses): New.
8594 (vectorize_loops): Handle simd hints and adjust simd builtins
8596 * tree-vectorizer.h (struct _stmt_vec_info): Add
8597 simd_lane_access_p field.
8598 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8599 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8600 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8601 (omp_clause_code_name): Same.
8602 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8603 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8604 * tree.def (OMP_SIMD): New entry.
8605 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8606 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8607 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8608 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8609 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8610 (OMP_CLAUSE_LINEAR_STEP): New.
8611 (OMP_CLAUSE_SAFELEN_EXPR): New.
8612 (OMP_CLAUSE__SIMDUID__DECL): New.
8613 (find_omp_clause): New prototype.
8615 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8617 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8618 Haswell processor detection.
8620 2013-08-27 Christian Widmer <shadow@umbrox.de>
8623 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8624 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8625 AVX2 capable processors.
8627 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8629 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8630 of vget_low_* with implementations in terms of other intrinsics.
8632 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8635 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8636 values to -1, 0 and 1.
8638 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8640 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8641 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8642 (<optab><mode>3_insn): Remove.
8643 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8644 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8645 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8646 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8647 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8648 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8649 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8650 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8651 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8652 (ror<mode>3_insn): Likewise.
8653 * config/aarch64/predicates.md (aarch64_simd_register): New.
8655 2013-08-27 Richard Biener <rguenther@suse.de>
8657 PR tree-optimization/57521
8658 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8659 one edge is non-critical.
8660 (find_phi_replacement_condition): Make sure to use a non-critical
8661 edge. Cleanup and remove old bug workarounds.
8662 (bb_postdominates_preds): Remove.
8663 (if_convertible_loop_p_1): Do not compute post-dominators.
8664 (combine_blocks): Do not free post-dominators.
8665 (main_tree_if_conversion): Likewise.
8666 (pass_data_if_conversion): Add TODO_verify_ssa.
8668 2013-08-27 DJ Delorie <dj@redhat.com>
8670 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8672 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8674 * function.c (assign_parm_find_data_types): Set passed_mode and
8675 nominal_mode to the TYPE_MODE of nominal_type for the built
8676 pointer type in case of the struct-pass-by-reference.
8678 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8680 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8681 (UINT16_TYPE): Change default to "unsigned int".
8683 * config/avr/avr.opt (mfract-convert-truncate): New option.
8684 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8685 is set, round negative fractional integers according to n1169
8686 when converting to integer types.
8688 2013-08-26 Jan Hubicka <jh@suse.cz>
8690 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8691 methods can not be called indirectly when their address is not taken.
8693 2013-08-26 Jan Hubicka <jh@suse.cz>
8695 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8698 2013-08-26 Jan Hubicka <jh@suse.cz>
8700 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8703 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8705 * reload.c (find_valid_class): Allow classes that do not include
8706 FIRST_PSEUDO_REGISTER - 1.
8708 2013-08-26 Jan Hubicka <jh@suse.cz>
8710 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8711 fix edge count/frequency when speculation failed; fix type check
8712 for the direct call.
8714 2013-08-26 Jan Hubicka <jh@suse.cz>
8716 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8718 2013-08-26 Jan Hubicka <jh@suse.cz>
8720 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8721 to be changed by edge redirection.
8723 2013-08-26 Jan Hubicka <jh@suse.cz>
8725 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8726 formating; add sanity check.
8727 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8728 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8729 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8730 (resolve_noninline_speculation): Update callee keys, too.
8732 2013-08-26 Jan Hubicka <jh@suse.cz>
8734 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8735 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8737 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8739 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8742 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8744 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8745 (x86_64_elf_select_section): Ditto.
8746 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8748 (ix86_pass_by_reference): Ditto.
8749 (output_set_got): Ditto.
8750 (ix86_unary_operator_ok): Ditto.
8751 (ix86_expand_builtin): Ditto.
8753 2013-08-23 Jan Hubicka <jh@suse.cz>
8755 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8757 2013-08-23 Jan Hubicka <jh@suse.cz>
8759 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8760 (tree_decl_with_vis): Add FINAL field.
8762 2013-08-23 Jeff Law <law@redhat.com>
8764 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8765 the debugging dump when the expression is fully redundant.
8767 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8769 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8770 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8771 * pretty-print.c (pp_formatted_text_data): Likewise.
8772 (pp_write_text_to_stream): Likewise.
8773 (pp_write_text_as_dot_label_to_stream): Likewise.
8774 (pp_append_r): Likewise.
8775 (pp_format): Likewise.
8776 (pp_flush): Likewise.
8777 (pp_clear_output_area): Likewise.
8778 (pp_append_text): Likewise.
8779 (pp_formatted_text): Likewise.
8780 (pp_remaining_character_count_for_line): Likewise.
8781 (pp_newline): Likewise.
8782 (pp_character): Likewise.
8783 (output_buffer::~output_buffer): Define.
8784 (pretty_printer::~pretty_printer): Destruct output buffer.
8785 * pretty-print.h (output_buffer::~output_buffer): Declare.
8786 (pretty_printer::~pretty_printer): Declare virtual.
8788 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8791 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8792 HOST_WIDE_INT_M1U): New macros.
8793 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8794 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8795 unsigned -1 for lshift.
8796 * cse.c (cse_insn): Likewise.
8797 * double-int.c (rshift_double, lshift_double): Likewise.
8798 * builtins.c (fold_builtin_bitop): Likewise.
8799 * combine.c (force_to_mode): Likewise.
8800 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8801 * simplify-rtx.c (simplify_const_unary_operation,
8802 simplify_const_binary_operation): Likewise.
8803 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8804 check_va_list_escapes): Likewise.
8805 * rtlanal.c (nonzero_bits1): Likewise.
8806 * expmed.c (expand_smod_pow2): Likewise.
8807 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8809 2013-08-23 Jan Hubicka <jh@suse.cz>
8811 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8812 as having address taken.
8814 2013-08-23 Jan Hubicka <jh@suse.cz>
8816 * ipa-utils.h (method_class_type): Declare.
8817 * ipa-devirt.c (method_class_type): Export.
8819 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8820 do not walk base classes of anonymous types.
8822 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8824 PR rtl-optimization/58220
8826 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8827 handle SEQUENCE insns properly.
8829 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8831 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8833 (pp_newline_and_indent): Likewise.
8834 (pp_separate_with): Likewise.
8835 * pretty-print.c (pp_newline_and_flush): Define.
8836 (pp_newline_and_indent): Likewise.
8837 (pp_separate_with): Likewise.
8839 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8842 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8843 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8845 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8847 * config/i386/predicates.md (ext_sse_reg_operand): New.
8848 * config/i386/i386.md (*movti_internal): Use
8849 predicate to determine if EVEX is needed.
8850 (*movsi_internal): Ditto.
8851 (*movdf_internal): Ditto.
8852 (*movsf_internal): Ditto.
8853 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8855 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8857 PR tree-optimization/58209
8858 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8859 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8860 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8861 emit POINTER_PLUS_EXPR.
8862 (create_tailcall_accumulator): For pointer result type accumulate in
8865 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8867 * configure.ac: Add backslashes missing from the last change.
8868 * configure: Regenerate.
8870 2013-08-22 Jan Hubicka <jh@susue.cz>
8872 * ipa.c (function_and_variable_visibility): First remember function
8873 was global and then make it local.
8875 2013-08-22 Julian Brown <julian@codesourcery.com>
8877 * configure.ac: Add aarch64 to list of arches which use "nop" in
8879 * configure: Regenerate.
8881 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8883 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8884 gnu_libc_has_function.
8885 * config/s390/tpf.h: Likewise.
8887 2013-08-22 Jan Hubicka <jh@susue.cz>
8889 * timevar.c (validate_phases): Add cast.
8891 2013-08-22 Jan Hubicka <jh@susue.cz>
8893 * timevar.c (validate_phases): Use size_t for memory.
8894 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8896 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8898 * pretty-print.h (output_buffer::output_buffer): Declare.
8899 (pretty_printer::pretty_printer): Likewise.
8900 (pp_construct): Remove.
8901 * pretty-print.c (output_buffer::output_buffer): Define.
8902 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8903 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8904 (print_gimple_expr): Likewise.
8905 (print_gimple_seq): Likewise.
8906 (gimple_dump_bb): Likewise.
8907 * sched-vis.c (dump_value_slim): Likewise.
8908 (dump_insn_slim): Likewise.
8909 (dump_rtl_slim): Likewise.
8910 (str_pattern_slim): Likewise.
8911 * tree-mudflap.c (mf_varname_tree): Likewise.
8912 * graph.c (print_graph_cfg): Likewise.
8913 (start_graph_dump): Likewise.
8914 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8916 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8917 pretty printer initialization.
8918 * coretypes.h (diagnostic_context): Remove superflous type alias
8920 (pretty_printer): Likewise. Declare directly as a class.
8921 (pretty_print_info): Remove declaration as class.
8922 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8923 and pp_clear_output_area.
8924 (asan_add_global): Likewise.
8926 2013-08-22 Jan Hubicka <jh@suse.cz>
8928 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8929 * ipa-utils.h (update_type_inheritance_graph): Declare.
8930 (possible_polymorphic_call_target_p): Declare.
8931 (possible_polymorphic_call_target_p): New.
8932 * ipa-devirt.c: Update toplevel comments.
8933 (cached_polymorphic_call_targets): Move up.
8934 (odr_type_d): Move ID down.
8935 (polymorphic_type_binfo_p): Update comment.
8936 (odr_hasher::remove): Likewise;
8937 (get_odr_type): Set anonymous_namespace.
8938 (dump_odr_type): Dump it.
8939 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8940 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8941 (record_binfo): Add comment.
8942 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8944 (devirt_node_removal_hook): Do not iCE when cache is freed.
8945 (possible_polymorphic_call_target_p): New predicate.
8946 (update_type_inheritance_graph): New function.
8948 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8949 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8950 Sergey Lega <sergey.s.lega@intel.com>
8951 Anna Tikhonova <anna.tikhonova@intel.com>
8952 Ilya Tocar <ilya.tocar@intel.com>
8953 Andrey Turetskiy <andrey.turetskiy@intel.com>
8954 Ilya Verbin <ilya.verbin@intel.com>
8955 Kirill Yukhin <kirill.yukhin@intel.com>
8956 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8958 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
8959 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
8960 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
8961 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
8962 (OPTION_MASK_ISA_AVX2_UNSET): Update.
8963 (OPTION_MASK_ISA_AVX512F_UNSET): New.
8964 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
8965 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
8966 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
8967 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
8968 OPT_mavx512pf, OPT_mavx512er cases.
8969 * config/i386/constraints.md (v): New constraint.
8970 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
8971 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
8972 (bit_AVX512CD): New.
8973 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
8974 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
8975 * config/i386/i386-c.c (ix86_target_macros_internal):
8976 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
8978 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
8979 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
8980 * config/i386/i386.c (regclass_map, dbx_register_map)
8981 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
8982 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
8983 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
8985 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
8986 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
8987 -mavx512pf options. Fix formatting.
8988 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
8989 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
8990 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
8991 -mavx512cd, -mavx512pf options.
8992 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
8993 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
8994 (ix86_preferred_output_reload_class): Replace SSE_REGS with
8996 (ix86_hard_regno_mode_ok): Support 512-bit registers.
8997 (ix86_set_reg_reg_cost): Ditto.
8998 (x86_order_regs_for_local_alloc): Ditto.
8999 (MAX_VECT_LEN): Extend to 64-byte.
9000 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
9001 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
9002 (TARGET_AVX512ER, TARGET_AVX512CD): New.
9003 (BIGGEST_ALIGNMENT): Extend to 512-bits.
9004 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
9005 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
9006 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
9007 (SSE_REG_MODE_P): Support new modes.
9008 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
9009 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
9010 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
9011 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
9012 (REG_CLASS_CONTENTS): Add new registers.
9013 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
9014 (EXT_REX_SSE_REGNO_P): New.
9015 (HI_REGISTER_NAMES): Add new registers.
9016 * config/i386/i386.md: Define constants for new registers.
9017 (mode): Add new 512-bit modes.
9018 (prefix): Support evex prefix.
9019 (isa): Support avx512f, noavx512f, fma_avx512f.
9020 (ssemodesuffix): Add new 512-bit modes.
9022 (*movxi_internal_avx512f): Ditto.
9023 (*movdi_internal): Replace constraint "x" with the new constraint "v".
9025 (*movsi_internal): Likewise.
9026 (*movdf_internal): Likewise.
9027 (*movsf_internal): Likewise.
9028 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
9029 (<code><mode>3): Likewise.
9030 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
9032 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
9033 with the new constraint "v".
9034 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
9036 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
9037 with the new constraint "v".
9038 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
9039 (<sse2>_storedqu<avxsizesuffix>): Likewise.
9040 (*<plusminus_insn><mode>3): Likewise.
9041 (<sse>_vm<plusminus_insn><mode>3): Likewise.
9042 (*mul<mode>3): Likewise.
9043 (<sse>_vmmul<mode>3): Likewise.
9044 (<sse>_div<mode>3): Likewise.
9045 (<sse>_vmdiv<mode>3): Likewise.
9046 (<sse>_sqrt<mode>2): Likewise.
9047 (<sse>_vmsqrt<mode>2): Likewise.
9048 (*<code><mode>3_finite): Likewise.
9049 (*<code><mode>3) <smaxmin>: Likewise.
9050 (<sse>_vm<code><mode>3): Likewise.
9051 (*<code><mode>3) <any_logic>: Likewise.
9052 (*fma_fmadd_<mode>): Likewise.
9053 (*fma_fmsub_<mode>): Likewise.
9054 (*fma_fnmadd_<mode>): Likewise.
9055 (*fma_fnmsub_<mode>): Likewise.
9056 (*fma_fmaddsub_<mode>): Likewise.
9057 (*fma_fmsubadd_<mode>): Likewise.
9058 (*fmai_fmadd_<mode>): Likewise.
9059 (*fmai_fmsub_<mode>): Likewise.
9060 (*fmai_fnmadd_<mode>): Likewise.
9061 (*fmai_fnmsub_<mode>): Likewise.
9062 (sse_cvtsi2ss): Likewise.
9063 (sse_cvtsi2ssq): Likewise.
9064 (sse_cvtss2si): Likewise.
9065 (sse_cvtss2si_2): Likewise.
9066 (sse_cvtss2siq): Likewise.
9067 (sse_cvtss2siq_2): Likewise.
9068 (sse_cvttss2si): Likewise.
9069 (sse_cvtss2siq_2): Likewise.
9070 (float<sseintvecmodelower><mode>2): Likewise.
9071 (sse2_cvtsd2si_2): Likewise.
9072 (sse2_cvtsd2siq_2): Likewise.
9073 (*<plusminus_insn><mode>3): Likewise.
9074 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
9075 (*<sse4_1_avx2>_mul<mode>3): Likewise.
9076 (ashr<mode>3): Likewise.
9077 (<shift_insn><mode>3): Likewise.
9078 (avx2_<code><mode>3): Likewise.
9079 (*avx2_<code><mode>3): Likewise.
9080 (*andnot<mode>3): Likewise.
9081 (*<code><mode>3) <any_logic>: Likewise.
9082 (abs<mode>2): Likewise.
9083 (avx2_permvar<mode>): Likewise.
9084 (avx2_perm<mode>_1): Likewise.
9085 (*avx_vpermilp<mode>): Likewise.
9086 (avx_vpermilvar<mode>3): Likewise.
9087 (avx2_ashrv<mode>): Likewise.
9088 (avx2_<shift_insn>v<mode>): Likewise.
9089 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
9091 * doc/rtl.texi: Document XImode.
9093 2013-08-21 Jeff Law <law@redhat.com>
9095 * tree-flow.h (register_jump_thread): Pass vector of edges
9096 instead of each important edge.
9097 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
9098 thread path into a vector and pass that to register_jump_thread.
9099 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
9100 passed in edge vector to the current 3-edge form.
9103 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9105 * dce.c (fini_dce): Call df_analyze again just in case
9106 delete_unmarked_insns removed anything.
9108 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
9110 * reload.h (struct reg_equivs): Rename to ..
9111 (struct reg_equivs_s): .. this.
9113 2013-08-20 Martin Liska <marxin.liska@gmail.com>
9115 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
9117 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9119 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
9121 2013-08-21 Jeff Law <law@redhat.com>
9123 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
9124 simplify assignments too. If the RHS collapses to a singleton
9125 range, then return the value for the range.
9127 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
9129 * config/i386/sse.md (V16): Rename to...
9131 (mov<mode>): Update iterator name.
9132 (*mov<mode>_internal): Ditto.
9133 (push<mode>1): Ditto.
9134 (movmisalign<mode>): Ditto.
9136 2013-08-20 Jan Hubicka <jh@suse.cz>
9139 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9140 entry for direct edges.
9141 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9143 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9145 Revert my last two changes, r201865 and r201864:
9148 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9150 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9151 instances can own GC refs.
9153 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9154 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9155 (gcc::context::gt_pch_nx): Likewise.
9156 (gcc::context::gt_pch_nx): Likewise.
9157 * ggc.h (gt_ggc_mx <T>): New.
9158 (gt_pch_nx_with_op <T>): New.
9159 (gt_pch_nx <T>): New.
9160 * passes.c (opt_pass::gt_ggc_mx): New.
9161 (opt_pass::gt_pch_nx): New.
9162 (opt_pass::gt_pch_nx_with_op): New.
9163 (pass_manager::gt_ggc_mx): New.
9164 (pass_manager::gt_pch_nx): New.
9165 (pass_manager::gt_pch_nx_with_op): New.
9166 (pass_manager::operator new): Use
9167 ggc_internal_cleared_alloc_stat rather than xcalloc.
9168 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9169 (pass_manager::gt_ggc_mx): New.
9170 (pass_manager::gt_pch_nx): New.
9171 (pass_manager::gt_pch_nx_with_op): New.
9172 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9173 (opt_pass::operator new): New.
9174 (opt_pass::gt_ggc_mx): New.
9175 (opt_pass::gt_pch_nx): New.
9176 (opt_pass::gt_pch_nx_with_op): New.
9179 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9181 * Makefile.in (GTFILES): Add context.h.
9182 * context.c (gcc::context::operator new): New.
9183 (gcc::context::gt_ggc_mx): New.
9184 (gcc::context::gt_pch_nx): New.
9185 (gcc::context::gt_pch_nx): New.
9186 * context.h (gcc::context): Add GTY((user)) marking.
9187 (gcc::context::operator new): New.
9188 (gcc::context::gt_ggc_mx): New.
9189 (gcc::context::gt_pch_nx): New.
9190 (gcc::context::gt_pch_nx): New.
9191 (g): Add GTY marking.
9192 (gt_ggc_mx (gcc::context *)): New.
9193 (gt_pch_nx (gcc::context *)): New.
9194 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9195 void *cookie)): New.
9196 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9198 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9200 * dce.c (fini_dce): Call df_analyze again just in case
9201 delete_unmarked_insns removed anything.
9203 2013-08-20 Teresa Johnson <tejohnson@google.com>
9205 PR rtl-optimizations/57451
9206 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9207 from crossing split section boundaries.
9209 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9211 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9212 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9213 with MULTLIB_DEFAULTS.
9215 2013-08-20 Nick Clifton <nickc@redhat.com>
9217 * target.def (narrow_volatile_bitfield): Note that the default
9218 value is false, not !TARGET_STRICT_ALIGN.
9219 * doc/tm.texi: Regenerate.
9221 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9223 Fix LIB_SPEC for systems without libpthread.
9225 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9226 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9228 * config/i386/linux-common.h: Likewise.
9229 * config/mips/linux-common.h: Likewise.
9231 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9233 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9236 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9238 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9239 instances can own GC refs.
9241 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9242 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9243 (gcc::context::gt_pch_nx): Likewise.
9244 (gcc::context::gt_pch_nx): Likewise.
9245 * ggc.h (gt_ggc_mx <T>): New.
9246 (gt_pch_nx_with_op <T>): New.
9247 (gt_pch_nx <T>): New.
9248 * passes.c (opt_pass::gt_ggc_mx): New.
9249 (opt_pass::gt_pch_nx): New.
9250 (opt_pass::gt_pch_nx_with_op): New.
9251 (pass_manager::gt_ggc_mx): New.
9252 (pass_manager::gt_pch_nx): New.
9253 (pass_manager::gt_pch_nx_with_op): New.
9254 (pass_manager::operator new): Use
9255 ggc_internal_cleared_alloc_stat rather than xcalloc.
9256 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9257 (pass_manager::gt_ggc_mx): New.
9258 (pass_manager::gt_pch_nx): New.
9259 (pass_manager::gt_pch_nx_with_op): New.
9260 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9261 (opt_pass::operator new): New.
9262 (opt_pass::gt_ggc_mx): New.
9263 (opt_pass::gt_pch_nx): New.
9264 (opt_pass::gt_pch_nx_with_op): New.
9266 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9268 * Makefile.in (GTFILES): Add context.h.
9269 * context.c (gcc::context::operator new): New.
9270 (gcc::context::gt_ggc_mx): New.
9271 (gcc::context::gt_pch_nx): New.
9272 (gcc::context::gt_pch_nx): New.
9273 * context.h (gcc::context): Add GTY((user)) marking.
9274 (gcc::context::operator new): New.
9275 (gcc::context::gt_ggc_mx): New.
9276 (gcc::context::gt_pch_nx): New.
9277 (gcc::context::gt_pch_nx): New.
9278 (g): Add GTY marking.
9279 (gt_ggc_mx (gcc::context *)): New.
9280 (gt_pch_nx (gcc::context *)): New.
9281 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9282 void *cookie)): New.
9283 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9285 2013-08-20 Alan Modra <amodra@gmail.com>
9288 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9289 (rs6000_emit_epilogue): Likewise.
9291 2013-08-19 Dehao Chen <dehao@google.com>
9293 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9295 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9296 Jakub Jelinek <jakub@redhat.com>
9298 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9299 (BUILT_IN_FABSD64): Likewise.
9300 (BUILT_IN_FABSD128): Likewise.
9301 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9302 (fold_builtin_1): Likewise.
9303 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9304 destination and source operands.
9305 (*abstd2_fpr): Likewise.
9306 (*nabstd2_fpr): Likewise.
9308 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9310 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9313 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9315 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9318 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9320 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9321 * builtins.c (default_libc_has_function): New.
9322 (gnu_libc_has_function): Ditto.
9323 (no_c99_libc_has_function): Ditto.
9324 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9325 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9326 (fold_builtin_sincos): Likewise.
9327 (fold_builtin_cexp): Likewise.
9328 * builtins.def (DEF_C94_BUILTIN): Likewise.
9329 (DEF_C99_BUILTIN): Likewise.
9330 (DEF_C99_C90RES_BUILTIN): Likewise.
9331 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9332 definitions to using this define.
9333 * config/darwin-protos.h (darwin_libc_has_function): New.
9334 * config/darwin.c (darwin_libc_has_function): Ditto.
9335 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9336 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9337 * config/darwin.h: Ditto.
9338 * config/elfos.h: Ditto.
9339 * config/freebsd.h: Ditto.
9340 * config/i386/cygming.h: Ditto.
9341 * config/i386/djgpp.h: Ditto.
9342 * config/i386/i386-interix.h: Ditto.
9343 * config/microblaze/microblaze.h: Ditto.
9344 * config/mmix/mmix.h: Ditto.
9345 * config/gnu-user.h: Ditto.
9346 * config/ia64/hpux.h: Ditto.
9347 * config/pa/pa-hpux.h: Ditto.
9348 * config/pdp11/pdp11.h: Ditto.
9349 * config/picochip/picochip.h: Ditto.
9350 * config/linux.h: Ditto.
9351 * config/netbsd.h: Ditto.
9352 * config/openbsd.h: Ditto.
9353 * config/rs6000/aix43.h: Ditto.
9354 * config/rs6000/aix51.h: Ditto.
9355 * config/rs6000/aix52.h: Ditto.
9356 * config/rs6000/aix53.h: Ditto.
9357 * config/rs6000/aix61.h: Ditto.
9358 * config/rs6000/darwin.h: Ditto.
9359 * config/rs6000/linux.h: Ditto.
9360 * config/rs6000/linux64.h: Ditto.
9361 * config/s390/tpf.h: Ditto.
9362 * config/sol2-10.h: Ditto.
9363 * config/sol2.h: Ditto.
9364 * config/vms/vms.h: Ditto.
9365 * config/vxworks.h: Ditto.
9366 * config/linux-android.c (linux_android_libc_has_function):
9367 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9368 * config/linux-protos.h (linux_android_libc_has_function):
9370 * config/i386/i386.c (ix86_libc_has_function): New.
9371 * config/i386/i386-protos.h
9372 (ix86_libc_has_function): New declaration.
9373 * config/i386/i386.md
9374 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9375 ("isinf<mode>2): Likewise.
9376 * convert.c (convert_to_integer): Using new target hook
9377 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9378 TARGET_C99_FUNCTIONS.
9379 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9380 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9381 * coretypes.h (function_class): New enum for different
9382 classes of functions.
9383 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9384 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9385 (TARGET_HAS_SINCOS): Likewise.
9386 (TARGET_LIBC_HAS_FUNCTION): New.
9387 * doc/tm.texi: Regenerated.
9388 * targhooks.h (default_libc_has_function): New declaration.
9389 (no_c99_libc_has_function): Ditto.
9390 (gnu_libc_has_function): Ditto.
9391 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9392 and TARGET_HAS_SINCOS.
9394 2013-08-18 Jan Hubicka <jh@suse.cz>
9396 * Makeifle-in (ipa-devirt.o): New.
9397 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9398 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9399 (analyze_functions): Look into possible targets of polymorphic call.
9400 * dumpfile.c (dump_files): Add type-inheritance dump.
9401 * dumpfile.h (TDI_inheritance): New.
9402 * ipa-devirt.c: New file.
9403 * ipa-utils.h (odr_type_d): Forward declare.
9404 (odr_type): New type.
9405 (build_type_inheritance_graph): Declare.
9406 (possible_polymorphic_call_targets): Declare and introduce inline
9407 variant when only edge is pased.
9408 (dump_possible_polymorphic_call_targets): Likewise.
9409 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9410 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9411 (types_same_for_odr): ... here.
9412 * tree.h (type_in_anonymous_namespace_p): Declare.
9414 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9416 PR tree-optimization/58006
9417 * tree-parloops.c (take_address_of): Don't ICE if get_name
9419 (eliminate_local_variables_stmt): Remove clobber stmts.
9421 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9423 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9424 error message is issued for an alias to undefined symbol.
9426 2013-08-18 Jan Hubicka <jh@suse.cz>
9428 * cgraph.c (cgraph_create_indirect_edge): Discover
9429 polymorphic calls and record basic info into indirect_info.
9430 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9431 devirtualization, ignore objc function calls.
9432 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9433 call with no parm index info.
9434 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9435 * tree.c (virtual_method_call_p): New function.
9436 * tree.h (virtual_method_call_p): Declare.
9438 2013-08-16 Jan Hubicka <jh@suse.cz>
9441 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9443 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9445 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9448 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9450 * gengtype.c (type_for_name): Add special-case support for
9451 locating types within the "gcc::" namespace.
9452 (open_base_files): Emit a "using namespace gcc" directive.
9454 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9457 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9458 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9460 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9461 array instead of each individual operand as a separate argument.
9462 (emit_fusion_gpr_load): Likewise.
9463 (expand_fusion_gpr_load): Add new function declaration.
9465 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9466 signature to have the operands passed as an array, instead of as
9467 separate arguments. Allow ZERO_EXTEND to be in the memory
9468 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9469 depend on the register live/dead flags when peepholes are run.
9470 (expand_fusion_gpr_load): New function to be called from the
9471 peephole2 pass, to change the register that addis sets to be the
9473 (emit_fusion_gpr_load): Change the calling signature to have the
9474 operands passed as an array, instead of as separate arguments.
9475 Allow ZERO_EXTEND to be in the memory address, and also
9476 SIGN_EXTEND if -mpower8-fusion-sign.
9478 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9480 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9481 adjust the register addis loads up in the peephole2 pass. Do not
9482 depend on the register live/dead state when the peephole pass is done.
9484 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9486 * gengtype.c (create_user_defined_type): Ensure that the kind
9487 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9488 declaration is seen before the GTY((user)) marking.
9490 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9493 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9495 2013-08-16 Jan Hubicka <jh@suse.cz>
9497 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9498 arugment expected_type.
9499 (gimple_fold_call): Use it.
9500 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9501 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9502 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9503 (try_make_edge_direct_virtual_call): Likewise.
9504 * tree.c (obj_type_ref_class): New.
9505 * tree.h (obj_type_ref_class): Use it.
9507 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9509 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9510 (rtl_slim_pp): Likewise.
9511 (init_rtl_slim_pretty_print): Likewise.
9512 (dump_value_slim): Don't call it. Use local pretty printer.
9513 (dump_insn_slim): Likewise.
9514 (dump_rtl_slim): Likewise.
9515 (str_pattern_slim): Likewise.
9516 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9519 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9521 PR tree-optimization/58164
9522 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9523 walk gimple_goto_dest of GIMPLE_GOTO.
9525 PR tree-optimization/58165
9526 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9527 bi_call must be the last stmt in a bb, don't split_block, instead
9528 use fallthru edge from it and give up if there is none.
9529 Release conds vector when returning early.
9531 2013-08-14 Xinliang David Li <davidxl@google.com>
9533 * config/i386/i386.c (ix86_option_override_internal):
9534 Remove unused variable and field.
9536 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9539 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9540 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9541 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9542 and Linux, correct BLKmode alignment when 128-bit alignment is
9543 required and compatibility flag is not set.
9544 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9545 for zero-size arguments when compatibility flag is not set.
9547 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9549 PR tree-optimization/58145
9550 * tree-sra.c (build_ref_for_offset): If prev_base has
9551 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9553 2013-08-14 Xinliang David Li <davidxl@google.com>
9555 * config/i386/i386.c (ix86_option_override_internal):
9556 Fix uninitialized variable error.
9558 2013-08-14 Xinliang David Li <davidxl@google.com>
9560 * config/i386/i386.opt: Define two new options.
9561 * config/i386/x86-tune.def: Add arch selector field in macros.
9562 * config/i386/i386.h: Adjust macro definition.
9563 * config/i386/i386.c (ix86_option_override_internal):
9565 (parse_mtune_ctrl_str): New function.
9566 (set_ix86_tune_features): New function.
9567 (ix86_function_specific_restore): Call the new helper function.
9569 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9571 PR rtl-optimization/57662
9572 * sel-sched.c (code_motion_process_successors): When the current insn
9573 is removed after the recursive traversal, break from the loop.
9574 Add comments and debug printouts.
9576 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9577 Alexandre Oliva <aoliva@redhat.com>
9580 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9581 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9582 there also UNSPEC_PLTOFF.
9584 2013-08-14 Marek Polacek <polacek@redhat.com>
9586 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9587 NUM_CONDITIONS bit positions.
9589 2013-08-13 Cary Coutant <ccoutant@google.com>
9591 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9592 (attr_checksum): Hash vector contents instead of pointer.
9593 (attr_checksum_ordered): Likewise.
9595 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9597 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9598 when Pmode != word_mode. Add length_address attribute.
9599 (sse3_monitor_<mode>): Merge from sse3_monitor and
9600 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9601 Pmode != word_mode. Update insn length attribute.
9602 * config/i386/i386.c (ix86_option_override_internal): Update
9603 ix86_gen_monitor selection for merged sse3_monitor insn.
9605 2013-08-13 Julian Brown <julian@codesourcery.com>
9607 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9608 perform invalid legitimization on greater-than-word-size modes for
9611 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9613 * ira.c (setup_class_translate_array): Use aclass instead of cl
9614 for classes not fully covered by allocno classes.
9616 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9618 PR tree-optimization/57661
9619 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9620 * tree-inline.c (tree_function_versioning): Initialize it.
9621 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9622 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9623 that is not being copied.
9626 * asan.c (instrument_strlen_call): Fix typo in comment.
9627 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9629 2013-08-13 Steve Ellcey <sellcey@mips.com>
9631 * config/mips/mips.md (prefetch): Use lw instead of ld on
9632 loongson in 32bit mode.
9634 2013-08-13 Nick Clifton <nickc@redhat.com>
9636 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9638 2013-08-13 Jan Hubicka <jh@suse.cz>
9640 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9641 introduced edge; fix typo in sanity check.
9642 (cgraph_resolve_speculation): Export; improve diagnostic.
9643 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9644 speculation at type mismatch.
9645 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9646 (cgraph_resolve_speculation): Declare.
9647 (symtab_can_be_discarded): New function.
9648 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9649 * ipa-inline-transform.c (speculation_removed): New global var.
9650 (clone_inlined_nodes): See if speculation can be removed.
9651 (inline_call): If speculations was removed, we growths may not match.
9652 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9653 (speculation_useful_p): New function.
9654 (resolve_noninline_speculation): New function.
9655 (inline_small_functions): Resolve useless speculations.
9656 * ipa-inline.h (speculation_useful_p): Declare
9657 * ipa.c (can_replace_by_local_alias): Simplify.
9658 (ipa_profile): Produce speculative calls in non-lto, too;
9659 add simple cost model; produce local aliases.
9661 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9663 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9666 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9668 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9670 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9672 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9673 iterator instead of X87MODEF.
9675 2013-08-12 Perez Read <netfirewall@gmail.com>
9678 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9679 operand 0 for intel asm alternative.
9680 (*movabs<mode>_2): Ditto for operand 1.
9682 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9684 * config/aarch64/arm_none.h
9685 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9687 2013-08-12 Nick Clifton <nickc@redhat.com>
9689 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9691 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9693 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9694 expand for QI/HImode operand to produce more effictive code for
9695 unsigned char(short) --> float(double) conversion.
9697 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9699 * doc/invoke.texi: Mention that -ftls-model does not force the final
9702 2013-08-12 Marek Polacek <polacek@redhat.com>
9703 Marc Glisse <marc.glisse@inria.fr>
9705 PR tree-optimization/57980
9706 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9707 when creating -1 constant.
9709 2013-08-10 Jan Hubicka <jh@suse.cz>
9711 Workaround binutils PR14342.
9712 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9713 (gimple_init_edge_profiler): Likewise.
9714 (gimple_gen_ic_func_profiler): Likewise.
9716 2013-08-09 Jan Hubicka <jh@suse.cz>
9718 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9720 2013-08-09 Xinliang David Li <davidxl@google.com>
9722 * config/i386/stringop.def: New file.
9723 * config/i386/stringop.opt: New file.
9724 * config/i386/i386-opts.h: Include stringopt.def.
9725 * config/i386/i386.opt: Include stringopt.opt.
9726 * config/i386/i386.c (ix86_option_override_internal):
9727 Override default size based stringop inline strategies with options.
9728 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9731 2013-08-09 Jan Hubicka <jh@suse.cz>
9733 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9735 2013-08-09 Jan Hubicka <jh@suse.cz>
9737 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9739 (dump_cgraph_node): Dump profile-id.
9740 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9741 and common_target_probability.
9742 * lto-cgraph.c (lto_output_edge): Stream common targets.
9743 (lto_output_node): Stream profile ids.
9744 (input_node): Stream profile ids.
9745 (input_edge): Stream common targets.
9746 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9747 * ipa.c: Include value-prof.h
9748 (ipa_profile_generate_summary): Turn indirect call statement histograms
9749 into common targets.
9750 (ipa_profile): Turn common targets into speculative edges.
9752 2013-08-09 Jan Hubicka <jh@suse.cz>
9754 * cgraph.h (cgraph_node): Add profile_id.
9755 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9756 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9757 (del_node_map): Update.
9758 (find_func_by_funcdef_no): Replace by ...
9759 (find_func_by_profile_id): ... this one.
9760 (gimple_ic_transform): Do not remove useful histograms when
9761 speculation is not done; dump info when indirect call removal
9763 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9764 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9765 (__gcov_indirect_call_profiler_v2): .. this one.
9766 * profile.h (init_node_map): Update.
9767 * coverage.c (coverage_compute_profile_id): New function.
9768 * coverage.h (coverage_compute_profile_id): Declare.
9769 * tree-profile.c (init_ic_make_global_vars): Make
9770 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9771 (gimple_init_edge_profiler): Update prototype of
9772 __gcov_indirect_call_profiler.
9773 (gimple_gen_ic_func_profiler): Simplify.
9774 (tree_profiling): Use init_node_map
9776 2013-08-09 Jan Hubicka <jh@suse.cz>
9778 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9779 non-speculative refs.
9780 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9781 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9782 (cgraph_set_call_stmt): Likewise.
9783 (cgraph_create_edge_1): Fix release checking compilatoin;
9785 (cgraph_free_edge): Free indirect info.
9786 (cgraph_turn_edge_to_speculative): New function.
9787 (cgraph_speculative_call_info): New function.
9788 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9789 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9790 (dump_cgraph_node): Dump speculation.
9791 (verify_edge_count_and_frequency): Accept speculative edges.
9792 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9793 (verify_cgraph_node): Handle speculation.
9794 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9795 (cgraph_set_call_stmt): Update prototype.
9796 (cgraph_make_edge_direct): Update prototype.
9797 (cgraph_speculative_call_info): Declare.
9798 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9799 to change; update call of ipa_find_references.
9800 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9801 lto_stmt_uid and speculative flags.
9802 (ipa_dump_references): Dump speculation.
9803 (ipa_clone_references): Clone speculative flag.
9804 (ipa_clone_referring): Likewise.
9805 (ipa_clone_ref): New function.
9806 (ipa_find_reference): Look into lto_stmt_uids
9807 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9808 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9809 (ipa_find_reference): Update declaration.
9810 (ipa_clone_ref): Declare.
9811 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9812 stream speculative flag.
9813 (lto_output_ref): Stream statements uids and speculation.
9814 (input_ref): Likewise.
9815 (input_edge): Stream speuclation.
9816 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9817 (cgraph_set_call_stmt_including_clones): Handle speculation.
9818 * ipa-inline.c (heap_edge_removal_hook): New function.
9819 (inline_small_functions): Register it.
9820 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9821 also initialize refs.
9822 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9824 (try_make_edge_direct_simple_call): Likewise.
9825 (try_make_edge_direct_simple_call): Likewise.
9826 (update_indirect_edges_after_inlining): Likewise.
9827 (remove_described_reference): Look proper lto_stmt_uid.
9828 (propagate_controlled_uses): Likewise.
9829 (propagate_controlled_uses): Liekwise.
9830 * tree-inline.c (copy_bb): Copy speculative edges.
9831 (redirect_all_calls): New function.
9832 (copy_cfg_body): Do redirection after loop info is updated.
9833 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9835 2013-08-09 Jan Hubicka <jh@suse.cz>
9837 * lto-streamer-out.c (output_function): Renumber PHIs.
9838 * lto-streamer-in.c (input_function): Likewise.
9840 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9842 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9843 (get_lane_unsigned): Likewise.
9844 (dup_lane_scalar): Likewise.
9845 (get_lane): enable for VALL.
9846 * config/aarch64/aarch64-simd.md
9847 (aarch64_dup_lane_scalar<mode>): Remove.
9848 (aarch64_get_lane_signed<mode>): Likewise.
9849 (aarch64_get_lane_unsigned<mode>): Likewise.
9850 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9851 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9852 (aarch64_get_lane<mode>): Enable for all vector modes.
9853 (aarch64_get_lanedi): Remove misleading constraints.
9854 * config/aarch64/arm_neon.h
9855 (__aarch64_vget_lane_any): Define.
9856 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9857 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9858 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9859 * config/aarch64/iterators.md (VDQQH): New.
9863 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9865 * configure.ac: Add GAS check for LEON instructions on SPARC.
9866 * configure: Regenerate.
9867 * config.in: Likewise.
9868 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9870 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9871 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9873 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9874 (AS_LEON_FLAG): New macro.
9875 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9876 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9877 Deal with LEON and LEON3 for the memory model.
9878 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9879 (atomic_compare_and_swap<mode>_1): Likewise.
9880 (*atomic_compare_and_swap<mode>_1): Likewise.
9882 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9884 * config/arm/neon.md (vcond): Fix floating-point vector
9885 comparisons against 0.
9887 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9889 * lra-constraints.c (emit_spill_move): Remove assert.
9890 (process_alt_operands): Add more debugging
9891 output. Increase reject for spilling into memory. Decrease
9892 reject for reloading scratch.
9893 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9895 2013-08-08 Steve Ellcey <sellcey@mips.com>
9897 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9898 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9899 micromips incompatible. Add nan2008.
9900 (MULTILIB_DIRNAMES): Add nan2008.
9901 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9902 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9903 and micromips incompatible. Add nan2008.
9904 (MULTILIB_DIRNAMES): Add nan2008.
9905 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9907 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9909 PR rtl-optimization/58079
9910 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9911 simplify_comparison has widened a comparison with an integer.
9913 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9915 * config/arm/neon.md (movmisalign<mode>): Disable when we
9916 don't allow unaligned accesses.
9917 (*movmisalign<mode>_neon_store): Likewise.
9918 (*movmisalign<mode>_neon_load): Likewise.
9919 (*movmisalign<mode>_neon_store): Likewise.
9920 (*movmisalign<mode>_neon_load): Likewise.
9922 2013-08-08 Jan Hubicka <jh@suse.cz>
9924 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9925 (make_pass_rebuild_cgraph_edges): Also clear references.
9926 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9927 * ipa-inline-transform.c (inline_transform): Remove all references
9929 * cgraphunit.c (expand_function): Remove all references after
9931 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9932 (ipa_find_reference): Rewrite to iterator.
9933 (remove_stmt_references): Likewise.
9934 (ipa_clear_stmts_in_references): New function.
9935 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9936 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9938 * ipa-split.c (split_function): Remove references in split function.
9940 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9943 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9944 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9946 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9949 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9950 suggested mode for the assignment isn't compatible with the
9953 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
9956 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
9958 2013-08-07 Xinliang David Li <davidxl@google.com>
9960 * config/i386/i386.opt: New option -mtune-ctrl=.
9961 * config/i386/x86-tune.def: New file.
9962 * config/i386/i386.h: include x86-tune.def.
9963 * config/i386/i386.c (ix86_option_override_internal):
9964 Parsing -mtune-ctrl= option and set tune features.
9966 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
9969 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
9970 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
9973 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9975 * rtl.h (update_alignments): Declare.
9976 * final.c (grow_label_align): New function extracted from...
9977 (shorten_branches): ...here. Call it.
9978 (update_alignments): New function.
9979 * reorg.c (sibling_labels): New variable.
9980 (get_label_before): Add SIBLING parameter. If it is non-zero, push
9981 the new label along with it onto the sibling_labels vector.
9982 (fill_simple_delay_slots): Adjust call to get_label_before.
9983 (fill_slots_from_thread): Likewise.
9984 (relax_delay_slots): Likewise.
9985 (make_return_insns): Likewise.
9986 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
9988 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
9990 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
9991 document its semantics.
9992 (diagnostic_report_diagnostic): Adjust accordingly.
9994 2013-08-07 David Malcolm <dmalcolm@redhat.com>
9996 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
9997 (sparc_option_override): ...and port to new C++ pass API.
9998 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
10000 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
10002 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
10004 2013-08-06 Caroline Tice <cmtice@google.com>
10006 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
10007 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
10008 * tree-pass.h: Add pass_vtable_verify.
10009 * varasm.c (assemble_variable): Add code to properly set the comdat
10010 section and name for the .vtable_map_vars section.
10011 (assemble_vtyv_preinit_initializer): New function.
10012 (default_sectin_type_flags): Make sure .vtable_map_vars section has
10014 * output.h: Add function decl for assemble_vtv_preinit_initializer.
10015 * vtable-verify.c: New file.
10016 * vtable-verify.h: New file.
10017 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
10018 initialiation levels.
10019 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
10020 * passes.def: Insert pass_vtable_verify.
10021 * aclocal.m4: Reorder includes.
10022 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
10023 -fvtv-counts options.
10024 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
10025 as appropriate, if -fvtable-verify=... is used.
10026 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
10027 -fvtable-verify=... is used.
10028 * Makefile.in (OBJS): Add vtable-verify.o to list.
10029 (vtable-verify.o): Add new build rule.
10030 (GTFILES): Add vtable-verify.c to list.
10031 * common.opt (fvtable-verify=): New flag.
10032 (vtv_priority): Values for fvtable-verify= flag.
10033 (fvtv-counts): New flag.
10034 (fvtv-debug): New flag.
10035 * tree.h (save_vtable_map_decl): New extern function decl.
10037 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10039 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
10040 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
10041 (pass_data_rl78_devirt): ...new pass_data instance and...
10042 (make_pass_rl78_devirt): ...new function.
10043 (rl78_asm_file_start): Port pass registration to new C++ API.
10045 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10047 * coretypes.h (rtl_opt_pass): Add.
10048 (gcc::context): Add.
10049 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
10050 (epiphany_init): Port to new C++ pass API.
10051 (epiphany_optimize_mode_switching): Likewise.
10052 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
10053 (pass_manager::get_pass_mode_switching): New.
10054 (pass_manager::get_pass_peephole2): New.
10055 * mode-switching.c (pass_mode_switching): Add clone method.
10056 * recog.c (pass_peephole2): Add clone method.
10057 (pass_split_all_insns): Add clone method.
10059 2013-08-06 David Malcolm <dmalcolm@redhat.com>
10061 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
10062 (mips_option_override): ...here, porting to new C++ API for passes.
10064 2013-08-06 Jan Hubicka <jh@suse.cz>
10066 * cgraph.c (cgraph_get_body): New function based on lto.c
10068 * cgraph.h (cgraph_get_body): Declare.
10069 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
10071 * cgraphunit.c (expand_function): Get body prior expanding.
10072 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
10073 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
10075 * passes.c (do_per_function_toporder): Get body.
10076 * tree-inline.c (expand_call_inline): Get body prior inlining it.
10077 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
10079 2013-08-06 Martin Jambor <mjambor@suse.cz>
10082 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
10083 is not re-finalized. Rename second parameter to no_collect.
10085 2013-08-06 Martin Jambor <mjambor@suse.cz>
10087 PR middle-end/58041
10088 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
10089 MEM_REF has proper alignment information.
10091 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
10094 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
10096 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
10097 rtx (*) (rtx, ...) with insn_gen_fn.
10098 * genoutput.c (output_insn_data): Cast gen_? function pointers to
10099 insn_gen_fn::stored_funcptr. Add initializer braces.
10101 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10103 Rewrite how instances of passes are cloned to remove assumptions
10104 about their sizes (thus allowing pass subclasses to have
10105 additional data fields, albeit non-GC-managed ones at this point).
10107 * passes.c (make_pass_instance): Now that passes have clone
10108 methods, rewrite this function to eliminate XNEW and memcpy
10109 calls that used hardcoded sizes. Since this function no longer
10110 creates pass instances, rename it to...
10111 (add_pass_instance): ...this. Document the old way that passes were
10112 numbered and flagged, and rework this function to continue using it.
10113 (next_pass_1): Add an initial_pass argument for use by
10115 (position_pass): When adding multiple instances of a pass, use
10116 the pass's clone method, rather than relying on the XNEW/memcpy
10117 within the former make_pass_instance (now add_pass_instance).
10118 (pass_manager::pass_manager): When invoking next_pass_1, also supply
10119 the initial instance of the current pass within the pass manager.
10121 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10123 This is the automated part of the conversion of passes from C
10124 structs to C++ classes.
10126 Patch autogenerated by refactor_passes.py from
10127 https://github.com/davidmalcolm/gcc-refactoring-scripts
10128 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
10130 * asan.c (pass_asan): Convert from a global struct to a subclass of
10131 gimple_opt_pass along with...
10132 (pass_data_asan): ...new pass_data instance and...
10133 (make_pass_asan): ...new function.
10134 (pass_asan_O0): Convert from a global struct to a subclass of
10135 gimple_opt_pass along with...
10136 (pass_data_asan_O0): ...new pass_data instance and...
10137 (make_pass_asan_O0): ...new function.
10138 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10139 subclass of rtl_opt_pass along with...
10140 (pass_data_inc_dec): ...new pass_data instance and...
10141 (make_pass_inc_dec): ...new function.
10142 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10143 a subclass of rtl_opt_pass along with...
10144 (pass_data_reorder_blocks): ...new pass_data instance and...
10145 (make_pass_reorder_blocks): ...new function.
10146 (pass_duplicate_computed_gotos): Convert from a global struct to a
10147 subclass of rtl_opt_pass along with...
10148 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10149 (make_pass_duplicate_computed_gotos): ...new function.
10150 (pass_partition_blocks): Convert from a global struct to a subclass of
10151 rtl_opt_pass along with...
10152 (pass_data_partition_blocks): ...new pass_data instance and...
10153 (make_pass_partition_blocks): ...new function.
10154 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10155 struct to a subclass of rtl_opt_pass along with...
10156 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10158 (make_pass_branch_target_load_optimize1): ...new function.
10159 (pass_branch_target_load_optimize2): Convert from a global struct to a
10160 subclass of rtl_opt_pass along with...
10161 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10163 (make_pass_branch_target_load_optimize2): ...new function.
10164 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10165 of rtl_opt_pass along with...
10166 (pass_data_jump): ...new pass_data instance and...
10167 (make_pass_jump): ...new function.
10168 (pass_jump2): Convert from a global struct to a subclass of
10169 rtl_opt_pass along with...
10170 (pass_data_jump2): ...new pass_data instance and...
10171 (make_pass_jump2): ...new function.
10172 * cfgexpand.c (pass_expand): Convert from a global struct to a
10173 subclass of rtl_opt_pass along with...
10174 (pass_data_expand): ...new pass_data instance and...
10175 (make_pass_expand): ...new function.
10176 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10177 of rtl_opt_pass along with...
10178 (pass_data_free_cfg): ...new pass_data instance and...
10179 (make_pass_free_cfg): ...new function.
10180 (pass_into_cfg_layout_mode): Convert from a global struct to a
10181 subclass of rtl_opt_pass along with...
10182 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10183 (make_pass_into_cfg_layout_mode): ...new function.
10184 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10185 subclass of rtl_opt_pass along with...
10186 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10187 (make_pass_outof_cfg_layout_mode): ...new function.
10188 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10189 struct to a subclass of gimple_opt_pass along with...
10190 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10191 (make_pass_build_cgraph_edges): ...new function.
10192 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10193 subclass of gimple_opt_pass along with...
10194 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10195 (make_pass_rebuild_cgraph_edges): ...new function.
10196 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10197 subclass of gimple_opt_pass along with...
10198 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10200 (make_pass_remove_cgraph_callee_edges): ...new function.
10201 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10202 struct to a subclass of rtl_opt_pass along with...
10203 (pass_data_stack_adjustments): ...new pass_data instance and...
10204 (make_pass_stack_adjustments): ...new function.
10205 * combine.c (pass_combine): Convert from a global struct to a subclass
10206 of rtl_opt_pass along with...
10207 (pass_data_combine): ...new pass_data instance and...
10208 (make_pass_combine): ...new function.
10209 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10210 global struct to a subclass of rtl_opt_pass along with...
10211 (pass_data_compare_elim_after_reload): ...new pass_data instance
10213 (make_pass_compare_elim_after_reload): ...new function.
10214 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10215 of rtl_opt_pass along with...
10216 (pass_data_rtl_cprop): ...new pass_data instance and...
10217 (make_pass_rtl_cprop): ...new function.
10218 * cse.c (pass_cse): Convert from a global struct to a subclass of
10219 rtl_opt_pass along with...
10220 (pass_data_cse): ...new pass_data instance and...
10221 (make_pass_cse): ...new function.
10222 (pass_cse2): Convert from a global struct to a subclass of
10223 rtl_opt_pass along with...
10224 (pass_data_cse2): ...new pass_data instance and...
10225 (make_pass_cse2): ...new function.
10226 (pass_cse_after_global_opts): Convert from a global struct to a
10227 subclass of rtl_opt_pass along with...
10228 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10229 (make_pass_cse_after_global_opts): ...new function.
10230 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10231 of rtl_opt_pass along with...
10232 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10233 (make_pass_ud_rtl_dce): ...new function.
10234 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10235 rtl_opt_pass along with...
10236 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10237 (make_pass_fast_rtl_dce): ...new function.
10238 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10239 a subclass of rtl_opt_pass along with...
10240 (pass_data_df_initialize_opt): ...new pass_data instance and...
10241 (make_pass_df_initialize_opt): ...new function.
10242 (pass_df_initialize_no_opt): Convert from a global struct to a
10243 subclass of rtl_opt_pass along with...
10244 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10245 (make_pass_df_initialize_no_opt): ...new function.
10246 (pass_df_finish): Convert from a global struct to a subclass of
10247 rtl_opt_pass along with...
10248 (pass_data_df_finish): ...new pass_data instance and...
10249 (make_pass_df_finish): ...new function.
10250 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10251 rtl_opt_pass along with...
10252 (pass_data_rtl_dse1): ...new pass_data instance and...
10253 (make_pass_rtl_dse1): ...new function.
10254 (pass_rtl_dse2): Convert from a global struct to a subclass of
10255 rtl_opt_pass along with...
10256 (pass_data_rtl_dse2): ...new pass_data instance and...
10257 (make_pass_rtl_dse2): ...new function.
10258 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10259 subclass of rtl_opt_pass along with...
10260 (pass_data_dwarf2_frame): ...new pass_data instance and...
10261 (make_pass_dwarf2_frame): ...new function.
10262 * except.c (pass_set_nothrow_function_flags): Convert from a global
10263 struct to a subclass of rtl_opt_pass along with...
10264 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10266 (make_pass_set_nothrow_function_flags): ...new function.
10267 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10268 subclass of rtl_opt_pass along with...
10269 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10271 (make_pass_convert_to_eh_region_ranges): ...new function.
10272 * final.c (pass_compute_alignments): Convert from a global struct to a
10273 subclass of rtl_opt_pass along with...
10274 (pass_data_compute_alignments): ...new pass_data instance and...
10275 (make_pass_compute_alignments): ...new function.
10276 (pass_final): Convert from a global struct to a subclass of
10277 rtl_opt_pass along with...
10278 (pass_data_final): ...new pass_data instance and...
10279 (make_pass_final): ...new function.
10280 (pass_shorten_branches): Convert from a global struct to a subclass of
10281 rtl_opt_pass along with...
10282 (pass_data_shorten_branches): ...new pass_data instance and...
10283 (make_pass_shorten_branches): ...new function.
10284 (pass_clean_state): Convert from a global struct to a subclass of
10285 rtl_opt_pass along with...
10286 (pass_data_clean_state): ...new pass_data instance and...
10287 (make_pass_clean_state): ...new function.
10288 * function.c (pass_instantiate_virtual_regs): Convert from a global
10289 struct to a subclass of rtl_opt_pass along with...
10290 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10291 (make_pass_instantiate_virtual_regs): ...new function.
10292 (pass_leaf_regs): Convert from a global struct to a subclass of
10293 rtl_opt_pass along with...
10294 (pass_data_leaf_regs): ...new pass_data instance and...
10295 (make_pass_leaf_regs): ...new function.
10296 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10297 subclass of rtl_opt_pass along with...
10298 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10300 (make_pass_thread_prologue_and_epilogue): ...new function.
10301 (pass_match_asm_constraints): Convert from a global struct to a
10302 subclass of rtl_opt_pass along with...
10303 (pass_data_match_asm_constraints): ...new pass_data instance and...
10304 (make_pass_match_asm_constraints): ...new function.
10305 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10306 subclass of rtl_opt_pass along with...
10307 (pass_data_rtl_fwprop): ...new pass_data instance and...
10308 (make_pass_rtl_fwprop): ...new function.
10309 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10310 rtl_opt_pass along with...
10311 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10312 (make_pass_rtl_fwprop_addr): ...new function.
10313 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10314 rtl_opt_pass along with...
10315 (pass_data_rtl_pre): ...new pass_data instance and...
10316 (make_pass_rtl_pre): ...new function.
10317 (pass_rtl_hoist): Convert from a global struct to a subclass of
10318 rtl_opt_pass along with...
10319 (pass_data_rtl_hoist): ...new pass_data instance and...
10320 (make_pass_rtl_hoist): ...new function.
10321 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10322 subclass of gimple_opt_pass along with...
10323 (pass_data_lower_cf): ...new pass_data instance and...
10324 (make_pass_lower_cf): ...new function.
10325 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10326 from a global struct to a subclass of gimple_opt_pass along with...
10327 (pass_data_strength_reduction): ...new pass_data instance and...
10328 (make_pass_strength_reduction): ...new function.
10329 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10330 of rtl_opt_pass along with...
10331 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10332 (make_pass_rtl_ifcvt): ...new function.
10333 (pass_if_after_combine): Convert from a global struct to a subclass of
10334 rtl_opt_pass along with...
10335 (pass_data_if_after_combine): ...new pass_data instance and...
10336 (make_pass_if_after_combine): ...new function.
10337 (pass_if_after_reload): Convert from a global struct to a subclass of
10338 rtl_opt_pass along with...
10339 (pass_data_if_after_reload): ...new pass_data instance and...
10340 (make_pass_if_after_reload): ...new function.
10341 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10342 a subclass of rtl_opt_pass along with...
10343 (pass_data_initialize_regs): ...new pass_data instance and...
10344 (make_pass_initialize_regs): ...new function.
10345 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10346 of ipa_opt_pass_d along with...
10347 (pass_data_ipa_cp): ...new pass_data instance and...
10348 (make_pass_ipa_cp): ...new function.
10349 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10350 global struct to a subclass of gimple_opt_pass along with...
10351 (pass_data_inline_parameters): ...new pass_data instance and...
10352 (make_pass_inline_parameters): ...new function.
10353 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10354 subclass of gimple_opt_pass along with...
10355 (pass_data_early_inline): ...new pass_data instance and...
10356 (make_pass_early_inline): ...new function.
10357 (pass_ipa_inline): Convert from a global struct to a subclass of
10358 ipa_opt_pass_d along with...
10359 (pass_data_ipa_inline): ...new pass_data instance and...
10360 (make_pass_ipa_inline): ...new function.
10361 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10362 struct to a subclass of gimple_opt_pass along with...
10363 (pass_data_local_pure_const): ...new pass_data instance and...
10364 (make_pass_local_pure_const): ...new function.
10365 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10366 ipa_opt_pass_d along with...
10367 (pass_data_ipa_pure_const): ...new pass_data instance and...
10368 (make_pass_ipa_pure_const): ...new function.
10369 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10370 to a subclass of ipa_opt_pass_d along with...
10371 (pass_data_ipa_reference): ...new pass_data instance and...
10372 (make_pass_ipa_reference): ...new function.
10373 * ipa-split.c (pass_split_functions): Convert from a global struct to
10374 a subclass of gimple_opt_pass along with...
10375 (pass_data_split_functions): ...new pass_data instance and...
10376 (make_pass_split_functions): ...new function.
10377 (pass_feedback_split_functions): Convert from a global struct to a
10378 subclass of gimple_opt_pass along with...
10379 (pass_data_feedback_split_functions): ...new pass_data instance and...
10380 (make_pass_feedback_split_functions): ...new function.
10381 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10382 global struct to a subclass of simple_ipa_opt_pass along with...
10383 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10385 (make_pass_ipa_function_and_variable_visibility): ...new function.
10386 (pass_ipa_free_inline_summary): Convert from a global struct to a
10387 subclass of simple_ipa_opt_pass along with...
10388 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10389 (make_pass_ipa_free_inline_summary): ...new function.
10390 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10391 subclass of ipa_opt_pass_d along with...
10392 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10394 (make_pass_ipa_whole_program_visibility): ...new function.
10395 (pass_ipa_profile): Convert from a global struct to a subclass of
10396 ipa_opt_pass_d along with...
10397 (pass_data_ipa_profile): ...new pass_data instance and...
10398 (make_pass_ipa_profile): ...new function.
10399 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10400 ipa_opt_pass_d along with...
10401 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10402 (make_pass_ipa_cdtor_merge): ...new function.
10403 * ira.c (pass_ira): Convert from a global struct to a subclass of
10404 rtl_opt_pass along with...
10405 (pass_data_ira): ...new pass_data instance and...
10406 (make_pass_ira): ...new function.
10407 (pass_reload): Convert from a global struct to a subclass of
10408 rtl_opt_pass along with...
10409 (pass_data_reload): ...new pass_data instance and...
10410 (make_pass_reload): ...new function.
10411 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10412 subclass of rtl_opt_pass along with...
10413 (pass_data_cleanup_barriers): ...new pass_data instance and...
10414 (make_pass_cleanup_barriers): ...new function.
10415 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10416 of rtl_opt_pass along with...
10417 (pass_data_loop2): ...new pass_data instance and...
10418 (make_pass_loop2): ...new function.
10419 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10420 rtl_opt_pass along with...
10421 (pass_data_rtl_loop_init): ...new pass_data instance and...
10422 (make_pass_rtl_loop_init): ...new function.
10423 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10424 rtl_opt_pass along with...
10425 (pass_data_rtl_loop_done): ...new pass_data instance and...
10426 (make_pass_rtl_loop_done): ...new function.
10427 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10428 subclass of rtl_opt_pass along with...
10429 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10430 (make_pass_rtl_move_loop_invariants): ...new function.
10431 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10432 rtl_opt_pass along with...
10433 (pass_data_rtl_unswitch): ...new pass_data instance and...
10434 (make_pass_rtl_unswitch): ...new function.
10435 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10436 subclass of rtl_opt_pass along with...
10437 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10439 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10440 (pass_rtl_doloop): Convert from a global struct to a subclass of
10441 rtl_opt_pass along with...
10442 (pass_data_rtl_doloop): ...new pass_data instance and...
10443 (make_pass_rtl_doloop): ...new function.
10444 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10445 a subclass of rtl_opt_pass along with...
10446 (pass_data_lower_subreg): ...new pass_data instance and...
10447 (make_pass_lower_subreg): ...new function.
10448 (pass_lower_subreg2): Convert from a global struct to a subclass of
10449 rtl_opt_pass along with...
10450 (pass_data_lower_subreg2): ...new pass_data instance and...
10451 (make_pass_lower_subreg2): ...new function.
10452 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10453 struct to a subclass of ipa_opt_pass_d along with...
10454 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10455 (make_pass_ipa_lto_gimple_out): ...new function.
10456 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10457 of ipa_opt_pass_d along with...
10458 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10459 (make_pass_ipa_lto_finish_out): ...new function.
10460 * mode-switching.c (pass_mode_switching): Convert from a global struct
10461 to a subclass of rtl_opt_pass along with...
10462 (pass_data_mode_switching): ...new pass_data instance and...
10463 (make_pass_mode_switching): ...new function.
10464 * modulo-sched.c (pass_sms): Convert from a global struct to a
10465 subclass of rtl_opt_pass along with...
10466 (pass_data_sms): ...new pass_data instance and...
10467 (make_pass_sms): ...new function.
10468 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10469 subclass of gimple_opt_pass along with...
10470 (pass_data_expand_omp): ...new pass_data instance and...
10471 (make_pass_expand_omp): ...new function.
10472 (pass_lower_omp): Convert from a global struct to a subclass of
10473 gimple_opt_pass along with...
10474 (pass_data_lower_omp): ...new pass_data instance and...
10475 (make_pass_lower_omp): ...new function.
10476 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10477 of gimple_opt_pass along with...
10478 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10479 (make_pass_diagnose_omp_blocks): ...new function.
10480 * passes.c (pass_early_local_passes): Convert from a global struct to
10481 a subclass of simple_ipa_opt_pass along with...
10482 (pass_data_early_local_passes): ...new pass_data instance and...
10483 (make_pass_early_local_passes): ...new function.
10484 (pass_all_early_optimizations): Convert from a global struct to a
10485 subclass of gimple_opt_pass along with...
10486 (pass_data_all_early_optimizations): ...new pass_data instance and...
10487 (make_pass_all_early_optimizations): ...new function.
10488 (pass_all_optimizations): Convert from a global struct to a subclass
10489 of gimple_opt_pass along with...
10490 (pass_data_all_optimizations): ...new pass_data instance and...
10491 (make_pass_all_optimizations): ...new function.
10492 (pass_all_optimizations_g): Convert from a global struct to a subclass
10493 of gimple_opt_pass along with...
10494 (pass_data_all_optimizations_g): ...new pass_data instance and...
10495 (make_pass_all_optimizations_g): ...new function.
10496 (pass_rest_of_compilation): Convert from a global struct to a subclass
10497 of rtl_opt_pass along with...
10498 (pass_data_rest_of_compilation): ...new pass_data instance and...
10499 (make_pass_rest_of_compilation): ...new function.
10500 (pass_postreload): Convert from a global struct to a subclass of
10501 rtl_opt_pass along with...
10502 (pass_data_postreload): ...new pass_data instance and...
10503 (make_pass_postreload): ...new function.
10504 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10505 subclass of rtl_opt_pass along with...
10506 (pass_data_gcse2): ...new pass_data instance and...
10507 (make_pass_gcse2): ...new function.
10508 * postreload.c (pass_postreload_cse): Convert from a global struct to
10509 a subclass of rtl_opt_pass along with...
10510 (pass_data_postreload_cse): ...new pass_data instance and...
10511 (make_pass_postreload_cse): ...new function.
10512 * predict.c (pass_profile): Convert from a global struct to a subclass
10513 of gimple_opt_pass along with...
10514 (pass_data_profile): ...new pass_data instance and...
10515 (make_pass_profile): ...new function.
10516 (pass_strip_predict_hints): Convert from a global struct to a subclass
10517 of gimple_opt_pass along with...
10518 (pass_data_strip_predict_hints): ...new pass_data instance and...
10519 (make_pass_strip_predict_hints): ...new function.
10520 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10521 of rtl_opt_pass along with...
10522 (pass_data_peephole2): ...new pass_data instance and...
10523 (make_pass_peephole2): ...new function.
10524 (pass_split_all_insns): Convert from a global struct to a subclass of
10525 rtl_opt_pass along with...
10526 (pass_data_split_all_insns): ...new pass_data instance and...
10527 (make_pass_split_all_insns): ...new function.
10528 (pass_split_after_reload): Convert from a global struct to a subclass
10529 of rtl_opt_pass along with...
10530 (pass_data_split_after_reload): ...new pass_data instance and...
10531 (make_pass_split_after_reload): ...new function.
10532 (pass_split_before_regstack): Convert from a global struct to a
10533 subclass of rtl_opt_pass along with...
10534 (pass_data_split_before_regstack): ...new pass_data instance and...
10535 (make_pass_split_before_regstack): ...new function.
10536 (pass_split_before_sched2): Convert from a global struct to a subclass
10537 of rtl_opt_pass along with...
10538 (pass_data_split_before_sched2): ...new pass_data instance and...
10539 (make_pass_split_before_sched2): ...new function.
10540 (pass_split_for_shorten_branches): Convert from a global struct to a
10541 subclass of rtl_opt_pass along with...
10542 (pass_data_split_for_shorten_branches): ...new pass_data instance
10544 (make_pass_split_for_shorten_branches): ...new function.
10545 * ree.c (pass_ree): Convert from a global struct to a subclass of
10546 rtl_opt_pass along with...
10547 (pass_data_ree): ...new pass_data instance and...
10548 (make_pass_ree): ...new function.
10549 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10550 subclass of rtl_opt_pass along with...
10551 (pass_data_stack_regs): ...new pass_data instance and...
10552 (make_pass_stack_regs): ...new function.
10553 (pass_stack_regs_run): Convert from a global struct to a subclass of
10554 rtl_opt_pass along with...
10555 (pass_data_stack_regs_run): ...new pass_data instance and...
10556 (make_pass_stack_regs_run): ...new function.
10557 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10558 subclass of rtl_opt_pass along with...
10559 (pass_data_cprop_hardreg): ...new pass_data instance and...
10560 (make_pass_cprop_hardreg): ...new function.
10561 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10562 subclass of rtl_opt_pass along with...
10563 (pass_data_reginfo_init): ...new pass_data instance and...
10564 (make_pass_reginfo_init): ...new function.
10565 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10566 of rtl_opt_pass along with...
10567 (pass_data_regmove): ...new pass_data instance and...
10568 (make_pass_regmove): ...new function.
10569 * regrename.c (pass_regrename): Convert from a global struct to a
10570 subclass of rtl_opt_pass along with...
10571 (pass_data_regrename): ...new pass_data instance and...
10572 (make_pass_regrename): ...new function.
10573 * reorg.c (pass_delay_slots): Convert from a global struct to a
10574 subclass of rtl_opt_pass along with...
10575 (pass_data_delay_slots): ...new pass_data instance and...
10576 (make_pass_delay_slots): ...new function.
10577 (pass_machine_reorg): Convert from a global struct to a subclass of
10578 rtl_opt_pass along with...
10579 (pass_data_machine_reorg): ...new pass_data instance and...
10580 (make_pass_machine_reorg): ...new function.
10581 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10582 of rtl_opt_pass along with...
10583 (pass_data_sched): ...new pass_data instance and...
10584 (make_pass_sched): ...new function.
10585 (pass_sched2): Convert from a global struct to a subclass of
10586 rtl_opt_pass along with...
10587 (pass_data_sched2): ...new pass_data instance and...
10588 (make_pass_sched2): ...new function.
10589 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10590 to a subclass of rtl_opt_pass along with...
10591 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10592 (make_pass_stack_ptr_mod): ...new function.
10593 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10594 to a subclass of rtl_opt_pass along with...
10595 (pass_data_rtl_store_motion): ...new pass_data instance and...
10596 (make_pass_rtl_store_motion): ...new function.
10597 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10598 of gimple_opt_pass along with...
10599 (pass_data_tracer): ...new pass_data instance and...
10600 (make_pass_tracer): ...new function.
10601 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10602 to a subclass of gimple_opt_pass along with...
10603 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10604 (make_pass_diagnose_tm_blocks): ...new function.
10605 (pass_lower_tm): Convert from a global struct to a subclass of
10606 gimple_opt_pass along with...
10607 (pass_data_lower_tm): ...new pass_data instance and...
10608 (make_pass_lower_tm): ...new function.
10609 (pass_tm_init): Convert from a global struct to a subclass of
10610 gimple_opt_pass along with...
10611 (pass_data_tm_init): ...new pass_data instance and...
10612 (make_pass_tm_init): ...new function.
10613 (pass_tm_mark): Convert from a global struct to a subclass of
10614 gimple_opt_pass along with...
10615 (pass_data_tm_mark): ...new pass_data instance and...
10616 (make_pass_tm_mark): ...new function.
10617 (pass_tm_edges): Convert from a global struct to a subclass of
10618 gimple_opt_pass along with...
10619 (pass_data_tm_edges): ...new pass_data instance and...
10620 (make_pass_tm_edges): ...new function.
10621 (pass_tm_memopt): Convert from a global struct to a subclass of
10622 gimple_opt_pass along with...
10623 (pass_data_tm_memopt): ...new pass_data instance and...
10624 (make_pass_tm_memopt): ...new function.
10625 (pass_ipa_tm): Convert from a global struct to a subclass of
10626 simple_ipa_opt_pass along with...
10627 (pass_data_ipa_tm): ...new pass_data instance and...
10628 (make_pass_ipa_tm): ...new function.
10629 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10630 subclass of gimple_opt_pass along with...
10631 (pass_data_call_cdce): ...new pass_data instance and...
10632 (make_pass_call_cdce): ...new function.
10633 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10634 subclass of gimple_opt_pass along with...
10635 (pass_data_build_cfg): ...new pass_data instance and...
10636 (make_pass_build_cfg): ...new function.
10637 (pass_split_crit_edges): Convert from a global struct to a subclass of
10638 gimple_opt_pass along with...
10639 (pass_data_split_crit_edges): ...new pass_data instance and...
10640 (make_pass_split_crit_edges): ...new function.
10641 (pass_warn_function_return): Convert from a global struct to a
10642 subclass of gimple_opt_pass along with...
10643 (pass_data_warn_function_return): ...new pass_data instance and...
10644 (make_pass_warn_function_return): ...new function.
10645 (pass_warn_function_noreturn): Convert from a global struct to a
10646 subclass of gimple_opt_pass along with...
10647 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10648 (make_pass_warn_function_noreturn): ...new function.
10649 (pass_warn_unused_result): Convert from a global struct to a subclass
10650 of gimple_opt_pass along with...
10651 (pass_data_warn_unused_result): ...new pass_data instance and...
10652 (make_pass_warn_unused_result): ...new function.
10653 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10654 a subclass of gimple_opt_pass along with...
10655 (pass_data_merge_phi): ...new pass_data instance and...
10656 (make_pass_merge_phi): ...new function.
10657 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10658 a subclass of gimple_opt_pass along with...
10659 (pass_data_lower_complex): ...new pass_data instance and...
10660 (make_pass_lower_complex): ...new function.
10661 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10662 gimple_opt_pass along with...
10663 (pass_data_lower_complex_O0): ...new pass_data instance and...
10664 (make_pass_lower_complex_O0): ...new function.
10665 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10666 subclass of gimple_opt_pass along with...
10667 (pass_data_lower_eh): ...new pass_data instance and...
10668 (make_pass_lower_eh): ...new function.
10669 (pass_refactor_eh): Convert from a global struct to a subclass of
10670 gimple_opt_pass along with...
10671 (pass_data_refactor_eh): ...new pass_data instance and...
10672 (make_pass_refactor_eh): ...new function.
10673 (pass_lower_resx): Convert from a global struct to a subclass of
10674 gimple_opt_pass along with...
10675 (pass_data_lower_resx): ...new pass_data instance and...
10676 (make_pass_lower_resx): ...new function.
10677 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10678 of gimple_opt_pass along with...
10679 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10680 (make_pass_lower_eh_dispatch): ...new function.
10681 (pass_cleanup_eh): Convert from a global struct to a subclass of
10682 gimple_opt_pass along with...
10683 (pass_data_cleanup_eh): ...new pass_data instance and...
10684 (make_pass_cleanup_eh): ...new function.
10685 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10686 to a subclass of simple_ipa_opt_pass along with...
10687 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10688 (make_pass_ipa_lower_emutls): ...new function.
10689 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10690 a subclass of gimple_opt_pass along with...
10691 (pass_data_if_conversion): ...new pass_data instance and...
10692 (make_pass_if_conversion): ...new function.
10693 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10694 subclass of gimple_opt_pass along with...
10695 (pass_data_build_ssa): ...new pass_data instance and...
10696 (make_pass_build_ssa): ...new function.
10697 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10698 global struct to a subclass of gimple_opt_pass along with...
10699 (pass_data_loop_distribution): ...new pass_data instance and...
10700 (make_pass_loop_distribution): ...new function.
10701 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10702 subclass of gimple_opt_pass along with...
10703 (pass_data_mudflap_1): ...new pass_data instance and...
10704 (make_pass_mudflap_1): ...new function.
10705 (pass_mudflap_2): Convert from a global struct to a subclass of
10706 gimple_opt_pass along with...
10707 (pass_data_mudflap_2): ...new pass_data instance and...
10708 (make_pass_mudflap_2): ...new function.
10709 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10710 subclass of gimple_opt_pass along with...
10711 (pass_data_mudflap_1): ...new pass_data instance and...
10712 (make_pass_mudflap_1): ...new function.
10713 (pass_mudflap_2): Convert from a global struct to a subclass of
10714 gimple_opt_pass along with...
10715 (pass_data_mudflap_2): ...new pass_data instance and...
10716 (make_pass_mudflap_2): ...new function.
10717 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10718 gimple_opt_pass along with...
10719 (pass_data_nrv): ...new pass_data instance and...
10720 (make_pass_nrv): ...new function.
10721 (pass_return_slot): Convert from a global struct to a subclass of
10722 gimple_opt_pass along with...
10723 (pass_data_return_slot): ...new pass_data instance and...
10724 (make_pass_return_slot): ...new function.
10725 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10726 to a subclass of gimple_opt_pass along with...
10727 (pass_data_object_sizes): ...new pass_data instance and...
10728 (make_pass_object_sizes): ...new function.
10729 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10730 global struct to a subclass of gimple_opt_pass along with...
10731 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10733 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10734 (pass_fixup_cfg): Convert from a global struct to a subclass of
10735 gimple_opt_pass along with...
10736 (pass_data_fixup_cfg): ...new pass_data instance and...
10737 (make_pass_fixup_cfg): ...new function.
10738 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10739 (make_pass_mudflap_1): ...new function.
10740 (pass_mudflap_2): Replace declaration with that of...
10741 (make_pass_mudflap_2): ...new function.
10742 (pass_asan): Replace declaration with that of...
10743 (make_pass_asan): ...new function.
10744 (pass_asan_O0): Replace declaration with that of...
10745 (make_pass_asan_O0): ...new function.
10746 (pass_tsan): Replace declaration with that of...
10747 (make_pass_tsan): ...new function.
10748 (pass_tsan_O0): Replace declaration with that of...
10749 (make_pass_tsan_O0): ...new function.
10750 (pass_lower_cf): Replace declaration with that of...
10751 (make_pass_lower_cf): ...new function.
10752 (pass_refactor_eh): Replace declaration with that of...
10753 (make_pass_refactor_eh): ...new function.
10754 (pass_lower_eh): Replace declaration with that of...
10755 (make_pass_lower_eh): ...new function.
10756 (pass_lower_eh_dispatch): Replace declaration with that of...
10757 (make_pass_lower_eh_dispatch): ...new function.
10758 (pass_lower_resx): Replace declaration with that of...
10759 (make_pass_lower_resx): ...new function.
10760 (pass_build_cfg): Replace declaration with that of...
10761 (make_pass_build_cfg): ...new function.
10762 (pass_early_tree_profile): Replace declaration with that of...
10763 (make_pass_early_tree_profile): ...new function.
10764 (pass_cleanup_eh): Replace declaration with that of...
10765 (make_pass_cleanup_eh): ...new function.
10766 (pass_sra): Replace declaration with that of...
10767 (make_pass_sra): ...new function.
10768 (pass_sra_early): Replace declaration with that of...
10769 (make_pass_sra_early): ...new function.
10770 (pass_early_ipa_sra): Replace declaration with that of...
10771 (make_pass_early_ipa_sra): ...new function.
10772 (pass_tail_recursion): Replace declaration with that of...
10773 (make_pass_tail_recursion): ...new function.
10774 (pass_tail_calls): Replace declaration with that of...
10775 (make_pass_tail_calls): ...new function.
10776 (pass_tree_loop): Replace declaration with that of...
10777 (make_pass_tree_loop): ...new function.
10778 (pass_tree_loop_init): Replace declaration with that of...
10779 (make_pass_tree_loop_init): ...new function.
10780 (pass_lim): Replace declaration with that of...
10781 (make_pass_lim): ...new function.
10782 (pass_tree_unswitch): Replace declaration with that of...
10783 (make_pass_tree_unswitch): ...new function.
10784 (pass_predcom): Replace declaration with that of...
10785 (make_pass_predcom): ...new function.
10786 (pass_iv_canon): Replace declaration with that of...
10787 (make_pass_iv_canon): ...new function.
10788 (pass_scev_cprop): Replace declaration with that of...
10789 (make_pass_scev_cprop): ...new function.
10790 (pass_empty_loop): Replace declaration with that of...
10791 (make_pass_empty_loop): ...new function.
10792 (pass_record_bounds): Replace declaration with that of...
10793 (make_pass_record_bounds): ...new function.
10794 (pass_graphite): Replace declaration with that of...
10795 (make_pass_graphite): ...new function.
10796 (pass_graphite_transforms): Replace declaration with that of...
10797 (make_pass_graphite_transforms): ...new function.
10798 (pass_if_conversion): Replace declaration with that of...
10799 (make_pass_if_conversion): ...new function.
10800 (pass_loop_distribution): Replace declaration with that of...
10801 (make_pass_loop_distribution): ...new function.
10802 (pass_vectorize): Replace declaration with that of...
10803 (make_pass_vectorize): ...new function.
10804 (pass_slp_vectorize): Replace declaration with that of...
10805 (make_pass_slp_vectorize): ...new function.
10806 (pass_complete_unroll): Replace declaration with that of...
10807 (make_pass_complete_unroll): ...new function.
10808 (pass_complete_unrolli): Replace declaration with that of...
10809 (make_pass_complete_unrolli): ...new function.
10810 (pass_parallelize_loops): Replace declaration with that of...
10811 (make_pass_parallelize_loops): ...new function.
10812 (pass_loop_prefetch): Replace declaration with that of...
10813 (make_pass_loop_prefetch): ...new function.
10814 (pass_iv_optimize): Replace declaration with that of...
10815 (make_pass_iv_optimize): ...new function.
10816 (pass_tree_loop_done): Replace declaration with that of...
10817 (make_pass_tree_loop_done): ...new function.
10818 (pass_ch): Replace declaration with that of...
10819 (make_pass_ch): ...new function.
10820 (pass_ccp): Replace declaration with that of...
10821 (make_pass_ccp): ...new function.
10822 (pass_phi_only_cprop): Replace declaration with that of...
10823 (make_pass_phi_only_cprop): ...new function.
10824 (pass_build_ssa): Replace declaration with that of...
10825 (make_pass_build_ssa): ...new function.
10826 (pass_build_alias): Replace declaration with that of...
10827 (make_pass_build_alias): ...new function.
10828 (pass_build_ealias): Replace declaration with that of...
10829 (make_pass_build_ealias): ...new function.
10830 (pass_dominator): Replace declaration with that of...
10831 (make_pass_dominator): ...new function.
10832 (pass_dce): Replace declaration with that of...
10833 (make_pass_dce): ...new function.
10834 (pass_dce_loop): Replace declaration with that of...
10835 (make_pass_dce_loop): ...new function.
10836 (pass_cd_dce): Replace declaration with that of...
10837 (make_pass_cd_dce): ...new function.
10838 (pass_call_cdce): Replace declaration with that of...
10839 (make_pass_call_cdce): ...new function.
10840 (pass_merge_phi): Replace declaration with that of...
10841 (make_pass_merge_phi): ...new function.
10842 (pass_split_crit_edges): Replace declaration with that of...
10843 (make_pass_split_crit_edges): ...new function.
10844 (pass_pre): Replace declaration with that of...
10845 (make_pass_pre): ...new function.
10846 (pass_profile): Replace declaration with that of...
10847 (make_pass_profile): ...new function.
10848 (pass_strip_predict_hints): Replace declaration with that of...
10849 (make_pass_strip_predict_hints): ...new function.
10850 (pass_lower_complex_O0): Replace declaration with that of...
10851 (make_pass_lower_complex_O0): ...new function.
10852 (pass_lower_complex): Replace declaration with that of...
10853 (make_pass_lower_complex): ...new function.
10854 (pass_lower_vector): Replace declaration with that of...
10855 (make_pass_lower_vector): ...new function.
10856 (pass_lower_vector_ssa): Replace declaration with that of...
10857 (make_pass_lower_vector_ssa): ...new function.
10858 (pass_lower_omp): Replace declaration with that of...
10859 (make_pass_lower_omp): ...new function.
10860 (pass_diagnose_omp_blocks): Replace declaration with that of...
10861 (make_pass_diagnose_omp_blocks): ...new function.
10862 (pass_expand_omp): Replace declaration with that of...
10863 (make_pass_expand_omp): ...new function.
10864 (pass_expand_omp_ssa): Replace declaration with that of...
10865 (make_pass_expand_omp_ssa): ...new function.
10866 (pass_object_sizes): Replace declaration with that of...
10867 (make_pass_object_sizes): ...new function.
10868 (pass_strlen): Replace declaration with that of...
10869 (make_pass_strlen): ...new function.
10870 (pass_fold_builtins): Replace declaration with that of...
10871 (make_pass_fold_builtins): ...new function.
10872 (pass_stdarg): Replace declaration with that of...
10873 (make_pass_stdarg): ...new function.
10874 (pass_early_warn_uninitialized): Replace declaration with that of...
10875 (make_pass_early_warn_uninitialized): ...new function.
10876 (pass_late_warn_uninitialized): Replace declaration with that of...
10877 (make_pass_late_warn_uninitialized): ...new function.
10878 (pass_cse_reciprocals): Replace declaration with that of...
10879 (make_pass_cse_reciprocals): ...new function.
10880 (pass_cse_sincos): Replace declaration with that of...
10881 (make_pass_cse_sincos): ...new function.
10882 (pass_optimize_bswap): Replace declaration with that of...
10883 (make_pass_optimize_bswap): ...new function.
10884 (pass_optimize_widening_mul): Replace declaration with that of...
10885 (make_pass_optimize_widening_mul): ...new function.
10886 (pass_warn_function_return): Replace declaration with that of...
10887 (make_pass_warn_function_return): ...new function.
10888 (pass_warn_function_noreturn): Replace declaration with that of...
10889 (make_pass_warn_function_noreturn): ...new function.
10890 (pass_cselim): Replace declaration with that of...
10891 (make_pass_cselim): ...new function.
10892 (pass_phiopt): Replace declaration with that of...
10893 (make_pass_phiopt): ...new function.
10894 (pass_forwprop): Replace declaration with that of...
10895 (make_pass_forwprop): ...new function.
10896 (pass_phiprop): Replace declaration with that of...
10897 (make_pass_phiprop): ...new function.
10898 (pass_tree_ifcombine): Replace declaration with that of...
10899 (make_pass_tree_ifcombine): ...new function.
10900 (pass_dse): Replace declaration with that of...
10901 (make_pass_dse): ...new function.
10902 (pass_nrv): Replace declaration with that of...
10903 (make_pass_nrv): ...new function.
10904 (pass_rename_ssa_copies): Replace declaration with that of...
10905 (make_pass_rename_ssa_copies): ...new function.
10906 (pass_sink_code): Replace declaration with that of...
10907 (make_pass_sink_code): ...new function.
10908 (pass_fre): Replace declaration with that of...
10909 (make_pass_fre): ...new function.
10910 (pass_check_data_deps): Replace declaration with that of...
10911 (make_pass_check_data_deps): ...new function.
10912 (pass_copy_prop): Replace declaration with that of...
10913 (make_pass_copy_prop): ...new function.
10914 (pass_vrp): Replace declaration with that of...
10915 (make_pass_vrp): ...new function.
10916 (pass_uncprop): Replace declaration with that of...
10917 (make_pass_uncprop): ...new function.
10918 (pass_return_slot): Replace declaration with that of...
10919 (make_pass_return_slot): ...new function.
10920 (pass_reassoc): Replace declaration with that of...
10921 (make_pass_reassoc): ...new function.
10922 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10923 (make_pass_rebuild_cgraph_edges): ...new function.
10924 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10925 (make_pass_remove_cgraph_callee_edges): ...new function.
10926 (pass_build_cgraph_edges): Replace declaration with that of...
10927 (make_pass_build_cgraph_edges): ...new function.
10928 (pass_local_pure_const): Replace declaration with that of...
10929 (make_pass_local_pure_const): ...new function.
10930 (pass_tracer): Replace declaration with that of...
10931 (make_pass_tracer): ...new function.
10932 (pass_warn_unused_result): Replace declaration with that of...
10933 (make_pass_warn_unused_result): ...new function.
10934 (pass_diagnose_tm_blocks): Replace declaration with that of...
10935 (make_pass_diagnose_tm_blocks): ...new function.
10936 (pass_lower_tm): Replace declaration with that of...
10937 (make_pass_lower_tm): ...new function.
10938 (pass_tm_init): Replace declaration with that of...
10939 (make_pass_tm_init): ...new function.
10940 (pass_tm_mark): Replace declaration with that of...
10941 (make_pass_tm_mark): ...new function.
10942 (pass_tm_memopt): Replace declaration with that of...
10943 (make_pass_tm_memopt): ...new function.
10944 (pass_tm_edges): Replace declaration with that of...
10945 (make_pass_tm_edges): ...new function.
10946 (pass_split_functions): Replace declaration with that of...
10947 (make_pass_split_functions): ...new function.
10948 (pass_feedback_split_functions): Replace declaration with that of...
10949 (make_pass_feedback_split_functions): ...new function.
10950 (pass_strength_reduction): Replace declaration with that of...
10951 (make_pass_strength_reduction): ...new function.
10952 (pass_ipa_lower_emutls): Replace declaration with that of...
10953 (make_pass_ipa_lower_emutls): ...new function.
10954 (pass_ipa_function_and_variable_visibility): Replace declaration with
10956 (make_pass_ipa_function_and_variable_visibility): ...new function.
10957 (pass_ipa_tree_profile): Replace declaration with that of...
10958 (make_pass_ipa_tree_profile): ...new function.
10959 (pass_early_local_passes): Replace declaration with that of...
10960 (make_pass_early_local_passes): ...new function.
10961 (pass_ipa_whole_program_visibility): Replace declaration with that
10963 (make_pass_ipa_whole_program_visibility): ...new function.
10964 (pass_ipa_lto_gimple_out): Replace declaration with that of...
10965 (make_pass_ipa_lto_gimple_out): ...new function.
10966 (pass_ipa_increase_alignment): Replace declaration with that of...
10967 (make_pass_ipa_increase_alignment): ...new function.
10968 (pass_ipa_inline): Replace declaration with that of...
10969 (make_pass_ipa_inline): ...new function.
10970 (pass_ipa_free_lang_data): Replace declaration with that of...
10971 (make_pass_ipa_free_lang_data): ...new function.
10972 (pass_ipa_free_inline_summary): Replace declaration with that of...
10973 (make_pass_ipa_free_inline_summary): ...new function.
10974 (pass_ipa_cp): Replace declaration with that of...
10975 (make_pass_ipa_cp): ...new function.
10976 (pass_ipa_reference): Replace declaration with that of...
10977 (make_pass_ipa_reference): ...new function.
10978 (pass_ipa_pure_const): Replace declaration with that of...
10979 (make_pass_ipa_pure_const): ...new function.
10980 (pass_ipa_pta): Replace declaration with that of...
10981 (make_pass_ipa_pta): ...new function.
10982 (pass_ipa_lto_finish_out): Replace declaration with that of...
10983 (make_pass_ipa_lto_finish_out): ...new function.
10984 (pass_ipa_tm): Replace declaration with that of...
10985 (make_pass_ipa_tm): ...new function.
10986 (pass_ipa_profile): Replace declaration with that of...
10987 (make_pass_ipa_profile): ...new function.
10988 (pass_ipa_cdtor_merge): Replace declaration with that of...
10989 (make_pass_ipa_cdtor_merge): ...new function.
10990 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
10992 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10993 (pass_init_datastructures): Replace declaration with that of...
10994 (make_pass_init_datastructures): ...new function.
10995 (pass_fixup_cfg): Replace declaration with that of...
10996 (make_pass_fixup_cfg): ...new function.
10997 (pass_expand): Replace declaration with that of...
10998 (make_pass_expand): ...new function.
10999 (pass_instantiate_virtual_regs): Replace declaration with that of...
11000 (make_pass_instantiate_virtual_regs): ...new function.
11001 (pass_rtl_fwprop): Replace declaration with that of...
11002 (make_pass_rtl_fwprop): ...new function.
11003 (pass_rtl_fwprop_addr): Replace declaration with that of...
11004 (make_pass_rtl_fwprop_addr): ...new function.
11005 (pass_jump): Replace declaration with that of...
11006 (make_pass_jump): ...new function.
11007 (pass_jump2): Replace declaration with that of...
11008 (make_pass_jump2): ...new function.
11009 (pass_lower_subreg): Replace declaration with that of...
11010 (make_pass_lower_subreg): ...new function.
11011 (pass_cse): Replace declaration with that of...
11012 (make_pass_cse): ...new function.
11013 (pass_fast_rtl_dce): Replace declaration with that of...
11014 (make_pass_fast_rtl_dce): ...new function.
11015 (pass_ud_rtl_dce): Replace declaration with that of...
11016 (make_pass_ud_rtl_dce): ...new function.
11017 (pass_rtl_dce): Replace declaration with that of...
11018 (make_pass_rtl_dce): ...new function.
11019 (pass_rtl_dse1): Replace declaration with that of...
11020 (make_pass_rtl_dse1): ...new function.
11021 (pass_rtl_dse2): Replace declaration with that of...
11022 (make_pass_rtl_dse2): ...new function.
11023 (pass_rtl_dse3): Replace declaration with that of...
11024 (make_pass_rtl_dse3): ...new function.
11025 (pass_rtl_cprop): Replace declaration with that of...
11026 (make_pass_rtl_cprop): ...new function.
11027 (pass_rtl_pre): Replace declaration with that of...
11028 (make_pass_rtl_pre): ...new function.
11029 (pass_rtl_hoist): Replace declaration with that of...
11030 (make_pass_rtl_hoist): ...new function.
11031 (pass_rtl_store_motion): Replace declaration with that of...
11032 (make_pass_rtl_store_motion): ...new function.
11033 (pass_cse_after_global_opts): Replace declaration with that of...
11034 (make_pass_cse_after_global_opts): ...new function.
11035 (pass_rtl_ifcvt): Replace declaration with that of...
11036 (make_pass_rtl_ifcvt): ...new function.
11037 (pass_into_cfg_layout_mode): Replace declaration with that of...
11038 (make_pass_into_cfg_layout_mode): ...new function.
11039 (pass_outof_cfg_layout_mode): Replace declaration with that of...
11040 (make_pass_outof_cfg_layout_mode): ...new function.
11041 (pass_loop2): Replace declaration with that of...
11042 (make_pass_loop2): ...new function.
11043 (pass_rtl_loop_init): Replace declaration with that of...
11044 (make_pass_rtl_loop_init): ...new function.
11045 (pass_rtl_move_loop_invariants): Replace declaration with that of...
11046 (make_pass_rtl_move_loop_invariants): ...new function.
11047 (pass_rtl_unswitch): Replace declaration with that of...
11048 (make_pass_rtl_unswitch): ...new function.
11049 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
11050 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11051 (pass_rtl_doloop): Replace declaration with that of...
11052 (make_pass_rtl_doloop): ...new function.
11053 (pass_rtl_loop_done): Replace declaration with that of...
11054 (make_pass_rtl_loop_done): ...new function.
11055 (pass_web): Replace declaration with that of...
11056 (make_pass_web): ...new function.
11057 (pass_cse2): Replace declaration with that of...
11058 (make_pass_cse2): ...new function.
11059 (pass_df_initialize_opt): Replace declaration with that of...
11060 (make_pass_df_initialize_opt): ...new function.
11061 (pass_df_initialize_no_opt): Replace declaration with that of...
11062 (make_pass_df_initialize_no_opt): ...new function.
11063 (pass_reginfo_init): Replace declaration with that of...
11064 (make_pass_reginfo_init): ...new function.
11065 (pass_inc_dec): Replace declaration with that of...
11066 (make_pass_inc_dec): ...new function.
11067 (pass_stack_ptr_mod): Replace declaration with that of...
11068 (make_pass_stack_ptr_mod): ...new function.
11069 (pass_initialize_regs): Replace declaration with that of...
11070 (make_pass_initialize_regs): ...new function.
11071 (pass_combine): Replace declaration with that of...
11072 (make_pass_combine): ...new function.
11073 (pass_if_after_combine): Replace declaration with that of...
11074 (make_pass_if_after_combine): ...new function.
11075 (pass_ree): Replace declaration with that of...
11076 (make_pass_ree): ...new function.
11077 (pass_partition_blocks): Replace declaration with that of...
11078 (make_pass_partition_blocks): ...new function.
11079 (pass_match_asm_constraints): Replace declaration with that of...
11080 (make_pass_match_asm_constraints): ...new function.
11081 (pass_regmove): Replace declaration with that of...
11082 (make_pass_regmove): ...new function.
11083 (pass_split_all_insns): Replace declaration with that of...
11084 (make_pass_split_all_insns): ...new function.
11085 (pass_fast_rtl_byte_dce): Replace declaration with that of...
11086 (make_pass_fast_rtl_byte_dce): ...new function.
11087 (pass_lower_subreg2): Replace declaration with that of...
11088 (make_pass_lower_subreg2): ...new function.
11089 (pass_mode_switching): Replace declaration with that of...
11090 (make_pass_mode_switching): ...new function.
11091 (pass_sms): Replace declaration with that of...
11092 (make_pass_sms): ...new function.
11093 (pass_sched): Replace declaration with that of...
11094 (make_pass_sched): ...new function.
11095 (pass_ira): Replace declaration with that of...
11096 (make_pass_ira): ...new function.
11097 (pass_reload): Replace declaration with that of...
11098 (make_pass_reload): ...new function.
11099 (pass_clean_state): Replace declaration with that of...
11100 (make_pass_clean_state): ...new function.
11101 (pass_branch_prob): Replace declaration with that of...
11102 (make_pass_branch_prob): ...new function.
11103 (pass_value_profile_transformations): Replace declaration with that
11105 (make_pass_value_profile_transformations): ...new function.
11106 (pass_postreload_cse): Replace declaration with that of...
11107 (make_pass_postreload_cse): ...new function.
11108 (pass_gcse2): Replace declaration with that of...
11109 (make_pass_gcse2): ...new function.
11110 (pass_split_after_reload): Replace declaration with that of...
11111 (make_pass_split_after_reload): ...new function.
11112 (pass_branch_target_load_optimize1): Replace declaration with that
11114 (make_pass_branch_target_load_optimize1): ...new function.
11115 (pass_thread_prologue_and_epilogue): Replace declaration with that
11117 (make_pass_thread_prologue_and_epilogue): ...new function.
11118 (pass_stack_adjustments): Replace declaration with that of...
11119 (make_pass_stack_adjustments): ...new function.
11120 (pass_peephole2): Replace declaration with that of...
11121 (make_pass_peephole2): ...new function.
11122 (pass_if_after_reload): Replace declaration with that of...
11123 (make_pass_if_after_reload): ...new function.
11124 (pass_regrename): Replace declaration with that of...
11125 (make_pass_regrename): ...new function.
11126 (pass_cprop_hardreg): Replace declaration with that of...
11127 (make_pass_cprop_hardreg): ...new function.
11128 (pass_reorder_blocks): Replace declaration with that of...
11129 (make_pass_reorder_blocks): ...new function.
11130 (pass_branch_target_load_optimize2): Replace declaration with that
11132 (make_pass_branch_target_load_optimize2): ...new function.
11133 (pass_leaf_regs): Replace declaration with that of...
11134 (make_pass_leaf_regs): ...new function.
11135 (pass_split_before_sched2): Replace declaration with that of...
11136 (make_pass_split_before_sched2): ...new function.
11137 (pass_compare_elim_after_reload): Replace declaration with that of...
11138 (make_pass_compare_elim_after_reload): ...new function.
11139 (pass_sched2): Replace declaration with that of...
11140 (make_pass_sched2): ...new function.
11141 (pass_stack_regs): Replace declaration with that of...
11142 (make_pass_stack_regs): ...new function.
11143 (pass_stack_regs_run): Replace declaration with that of...
11144 (make_pass_stack_regs_run): ...new function.
11145 (pass_df_finish): Replace declaration with that of...
11146 (make_pass_df_finish): ...new function.
11147 (pass_compute_alignments): Replace declaration with that of...
11148 (make_pass_compute_alignments): ...new function.
11149 (pass_duplicate_computed_gotos): Replace declaration with that of...
11150 (make_pass_duplicate_computed_gotos): ...new function.
11151 (pass_variable_tracking): Replace declaration with that of...
11152 (make_pass_variable_tracking): ...new function.
11153 (pass_free_cfg): Replace declaration with that of...
11154 (make_pass_free_cfg): ...new function.
11155 (pass_machine_reorg): Replace declaration with that of...
11156 (make_pass_machine_reorg): ...new function.
11157 (pass_cleanup_barriers): Replace declaration with that of...
11158 (make_pass_cleanup_barriers): ...new function.
11159 (pass_delay_slots): Replace declaration with that of...
11160 (make_pass_delay_slots): ...new function.
11161 (pass_split_for_shorten_branches): Replace declaration with that of...
11162 (make_pass_split_for_shorten_branches): ...new function.
11163 (pass_split_before_regstack): Replace declaration with that of...
11164 (make_pass_split_before_regstack): ...new function.
11165 (pass_convert_to_eh_region_ranges): Replace declaration with that
11167 (make_pass_convert_to_eh_region_ranges): ...new function.
11168 (pass_shorten_branches): Replace declaration with that of...
11169 (make_pass_shorten_branches): ...new function.
11170 (pass_set_nothrow_function_flags): Replace declaration with that of...
11171 (make_pass_set_nothrow_function_flags): ...new function.
11172 (pass_dwarf2_frame): Replace declaration with that of...
11173 (make_pass_dwarf2_frame): ...new function.
11174 (pass_final): Replace declaration with that of...
11175 (make_pass_final): ...new function.
11176 (pass_rtl_seqabstr): Replace declaration with that of...
11177 (make_pass_rtl_seqabstr): ...new function.
11178 (pass_release_ssa_names): Replace declaration with that of...
11179 (make_pass_release_ssa_names): ...new function.
11180 (pass_early_inline): Replace declaration with that of...
11181 (make_pass_early_inline): ...new function.
11182 (pass_inline_parameters): Replace declaration with that of...
11183 (make_pass_inline_parameters): ...new function.
11184 (pass_update_address_taken): Replace declaration with that of...
11185 (make_pass_update_address_taken): ...new function.
11186 (pass_convert_switch): Replace declaration with that of...
11187 (make_pass_convert_switch): ...new function.
11188 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11189 to a subclass of simple_ipa_opt_pass along with...
11190 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11191 (make_pass_ipa_tree_profile): ...new function.
11192 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11193 subclass of gimple_opt_pass along with...
11194 (pass_data_sra_early): ...new pass_data instance and...
11195 (make_pass_sra_early): ...new function.
11196 (pass_sra): Convert from a global struct to a subclass of
11197 gimple_opt_pass along with...
11198 (pass_data_sra): ...new pass_data instance and...
11199 (make_pass_sra): ...new function.
11200 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11201 gimple_opt_pass along with...
11202 (pass_data_early_ipa_sra): ...new pass_data instance and...
11203 (make_pass_early_ipa_sra): ...new function.
11204 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11205 subclass of gimple_opt_pass along with...
11206 (pass_data_ccp): ...new pass_data instance and...
11207 (make_pass_ccp): ...new function.
11208 (pass_fold_builtins): Convert from a global struct to a subclass of
11209 gimple_opt_pass along with...
11210 (pass_data_fold_builtins): ...new pass_data instance and...
11211 (make_pass_fold_builtins): ...new function.
11212 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11213 subclass of gimple_opt_pass along with...
11214 (pass_data_copy_prop): ...new pass_data instance and...
11215 (make_pass_copy_prop): ...new function.
11216 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11217 global struct to a subclass of gimple_opt_pass along with...
11218 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11219 (make_pass_rename_ssa_copies): ...new function.
11220 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11221 subclass of gimple_opt_pass along with...
11222 (pass_data_dce): ...new pass_data instance and...
11223 (make_pass_dce): ...new function.
11224 (pass_dce_loop): Convert from a global struct to a subclass of
11225 gimple_opt_pass along with...
11226 (pass_data_dce_loop): ...new pass_data instance and...
11227 (make_pass_dce_loop): ...new function.
11228 (pass_cd_dce): Convert from a global struct to a subclass of
11229 gimple_opt_pass along with...
11230 (pass_data_cd_dce): ...new pass_data instance and...
11231 (make_pass_cd_dce): ...new function.
11232 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11233 subclass of gimple_opt_pass along with...
11234 (pass_data_dominator): ...new pass_data instance and...
11235 (make_pass_dominator): ...new function.
11236 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11237 gimple_opt_pass along with...
11238 (pass_data_phi_only_cprop): ...new pass_data instance and...
11239 (make_pass_phi_only_cprop): ...new function.
11240 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11241 subclass of gimple_opt_pass along with...
11242 (pass_data_dse): ...new pass_data instance and...
11243 (make_pass_dse): ...new function.
11244 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11245 a subclass of gimple_opt_pass along with...
11246 (pass_data_forwprop): ...new pass_data instance and...
11247 (make_pass_forwprop): ...new function.
11248 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11249 struct to a subclass of gimple_opt_pass along with...
11250 (pass_data_tree_ifcombine): ...new pass_data instance and...
11251 (make_pass_tree_ifcombine): ...new function.
11252 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11253 subclass of gimple_opt_pass along with...
11254 (pass_data_ch): ...new pass_data instance and...
11255 (make_pass_ch): ...new function.
11256 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11257 subclass of gimple_opt_pass along with...
11258 (pass_data_tree_loop): ...new pass_data instance and...
11259 (make_pass_tree_loop): ...new function.
11260 (pass_tree_loop_init): Convert from a global struct to a subclass of
11261 gimple_opt_pass along with...
11262 (pass_data_tree_loop_init): ...new pass_data instance and...
11263 (make_pass_tree_loop_init): ...new function.
11264 (pass_lim): Convert from a global struct to a subclass of
11265 gimple_opt_pass along with...
11266 (pass_data_lim): ...new pass_data instance and...
11267 (make_pass_lim): ...new function.
11268 (pass_tree_unswitch): Convert from a global struct to a subclass of
11269 gimple_opt_pass along with...
11270 (pass_data_tree_unswitch): ...new pass_data instance and...
11271 (make_pass_tree_unswitch): ...new function.
11272 (pass_predcom): Convert from a global struct to a subclass of
11273 gimple_opt_pass along with...
11274 (pass_data_predcom): ...new pass_data instance and...
11275 (make_pass_predcom): ...new function.
11276 (pass_vectorize): Convert from a global struct to a subclass of
11277 gimple_opt_pass along with...
11278 (pass_data_vectorize): ...new pass_data instance and...
11279 (make_pass_vectorize): ...new function.
11280 (pass_graphite): Convert from a global struct to a subclass of
11281 gimple_opt_pass along with...
11282 (pass_data_graphite): ...new pass_data instance and...
11283 (make_pass_graphite): ...new function.
11284 (pass_graphite_transforms): Convert from a global struct to a subclass
11285 of gimple_opt_pass along with...
11286 (pass_data_graphite_transforms): ...new pass_data instance and...
11287 (make_pass_graphite_transforms): ...new function.
11288 (pass_check_data_deps): Convert from a global struct to a subclass of
11289 gimple_opt_pass along with...
11290 (pass_data_check_data_deps): ...new pass_data instance and...
11291 (make_pass_check_data_deps): ...new function.
11292 (pass_iv_canon): Convert from a global struct to a subclass of
11293 gimple_opt_pass along with...
11294 (pass_data_iv_canon): ...new pass_data instance and...
11295 (make_pass_iv_canon): ...new function.
11296 (pass_scev_cprop): Convert from a global struct to a subclass of
11297 gimple_opt_pass along with...
11298 (pass_data_scev_cprop): ...new pass_data instance and...
11299 (make_pass_scev_cprop): ...new function.
11300 (pass_record_bounds): Convert from a global struct to a subclass of
11301 gimple_opt_pass along with...
11302 (pass_data_record_bounds): ...new pass_data instance and...
11303 (make_pass_record_bounds): ...new function.
11304 (pass_complete_unroll): Convert from a global struct to a subclass of
11305 gimple_opt_pass along with...
11306 (pass_data_complete_unroll): ...new pass_data instance and...
11307 (make_pass_complete_unroll): ...new function.
11308 (pass_complete_unrolli): Convert from a global struct to a subclass of
11309 gimple_opt_pass along with...
11310 (pass_data_complete_unrolli): ...new pass_data instance and...
11311 (make_pass_complete_unrolli): ...new function.
11312 (pass_parallelize_loops): Convert from a global struct to a subclass
11313 of gimple_opt_pass along with...
11314 (pass_data_parallelize_loops): ...new pass_data instance and...
11315 (make_pass_parallelize_loops): ...new function.
11316 (pass_loop_prefetch): Convert from a global struct to a subclass of
11317 gimple_opt_pass along with...
11318 (pass_data_loop_prefetch): ...new pass_data instance and...
11319 (make_pass_loop_prefetch): ...new function.
11320 (pass_iv_optimize): Convert from a global struct to a subclass of
11321 gimple_opt_pass along with...
11322 (pass_data_iv_optimize): ...new pass_data instance and...
11323 (make_pass_iv_optimize): ...new function.
11324 (pass_tree_loop_done): Convert from a global struct to a subclass of
11325 gimple_opt_pass along with...
11326 (pass_data_tree_loop_done): ...new pass_data instance and...
11327 (make_pass_tree_loop_done): ...new function.
11328 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11329 struct to a subclass of gimple_opt_pass along with...
11330 (pass_data_cse_reciprocals): ...new pass_data instance and...
11331 (make_pass_cse_reciprocals): ...new function.
11332 (pass_cse_sincos): Convert from a global struct to a subclass of
11333 gimple_opt_pass along with...
11334 (pass_data_cse_sincos): ...new pass_data instance and...
11335 (make_pass_cse_sincos): ...new function.
11336 (pass_optimize_bswap): Convert from a global struct to a subclass of
11337 gimple_opt_pass along with...
11338 (pass_data_optimize_bswap): ...new pass_data instance and...
11339 (make_pass_optimize_bswap): ...new function.
11340 (pass_optimize_widening_mul): Convert from a global struct to a
11341 subclass of gimple_opt_pass along with...
11342 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11343 (make_pass_optimize_widening_mul): ...new function.
11344 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11345 subclass of gimple_opt_pass along with...
11346 (pass_data_phiopt): ...new pass_data instance and...
11347 (make_pass_phiopt): ...new function.
11348 (pass_cselim): Convert from a global struct to a subclass of
11349 gimple_opt_pass along with...
11350 (pass_data_cselim): ...new pass_data instance and...
11351 (make_pass_cselim): ...new function.
11352 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11353 subclass of gimple_opt_pass along with...
11354 (pass_data_phiprop): ...new pass_data instance and...
11355 (make_pass_phiprop): ...new function.
11356 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11357 subclass of gimple_opt_pass along with...
11358 (pass_data_pre): ...new pass_data instance and...
11359 (make_pass_pre): ...new function.
11360 (pass_fre): Convert from a global struct to a subclass of
11361 gimple_opt_pass along with...
11362 (pass_data_fre): ...new pass_data instance and...
11363 (make_pass_fre): ...new function.
11364 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11365 subclass of gimple_opt_pass along with...
11366 (pass_data_reassoc): ...new pass_data instance and...
11367 (make_pass_reassoc): ...new function.
11368 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11369 subclass of gimple_opt_pass along with...
11370 (pass_data_sink_code): ...new pass_data instance and...
11371 (make_pass_sink_code): ...new function.
11372 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11373 subclass of gimple_opt_pass along with...
11374 (pass_data_strlen): ...new pass_data instance and...
11375 (make_pass_strlen): ...new function.
11376 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11377 struct to a subclass of gimple_opt_pass along with...
11378 (pass_data_build_alias): ...new pass_data instance and...
11379 (make_pass_build_alias): ...new function.
11380 (pass_build_ealias): Convert from a global struct to a subclass of
11381 gimple_opt_pass along with...
11382 (pass_data_build_ealias): ...new pass_data instance and...
11383 (make_pass_build_ealias): ...new function.
11384 (pass_ipa_pta): Convert from a global struct to a subclass of
11385 simple_ipa_opt_pass along with...
11386 (pass_data_ipa_pta): ...new pass_data instance and...
11387 (make_pass_ipa_pta): ...new function.
11388 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11389 subclass of gimple_opt_pass along with...
11390 (pass_data_uncprop): ...new pass_data instance and...
11391 (make_pass_uncprop): ...new function.
11392 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11393 global struct to a subclass of gimple_opt_pass along with...
11394 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11395 (make_pass_late_warn_uninitialized): ...new function.
11396 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11397 to a subclass of gimple_opt_pass along with...
11398 (pass_data_init_datastructures): ...new pass_data instance and...
11399 (make_pass_init_datastructures): ...new function.
11400 (pass_early_warn_uninitialized): Convert from a global struct to a
11401 subclass of gimple_opt_pass along with...
11402 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11403 (make_pass_early_warn_uninitialized): ...new function.
11404 (pass_update_address_taken): Convert from a global struct to a
11405 subclass of gimple_opt_pass along with...
11406 (pass_data_update_address_taken): ...new pass_data instance and...
11407 (make_pass_update_address_taken): ...new function.
11408 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11409 struct to a subclass of gimple_opt_pass along with...
11410 (pass_data_release_ssa_names): ...new pass_data instance and...
11411 (make_pass_release_ssa_names): ...new function.
11412 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11413 subclass of gimple_opt_pass along with...
11414 (pass_data_stdarg): ...new pass_data instance and...
11415 (make_pass_stdarg): ...new function.
11416 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11417 global struct to a subclass of gimple_opt_pass along with...
11418 (pass_data_convert_switch): ...new pass_data instance and...
11419 (make_pass_convert_switch): ...new function.
11420 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11421 to a subclass of gimple_opt_pass along with...
11422 (pass_data_tail_recursion): ...new pass_data instance and...
11423 (make_pass_tail_recursion): ...new function.
11424 (pass_tail_calls): Convert from a global struct to a subclass of
11425 gimple_opt_pass along with...
11426 (pass_data_tail_calls): ...new pass_data instance and...
11427 (make_pass_tail_calls): ...new function.
11428 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11429 struct to a subclass of gimple_opt_pass along with...
11430 (pass_data_lower_vector): ...new pass_data instance and...
11431 (make_pass_lower_vector): ...new function.
11432 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11433 gimple_opt_pass along with...
11434 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11435 (make_pass_lower_vector_ssa): ...new function.
11436 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11437 to a subclass of gimple_opt_pass along with...
11438 (pass_data_slp_vectorize): ...new pass_data instance and...
11439 (make_pass_slp_vectorize): ...new function.
11440 (pass_ipa_increase_alignment): Convert from a global struct to a
11441 subclass of simple_ipa_opt_pass along with...
11442 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11443 (make_pass_ipa_increase_alignment): ...new function.
11444 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11445 gimple_opt_pass along with...
11446 (pass_data_vrp): ...new pass_data instance and...
11447 (make_pass_vrp): ...new function.
11448 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11449 subclass of simple_ipa_opt_pass along with...
11450 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11451 (make_pass_ipa_free_lang_data): ...new function.
11452 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11453 gimple_opt_pass along with...
11454 (pass_data_tsan): ...new pass_data instance and...
11455 (make_pass_tsan): ...new function.
11456 (pass_tsan_O0): Convert from a global struct to a subclass of
11457 gimple_opt_pass along with...
11458 (pass_data_tsan_O0): ...new pass_data instance and...
11459 (make_pass_tsan_O0): ...new function.
11460 * var-tracking.c (pass_variable_tracking): Convert from a global
11461 struct to a subclass of rtl_opt_pass along with...
11462 (pass_data_variable_tracking): ...new pass_data instance and...
11463 (make_pass_variable_tracking): ...new function.
11464 * web.c (pass_web): Convert from a global struct to a subclass of
11465 rtl_opt_pass along with...
11466 (pass_data_web): ...new pass_data instance and...
11467 (make_pass_web): ...new function.
11468 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11469 declaration with that of...
11470 (make_pass_mode_switch_use): ...new function.
11471 (pass_resolve_sw_modes): Replace declaration with that of...
11472 (make_pass_resolve_sw_modes): ...new function.
11473 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11474 from a global struct to a subclass of rtl_opt_pass along with...
11475 (pass_data_mode_switch_use): ...new pass_data instance and...
11476 (make_pass_mode_switch_use): ...new function.
11477 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11478 from a global struct to a subclass of rtl_opt_pass along with...
11479 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11480 (make_pass_resolve_sw_modes): ...new function.
11481 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11482 struct to a subclass of rtl_opt_pass along with...
11483 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11484 (make_pass_insert_vzeroupper): ...new function.
11485 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11486 global struct to a subclass of rtl_opt_pass along with...
11487 (pass_data_work_around_errata): ...new pass_data instance and...
11488 (make_pass_work_around_errata): ...new function.
11489 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11490 struct to a subclass of rtl_opt_pass along with...
11491 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11492 (make_pass_mips_machine_reorg2): ...new function.
11494 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11496 * passes.c (pass_manager::operator new): New.
11498 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11500 Handwritten part of conversion of passes to C++ classes.
11502 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11503 (toplev.o): Add dep on PASS_MANAGER_H.
11504 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11505 of early local pases to reflect this moving from a global to a
11506 member of gcc::pass_manager.
11507 (cgraph_add_new_function): Likewise.
11508 * lto-cgraph.c (lto_output_node): Update for conversion of
11509 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11510 * passes.c (opt_pass::clone): New.
11511 (opt_pass::gate): New.
11512 (opt_pass::execute): New.
11513 (opt_pass::opt_pass): New.
11514 (pass_manager::execute_early_local_passes): New.
11515 (pass_manager::execute_pass_mode_switching): new.
11516 (finish_optimization_passes): Convert to...
11517 (pass_manager::finish_optimization_passes): ...this.
11518 (finish_optimization_passes): Update for conversion of passes to
11520 (register_dump_files_1): Use has_gate since we cannot portably
11521 check a vtable entry against NULL.
11522 (dump_one_pass): Likewise.
11523 (ipa_write_summaries_2): Likewise.
11524 (ipa_write_optimization_summaries_1): Likewise.
11525 (ipa_read_summaries_1): Likewise.
11526 (ipa_read_optimization_summaries_1): Likewise.
11527 (execute_ipa_stmt_fixups): Likewise.
11528 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11529 pass-creation functions rather than wiring up globals, and
11530 storing the results in fields of pass_manager generated using
11531 pass-instances.def.
11532 (pass_manager::dump_profile_report): Update for conversion of
11533 passes to C++ classes.
11534 (pass_manager::execute_ipa_summary_passes): Likewise.
11535 (execute_one_ipa_transform_pass): Likewise.
11536 (execute_one_pass): Use has_gate and has_execute since we cannot
11537 portably check a vtable entry against NULL.
11538 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11539 (pass_manager): Use pass-instances.def to add fields for the
11540 various pass instances.
11541 * toplev.c (finalize): Update for move of
11542 finish_optimization_passes to a method of gcc::pass_manager.
11543 * toplev.h (finish_optimization_passes): Move to method of class
11545 * tree-pass.h (struct pass_data): New.
11546 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11547 (opt_pass::gate): Convert to virtual function.
11548 (opt_pass::~opt_pass): New.
11549 (opt_pass::clone): New.
11550 (opt_pass::execute): Convert to virtual function.
11551 (opt_pass::opt_pass): New.
11552 (opt_pass::ctxt_): new.
11553 (gimple_opt_pass): Convert to subclass of opt_pass.
11554 (gimple_opt_pass::gimple_opt_pass): New.
11555 (rtl_opt_pass): Convert to subclass of opt_pass.
11556 (rtl_opt_pass::rtl_opt_pass): New.
11557 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11558 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11559 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11560 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11561 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11562 invocation of pass_mode_switching to reflect this moving from a
11563 global to a member of gcc::pass_manager.
11564 (ix86_option_override): Rework how pass_insert_vzeroupper is
11565 added to the pass_manager to reflect autogenerated changes.
11566 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11568 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11570 PR rtl-optimization/57708
11571 * recog.c (peep2_find_free_register): Validate all regs in a
11574 2013-08-05 Jan Hubicka <jh@suse.cz>
11577 * cgraph.c (verify_cgraph_node): Accept local flags from other
11579 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11580 (function_and_variable_visibility): Likewise.
11581 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11583 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11585 * graph.c (init_graph_slim_pretty_print): Remove.
11586 (print_graph_cfg): Do not call it. Use local pretty printer.
11587 (start_graph_dump): Likewise.
11589 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11591 * gimple-pretty-print.c (buffer): Remove.
11592 (initialized): Likewise.
11593 (maybe_init_pretty_print): Likewise.
11594 (print_gimple_stmt): Do not call it. Use non-static local
11595 pretty_printer variable.
11596 (print_gimple_expr): Likewise.
11597 (print_gimple_seq): Likewise.
11598 (gimple_dump_bb): Likewise.
11600 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11602 * asan.c (asan_pp): Remove.
11603 (asan_pp_initialized): Likewise.
11604 (asan_pp_initialize): Likewise.
11605 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11606 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11607 (asan_add_global): Likewise.
11609 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11611 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11612 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11613 * pretty-print.c (pp_formatted_text_data): Likewise.
11614 (pp_indent): Rename from pp_base_indent.
11615 (pp_format): Rename from pp_base_format.
11616 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11617 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11618 (pp_flush): Rename from pp_base_flush.
11619 (pp_set_line_maximum_length): Rename from
11620 pp_base_set_line_maximum_length.
11621 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11622 (pp_set_prefix): Rename from pp_base_set_prefix.
11623 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11624 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11625 (pp_append_text): Rename from pp_base_append_text.
11626 (pp_formatted_text): Rename from pp_base_formatted_text.
11627 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11628 (pp_remaining_character_count_for_line): Rename from
11629 pp_base_remaining_character_count_for_line.
11630 (pp_newline): Rename from pp_base_newline.
11631 (pp_character): Rename from pp_base_character.
11632 (pp_string): Rename from pp_base_string.
11633 (pp_maybe_space): Rename from pp_base_maybe_space.
11634 * asan.c (asan_pp_string): Adjust.
11635 (asan_emit_stack_protection): Likewise.
11636 (asan_add_global): Likewise.
11637 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11638 * tree-mudflap.c (mf_varname_tree): Likewise.
11639 * tree-pretty-print.c (pp_tree_identifier): Rename from
11640 pp_base_tree_identifier.
11641 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11642 Declare as function.
11644 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11646 * pretty-print.h (pp_bar_bar): New.
11647 (pp_ampersand_ampersand): Likewise.
11648 (pp_less_equal): Likewise.
11649 (pp_greater_equal): Likewise.
11650 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11651 printer functions instead of pp_string or operators and punctuators.
11652 (dump_gimple_call): Likewise.
11653 (dump_gimple_omp_for): Likewise.
11654 (dump_gimple_transaction): Likewise.
11655 (dump_gimple_phi): Likewise.
11656 (pp_gimple_stmt_1): Likewise.
11657 * sched-vis.c (print_insn): Likewise.
11658 * tree-mudflap.c (mf_varname_tree): Likewise.
11659 * tree-pretty-print.c (dump_block_node): Likewise.
11660 (dump_generic_node): Likewise.
11662 2013-08-02 Jan Hubicka <jh@suse.cz>
11664 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11666 * lto-streamer-out.c (tree_is_indexable): Results decls and
11667 parm decls are not indexable.
11668 (DFS_write_tree_body): Do not follow args and results.
11669 (hash_tree): Likewise.
11670 (output_functions): Rearrange so struct function is needed
11671 only when real body is output; be able to also ouptut abstract
11672 functions; output DECL_ARGUMENTS and DECL_RESULT.
11673 (lto_output): When not in WPA, ale store abstract functions.
11674 (write_symbol): Do not care about RESULT_DECL.
11675 (output_symbol_p): Handle correctly sbtract decls.
11676 * lto-streamer-in.c (input_function): Rearrange so struct
11677 function can be NULL at entry; allow streaming of
11678 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11679 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11680 sanity check during LTO.
11681 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11682 RESULT_DECl and DECL_ARGUMENTS.
11683 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11686 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11688 * pretty-print.h (pp_underscore): New.
11690 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11691 printer functions instead of pp_character.
11692 (dump_binary_rhs): Likewise.
11693 (dump_ternary_rhs): Likewise.
11694 (dump_gimple_call_args): Likewise.
11695 (pp_points_to_solution): Likewise.
11696 (dump_gimple_call): Likewise.
11697 (dump_gimple_switch): Likewise.
11698 (dump_gimple_cond): Likewise.
11699 (dump_gimple_bind): Likewise.
11700 (dump_gimple_try): Likewise.
11701 (dump_gimple_omp_for): Likewise.
11702 (dump_gimple_omp_continue): Likewise.
11703 (dump_gimple_omp_single): Likewise.
11704 (dump_gimple_omp_sections): Likewise.
11705 (dump_gimple_omp_block): Likewise.
11706 (dump_gimple_omp_critical): Likewise.
11707 (dump_gimple_transaction): Likewise.
11708 (dump_gimple_asm): Likewise.
11709 (dump_gimple_phi): Likewise.
11710 (dump_gimple_omp_parallel): Likewise.
11711 (dump_gimple_omp_task): Likewise.
11712 (dump_gimple_omp_atomic_load): Likewise.
11713 (dump_gimple_omp_atomic_store): Likewise.
11714 (dump_gimple_mem_ops): Likewise.
11715 (pp_gimple_stmt_1): Likewise.
11716 (pp_cfg_jump): Likewise.
11717 (dump_implicit_edges): Likewise.
11718 (gimple_dump_bb_for_graph): Likewise.
11719 * graph.c (draw_cfg_node): Likewise.
11720 * langhooks.c (lhd_print_error_function): Likewise.
11721 * sched-vis.c (print_exp): Likewise.
11722 (print_value): Likewise.
11723 (print_pattern): Likewise.
11724 (print_insn): Likewise.
11725 (rtl_dump_bb_for_graph): Likewise.
11726 * tree-pretty-print.c (dump_function_declaration): Likewise.
11727 (dump_array_domain): Likewise.
11728 (dump_omp_clause): Likewise.
11729 (dump_location): Likewise.
11730 (dump_generic_node): Likewise.
11731 (print_struct_decl): Likewise.
11732 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11734 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11736 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11737 candidate table when replacing a candidate statement.
11738 (replace_rhs_if_not_dup): Likewise.
11739 (replace_one_candidate): Likewise.
11741 2013-08-02 Jan Hubicka <jh@suse.cz>
11742 Martin Liska <marxin.liska@gmail.com>
11744 * cgraphunit.c (add_new_function): Fix logic when adding from
11746 (assemble_thunk): Rename to ...
11747 (expand_thunk); .. this one; export; get it working with
11748 general functions; make produced gimple valid.
11749 * cgraph.h (expand_thunk): Declare.
11751 2013-08-02 Jan Hubicka <jh@suse.cz>
11753 * ipa-cp.c (gather_context_independent_values): Use
11754 ipa_get_param_move_cost.
11755 (get_replacement_map): Remove PARAM; move parameter folding
11757 (create_specialized_node): Update.
11758 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11759 assert that we have gimple body; update move_cost.
11760 (count_formal_params): Assert that we have gimple body.
11761 (ipa_dump_param): New function.
11762 (ipa_alloc_node_params): Break out from ...
11763 (ipa_initialize_node_params): ... here.
11764 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11765 (ipa_write_node_info): Stream move costs.
11766 (ipa_read_node_info): Read move costs.
11767 (ipa_update_after_lto_read): Do not recompute node params.
11768 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11769 (ipa_get_param): Check we are not in WPA.
11770 (ipa_get_param_move_cost): New.
11771 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11772 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11773 parm numbers to be present.
11775 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11777 PR rtl-optimization/58048
11778 * lra-constraints.c (process_alt_operands): Don't check asm
11779 operand on register.
11781 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11783 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11784 the implied StoreLoad barrier for atomic operations if before.
11786 2013-08-02 Jan Hubicka <jh@suse.cz>
11787 Martin Liska <marxin.liska@gmail.com>
11789 * cgraph.c (cgraph_function_body_availability): Do not check
11791 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11792 symtab_node_availability): Declare.
11793 * ipa.c (can_replace_by_local_alias): New.
11794 (function_and_variable_visibility): Use it.
11795 * symtab.c (symtab_for_node_and_aliases,
11796 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11798 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11800 PR rtl-optimization/57963
11801 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11802 (lra_constraints): Use them.
11804 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11806 * config/arm/types.md (define_attr "type"): Add "load_acq"
11808 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11810 (cortex_a53_store1): Likewise.
11812 2013-08-01 Jan Hubicka <jh@suse.cz>
11814 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11815 partitions are not needed.
11817 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11819 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11820 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11821 MAYBE_NON_Q_CLASS_P where appropriate.
11823 2013-08-01 Jan Hubicka <jh@suse.cz>
11825 * cgraph.h (release_function_body): Declare.
11826 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11827 of unused delcarations.
11829 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11831 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11832 RTL form when subtracting a constant.
11834 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11836 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11837 Generate canonical plus rtx with negated immediate instead of minus
11839 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11841 2013-08-01 Jan Hubicka <jh@suse.cz>
11843 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11844 (cgraph_release_function_body): Likewise.
11845 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11846 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11847 to used_as_abstract_origin.
11848 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11849 symbols used as abstract origins.
11850 * cgraphunit.c (analyze_functions): Update.
11851 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11852 used_as_abstract_origin.
11853 * tree-inline.c (tree_function_versioning): Update
11854 used_as_abstract_origin; be ready for DECL_RESULT and
11855 DECL_ARGUMENTS to be NULL.
11857 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11858 for abstract functions.
11859 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11862 2013-08-01 Jan Hubicka <jh@suse.cz>
11864 * profile.c (compute_value_histograms): Fix thinko.
11866 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11868 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11869 aarch-common-protos.h to extra_headers.
11870 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11871 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11872 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11874 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11876 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11877 Include "../arm/types.md". Define "type" attribute for all patterns.
11878 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11881 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11883 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11884 to support power8 load fusion.
11885 (fusion_gpr_mem_load): Likewise.
11887 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11889 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11890 declarations for power8 load fusion.
11891 (emit_fusion_gpr_load): Likewise.
11893 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11894 tuning for power8, turn on fusion mode by default. Turn on sign
11895 extending fusion mode if normal fusion mode is on, and we are at
11897 (fusion_gpr_load_p): New function, return true if we can fuse an
11898 addis instruction with a dependent load to a GPR.
11899 (emit_fusion_gpr_load): Emit the instructions for power8 load
11902 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11903 (VSX load fusion peepholes): New peepholes to fuse together an
11904 addi instruction with a VSX load instruction.
11906 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11907 peepholes to fuse an addis instruction with a load to a GPR base
11908 register. If we are supporting sign extending fusions, convert
11909 sign extending loads to zero extending loads and add an explicit
11912 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11914 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11915 aarch-common-protos.h to extra_headers.
11916 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11917 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11918 (arm_early_store_addr_dep): Likewise.
11919 (arm_no_early_alu_shift_dep): Likewise.
11920 (arm_no_early_alu_shift_value_dep): Likewise.
11921 (arm_no_early_mul_dep): Likewise.
11922 (arm_no_early_store_addr_dep): Likewise.
11923 (arm_mac_accumulator_is_mul_result): Likewise.
11924 (arm_mac_accumulator_is_result): Likewise.
11925 * config/arm/aarch-common.c: ... here. New file.
11926 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11928 (arm_early_store_addr_dep): Likewise.
11929 (arm_no_early_alu_shift_dep): Likewise.
11930 (arm_no_early_alu_shift_value_dep): Likewise.
11931 (arm_no_early_mul_dep): Likewise.
11932 (arm_no_early_store_addr_dep): Likewise.
11933 (arm_mac_accumulator_is_mul_result): Likewise.
11934 (arm_mac_accumulator_is_result): Likewise.
11935 * config/arm/aarch-common-protos.h: ... here. New file.
11936 * config/arm/t-arm (aarch-common.o): Define.
11938 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11940 * config/arm/arm.md: Include new file "types.md".
11941 (define_attr "type"): Move from here to ...
11942 (define_attr "mul32"): Likewise.
11943 (define_attr "mul64"): Likewise.
11944 * config/arm/types.md: ... here. New file.
11946 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11948 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11949 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11951 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11953 * gen-pass-instances.awk: Fix offset of substr().
11955 2013-07-31 David Malcolm <dmalcolm@redhat.com>
11957 * Makefile.in (pass-instances.def): New.
11958 (passes.o): Replace dependency on passes.def with one on
11961 * gen-pass-instances.awk: New.
11963 * passes.c (pass_manager::pass_manager): Use pass-instances.def
11964 rather than passes.def, updating local definition of NEXT_PASS
11965 macro to add an extra NUM parameter (currently unused).
11967 2013-07-30 David Malcolm <dmalcolm@redhat.com>
11969 * Makefile.in (PASS_MANAGER_H): New.
11970 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
11971 (passes.o): Likewise.
11972 (statistics.o): Likewise.
11973 (cgraphunit.o): Likewise.
11974 (context.o): Depend on PASS_MANAGER_H.
11976 * pass_manager.h: New.
11978 * cgraphunit.c (cgraph_add_new_function): Update for moves
11979 of globals to fields of pass_manager.
11980 (analyze_function): Likewise.
11981 (expand_function): Likewise.
11982 (ipa_passes): Likewise.
11983 (compile): Likewise.
11985 * context.c (context::context): New.
11986 * context.h (context::context): New.
11987 (context::get_passes): New.
11988 (context::passes_): New.
11990 * lto-cgraph.c (input_node): Update for moves of globals to
11991 fields of pass_manager.
11993 * passes.c (all_passes): Remove, in favor of a field of the
11994 same name within the new class pass_manager.
11995 (all_small_ipa_passes): Likewise.
11996 (all_lowering_passes): Likewise.
11997 (all_regular_ipa_passes): Likewise.
11998 (all_late_ipa_passes): Likewise.
11999 (all_lto_gen_passes): Likewise.
12000 (passes_by_id): Likewise.
12001 (passes_by_id_size): Likewise.
12002 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12003 the new class pass_manager.
12004 (set_pass_for_id): Convert to...
12005 (pass_manager::set_pass_for_id): ...method.
12006 (get_pass_for_id): Convert to...
12007 (pass_manager::get_pass_for_id): ...method.
12008 (register_one_dump_file): Move body of implementation into...
12009 (pass_manager::register_one_dump_file): ...here.
12010 (register_dump_files_1): Convert to...
12011 (pass_manager::register_dump_files_1): ...method.
12012 (register_dump_files): Convert to...
12013 (pass_manager::register_dump_files): ...method.
12014 (create_pass_tab): Update for moves of globals to fields of
12016 (dump_passes): Move body of implementation into...
12017 (pass_manager::dump_passes): ...here.
12018 (register_pass): Move body of implementation into...
12019 (pass_manager::register_pass): ...here.
12020 (init_optimization_passes): Convert into...
12021 (pass_manager::pass_manager): ...constructor for new
12022 pass_manager class, and initialize the pass_lists array.
12023 (check_profile_consistency): Update for moves of globals to
12024 fields of pass_manager.
12025 (dump_profile_report): Move body of implementation into...
12026 (pass_manager::dump_profile_report): ...here.
12027 (ipa_write_summaries_1): Update for moves of pass lists from
12028 being globals to fields of pass_manager.
12029 (ipa_write_optimization_summaries): Likewise.
12030 (ipa_read_summaries): Likewise.
12031 (ipa_read_optimization_summaries): Likewise.
12032 (execute_all_ipa_stmt_fixups): Likewise.
12034 * statistics.c (statistics_fini): Update for moves of globals to
12035 fields of pass_manager.
12037 * toplev.c (general_init): Replace call to
12038 init_optimization_passes with construction of the pass_manager
12041 * tree-pass.h (all_passes): Remove, in favor of a field of the
12042 same name within the new class pass_manager.
12043 (all_small_ipa_passes): Likewise.
12044 (all_lowering_passes): Likewise.
12045 (all_regular_ipa_passes): Likewise.
12046 (all_lto_gen_passes): Likewise.
12047 (all_late_ipa_passes): Likewise.
12048 (passes_by_id): Likewise.
12049 (passes_by_id_size): Likewise.
12050 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12051 the new class pass_manager.
12052 (get_pass_for_id): Remove.
12054 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12056 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
12059 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12061 * arm.md (mulhi3): New expand pattern.
12063 2013-07-30 Jan Hubicka <jh@suse.cz>
12064 Martin Liska <marxin.liska@gmail.com>
12066 * profile.c (compute_value_histograms): Do not ICE when
12067 there is mismatch only on some counters.
12069 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12071 PR rtl-optimization/57637
12072 * function.c (move_insn_for_shrink_wrap): Also check the
12073 GEN set of the LIVE problem for the liveness analysis
12074 if it exists, otherwise give up.
12076 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12078 PR tree-optimization/57993
12079 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
12080 replaced statement in the candidate table.
12081 (phi_add_costs): Return infinite cost when the hidden basis does
12082 not dominate all phis on which the candidate is dependent.
12083 (replace_one_candidate): Record replaced statement in the
12086 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
12088 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
12089 (ashlv2si3): New expander.
12090 (*ashlv2si3_i): New define_insn_and_split.
12091 * predicates.md (float_operation): Allow patterns with three
12092 basic sub-patterns.
12094 PR rtl-optimization/58021
12095 * mode-switching.c (create_pre_exit): Always split off preceding
12096 insns if we are not at the basic block head.
12098 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
12100 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
12101 (UCLIBC_DYNAMIC_LINKER): New macro.
12102 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
12104 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
12105 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
12107 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
12108 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
12109 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
12110 for SF and DF modes. Use ieee_quad_format for TF mode.
12111 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
12112 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
12113 (mips_option_override): Handle `-mnan=legacy'.
12114 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
12115 `-mabs=2008' and `-mnan=2008'.
12116 (OPTION_DEFAULT_SPECS): Add "nan" default.
12117 (ASM_SPEC): Handle `-mnan='.
12118 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
12119 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
12120 comment accordingly.
12121 (neg<mode>2): Likewise.
12122 * config/mips/mips.opt (mabs, mnan): New options.
12123 * doc/install.texi (Configuration): Document `--with-nan=' option.
12124 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
12126 (MIPS Options): Document them.
12127 * config.gcc <mips*-*-*>: Handle `--with-nan='.
12128 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
12129 * configure: Regenerate.
12130 * config.in: Regenerate.
12132 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12134 * config/i386/i386.md (float post-reload splitters): Do not check
12135 for subregs of SSE registers.
12137 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12138 H.J. Lu <hongjiu.lu@intel.com>
12142 * config/i386/i386.md (post-reload splitter
12143 to avoid partial SSE reg dependency stalls): New pattern.
12145 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12147 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12149 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12151 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12152 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12154 2013-07-26 Cary Coutant <ccoutant@google.com>
12156 * dwarf2out.c (die_checksum_ordered): Don't include template
12157 instantiations in signature.
12158 (is_template_parameter): New function.
12159 (is_template_instantiation): New function.
12160 (generate_skeleton_bottom_up): Don't include template instantiations
12162 (generate_skeleton): Likewise.
12163 (break_out_comdat_types): Move recursive call to break out nested
12165 (prune_unused_types_mark_generic_parms_dies): Call
12166 is_template_parameter.
12168 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12170 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12171 uses vector registers.
12172 * config/aarch64/iterators.md: Add attributes rtn and vas.
12174 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12175 Richard Earnshaw <richard.earnshaw@arm.com>
12177 * combine.c (simplify_comparison): Re-canonicalize operands
12179 * config/arm/arm.md (movcond_addsi): New splitter.
12181 2013-07-25 Sterling Augustine <saugustine@google.com>
12183 * dwarf2out.c (size_of_pubnames): Move code to...
12184 (include_pubname_in_output): ...here. New.
12185 (want_pubnames): Rearrange.
12186 (output_pubnames): Call include_pubname_in_output. Move assertion.
12188 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12190 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12192 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12195 * doc/extend.texi: Remove obsolete builtins. Fix
12196 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12198 2013-07-25 Jan Hubicka <jh@suse.cz>
12200 * cgraph.c (release_function_body): Break out from ...
12201 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12202 and DECL_ARGUMENTS.
12203 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12204 old_tree in the map.
12205 (create_specialized_node): Update.
12206 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12208 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12209 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12210 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12211 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12214 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12216 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12217 addsi3_carryin_alt2_<optab>): Correct output template.
12219 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12221 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12222 Adjust for arm_restrict_it.
12223 Remove trailing whitespace.
12225 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12227 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12228 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12230 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12232 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12234 PR rtl-optimization/57960
12235 * lra-constraints.c (process_alt_operands): Use the right mode
12236 when checking strict_low.
12238 2013-07-25 Jan Hubicka <jh@suse.cz>
12240 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12241 * cgraph.c (cgraph_remove_node): Do not release function body
12242 when in cgraph streaming.
12243 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12244 in other partitions are not considered reachable; fix handling of
12247 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12249 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12251 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12256 * config/arm/arm.md ("*sibcall_insn): Replace use of
12257 Ss with US. Adjust output for v5 and v4t.
12258 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12260 * config/arm/constraints.md ("Ss"): Rename to US.
12262 2013-07-25 Terry Guo <terry.guo@arm.com>
12264 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12265 shift_add/shift_sub0/shift_sub1 RTXs.
12267 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12268 Anton Blanchard <anton@au1.ibm.com>
12270 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12271 (altivec_vpks<VI_char>ss): Likewise.
12272 (altivec_vpks<VI_char>us): Likewise.
12273 (altivec_vpku<VI_char>us): Likewise.
12274 (altivec_vpku<VI_char>um): Likewise.
12276 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12278 Introduce context class.
12280 * Makefile.in (CONTEXT_H): New.
12281 (OBJS): Add context.o.
12282 (toplev.o): Add CONTEXT_H to dependencies.
12285 * toplev.c (general_init): Create the singleton gcc::context instance.
12291 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12293 PR rtl-optimization/57968
12294 * mode-switching.c (create_pre_exit): Allow instructions that
12295 don't set a return register to need a non-exit mode.
12297 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12298 Anton Blanchard <anton@au1.ibm.com>
12300 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12301 operands to vperm for little endian.
12302 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12303 of lvsl to create the control mask for a vperm for little endian.
12305 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12306 Anton Blanchard <anton@au1.ibm.com>
12308 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12309 two operands for little-endian.
12311 2013-07-23 Steve Ellcey <sellcey@mips.com>
12313 * config/mips/mips.c (mips_case_values_threshold): New.
12314 (TARGET_CASE_VALUES_THRESHOLD): Define.
12316 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12317 Anton Blanchard <anton@au1.ibm.com>
12319 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12320 selection of field for vector splat in little endian mode.
12322 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12324 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12325 expanders to rs6000.md.
12326 (ior<mode>3): Likewise.
12327 (and<mode>3): Likewise.
12328 (one_cmpl<mode>2): Likewise.
12329 (nor<mode>3): Likewise.
12330 (andc<mode>3): Likewise.
12331 (eqv<mode>3): Likewise.
12332 (nand<mode>3): Likewise.
12333 (orc<mode>3): Likewise.
12335 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12338 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12339 to split multi-word logical operations.
12340 (rs6000_split_logical_di): Likewise.
12341 (rs6000_split_logical): Likewise.
12343 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12344 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12345 and allow TImode operations in 32-bit.
12346 (vsx_and<mode>3_64bit): Likewise.
12347 (vsx_ior<mode>3_32bit): Likewise.
12348 (vsx_ior<mode>3_64bit): Likewise.
12349 (vsx_xor<mode>3_32bit): Likewise.
12350 (vsx_xor<mode>3_64bit): Likewise.
12351 (vsx_one_cmpl<mode>2_32bit): Likewise.
12352 (vsx_one_cmpl<mode>2_64bit): Likewise.
12353 (vsx_nor<mode>3_32bit): Likewise.
12354 (vsx_nor<mode>3_64bit): Likewise.
12355 (vsx_andc<mode>3_32bit): Likewise.
12356 (vsx_andc<mode>3_64bit): Likewise.
12357 (vsx_eqv<mode>3_32bit): Likewise.
12358 (vsx_eqv<mode>3_64bit): Likewise.
12359 (vsx_nand<mode>3_32bit): Likewise.
12360 (vsx_nand<mode>3_64bit): Likewise.
12361 (vsx_orc<mode>3_32bit): Likewise.
12362 (vsx_orc<mode>3_64bit): Likewise.
12364 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12365 logical types in GPRs.
12367 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12368 logical insns to rs6000.md, and allow TImode operations in
12370 (altivec_ior<mode>3): Likewise.
12371 (altivec_xor<mode>3): Likewise.
12372 (altivec_one_cmpl<mode>2): Likewise.
12373 (altivec_nor<mode>3): Likewise.
12374 (altivec_andc<mode>3): Likewise.
12376 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12377 attributes for moving the 128-bit logical operations into
12379 (BOOL_REGS_OUTPUT): Likewise.
12380 (BOOL_REGS_OP1): Likewise.
12381 (BOOL_REGS_OP2): Likewise.
12382 (BOOL_REGS_UNARY): Likewise.
12383 (BOOL_REGS_AND_CR0): Likewise.
12384 (one_cmpl<mode>2): Add support for DI logical operations on
12385 32-bit, splitting the operations to 32-bit.
12386 (anddi3): Likewise.
12387 (iordi3): Likewise.
12388 (xordi3): Likewise.
12389 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12390 changes to combine the 32/64-bit code, allow logical operations on
12391 TI mode in 32-bit, and to use similar match_operator patterns like
12392 scalar mode uses. Combine the Altivec and VSX code for logical
12393 operations, and move it here.
12394 (ior<mode>3, 128-bit types): Likewise.
12395 (xor<mode>3, 128-bit types): Likewise.
12396 (one_cmpl<mode>3, 128-bit types): Likewise.
12397 (nor<mode>3, 128-bit types): Likewise.
12398 (andc<mode>3, 128-bit types): Likewise.
12399 (eqv<mode>3, 128-bit types): Likewise.
12400 (nand<mode>3, 128-bit types): Likewise.
12401 (orc<mode>3, 128-bit types): Likewise.
12402 (and<mode>3_internal): Likewise.
12403 (bool<mode>3_internal): Likewise.
12404 (boolc<mode>3_internal1): Likewise.
12405 (boolc<mode>3_internal2): Likewise.
12406 (boolcc<mode>3_internal1): Likewise.
12407 (boolcc<mode>3_internal2): Likewise.
12408 (eqv<mode>3_internal1): Likewise.
12409 (eqv<mode>3_internal2): Likewise.
12410 (one_cmpl1<mode>3_internal): Likewise.
12412 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12414 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12415 Rename flag_stack_usage to flag_stack_usage_info.
12417 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12419 * config/microblaze/sync.md: New file.
12420 * config/microblaze/microblaze.md: Include sync.md
12421 * config/microblaze/microblaze.c: Add print_operand 'y'.
12422 * config/microblaze/constraints.md: Add memory_contraint
12423 'Q' which is a single register.
12425 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12427 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12429 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12431 * reload.c (find_reloads): Exit loop once we find this operand
12432 cannot be reloaded somehow for this alternative.
12434 * reload.c (find_reloads): Exit loop once we find a hard register.
12436 * rtlanal.c (computed_jump_p): Exit loop once we find label
12439 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12441 * cfgloopmanip.c (remove_path): Exit loop after setting
12444 * gensupport.c (subst_dup): Avoid loop if code is not
12445 MATCH_DUP nor MATCH_OP_DUP.
12447 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12449 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12451 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12453 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12454 true for SP_REGNUM if mode == ptr_mode.
12455 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12456 with value R0_REGNUM + 31.
12458 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12460 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12461 pad pointer-typed argument downward.
12463 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12465 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12466 and __ILP32__ when the ILP32 model is in use.
12468 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12470 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12471 (aarch64_load_symref_appropriately): In the case of
12472 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12473 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12474 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12475 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12476 if 'mode' doesn't equal to 'ptr_mode'.
12477 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12478 'vcall_offset'; change to call aarch64_emit_move differently depending
12479 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12480 to calculate the upper bound of 'vcall_offset'.
12481 (aarch64_cannot_force_const_mem): Change to also return true if
12483 (aarch64_legitimize_reload_address): In the case of large
12484 displacements, add new local variable 'xmode' and an assertion
12485 based on it; change to use 'xmode' to generate the new rtx and
12487 (aarch64_asm_trampoline_template): Change to generate the template
12488 differently depending on TARGET_ILP32 or not; change to use
12489 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12490 (aarch64_trampoline_size): Removed.
12491 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12492 and replace immediate literals with it. Change to use 'ptr_mode'
12493 instead of 'DImode' and call convert_memory_address if the mode
12494 of 'fnaddr' doesn't equal to 'ptr_mode'.
12495 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12497 (aarch64_elf_asm_destructor): Likewise.
12498 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12499 on TARGET_ILP32 instead of aarch64_trampoline_size.
12500 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12501 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12502 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12503 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12504 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12505 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12506 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12507 depending on the value of 'mode'.
12508 (add_losym_<mode>): New.
12509 (ldr_got_small_<mode>): New, based on ldr_got_small.
12510 (ldr_got_small): Remove.
12511 (ldr_got_small_sidi): New.
12512 * config/aarch64/iterators.md (P): New.
12513 (PTR): Change to 'ptr_mode' in the condition.
12515 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12517 * config.gcc (aarch64*-*-*): Support --with-abi.
12518 (aarch64*-*-elf): Support --with-multilib-list.
12519 (aarch64*-*-linux*): Likewise.
12520 (supported_defaults): Add abi to aarch64*-*-*.
12521 * configure.ac: Mention AArch64 for --with-multilib-list.
12522 * configure: Re-generated.
12523 * config/aarch64/biarchilp32.h: New file.
12524 * config/aarch64/biarchlp64.h: New file.
12525 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12527 (MULTILIB_DEFAULTS): Ditto.
12528 (DRIVER_SELF_SPECS): Ditto.
12529 (ASM_SPEC): Update to also substitute -mabi.
12530 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12531 file whose name depends on -mabi= and -mbig-endian.
12532 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12534 (POINTER_SIZE): New define.
12535 (POINTERS_EXTEND_UNSIGNED): Ditto.
12536 (enum aarch64_abi_type): New enumeration tag.
12537 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12538 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12539 (TARGET_ILP32): New define.
12540 * config/aarch64/aarch64.opt (mabi): New.
12541 (aarch64_abi): New.
12542 (ilp32, lp64): New values for -mabi.
12543 * config/aarch64/t-aarch64 (comma): New define.
12544 (MULTILIB_OPTIONS): Ditto.
12545 (MULTILIB_DIRNAMES): Ditto.
12546 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12547 * doc/invoke.texi: Document -mabi for AArch64.
12549 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12551 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12553 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12554 Anton Blanchard <anton@au1.ibm.com>
12556 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12557 endianness when selecting field to splat.
12559 2013-07-22 Eric Christopher <echristo@gmail.com>
12561 * dwarf2out.c (die_odr_checksum): New function to use
12562 CHECKSUM_ macros and ULEB128 for DIE tag.
12563 (generate_type_signature): Use.
12565 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12567 * config.gcc (sparc*-*-*): Accept leon3 processor.
12568 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12569 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12570 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12571 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12572 (mfix-ut699): Adjust comment.
12573 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12574 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12575 (CPP_CPU_SPEC): Likewise.
12576 (ASM_CPU_SPEC): Likewise.
12577 * config/sparc/sparc.c (leon3_cost): New constant.
12578 (sparc_option_override): Add leon3 support.
12579 (mem_ref): New function.
12580 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12581 (sparc_do_work_around_errata): Look into the instruction in the delay
12582 slot and adjust accordingly. Add fix for the data cache nullify issues
12583 of the UT699. Change insertion position for the NOP.
12584 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12585 (leon3_load): New reservation.
12586 (leon_store): Bump latency to 2.
12587 (grfpu): New automaton.
12588 (grfpu_alu): New unit.
12589 (grfpu_ds): Likewise.
12590 (leon_fp_alu): Adjust.
12591 (leon_fp_mult): Delete.
12592 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12593 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12594 * config/sparc/sparc.md (cpu): Add leon3.
12595 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12596 (swapsi): Likewise.
12597 (atomic_test_and_set): Likewise.
12598 (ldstub): Likewise.
12600 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12602 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12603 default for R5900 targets.
12604 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12605 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12606 * config/mips/mips.c (mips_option_override): Report an error for
12607 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12608 for -march=r5900 -mhard-float.
12610 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12612 * df-problems.c (can_move_insns_across): Exit loop once we
12613 find a non-fixed, non-global register.
12615 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12618 * omega.c (omega_eliminate_red): Break after setting red_found.
12619 (omega_problem_has_red_equations): Similarly after setting found.
12620 (omega_query_variable): Similarly after setting coupled.
12622 2013-07-22 Marek Polacek <polacek@redhat.com>
12624 * gimplify.c: Don't include gimple.h twice.
12626 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12628 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12629 instead of TARGET_THUMB1.
12630 (Pz): New constraint.
12631 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12633 (compare_negsi_si): Likewise.
12634 (compare_addsi2_op0): Likewise.
12635 (compare_addsi2_op1): Likewise.
12636 (addsi3_carryin_<optab>): Likewise.
12637 (addsi3_carryin_alt2_<optab>): Likewise.
12638 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12639 for arm_restrict_it.
12640 (subsi3_carryin): Likewise.
12641 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12642 (minmax_arithsi): Disable for arm_restrict_it.
12643 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12644 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12645 (satsi_<SAT:code>_shift): Likewise.
12646 (arm_shiftsi3): Add alternative for 16-bit encoding.
12647 (arm32_movhf): Disable for arm_restrict_it.
12648 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12649 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12651 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12653 * config/arm/arm.md (attribute "insn"): Delete.
12654 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12655 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12656 (not_shiftsi): Update for attribute change.
12657 (not_shiftsi_compare0): Likewise.
12658 (not_shiftsi_compare0_scratch): Likewise.
12659 (arm_one_cmplsi2): Likewise.
12660 (thumb1_one_cmplsi2): Likewise.
12661 (notsi_compare0): Likewise.
12662 (notsi_compare0_scratch): Likewise.
12663 (thumb1_movdi_insn): Likewise.
12664 (arm_movsi_insn): Likewise.
12665 (movhi_insn_arch4): Likewise.
12666 (movhi_bytes): Likewise.
12667 (arm_movqi_insn): Likewise.
12668 (thumb1_movqi_insn): Likewise.
12669 (arm32_movhf): Likewise.
12670 (thumb1_movhf): Likewise.
12671 (arm_movsf_soft_insn): Likewise.
12672 (thumb1_movsf_insn): Likewise.
12673 (thumb_movdf_insn): Likewise.
12674 (movsicc_insn): Likewise.
12675 (movsfcc_soft_insn): Likewise.
12676 (and_scc): Likewise.
12677 (cond_move): Likewise.
12678 (if_move_not): Likewise.
12679 (if_not_move): Likewise.
12680 (if_shift_move): Likewise.
12681 (if_move_shift): Likewise.
12682 (if_shift_shift): Likewise.
12683 (if_not_arith): Likewise.
12684 (if_arith_not): Likewise.
12685 (cond_move_not): Likewise.
12686 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12687 (neon_mov<mode>): Likewise.
12688 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12689 (thumb2_movsi_vfp): Likewise.
12690 (movsf_vfp): Likewise.
12691 (thumb2_movsf_vfp): Likewise.
12692 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12694 (cortexa7_older_only): Likewise.
12695 (cortexa7_younger): Likewise.
12696 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12697 (1020alu_shift_op): Likewise.
12698 (1020alu_shift_reg_op): Likewise.
12699 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12700 (alu_shift_op): Likewise.
12701 (alu_shift_reg_op): Likewise.
12702 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12703 (11_alu_shift_op): Likewise.
12704 (11_alu_shift_reg_op): Likewise.
12705 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12706 (9_alu_shift_reg_op): Likewise.
12707 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12709 (cortex_a15_alu_shift): Likewise.
12710 (cortex_a15_alu_shift_reg): Likewise.
12711 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12712 (cortex_a5_alu_shift): Likewise.
12713 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12715 (cortex_a53_alu_shift): Likewise.
12716 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12718 (cortex_a7_alu_reg): Likewise.
12719 (cortex_a7_alu_shift): Likewise.
12720 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12721 (cortex_a8_alu_shift): Likewise.
12722 (cortex_a8_alu_shift_reg): Likewise.
12723 (cortex_a8_mov): Likewise.
12724 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12725 (cortex_a9_dp_shift): Likewise.
12726 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12727 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12728 (cortex_r4_mov): Likewise.
12729 (cortex_r4_alu_shift): Likewise.
12730 (cortex_r4_alu_shift_reg): Likewise.
12731 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12732 (526_alu_shift_op): Likewise.
12733 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12734 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12735 (626te_alu_shift_op): Likewise.
12736 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12737 (726te_alu_op): Likewise.
12738 (726te_alu_shift_op): Likewise.
12739 (726te_alu_shift_reg_op): Likewise.
12740 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12741 (mp626_alu_shift_op): Likewise.
12742 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12743 (pj4_alu_e1_conds): Likewise.
12744 (pj4_alu): Likewise.
12745 (pj4_alu_conds): Likewise.
12746 (pj4_shift): Likewise.
12747 (pj4_shift_conds): Likewise.
12748 (pj4_alu_shift): Likewise.
12749 (pj4_alu_shift_conds): Likewise.
12751 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12753 * config/arm/predicates.md (shiftable_operator_strict_it):
12755 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12756 Disable cond_exec version for arm_restrict_it.
12757 (thumb2_smaxsi3): Convert to generate cond_exec.
12758 (thumb2_sminsi3): Likewise.
12759 (thumb32_umaxsi3): Likewise.
12760 (thumb2_uminsi3): Likewise.
12761 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12762 (thumb2_neg_abssi2): Likewise.
12763 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12764 (thumb2_movsicc_insn): Adjust alternatives.
12765 (thumb2_mov_negscc): Disable for arm_restrict_it.
12766 (thumb2_mov_negscc_strict_it): New pattern.
12767 (thumb2_mov_notscc_strict_it): New pattern.
12768 (thumb2_mov_notscc): Disable for arm_restrict_it.
12769 (thumb2_ior_scc): Likewise.
12770 (thumb2_ior_scc_strict_it): New pattern.
12771 (thumb2_cond_move): Adjust for arm_restrict_it.
12772 (thumb2_cond_arith): Disable for arm_restrict_it.
12773 (thumb2_cond_arith_strict_it): New pattern.
12774 (thumb2_cond_sub): Adjust for arm_restrict_it.
12775 (thumb2_movcond): Likewise.
12776 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12777 (thumb2_zero_extendhisi2_v6): Likewise.
12778 (thumb2_zero_extendqisi2_v6): Likewise.
12779 (orsi_notsi_si): Likewise.
12780 (orsi_not_shiftsi_si): Likewise.
12782 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12784 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12785 instruction sequence is 1 byte shorter.
12787 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12789 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12790 it is not needed after split.
12792 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12795 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12796 second label for nonlocal goto receivers. Don't output pic base labels
12797 unless we're producing PIC; mark that action unreachable().
12798 (ix86_save_reg): If the function contains a nonlocal label, save the
12800 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12801 * config/darwin.c (emitted_pic_label_num): New GTY.
12802 (update_pic_label_number_if_needed): New.
12803 (machopic_output_function_base_name): Adjust for nonlocal receiver
12805 (machopic_should_output_picbase_label): New.
12806 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12807 (nonlocal_goto_receiver): New insn and split.
12809 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12811 * config/aarch64/aarch64-builtins.c
12812 (aarch64_fold_builtin): Fold abs in all modes.
12813 * config/aarch64/aarch64-simd-builtins.def
12814 (abs): Enable for all modes.
12815 * config/aarch64/arm_neon.h
12816 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12817 (vabs_f64): Add missing intrinsic.
12819 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12821 * config/aarch64/arm_neon.h (vabs_s64): New function
12823 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12826 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12827 * config/avr/avr.md (adjust_len): Add `round'.
12828 * config/avr/avr-protos.h (avr_out_round): New prototype.
12829 (avr_out_plus): Add `out_label' argument.
12830 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12831 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12832 Handle the case where `insn' is just a pattern.
12833 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12834 (avr_out_round): New function.
12835 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12837 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12839 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12840 Add check for flag_stack_usage to handle -fstack-usage support
12842 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12844 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12845 interaction for new Power8 flags and VSX.
12847 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12849 PR middle-end/57698
12850 * tree-inline.c (expand_call_inline): Emit errors during
12851 early_inlining only if optimization is not turned on.
12853 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12857 * passes.c (init_optimization_passes): Move the construction of
12858 the pass hierarchy into a new passes.def file.
12860 * Makefile.in (passes.o): Add dependency on passes.def.
12862 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12864 * passes.c (init_optimization_passes): Introduce macros for
12865 constructing the tree of passes (INSERT_PASSES_AFTER,
12866 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12867 TERMINATE_PASS_LIST).
12869 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12870 Wei Mi <wmi@google.com>
12872 PR rtl-optimization/57878
12873 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12875 (reload_pseudo_compare_func): Check nregs first for reload
12878 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12880 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12882 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12884 * read-rtl.c (validate_const_int): Once an invalid character is
12885 seen, quit the loop.
12887 * gengtype.c (write_roots): Similarly once we find the "deletable"
12888 or "if_marked" option.
12890 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12892 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12893 "xtab" and "sat". Move value "clz" from here to ...
12894 (attriubte "type"): ... here.
12895 (satsi_<SAT:code>): Delete "insn" attribute.
12896 (satsi_<SAT:code>_shift): Likewise.
12897 (arm_zero_extendqisi2addsi): Likewise.
12898 (arm_extendqisi2addsi): Likewise.
12899 (clzsi2): Update for attribute changes.
12900 (rbitsi2): Likewise.
12901 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12903 (arm_usatsihi): Likewise.
12904 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12906 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12908 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12909 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12910 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12911 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12912 in alphabetical order.
12913 (attribute "core_cycles"): Update for attribute changes.
12914 (arm_addsi3): Likewise.
12915 (addsi3_compare0): Likewise.
12916 (addsi3_compare0_scratch): Likewise.
12917 (addsi3_compare_op1): Likewise.
12918 (addsi3_compare_op2): Likewise.
12919 (compare_addsi2_op0): Likewise.
12920 (compare_addsi2_op1): Likewise.
12921 (addsi3_carryin_shift_<optab>): Likewise.
12922 (subsi3_carryin_shift): Likewise.
12923 (rsbsi3_carryin_shift): Likewise.
12924 (arm_subsi3_insn): Likewise.
12925 (subsi3_compare0): Likewise.
12926 (subsi3_compare): Likewise.
12927 (arm_andsi3_insn): Likewise.
12928 (thumb1_andsi3_insn): Likewise.
12929 (andsi3_compare0): Likewise.
12930 (andsi3_compare0_scratch): Likewise.
12931 (zeroextractsi_compare0_scratch
12932 (andsi_not_shiftsi_si): Likewise.
12933 (iorsi3_insn): Likewise.
12934 (iorsi3_compare0): Likewise.
12935 (iorsi3_compare0_scratch): Likewise.
12936 (arm_xorsi3): Likewise.
12937 (thumb1_xorsi3_insn): Likewise.
12938 (xorsi3_compare0): Likewise.
12939 (xorsi3_compare0_scratch): Likewise.
12940 (satsi_<SAT:code>_shift): Likewise.
12942 (arm_shiftsi3): Likewise.
12943 (shiftsi3_compare0): Likewise.
12944 (not_shiftsi): Likewise.
12945 (not_shiftsi_compare0): Likewise.
12946 (not_shiftsi_compare0_scratch): Likewise.
12947 (arm_one_cmplsi2): Likewise.
12948 (thumb_one_complsi2): Likewise.
12949 (notsi_compare0): Likewise.
12950 (notsi_compare0_scratch): Likewise.
12951 (thumb1_zero_extendhisi2): Likewise.
12952 (arm_zero_extendhisi2): Likewise.
12953 (arm_zero_extendhisi2_v6): Likewise.
12954 (arm_zero_extendhisi2addsi): Likewise.
12955 (thumb1_zero_extendqisi2): Likewise.
12956 (thumb1_zero_extendqisi2_v6): Likewise.
12957 (arm_zero_extendqisi2): Likewise.
12958 (arm_zero_extendqisi2_v6): Likewise.
12959 (arm_zero_extendqisi2addsi): Likewise.
12960 (thumb1_extendhisi2): Likewise.
12961 (arm_extendhisi2): Likewise.
12962 (arm_extendhisi2_v6): Likewise.
12963 (arm_extendqisi): Likewise.
12964 (arm_extendqisi_v6): Likewise.
12965 (arm_extendqisi2addsi): Likewise.
12966 (thumb1_extendqisi2): Likewise.
12967 (thumb1_movdi_insn): Likewise.
12968 (arm_movsi_insn): Likewise.
12969 (movsi_compare0): Likewise.
12970 (movhi_insn_arch4): Likewise.
12971 (movhi_bytes): Likewise.
12972 (arm_movqi_insn): Likewise.
12973 (thumb1_movqi_insn): Likewise.
12974 (arm32_movhf): Likewise.
12975 (thumb1_movhf): Likewise.
12976 (arm_movsf_soft_insn): Likewise.
12977 (thumb1_movsf_insn): Likewise.
12978 (movdf_soft_insn): Likewise.
12979 (thumb_movdf_insn): Likewise.
12980 (arm_cmpsi_insn): Likewise.
12981 (cmpsi_shiftsi): Likewise.
12982 (cmpsi_shiftsi_swp): Likewise.
12983 (arm_cmpsi_negshiftsi_si): Likewise.
12984 (movsicc_insn): Likewise.
12985 (movsfcc_soft_insn): Likewise.
12986 (arith_shiftsi): Likewise.
12987 (arith_shiftsi_compare0
12988 (arith_shiftsi_compare0_scratch
12989 (sub_shiftsi): Likewise.
12990 (sub_shiftsi_compare0
12991 (sub_shiftsi_compare0_scratch
12992 (and_scc): Likewise.
12993 (cond_move): Likewise.
12994 (if_plus_move): Likewise.
12995 (if_move_plus): Likewise.
12996 (if_move_not): Likewise.
12997 (if_not_move): Likewise.
12998 (if_shift_move): Likewise.
12999 (if_move_shift): Likewise.
13000 (if_shift_shift): Likewise.
13001 (if_not_arith): Likewise.
13002 (if_arith_not): Likewise.
13003 (cond_move_not): Likewise.
13004 (thumb1_ashlsi3): Set type attribute.
13005 (thumb1_ashrsi3): Likewise.
13006 (thumb1_lshrsi3): Likewise.
13007 (thumb1_rotrsi3): Likewise.
13008 (shiftsi3_compare0_scratch): Likewise.
13009 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
13010 (neon_mov<mode>): Likewise.
13011 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
13013 (thumb2_movsi_insn): Likewise.
13014 (thumb2_cmpsi_neg_shiftsi): Likewise.
13015 (thumb2_extendqisi_v6): Likewise.
13016 (thumb2_zero_extendhisi2_v6): Likewise.
13017 (thumb2_zero_extendqisi2_v6): Likewise.
13018 (thumb2_shiftsi3_short): Likewise.
13019 (thumb2_addsi3_compare0_scratch): Likewise.
13020 (orsi_not_shiftsi_si): Likewise.
13021 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
13022 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
13024 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
13025 (1020alu_shift_op): Likewise.
13026 (1020alu_shift_reg_op): Likewise.
13027 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13028 (alu_shift_op): Likewise.
13029 (alu_shift_reg_op): Likewise.
13030 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13031 (11_alu_shift_op): Likewise.
13032 (11_alu_shift_reg_op): Likewise.
13033 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13034 (9_alu_shift_reg_op): Likewise.
13035 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13037 (cortex_a15_alu_shift): Likewise.
13038 (cortex_a15_alu_shift_reg): Likewise.
13039 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
13041 (cortex_a5_alu_shift): Likewise.
13042 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
13044 (cortex_a53_alu_shift): Likewise.
13045 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13047 (cortex_a7_alu_reg): Likewise.
13048 (cortex_a7_alu_shift): Likewise.
13049 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
13051 (cortex_a8_alu_shift): Likewise.
13052 (cortex_a8_alu_shift_reg): Likewise.
13053 (cortex_a8_mov): Likewise.
13054 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13055 (cortex_a9_dp_shift): Likewise.
13056 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
13058 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
13060 (cortex_r4_mov): Likewise.
13061 (cortex_r4_alu_shift): Likewise.
13062 (cortex_r4_alu_shift_reg): Likewise.
13063 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13064 (526_alu_shift_op): Likewise.
13065 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13066 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13067 (626te_alu_shift_op): Likewise.
13068 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
13069 (726te_alu_op): Likewise.
13070 (726te_alu_shift_op): Likewise.
13071 (726te_alu_shift_reg_op): Likewise.
13072 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13073 (mp626_alu_shift_op): Likewise.
13074 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
13075 (pj4_alu_e1_conds): Likewise.
13076 (pj4_alu): Likewise.
13077 (pj4_alu_conds): Likewise.
13078 (pj4_shift): Likewise.
13079 (pj4_shift_conds): Likewise.
13080 (pj4_alu_shift): Likewise.
13081 (pj4_alu_shift_conds): Likewise.
13082 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13084 (cortexa7_older_only): Likewise.
13085 (cortexa7_younger): Likewise.
13087 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13089 * ipa-pure-const.c (generate_summary): Rename to...
13090 (pure_const_generate_summary): ... this.
13092 2013-07-17 Iain Sandoe <iain@codesourcery.com>
13094 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
13096 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
13099 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
13102 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13104 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
13105 enabled without -march=zEC12.
13106 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
13109 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13111 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
13112 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
13113 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
13114 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
13115 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
13116 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
13117 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
13119 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
13120 and ISA_HAS_NMADD3_NMSUB3.
13121 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
13122 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
13123 (nmsub4<mode>, nmsub3<mode>): Likewise.
13124 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
13126 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13128 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
13129 TARGET_MIPS5400 checking.
13131 2013-07-16 Jakub Jelinek <jakub@redhat.com>
13132 Peter Bergner <bergner@vnet.ibm.com>
13134 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13135 registers in the comment.
13136 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13137 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13138 rather than FIRST_PSEUDO_REGISTERS.
13140 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13142 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13143 enable extra ISA flags with TARGET_HTM.
13145 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13147 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13150 2013-07-15 Cong Hou <congh@google.com>
13152 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13153 in compare function for sorting.
13155 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13157 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13158 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13159 * config/rs6000/rs6000.opt: Add -mhtm option.
13160 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13161 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13162 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13163 __HTM__ if the HTM instructions are available.
13164 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13165 htm_spr_reg_operand): New define_predicates.
13166 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13167 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13169 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13170 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13171 HTM builtin functions.
13172 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13173 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13174 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13175 (rs6000_builtin_mask_calculate): Likewise.
13176 (rs6000_option_override_internal): Likewise.
13177 (bdesc_htm): Add new HTM builtin support.
13178 (htm_spr_num): New function.
13179 (htm_spr_regno): Likewise.
13180 (rs6000_htm_spr_icode): Likewise.
13181 (htm_expand_builtin): Likewise.
13182 (htm_init_builtins): Likewise.
13183 (rs6000_expand_builtin): Add support for HTM builtin functions.
13184 (rs6000_init_builtins): Likewise.
13185 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13187 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13188 (TARGET_HTM, MASK_HTM): Define macros.
13189 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13190 (FIXED_REGISTERS): Likewise.
13191 (CALL_USED_REGISTERS): Likewise.
13192 (CALL_REALLY_USED_REGISTERS): Likewise.
13193 (REG_ALLOC_ORDER): Likewise.
13194 (enum reg_class): Likewise.
13195 (REG_CLASS_NAMES): Likewise.
13196 (REG_CLASS_CONTENTS): Likewise.
13197 (REGISTER_NAMES): Likewise.
13198 (ADDITIONAL_REGISTER_NAMES): Likewise.
13199 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13200 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13201 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13202 * config/rs6000/htm.md: New file.
13203 * config/rs6000/htmintrin.h: New file.
13204 * config/rs6000/htmxlintrin.h: New file.
13206 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13208 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13209 Define SYMBOL_TINY_GOT, update comment.
13210 * config/aarch64/aarch64.c
13211 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13212 (aarch64_expand_mov_immediate): Likewise.
13213 (aarch64_print_operand): Likewise.
13214 (aarch64_classify_symbol): Likewise.
13215 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13216 (ldr_got_tiny): Define.
13218 2013-07-13 Tobias Grosser <tobias@grosser.es>
13220 PR tree-optimization/54094
13221 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13222 scheduling dimension for the parallelism check from the polyhedral
13223 information in the AST.
13224 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13227 2013-07-13 Jason Merrill <jason@redhat.com>
13229 * print-tree.c (debug_vec_tree): Use debug_raw.
13230 (debug_raw (vec<tree, va_gc> &)): New.
13231 (debug_raw (vec<tree, va_gc> *)): New.
13232 * tree.h: Declare them.
13234 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13236 * ifcvt.c (ifcvt_after_combine): New static variable.
13237 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13239 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13240 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13241 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13243 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13245 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13247 2013-07-12 Michael Matz <matz@suse.de>
13249 PR middle-end/55771
13250 * convert.c (convert_to_real): Reject non-float inner types.
13252 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13254 * config/aarch64/aarch64-protos.h
13255 (aarch64_simd_immediate_valid_for_move): Remove.
13256 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13257 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13259 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13261 2013-07-11 Steve Ellcey <sellcey@mips.com>
13263 * config/mips/mips.c (mips_conditional_register_usage): Do not
13264 use t[0-7] registers in MIPS16 mode when optimizing for size.
13266 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13268 * config/i386/i386.c (dispatch_function_versions): Fix array
13269 indexing of function_version_info to match actual_versions.
13271 2013-07-11 Teresa Johnson <tejohnson@google.com>
13273 * vec.h (struct va_gc): Move release out-of-line.
13274 (va_gc::release): Call ggc_free on released vec.
13276 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13278 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13279 Require GOT register as additional operand in UNSPEC.
13280 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13281 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13282 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13283 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13284 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13285 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13286 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13288 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13291 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13292 name seen by assembler/linker rather if available.
13294 2013-07-11 Andreas Schwab <schwab@suse.de>
13296 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13298 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13300 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13302 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13304 * doc/tm.texi.in: Move hook documentation to ....
13305 * target.def: ... here.
13307 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13308 text on @hook line.
13309 * doc/tm.texi: Regenerate.
13311 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13314 * doc/invoke.texi: Document -Wconditionally-supported.
13316 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13319 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13322 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13325 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13326 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13327 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13328 Remove duplicate devices.
13329 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13330 * config/avr/t-multilib: Regenerate.
13331 * config/avr/avr-tables.opt: Regenerate.
13332 * doc/avr-mmcu.texi: Regenerate.
13334 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13337 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13339 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13341 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13342 the cost of MULT when optimizing for size.
13344 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13346 * config/cr16/cr16-protos.h: Don't include target.h.
13348 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13350 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13351 adjust register size for TDmode and TFmode for VSX registers.
13353 2013-07-08 Kai Tietz <ktietz@redhat.com>
13356 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13357 hook_bool_const_tree_true.
13359 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13361 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13362 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13363 * config/s390/s390.md: Define FPR*_REGNUM constants.
13364 Fix FPR2_REGNUM constant (18 -> 17).
13365 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13366 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13367 ("trunc<BFP:mode><DFP_ALL:mode>2")
13368 ("trunc<DFP_ALL:mode><BFP:mode>2")
13369 ("*extend<BFP:mode><DFP_ALL:mode>2")
13370 ("*extend<DFP_ALL:mode><BFP:mode>2")
13371 ("extend<BFP:mode><DFP_ALL:mode>2")
13372 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13375 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13377 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13379 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13381 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13382 and cfun_fpr_bit_p to cfun_fpr_save_p.
13383 (s390_frame_area, s390_register_info, s390_frame_info)
13384 (s390_emit_prologue, s390_emit_epilogue)
13385 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13387 * config/s390/s390.h: Define *_REGNUM macros for floating point
13390 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13392 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13394 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13396 PR rtl-optimization/57786
13397 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13398 and break out of the loop when it is set to false.
13400 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13403 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13404 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13405 (const_int 63)) 0)).
13406 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13407 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13408 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13410 PR rtl-optimization/57829
13411 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13412 mask bits outside of mode are just sign-extension from mode to HWI.
13414 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13416 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13417 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13418 adjust_address instead of change_address to keep info about alignment.
13419 (emit_strmov): Remove.
13420 (emit_memmov): New function.
13421 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13422 (expand_movmem_epilogue): Likewise and return updated rtx for
13424 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13425 destination and source.
13426 (decide_alignment): Refactor, handle vector_loop.
13427 (ix86_expand_movmem): Likewise.
13428 (ix86_expand_setmem): Likewise.
13429 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13431 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13433 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13434 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13436 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13438 * config/i386/sse.md (sse_movlhps): Change alternative 3
13439 of operand 2 to "m".
13441 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13444 * config/i386/sse.md (iptr): New mode attribute.
13445 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13446 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13447 (<sse>_vmmul<mode>3): Ditto.
13448 (<sse>_vmdiv<mode>3): Ditto.
13449 (sse_vmrcpv4sf2): Ditto.
13450 (<sse>_vmsqrt<mode>2): Ditto.
13451 (sse_vmrsqrtv4sf2): Ditto.
13452 (<sse>_vm<code><mode>3): Ditto.
13453 (avx_vmcmp<mode>3): Ditto.
13454 (<sse>_vmmaskcmp<mode>3): Ditto.
13455 (<sse>_comi): Ditto.
13456 (<sse>_ucomi): Ditto.
13457 (*xop_vmfrcz_<mode>): Ditto.
13458 (*fmai_fmadd_<mode>): Ditto.
13459 (*fmai_fmsub_<mode>): Ditto.
13460 (*fmai_fnmadd_<mode>): Ditto.
13461 (*fmai_fnmsub_<mode>): Ditto.
13462 (*fma4i_vmfmadd_<mode>): Ditto.
13463 (*fma4i_vmfmsub_<mode>): Ditto.
13464 (*fma4i_vmfnmadd_<mode>): Ditto.
13465 (*fma4i_vmfnmsub_<mode>): Ditto.
13466 (*xop_vmfrcz_<mode>): Ditto.
13467 (sse_cvtps2pi): Ditto.
13468 (sse_cvttps2pi): Ditto.
13469 (sse_cvtss2si): Ditto.
13470 (sse_cvtss2si_2): Ditto.
13471 (sse_cvtss2siq_2): Ditto.
13472 (sse_cvttss2si): Ditto.
13473 (sse_cvttss2siq): Ditto.
13474 (sse_cvtsd2si): Ditto.
13475 (sse_cvtsd2si_2): Ditto.
13476 (sse_cvtsd2siq_2): Ditto.
13477 (sse_cvttsd2si): Ditto.
13478 (sse_cvttsd2siq): Ditto.
13479 (sse_cvtsd2ss): Ditto.
13480 (sse_cvtss2sd): Ditto.
13481 (avx2_pbroadcast<mode>): Ditto.
13482 (avx2_pbroadcast<mode>_1): Ditto.
13483 (*avx_vperm_broadcast_v4sf): Ditto.
13485 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13486 (sse_movlhps): Ditto.
13487 (sse_storehps): Ditto.
13488 (sse_loadhps): Ditto.
13489 (sse_storelps): Ditto.
13490 (sse_loadlps): Ditto.
13491 (*vec_concatv4sf): Ditto.
13492 (*vec_interleave_highv2df): Ditto.
13493 (*vec_interleave_lowv2df): Ditto.
13494 (*vec_extractv2df_1_sse): Ditto.
13495 (*vec_extractv2df_0_sse): Ditto.
13496 (sse2_storelpd): Ditto.
13497 (sse2_loadlpd): Ditto.
13498 (sse2_movsd): Ditto.
13499 (*vec_concatv4si): Ditto.
13500 (vec_concatv2di): Ditto.
13502 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13503 for Intel asm dialect.
13504 (mmx_punpcklwd): Ditto.
13505 (mmx_punpckldq): Ditto.
13507 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13508 for intel assembler dialect.
13510 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13513 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13514 for BUILT_IN_C{LZ,LRSB}*.
13515 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13516 * tree-vrp.c (extract_range_basic): Handle
13517 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13518 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13519 fall thru to code calling set_value*.
13520 * builtins.c (expand_builtin): Remove *IMAX cases.
13521 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13522 if width is bigger than 2*HWI.
13524 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13526 PR rtl-optimization/55342
13527 * lra-int.h (lra_subreg_reload_pseudos): New.
13528 * lra.c: Add undoing optional reloads to the block diagram.
13529 (lra_subreg_reload_pseudos): New.
13530 (lra_optional_reload_pseudos): Change comments.
13531 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13532 lra_optional_reload_pseudos after undo transformations.
13533 * lra-assigns.c (pseudo_prefix_title): New.
13534 (lra_setup_reg_renumber): Use it.
13535 (spill_for): Ditto. Check subreg reload pseudos too.
13536 (assign_by_spills): Consider subreg reload pseudos too.
13537 * lra-constraints.c (simplify_operand_subreg): Use
13538 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13539 (curr_insn_transform): Recognize and do optional reloads.
13540 (undo_optional_reloads): New.
13541 (lra_undo_inheritance): Call undo_optional_reloads.
13543 2013-07-05 Thomas Quinot <quinot@adacore.com>
13545 * tree-complex.c (expand_complex_operations_1): Fix typo.
13547 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13549 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13550 (tune_params): New member 'const vec_costs'.
13551 * config/aarch64/aarch64.c (generic_vector_cost): New.
13552 (generic_tunings): New member 'generic_vector_cost'.
13553 (aarch64_builtin_vectorization_cost): New.
13554 (aarch64_add_stmt_cost): New.
13555 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13556 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13558 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13561 * config/i386/predicates.md (vsib_address_operand): Disallow
13562 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13564 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13566 PR middle-end/55030
13567 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13568 expand_builtin_setjmp_receiver.
13569 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13570 with NULL for the label parameter.
13571 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13572 the frame-pointer. Adjust comments.
13573 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13574 only if LABEL is non-NULL.
13576 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13578 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13579 (ARM_ABI_AAPCS64): Ditto.
13581 (ARM_DEFAULT_ABI): Ditto.
13583 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13585 * config/aarch64/aarch64-builtins.c
13586 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13587 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13589 * config/aarch64/aarch64-simd.md
13590 (aarch64_ld1<VALL:mode>): New.
13591 (aarch64_st1<VALL:mode>): Likewise.
13592 * config/aarch64/arm_neon.h
13593 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13595 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13597 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13599 (ix86_option_override_internal):Turn on all -mavx target flags by
13600 default as they are dependent on AVX anyway.
13602 2013-07-02 Cary Coutant <ccoutant@google.com>
13604 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13605 deterministic hash.
13606 (loc_checksum_ordered): Likewise.
13607 (hash_loc_operands): Remove inline keyword.
13609 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13611 PR tree-optimization/57741
13612 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13613 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13614 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13615 Allow REAL_CST step_exprs if flag_associative_math.
13616 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13618 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13620 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13622 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13624 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13626 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13628 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13630 (iorsi3_insn): Likewise.
13631 (arm_xorsi3): Likewise.
13633 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13635 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13637 (attribute "type"): ... here, and prefix with "wmmx_".
13638 (attribute "core_cycles"): Update for attribute changes.
13639 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13640 (tbcstv4hi): Likewise.
13641 (tbcstv2si): Likewise.
13642 (iwmmxt_iordi3): Likewise.
13643 (iwmmxt_xordi3): Likewise.
13644 (iwmmxt_anddi3): Likewise.
13645 (iwmmxt_nanddi3): Likewise.
13646 (iwmmxt_arm_movdi): Likewise.
13647 (iwmmxt_movsi_insn): Likewise.
13648 (mov<mode>_internal): Likewise.
13649 (and<mode>3_iwmmxt): Likewise.
13650 (ior<mode>3_iwmmxt): Likewise.
13651 (xor<mode>3_iwmmxt): Likewise.
13652 (add<mode>3_iwmmxt): Likewise.
13653 (ssaddv8qi3): Likewise.
13654 (ssaddv4hi3): Likewise.
13655 (ssaddv2si3): Likewise.
13656 (usaddv8qi3): Likewise.
13657 (usaddv4hi3): Likewise.
13658 (usaddv2si3): Likewise.
13659 (sub<mode>3_iwmmxt): Likewise.
13660 (sssubv8qi3): Likewise.
13661 (sssubv4hi3): Likewise.
13662 (sssubv2si3): Likewise.
13663 (ussubv8qi3): Likewise.
13664 (ussubv4hi3): Likewise.
13665 (ussubv2si3): Likewise.
13666 (mulv4hi3_iwmmxt): Likewise.
13667 (smulv4hi3_highpart): Likewise.
13668 (umulv4hi3_highpart): Likewise.
13669 (iwmmxt_wmacs): Likewise.
13670 (iwmmxt_wmacsz): Likewise.
13671 (iwmmxt_wmacu): Likewise.
13672 (iwmmxt_wmacuz): Likewise.
13673 (iwmmxt_clrdi): Likewise.
13674 (iwmmxt_clrv8qi): Likewise.
13675 (iwmmxt_clr4hi): Likewise.
13676 (iwmmxt_clr2si): Likewise.
13677 (iwmmxt_uavgrndv8qi3): Likewise.
13678 (iwmmxt_uavgrndv4hi3): Likewise.
13679 (iwmmxt_uavgv8qi3): Likewise.
13680 (iwmmxt_uavgv4hi3): Likewise.
13681 (iwmmxt_tinsrb): Likewise.
13682 (iwmmxt_tinsrh): Likewise.
13683 (iwmmxt_tinsrw): Likewise.
13684 (iwmmxt_textrmub): Likewise.
13685 (iwmmxt_textrmsb): Likewise.
13686 (iwmmxt_textrmuh): Likewise.
13687 (iwmmxt_textrmsh): Likewise.
13688 (iwmmxt_textrmw): Likewise.
13689 (iwmxxt_wshufh): Likewise.
13690 (eqv8qi3): Likewise.
13691 (eqv4hi3): Likewise.
13692 (eqv2si3): Likewise.
13693 (gtuv8qi3): Likewise.
13694 (gtuv4hi3): Likewise.
13695 (gtuv2si3): Likewise.
13696 (gtv8qi3): Likewise.
13697 (gtv4hi3): Likewise.
13698 (gtv2si3): Likewise.
13699 (smax<mode>3_iwmmxt): Likewise.
13700 (umax<mode>3_iwmmxt): Likewise.
13701 (smin<mode>3_iwmmxt): Likewise.
13702 (umin<mode>3_iwmmxt): Likewise.
13703 (iwmmxt_wpackhss): Likewise.
13704 (iwmmxt_wpackwss): Likewise.
13705 (iwmmxt_wpackdss): Likewise.
13706 (iwmmxt_wpackhus): Likewise.
13707 (iwmmxt_wpackwus): Likewise.
13708 (iwmmxt_wpackdus): Likewise.
13709 (iwmmxt_wunpckihb): Likewise.
13710 (iwmmxt_wunpckihh): Likewise.
13711 (iwmmxt_wunpckihw): Likewise.
13712 (iwmmxt_wunpckilb): Likewise.
13713 (iwmmxt_wunpckilh): Likewise.
13714 (iwmmxt_wunpckilw): Likewise.
13715 (iwmmxt_wunpckehub): Likewise.
13716 (iwmmxt_wunpckehuh): Likewise.
13717 (iwmmxt_wunpckehuw): Likewise.
13718 (iwmmxt_wunpckehsb): Likewise.
13719 (iwmmxt_wunpckehsh): Likewise.
13720 (iwmmxt_wunpckehsw): Likewise.
13721 (iwmmxt_wunpckelub): Likewise.
13722 (iwmmxt_wunpckeluh): Likewise.
13723 (iwmmxt_wunpckeluw): Likewise.
13724 (iwmmxt_wunpckelsb): Likewise.
13725 (iwmmxt_wunpckelsh): Likewise.
13726 (iwmmxt_wunpckelsw): Likewise.
13727 (ror<mode>3): Likewise.
13728 (ashr<mode>3_iwmmxt): Likewise.
13729 (lshr<mode>3_iwmmxt): Likewise.
13730 (ashl<mode>3_iwmmxt): Likewise.
13731 (ror<mode>3_di): Likewise.
13732 (ashr<mode>3_di): Likewise.
13733 (lshr<mode>3_di): Likewise.
13734 (ashl<mode>3_di): Likewise.
13735 (iwmmxt_wmadds): Likewise.
13736 (iwmmxt_wmaddu): Likewise.
13737 (iwmmxt_tmia): Likewise.
13738 (iwmmxt_tmiaph): Likewise.
13739 (iwmmxt_tmiabb): Likewise.
13740 (iwmmxt_tmiatb): Likewise.
13741 (iwmmxt_tmiabt): Likewise.
13742 (iwmmxt_tmiatt): Likewise.
13743 (iwmmxt_tmovmskb): Likewise.
13744 (iwmmxt_tmovmskh): Likewise.
13745 (iwmmxt_tmovmskw): Likewise.
13746 (iwmmxt_waccb): Likewise.
13747 (iwmmxt_wacch): Likewise.
13748 (iwmmxt_waccw): Likewise.
13749 (iwmmxt_waligni): Likewise.
13750 (iwmmxt_walignr): Likewise.
13751 (iwmmxt_walignr0): Likewise.
13752 (iwmmxt_walignr1): Likewise.
13753 (iwmmxt_walignr2): Likewise.
13754 (iwmmxt_walignr3): Likewise.
13755 (iwmmxt_wsadb): Likewise.
13756 (iwmmxt_wsadh): Likewise.
13757 (iwmmxt_wsadbz): Likewise.
13758 (iwmmxt_wsadhz): Likewise.
13759 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13760 (iwmmxt_wabsdiffb): Likewise.
13761 (iwmmxt_wabsdiffh): Likewise.
13762 (iwmmxt_wabsdiffw): Likewise.
13763 (iwmmxt_waddsubhx): Likewise
13764 (iwmmxt_wsubaddhx): Likewise.
13765 (addc<mode>3): Likewise.
13766 (iwmmxt_avg4): Likewise.
13767 (iwmmxt_avg4r): Likewise.
13768 (iwmmxt_wmaddsx): Likewise.
13769 (iwmmxt_wmaddux): Likewise.
13770 (iwmmxt_wmaddsn): Likewise.
13771 (iwmmxt_wmaddun): Likewise.
13772 (iwmmxt_wmulwsm): Likewise.
13773 (iwmmxt_wmulwum): Likewise.
13774 (iwmmxt_wmulsmr): Likewise.
13775 (iwmmxt_wmulumr): Likewise.
13776 (iwmmxt_wmulwsmr): Likewise.
13777 (iwmmxt_wmulwumr): Likewise.
13778 (iwmmxt_wmulwl): Likewise.
13779 (iwmmxt_wqmulm): Likewise.
13780 (iwmmxt_wqmulwm): Likewise.
13781 (iwmmxt_wqmulmr): Likewise.
13782 (iwmmxt_wqmulwmr): Likewise.
13783 (iwmmxt_waddbhusm): Likewise.
13784 (iwmmxt_waddbhusl): Likewise.
13785 (iwmmxt_wqmiabb): Likewise.
13786 (iwmmxt_wqmiabt): Likewise.
13787 (iwmmxt_wqmiatb): Likewise.
13788 (iwmmxt_wqmiatt): Likewise.
13789 (iwmmxt_wqmiabbn): Likewise.
13790 (iwmmxt_wqmiabtn): Likewise.
13791 (iwmmxt_wqmiatbn): Likewise.
13792 (iwmmxt_wqmiattn): Likewise.
13793 (iwmmxt_wmiabb): Likewise.
13794 (iwmmxt_wmiabt): Likewise.
13795 (iwmmxt_wmiatb): Likewise.
13796 (iwmmxt_wmiatt): Likewise.
13797 (iwmmxt_wmiabbn): Likewise.
13798 (iwmmxt_wmiabtn): Likewise.
13799 (iwmmxt_wmiatbn): Likewise.
13800 (iwmmxt_wmiattn): Likewise.
13801 (iwmmxt_wmiawbb): Likewise.
13802 (iwmmxt_wmiawbt): Likewise.
13803 (iwmmxt_wmiawtb): Likewise.
13804 (iwmmxt_wmiawtt): Likewise.
13805 (iwmmxt_wmiawbbn): Likewise.
13806 (iwmmxt_wmiawbtn): Likewise.
13807 (iwmmxt_wmiawtbn): Likewise.
13808 (iwmmxt_wmiawttn): Likewise.
13809 (iwmmxt_wmerge): Likewise.
13810 (iwmmxt_tandc<mode>3): Likewise.
13811 (iwmmxt_torc<mode>3): Likewise.
13812 (iwmmxt_torvsc<mode>3): Likewise.
13813 (iwmmxt_textrc<mode>3): Likewise.
13814 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13815 (wmmxt_pack): Likewise.
13816 (wmmxt_mult_c1): Likewise.
13817 (wmmxt_mult_c2): Likewise.
13818 (wmmxt_alu_c1): Likewise.
13819 (wmmxt_alu_c2): Likewise.
13820 (wmmxt_alu_c3): Likewise.
13821 (wmmxt_transfer_c1): Likewise.
13822 (wmmxt_transfer_c2): Likewise.
13823 (wmmxt_transfer_c3): Likewise.
13824 (marvell_f_iwmmxt_wstr): Likewise.
13825 (marvell_f_iwmmxt_wldr): Likewise.
13827 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13829 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13831 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13834 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13835 * lra-constraints.c (need_for_split_p): Check call used hard regs
13836 living through calls.
13838 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13839 call used regs for call insn.
13841 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13844 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13845 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13847 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13849 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13851 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13853 * lra-constraints.c (need_for_split_p): Check call used hard regs
13854 living through calls.
13856 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13859 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13860 to tie with any other modes. Eliminate Altivec vector mode tests,
13861 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13862 code, to return 0 if testing MODE2 for a condition, if we've
13863 already tested MODE1 for the same condition.
13865 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13867 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13870 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13872 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13873 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13875 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13877 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13879 (aarch64_symbolic_constant_p): Remove.
13880 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13881 static. Fix line length and white space.
13882 (aarch64_symbolic_constant_p): Remove.
13883 * config/aarch64/predicates.md (aarch64_valid_symref):
13884 Use aarch64_classify_symbol_expression.
13886 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13888 * config/arm/constraints.md (Ts): New constraint.
13889 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13891 (compare_scc): Use "Ts" constraint for operand 0.
13892 (ior_scc_scc): Likewise.
13893 (and_scc_scc): Likewise.
13894 (and_scc_scc_nodom): Likewise.
13895 (ior_scc_scc_cmp): Likewise for operand 7.
13896 (and_scc_scc_cmp): Likewise.
13897 * config/arm/thumb2.md (thumb2_movsi_insn):
13898 Add alternatives for 16-bit encodings.
13899 (thumb2_movhi_insn): Likewise.
13900 (thumb2_movsicc_insn): Likewise.
13901 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13902 (thumb2_negscc): Use "Ts" constraint.
13903 Move mvn instruction outside cond_exec block.
13904 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13905 for 16-bit encodings.
13907 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13909 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13911 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13912 (mulsi3subsi): Likewise.
13913 (mulsidi3adddi): Likewise.
13914 (mulsidi3_v6): Likewise.
13915 (umulsidi3_v6): Likewise.
13916 (umulsidi3adddi_v6): Likewise.
13917 (smulsi3_highpart_v6): Likewise.
13918 (umulsi3_highpart_v6): Likewise.
13919 (mulhisi3tb): Likewise.
13920 (mulhisi3bt): Likewise.
13921 (mulhisi3tt): Likewise.
13922 (maddhisi4): Likewise.
13923 (maddhisi4tb): Likewise.
13924 (maddhisi4tt): Likewise.
13925 (maddhidi4): Likewise.
13926 (maddhidi4tb): Likewise.
13927 (maddhidi4tt): Likewise.
13928 (zeroextractsi_compare0_scratch): Likewise.
13929 (insv_zero): Likewise.
13930 (insv_t2): Likewise.
13931 (anddi_notzesidi_di): Likewise.
13932 (anddi_notsesidi_di): Likewise.
13933 (andsi_notsi_si): Likewise.
13934 (iordi_zesidi_di): Likewise.
13935 (xordi_zesidi_di): Likewise.
13936 (andsi_iorsi3_notsi): Likewise.
13937 (smax_0): Likewise.
13938 (smax_m1): Likewise.
13939 (smin_0): Likewise.
13940 (not_shiftsi): Likewise.
13941 (unaligned_loadsi): Likewise.
13942 (unaligned_loadhis): Likewise.
13943 (unaligned_loadhiu): Likewise.
13944 (unaligned_storesi): Likewise.
13945 (unaligned_storehi): Likewise.
13946 (extv_reg): Likewise.
13947 (extzv_t2): Likewise.
13948 (divsi3): Likewise.
13949 (udivsi3): Likewise.
13950 (arm_zero_extendhisi2addsi): Likewise.
13951 (arm_zero_extendqisi2addsi): Likewise.
13952 (compareqi_eq0): Likewise.
13953 (arm_extendhisi2_v6): Likewise.
13954 (arm_extendqisi2addsi): Likewise.
13955 (arm_movt): Likewise.
13956 (thumb2_ldrd): Likewise.
13957 (thumb2_ldrd_base): Likewise.
13958 (thumb2_ldrd_base_neg): Likewise.
13959 (thumb2_strd): Likewise.
13960 (thumb2_strd_base): Likewise.
13961 (thumb2_strd_base_neg): Likewise.
13962 (arm_negsi2): Add alternative for 16-bit encoding.
13963 (arm_one_cmplsi2): Likewise.
13965 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13967 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
13968 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
13969 (movdfcc): Likewise.
13970 * config/arm/vfp.md (*thumb2_movsf_vfp):
13971 Disable predication for arm_restrict_it.
13972 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
13973 (*thumb2_movdfcc_vfp): Likewise.
13974 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
13975 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
13976 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
13977 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
13978 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
13979 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
13980 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
13981 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
13982 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
13983 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
13984 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
13985 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
13986 Disable predication for arm_restrict_it.
13988 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
13990 * config/i386/bmiintrin.h (_bextr_u32): New.
13991 (_bextr_u64): Ditto.
13993 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
13995 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
13996 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
13997 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
13998 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
13999 * config/mips/n32-elf.h: ...this new file.
14001 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
14004 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
14005 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
14007 2013-06-27 Catherine Moore <clm@codesourcery.com>
14009 * config/mips/mips-tables.opt: Regenerate.
14010 * config/mips/mips-cpus.def: Add m14ke and m14kec.
14011 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
14012 * doc/invoke.texi: Add -m14kc.
14014 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14017 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
14018 constraints of operand 1 and 2.
14021 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
14022 to match RTL canonicalization. Swap predicates and
14023 constraints of operand 1 and 2.
14025 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
14027 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
14028 Process OP_INOUT regs for splitting too.
14030 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14032 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
14033 decl before the loop, initialize to NULL.
14034 (vectorizable_load): Initialize ptr_incr to NULL.
14036 2013-06-27 Martin Jambor <mjambor@suse.cz>
14039 * ipa-ref.h (ipa_maybe_record_reference): Declare.
14040 * ipa-ref.c (ipa_maybe_record_reference): New function.
14041 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
14042 * ipa-cp.c (create_specialized_node): Record potential references from
14044 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
14046 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14048 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
14049 parameter 'mode' of type 'enum machine_mode mode'; change to pass
14050 'mode' to force_reg.
14051 (aarch64_add_offset): Update calls to aarch64_force_temporary.
14052 (aarch64_expand_mov_immediate): Likewise.
14054 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14056 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
14057 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
14059 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14061 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
14062 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14063 (struct machine_function): Add tbegin_p.
14064 (s390_canonicalize_comparison): Fold CC mode compares to
14065 conditional jump if possible.
14066 (s390_emit_jump): Return the emitted jump.
14067 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
14068 Handle CCRAWmode compares.
14069 (s390_option_override): Default to -mhtm if available.
14070 (s390_reg_clobbered_rtx): Handle floating point regs as well.
14071 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
14072 FPRs instead of df_regs_ever_live_p.
14073 (s390_optimize_nonescaping_tx): New function.
14074 (s390_init_frame_layout): Extend clobbered_regs array to cover
14076 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
14077 (s390_expand_tbegin): New function.
14078 (enum s390_builtin): New enum definition.
14079 (code_for_builtin): New array definition.
14080 (s390_init_builtins): New function.
14081 (s390_expand_builtin): New function.
14082 (TARGET_INIT_BUILTINS): Define.
14083 (TARGET_EXPAND_BUILTIN): Define.
14084 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
14085 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
14086 (s390_alc_comparison): Likewise.
14087 * config/s390/s390-modes.def: Add CCRAWmode.
14088 * config/s390/s390.h (processor_flags): Add PF_TX.
14089 (TARGET_CPU_HTM): Define macro.
14090 (TARGET_HTM): Define macro.
14091 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
14092 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
14093 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14094 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
14095 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
14096 (TBEGIN_MASK, TBEGINC_MASK): New constants.
14097 ("*cc_to_int"): Move up.
14098 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
14099 constants other than 0.
14100 ("*ccraw_to_int"): New insn and splitter definition.
14101 ("tbegin", "tbegin_nofloat", "tbegin_retry")
14102 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
14103 ("tx_assist"): New expander.
14104 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
14105 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
14106 * config/s390/s390.opt: Add -mhtm option.
14107 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
14108 * config/s390/htmxlintrin.h: New file.
14109 * config/s390/htmintrin.h: New file.
14110 * config/s390/s390intrin.h: New file.
14111 * doc/extend.texi: Document htm builtins.
14112 * config.gcc: Add the new header files to extra_headers.
14114 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
14116 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
14117 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
14119 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
14120 Pat Haugen <pthaugen@us.ibm.com>
14121 Peter Bergner <bergner@vnet.ibm.com>
14123 * config/rs6000/power8.md: New.
14124 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
14125 setting for power8 entry.
14126 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
14127 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
14128 test for Power4/Power5 only.
14129 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
14131 (force_new_group): Adjust comment.
14132 * config/rs6000/rs6000.md: Include power8.md.
14134 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14136 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14137 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14139 (tune_params): Update comment.
14140 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14141 (arm_max_conditional_execute): New function.
14142 (thumb2_final_prescan_insn): Use max_insn_skipped and
14143 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14145 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14147 PR tree-optimization/57705
14148 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14149 SSA_NAME step, provided that it is not defined inside the loop.
14150 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14151 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14153 2013-06-25 Martin Jambor <mjambor@suse.cz>
14155 PR middle-end/57670
14156 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14157 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14159 (ipa_note_param_call): Initialize member_ptr flag.
14160 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14161 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14162 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14163 (ipa_read_indirect_edge_info): Likewise.
14165 2013-06-25 Richard Biener <rguenther@suse.de>
14167 PR middle-end/56977
14168 * passes.c (init_optimization_passes): Move pass_fold_builtins
14169 and pass_dce earlier with -Og.
14171 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14173 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14174 <BIT_FIELD_REF>: Remove trailing TAB.
14175 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14178 2013-06-24 Martin Jambor <mjambor@suse.cz>
14180 PR tree-optimization/57358
14181 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14182 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14183 (ipa_analyze_params_uses): Generate pessimistic info when true.
14185 2013-06-24 Martin Jambor <mjambor@suse.cz>
14187 PR tree-optimization/57539
14188 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14189 global.inlined_to of the new node to it. All callers changed.
14190 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14191 inlining_into, pass it to cgraph_clone_node.
14192 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14193 ipa_free_edge_args_substructures.
14194 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14195 rdesc linked list. Do not assert rdesc edges have inlined caller.
14196 Assert we have found an rdesc in the rdesc list.
14198 2013-06-24 Richard Biener <rguenther@suse.de>
14200 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14201 (pointer_set_lookup): Declare.
14202 (class pointer_map): New template class implementing a
14203 generic pointer to T map.
14204 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14205 pointer_map<T>::contains, pointer_map<T>::insert,
14206 pointer_map<T>::traverse): New functions.
14207 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14208 (pointer_set_lookup): New function.
14209 (pointer_set_contains): Use pointer_set_lookup.
14210 (pointer_set_insert): Likewise.
14211 (insert_aux): Remove.
14212 (struct pointer_map_t): Embed a pointer_set_t.
14213 (pointer_map_create): Adjust.
14214 (pointer_map_destroy): Likewise.
14215 (pointer_map_contains): Likewise.
14216 (pointer_map_insert): Likewise.
14217 (pointer_map_traverse): Likewise.
14218 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14219 pointer_map<unsigned> instead of a pointer_map_t.
14220 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14221 (streamer_tree_cache_lookup): Likewise.
14222 (streamer_tree_cache_create): Likewise.
14223 (streamer_tree_cache_delete): Likewise.
14224 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14225 pointer_map<unsigned> instead of a pointer_map_t.
14226 (lto_init_tree_ref_encoder): Adjust.
14227 (lto_destroy_tree_ref_encoder): Likewise.
14228 * lto-section-out.c (lto_output_decl_index): Likewise.
14229 (lto_record_function_out_decl_state): Likewise.
14230 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14232 2013-06-24 Richard Biener <rguenther@suse.de>
14234 PR tree-optimization/57488
14235 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14237 2013-06-24 Alan Modra <amodra@gmail.com>
14239 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14240 (gen_easy_altivec_constant): Likewise.
14241 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14242 easy_vector_constant_msb): Likewise.
14244 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14247 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14248 add missing return true.
14250 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14253 * config/sh/predicates.md (general_extend_operand): Invoke
14254 general_movsrc_operand for memory operands.
14255 (general_movsrc_operand): Allow reg+reg addressing, do not use
14256 general_operand for memory operands.
14258 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14260 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14261 when current target options does not apply.
14262 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14263 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14264 * config/i386/bmiintrin.h: Pass appropriate target
14265 attributes to header.
14266 * config/i386/mmintrin.h: Ditto.
14267 * config/i386/nmmintrin.h: Ditto.
14268 * config/i386/avx2intrin.h: Ditto.
14269 * config/i386/fxsrintrin.h: Ditto.
14270 * config/i386/tbmintrin.h: Ditto.
14271 * config/i386/xsaveintrin.h: Ditto.
14272 * config/i386/f16cintrin.h: Ditto.
14273 * config/i386/xtestintrin.h: Ditto.
14274 * config/i386/xsaveoptintrin.h: Ditto.
14275 * config/i386/bmi2intrin.h: Ditto.
14276 * config/i386/lzcntintrin.h: Ditto.
14277 * config/i386/smmintrin.h: Ditto.
14278 * config/i386/wmmintrin.h: Ditto.
14279 * config/i386/x86intrin.h: Remove all header include guards.
14280 * config/i386/prfchwintrin.h: Ditto.
14281 * config/i386/pmmintrin.h: Ditto.
14282 * config/i386/tmmintrin.h: Ditto.
14283 * config/i386/xmmintrin.h: Ditto.
14284 * config/i386/popcntintrin.h: Ditto.
14285 * config/i386/rdseedintrin.h: Ditto.
14286 * config/i386/ammintrin.h: Ditto.
14287 * config/i386/emmintrin.h: Ditto.
14288 * config/i386/immintrin.h: Remove all header include guards.
14289 * config/i386/fma4intrin.h: Ditto.
14290 * config/i386/lwpintrin.h: Ditto.
14291 * config/i386/xopintrin.h: Ditto.
14292 * config/i386/ia32intrin.h: Ditto.
14293 * config/i386/avxintrin.h: Ditto.
14294 * config/i386/rtmintrin.h: Ditto.
14295 * config/i386/fmaintrin.h: Ditto.
14296 * config/i386/mm3dnow.h: Ditto.
14298 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14300 * common/config/i386/i386-common.c: Handle LZCNT.
14302 2013-06-22 Andi Kleen <ak@linux.intel.com>
14304 * doc/extend.texi: Use __atomic_store_n instead of
14305 __atomic_store in HLE example.
14307 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14309 * config/sh/sh.c: Remove <cstdlib> workaround.
14311 2013-06-21 Andi Kleen <ak@linux.intel.com>
14313 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14315 2013-06-21 Andi Kleen <ak@linux.intel.com>
14317 * doc/extend.texi: Document that __atomic_clear and
14318 __atomic_test_and_set should only be used with bool.
14320 2013-06-20 Jan Hubicka <jh@suse.cz>
14322 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14323 types_same_for_odr.
14324 * tree.c (decls_same_for_odr): New function.
14325 (same_for_edr): New function.
14326 (types_same_for_odr): New function.
14327 (get_binfo_at_offset): Use it.
14328 * tree.h (types_same_for_odr): Declare.
14330 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14331 Jason Merrill <jason@redhat.com>
14333 * system.h: Include <cstdlib> as well as <stdlib.h>.
14335 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14338 * config/i386/i386.c (construct_container): Report error if
14339 long double is used with disabled x87 float returns.
14341 2013-06-20 Jan Hubicka <jh@suse.cz>
14343 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14345 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14347 PR rtl-optimization/57425
14348 PR rtl-optimization/57569
14349 * alias.c (write_dependence_p): Remove parameters mem_mode and
14350 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14351 Changed all callers.
14352 (canon_anti_dependence): Get comments and semantics in sync.
14353 Add parameter mem_canonicalized. Changed all callers.
14354 * rtl.h (canon_anti_dependence): Update prototype.
14356 2013-06-20 Richard Biener <rguenther@suse.de>
14358 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14359 byte case, inline streamer_read_uchar and defer section
14362 2013-06-20 Richard Biener <rguenther@suse.de>
14364 PR tree-optimization/57584
14365 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14366 SSA names into the expanded expression that take part in
14367 abnormal coalescing.
14369 2013-06-19 Sharad Singhai <singhai@google.com>
14371 * gcov.c (print_usage): Handle new option.
14372 (process_args): Ditto.
14373 (get_gcov_intermediate_filename): New function.
14374 (output_intermediate_file): New function.
14375 (output_gcov_file): New function
14376 (generate_results): Handle new option.
14377 (release_function): Relase demangled name.
14378 (read_graph_file): Handle demangled name.
14379 (output_lines): Ditto.
14380 * doc/gcov.texi: Document gcov intermediate format.
14382 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14385 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14386 (lra_emit_add): Use the functions. Add comment about Y as an
14389 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14392 * collect2.c (collect_atexit): New.
14393 (collect_exit): Delete.
14394 (main): Register collect_atexit with atexit.
14395 (collect_wait): Change collect_exit to exit.
14397 * collect2.h (collect_exit): Delete.
14398 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14400 2013-06-19 Wei Mi <wmi@google.com>
14402 PR rtl-optimization/57518
14403 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14404 if regno is used in paradoxical subreg.
14405 (update_equiv_regs): Check pdx_subregs[regno] before
14406 set a reg to be equivalent with a mem.
14408 2013-06-19 Matthias Klose <doko@ubuntu.com>
14411 * file-find.h (find_a_file): Add a mode parameter.
14412 * file-find.c (find_a_file): Likewise.
14413 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14414 with X_OK for the executables.
14415 * collect2.c (main): Call find_a_file with X_OK.
14417 2013-06-19 Steve Ellcey <sellcey@mips.com>
14420 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14421 Use NEXT_INSN instead of next_real_insn.
14423 2013-06-19 Jan Hubicka <jh@suse.cz>
14425 * cgraph.h (const_value_known_p): Replace by ...
14426 (ctor_for_folding): .. this one.
14427 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14428 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14429 * expr.c (expand_expr_real_1): Likewise.
14430 (string_constant): Likewise.
14431 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14432 * ipa.c (process_references): Likewise.
14433 (symtab_remove_unreachable_nodes): Likewise.
14434 * ipa-inline-analysis.c (param_change_prob): Likewise.
14435 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14436 (get_base_constructor): Likwise.
14437 * varpool.c (varpool_remove_node): Likewise.
14438 (varpool_remove_initializer): LIkewise.
14439 (dump_varpool_node): LIkwise.
14440 (const_value_known_p): Rewrite to ...
14441 (ctor_for_folding): ... this one.
14443 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14446 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14447 PERSONALITY in $PATH derived prefixes.
14449 2013-06-19 Jeff Law <law@redhat.com>
14451 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14454 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14455 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14458 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14460 * config/arm/vfp.md: Move VFP instruction classification documentation
14462 * config/arm/arm.md: ... here. Update instruction classification
14465 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14467 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14468 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14469 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14471 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14474 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14475 with NEXT_INSN. Use tablejump_p to check for jump table data
14478 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14481 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14482 that now in C++ the value is correct per the C++ standards.
14484 2013-06-19 Richard Biener <rguenther@suse.de>
14486 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14487 for global context.
14489 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14492 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14495 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14496 with next_active_insn.
14498 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14500 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14501 functions are inlined during failures to flag an error.
14502 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14503 in early inline pass.
14505 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14507 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14510 2013-06-18 Julian Brown <julian@codesourcery.com>
14512 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14513 Permit virtual register pre-reload if !strict.
14514 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14516 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14518 * config/arm/neon.md (movmisalign<mode>): Use
14519 neon_perm_struct_or_reg_operand instead of
14520 neon_struct_or_register_operand.
14521 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14522 neon_permissive_struct_operand instead of neon_struct_operand.
14523 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14524 neon_vector_mem_operand.
14525 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14526 neon_vector_mem_operand.
14527 (neon_permissive_struct_operand): New.
14528 (neon_struct_or_register_operand): Rename to...
14529 (neon_perm_struct_or_reg_operand): This. Adjust call to
14530 neon_vector_mem_operand.
14532 2013-06-18 Richard Biener <rguenther@suse.de>
14534 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14535 * lto-streamer.h: Include pointer-set.h.
14536 (struct lto_decl_slot): Remove.
14537 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14538 Remove next_index entry.
14539 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14540 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14541 (lto_init_tree_ref_encoder): Adjust.
14542 (lto_destroy_tree_ref_encoder): Likewise.
14543 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14544 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14545 (lto_output_decl_index): Adjust.
14546 (lto_new_out_decl_state): Likewise.
14547 (lto_record_function_out_decl_state): Likewise.
14548 * lto-streamer-out.c (copy_function): Likewise.
14550 2013-06-18 Richard Biener <rguenther@suse.de>
14552 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14553 * cgraphunit.c: Include cfgloop.h.
14554 (init_lowered_empty_function): Initialize the loop tree.
14555 (assemble_thunk): Insert new BBs into loops.
14557 2013-06-18 Richard Biener <rguenther@suse.de>
14559 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14560 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14561 the map from cache entry to cache index optional.
14562 (streamer_tree_cache_replace_tree): Adjust accordingly.
14563 (streamer_tree_cache_append): Likewise.
14564 (streamer_tree_cache_delete): Likewise.
14565 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14566 streamer cache map from cache entry to cache index.
14567 * lto-streamer-out.c (create_output_block): Adjust.
14569 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14571 * config/arm/arm.md (attribute "insn"): Move multiplication and
14572 division attributes to...
14573 (attribute "type"): ... here. Remove mult.
14574 (attribute "mul32"): New attribute.
14575 (attribute "mul64"): Add umaal.
14576 (*arm_mulsi3): Update attributes.
14577 (*arm_mulsi3_v6): Likewise.
14578 (*thumb_mulsi3): Likewise.
14579 (*thumb_mulsi3_v6): Likewise.
14580 (*mulsi3_compare0): Likewise.
14581 (*mulsi3_compare0_v6): Likewise.
14582 (*mulsi_compare0_scratch): Likewise.
14583 (*mulsi_compare0_scratch_v6): Likewise.
14584 (*mulsi3addsi): Likewise.
14585 (*mulsi3addsi_v6): Likewise.
14586 (*mulsi3addsi_compare0): Likewise.
14587 (*mulsi3addsi_compare0_v6): Likewise.
14588 (*mulsi3addsi_compare0_scratch): Likewise.
14589 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14590 (*mulsi3subsi): Likewise.
14591 (*mulsidi3adddi): Likewise.
14592 (*mulsi3addsi_v6): Likewise.
14593 (*mulsidi3adddi_v6): Likewise.
14594 (*mulsidi3_nov6): Likewise.
14595 (*mulsidi3_v6): Likewise.
14596 (*umulsidi3_nov6): Likewise.
14597 (*umulsidi3_v6): Likewise.
14598 (*umulsidi3adddi): Likewise.
14599 (*umulsidi3adddi_v6): Likewise.
14600 (*smulsi3_highpart_nov6): Likewise.
14601 (*smulsi3_highpart_v6): Likewise.
14602 (*umulsi3_highpart_nov6): Likewise.
14603 (*umulsi3_highpart_v6): Likewise.
14604 (mulhisi3): Likewise.
14605 (*mulhisi3tb): Likewise.
14606 (*mulhisi3bt): Likewise.
14607 (*mulhisi3tt): Likewise.
14608 (maddhisi4): Likewise.
14609 (*maddhisi4tb): Likewise.
14610 (*maddhisi4tt): Likewise.
14611 (maddhidi4): Likewise.
14612 (*maddhidi4tb): Likewise.
14613 (*maddhidi4tt): Likewise.
14614 (divsi3): Likewise.
14615 (udivsi3): Likewise.
14616 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14617 (thumb2_mulsi_short_compare0): Likewise.
14618 (thumb2_mulsi_short_compare0_scratch): Likewise.
14619 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14620 (1020mult2): Likewise.
14621 (1020mult3): Likewise.
14622 (1020mult4): Likewise.
14623 (1020mult5): Likewise.
14624 (1020mult6): Likewise.
14625 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14627 (cortex_a15_mult64): Likewise.
14628 (cortex_a15_sdiv): Likewise.
14629 (cortex_a15_udiv): Likewise.
14630 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14636 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14637 (pj4_ir_div): Likewise.
14638 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14639 (11_mult2): Likewise.
14640 (11_mult3): Likewise.
14641 (11_mult4): Likewise.
14642 (11_mult5): Likewise.
14643 (11_mult6): Likewise.
14644 (11_mult7): Likewise.
14645 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14646 (cortex_a8_mla): Likewise.
14647 (cortex_a8_mull): Likewise.
14648 (cortex_a8_smulwy): Likewise.
14649 (cortex_a8_smlald): Likewise.
14650 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14651 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14652 (cortex_r4_mul_3): Likewise.
14653 (cortex_r4_mla_4): Likewise.
14654 (cortex_r4_mla_3): Likewise.
14655 (cortex_r4_smlald): Likewise.
14656 (cortex_r4_mull): Likewise.
14657 (cortex_r4_sdiv): Likewise.
14658 (cortex_r4_udiv): Likewise.
14659 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14660 (cortex_a7_idiv): Likewise.
14661 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14662 (9_mult2): Likewise.
14663 (9_mult3): Likewise.
14664 (9_mult4): Likewise.
14665 (9_mult5): Likewise.
14666 (9_mult6): Likewise.
14667 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14668 (cortex_a53_sdiv): Likewise.
14669 (cortex_a53_udiv): Likewise.
14670 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14671 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14672 (mp626_mult2): Likewise.
14673 (mp626_mult3): Likewise.
14674 (mp626_mult4): Likewise.
14675 * config/arm/fa526.md (526_mult1): Update attribute change.
14676 (526_mult2): Likewise.
14677 * config/arm/arm-generic.md (mult): Update attribute change.
14678 (mult_ldsched_strongarm): Likewise.
14679 (mult_ldsched): Likewise.
14680 (multi_cycle): Likewise.
14681 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14682 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14683 (606te_mult2): Likewise.
14684 (606te_mult3): Likewise.
14685 (606te_mult4): Likewise.
14686 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14687 (cortex_a9_mac16): Likewise.
14688 (cortex_a9_multiply): Likewise.
14689 (cortex_a9_mac): Likewise.
14690 (cortex_a9_multiply_long): Likewise.
14691 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14692 (626te_mult2): Likewise.
14693 (626te_mult3): Likewise.
14694 (626te_mult4): Likewise.
14696 2013-06-18 Richard Biener <rguenther@suse.de>
14699 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14701 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14704 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14705 with next_active_insn.
14707 2013-06-18 Alan Modra <amodra@gmail.com>
14709 * config/rs6000/rs6000.h (enum data_align): New.
14710 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14711 (DATA_ABI_ALIGNMENT): Define.
14712 (CONSTANT_ALIGNMENT): Correct comment.
14713 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14714 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14716 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14718 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14719 ATTRIBUTE_UNUSED marking.
14721 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14723 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14724 alternative and update.
14725 (aarch64_dup_lanedi): Delete.
14726 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14727 * config/aarch64/aarch64-simd-builtins.def: Update.
14729 2013-06-17 Richard Biener <rguenther@suse.de>
14731 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14732 (lto_input_scc): Declare.
14733 (lto_input_tree_1): Likewise.
14734 (struct lto_stats_d): Add num_tree_bodies_output and
14735 num_pickle_refs_output.
14736 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14737 (lto_read_tree_1): Split out from ...
14738 (lto_read_tree): ... this.
14739 (lto_input_scc): New function.
14740 (lto_input_tree_1): Split out from ...
14741 (lto_input_tree): ... this. Handle LTO_tree_scc.
14742 (lto_data_in_create): Create the streamer cache without hashes.
14743 * lto-streamer-out.c (create_output_block): Create the streamer
14744 cache with hashes when not doing WPA.
14745 (lto_write_tree_1): Split out from ...
14746 (lto_write_tree): ... this.
14747 (get_symbol_initial_value): New function.
14748 (lto_output_tree_1): Split out from ...
14749 (lto_output_tree): ... this. Write trees as series of SCCs
14750 using a DFS walk via DFS_write_tree.
14751 (struct sccs, struct scc_entry): New types.
14752 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14753 (DFS_write_tree_body): New function.
14754 (DFS_write_tree): Likewise.
14755 (hash_tree): Likewise.
14756 (scc_entry_compare): Likewise.
14757 (hash_scc): Likewise.
14758 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14759 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14760 TREE_CHAIN as regular reference.
14761 (streamer_read_integer_cst): Remove.
14762 (streamer_get_pickled_tree): Adjust.
14763 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14764 of DECL_EXTERNALs in BLOCK_VARS for now.
14765 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14767 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14768 Add hash value argument and record that if hashes are recorded
14770 (streamer_tree_cache_insert_1): Adjust.
14771 (streamer_tree_cache_insert): Likewise.
14772 (streamer_tree_cache_insert_at): Rename to ...
14773 (streamer_tree_cache_replace_tree): ... this and adjust.
14774 (streamer_tree_cache_append): Adjust.
14775 (record_common_node): Likewise.
14776 (streamer_tree_cache_create): Add argument whether to
14777 record hash values together with trees.
14778 (streamer_tree_cache_delete): Adjust.
14779 * tree-streamer.h (struct streamer_tree_cache_d): Add
14781 (streamer_read_integer_cst): Remove.
14782 (streamer_tree_cache_insert): Adjust.
14783 (streamer_tree_cache_append): Likewise.
14784 (streamer_tree_cache_insert_at): Rename to ...
14785 (streamer_tree_cache_replace_tree): ... this and adjust.
14786 (streamer_tree_cache_create): Add argument whether to record hashes.
14787 (streamer_tree_cache_get): Rename to ...
14788 (streamer_tree_cache_get_tree): ... this.
14789 (streamer_tree_cache_get_hash): New function.
14790 * tree.c (cache_integer_cst): New function.
14791 * tree.h (cache_integer_cst): Declare.
14792 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14793 * lto-symtab.c (lto_varpool_replace_node): Only release
14794 DECL_INITIAL of non-prevailing decls.
14795 * varpool.c (varpool_remove_initializer): Do not release
14796 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14798 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14800 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14801 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14802 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14803 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14804 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14805 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14806 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14807 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14808 instead of TARGET_64BIT.
14809 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14810 Require ISA_HAS_<D>DIV.
14812 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14814 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14815 (mips*-*-linux*): Move default with_llsc setting to where other
14817 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14819 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14820 Likewise. Remove default with_tune setting. Move default float
14821 setting to its own block. Handle with_llsc in the same block as above.
14823 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14825 PR rtl-optimization/57425
14826 PR rtl-optimization/57569
14827 * alias.c (write_dependence_p): Add new parameters mem_mode,
14828 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14829 Changed all callers.
14830 (canon_anti_dependence): New function.
14831 * cse.c (check_dependence): Use canon_anti_dependence.
14832 * cselib.c (cselib_invalidate_mem): Likewise.
14833 * rtl.h (canon_anti_dependence): Declare.
14835 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14837 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14838 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14839 ".set mips3" for 64-bit targets.
14841 2013-06-15 Dehao Chen <dehao@google.com>
14843 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14844 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14845 (gimple_check_call_args): Likewise.
14846 * value-prof.c (check_ic_target): Likewise.
14847 * ipa-inline.c (early_inliner): Likewise.
14848 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14849 * cgraph.c (cgraph_create_edge_1): Likewise.
14850 (cgraph_make_edge_direct): Likewise.
14852 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14855 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14856 rs6000_output_move_128bit to handle emitting quad memory
14857 operations. Set attribute length to 8 bytes.
14859 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14861 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14863 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14864 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14865 (aarch64_<su>mlsl<mode>): Likewise.
14867 2013-06-14 Mike Stump <mikestump@comcast.net>
14869 * Makefile.in (TARGET_H): Add insn-codes.h.
14871 2013-06-14 Alan Modra <amodra@gmail.com>
14873 PR middle-end/57134
14874 PR middle-end/57586
14875 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14876 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14877 bitfield expansion when EXPAND_MEMORY.
14878 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14880 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14882 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14883 test for clearing quad memory on 32-bit later.
14885 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14887 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14888 (fold_negate_expr): Likewise.
14889 (fold_real_zero_addition_p): Handle vectors.
14890 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14892 2013-06-14 Alan Modra <amodra@gmail.com>
14894 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14896 2013-06-13 Jan Hubicka <jh@suse.cz>
14898 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14899 Local comdats are not externally visible.
14900 * symtab.c (dump_symtab_base): Dump externally visible.
14901 (verify_symtab_base): Verify back links in the symtab hash.
14903 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14905 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14906 CONVERT_EXPR as equal nodes.
14908 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14910 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14912 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14914 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14915 Generalize to complex and vector.
14916 * tree.c (build_all_ones_cst): New function.
14917 * tree.h (build_all_ones_cst): Declare it.
14919 2013-06-13 Alan Modra <amodra@gmail.com>
14921 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14922 * config/rs6000/rs6000.md (signbittf2): New insn.
14923 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14924 (abstf2_internal, cmptf_internal2): Likewise.
14925 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14927 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14928 Pat Haugen <pthaugen@us.ibm.com>
14929 Peter Bergner <bergner@vnet.ibm.com>
14931 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14932 power8 byte, half-word, and quad-word atomic instructions.
14933 (emit_store_conditional): Likewise.
14934 (rs6000_expand_atomic_compare_and_swap): Likewise.
14935 (rs6000_expand_atomic_op): Likewise.
14937 * config/rs6000/sync.md (larx): Add new modes for power8.
14939 (AINT): New mode iterator to include TImode as well as normal
14940 integer modes on power8.
14941 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14942 that VSX registers are not considered. Use AINT mode iterator
14943 instead of INT1 to allow inclusion of quad word atomic operations
14945 (load_locked<mode>): Likewise.
14946 (store_conditional<mode>): Likewise.
14947 (atomic_compare_and_swap<mode>): Likewise.
14948 (atomic_exchange<mode>): Likewise.
14949 (atomic_nand<mode>): Likewise.
14950 (atomic_fetch_<fetchop_name><mode>): Likewise.
14951 (atomic_nand_fetch<mode>): Likewise.
14952 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14954 (ATOMIC): On power8, add QImode, HImode modes.
14955 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
14956 modes that promote to SImode.
14957 (load_lockedti): Convert TImode arguments to PTImode, so that we
14958 get a guaranteed even/odd register pair.
14959 (load_lockedpti): Likewise.
14960 (store_conditionalti): Likewise.
14961 (store_conditionalpti): Likewise.
14963 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
14964 atomic load/store instructions.
14967 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
14969 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
14971 (insn_count): New attribute, with most cases extracted from...
14972 (length): ...here. Redefine most cases in terms of insn_count.
14973 (single_insn): Delete.
14974 (can_delay): Use insn_count to check for single instructions.
14975 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
14976 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
14977 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
14978 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
14979 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
14980 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
14981 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
14982 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
14983 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
14984 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
14985 rather than "length".
14986 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
14987 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
14988 Use "insn_count" rather than "length".
14989 * config/mips/mips-dsp.md
14990 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
14991 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
14994 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
14996 PR tree-optimization/57361
14997 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
14999 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
15001 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
15003 (aarch64_simd_combine<mode>): New instruction expansion.
15004 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
15005 function prototype.
15006 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
15007 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
15009 2013-06-12 Jan Hubicka <jh@suse.cz>
15011 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
15012 decl has when in streaming stage.
15013 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
15014 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
15016 2013-06-12 Roland Stigge <stigge@antcom.de>
15019 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
15021 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15023 PR tree-optimization/57537
15024 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
15025 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
15027 2013-06-12 Richard Biener <rguenther@suse.de>
15029 * data-streamer.h (streamer_write_char_stream): CSE
15030 obs->current_pointer.
15031 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
15032 streamer_write_char_stream manually and optimize the resulting loop.
15033 (streamer_write_hwi_stream): Likewise.
15035 2013-06-12 Jan Hubicka <jh@suse.cz>
15037 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
15038 * cgraph.h (varpool_create_empty_node): Declare.
15039 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
15041 * symtab.c (symtab_unregister_node): Be lax about missin entries
15043 (symtab_get_node): Update comment.
15044 * varpool.c (varpool_create_empty_node): Break out from ...
15045 (varpool_node_for_decl): ... here.
15046 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
15048 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
15050 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
15051 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
15052 part. Use straight-line flow at the end.
15053 <COMPONENT_REF>: Remove superfluous else.
15054 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
15056 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15059 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
15060 target hook even for !TREE_PUBLIC decls. If no resolution info
15061 is available, return false for common and external decls.
15063 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15065 * config/rl78/constraints.md (U): New constraint.
15066 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
15069 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15072 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
15073 to allow returning address to AT_PLATFORM name.
15075 2013-06-11 Jan Hubicka <jh@suse.cz>
15077 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
15078 * cgraph.h (symtab_node_base): Add weakref flag.
15079 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
15080 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
15081 (output_weakrefs): Use weakref flag.
15082 * fold-const.c (simple_operand_p): Handle WEAK.
15083 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
15084 * ipa.c (varpool_externally_visible_p): Drop weakref.
15085 (function_and_variable_visibility): Update comment; fix weakref
15086 sanity checks; do not clear DECL_WEAK on them.
15087 * lto-cgraph.c (lto_output_node): update.
15088 (lto_output_varpool_node): Update.
15089 (input_overwrite_node): Update.
15090 (input_node): Update.
15091 (input_varpool_node): Update.
15092 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
15093 (lto_symtab_merge_symbols): Add sanity check.
15094 (lto_symtab_prevailing_decl): Do not special case weakrefs.
15095 * passes.c (rest_of_decl_compilation): Set static flag, too.
15096 * symtab.c (dump_symtab_base): Dump weakref.
15097 (verify_symtab_base): Sanity check weakrefs.
15098 (symtab_make_decl_local): Remove duplicated code.
15099 (symtab_alias_ultimate_target): Simplify.
15100 * varpool.c (varpool_create_variable_alias): Set weakref flag.
15102 2013-06-11 Tom de Vries <tom@codesourcery.com>
15104 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
15105 sequence_vect == NULL.
15107 2013-06-11 DJ Delorie <dj@redhat.com>
15109 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
15110 (rl78_unwind_word_mode): New.
15112 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15114 * final.c (debug_prefix_maps): Make static.
15116 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15118 * function.c (initial_trampoline): Remove stray copy.
15120 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
15122 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
15124 2013-06-11 Martin Jambor <mjambor@suse.cz>
15126 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
15127 within bounds at the beginning of the function.
15129 2013-06-11 Alan Modra <amodra@gmail.com>
15131 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15133 (get_named_section): Don't NULL !DECL_P decl.
15135 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15137 * doc/invoke.texi (core-avx2): Document.
15139 (atom): Updated with MOVBE.
15141 2013-06-11 Richard Biener <rguenther@suse.de>
15143 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15145 2013-06-11 Anton Blanchard <anton@samba.org>
15147 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15148 correct shift value in little-endian mode.
15150 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15153 * varasm.c (get_variable_align): Move #endif to the right place.
15155 2013-06-10 Cary Coutant <ccoutant@google.com>
15157 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15158 for hash so that hash table traversal order is deterministic.
15160 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15161 Pat Haugen <pthaugen@us.ibm.com>
15162 Peter Bergner <bergner@vnet.ibm.com>
15164 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15165 of vectors in GPRS if they are direct moves or quad word load or
15168 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15170 (direct_move_p): Likewise.
15171 (quad_load_store_p): Likewise.
15173 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15174 classes into bins based on the physical register type.
15175 (reg_class_to_reg_type): Likewise.
15176 (IS_STD_REG_TYPE): Likewise.
15177 (IS_FP_VECT_REG_TYPE): Likewise.
15178 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15179 use direct move instructions.
15180 (reload_gpr_vsx): Likewise.
15181 (reload_vsx_gpr): Likewise.
15182 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15183 information that is a simplification of register classes. Also
15184 precalculate direct move reload helpers.
15185 (direct_move_p): New function to return true if the operation can
15186 be done as a direct move instruciton.
15187 (quad_load_store_p): New function to return true if the operation
15188 is a quad memory operation.
15189 (rs6000_legitimize_address): If quad memory, only allow register
15190 indirect for TImode addresses.
15191 (rs6000_legitimate_address_p): Likewise.
15192 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15193 (rs6000_reload_register_type): Likewise.
15194 (register_to_reg_type): Return register type.
15195 (rs6000_secondary_reload_simple_move): New helper function for
15196 secondary reload and secondary memory needed to identify anything
15197 that is a simple move, and does not need reloading.
15198 (rs6000_secondary_reload_direct_move): New helper function for
15199 secondary reload to identify cases that can be done with several
15200 instructions via the direct move instructions.
15201 (rs6000_secondary_reload_move): New helper function for secondary
15202 reload to identify moves between register types that can be done.
15203 (rs6000_secondary_reload): Add support for quad memory operations
15204 and for direct move.
15205 (rs6000_secondary_memory_needed): Likewise.
15206 (rs6000_debug_secondary_memory_needed): Change argument names.
15207 (rs6000_output_move_128bit): New function to return the move to
15208 use for 128-bit moves, including knowing about the various
15209 limitations of quad memory operations.
15211 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15212 memory operations. call rs6000_output_move_128bit for the actual
15213 instruciton(s) to generate.
15214 (vsx_movti_64bit): Likewise.
15216 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15217 (UNSPEC_P8V_MTVSRWZ): Likewise.
15218 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15219 (UNSPEC_P8V_MTVSRD): Likewise.
15220 (UNSPEC_P8V_XXPERMDI): Likewise.
15221 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15222 (UNSPEC_FUSION_GPR): Likewise.
15223 (FMOVE128_GPR): New iterator for direct move.
15224 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15225 (f32_sv): Likewise.
15226 (f32_dm): Likewise.
15227 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15228 loads and direct move instructions.
15229 (zero_extendsidi2_lfiwzx): Likewise.
15230 (extendsidi2_lfiwax): Likewise.
15231 (extendsidi2_nocell): Likewise.
15232 (floatsi<mode>2_lfiwax): Likewise.
15233 (lfiwax): Likewise.
15234 (floatunssi<mode>2_lfiwzx): Likewise.
15235 (lfiwzx): Likewise.
15236 (fix_trunc<mode>_stfiwx): Likewise.
15237 (fixuns_trunc<mode>_stfiwx): Likewise.
15238 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15239 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15240 (parity<mode>2_cmpb): Set length/type attr.
15241 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15242 for 'mr.' to fast_compare.
15243 (bpermd_<mode>): Change type attr to popcnt.
15244 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15245 (p8_mtvsrwz_1): Likewise.
15246 (p8_mtvsrwz_2): Likewise.
15247 (reload_fpr_from_gpr<mode>): Likewise.
15248 (p8_mtvsrd_1): Likewise.
15249 (p8_mtvsrd_2): Likewise.
15250 (p8_xxpermdi_<mode>): Likewise.
15251 (reload_vsx_from_gpr<mode>): Likewise.
15252 (reload_vsx_from_gprsf): Likewise.
15253 (p8_mfvsrd_3_<mode>): LIkewise.
15254 (reload_gpr_from_vsx<mode>): Likewise.
15255 (reload_gpr_from_vsxsf): Likewise.
15256 (p8_mfvsrd_4_disf): Likewise.
15257 (multi-word GPR splits): Do not split direct moves or quad memory
15260 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15262 * tree-into-ssa.c (interesting_blocks): Make static.
15264 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15267 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15268 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15269 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15270 (get_variable_align): New function.
15271 (get_variable_section, emit_bss, emit_common,
15272 assemble_variable_contents, place_block_symbol): Use
15273 get_variable_align instead of DECL_ALIGN.
15274 (assemble_noswitch_variable): Add align argument, use it
15275 instead of DECL_ALIGN.
15276 (assemble_variable): Adjust caller. Use get_variable_align
15277 instead of DECL_ALIGN.
15278 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15280 (DATA_ABI_ALIGNMENT): Define.
15281 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15282 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15283 opt is false, only return the psABI mandated alignment increase.
15284 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15285 (DATA_ABI_ALIGNMENT): ... this.
15286 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15287 (DATA_ABI_ALIGNMENT): ... this.
15288 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15289 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15290 (DATA_ABI_ALIGNMENT): ... this.
15291 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15292 * doc/tm.texi: Regenerated.
15294 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15296 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15297 cmp_code to construct REG_EQUAL note.
15299 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15302 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15303 that operands[2] doesn't overlap with operands[0].
15305 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15306 Jan Hubicka <jh@suse.cz>
15308 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15309 hack to mark symbols as used.
15311 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15313 PR rtl-optimization/57559
15314 * lra-constraints.c (process_alt_operands): Don't discourage
15315 memory with known offset for offsetable memory constraint.
15316 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15318 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15320 * varasm.c (struct oc_local_state): Reorder fields.
15321 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15322 and adjust accordingly.
15323 (output_constructor): Reorder initialization code and adjust call to
15324 output_constructor_bitfield.
15326 2013-06-07 Jan Hubicka <jh@suse.cz>
15328 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15330 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15332 * tree-object-size.c (unknown): Make const.
15334 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15336 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15337 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15338 for last alternative in the cpu_facility attribute.
15340 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15343 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15344 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15345 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15346 * config/arm/constraints.md (Dg): New constraint.
15347 * config/arm/neon.md (xordi3_neon): Remove.
15348 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15349 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15351 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15353 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15354 Clean up alternatives.
15356 2013-06-07 Alan Modra <amodra@gmail.com>
15358 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15361 2013-06-07 Alan Modra <amodra@gmail.com>
15363 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15365 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15367 * config/arm/constraints.md (Df): New constraint.
15368 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15369 Correct length attribute for last two alternatives.
15371 2013-06-07 Alan Modra <amodra@gmail.com>
15373 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15374 override user -mfp-in-toc.
15375 (offsettable_ok_by_alignment): Consider just the current access
15376 rather than the whole object, unless BLKmode. Handle
15377 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15378 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15379 for -mcmodel=medium.
15380 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15381 override user -mfp-in-toc or -msum-in-toc. Default to
15382 -mno-fp-in-toc for -mcmodel=medium.
15384 2013-06-06 DJ Delorie <dj@redhat.com>
15386 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15387 TARGET_VALID_POINTER_MODE.
15389 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15390 Pat Haugen <pthaugen@us.ibm.com>
15391 Peter Bergner <bergner@vnet.ibm.com>
15393 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15394 Document new power8 builtins.
15396 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15397 condition code register, to allow 128-bit logical operations to be
15398 done in the VSX or GPR registers.
15399 (nor<mode>3): Use the canonical form for nor.
15400 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15401 vclz*, and vpopcnt* vector instructions.
15402 (nand<mode>3): Likewise.
15403 (orc<mode>3): Likewise.
15404 (clz<mode>2): LIkewise.
15405 (popcount<mode>2): Likewise.
15407 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15408 that only the GPRs are recognized.
15410 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15411 support for new power8 builtins.
15413 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15415 (xscvdpspn): Likewise.
15421 (vpopcnt): Likewise.
15422 (vpopcntb): Likewise.
15423 (vpopcnth): Likewise.
15424 (vpopcntw): Likewise.
15425 (vpopcntd): Likewise.
15427 (vmrgew): Likewise.
15428 (vmrgow): Likewise.
15430 (eqv_v16qi3): Likewise.
15431 (eqv_v8hi3): Likewise.
15432 (eqv_v4si3): Likewise.
15433 (eqv_v2di3): Likewise.
15434 (eqv_v4sf3): Likewise.
15435 (eqv_v2df3): Likewise.
15437 (nand_v16qi3): Likewise.
15438 (nand_v8hi3): Likewise.
15439 (nand_v4si3): Likewise.
15440 (nand_v2di3): Likewise.
15441 (nand_v4sf3): Likewise.
15442 (nand_v2df3): Likewise.
15444 (orc_v16qi3): Likewise.
15445 (orc_v8hi3): Likewise.
15446 (orc_v4si3): Likewise.
15447 (orc_v2di3): Likewise.
15448 (orc_v4sf3): Likewise.
15449 (orc_v2df3): Likewise.
15451 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15452 allow power8 quad mode in 64-bit.
15453 (rs6000_builtin_vectorized_function): Add support to vectorize
15454 ISA 2.07 count leading zeros, population count builtins.
15455 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15456 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15457 (builtin_function_type): Add vgbbd builtin function which takes an
15459 (altivec_expand_vec_perm_const): Add support for new power8 merge
15462 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15463 that does not include TImdoe for use with 32-bit.
15464 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15466 (UNSPEC_VSX_CVDPSPN): Likewise.
15467 (vsx_xscvdpspn): Likewise.
15468 (vsx_xscvspdpn): Likewise.
15469 (vsx_xscvdpspn_scalar): Likewise.
15470 (vsx_xscvspdpn_directmove): Likewise.
15471 (vsx_and<mode>3): Split logical operations into 32-bit and
15472 64-bit. Add support to do logical operations on TImode as well as
15473 VSX vector types. Allow logical operations to be done in either
15474 VSX registers or in general purpose registers in 64-bit mode. Add
15475 splitters if GPRs were used. For AND, add clobber of CCmode to
15476 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15478 (vsx_and<mode>3_32bit): Likewise.
15479 (vsx_and<mode>3_64bit): Likewise.
15480 (vsx_ior<mode>3): Likewise.
15481 (vsx_ior<mode>3_32bit): Likewise.
15482 (vsx_ior<mode>3_64bit): Likewise.
15483 (vsx_xor<mode>3): Likewise.
15484 (vsx_xor<mode>3_32bit): Likewise.
15485 (vsx_xor<mode>3_64bit): Likewise.
15486 (vsx_one_cmpl<mode>2): Likewise.
15487 (vsx_one_cmpl<mode>2_32bit): Likewise.
15488 (vsx_one_cmpl<mode>2_64bit): Likewise.
15489 (vsx_nor<mode>3): Likewise.
15490 (vsx_nor<mode>3_32bit): Likewise.
15491 (vsx_nor<mode>3_64bit): Likewise.
15492 (vsx_andc<mode>3): Likewise.
15493 (vsx_andc<mode>3_32bit): Likewise.
15494 (vsx_andc<mode>3_64bit): Likewise.
15495 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15496 and xxlorc instructions.
15497 (vsx_eqv<mode>3_64bit): Likewise.
15498 (vsx_nand<mode>3_32bit): Likewise.
15499 (vsx_nand<mode>3_64bit): Likewise.
15500 (vsx_orc<mode>3_32bit): Likewise.
15501 (vsx_orc<mode>3_64bit): Likewise.
15503 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15505 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15507 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15508 (p8_vmrgow): Likewise.
15509 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15510 GPRs to be split under VSX.
15511 (p8v_clz<mode>2): Add power8 count leading zero support.
15512 (p8v_popcount<mode>2): Add power8 population count support.
15513 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15516 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15519 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15521 (vec_nand): Likewise.
15522 (vec_vclz): Likewise.
15523 (vec_vclzb): Likewise.
15524 (vec_vclzd): Likewise.
15525 (vec_vclzh): Likewise.
15526 (vec_vclzw): Likewise.
15527 (vec_vgbbd): Likewise.
15528 (vec_vmrgew): Likewise.
15529 (vec_vmrgow): Likewise.
15530 (vec_vpopcnt): Likewise.
15531 (vec_vpopcntb): Likewise.
15532 (vec_vpopcntd): Likewise.
15533 (vec_vpopcnth): Likewise.
15534 (vec_vpopcntw): Likewise.
15536 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15538 PR rtl-optimization/57468
15539 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15542 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15544 PR rtl-optimization/57459
15545 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15546 type when setting live regs.
15548 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15550 * config/s390/s390.opt (mlra): New option.
15551 * config/s390/s390.c (s390_decompose_address): Check displacement
15552 for all registers for LRA.
15553 (s390_secondary_reload): Don't used secondary reloads for LRA.
15554 (s390_lra_p): New function.
15555 (TARGET_LRA_P): Define.
15556 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15557 of attribute cpu_facility to zarch for the last alternative.
15558 (*cmpmem_short): Ditto.
15560 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15562 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15563 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15564 (arm_expand_prologue): Likewise.
15566 2013-06-06 Teresa Johnson <tejohnson@google.com>
15569 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15570 as this is now done by redirect_edge_and_branch_force.
15571 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15572 barriers, and fix interaction with splitting.
15573 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15574 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15575 reflect changes made in the routine.
15576 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15577 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15578 since this is called in cfglayout mode, and replace partition fixup
15579 with assert as that is now done by force_nonfallthru_and_redirect.
15580 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15581 already be marked with region crossing note.
15582 (insert_section_boundary_note): Make non-static, gate on flag
15583 has_bb_partition, rewrite to also check for multiple partitions.
15584 (rest_of_handle_reorder_blocks): Remove call to
15585 insert_section_boundary_note, now done later during free_cfg.
15586 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15587 * bb-reorder.h (insert_section_boundary_note): Declare.
15588 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15589 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15590 invoke insert_section_boundary_note.
15591 (try_redirect_by_replacing_jump): Remove unnecessary
15592 check for region crossing note.
15593 (fixup_partition_crossing): New function.
15594 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15595 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15596 in non-cfglayout mode.
15597 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15598 remove old code that tried to do this. Emit barrier correctly
15599 when we are in cfglayout mode.
15600 (last_bb_in_partition): New function.
15601 (rtl_split_edge): Correctly fixup partition boundaries.
15602 (commit_one_edge_insertion): Remove old code that tried to
15603 fixup region crossing edge since this is now handled in
15604 split_block, and set up insertion point correctly since
15605 block may now end in a jump.
15606 (verify_hot_cold_block_grouping): Guard against checking when not in
15607 linearized RTL mode.
15608 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15610 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15611 rtl_verify_flow_info, so not called in cfglayout mode.
15612 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15613 (fixup_reorder_chain): Remove old code that attempted to fixup region
15614 crossing note as this is now handled in force_nonfallthru_and_redirect.
15615 (duplicate_insn_chain): Don't duplicate switch section notes.
15616 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15618 * basic-block.h (emit_barrier_after_bb): Declare.
15620 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15622 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15623 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15624 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15626 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15628 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15630 * config/arm/ldmstm.md: Regenerate.
15632 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15634 * config/arm/sync.md (atomic_loaddi_1):
15635 Disable predication for arm_restrict_it.
15636 (arm_load_exclusive<mode>): Likewise.
15637 (arm_load_exclusivesi): Likewise.
15638 (arm_load_exclusivedi): Likewise.
15639 (arm_load_acquire_exclusive<mode>): Likewise.
15640 (arm_load_acquire_exclusivesi): Likewise.
15641 (arm_load_acquire_exclusivedi): Likewise.
15642 (arm_store_exclusive<mode>): Likewise.
15643 (arm_store_exclusive<mode>): Likewise.
15644 (arm_store_release_exclusivedi): Likewise.
15645 (arm_store_release_exclusive<mode>): Likewise.
15647 2013-06-06 Richard Biener <rguenther@suse.de>
15649 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15651 (lto_tag_is_tree_code_p): Adjust.
15652 (lto_tag_is_gimple_code_p): Likewise.
15653 (lto_gimple_code_to_tag): Likewise.
15654 (lto_tag_to_gimple_code): Likewise.
15655 (lto_tree_code_to_tag): Likewise.
15656 (lto_tag_to_tree_code): Likewise.
15657 * data-streamer.h (streamer_write_hwi_in_range): Use
15658 uhwi streaming to stream the normalized range.
15659 (streamer_read_hwi_in_range): Likewise.
15661 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15663 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15664 (predicable_short_it): Likewise.
15665 (predicated): Likewise.
15666 (enabled): Handle above.
15667 (define_cond_exec): Set predicated attribute to yes.
15669 2013-06-05 Mike Stump <mikestump@comcast.net>
15671 * gdbinit.in (__FUNCTION__): Add.
15673 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15675 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15676 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15678 2013-06-05 Jan Hubicka <jh@suse.cz>
15680 * varasm.c (mark_decl_referenced): Revert the removal until targets
15683 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15685 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15686 instead of mark_decl_referenced.
15688 2013-06-05 Jan Hubicka <jh@suse.cz>
15690 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15691 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15692 and symtab_used_from_object_file_p.
15693 (cgraph_make_node_local_1): Clear forced_by_abi.
15694 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15695 * cgraph.h (symtab_node_base): Add forced_by_abi.
15696 (decide_is_variable_needed): Remove.
15697 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15698 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15699 (decide_is_symbol_needed): ... this one; handle symbols in general;
15700 always analyze virtuals; honnor forced_by_abi.
15701 (cgraph_finalize_function): Update.
15702 (varpool_finalize_decl): Update.
15703 (symbol_defined_and_needed): Remove.
15704 (analyze_functions): Update.
15705 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15706 output_refs, input_overwrite_node): Handle forced_by_abi.
15707 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15708 (address_taken_from_non_vtable_p): ... this one.
15709 (comdat_can_be_unshared_p_1): New function.
15710 (cgraph_comdat_can_be_unshared_p): Rename to ...
15711 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15712 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15713 (function_and_variable_visibility): Clear forced_by_abi as needed.
15714 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15715 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15716 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15717 * varpool.c (decide_is_variable_needed): Remove.
15719 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15721 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15722 (arm_option_override): Override arm_restrict_it where appropriate.
15723 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15724 * config/arm/arm.opt (mrestrict-it): New command-line option.
15725 * doc/invoke.texi: Document -mrestrict-it.
15727 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15729 * tsan.c (tsan_atomic_table): Make const.
15731 2013-06-05 Richard Biener <rguenther@suse.de>
15733 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15734 index associated with the tree we are supposed to replace.
15735 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15736 TREE_ASM_WRITTEN as zero for everything but SSA names.
15738 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15740 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15742 2013-06-04 Jan Hubicka <jh@suse.cz>
15744 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15745 (input_node, input_varpool_node): Handle correctly external same
15747 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15748 nodes at ltrans stage.
15750 2013-06-04 Jan Hubicka <jh@suse.cz>
15752 * ipa-inline.c (update_caller_keys): Fix availability test.
15753 (update_callee_keys): Likewise.
15754 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15755 to follow ELF standard.
15757 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15759 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15760 (mips64r5900el-*-elf*): New configurations.
15761 * config/mips/mips-cpus.def (r5900): New processor.
15762 * config/mips/mips-tables.opt: Regenerate.
15763 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15764 (mips_issue_rate): Handle PROCESSOR_R5900.
15765 (mips_reorg_process_insns): Force reorder mode for the R5900.
15766 * config/mips/mips.h (TARGET_MIPS5900): Define.
15767 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15769 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15771 * config/mips/mips.md (processor): Add r5900.
15772 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15774 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15776 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15777 into function to generate MOVI instruction.
15778 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15779 (aarch64_preferred_simd_mode): Turn into wrapper.
15780 (aarch64_output_scalar_simd_mov_immediate): New function.
15781 * config/aarch64/aarch64-protos.h: Add prototype for above.
15783 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15785 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15786 element_char member.
15787 (sizetochar): Return signed char.
15788 (aarch64_simd_valid_immediate): Remove elchar and other
15789 unnecessary variables.
15790 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15791 Calculate element_char as required.
15792 * config/aarch64/aarch64-protos.h: Update and move prototype
15793 for aarch64_output_simd_mov_immediate.
15794 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15797 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15799 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15800 information completed by aarch64_simd_valid_immediate.
15801 (aarch64_legitimate_constant_p): Update arguments.
15802 (aarch64_simd_valid_immediate): Work with struct rather than many
15804 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15805 (aarch64_simd_make_constant): Update arguments.
15806 (aarch64_output_simd_mov_immediate): Work with struct rather than
15807 many pointers. Output immediate directly rather than as operand.
15808 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15810 * config/aarch64/constraints.md (Dn): Update arguments.
15812 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15814 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15816 (aarch64_simd_immediate_valid_for_move): Remove.
15817 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15818 (aarch64_simd_make_constant): Update call.
15819 (aarch64_output_simd_mov_immediate): Update call.
15820 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15822 * config/aarch64/constraints.md (Dn): Update call.
15824 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15826 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15827 return type to bool for prototype.
15828 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15829 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15830 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15832 2013-06-04 Catherine Moore <clm@codesourcery.com>
15834 * config/mips/mips.opt (meva): New.
15835 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15836 (ASM_SPEC): Handle -meva.
15837 * doc/invoke.texi (meva): Document.
15839 2013-06-04 Alan Modra <amodra@gmail.com>
15841 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15844 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15846 * rtl.def: Add extra fourth optional field to define_cond_exec.
15847 * gensupport.c (process_one_cond_exec): Process attributes from
15849 * doc/md.texi: Document fourth field in define_cond_exec.
15851 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15853 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15854 out the processing order as in store_bit_field_1.
15856 2013-06-04 Jan Hubicka <jh@suse.cz>
15858 PR middle-end/57500
15859 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15860 non-VAR_DECL node if it does not exist yet.
15862 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15864 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15865 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15866 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15867 target_cpu_default setting.
15869 2013-06-03 Teresa Johnson <tejohnson@google.com>
15871 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15872 default to -fopt-info=optimized instead of all.
15873 * doc/invoke.texi: Ditto.
15874 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15875 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15876 (execute_vect_slp): Emit BB vectorization success under
15877 MSG_OPTIMIZED_LOCATIONS.
15878 * tree-vect-slp.c (vect_slp_transform_bb): Change
15879 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15880 * tree-vect-loop.c (vect_transform_loop): Ditto.
15882 2013-06-03 Jason Merrill <jason@redhat.com>
15885 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15886 Use TARGET_EXPR for C++.
15888 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15890 PR rtl-optimization/57268
15891 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15892 if DEBUG_INSN_P (insn).
15895 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15897 PR rtl-optimization/57268
15898 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15899 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15901 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15903 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15904 (ix86_avoid_lea_for_addr): Likewise.
15905 (exact_dependency_1): Likewise.
15906 (ix86_adjust_cost): Likewise.
15907 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15909 (do_reorder_for_imul): Fix typo, formatting and
15910 !reload_completed check removed.
15911 (ix86_sched_reorder): Fix typo and formatting.
15912 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15915 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15917 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15919 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15921 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15922 <STRING_CST>: Likewise.
15923 <VECTOR_CST>: Likewise.
15925 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15926 Mikael Morin <mikael@gcc.gnu.org>
15928 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15929 * config.in: Regenerated.
15930 * configure: Regenerated.
15932 2013-06-01 Jan Hubicka <jh@suse.cz>
15934 PR middle-end/57366
15935 * cgraphunit.c (compile): When weakref is not supported,
15936 set up transparent aliases before final output pass.
15937 * varasm.c (assemble_alias): Do not try to do it here.
15939 2013-06-01 Jan Hubicka <jh@suse.cz>
15941 PR middle-end/57467
15942 * passes.c (for_per_function): Skip unanalyzed functions.
15944 2013-06-01 Jan Hubicka <jh@suse.cz>
15946 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15947 (lto_symtab_merge_symbols_1): ... this one.
15948 (lto_symtab_merge_cgraph_nodes): Rename to ...
15949 (lto_symtab_merge_symbols): ... this one; simplify.
15950 * cgraph.c (same_body_aliases_done): Rename to ...
15951 (cpp_implicit_aliases_done): ... this one.
15952 (cgraph_create_function_alias): Update.
15953 (cgraph_same_body_alias): Update.
15954 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15955 (verify_edge_corresponds_to_fndecl): Simplify.
15956 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
15957 (cgraph_node): Remove same_body_alias.
15958 (varpool_node): Remove alias_of and extra_name_alias.
15959 (same_body_aliases_done): Rename to ..
15960 (cpp_implicit_aliases_done): ... this one.
15961 (symtab_alias_ultimate_target): Add default parameter.
15962 (symtab_resolve_alias): New function.
15963 (fixup_same_cpp_alias_visibility): Declare.
15964 (cgraph_function_node): Add default parameter.
15965 (cgraph_node_asm_name): Likewise.
15966 (cgraph_function_or_thunk_node): Add default parameter; do
15967 not ICE when it is NULL.
15968 (varpool_variable_node): Likewise.
15969 * tree-emutls.c (create_emultls_var): Update.
15970 (ipa_lower_emutls): Update.
15971 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
15972 (cgraph_reset_node): Reset alias info.
15973 (cgraph_finalize_function): Update.
15974 (fixup_same_cpp_alias_visibility): Move to symtab.c.
15975 (analyze_function): Simplify.
15976 (cgraph_process_same_body_aliases): Simplify.
15977 (analyze_functions): Fixup same body aliases.
15978 (handle_alias_pairs): Simplify.
15979 (assemble_thunk): Update.
15980 (assemble_thunks_and_aliases): Update.
15981 (output_weakrefs): Rewrite.
15982 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
15983 (lto_output_varpool_node): Likewise.
15984 (compute_ltrans_boundary): Remve assert.
15985 (get_alias_symbol): New functoin.
15986 (input_node): Rewrite alias handling.
15987 (input_varpool_node): Likewise.
15988 * ipa-pure-const.c (propagate_pure_const): Fix formating.
15989 * ipa.c (process_references): Handle weakrefs correctly.
15990 (symtab_remove_unreachable_nodes): Likewise.
15991 * trans-mem.c (get_cg_data): Update.
15992 (ipa_tm_create_version_alias): Update.
15993 (ipa_tm_execute): Update.
15994 * symtab.c (dump_symtab_base): Dump aliases.
15995 (verify_symtab_base): Verify aliases.
15996 (symtab_node_availability): New function.
15997 (symtab_alias_ultimate_target): Simplify.
15998 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
15999 handle all the fixup cases.
16000 (symtab_resolve_alias): New function.
16001 * passes.c (ipa_write_summaries): Handle weakrefs.
16002 * varpool.c (varpool_analyze_node): Simplify.
16003 (assemble_aliases): Update.
16004 (varpool_create_variable_alias): Simplify.
16005 (varpool_extra_name_alias): Simplify.
16006 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
16007 (lto_symtab_merge_symbols): ... this one.
16009 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
16012 PR rtl-optimization/57268
16013 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16014 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16016 2013-06-01 Tobias Burnus <burnus@net-b.de>
16018 Partially reverted:
16019 2013-05-31 Tobias Burnus <burnus@net-b.de>
16021 PR middle-end/57073
16022 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
16025 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16027 PR rtl-optimization/57268
16028 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16029 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16031 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16033 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
16034 unordered comparison operators when -fno-trapping-math is in effect
16036 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
16037 and implement unordered comparison operators properly on the e500.
16039 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16041 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
16042 for constant scalar integers.
16043 (simplify_relational_operation_1): Likewise.
16045 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
16047 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
16048 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
16051 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16052 Igor Zamyatin <igor.zamyatin@intel.com>
16054 Silvermont (SLM) architecture performance tuning.
16055 * config/i386/i386.h (enum ix86_tune_indices): Add
16056 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
16057 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
16059 * config/i386/i386.c (initial_ix86_tune_features)
16060 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
16061 (ix86_lea_outperforms): Handle Silvermont tuning.
16062 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
16064 (ix86_use_lea_for_mov): Likewise.
16065 (ix86_avoid_lea_for_addr): Likewise.
16066 (ix86_lea_for_add_ok): Likewise.
16067 (exact_dependency_1): New function.
16068 (exact_store_load_dependency): Likewise.
16069 (ix86_adjust_cost): Handle Silvermont tuning.
16070 (do_reoder_for_imul): Likewise.
16071 (swap_top_of_ready_list): New function.
16072 (ix86_sched_reorder): Changed to handle Silvermont tuning.
16074 * config/i386/i386.md (peepholes that split memory operand in fp
16077 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16079 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
16080 Remove un-necessary braces.
16082 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16084 * config/aarch64/aarch64.c (aarch64_classify_symbol):
16085 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
16087 2013-05-31 Tobias Burnus <burnus@net-b.de>
16089 PR middle-end/57073
16090 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
16092 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16095 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
16096 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
16097 * config/arm/neon.md (iordi3_neon): Remove.
16098 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
16099 * config/arm/predicates.md (imm_for_neon_logic_operand):
16100 Move to earlier in the file.
16101 (neon_logic_op2): Likewise.
16102 (arm_iordi_operand_neon): New predicate.
16104 2013-05-31 Richard Biener <rguenther@suse.de>
16106 PR tree-optimization/57478
16107 PR tree-optimization/57453
16108 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
16111 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16113 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
16114 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
16116 2013-05-30 Tobias Burnus <burnus@net-b.de>
16117 Thomas Koenig <tkoenig@gcc.gnu.org>
16119 PR middle-end/57073
16120 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
16121 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
16123 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
16125 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
16127 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
16129 * target.def (register_usage_leveling_p): New hook.
16130 * targhooks.c (default_register_usage_leveling_p): New.
16131 * targhooks.h (default_register_usage_leveling_p): New prototype.
16132 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16133 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16134 * doc/tm.texi: Update.
16135 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16137 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16139 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16140 (*insv_reg<mode>): New define_insn.
16142 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16144 PR rtl-optimization/57439
16145 * postreload.c (move2add_valid_value_p): Check that we have
16146 a zero subreg_regno_offset when accessing the register in
16147 the requested mode.
16149 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16150 Igor Zamyatin <igor.zamyatin@intel.com>
16152 Silvermont (SLM) architecture pipeline model, tuning and
16154 * config.gcc: Add slm config options and target.
16156 * config/i386/slm.md: New.
16158 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16160 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16162 (ix86_target_macros_internal): Likewise.
16164 * config/i386/i386.c (slm_cost): New cost.
16165 (m_SLM): New macro flag.
16166 (initial_ix86_tune_features): Set m_SLM.
16167 (x86_accumulate_outgoing_args): Likewise.
16168 (x86_arch_always_fancy_math_387): Likewise.
16169 (processor_target_table): Add slm cost.
16170 (cpu_names): Add slm cpu name.
16171 (x86_option_override_internal): Set SLM ISA.
16172 (ix86_issue_rate): New case PROCESSOR_SLM.
16173 (ia32_multipass_dfa_lookahead): Likewise.
16174 (fold_builtin_cpu): Add slm.
16176 * config/i386/i386.h (TARGET_SLM): New target macro.
16177 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16178 (processor_type): Add PROCESSOR_SLM.
16180 * config/i386/i386.md (cpu): Add new value "slm".
16181 (slm.md): Include slm.md.
16183 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16184 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16186 * config/arm/arm-protos.h: Add and update function protos.
16187 * config/arm/arm.c (use_simple_return_p): New added.
16188 (thumb2_expand_return): Check simple_return flag.
16189 * config/arm/arm.md: Add simple_return and conditional simple_return.
16190 * config/arm/iterators.md: Add iterator for return and simple_return.
16192 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16194 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16195 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16196 (arm_emit_vfp_multi_reg_pop): Likewise.
16197 (thumb2_emit_ldrd_pop): Likewise.
16198 (arm_expand_epilogue): Add misc REG_CFA notes.
16199 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16201 2013-05-29 Lawrence Crowl <crowl@google.com>
16203 * config/arm/t-arm: Update for below.
16205 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16206 Change type to hash_table. Update dependent calls and types.
16208 * config/i386/t-cygming: Update for below.
16210 * config/i386/t-interix: Update for below.
16212 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16213 Change type to hash_table. Update dependent calls and types.
16214 (i386_find_on_wrapper_list::wrappers): Likewise.
16216 * config/ia64/t-ia64: Update for below.
16218 * config/ia64/ia64.c (bundle_state_table):
16219 Change type to hash_table. Update dependent calls and types.
16221 * config/mips/mips.c (mips_reorg_process_insns::htab):
16222 Change type to hash_table. Update dependent calls and types.
16224 * config/sol2.c (solaris_comdat_htab):
16225 Change type to hash_table. Update dependent calls and types.
16227 * config/t-sol2: Update for above.
16229 2013-05-29 Teresa Johnson <tejohnson@google.com>
16231 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16232 functions are not yet marked as defined.
16234 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16235 Pat Haugen <pthaugen@us.ibm.com>
16236 Peter Bergner <bergner@vnet.ibm.com>
16238 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16242 (vrotl<mode>3): Likewise.
16243 (vashl<mode>3): Likewise.
16244 (vlshr<mode>3): Likewise.
16245 (vashr<mode>3): Likewise.
16247 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16248 support for power8 V2DI builtins.
16250 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16251 power8 V2DI builtins.
16252 (vupkhsw): Likewise.
16253 (vupklsw): Likewise.
16254 (vaddudm): Likewise.
16255 (vminsd): Likewise.
16256 (vmaxsd): Likewise.
16257 (vminud): Likewise.
16258 (vmaxud): Likewise.
16259 (vpkudum): Likewise.
16260 (vpksdss): Likewise.
16261 (vpkudus): Likewise.
16262 (vpksdus): Likewise.
16267 (vsubudm): Likewise.
16268 (vcmpequd): Likewise.
16269 (vcmpgtsd): Likewise.
16270 (vcmpgtud): Likewise.
16271 (vcmpequd_p): Likewise.
16272 (vcmpgtsd_p): Likewise.
16273 (vcmpgtud_p): Likewise.
16274 (vupkhsw): Likewise.
16275 (vupklsw): Likewise.
16276 (vaddudm): Likewise.
16277 (vmaxsd): Likewise.
16278 (vmaxud): Likewise.
16279 (vminsd): Likewise.
16280 (vminud): Likewise.
16281 (vpksdss): Likewise.
16282 (vpksdus): Likewise.
16283 (vpkudum): Likewise.
16284 (vpkudus): Likewise.
16289 (vsubudm): Likewise.
16291 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16292 support for power8 V2DI instructions.
16294 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16295 power8 V2DI instructions. Combine pack and unpack insns to use an
16296 iterator for each mode. Check whether a particular mode supports
16297 Altivec instructions instead of just checking TARGET_ALTIVEC.
16298 (UNSPEC_VPKUWUM): Likewise.
16299 (UNSPEC_VPKSHSS): Likewise.
16300 (UNSPEC_VPKSWSS): Likewise.
16301 (UNSPEC_VPKUHUS): Likewise.
16302 (UNSPEC_VPKSHUS): Likewise.
16303 (UNSPEC_VPKUWUS): Likewise.
16304 (UNSPEC_VPKSWUS): Likewise.
16305 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16306 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16307 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16308 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16309 (UNSPEC_VUPKHSB): Likewise.
16310 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16311 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16312 (UNSPEC_VUPKHSH): Likewise.
16313 (UNSPEC_VUPKLSB): Likewise.
16314 (UNSPEC_VUPKLSH): Likewise.
16316 (VI_char): Likewise.
16317 (VI_scalar): Likewise.
16318 (VI_unit): Likewise.
16320 (VP_small): Likewise.
16321 (VP_small_lc): Likewise.
16322 (VU_char): Likewise.
16323 (add<mode>3): Likewise.
16324 (altivec_vaddcuw): Likewise.
16325 (altivec_vaddu<VI_char>s): Likewise.
16326 (altivec_vadds<VI_char>s): Likewise.
16327 (sub<mode>3): Likewise.
16328 (altivec_vsubcuw): Likewise.
16329 (altivec_vsubu<VI_char>s): Likewise.
16330 (altivec_vsubs<VI_char>s): Likewise.
16331 (altivec_vavgs<VI_char>): Likewise.
16332 (altivec_vcmpbfp): Likewise.
16333 (altivec_eq<mode>): Likewise.
16334 (altivec_gt<mode>): Likewise.
16335 (altivec_gtu<mode>): Likewise.
16336 (umax<mode>3): Likewise.
16337 (smax<mode>3): Likewise.
16338 (umin<mode>3): Likewise.
16339 (smin<mode>3): Likewise.
16340 (altivec_vpkuhum): Likewise.
16341 (altivec_vpkuwum): Likewise.
16342 (altivec_vpkshss): Likewise.
16343 (altivec_vpkswss): Likewise.
16344 (altivec_vpkuhus): Likewise.
16345 (altivec_vpkshus): Likewise.
16346 (altivec_vpkuwus): Likewise.
16347 (altivec_vpkswus): Likewise.
16348 (altivec_vpks<VI_char>ss): Likewise.
16349 (altivec_vpks<VI_char>us): Likewise.
16350 (altivec_vpku<VI_char>us): Likewise.
16351 (altivec_vpku<VI_char>um): Likewise.
16352 (altivec_vrl<VI_char>): Likewise.
16353 (altivec_vsl<VI_char>): Likewise.
16354 (altivec_vsr<VI_char>): Likewise.
16355 (altivec_vsra<VI_char>): Likewise.
16356 (altivec_vsldoi_<mode>): Likewise.
16357 (altivec_vupkhsb): Likewise.
16358 (altivec_vupkhs<VU_char>): Likewise.
16359 (altivec_vupkls<VU_char>): Likewise.
16360 (altivec_vupkhsh): Likewise.
16361 (altivec_vupklsb): Likewise.
16362 (altivec_vupklsh): Likewise.
16363 (altivec_vcmpequ<VI_char>_p): Likewise.
16364 (altivec_vcmpgts<VI_char>_p): Likewise.
16365 (altivec_vcmpgtu<VI_char>_p): Likewise.
16366 (abs<mode>2): Likewise.
16367 (vec_unpacks_hi_v16qi): Likewise.
16368 (vec_unpacks_hi_v8hi): Likewise.
16369 (vec_unpacks_lo_v16qi): Likewise.
16370 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16371 (vec_unpacks_lo_v8hi): Likewise.
16372 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16373 (vec_pack_trunc_v8h): Likewise.
16374 (vec_pack_trunc_v4si): Likewise.
16375 (vec_pack_trunc_<mode>): Likewise.
16377 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16379 (vec_vmaxsd): Likewise.
16380 (vec_vmaxud): Likewise.
16381 (vec_vminsd): Likewise.
16382 (vec_vminud): Likewise.
16383 (vec_vpksdss): Likewise.
16384 (vec_vpksdus): Likewise.
16385 (vec_vpkudum): Likewise.
16386 (vec_vpkudus): Likewise.
16387 (vec_vrld): Likewise.
16388 (vec_vsld): Likewise.
16389 (vec_vsrad): Likewise.
16390 (vec_vsrd): Likewise.
16391 (vec_vsubudm): Likewise.
16392 (vec_vupkhsw): Likewise.
16393 (vec_vupklsw): Likewise.
16395 2013-05-29 Jan Hubicka <jh@suse.cz>
16397 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16398 flags; reorder rest of fields in more consistent way.
16399 (varpool_node): Remove analyzed, finalized and alias.
16400 (cgraph_ndoe): Likewise.
16401 (symtab_alias_ultimate_target): New function.
16402 (cgraph_function_node): Move offline.
16403 (cgraph_reset_node): Declare.
16404 (cgraph_comdat_can_be_unshared_p): Remove.
16405 (varpool_remove_initializer): Declare.
16406 (varpool_first_defined_variable, varpool_next_defined_variable
16407 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16408 (cgraph_function_with_gimple_body_p): Update.
16409 (varpool_all_refs_explicit_p): Update.
16410 (symtab_alias_target): New function.
16411 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16412 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16413 (cgraph_function_or_thunk_node): Simplify using
16414 symtab_alias_ultimate_target.
16415 (varpool_variable_node): Likewise.
16416 * cgraph.c (cgraph_create_function_alias): Update.
16417 (cgraph_add_thunk): Update.
16418 (cgraph_remove_node): Update.
16419 (dump_cgraph_node): Do not dump removed flags.
16420 (cgraph_function_body_availability): Update.
16421 (cgraph_propagate_frequency): Update.
16422 (verify_cgraph_node): Check sanity of local flag.
16423 (cgraph_function_node): Move here from cgraph.h; revamp for
16424 cgraph_function_or_thunk_node.
16425 * lto-symtab.c (lto_varpool_replace_node): Update.
16426 (lto_symtab_resolve_can_prevail_p): Update.
16427 (lto_symtab_merge_cgraph_nodes): Update.
16428 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16429 propagate_constants_accross_call, devirtualization_time_bonus,
16430 ipcp_propagate_stage): Update.
16431 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16432 * ipa-inline-transform.c (clone_inlined_nodes,
16433 preserve_function_body_p): Update.
16434 * ipa-reference.c (propagate): Update.
16435 (write_node_summary_p): Update.
16436 * toplev.c (wrapup_global_declaration_2): Update.
16437 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16438 (analyze_function) ... this one.
16439 (cgraph_process_new_functions): Update.
16440 (cgraph_reset_node): Export.
16441 (cgraph_finalize_function): Update.
16442 (cgraph_add_new_function): Update.
16443 (process_function_and_variable_attributes): Update.
16444 (varpool_finalize_decl): Update.
16445 (symbol_finalized): Remove.
16446 (symbol_finalized_and_needed): Rename to ...
16447 (symbol_defined_and_needed): ... update.
16448 (cgraph_analyze_functions): Update.
16449 (handle_alias_pairs): Update.
16450 (mark_functions_to_output): Update.
16451 (assemble_thunk): Update.
16452 (output_in_order): Update.
16453 (output_weakrefs): Update.
16454 (finalize_compilation_unit): Update.
16455 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16456 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16457 input_node, input_varpool_node): Update.
16458 * dbxout.c (dbxout_expand_expr): Update.
16459 * cgraphclones.c (cgraph_clone_node): Update.
16460 (cgraph_copy_node_for_versioning): Update.
16461 (cgraph_materialize_clone): Update.
16462 (cgraph_materialize_all_clones): Update.
16463 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16464 propagate_pure_const, propagate_nothrow): Update.
16465 * lto-streamer-out.c (lto_output, write_symbol): Update.
16466 * ipa-utils.c (ipa_reverse_postorder): Update.
16467 * ipa-inline.c (can_inline_edge_p): Update.
16468 (update_caller_keys, ipa_inline): Update.
16469 * dwarf2out.c (reference_to_unused,
16470 premark_types_used_by_global_vars_helper): Update.
16471 * tree-eh.c (tree_could_trap_p): Update.
16472 * ipa-split.c (consider_split, execute_split_functions): Update.
16473 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16474 has_addr_references_p): Update; move ahead in file for better
16476 (process_references): Simplify.
16477 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16478 bodies are removed.
16479 (cgraph_comdat_can_be_unshared_p): Make static.
16480 (cgraph_externally_visible_p): Update.
16481 (varpool_externally_visible_p): Update.
16482 (function_and_variable_visibility): Update.
16483 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16484 ipa_tm_mark_force_output_node): Update.
16485 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16486 estimate_edge_devirt_benefit, inline_generate_summary,
16487 inline_write_summary): Update.
16488 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16489 * ipa-prop.c (ipa_compute_jump_functions): Update.
16490 (ipa_print_node_params, ipa_prop_read_section,
16491 ipa_update_after_lto_read, read_replacements_section): Update.
16492 * varasm.c (mark_decl_referenced): Update.
16493 (assemble_alias, dump_tm_clone_pairs): Update.
16494 * tree-inline.c (copy_bb): Update.
16495 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16497 * symtab.c (dump_symtab_base): Print new flags.
16498 (verify_symtab_base): Verify new flags.
16499 (symtab_alias_ultimate_target): New function.
16500 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16501 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16503 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16505 * i386.c (ix86_get_function_versions_dispatcher,
16506 ix86_generate_version_dispatcher_body): Update.
16507 (fold_builtin_cpu): Use varpool_add_new_variable.
16508 * varpool.c (varpool_remove_initializer): Break out from ...
16509 (varpool_remove_node): ... this one.
16510 (dump_varpool_node, varpool_node_for_asm,
16511 cgraph_variable_initializer_availability, varpool_analyze_node,
16512 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16513 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16516 2013-05-29 Jan Hubicka <jh@suse.cz>
16518 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16520 2013-05-29 Easwaran Raman <eraman@google.com>
16522 PR tree-optimization/57442
16523 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16524 when control exits the main loop.
16526 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16528 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16530 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16531 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16532 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16535 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16537 PR tree-optimization/57441
16538 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16539 Don't limit size of incr_vec to number of candidates.
16541 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16543 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16544 and mips16 directories.
16545 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16546 (MULTILIB_DIRNAMES): Ditto.
16547 (MULTILIB_EXCEPTIONS): Add new exceptions.
16548 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16549 (MULTILIB_DIRNAMES): Ditto.
16550 (MULTILIB_EXCEPTIONS): Add new exceptions.
16552 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16553 Marcus Shawcroft <marcus.shawcroft@arm.com>
16555 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16556 SYMBOL_TINY_ABSOLUTE.
16557 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16558 SYMBOL_TINY_ABSOLUTE.
16559 (aarch64_expand_mov_immediate): Likewise.
16560 (aarch64_classify_symbol): Likewise.
16561 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16562 Permit SYMBOL_TINY_ABSOLUTE.
16563 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16565 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16566 Marcus Shawcroft <marcus.shawcroft@arm.com>
16568 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16569 Refactor if/switch. Replace gcc_assert with if.
16571 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16573 * config/i386/i386.c (initial_ix86_tune_features): Enable
16574 FP Reassociation for AMD bdver1 and bdver2.
16576 2013-05-29 Martin Jambor <mjambor@suse.cz>
16578 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16579 and IMAGPART_EXPR do not occur within other handled_components.
16581 2013-05-29 Richard Biener <rguenther@suse.de>
16583 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16584 access on whether the use is in the BB we currently try to
16586 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16587 vectorize to vect_bb_slp_scalar_cost.
16589 2013-05-29 Richard Biener <rguenther@suse.de>
16591 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16592 computing scalar cost offsetted by stmts that are kept live
16594 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16595 for computation of scalar cost.
16597 2013-05-28 Steve Ellcey <sellcey@mips.com>
16599 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16601 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16603 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16604 array notation built-in reduction functions.
16605 * doc/passes.texi (Passes): Added documentation about changes done
16607 * doc/invoke.texi (C Dialect Options): Added documentation about
16608 the -fcilkplus flag.
16609 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16610 (BUILTINS_DEF): Depend on cilkplus.def.
16611 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16612 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16613 * cilkplus.def: New file.
16615 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16617 PR rtl-optimization/57439
16618 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16620 2013-05-28 Easwaran Raman <eraman@google.com>
16622 PR tree-optimization/57337
16623 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16624 (find_insert_point): Correctly identify the insertion point
16625 when two statements with the same UID is compared.
16627 2013-05-28 Richard Biener <rguenther@suse.de>
16629 PR tree-optimization/56787
16630 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16631 from the list of data references.
16632 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16634 (vect_analyze_loop_operations): Likewise.
16635 (vect_transform_loop): Remove clobbers.
16637 2013-05-28 Martin Jambor <mjambor@suse.cz>
16639 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16640 and REALPART_EXPRs have scalar type.
16642 2013-05-28 Richard Biener <rguenther@suse.de>
16644 PR tree-optimization/57411
16645 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16647 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16648 virtual operand propagation.
16650 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16652 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16653 destination register for bmasksi_vis.
16654 (vector_init_bshuffle): Likewise.
16655 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16657 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16659 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16660 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16661 mode if the instruction isn't available in the original mode.
16662 * config/sparc/sparc.opt (mfix-ut699): New option.
16663 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16664 (divdf3): Turn into expander.
16665 (divdf3_nofix): New insn.
16666 (divdf3_fix): Likewise.
16667 (divsf3): Disable if -mfix-ut699.
16668 (sqrtdf2): Turn into expander.
16669 (sqrtdf2_nofix): New insn.
16670 (sqrtdf2_fix): Likewise.
16671 (sqrtsf2): Disable if -mfix-ut699.
16673 2013-05-27 Richard Biener <rguenther@suse.de>
16675 PR middle-end/57412
16676 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16677 block for the new loop.
16679 2013-05-27 Richard Biener <rguenther@suse.de>
16681 PR tree-optimization/57343
16682 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16683 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16684 (number_of_iterations_cond): Do not build the folded tree.
16686 2013-05-27 Richard Biener <rguenther@suse.de>
16689 PR middle-end/57381
16690 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16691 OEP_CONSTANT_ADDRESS_OF retained.
16693 PR tree-optimization/57417
16694 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16695 for unchanged base.
16696 (set_ssa_val_to): Compare addresses using
16697 get_addr_base_and_unit_offset.
16699 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16701 PR rtl-optimization/56833
16702 * postreload.c (move2add_record_mode): New function.
16703 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16704 (move2add_use_add2_insn): Use move2add_record_sym_value.
16705 (move2add_use_add3_insn): Likewise.
16706 (reload_cse_move2add): Use move2add_valid_value_p and
16707 move2add_record_mode. Invalidate call-clobbered and REG_INC
16708 affected regs by setting reg_mode to VOIDmode.
16709 (move2add_note_store): Don't pretend the inside of a SUBREG is
16710 the actual destination. Invalidate single/leading registers by
16711 setting reg_mode to VOIDmode.
16712 Use move2add_record_sym_value, move2add_valid_value_p and
16713 move2add_record_mode.
16715 2013-05-27 Richard Biener <rguenther@suse.de>
16717 PR tree-optimization/57396
16718 * tree-affine.c (double_int_constant_multiple_p): Properly
16719 return false for val == 0 and div != 0.
16721 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16723 * config/mips/mips.h: Use #elif in preprocessor conditions.
16725 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16728 * config/mips/constraints.md (kl): New constraint.
16729 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16730 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16731 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16732 from LO for MIPS16.
16733 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16735 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16738 * config/mips/mips.c (mips_can_inline_p): New function.
16739 (TARGET_CAN_INLINE_P): Define.
16741 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16743 * sched-int.h (ds_t, dw_t): Make unsigned int.
16744 Fix documentation that describes how all the ds_t bits are used.
16745 Reserve the last bit for delayed-branch scheduling.
16746 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16747 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16748 (gen_dep_weak_1): Remove prototype.
16749 * sched-deps.c (get_dep_weak_1): Make static.
16750 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16751 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16752 * doc/tm.texi: Regenerate.
16753 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16755 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16758 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16760 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16761 Sandra Loosemore <sandra@codesourcery.com>
16763 * config.gcc (powerpc-*): Allow native for with-cpu.
16765 2013-05-24 Jeff Law <law@redhat.com>
16767 PR tree-optimization/57124
16768 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16769 conversion feeding a condition if the range has an overflow
16770 if -fstrict-overflow. Add warnings for when we do make the
16773 2013-05-24 Dehao Chen <dehao@google.com>
16775 * tree-cfg.c (locus_discrim_map): Fix the typo.
16776 (locus_discrim_hasher): Likewise.
16777 (locus_discrim_hasher::hash): Likewise.
16778 (locus_discrim_hasher::equal): Likewise.
16780 2013-05-24 Martin Jambor <mjambor@suse.cz>
16782 PR tree-optimization/57294
16783 * cgraph.h (ipa_record_stmt_references): Declare.
16784 * cgraphbuild.c (ipa_record_stmt_references): New function.
16785 (build_cgraph_edges): Use ipa_record_stmt_references.
16786 (rebuild_cgraph_edges): Likewise.
16787 (cgraph_rebuild_references): Likewise.
16788 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16789 associated with the old statement and build references from the
16790 newly built statements.
16791 * ipa-ref.c (ipa_remove_stmt_references): New function.
16792 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16794 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16796 * lra-constraints.c (emit_spill_move): Use smaller mode for
16798 (check_and_process_move): Consider mem-reg moves for secondary
16800 (curr_insn_transform): Don't lose insns emitted before for
16801 secondary memory moves.
16802 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16803 reg set up in the current insn.
16805 2013-05-24 Dehao Chen <dehao@google.com>
16807 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16809 (locus_descrim_hasher::equal): Likewise.
16810 (build_gimple_cfg): New discriminator assignment algorithm.
16811 (make_edges): Likewise.
16812 (next_discriminator_for_locus): Likewise.
16813 (same_line_p): Likewise.
16814 (assign_discriminators): Likewise.
16815 (make_cond_expr_edges): Likewise.
16816 (make_gimple_switch_edges): Likewise.
16817 (make_goto_expr_edges): Likewise.
16818 (make_gimple_asm_edges): Likewise.
16820 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16822 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16823 X format specifier to only display bottom 16 bits.
16824 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16825 immediate to match for operand 2, since it will be masked.
16827 2013-05-24 Richard Biener <rguenther@suse.de>
16829 PR tree-optimization/57287
16830 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16831 all SSA names that occur in abnormal PHIs.
16833 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16836 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16837 that index is not negative.
16839 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16841 PR rtl-optimization/55177
16842 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16843 (simplify_byte_swapping_operation): New.
16844 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16845 (simplify_relational_operation_1): Deal with BSWAP.
16847 2013-05-23 Richard Henderson <rth@redhat.com>
16850 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16851 (ix86_reorg): Call it.
16853 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16856 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16857 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16858 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16860 2013-05-23 Christian Bruel <christian.bruel@st.com>
16863 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16865 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16866 Marcus Shawcroft <marcus.shawcroft@arm.com>
16868 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16869 * config/aarch64/constraints.md (Usa): Remove.
16870 * doc/md.texi (AArch64 Usa): Remove.
16872 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16873 Marcus Shawcroft <marcus.shawcroft@arm.com>
16875 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16876 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16877 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16878 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16880 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16882 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16883 instruction (AdvSIMD).
16884 * config/aarch64/aarch64-builtins.c
16885 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16886 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16888 2013-05-23 Martin Jambor <mjambor@suse.cz>
16890 PR middle-end/57347
16891 * tree.h (contains_bitfld_component_ref_p): Declare.
16892 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16893 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16895 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16896 not access a bit-field. Assert all final offsets are byte-aligned.
16898 2013-05-23 Richard Biener <rguenther@suse.de>
16900 PR tree-optimization/57380
16901 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16902 least one invariant or re-used load.
16903 * passes.c (init_optimization_passes): Move pass_phiprop before
16906 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16908 * config/aarch64/aarch64-simd.md
16909 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16911 2013-05-23 Richard Biener <rguenther@suse.de>
16913 PR middle-end/57381
16914 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16915 OEP_CONSTANT_ADDRESS_OF retained.
16917 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16919 PR middle-end/57344
16920 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16921 don't lower unit. Handle unit not being always BITS_PER_WORD.
16923 2013-05-23 Richard Biener <rguenther@suse.de>
16925 PR rtl-optimization/57341
16926 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16927 instead of true_dependence.
16929 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16931 * bb-reorder.c (branch_threshold): Make const.
16932 (exec_threshold): Ditto.
16934 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16935 Pat Haugen <pthaugen@us.ibm.com>
16936 Peter Bergner <bergner@vnet.ibm.com>
16938 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16939 documentation for the power8 crypto builtins.
16941 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16943 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16944 macros for defining power8 builtin functions.
16945 (BU_P8V_AV_2): Likewise.
16946 (BU_P8V_AV_P): Likewise.
16947 (BU_P8V_VSX_1): Likewise.
16948 (BU_P8V_OVERLOAD_1): Likewise.
16949 (BU_P8V_OVERLOAD_2): Likewise.
16950 (BU_CRYPTO_1): Likewise.
16951 (BU_CRYPTO_2): Likewise.
16952 (BU_CRYPTO_3): Likewise.
16953 (BU_CRYPTO_OVERLOAD_1): Likewise.
16954 (BU_CRYPTO_OVERLOAD_2): Likewise.
16955 (XSCVSPDP): Fix typo, point to the correct instruction.
16956 (VCIPHER): Add power8 crypto builtins.
16957 (VCIPHERLAST): Likewise.
16958 (VNCIPHER): Likewise.
16959 (VNCIPHERLAST): Likewise.
16960 (VPMSUMB): Likewise.
16961 (VPMSUMH): Likewise.
16962 (VPMSUMW): Likewise.
16963 (VPERMXOR_V2DI): Likewise.
16964 (VPERMXOR_V4SI: Likewise.
16965 (VPERMXOR_V8HI: Likewise.
16966 (VPERMXOR_V16QI: Likewise.
16967 (VSHASIGMAW): Likewise.
16968 (VSHASIGMAD): Likewise.
16969 (VPMSUM): Likewise.
16970 (VPERMXOR): Likewise.
16971 (VSHASIGMA): Likewise.
16973 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16974 __CRYPTO__ if the crypto instructions are available.
16975 (altivec_overloaded_builtins): Add support for overloaded power8
16978 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
16979 support for power8 crypto builtins.
16980 (builtin_function_type): Likewise.
16981 (altivec_init_builtins): Add support for builtins that take vector
16982 long long (V2DI) arguments.
16984 * config/rs6000/crypto.md: New file, define power8 crypto
16987 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16988 Pat Haugen <pthaugen@us.ibm.com>
16989 Peter Bergner <bergner@vnet.ibm.com>
16991 * doc/invoke.texi (Option Summary): Add power8 options.
16992 (RS/6000 and PowerPC Options): Likewise.
16994 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
16995 constraints.md instead of rs6000.h. Reorder w* constraints. Add
16996 wm, wn, wr documentation.
16998 * config/rs6000/constraints.md (wm): New constraint for VSX
16999 registers if direct move instructions are enabled.
17000 (wn): New constraint for no registers.
17001 (wq): New constraint for quad word even GPR registers.
17002 (wr): New constraint if 64-bit instructions are enabled.
17003 (wv): New constraint if power8 vector instructions are enabled.
17004 (wQ): New constraint for quad word memory locations.
17006 * config/rs6000/predicates.md (const_0_to_15_operand): New
17007 constraint for 0..15 for crypto instructions.
17008 (gpc_reg_operand): If VSX allow registers in VSX registers as well
17009 as GPR and floating point registers.
17010 (int_reg_operand): New predicate to match only GPR registers.
17011 (base_reg_operand): New predicate to match base registers.
17012 (quad_int_reg_operand): New predicate to match even GPR registers
17013 for quad memory operations.
17014 (vsx_reg_or_cint_operand): New predicate to allow vector logical
17015 operations in both GPR and VSX registers.
17016 (quad_memory_operand): New predicate for quad memory operations.
17017 (reg_or_indexed_operand): New predicate for direct move support.
17019 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
17020 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
17021 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
17022 (POWERPC_MASKS): Add power8 options.
17023 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
17026 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
17027 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
17029 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
17030 (-mpower8-fusion): New power8 options.
17031 (-mpower8-fusion-sign): Likewise.
17032 (-mpower8-vector): Likewise.
17033 (-mcrypto): Likewise.
17034 (-mdirect-move): Likewise.
17035 (-mquad-memory): Likewise.
17037 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
17038 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
17040 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
17041 (rs6000_debug_vector_unit): Add p8_vector.
17042 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
17043 definitions. Also print fusion state.
17044 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
17045 (rs6000_builtin_mask_calculate): Add power8 builtin support.
17046 (rs6000_option_override_internal): Add support for power8.
17047 (rs6000_common_init_builtins): Add debugging for skipped builtins
17048 if -mdebug=builtin.
17049 (rs6000_adjust_cost): Add power8 support.
17050 (rs6000_issue_rate): Likewise.
17051 (insn_must_be_first_in_group): Likewise.
17052 (insn_must_be_last_in_group): Likewise.
17053 (force_new_group): Likewise.
17054 (rs6000_register_move_cost): Likewise.
17055 (rs6000_opt_masks): Likewise.
17057 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
17058 power8 capable assembler, default to power7 options.
17059 (TARGET_DIRECT_MOVE): Likewise.
17060 (TARGET_CRYPTO): Likewise.
17061 (TARGET_P8_VECTOR): Likewise.
17062 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
17063 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
17064 (VECTOR_MEM_P8_VECTOR_P): Likewise.
17065 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
17066 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
17067 (TARGET_XSCVDPSPN): Likewise.
17068 (TARGET_XSCVSPDPN): Likewsie.
17069 (TARGET_SYNC_HI_QI): Likewise.
17070 (TARGET_SYNC_TI): Likewise.
17071 (MASK_CRYPTO): Likewise.
17072 (MASK_DIRECT_MOVE): Likewise.
17073 (MASK_P8_FUSION): Likewise.
17074 (MASK_P8_VECTOR): Likewise.
17075 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
17076 temporary used by some of the direct move instructions to get two FP
17077 temporary registers does not force creation of a stack frame.
17078 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
17079 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
17080 that any VSX registers are tieable, even if they are also an
17081 Altivec vector mode.
17082 (r6000_reg_class_enum): Add wm, wr, wv constraints.
17083 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
17084 (RS6000_BTM_CRYPTO): Likewise.
17085 (RS6000_BTM_COMMON): Likewise.
17087 * config/rs6000/rs6000.md (cpu attribute): Add power8.
17088 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
17089 (enum rs6000_vector): Add power8 vector support.
17091 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17095 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
17096 (any_sibcall_could_use_r3): this and handle indirect calls.
17097 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
17099 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17101 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
17103 2013-05-22 Richard Biener <rguenther@suse.de>
17105 PR middle-end/57349
17106 * profile.c (branch_prob): Do not split blocks that are
17107 abnormally receiving from ECF_RETURNS_TWICE functions.
17109 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17111 * recog.c (offsettable_address_addr_space_p): Fix calculation of
17112 address mode. Move pointer mode initialization to the same place.
17114 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
17116 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
17117 while it has any effect.
17119 2013-05-21 Easwaran Raman <eraman@google.com>
17121 PR tree-optimization/57322
17122 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
17123 UID of the statement added to the BB to be 1.
17125 2013-05-21 Jakub Jelinek <jakub@redhat.com>
17127 PR tree-optimization/57331
17128 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
17129 of conversion from pointer type to integral type with integer.
17131 2013-05-21 Martin Jambor <mjambor@suse.cz>
17134 * ipa-prop.c (ipa_read_node_info): Process param_used and
17135 controlled_uses in the same order as when writing.
17137 2013-05-21 Magnus Granberg <baldrick@free.fr>
17140 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17142 2013-05-21 Richard Biener <rguenther@suse.de>
17144 PR tree-optimization/57318
17145 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17146 estimate stmts with side-effects as likely eliminated.
17148 2013-05-21 Richard Biener <rguenther@suse.de>
17150 PR tree-optimization/57330
17151 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17152 preserve the call stmts fntype.
17154 2013-05-21 Richard Biener <rguenther@suse.de>
17156 PR tree-optimization/57303
17157 * tree-ssa-sink.c (statement_sink_location): Improve killing
17158 stmt detection and properly handle self-assignments.
17160 2013-05-21 Christian Bruel <christian.bruel@st.com>
17162 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17163 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17164 registers. Set register size out of the PARALLEL loop.
17166 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17169 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17170 (*fmasf4, *fmasf4_media): New insns.
17172 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17174 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17175 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17176 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17177 (mips_idiv_insns): Update the comments to say that the returned
17178 instruction counts are in units of BASE_INSN_LENGTH.
17179 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17180 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17181 using 2 rather than 4 as the length of indirect MIPS16 and
17182 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17183 length of a NOP. Don't divide MIPS16 lengths by 2.
17184 (mips16_split_long_branches): Assume a branch is long if the
17185 length is greater than 4 rather than 8.
17186 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17187 rather than multiplying them by 2. Multiply instruction counts
17188 by BASE_INSN_LENGTH rather than 4.
17189 (*jump_mips16, tls_get_tp_mips16_<mode>)
17190 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17192 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17194 * config/mips/mips.md (extended_mips16): Remove branch case.
17195 (length): Remove duplicated extended_mips16 test.
17197 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17199 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17201 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17203 * recog.h (Recog_data): Rename to...
17204 (recog_data_d): ...this.
17205 (recog_data): Update accordingly.
17206 * recog.c (recog_data): Likewise.
17207 * reload.c (save_recog_data): Likewise.
17208 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17209 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17211 2013-05-17 Julian Brown <julian@codesourcery.com>
17213 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17214 found in a REG_EQUAL note, invalidate it.
17216 2013-05-17 Easwaran Raman <eraman@google.com>
17218 * tree-ssa-reassoc.c (find_insert_point): New function.
17219 (insert_stmt_after): Likewise.
17220 (get_def_stmt): Likewise.
17221 (ensure_ops_are_available): Likewise.
17222 (not_dominated_by): Likewise.
17223 (rewrite_expr_tree): Do not move statements beyond what is
17224 necessary. Remove call to swap_ops_for_binary_stmt...
17225 (reassociate_bb): ... and move it here.
17226 (build_and_add_sum): Assign UIDs for new statements.
17227 (linearize_expr): Likewise.
17228 (do_reassoc): Renumber gimple statement UIDs.
17230 2013-05-17 Jan Hubicka <jh@suse.cz>
17232 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17234 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17235 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17237 (output_weakrefs): Update.
17239 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17240 Martin Jambor <mjambor@suse.cz>
17242 PR middle-end/57276
17243 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17244 value that corresponds to the given aggval is found in values vector.
17246 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17248 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17249 sse, sse2, sse3, ssse3 and sse4a flags to options.
17251 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17253 * gengtype-state.c: (s_expr_writer): New class, to handle
17254 prettifying of output layout of s-expressions.
17255 (state_writer): New class, to write out gtype.state.
17256 (state_written_type_count): Move this variable into member data of
17258 (s_expr_writer::s_expr_writer): New code: constructor for new class
17259 (state_writer::state_writer(): ditto
17260 (s_expr_writer::write_new_line): New function
17261 (s_expr_writer::write_any_indent): ditto
17262 (s_expr_writer::begin_s_expr): ditto
17263 (s_expr_writer::end_s_expr): ditto
17264 (write_state_fileloc): convert to method of state_writer...
17265 (state_writer:: write_state_fileloc): ...and use methods of
17266 s_expr_writer to write indentation into the gtype.state output file
17267 to visually represent the hierarchical structure of the list
17269 (write_state_fields): ditto, renaming to...
17270 (state_writer::write_state_fields)
17271 (write_state_a_string): ditto, renaming to...
17272 (state_writer::write_state_a_string)
17273 (write_state_string_option): ditto, renaming to...
17274 (state_writer::write_state_string_option)
17275 (write_state_type_option): ditto, renaming to...
17276 (state_writer::write_state_type_option)
17277 (write_state_nested_option): ditto, renaming to...
17278 (state_writer::write_state_nested_option)
17279 (write_state_option): ditto, renaming to...
17280 (state_writer::write_state_option)
17281 (write_state_options): ditto, renaming to...
17282 (state_writer::write_state_options)
17283 (write_state_lang_bitmap): ditto, renaming to...
17284 (state_writer::write_state_lang_bitmap)
17285 (write_state_version): ditto, renaming to...
17286 (state_writer::write_state_version)
17287 (write_state_scalar_type): ditto, renaming to...
17288 (state_writer::write_state_scalar_type)
17289 (write_state_string_type): ditto, renaming to...
17290 (state_writer::write_state_string_type)
17291 (write_state_undefined_type): ditto, renaming to...
17292 (state_writer::write_state_undefined_type)
17293 (write_state_struct_union_type): ditto, renaming to...
17294 (state_writer::write_state_struct_union_type)
17295 (write_state_struct_type): ditto, renaming to...
17296 (state_writer::write_state_struct_type)
17297 (write_state_user_struct_type): ditto, renaming to...
17298 (state_writer::write_state_user_struct_type)
17299 (write_state_lang_struct_type): ditto, renaming to...
17300 (state_writer::write_state_lang_struct_type)
17301 (write_state_param_struct_type): ditto, renaming to...
17302 (state_writer::write_state_param_struct_type)
17303 (write_state_pointer_type): ditto, renaming to...
17304 (state_writer::write_state_pointer_type)
17305 (write_state_array_type): ditto, renaming to...
17306 (state_writer::write_state_array_type)
17307 (write_state_gc_used): ditto, renaming to...
17308 (state_writer::write_state_gc_used)
17309 (write_state_common_type_content): ditto, renaming to...
17310 (state_writer::write_state_common_type_content)
17311 (write_state_type): ditto, renaming to...
17312 (state_writer::write_state_type)
17313 (write_state_pair_list): ditto, renaming to...
17314 (state_writer::write_state_pair_list)
17315 (write_state_pair): ditto, renaming to...
17316 (state_writer::write_state_pair)
17317 (write_state_typedefs): ditto, renaming to...
17318 (state_writer::write_state_typedefs)
17319 (write_state_structures): ditto, renaming to...
17320 (state_writer::write_state_structures)
17321 (write_state_param_structs): ditto, renaming to...
17322 (state_writer::write_state_param_structs)
17323 (write_state_variables): ditto, renaming to...
17324 (state_writer::write_state_variables)
17325 (write_state_srcdir): ditto, renaming to...
17326 (state_writer::write_state_srcdir)
17327 (write_state_files_list): ditto, renaming to...
17328 (state_writer::write_state_files_list)
17329 (write_state_languages): ditto, renaming to...
17330 (state_writer::write_state_languages)
17331 (write_state): create a state_writer instance and use it when
17332 writing out the state file
17334 2013-05-17 Mike Stump <mikestump@comcast.net>
17336 PR rtl-optimization/57304
17337 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17338 accessing DF_REF_REAL_LOC.
17340 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17342 PR rtl-optimization/57281
17343 PR rtl-optimization/57300
17344 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17345 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17346 what the other splitter did if the registers are dead.
17348 2013-05-17 Richard Biener <rguenther@suse.de>
17350 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17353 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17355 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17358 2013-05-17 Marek Polacek <polacek@redhat.com>
17360 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17361 length when doing non-zero store of storing '\0' to '\0'.
17363 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17365 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17366 vect_external_def oprnd1 with loop_vinfo, try to emit
17367 optional cast, negation and and stmts on the loop preheader
17368 edge instead of into the pattern def seq.
17370 PR tree-optimization/57051
17371 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17372 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17374 2013-05-16 Nick Clifton <nickc@redhat.com>
17376 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17377 (rl78_is_naked_func): New function.
17378 (rl78_expand_prologue): Skip prologue generation for naked functions.
17379 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17380 * doc/extend.texi (naked): Add RL78 to the list of processors
17381 that supports this attribute.
17383 2013-05-16 Jeff Law <law@redhat.com>
17385 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17387 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17389 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17390 cache parameters using detect_caches_amd also for CYRIX,
17391 NSC and TM2 signatures.
17393 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17394 Dzianis Kahanovich <mahatma@eu.by>
17398 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17399 VIA/Centaur processors and determine their cache parameters
17400 using detect_caches_amd.
17402 2013-05-16 Teresa Johnson <tejohnson@google.com>
17404 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17405 (rtl_verify_edges): New function.
17406 (rtl_verify_bb_insns): Ditto.
17407 (rtl_verify_bb_pointers): Ditto.
17408 (rtl_verify_bb_insn_chain): Ditto.
17409 (rtl_verify_fallthru): Ditto.
17410 (rtl_verify_bb_layout): Ditto.
17411 (rtl_verify_flow_info_1): Outline checks into new functions.
17412 (rtl_verify_flow_info): Ditto.
17414 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17416 * cfghooks.c (copy_bbs): Add update_dominance argument.
17417 * cfghooks.h (copy_bbs): Update prototype.
17418 * tree-cfg.c (gimple_duplicate_sese_region):
17419 Add update_dominance argument.
17420 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17421 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17422 gimple_duplicate_sese_region call.
17423 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17424 Update copy_bbs call.
17425 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17426 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17428 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17430 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17431 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17432 vect_recog_rotate_pattern.
17433 (vect_recog_rotate_pattern): New function.
17435 2013-05-16 Jason Merrill <jason@redhat.com>
17437 * Makefile.in (LLINKER): New variable.
17438 (mostlyclean): Remove link mutex.
17439 * configure.ac: Handle --enable-link-mutex.
17440 * lock-and-run.sh: New script.
17442 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17445 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17448 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17450 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17452 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17454 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17455 * config/arm/arm.c (next_consecutive_mem): New function.
17456 (gen_movmem_ldrd_strd): Likewise.
17457 * config/arm/arm.md (movmemqi): Update condition and code.
17458 (unaligned_loaddi, unaligned_storedi): New patterns.
17460 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17462 * config.gcc: Obsolete *-*-solaris2.9*.
17463 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17465 2013-05-16 Richard Biener <rguenther@suse.de>
17467 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17468 earlier, after GRAPHITE transforms and IV canonicalization.
17470 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17472 * omp-low.c (extract_omp_for_data): For collapsed loops,
17473 if at least one of the loops is known at compile time to
17474 iterate zero times, set count to 0.
17475 (expand_omp_regimplify_p): New function.
17476 (expand_omp_for_generic): For collapsed loops, if at least
17477 one of the loops isn't known to iterate at least once,
17478 add runtime check with setting count to 0.
17479 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17480 For unsigned types if it isn't known at compile time that
17481 the loop will iterate at least once, add runtime check to bypass
17482 the whole loop if initial condition isn't true.
17484 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17486 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17488 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17490 PR middle-end/57286
17491 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17492 transformations to avoid an infinite loop.
17494 2013-05-16 Marek Polacek <polacek@redhat.com>
17496 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17498 2013-05-15 Leif Ekblad <leif@rdos.net>
17500 * config/i386/i386.c (ix86_decompose_address): Use
17501 DEFAULT_TLS_SEG_REG to access TLS segment register.
17502 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17503 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17504 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17506 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17509 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17510 sibling calls to functions that would normally be lazily bound,
17511 unless $gp is call-clobbered.
17513 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17515 * config/i386/i386.c (ix86_option_override_internal): Update
17516 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17517 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17518 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17519 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17521 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17523 2013-05-15 Andreas Schwab <schwab@suse.de>
17525 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17526 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17529 2013-05-15 Teresa Johnson <tejohnson@google.com>
17531 * loop-unroll.c (report_unroll_peel): Check decision before
17532 emitting unroll/peel message.
17534 2013-05-15 Teresa Johnson <tejohnson@google.com>
17536 * function.h (has_bb_partition): New rtl_data flag.
17537 (bb_reorder_complete): Ditto.
17538 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17539 instead of flag_reorder_blocks_and_partition.
17540 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17541 with some enhancements.
17542 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17543 * bb-reorder.c (connect_traces): Check for has_bb_partition
17544 instead of flag_reorder_blocks_and_partition.
17545 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17546 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17547 verify_hot_cold_block_grouping.
17548 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17550 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17553 * config/arm/predicates.md (call_insn_operand): New predicate.
17554 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17555 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17556 if insn is not a tail call.
17557 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17559 * config/arm/arm.h (enum reg_class): New caller save register class.
17560 (REG_CLASS_NAMES): Likewise.
17561 (REG_CLASS_CONTENTS): Likewise.
17562 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17565 2013-05-15 Richard Biener <rguenther@suse.de>
17567 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17568 of MSG_OPTIMIZED_LOCATIONS.
17569 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17570 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17572 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17573 of MSG_OPTIMIZED_LOCATIONS.
17574 (execute_vect_slp): Likewise.
17575 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17576 (vect_create_cond_for_alias_checks): Likewise.
17577 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17578 (vect_recog_widen_mult_pattern): Likewise.
17579 (vect_recog_widen_sum_pattern): Likewise.
17580 (vect_recog_over_widening_pattern): Likewise.
17581 (vect_recog_widen_shift_pattern): Likewise.
17582 (vect_recog_vector_vector_shift_pattern): Likewise.
17583 (vect_recog_divmod_pattern): Likewise.
17584 (vect_recog_mixed_size_cond_pattern): Likewise.
17585 (vect_recog_bool_pattern): Likewise.
17586 (vect_pattern_recog_1): Likewise.
17588 2013-05-15 Martin Jambor <mjambor@suse.cz>
17590 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17591 non-functions to builtin_unreachable.
17592 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17593 correct when new direct edges were discovered.
17595 2013-05-15 Martin Jambor <mjambor@suse.cz>
17597 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17598 header, print symbol order instead of node uid, print more information
17599 about indirect edge targets.
17600 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17602 (ipa_make_edge_direct_to_target): Likewise.
17603 (remove_described_reference): Likewise.
17604 (propagate_controlled_uses): Likewise.
17605 (ipa_print_node_params): Also print symbol order.
17606 (ipcp_transform_function): Print symbol order instead of node uids.
17607 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17608 (cgraph_get_create_real_symbol_node): Likewise.
17609 * ipa-cp.c (print_lattice): Likewise.
17610 (print_all_lattices): Likewise.
17611 (determine_versionability): Likewise.
17612 (initialize_node_lattices): Likewise.
17613 (estimate_local_effects): Likewise.
17614 (update_profiling_info): Likewise.
17615 (create_specialized_node): Likewise.
17616 (perhaps_add_new_callers): Likewise.
17617 (decide_about_value): Likewise.
17618 (decide_whether_version_node): Likewise.
17619 (identify_dead_nodes): Likewise.
17620 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17621 (dump_inline_summary): Likewise.
17622 (estimate_node_size_and_time): Likewise.
17623 (inline_analyze_function): Likewise.
17624 * ipa-inline.c (report_inline_failed_reason): Likewise.
17625 (want_early_inline_function_p): Likewise.
17626 (edge_badness): Likewise.
17627 (update_edge_key): Likewise.
17628 (inline_small_functions): Likewise. Add dumping of order to two other
17630 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17631 instead of node uids.
17632 (propagate_pure_const): Likewise.
17633 (propagate_pure_const): Likewise.
17634 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17635 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17636 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17638 * tree-pretty-print.c (dump_function_header): Likewise.
17639 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17640 Print symbol order instead of node uids.
17642 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17644 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17645 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17647 2013-05-15 Richard Biener <rguenther@suse.de>
17649 PR tree-optimization/57275
17650 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17651 return value for fail to do runtime alias checks for gather loads.
17653 2013-05-15 Jan Hubicka <jh@suse.cz>
17657 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17658 weakrefs are not external.
17659 (lto_symtab_merge_decls): Fix thinko when dealing with
17660 non-lto_symtab decls.
17661 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17662 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17663 * varpool.c (dump_varpool_node): Dump more flags.
17665 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17667 * config/i386/i386.c (processor_alias_table): Add instruction
17668 FSGSBASE for AMD bdver3 architecture.
17670 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17672 * tree.c (warn_deprecated_use): Print file:line using locus color.
17673 * diagnostic.c (diagnostic_report_current_module): Print file:line
17674 and file:line:column using locus color.
17676 2013-05-14 Mike Stump <mikestump@comcast.net>
17678 * gdbinit.in: Add __null.
17680 2013-05-14 Mike Stump <mikestump@comcast.net>
17682 * recog.h: Rename struct recog_data to Recog_data.
17683 * recog.c: Likewise.
17684 * reload.c (can_reload_into): Likewise.
17685 * config/picochip/picochip.c: Likewise.
17687 2013-05-14 Mike Stump <mikestump@comcast.net>
17689 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17691 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17693 * resource.h (struct resources): Remove unch_memory member.
17694 (CLEAR_RESOURCE): Don't clear unch_memory.
17695 * resource.c (mark_referenced_resources): Don't set it.
17696 (mark_set_resources): Likewise.
17697 (mark_target_live_regs): Don't clear it.
17698 (init_resource_info): Likewise.
17699 * reorg.c (resource_conflicts_p): Don't compare it.
17700 (redundant_insn): Don't set it.
17702 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17704 * emit-rtl.c (next_label): Remove unused function.
17705 (skip_consecutive_labels, link_cc0_insns): Move to ...
17706 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17707 only place where these functions are used, and make them static.
17709 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17711 * fold-const.c (fold_negate_expr): Handle vectors.
17712 (fold_truth_not_expr): Make it static.
17713 (fold_invert_truthvalue): New static function.
17714 (invert_truthvalue_loc): Handle vectors. Do not call
17715 fold_truth_not_expr directly.
17716 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17717 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17718 (fold_comparison): Handle vector constants.
17719 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17720 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17721 * tree.h (fold_truth_not_expr): Remove declaration.
17723 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17725 * config/aarch64/aarch64-simd.md
17726 (aarch64_vcond_internal<mode>): Rename to...
17727 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17728 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17729 float modes. Clarify all iterator modes.
17730 (vcond<mode><mode>): Use new name for vcond expanders.
17731 (vcond<v_cmp_result><mode>): Likewise.
17732 (vcondu<mode><mode>: Likewise.
17733 * config/aarch64/iterators.md (VDQF_COND): New.
17735 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17738 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17739 variable for the shift amount. Check that we shift by non-negative
17742 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17745 * config/arm/arm.h (EPILOGUE_USES): Only return true
17746 for LR_REGNUM after epilogue_completed.
17748 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17750 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17751 is error_mark_node.
17753 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17756 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17757 and Solaris 11+/x86 with gld.
17758 * configure: Regenerate.
17760 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17762 * expmed.c (expand_shift_1): Canonicalize rotates by
17763 constant bitsize / 2 to bitsize - 1.
17764 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17765 case ROTATERT>: Likewise.
17768 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17770 * config/i386/i386.md (rotateinv): New code attr.
17771 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17772 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17773 roll $31, %eax, etc.
17775 2013-05-14 Richard Biener <rguenther@suse.de>
17777 PR middle-end/57235
17778 * tree-eh.c (sink_clobbers): Give up for successors with
17779 multiple predecessors and no virtual uses.
17781 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17783 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17784 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17786 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17788 PR middle-end/57251
17789 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17790 the case when both op0 and op1 have VOIDmode.
17792 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17794 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17795 in multiply-accumulate mode.
17797 2013-05-13 Guozhi Wei <carrot@google.com>
17799 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17801 2013-05-13 Kai Tietz <ktietz@redhat.com>
17804 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17805 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17806 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17807 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17808 (ix86_expand_prologue): Likewise.
17809 (ix86_expand_split_stack_prologue): Likewise.
17810 (legitimate_pic_address_disp_p): Likewise.
17811 (legitimize_pic_address): Likewise.
17812 (legitimize_tls_address): Likewise.
17813 (legitimize_pe_coff_symbol): Likewise.
17814 (output_pic_addr_const): Likewise.
17815 (construct_plt_address): Likewise.
17816 (ix86_expand_call): Likewise.
17817 (x86_output_mi_thunk): Likewise.
17818 (x86_function_profiler): Likewise.
17820 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17822 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17823 similar switch cases.
17824 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17825 (aarch64_simd_mov_to_<mode>low): Delete.
17826 (aarch64_simd_mov_to_<mode>high): Delete.
17827 (move_lo_quad_<mode>): Add w<-r alternative.
17828 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17829 (aarch64_simd_mov_from_*): Update type attribute.
17830 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17833 2013-05-13 Jan Hubicka <jh@suse.cz>
17835 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17836 * config/i386/i386.c (ix86_compute_frame_layout,
17837 ix86_expand_epilogue, emit_i387_cw_initialization,
17838 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17839 ix86_local_alignment): Fix use of size/speed predicates.
17841 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17843 PR tree-optimization/45216
17844 PR tree-optimization/57157
17845 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17846 the (-Y) & (B - 1) variant if OP is |.
17847 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17848 return shifted. Use (-op1) & (prec - 1) as other_amount
17849 instead of prec - op1.
17851 2013-05-13 Martin Jambor <mjambor@suse.cz>
17853 PR middle-end/42371
17854 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17855 (ipa_constant_data): New type.
17856 (ipa_jump_func): Use ipa_constant_data to hold information about
17857 constant jump functions.
17858 (ipa_get_jf_constant): Adjust to jump function type changes.
17859 (ipa_get_jf_constant_rdesc): New function.
17860 (ipa_param_descriptor): New field controlled_uses.
17861 (ipa_get_controlled_uses): New function.
17862 (ipa_set_controlled_uses): Likewise.
17863 * ipa-ref.h (ipa_find_reference): Declare.
17864 * ipa-prop.c (ipa_cst_ref_desc): New type.
17865 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17867 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17868 New parameter cs. Adjust all callers.
17869 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17870 (remove_described_reference): New function.
17871 (jfunc_rdesc_usable): Likewise.
17872 (try_make_edge_direct_simple_call): Decrement controlled use count,
17873 attempt to remove reference if it hits zero.
17874 (combine_controlled_uses_counters): New function.
17875 (propagate_controlled_uses): Likewise.
17876 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17877 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17878 (ipa_print_node_params): Print described use counter.
17879 (ipa_write_jump_function): Adjust to jump function type changes.
17880 (ipa_read_jump_function): New parameter CS, pass it to
17881 ipa_set_jf_constant. Adjust caller.
17882 (ipa_write_node_info): Stream controlled use count
17883 (ipa_read_node_info): Likewise.
17884 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17886 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17887 count. Remove cloning-added reference if it reaches zero.
17888 * ipa-ref.c (ipa_find_reference): New function.
17890 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17892 * config/i386/i386.c (processor_target_table): Modified default
17893 alignment values for AMD BD and BT architectures.
17895 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17897 * tree-vect-generic.c (uniform_vector_p): Move ...
17898 * tree.c (uniform_vector_p): ... here.
17899 * tree.h (uniform_vector_p): Declare it.
17900 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17903 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17905 PR tree-optimization/57230
17906 * tree-ssa-strlen.c (handle_char_store): Record length for
17907 array store from STRING_CST.
17909 PR tree-optimization/57230
17910 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17913 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17915 * config/epiphany/epiphany.c (epiphany_init): Check size of
17916 NUM_MODES_FOR_MODE_SWITCHING.
17917 (epiphany_expand_prologue):
17918 Remove CONFIG_REGNUM initial value handling code.
17919 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17920 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17921 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17922 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17923 Don't return 1 for FP_MODE_NONE.
17924 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17925 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17926 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17927 * config/epiphany/epiphany.md (save_config): New pattern.
17929 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17931 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17933 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17935 * config/i386/i386.md (memory): Handle sseishft1.
17936 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17937 (*vec_extractv2di_1): Ditto.
17939 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17941 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17944 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17946 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17947 Add mthumb/march=armv7-a multilib.
17948 Add mthumb/march=armv7-r multilib.
17949 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17951 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17953 * config/v850/t-rtems: Add more multilibs.
17955 2013-05-10 Richard Biener <rguenther@suse.de>
17957 PR tree-optimization/57214
17958 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
17959 not propagate from SSA names that occur in abnormal PHI nodes.
17961 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
17963 * stor-layout.c (element_precision): New function.
17964 * machmode.h (element_precision): Declare it.
17965 * tree.c (build_minus_one_cst): New function.
17966 (element_precision): Likewise.
17967 * tree.h (build_minus_one_cst): Declare new function.
17968 (element_precision): Likewise.
17969 * fold-const.c (operand_equal_p): Use element_precision.
17970 (fold_binary_loc): Handle vector types.
17971 * convert.c (convert_to_integer): Use element_precision.
17972 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
17975 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
17977 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
17978 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
17979 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
17980 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17981 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
17982 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
17983 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
17984 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
17985 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
17986 (Uuw8): New constraints.
17987 (Usb4): Move into alphabetical order.
17988 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
17989 (sd8_operand, ub8_operand, uw8_operand): New predicates.
17990 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
17991 previously unnamed patterns.
17992 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
17993 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
17994 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
17995 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
17996 of set_attr_alternative/if_then_else. Use extended_mips16 instead
17997 of specific lengths.
17999 2013-05-10 Jakub Jelinek <jakub@redhat.com>
18001 * config/i386/i386.md (rotateinv): New code attr.
18002 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
18003 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
18004 roll $31, %eax, etc.
18006 PR tree-optimization/45216
18007 PR tree-optimization/57157
18008 * tree-ssa-forwprop.c (simplify_rotate): New function.
18009 (ssa_forward_propagate_and_combine): Call it.
18011 2013-05-10 Richard Biener <rguenther@suse.de>
18013 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
18014 disable peeling when we version for aliasing.
18015 (vector_alignment_reachable_p): Honor explicit user alignment.
18016 (vect_supportable_dr_alignment): Likewise.
18017 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
18018 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
18019 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
18020 then peeling to arrange for the cost-model check to come first.
18022 2013-05-10 Alan Modra <amodra@gmail.com>
18024 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
18025 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
18026 * configure: Regenerate.
18028 2013-05-10 Alan Modra <amodra@gmail.com>
18031 * varasm.c (default_elf_select_section): Move !DECL_P check..
18032 (get_named_section): ..to here before calling get_section_name.
18034 (default_section_type_flags): Add DECL_P check.
18035 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
18036 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
18038 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
18040 * config/epiphany/epiphany.c (epiphany_expand_prologue):
18041 When using gen_stack_adjust_str with a register offset, add a
18042 REG_FRAME_RELATED_EXPR note.
18044 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
18046 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
18047 (*vec_extractv4si_zext_mem): Ditto.
18048 (*vec_extractv2di): Add 0->x and x->x alternatives.
18049 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
18050 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
18052 2013-05-09 Jason Merrill <jason@redhat.com>
18054 N3639 C++1y VLA support
18055 * gimplify.c (gimplify_vla_decl): Don't touch an existing
18058 * tree.c (build_constructor_va): New.
18059 * tree.h: Declare it.
18061 2013-05-09 Martin Jambor <mjambor@suse.cz>
18064 * gimple-fold.c (canonicalize_constructor_val): Call
18065 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
18067 2013-05-09 Jan Hubicka <jh@suse.cz>
18068 Richard Biener <rguenther@suse.de>
18071 * symtab.c (symtab_make_decl_local): Do not add private names.
18073 2013-05-09 Jan Hubicka <jh@suse.cz>
18076 * symtab.c (insert_to_assembler_name_hash): Handle clones.
18077 (unlink_from_assembler_name_hash): Likewise.
18078 (symtab_prevail_in_asm_name_hash, symtab_register_node,
18079 symtab_unregister_node, symtab_initialize_asm_name_hash,
18080 change_decl_assembler_name): Update.
18082 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
18084 * config/aarch64/aarch64.md: New movtf split.
18085 (*movtf_aarch64): Update.
18086 (aarch64_movdi_tilow): Handle TF modes and rename to
18087 aarch64_movdi_<mode>low.
18088 (aarch64_movdi_tihigh): Handle TF modes and rename to
18089 aarch64_movdi_<mode>high
18090 (aarch64_movtihigh_di): Handle TF modes and rename to
18091 aarch64_mov<mode>high_di
18092 (aarch64_movtilow_di): Handle TF modes and rename to
18093 aarch64_mov<mode>low_di
18094 (aarch64_movtilow_tilow): Remove spurious whitespace.
18095 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
18097 (aarch64_print_operand): Update.
18099 2013-05-09 Alan Modra <amodra@gmail.com>
18101 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
18103 * configure: Regenerate.
18105 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18107 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
18108 splitter preparation statements.
18109 * config/i386/sse.md (*vec_extract* splitters): Ditto.
18110 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
18113 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18115 * gimple-ssa-strength-reduction.c (count_candidates): Change
18116 return value to int.
18117 (analyze_candidates_and_replace): Change type of length to int.
18119 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18121 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
18122 (*vec_extract<mode>): Use VI12_128 mode iterator.
18123 (*vec_extract<mode>_mem): Ditto.
18124 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
18127 2013-05-08 Diego Novillo <dnovillo@google.com>
18132 2012-08-17 Diego Novillo <dnovillo@google.com>
18135 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18136 * config.in: Regenerate.
18137 * configure: Regenerate.
18138 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18140 2013-05-08 Jan Hubicka <jh@suse.cz>
18143 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18144 * cgraph.h (symtab_node_base): Add unique_name.
18145 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18146 input_overwrite_node, input_varpool_node): Stream unique_name.
18147 * cgraphclones.c (cgraph_create_virtual_clone,
18148 cgraph_function_versioning): Set unique_name.
18149 * ipa.c (function_and_variable_visibility): Set unique_name.
18151 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18153 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18154 (alloc_cand_and_find_basis): Restrict conditional candidate
18155 processing to CAND_MULTs.
18157 2013-05-08 Jan Hubicka <jh@suse.cz>
18160 lto-symtab.c (lto_symtab_symbol_p): New function.
18161 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18162 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18163 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18164 Skip static symbols.
18166 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18168 PR tree-optimization/57200
18169 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18170 Only call inform if the preceding warning_at returns true.
18172 2013-05-07 Han Shen <shenhan@google.com>
18174 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18175 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18176 * common.opt (fstack-protector-strong): New option.
18177 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18178 * doc/invoke.texi (Optimization Options): Document
18179 "-fstack-protector-strong".
18180 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18182 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18184 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18186 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18188 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18189 it is not lhs of the insn.
18190 (setup_reg_equiv): Remove insn having equiv memory which it is not
18192 * lra-constraints.c (process_address): Try to improve generation
18193 code for address base + disp.
18194 (lra_constraints): Make correct the code for checking insn setting
18195 up backward equivalence. Remove insn only if it is in the init
18197 * lra-eliminations.c (update_reg_eliminate): Change return value.
18198 (lra_eliminate): Use the result.
18200 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18202 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18203 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18204 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18205 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18206 register target operands.
18207 (*vec_extractv8hi_sse2): New pattern.
18208 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18209 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18210 (*vec_extract<mode>_mem): New insn and split pattern.
18212 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18214 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18215 (TARGET_ASAN_SHADOW_OFFSET): Define.
18216 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18217 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18219 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18221 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18222 (incr_vec_index): Return -1 if increment not found.
18223 (create_add_on_incoming_edge): Assert if increment not found.
18224 (record_increment): Limit number of increments recorded.
18225 (all_phi_incrs_profitable): Return false if an increment not found.
18226 (replace_profitable_candidates): Don't process increments that were
18228 (analyze_candidates_and_replace): Limit size of incr_vec.
18230 2013-05-07 Richard Biener <rguenther@suse.de>
18232 * calls.c (special_function_p): setjmp-like functions are leaf.
18233 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18234 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18236 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18238 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18239 (aarch64_simd_mov<mode>): New expander.
18240 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18241 (aarch64_simd_mov_to_<mode>high): Likewise.
18242 (aarch64_simd_mov_from_<mode>low): Likewise.
18243 (aarch64_simd_mov_from_<mode>high): Likewise.
18244 (aarch64_dup_lane<mode>): Update.
18245 (aarch64_dup_lanedi): New instruction pattern.
18246 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18247 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18249 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18251 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18252 (replace_mult_candidate): Remove unnecessary argument; remove
18253 unnecessary parameter from call to introduce_cast_before_cand.
18254 (replace_unconditional_candidate): Remove unnecessary parameter
18255 from call to replace_mult_candidate.
18256 (replace_conditional_candidate): Likewise.
18257 (insert_initializers): Use make_temp_ssa_name.
18258 (introduce_cast_before_cand): Remove unnecessary argument; use
18259 make_temp_ssa_name.
18260 (replace_one_candidate): Remove unnecessary argument; remove
18261 unnecessary parameter from calls to introduce_cast_before_cand.
18262 (replace_profitable_candidates): Remove unnecessary parameters
18263 from calls to replace_one_candidate.
18265 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18267 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18268 phi def as possibly hiding a basis for a CAND_ADD whose operands
18269 have been commuted in the analysis.
18270 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18272 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18274 * config/aarch64/aarch64.md
18275 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18276 shift value between 0-4.
18278 2013-05-07 Richard Biener <rguenther@suse.de>
18280 * double-int.h (rshift): New overload.
18281 * double-int.c (rshift): New function.
18282 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18283 (create_reference_ops_from_ref): Remove.
18284 (vn_reference_insert): Use shared ops for constructing the
18285 reference and copy it.
18287 2013-05-07 Richard Biener <rguenther@suse.de>
18289 PR middle-end/57190
18290 * tree-eh.c (sink_clobbers): Properly propagate
18291 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18293 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18295 PR tree-optimization/57149
18296 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18297 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18298 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18299 uninit_undefined_value_p instead of ssa_undefined_value_p.
18302 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18303 for modifier == EXPAND_INITIALIZER.
18305 2013-05-07 Anton Blanchard <anton@samba.org>
18307 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18308 for powerpc64 little endian.
18309 * configure: Regenerate.
18311 2013-05-06 Graham Stott <grahams@btinternet.com>
18313 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18314 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18315 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18316 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18318 2013-05-06 Graham Stott <grahams@btinternet.com>
18320 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18321 codes which allow non-lvalues.
18323 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18325 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18326 components are all 1s.
18327 (integer_minus_onep): New function.
18328 * tree.h (integer_minus_onep): Declare it.
18329 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18330 integer_minus_onep instead of integer_all_onesp.
18332 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18335 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18336 variations of these patterns.
18338 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18340 * config/i386/i386.md (isa): Add x64_sse4 member.
18341 (enabled): Handle x64_sse4.
18342 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18343 instruction for 64bit SSE4_1 targets. Update insn attributes.
18344 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18345 instruction for SSE4_1 targets. Update insn attributes.
18346 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18347 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18348 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18350 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18351 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18352 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18354 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18357 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18359 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18361 * final.c (do_assembler_dialects): Don't handle curly braces and
18362 vertical bar escaped by % as dialect delimiters.
18363 (output_asm_insn): Print curly braces and vertical bar if escaped
18364 by % and ASSEMBLER_DIALECT defined.
18365 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18366 * doc/tm.texi: Regenerated.
18368 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18370 * config/mips/mips.c: Include tree-pass.h.
18371 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18372 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18373 (pass_mips_machine_reorg2): New machine specific pass.
18374 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18375 (mips_option_override): Register the new pass.
18376 * rtl.h (cleanup_barriers): Remove prototype.
18377 (dbr_schedule): Likewise.
18378 * jump.c (cleanup_barriers): Make static.
18379 * reorg.c (dbr_schedule): Likewise.
18381 2013-05-06 Richard Biener <rguenther@suse.de>
18383 PR tree-optimization/57185
18384 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18385 SSA names properly.
18387 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18390 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18391 between "first_arg" and "insn", not "last" and "insn".
18393 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18395 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18396 (find_candidates_in_block): Re-enable slsr_process_phi.
18397 (create_phi_basis): Fix double counting of candidate adjustment.
18399 2013-05-06 Richard Biener <rguenther@suse.de>
18401 PR middle-end/57147
18402 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18403 the edge is also fallthru, preserve it and just clear the
18405 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18406 also complex, preserve that and just clear the fallthru flag.
18407 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18408 update virtual operands.
18410 2013-05-06 Alan Modra <amodra@gmail.com>
18412 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18413 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18414 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18415 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18416 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18417 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18419 2013-05-06 Alan Modra <amodra@gmail.com>
18421 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18422 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18423 (DEFAULT_ASM_ENDIAN): Define.
18424 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18425 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18426 Update -K PIC clause from sysv4.h.
18427 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18428 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18430 2013-05-06 Alan Modra <amodra@gmail.com>
18432 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18433 twice for little-endian.
18434 (ashrdi3_no_power, ashrdi3): Support little-endian.
18436 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18439 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18440 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18442 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18443 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18444 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18447 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18448 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18450 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18451 * config/pa/pa.opt: Make mbig-switch a no-op.
18452 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18453 (CASE_VECTOR_MODE): Always return SImode.
18454 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18455 for the !TARGET_BIG_SWITCH case.
18456 * config/pa/pa-linux.h: Likewise.
18457 * config/pa/pa-openbsd.h: Likewise.
18458 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18459 * config/pa/pa.md (short_jump): Remove define_insn.
18460 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18461 (casesi0): Remove define_insn.
18462 (type): Remove btable_branch.
18463 (pa_combine_type): Likewise.
18464 (in_nullified_branch_delay): Likewise.
18465 (in_call_delay): Likewise.
18466 (define_delay): Likewise.
18467 (define_insn_reservation "Z3"): Likewise.
18468 (define_insn_reservation "Z4"): Likewise.
18469 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18470 (pa_adjust_insn_length): Remove adjustment for btable branches.
18471 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18474 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18476 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18477 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18478 Add m->r,x alternatives.
18479 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18480 splitters using SWI48x mode iterator.
18481 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18482 TARGET_64BIT. Add m->x alternative.
18483 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18484 Add o->x alternative. Enable for TARGET_SSE.
18485 (sse_storeq): Remove expander.
18486 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18487 with memory input operand.
18488 (*vec_extractv2di_1 splitter): New.
18489 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18490 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18492 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18494 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18495 (INT_LOWPART): Delete.
18496 (extract_MB): Adjust.
18497 (extract_ME): Adjust.
18498 (print_operand): Adjust.
18500 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18502 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18503 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18504 (reg_or_logical_cint_operand, easy_fp_constant,
18505 logical_const_operand): Delete "CONST_DOUBLE" case.
18506 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18507 "HOST_BITS_PER_WIDE_INT == 64" test.
18508 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18509 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18510 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18511 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18513 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18514 CONST_DOUBLE DImode/VOIDmode case.
18515 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18516 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18517 CONST_DOUBLE VOIDmode case.
18518 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18519 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18520 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18521 Delete CONST_DOUBLE case.
18522 (splitters for mov FMOVE64 const_double): Delete
18523 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18524 "HOST_BITS_PER_WIDE_INT >= 64" test.
18525 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18527 (mov DI const_double): Delete.
18529 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18531 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18532 on op shows all bits zero in mode of a lowpart subreg, return zero.
18534 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18537 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18538 to save TFmode registers and DImode to save TImode registers for
18539 caller save operations.
18540 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18541 mark being partially clobbered since they only use the first
18544 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18545 and TDmode only use the upper 64-bits of each VSX register.
18547 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18549 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18550 (find_candidates_in_block): Disable slsr_process_phi.
18552 2013-05-03 Guozhi Wei <carrot@google.com>
18554 * coverage.c (coverage_obj_init): Move the construction of gcov
18556 (build_init_ctor): ... here.
18558 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18560 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18561 (slsr_cand_d): Redefine def_phi.
18562 (stride_status, phi_adjust_status, count_phis_status): New enums.
18563 (find_phi_def): New.
18564 (find_basis_for_base_expr): New.
18565 (find_basis_for_candidate): Handle hidden bases.
18566 (alloc_cand_and_find_basis): Handle phi candidates.
18567 (slsr_process_phi): New.
18568 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18569 (create_mul_imm_cand): Likewise.
18570 (create_add_ssa_cand): Exclude phi base candidates.
18571 (create_add_imm_cand): Likewise.
18572 (slsr_process_cast): Likewise.
18573 (slsr_process_copy): Likewise.
18574 (find_candidates_in_block): Handle phi candidates.
18575 (dump_candidate): Likewise.
18576 (unconditional_cands): Delete.
18577 (unconditional_cands_with_known_stride_p): Delete.
18578 (phi_dependent_cand_p): New.
18579 (cand_increment): Handle phi-dependent candidates.
18580 (replace_dependent): Delete.
18581 (replace_mult_candidate): New.
18582 (replace_unconditional_candidate): New.
18583 (incr_vec_index): Move to avoid forward reference.
18584 (create_add_on_incoming_edge): New.
18585 (create_phi_basis): New.
18586 (replace_dependents): Delete.
18587 (replace_conditional_candidate): New.
18588 (phi_add_costs): New.
18589 (replace_uncond_cands_and_profitable_phis): New.
18590 (record_increment): Handle phi adjustments.
18591 (record_phi_increments): New.
18592 (record_increments): Handle phi adjustments.
18593 (phi_incr_cost): New.
18594 (lowest_cost_path): Handle phis.
18595 (total_savings): Likewise.
18596 (analyze_increments): Likewise.
18597 (ncd_with_phi): New.
18598 (ncd_of_cand_and_phis): New.
18599 (nearest_common_dominator_for_cands): Handle phi increments.
18600 (all_phi_incrs_profitable): New.
18601 (replace_profitable_candidates): Handle phi-dependent candidates.
18602 (analyze_candidates_and_replace): Likewise.
18604 2013-05-03 Teresa Johnson <tejohnson@google.com>
18607 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18608 do not exceed REG_BR_PROB_BASE.
18610 2013-05-03 Jeff Law <law@redhat.com>
18612 PR tree-optimization/57144
18613 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18614 operand of the condition will bit into the new type when eliminating
18615 a cast feeding a condition.
18617 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18619 PR rtl-optimization/57130
18620 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18621 of COMPARE as in_code to the recursive call if needed.
18623 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18625 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18626 (enabled): Handle new members.
18627 * config/i386/sse.md (*vec_concatv2si): Merge from
18628 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18629 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18631 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18633 PR tree-optimization/57027
18634 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18635 for fnms opportunity, check we got the prerequisite kind
18636 of tree / gimple before using accessor functions.
18638 2013-05-03 Richard Biener <rguenther@suse.de>
18640 * double-int.h (lshift): New overload without precision
18641 and arith argument.
18642 (operator *=, operator +=, operator -=): Move ...
18643 * double-int.c (operator *=, operator +=, operator -=): ... here
18644 and implement more efficiently.
18645 (mul_double_with_sign): Remove.
18646 (lshift_double): Adjust to take unsinged shift argument, push
18647 dispatching code to callers.
18648 (mul_double_wide_with_sign): Add early out for callers that
18649 are not interested in high parts or overflow.
18650 (lshift): New function.
18651 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18652 dispatch code here.
18653 (lrotate, rrotate): Use logical shifts.
18654 * expr.c (get_inner_reference): Use lshift.
18655 * fixed-value.c (do_fixed_divide): Likewise.
18656 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18657 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18658 (indirect_refs_may_alias_p): Likewise.
18659 (stmt_kills_ref_p_1): Likewise.
18661 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18663 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18665 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18667 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18668 scalar form of FABD instruction.
18670 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18672 * lra-constraints.c (process_alt_operands): Add checking alt
18673 number to choose the best alternative.
18675 2013-05-02 Richard Biener <rguenther@suse.de>
18677 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18678 bitmap and its handling.
18679 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18681 2013-05-02 Richard Biener <rguenther@suse.de>
18683 PR middle-end/57140
18684 * tree-inline.c (copy_loops): Properly handle removed loops.
18685 (copy_cfg_body): Mark destination loops for fixup if source
18686 loops needed fixup.
18688 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18691 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18692 generating simple_return for naked functions.
18694 2013-05-02 Martin Jambor <mjambor@suse.cz>
18696 PR middle-end/56988
18697 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18698 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18700 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18701 ipa_agg_replacement_value structures.
18702 (known_aggs_to_agg_replacement_list): Likewise.
18703 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18704 (read_agg_replacement_chain): Likewise.
18705 (ipcp_transform_function): Also check that by_ref flags match.
18707 2013-05-02 Richard Biener <rguenther@suse.de>
18709 * graphds.h (struct graph): Add obstack member.
18710 * graphds.c (new_graph): Initialize obstack and allocate
18712 (add_edge): Allocate edge from the obstack.
18713 (free_graph): Free the obstack instead of all edges and vertices.
18715 2013-05-02 Teresa Johnson <tejohnson@google.com>
18717 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18719 * cfg.c (update_bb_profile_for_threading): Ditto.
18720 * tree-inline.c (copy_bb): Ditto.
18721 (copy_edges_for_bb): Ditto.
18722 (initialize_cfun): Ditto.
18723 (copy_cfg_body): Ditto.
18724 (expand_call_inline): Ditto.
18725 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18726 (estimate_node_size_and_time): Ditto.
18727 (inline_merge_summary): Ditto.
18728 * cgraphclones.c (cgraph_clone_edge): Ditto.
18729 (cgraph_clone_node): Ditto.
18730 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18731 (compute_trg_info): Ditto.
18733 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18735 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18736 S reg when fp attribute set.
18737 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18739 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18741 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18743 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18744 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18745 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18747 2013-05-02 Richard Biener <rguenther@suse.de>
18749 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18750 (struct instantiate_cache_entry): New type.
18751 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18752 (struct instantiate_cache_type): New type.
18753 (set_instantiated_value, get_instantiated_value): Remove.
18754 (get_instantiated_value_entry): New function.
18755 (instantiate_scev_name): Use the new cache and adjust.
18756 (instantiate_scev_poly): Adjust.
18757 (instantiate_scev_binary): Likewise.
18758 (instantiate_array_ref): Likewise.
18759 (instantiate_scev_convert): Likewise.
18760 (instantiate_scev_not): Likewise.
18761 (instantiate_scev_3): Likewise.
18762 (instantiate_scev_2): Likewise.
18763 (instantiate_scev_r): Likewise.
18764 (instantiate_scev): Likewise.
18765 (resolve_mixers): Likewise.
18767 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18770 * lra-constraints.c (best_small_class_operands_num): Remove.
18771 (process_alt_operands): Remove small_class_operands_num. Take
18772 small classes operands into losers and only if the operand is not
18773 matched. Modify debugging output.
18774 (curr_insn_transform): Remove best_small_class_operands_num.
18777 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18779 * config/aarch64/aarch64-builtins.c
18780 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18781 * config/aarch64/aarch64-simd-builtins.def
18782 (reduc_splus_): Add new modes.
18783 (reduc_uplus_): New.
18784 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18785 (reduc_uplus_v4sf): Likewise.
18786 (reduc_splus_v4sf): Likewise.
18787 (aarch64_addv<mode>): Likewise.
18788 (reduc_uplus_<mode>): Likewise.
18789 (reduc_splus_<mode>): Likewise.
18790 (aarch64_addvv2di): Likewise.
18791 (reduc_uplus_v2di): Likewise.
18792 (reduc_splus_v2di): Likewise.
18793 (aarch64_addvv2si): Likewise.
18794 (reduc_uplus_v2si): Likewise.
18795 (reduc_splus_v2si): Likewise.
18796 (reduc_<sur>plus_<mode>): New.
18797 (reduc_<sur>plus_v2di): Likewise.
18798 (reduc_<sur>plus_v2si): Likewise.
18799 (reduc_<sur>plus_v4sf): Likewise.
18800 (aarch64_addpv4sf): Likewise.
18801 * config/aarch64/arm_neon.h
18802 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18803 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18804 add UNSPEC_SADDV, UNSPEC_UADDV.
18806 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18808 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18810 * config/aarch64/arm_neon.h
18811 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18813 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18815 * config/aarch64/aarch64-builtins
18816 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18818 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18820 * config/aarch64/aarch64-simd-builtins.def
18821 (reduc_smax_): New.
18822 (reduc_smin_): Likewise.
18823 (reduc_umax_): Likewise.
18824 (reduc_umin_): Likewise.
18825 (reduc_smax_nan_): Likewise.
18826 (reduc_smin_nan_): Likewise.
18829 (smax): Update for V2SF, V4SF and V2DF modes.
18832 (smin_nan): Likewise.
18833 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18834 (<su><maxmin><mode>3): ...This, refactor.
18835 (s<maxmin><mode>3): New.
18836 (<maxmin_uns><mode>3): Likewise.
18837 (reduc_<maxmin_uns>_<mode>): Refactor.
18838 (reduc_<maxmin_uns>_v4sf): Likewise.
18839 (reduc_<maxmin_uns>_v2si): Likewise.
18840 (aarch64_<fmaxmin><mode>: Remove.
18841 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18843 (vmin<q>_f<32,64>): Likewise.
18844 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18846 (su): Add mappings for smax, smin, umax, umin.
18848 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18849 (FMAXMIN): Rename as...
18850 (FMAXMIN_UNS): ...This.
18852 (fmaxminv): Likewise.
18853 (fmaxmin): Likewise.
18855 (maxmin_uns_op): Likewise.
18857 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18859 * config/aarch64/arm_neon.h
18860 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18861 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18862 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18864 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18866 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18867 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18869 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18871 * config/aarch64/aarch64-simd.md
18872 (vcond<mode>_internal): Handle special cases for constant masks.
18873 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18874 (vcondu<mode><mode>): Likewise.
18875 (vcond<v_cmp_result><mode>): New.
18877 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18879 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18880 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18881 * config/aarch64/aarch64-simd-builtins.def
18882 (cmeq): Update to BUILTIN_VALLDI.
18887 * config/aarch64/arm_neon.h
18888 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18889 to builtins or C as appropriate.
18891 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18893 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18895 (cmhi): Rename to...
18897 * config/aarch64/aarch64-simd.md
18898 (simd_mode): Add SF.
18899 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18900 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18901 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18902 (cstore<mode>_neg): ...This.
18903 * config/aarch64/iterators.md
18905 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18906 (COMPARISONS): New.
18907 (UCOMPARISONS): Likewise.
18908 (optab): Add missing comparisons.
18914 (VCMP_S): Likewise.
18915 (VCMP_U): Likewise.
18916 (V_cmp_result): Add DF, SF modes.
18917 (v_cmp_result): Likewise.
18919 (vmtype): Likewise.
18920 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18922 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18924 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18925 define_insn to define_insn_and_split.
18926 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18927 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18928 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18929 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18930 (thumb2_negscc): Likewise.
18932 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18934 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18936 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18938 * config/arm/thumb2.md: Remove trailing whitespaces.
18940 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18942 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18943 Use gen_int_mode rather than GEN_INT.
18945 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18947 * value-prof.c (stream_in_histogram_value): Remove the strayed
18950 2013-04-30 Richard Biener <rguenther@suse.de>
18952 PR middle-end/57122
18953 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18955 2013-04-30 Richard Biener <rguenther@suse.de>
18957 PR middle-end/57107
18958 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
18960 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
18962 PR rtl-optimization/56957
18963 PR rtl-optimization/57105
18964 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
18965 variable. Use just INSN_UID for determining whether an insn
18966 should be only disconnected from the insn stream.
18967 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
18969 2013-04-30 Jakub Jelinek <jakub@redhat.com>
18971 PR tree-optimization/57104
18972 * tsan.c (instrument_expr): Don't instrument accesses to
18973 DECL_HARD_REGISTER VAR_DECLs.
18975 2013-04-30 Richard Biener <rguenther@suse.de>
18977 * function.h (loops_for_fn): New inline function.
18978 (set_loops_for_fn): Likewise.
18979 * cfgloop.h (place_new_loop): Add struct function parameter.
18980 (get_loop): Likewise.
18981 (get_loops): Likewise.
18982 (number_of_loops): Likewise.
18983 (fel_next): Adjust.
18984 (fel_init): Likewise.
18985 * cfg.c (get_loop_copy): Adjust.
18986 * cfgloop.c (flow_loops_dump): Likewise.
18987 (record_loop_exits): Likewise.
18988 (verify_loop_structure): Likewise.
18989 * cfgloopanal.c (mark_irreducible_loops): Likewise.
18990 (estimate_reg_pressure_cost): Likewise.
18991 (mark_loop_exit_edges): Likewise.
18992 * cfgloopmanip.c (place_new_loop): Likewise.
18993 (add_loop): Likewise.
18994 (duplicate_loop): Likewise.
18995 * graph.c (draw_cfg_nodes): Likewise.
18996 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
18997 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
18998 (extract_affine_chrec): Likewise.
18999 (build_scop_iteration_domain): Likewise.
19000 * graphite.c (graphite_initialize): Likewise.
19001 * ira-build.c (create_loop_tree_nodes): Likewise.
19002 (more_one_region_p): Likewise.
19003 (rebuild_regno_allocno_maps): Likewise.
19004 (mark_loops_for_removal): Likewise.
19005 (mark_all_loops_for_removal): Likewise.
19006 (remove_unnecessary_regions): Likewise.
19007 (ira_build): Likewise.
19008 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19009 * loop-init.c (fix_loop_structure): Likewise.
19010 (gate_rtl_move_loop_invariants): Likewise.
19011 (gate_rtl_unswitch): Likewise.
19012 (gate_rtl_unroll_and_peel_loops): Likewise.
19013 (rtl_doloop): Likewise.
19014 * lto-streamer-in.c (input_cfg): Likewise.
19015 * lto-streamer-out.c (output_cfg): Likewise.
19016 * modulo-sched.c (sms_schedule): Likewise.
19017 * predict.c (tree_estimate_probability): Likewise.
19018 (tree_estimate_probability_driver): Likewise.
19019 (estimate_loops): Likewise.
19020 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
19021 (move_sese_region_to_fn): Likewise.
19022 (debug_loop_num): Likewise.
19023 * tree-chrec.c (chrec_evaluate): Likewise.
19024 (hide_evolution_in_other_loops_than_loop): Likewise.
19025 (chrec_component_in_loop_num): Likewise.
19026 (reset_evolution_in_loop): Likewise.
19027 (evolution_function_is_invariant_rec_p): Likewise.
19028 * tree-if-conv.c (main_tree_if_conversion): Likewise.
19029 * tree-inline.c (copy_loops): Likewise.
19030 (copy_cfg_body): Likewise.
19031 (tree_function_versioning): Likewise.
19032 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
19033 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
19035 (add_to_evolution_1): Likewise.
19036 (scev_const_prop): Likewise.
19037 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
19038 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
19039 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
19040 (tree_ssa_lim_initialize): Likewise.
19041 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
19042 (verify_loop_closed_ssa): Likewise.
19043 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
19044 (tree_ssa_loop_im): Likewise.
19045 (tree_ssa_loop_unswitch): Likewise.
19046 (tree_vectorize): Likewise.
19047 (check_data_deps): Likewise.
19048 (tree_ssa_loop_ivcanon): Likewise.
19049 (tree_ssa_loop_bounds): Likewise.
19050 (tree_complete_unroll): Likewise.
19051 (tree_complete_unroll_inner): Likewise.
19052 (tree_parallelize_loops): Likewise.
19053 (tree_ssa_loop_prefetch): Likewise.
19054 (tree_ssa_loop_ivopts): Likewise.
19055 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
19056 * tree-vectorizer.c (vectorize_loops): Likewise.
19058 2013-04-29 Mike Frysinger <vapier@gentoo.org>
19060 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
19061 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
19062 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
19063 with EABI_LINK_SPEC.
19065 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19068 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
19071 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
19074 * lra-constraints.c (process_alt_operands): Discourage a bit more
19075 using memory for pseudos. Print cost dump for alternatives.
19076 Modify cost values for conflicts with early clobbers.
19077 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
19079 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19082 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
19084 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19086 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
19087 from/to S register.
19088 (movdi_aarch64): Support LDR/STR from/to D register.
19090 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19092 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
19093 or higher by default.
19095 2013-04-29 Richard Biener <rguenther@suse.de>
19097 PR middle-end/57075
19098 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
19099 even if not adding abnormal edges for calls that can make
19102 2013-04-29 Richard Biener <rguenther@suse.de>
19104 PR middle-end/57103
19105 * tree-cfg.c (move_stmt_op): Fix condition under which to update
19107 (move_stmt_r): Remove redundant checking.
19109 2013-04-29 Teresa Johnson <tejohnson@google.com>
19112 * basic-block.h (apply_scale): New function.
19113 (apply_probability): Use apply_scale.
19114 * gimple-streamer-in.c (input_bb): Ditto.
19115 * lto-streamer-in.c (input_cfg): Ditto.
19116 * lto-cgraph.c (merge_profile_summaries): Ditto.
19117 * tree-optimize.c (execute_fixup_cfg): Ditto.
19118 * tree-inline.c (copy_bb): Update comment to use apply_scale.
19119 (copy_edges_for_bb): Ditto.
19120 (copy_cfg_body): Ditto.
19122 2013-04-29 Tom de Vries <tom@codesourcery.com>
19124 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
19125 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
19126 (tail_merge_optimize): Handle current_loops == NULL.
19128 2013-04-26 Jeff Law <law@redhat.com>
19130 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
19131 (simplify_cond_using_ranges): Generalize code to simplify
19132 COND_EXPRs where one argument is a constant and the other
19133 is an SSA_NAME created by an integral type conversion.
19135 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19137 * config/arm/arm.md (store_minmaxsi): Use only when
19138 optimize_insn_for_size_p.
19140 2013-04-29 Christian Bruel <christian.bruel@st.com>
19143 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19145 2013-04-29 Richard Biener <rguenther@suse.de>
19147 PR middle-end/57089
19148 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19149 loop tree make sure to schedule a fixup for the child as well.
19150 (expand_omp_for_generic): Properly add loops.
19151 (expand_omp_for_static_nochunk): Likewise.
19152 (expand_omp_for_static_chunk): Likewise.
19153 (expand_omp_for): For the degenerate case fixup loops.
19154 (expand_omp_sections): Fix default bb placement in loops.
19155 (expand_omp_atomic_pipeline): Properly add loops.
19157 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19159 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19161 2013-04-29 Tom de Vries <tom@codesourcery.com>
19163 * tree-ssa-tail-merge.c: Update header comment.
19165 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19167 * config/aarch64/arm_neon.h
19168 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19169 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19170 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19171 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19172 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19173 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19174 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19175 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19177 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19179 * config/aarch64/aarch64-simd.md
19180 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19181 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19182 fix_trunc, fixuns_trunc.
19183 (ftrunc<VDQF:mode>2): New.
19184 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19185 (fix_trunc_optab): New.
19187 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19189 * config/aarch64/aarch64-builtins.c
19190 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19191 iceilf, lround, iroundf.
19193 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19196 * config/i386/i386.h (enum ix86_tune_indices)
19197 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19198 New, split from X86_TUNE_INTER_UNIT_MOVES.
19199 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19200 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19201 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19202 (TARGET_INTER_UNIT_MOVES): Remove.
19203 * config/i386/i386.c (initial_ix86_tune_features): Update.
19204 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19205 (ix86_expand_convert_uns_didf_sse): Use
19206 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19207 (ix86_expand_vector_init_one_nonzero): Ditto.
19208 (ix86_expand_vector_init_interleave): Ditto.
19209 (inline_secondary_memory_needed): Return true for moves from SSE class
19210 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19211 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19212 * config/i386/constraints.md (Yi, Ym): Depend on
19213 TARGET_INTER_UNIT_MOVES_TO_VEC.
19214 (Yj, Yn): New constraints.
19215 * config/i386/i386.md (*movdi_internal): Change constraints of
19216 operand 1 from Yi to Yj and from Ym to Yn.
19217 (*movsi_internal): Ditto.
19218 (*movdf_internal): Ditto.
19219 (*movsf_internal): Ditto.
19220 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19221 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19222 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19223 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19224 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19225 * config/i386/sse.md (movdi_to_sse): Ditto.
19226 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19227 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19228 TARGET_INTER_UNIT_MOVES.
19229 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19230 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19231 instead of TARGET_INTER_UNIT_MOVES.
19232 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19233 operand 1 from Yi to Yj and from Ym to Yn.
19235 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19237 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19238 (float_truncate_hi_): Likewise.
19239 (float_extend_lo_): Likewise.
19240 (float_truncate_lo_): Likewise.
19241 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19242 (aarch64_float_extend_lo_v2df): Likewise.
19243 (vec_unpacks_hi_v4sf): Likewise.
19244 (aarch64_float_truncate_lo_v2sf): Likewise.
19245 (aarch64_float_truncate_hi_v4sf): Likewise.
19246 (vec_pack_trunc_v2df): Likewise.
19247 (vec_pack_trunc_df): Likewise.
19249 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19251 * config/aarch64/aarch64-builtins.c
19252 (aarch64_fold_builtin): Fold float conversions.
19253 * config/aarch64/aarch64-simd-builtins.def
19254 (floatv2si, floatv4si, floatv2di): New.
19255 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19256 * config/aarch64/aarch64-simd.md
19257 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19258 * config/aarch64/iterators.md (FLOATUORS): New.
19259 (optab): Add float, floatuns.
19260 (su_optab): Likewise.
19262 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19264 * config/aarch64/aarch64-builtins.c
19265 (aarch64_builtin_vectorized_function): Use new names for
19267 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19268 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19269 (fcvtzu): Split as...
19270 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19271 (fcvtas): Split as...
19272 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19273 (fcvtau): Split as...
19274 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19275 (fcvtps): Split as...
19276 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19277 (fcvtpu): Split as...
19278 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19279 (fcvtms): Split as...
19280 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19281 (fcvtmu): Split as...
19282 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19283 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19284 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19285 (lfrintnusf, lfrintnudf): Likewise.
19286 * config/aarch64/aarch64-simd.md
19287 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19289 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19290 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19291 (fcvt_pattern): Likewise.
19293 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19295 * config/aarch64/aarch64-simd.md
19296 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19297 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19299 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19301 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19302 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19303 (vrnd<a,m,n,p>_f32): Implement using builtins.
19304 (vrnd<i,x><q>_f<32, 64>): New.
19306 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19308 * config/aarch64/aarch64-builtins.c
19309 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19310 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19311 (frintz): Rename to...
19313 (frintp): Rename to...
19315 (frintm): Rename to...
19317 (frinti): Rename to...
19318 (nearbyint): ...this.
19319 (frintx): Rename to...
19321 (frinta): Rename to...
19323 * config/aarch64/aarch64-simd.md
19324 (aarch64_frint<frint_suffix><mode>): Delete.
19325 (<frint_pattern><mode>2): Convert to insn.
19326 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19327 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19328 (frint_pattern): Likewise.
19329 (frint_suffix): Likewise.
19331 2013-04-29 Richard Biener <rguenther@suse.de>
19333 PR tree-optimization/57081
19334 * loop-init.c: Include tree-flow.h.
19335 (loop_optimizer_finalize): Free number of iteration estimates.
19336 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19338 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19340 PR tree-optimization/57083
19341 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19342 non-singleton shift count range, zero extend low_bound for uns case.
19344 * config/i386/predicates.md (general_vector_operand): New predicate.
19345 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19346 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19347 if they aren't nonimmediate operands. If their original values
19348 satisfy const_vector_equal_evenodd_p, don't shift them.
19349 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19350 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19351 if not nonimmediate_operand.
19352 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19353 instead of register_operand.
19354 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19356 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19358 * stor-layout.c (finalize_size_functions): Allocate a structure and
19359 reset cfun before dumping the functions.
19361 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19363 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19366 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19367 use xop_pmacsdqh if uns_p.
19368 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19369 the immediate rotate count.
19371 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19373 * rtl.h (struct rtx_def): Add comment for field jump.
19374 (LRA_SUBREG_P): New macro.
19375 * recog.c (register_operand): Check LRA_SUBREG_P.
19376 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19378 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19379 (lra_final_code_change): Skip subreg change for operators.
19380 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19381 if there are no operand changes.
19382 * lra-constraints.c (curr_insn_set): New.
19383 (match_reload): Set LRA_SUBREG_P.
19384 (emit_spill_move): Ditto.
19385 (check_and_process_move): Use curr_insn_set. Process only single
19386 set insns. Don't initialize sec_mem_p and change_p.
19387 (simplify_operand_subreg): Use LRA_SUBREG_P.
19388 (reg_in_class_p): New function.
19389 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19390 of #ifdef. Add code to remove cycling.
19391 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19392 non-null disp. Reload inner instead of disp when base and index
19393 are null. Try to put lo_sum into register.
19394 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19395 (check_and_process_move): Move code for move cost check to
19396 simple_move_p. Remove equiv_substitution.
19397 (simple_move_p): New function.
19398 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19399 curr_insn_set. Call check_and_process_move only for single set
19400 insns. Use the new function. Move call of check_and_process_move
19401 after operand equiv substitution and address process.
19403 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19406 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19407 with nonlocal goto receivers or returns twice calls, ignore
19408 unininitialized values from abnormal edges to nl goto receiver
19409 or returns twice call.
19411 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19413 PR tree-optimization/57051
19414 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19415 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19418 2013-04-26 Richard Biener <rguenther@suse.de>
19420 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19421 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19422 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19423 (fixup_loop_arrays_after_move): New function.
19424 (move_sese_region_to_fn): Properly outline the loop tree parts
19425 of the SESE region.
19427 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19429 * config/i386/i386.md (type, unit): Fix long lines.
19431 2013-04-26 Richard Biener <rguenther@suse.de>
19433 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19434 (lto-streamer-out.o): Likewise.
19435 * cfgloop.c (init_loops_structure): Export, add struct function
19436 argument and adjust.
19437 (flow_loops_find): Adjust.
19438 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19439 (init_loops_structure): Declare.
19440 * lto-streamer-in.c: Include cfgloop.h.
19441 (input_cfg): Input the loop tree.
19442 * lto-streamer-out.c: Include cfgloop.h.
19443 (output_cfg): Output the loop tree.
19444 (output_struct_function_base): Do not drop PROP_loops.
19446 2013-03-26 Richard Biener <rguenther@suse.de>
19448 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19449 (pass_build_cfg): Provide PROP_loops.
19450 (move_sese_region_to_fn): Remove loops that are outlined into fn
19452 * tree-inline.c: Include cfgloop.h.
19453 (initialize_cfun): Do not drop PROP_loops.
19454 (copy_loops): New function.
19455 (copy_cfg_body): Copy loop structure.
19456 (tree_function_versioning): Initialize destination loop tree.
19457 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19458 (pass_parallelize_loops): Do IL verification.
19459 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19460 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19461 the CFG make sure we fixup loops as well.
19462 * tree-ssa-tail-merge.c: Include cfgloop.h.
19463 (replace_block_by): When merging loop latches mark loops for fixup.
19464 * lto-streamer-out.c (output_struct_function_base): Drop
19465 PROP_loops for now.
19466 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19467 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19468 * ipa-split.c: Include cfgloop.h.
19469 (split_function): Add the new return block to the loop tree root.
19470 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19471 whether we have removed the forwarder block.
19472 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19473 * cfgloop.h (place_new_loop): Declare.
19474 * cfgloopmanip.c (place_new_loop): Export.
19475 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19476 (tree-switch-conversion.o): Likewise.
19477 (tree-complex.o): Likewise.
19478 (tree-inline.o): Likewise.
19479 (tree-ssa-tailmerge.o): Likewise.
19480 (ipa-split.o): Likewise.
19481 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19482 (tree-ssa-copy.o): Likewise.
19483 * tree-switch-conversion.c: Include cfgloop.h
19484 (process_switch): If we emit a bit-test cascade, schedule loops
19486 * tree-complex.c: Include cfgloop.h.
19487 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19488 * asan.c: Include cfgloop.h.
19489 (create_cond_insert_point): Properly add new basic-blocks to
19490 loops, schedule loop fixup.
19491 * cfgloop.c (verify_loop_structure): Check that looks are not
19493 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19495 (expand_omp_for_generic): Likewise.
19496 (expand_omp_sections): Likewise.
19497 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19498 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19499 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19500 is initialized, not when loops are present.
19501 * tree-parloops.c (parallelize_loops): Remove checking here.
19502 * passes.c (init_optimization_passes): Schedule a copy-propagation
19503 pass before complete unrolling of inner loops.
19505 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19507 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19508 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19509 (colorize_init): Add argument to _WIN32 version.
19510 * toplev.c: Include diagnostic-color.h.
19511 (process_options): Default to -fdiagnostics-color=auto if
19512 GCC_COLORS env var is in the environment.
19513 * common.opt (fdiagnostics-color=): Add Var and Init.
19514 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19515 env var is in the environment, the default is auto rather than never.
19517 * diagnostic.h (file_name_as_prefix): Add context argument.
19518 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19519 the string as locus.
19520 * langhooks.c (lhd_print_error_function): Adjust caller.
19522 2013-04-25 Lawrence Crowl <crowl@google.com>
19524 * var-tracking.c (shared_hash_def::htab):
19525 Change type to hash_table. Update dependent calls and types.
19527 2013-04-25 Lawrence Crowl <crowl@google.com>
19529 * Makefile.in: Update as needed below.
19531 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19532 Move declaration to after the type's method definitons.
19534 * attribs.c (htab_t scoped_attributes::attribute_hash):
19535 Change type to hash_table. Update dependent calls and types.
19537 * bitmap.c (htab_t bitmap_desc_hash):
19538 Change type to hash_table. Update dependent calls and types.
19540 * cselib.c (htab_t cselib_hash_table):
19541 Change type to hash_table. Update dependent calls and types.
19543 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19544 (hash_string_slot_node): Move implementation into lto-streamer.h
19545 struct string_slot_hasher.
19546 (eq_string_slot_node): Likewise.
19548 * data-streamer-out.c: Update output_block::string_hash_table
19549 dependent calls and types.
19551 * dwarf2cfi.c (htab_t trace_index):
19552 Change type to hash_table. Update dependent calls and types.
19554 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19555 Change type to hash_table. Update dependent calls and types.
19556 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19557 (htab_t optimize_external_refs::map): Likewise.
19558 (htab_t output_comp_unit::extern_map): Likewise.
19559 (htab_t output_comdat_type_unit::extern_map): Likewise.
19560 (htab_t output_macinfo::macinfo_htab): Likewise.
19561 (htab_t optimize_location_lists::htab): Likewise.
19562 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19564 * except.c (htab_t ehspec_hash_type):
19565 Change type to hash_table. Update dependent calls and types.
19566 (assign_filter_values::ttypes): Likewise.
19567 (assign_filter_values::ehspec): Likewise.
19568 (sjlj_assign_call_site_values::ar_hash): Likewise.
19569 (convert_to_eh_region_ranges::ar_hash): Likewise.
19571 * gcse.c (htab_t pre_ldst_table):
19572 Change type to hash_table. Update dependent calls and types.
19574 * ggc-common.c (htab_t saving_htab):
19575 Change type to hash_table. Update dependent calls and types.
19576 (htab_t loc_hash): Likewise.
19577 (htab_t ptr_hash): Likewise.
19578 (call_count): Rename ggc_call_count.
19579 (call_alloc): Rename ggc_call_alloc.
19580 (loc_descriptor): Rename make_loc_descriptor.
19581 (add_statistics): Rename ggc_add_statistics.
19583 * ggc-common.c (saving_htab):
19584 Change type to hash_table. Update dependent calls and types.
19586 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19587 (push_gimplify_context): Likewise.
19588 (pop_gimplify_context): Likewise.
19589 (struct gimple_temp_hash_elt): Added.
19590 (struct gimplify_hasher): Likewise.
19591 (struct gimplify_ctx.temp_htab):
19592 Change type to hash_table. Update dependent calls and types.
19594 * gimple-fold.c: Include gimplify-ctx.h.
19596 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19597 Change type to hash_table. Update dependent calls and types.
19598 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19599 avoid potential global name collision.
19601 * gimplify.c: Include gimplify-ctx.h.
19602 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19603 (htab_t gimplify_ctx::temp_htab):
19604 Update dependent calls and types for new type hash_table.
19605 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19606 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19608 * gimplify-ctx.h: New.
19609 (struct gimple_temp_hash_elt): Move from gimplify.c.
19610 (class gimplify_hasher): New.
19611 (struct gimplify_ctx): Move from gimple.h.
19612 (htab_t gimplify_ctx::temp_htab):
19613 Change type to hash_table. Update dependent calls and types.
19615 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19616 (htab_t ivs_params::newivs_index):
19617 Change type to hash_table. Update dependent calls and types.
19618 (htab_t ivs_params::params_index): Likewise.
19619 (htab_t print_generated_program::params_index): Likewise.
19620 (htab_t gloog::newivs_index): Likewise.
19621 (htab_t gloog::params_index): Likewise.
19623 * graphite.c: Include graphite-htab.h.
19624 4htab_t graphite_transform_loops::bb_pbb_mapping):
19625 Change type to hash_table. Update dependent calls and types.
19627 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19628 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19629 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19631 * graphite-dependences.c: Include graphite-htab.h.
19632 (loop_is_parallel_p): Change hash table type of parameter.
19634 * graphite-htab.h: New.
19635 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19636 (extern find_pbb_via_hash): Move from graphite-poly.h.
19637 (extern loop_is_parallel_p): Move from graphite-poly.h.
19638 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19640 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19641 (extern loop_is_parallel_p): Move to graphite-htab.h.
19642 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19644 * haifa-sched.c (htab_t delay_htab):
19645 Change type to hash_table. Update dependent calls and types.
19646 (htab_t delay_htab_i2): Likewise.
19648 * ira-color.c (htab_t allocno_hard_regs_htab):
19649 Change type to hash_table. Update dependent calls and types.
19651 * ira-costs.c (htab_t cost_classes_htab):
19652 Change type to hash_table. Update dependent calls and types.
19654 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19655 Change type to hash_table. Update dependent calls and types.
19657 * loop-iv.c (htab_t bivs):
19658 Change type to hash_table. Update dependent calls and types.
19660 * loop-unroll.c (htab_t opt_info::insns_to_split):
19661 Change type to hash_table. Update dependent calls and types.
19662 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19664 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19665 (struct string_slot_hasher): New.
19666 (htab_t output_block::string_hash_table):
19667 Change type to hash_table. Update dependent calls and types.
19669 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19670 (htab_t file_name_hash_table):
19671 Change type to hash_table. Update dependent calls and types.
19673 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19676 * lto-streamer.c (htab_t tree_htab):
19677 Change type to hash_table. Update dependent calls and types.
19679 * omp-low.c: Include gimplify-ctx.h.
19681 * passes.c (htab_t name_to_pass_map):
19682 Change type to hash_table. Update dependent calls and types.
19683 (pass_traverse): Rename to passes_pass_traverse.
19685 * plugin.c (htab_t event_tab):
19686 Change type to hash_table. Update dependent calls and types.
19688 * postreload-gcse.c (htab_t expr_table):
19689 Change type to hash_table. Update dependent calls and types.
19690 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19692 * sese.c (debug_rename_map_1): Make extern.
19693 (htab_t copy_bb_and_scalar_dependences::rename_map):
19694 Change type to hash_table. Update dependent calls and types.
19696 * sese.h (extern debug_rename_map): Move to .c file.
19698 * store-motion.c (htab_t store_motion_mems_table):
19699 Change type to hash_table. Update dependent calls and types.
19701 * trans-mem.c (htab_t tm_new_mem_hash):
19702 Change type to hash_table. Update dependent calls and types.
19704 * tree-browser.c (htab_t TB_up_ht):
19705 Change type to hash_table. Update dependent calls and types.
19707 * tree-cfg.c (htab_t discriminator_per_locus):
19708 Change type to hash_table. Update dependent calls and types.
19710 * tree-complex.c: Include tree-hasher.h
19711 (htab_t complex_variable_components):
19712 Change type to hash_table. Update dependent calls and types.
19714 * tree-eh.c (htab_t finally_tree):
19715 Change type to hash_table. Update dependent calls and types.
19717 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19718 struct int_tree_hasher.
19719 (extern int_tree_map_eq): Likewise.
19720 (uid_decl_map_hash): Removed.
19721 (extern decl_tree_map_eq): Likewise.
19723 * tree-hasher.h: New.
19724 (struct int_tree_hasher): New.
19725 (typedef int_tree_htab_type): New.
19727 * tree-inline.c: Include gimplify-ctx.h.
19729 * tree-mudflap.c: Include gimplify-ctx.h.
19731 * tree-parloops.c: Include tree-hasher.h.
19732 (htab_t eliminate_local_variables_stmt::decl_address):
19733 Change type to hash_table. Update dependent calls and types.
19734 (htab_t separate_decls_in_region::decl_copies): Likewise.
19736 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19737 Change type to hash_table. Update dependent calls and types.
19739 * tree-sra.c (candidates):
19740 Change type to hash_table. Update dependent calls and types.
19742 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19744 (int_tree_map_hash): Likewise.
19746 * tree-ssa-dom.c (htab_t avail_exprs):
19747 Change type to hash_table. Update dependent calls and types.
19749 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19750 Change type to hash_table. Update dependent calls and types.
19752 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19753 Change type to hash_table. Update dependent calls and types.
19755 * tree-ssa-phiopt.c (seen_ssa_names):
19756 Change type to hash_table. Update dependent calls and types.
19758 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19759 Change type to hash_table. Update dependent calls and types.
19761 * tree-ssa-uncprop.c (equiv):
19762 Change type to hash_table. Update dependent calls and types.
19764 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19766 PR rtl-optimization/57003
19767 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19768 call note_stores with kill_clobbered_value callback again after
19769 killing regs_invalidated_by_call.
19771 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19773 * config/aarch64/aarch64-simd.md
19774 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19775 (aarch64_simd_bsl<mode>): Likewise.
19776 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19778 2013-04-25 Marek Polacek <polacek@redhat.com>
19780 PR tree-optimization/57066
19781 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19783 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19785 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19787 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19789 * config/aarch64/aarch64-builtins.c
19790 (aarch64_fold_builtin): New.
19791 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19792 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19793 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19794 * config/aarch64/arm_neon.h
19795 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19797 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19798 Tejas Belagod <tejas.belagod@arm.com>
19800 * config/aarch64/aarch64-builtins.c
19801 (aarch64_gimple_fold_builtin): New.
19802 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19803 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19804 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19805 (addvv4sf): Update.
19806 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19808 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19810 * config/aarch64/aarch64.md
19811 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19813 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19815 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19816 (*ngcsi_uxtw): New pattern.
19818 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19819 Julian Brown <julian@codesourcery.com>
19821 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19822 (TB_DREG): Add T_V4HF.
19823 (v4hf_UP): New macro.
19824 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19825 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19826 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19828 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19829 (arm_vector_mode_supported_p): Handle V4HF.
19830 (arm_mangle_map): Handle V4HFmode.
19831 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19832 * config/arm/arm_neon_builtins.def: Add entries for
19833 vcvtv4hfv4sf, vcvtv4sfv4hf.
19834 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19835 (neon_vcvtv4hfv4sf): Likewise.
19836 * config/arm/neon-gen.ml: Handle half-precision floating point
19838 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19839 * config/arm/arm_neon.h: Regenerate.
19840 * config/arm/neon.ml (type elts): Add F16.
19841 (type vectype): Add T_float16x4, T_floatHF.
19842 (type vecmode): Add V4HF.
19843 (type features): Add Requires_FP_bit feature.
19844 (elt_width): Handle F16.
19845 (elt_class): Likewise.
19846 (elt_of_class_width): Likewise.
19847 (mode_of_elt): Refactor.
19848 (type_for_elt): Handle F16, fix error messages.
19849 (vectype_size): Handle T_float16x4.
19850 (vcvt_sh): New function.
19851 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19852 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19853 (string_of_mode): Handle V4HF.
19854 * doc/arm-neon-intrinsics.texi: Regenerate.
19856 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19858 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19859 format specifier in 'X' case.
19861 2013-04-25 Alan Modra <amodra@gmail.com>
19864 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19865 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19866 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19867 Repeat for many other rotate/shift and mask patterns using subregs.
19868 Name lshiftrt insns.
19869 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19870 on WORDS_BIG_ENDIAN.
19872 2013-04-25 Alan Modra <amodra@gmail.com>
19874 * config.gcc: Support little-endian powerpc-linux targets.
19875 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19876 (LINK_OS_LINUX_SPEC): Define.
19877 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19878 Preserve MASK_LITTLE_ENDIAN.
19879 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19880 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19881 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19882 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19883 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19884 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19885 Correct fp word order for little-endian. Don't shift toc entries
19886 smaller than a word for little-endian.
19887 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19888 (bswapdi2 splits): Correct low-part subreg for little-endian.
19889 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19890 low/high where such is correct only for be.
19891 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19892 little-endian for -mcall-aixdesc.
19894 2013-04-25 Alan Modra <amodra@gmail.com>
19896 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19897 replace_equiv_address_nv.
19899 2013-04-25 Alan Modra <amodra@gmail.com>
19901 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19903 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19906 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19907 * rtl.h (struct rtx_def): ...
19909 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19911 PR rtl-optimizations/57046
19912 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19913 for multi-reg splits.
19915 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19917 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19919 2013-04-24 Sterling Augustine <saugustine@google.com>
19921 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19922 (comp_dir_string, debug_str_dwo_section): New.
19923 (DEBUG_STR_DWO_SECTION): Rename to ...
19924 (DEBUG_DWO_STR_SECTION): ... this.
19925 (DEBUG_NORM_STR_SECTION): Delete.
19926 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19927 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19928 (find_AT_string): Move most logic to ...
19929 (find_AT_string_in_table): ... here. New.
19930 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19931 add_skeleton_AT_string. Delete logic.
19932 (output_skeleton_debug_sections): Remove call to
19933 add_top_level_skeleton_die_attrs.
19934 (add_comp_dir_attribute): Move logic to comp_dir_string.
19935 (dwarf2out_init): Initialize debug_str_dwo_section.
19936 (output_indirect_string): Call find_string_form.
19937 (output_indirect_strings): Rewrite.
19938 (prune_unused_types): Empty skeleton_debug_str_hash.
19939 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19940 (dwarf2out_finish): Call output_indirect_strings.
19942 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19944 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19946 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19948 * rtl.h (struct rtx_def): Add comment for field jump.
19949 (LRA_SUBREG_P): New macro.
19950 * recog.c (register_operand): Check LRA_SUBREG_P.
19951 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19953 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19954 (lra_final_code_change): Skip subreg change for operators.
19955 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19956 if there are no operand changes.
19957 * lra-constraints.c (curr_insn_set): New.
19958 (match_reload): Set LRA_SUBREG_P.
19959 (emit_spill_move): Ditto.
19960 (check_and_process_move): Use curr_insn_set. Process only single
19961 set insns. Don't initialize sec_mem_p and change_p.
19962 (simplify_operand_subreg): Use LRA_SUBREG_P.
19963 (reg_in_class_p): New function.
19964 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19965 of #ifdef. Add code to remove cycling.
19966 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19967 non-null disp. Reload inner instead of disp when base and index
19968 are null. Try to put lo_sum into register.
19969 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19970 (check_and_process_move): Move code for move cost check to
19971 simple_move_p. Remove equiv_substitution.
19972 (simple_move_p): New function.
19973 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19974 curr_insn_set. Call check_and_process_move only for single set
19975 insns. Use the new function. Move call of check_and_process_move
19976 after operand equiv substitution and address process.
19978 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
19980 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
19981 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
19982 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
19984 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19986 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
19988 2013-04-24 Marek Polacek <polacek@redhat.com>
19990 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
19991 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
19992 (select_loops_exit_conditions): Likewise.
19993 (number_of_iterations_for_all_loops): Likewise.
19994 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
19995 (scev_analysis): Likewise.
19997 2013-04-02 Catherine Moore <clm@codesourcery.com>
19998 Chao-ying Fu <fu@mips.com>
20000 * config/mips/micromips.md (jraddiusp): New pattern.
20001 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
20002 instruction if possible.
20004 2013-04-24 Alan Modra <amodra@gmail.com>
20006 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
20008 2013-04-24 Julian Brown <julian@codesourcery.com>
20009 Chung-Lin Tang <cltang@codesourcery.com>
20011 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
20012 dependency behavior in enumeration type DIE generation. Add TODO note
20013 to comments about future DW_FORM_sdata/udata re-work of related code.
20015 2013-04-23 Lawrence Crowl <crowl@google.com>
20017 * Makefile.in: Update as needed below.
20019 * hash-table.h (class hash_table):
20020 Correct many methods with parameter types compare_type to the correct
20021 value_type. (Correct code was unlikely to notice the change.)
20022 (hash_table::elements_with_deleted) New.
20023 (class hashtable::iterator): New.
20024 (hashtable::begin()): New.
20025 (hashtable::end()): New.
20026 (FOR_EACH_HASH_TABLE_ELEMENT): New.
20028 * statistics.c (statistics_hashes):
20029 Change type to hash_table. Update dependent calls and types.
20031 * tree-into-ssa.c (var_infos):
20032 Change type to hash_table. Update dependent calls and types.
20034 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
20035 Change type to hash_table. Update dependent calls and types.
20037 * tree-ssa-loop-im.c (struct mem_ref.refs):
20038 Change type to hash_table. Update dependent calls and types.
20040 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
20041 Change type to hash_table. Update dependent calls and types.
20043 * tree-ssa-sccvn.c (vn_tables_s::nary):
20044 Change type to hash_table. Update dependent calls and types.
20045 (vn_tables_s::phis): Likewise.
20046 (vn_tables_s::references): Likewise.
20048 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
20049 (vn_reference_eq): Update parameter and return types.
20051 * tree-ssa-structalias.c (pointer_equiv_class_table):
20052 Change type to hash_table. Update dependent calls and types.
20053 (location_equiv_class_table): Likewise.
20055 * tree-vect-data-refs.c: Consequential changes for making
20056 peeling a hash_table.
20058 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
20059 (destroy_loop_vec_info): Dependent hash_table update.
20061 * tree-vectorizer.h (peeling_htab):
20062 Change type to hash_table. Update dependent calls and types.
20064 2013-04-23 Shiva Chen <shiva0217@gmail.com>
20066 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
20067 to check the register content is equal or not.
20068 * lra-constraints.c (match_reload): Use lra_assign_reg_val
20069 to assign register content record.
20070 * lra-eliminations.c (update_reg_eliminate): Use
20071 lra_update_reg_val_offset to update register content offset.
20072 * lra-int.h (struct lra_reg): Add offset member.
20073 (lra_reg_val_equal_p): New static inline function.
20074 (lra_update_reg_val_offset): New static inline function.
20075 (lra_assign_reg_val): New static inline function.
20076 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
20077 to assign register content record.
20078 (initialize_lra_reg_info_element): Initial offset to zero.
20080 2013-04-23 Catherine Moore <clm@codesourcery.com>
20082 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
20083 operands. Record compression.
20085 2013-04-23 Xinliang David Li <davidxl@google.com>
20087 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
20089 2013-04-23 Richard Biener <rguenther@suse.de>
20091 PR middle-end/57036
20092 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
20093 parameter, only add abnormal goto edges from the copied body
20094 if the call could perform abnormal gotos.
20095 (copy_cfg_body): Adjust.
20097 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
20099 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
20101 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
20103 * coretypes.h (gimple_stmt_iterator): Add struct to make
20106 2013-04-23 Richard Biener <rguenther@suse.de>
20108 PR tree-optimization/57026
20109 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
20110 from SSA names occuring in abnormal PHI nodes.
20112 2013-04-22 Andi Kleen <ak@linux.intel.com>
20114 * lto/lto.c (print_lto_report_1): Fix LTO report names.
20116 2013-04-22 Andi Kleen <ak@linux.intel.com>
20118 * lto/lto.c (print_lto_report_1): Declare early.
20119 (read_cgraph_and_symbols): Call print_lto_report_1 early.
20121 2013-04-22 Andi Kleen <ak@linux.intel.com>
20123 * common.opt (-flto-report-wpa): Add.
20124 * doc/invoke.texi (-flto-report-wpa): Add.
20125 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
20128 2013-04-22 Xinliang David Li <davidxl@google.com>
20130 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
20131 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
20132 * Makefile.in: New dependency
20134 David Daney <ddaney.cavm@gmail.com>
20136 * configure.ac (gcc_cv_as_micromips_support): Use the
20137 --fatal-warnings option.
20138 * configure: Regenerate.
20140 2013-04-22 Marek Polacek <polacek@redhat.com>
20143 * tsan.c (instrument_expr): Don't instrument expression
20144 in case its size is zero.
20146 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20150 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20152 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20154 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20156 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20157 (gimple_stmt_iterator): New typedef.
20158 * gimple.h (gimple_stmt_iterator): Rename to...
20159 (gimple_stmt_iterator_d): ... This.
20160 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20161 trees be valid for GIMPLE and GENERIC.
20162 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20163 * gimple-fold.c (gimple_fold_call): Call target hook
20164 gimple_fold_builtin.
20165 * hooks.c (hook_bool_gsiptr_false): New.
20166 * hooks.h (hook_bool_gsiptr_false): New.
20167 * target.def (fold_stmt): New.
20168 * doc/tm.texi: Regenerate.
20170 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20173 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20174 a set sp if no stack realignment.
20176 2013-04-22 Nick Clifton <nickc@redhat.com>
20178 * config.gcc (tilegx-linux): Extend extra_objs rather than
20180 (tilepro-linux): Likewise.
20182 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20184 * config/aarch64/aarch64-builtins.c
20186 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20187 (VAR<1-12>): Add MAP parameter.
20188 (BUILTIN_*): Likewise.
20189 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20190 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20191 (aarch64_ushl_n<mode>): Likewise.
20192 (aarch64_sshr_n<mode>): Likewise.
20193 (aarch64_ushr_n<mode>): Likewise.
20194 (aarch64_<maxmin><mode>): Likewise.
20195 (aarch64_sqrt<mode>): Likewise.
20196 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20197 (vshr<q>_n_*): Likewise.
20199 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20201 * config/aarch64/aarch64-builtins.c
20202 (aarch64_simd_builtin_type_mode): Handle SF types.
20204 (BUILTIN_GPF): Define.
20205 (aarch64_init_simd_builtins): Handle SF types.
20206 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20207 (frecps): Likewise.
20208 (frecpx): Likewise.
20209 * config/aarch64/aarch64-simd.md
20210 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20211 (aarch64_frecpe<mode>): New.
20212 (aarch64_frecps<mode>): Likewise.
20213 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20214 (v8type): Add frecp<esx>.
20215 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20216 (aarch64_frecps<mode>): Likewise.
20217 * config/aarch64/iterators.md (FRECP): New.
20218 (frecp_suffix): Likewise.
20219 * config/aarch64/arm_neon.h
20220 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20222 2013-04-22 Christian Bruel <christian.bruel@st.com>
20225 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20226 (REG_CLASS_NAMES): Idem.
20227 (REG_CLASS_CONTENTS): Idem.
20228 (REGCLASS_HAS_FP_REG): Idem.
20229 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20230 (sh_conditional_register_usage): Idem.
20232 2013-04-21 Jeff Law <law@redhat.com>
20234 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20235 (ssa_forward_propagate_and_combine): Use it.
20237 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20239 * lra.c: Update the flow chart diagram.
20241 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20243 PR rtl-optimization/56847
20244 * lra-constraints.c (process_alt_operands): Discourage alternative
20245 with non-matche doffsettable memory constraint fro memory with
20248 2013-04-19 Richard Biener <rguenther@suse.de>
20250 PR tree-optimization/56982
20251 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20253 * gimplify.c (gimplify_call_expr): Notice special calls.
20254 (gimplify_modify_expr): Likewise.
20255 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20256 abnormal control flow receivers.
20257 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20258 in the same way as cfun->has_nonlocal_labels.
20259 (gimple_purge_dead_abnormal_call_edges): Likewise.
20260 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20261 receivers start a basic-block.
20263 2013-04-19 Richard Biener <rguenther@suse.de>
20265 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20267 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20268 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20269 (SLP_TREE_LOAD_PERMUTATION): Add.
20270 (vect_transform_slp_perm_load): Adjust prototype.
20271 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20272 (vect_free_slp_instance): Likewise.
20273 (vect_create_new_slp_node): Likewise.
20274 (vect_supported_slp_permutation_p): Remove.
20275 (vect_slp_rearrange_stmts): Adjust.
20276 (vect_supported_load_permutation_p): Likewise. Inline
20277 vect_supported_slp_permutation_p here.
20278 (vect_analyze_slp_instance): Compute load permutations per
20279 slp node instead of per instance.
20280 (vect_get_slp_defs): Adjust.
20281 (vect_transform_slp_perm_load): Likewise.
20282 (vect_schedule_slp_instance): Remove redundant code.
20283 (vect_schedule_slp): Remove hack for PR56270, add it ...
20284 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20285 CSE loads for SLP. Adjust.
20287 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20289 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20290 spelling in two comments.
20292 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20295 * config/arm/arm.c (load_multiple_sequence): Require SP
20296 as base register for loads if SP is in the register list.
20298 2013-04-19 Martin Jambor <mjambor@suse.cz>
20300 PR tree-optimization/56718
20301 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20302 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20303 and made public. Adjusted all callers.
20304 (ipa_intraprocedural_devirtualization): New function.
20305 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20306 (ipa_intraprocedural_devirtualization): Likewise.
20307 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20309 2013-04-19 Richard Biener <rguenther@suse.de>
20311 PR tree-optimization/57000
20312 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20314 2013-04-19 Terry Guo <terry.guo@arm.com>
20316 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20318 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20319 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20320 (cortex_m4_fmacs): Use new reservations.
20321 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20323 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20325 PR rtl-optimization/56999
20326 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20328 (lra_coalesce): Remove split_origin_bitmap and related code.
20329 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20330 ranges if necessary.
20332 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20334 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20336 (ix86_expand_call): Remove clobbered_registers array and use
20337 x86_64_ms_sysv_extra_clobbered_registers instead.
20338 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20340 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20342 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20343 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20344 (*call_value_rex64_ms_sysv): Ditto.
20346 2013-04-18 Cary Coutant <ccoutant@google.com>
20348 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20349 parent instead of die_mark.
20351 2013-04-18 Diego Novillo <dnovillo@google.com>
20353 * gimple.c (create_gimple_tmp): New.
20354 (get_expr_type): New.
20355 (build_assign): New.
20356 (build_type_cast): New.
20357 * gimple.h (enum ssa_mode): Define.
20358 (gimple_seq_set_location): New.
20359 * asan.c (build_check_stmt): Change some gimple_build_* calls
20360 to use build_assign and build_type_cast.
20362 2013-04-18 Richard Biener <rguenther@suse.de>
20364 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20365 handle negative step. Remove redundant checks.
20366 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20367 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20368 for negative step and grouped loads fail to vectorize.
20370 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20372 * emit-rtl.c (reset_insn_used_flags): New function.
20373 (reset_all_used_flags): Use it.
20374 (verify_insn_sharing): New function.
20375 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20377 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20379 PR tree-optimization/56984
20380 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20381 and (x >> M) >= N don't register any assertion if N << M is the
20384 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20386 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20387 do not delete_insn insns that have not yet been emitted, only
20388 unlink them with remove_insn.
20389 * df-scan.c (df_insn_delete): Revert r197492.
20391 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20393 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20394 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20396 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20398 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20399 define_insn_and_split.
20400 (and_scc,ior_scc,negscc): Likewise.
20401 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20403 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20405 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20406 can benefit from using a sequence of LDRD instructions in epilogue
20407 instead of a single LDM instruction.
20409 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20412 * doc/extend.texi: Fix typo.
20414 2013-04-17 Richard Biener <rguenther@suse.de>
20416 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20417 (vect_build_slp_tree): ... here.
20418 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20419 match. Remove special-casing of mismatched complex loads.
20420 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20421 re-try the match with swapped commutative operands.
20422 (vect_supported_load_permutation_p): Remove special-casing of
20423 mismatched complex loads.
20424 (vect_analyze_slp_instance): Adjust.
20426 2013-04-17 Richard Biener <rguenther@suse.de>
20428 PR rtl-optimization/56921
20429 * cfgloop.h (struct loop): Add simple_loop_desc member.
20430 (struct niter_desc): Mark with GTY(()).
20431 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20432 * loop-iv.c (get_simple_loop_desc): Likewise.
20433 (free_simple_loop_desc): Likewise.
20436 2013-04-16 Richard Biener <rguenther@suse.de>
20438 PR rtl-optimization/56921
20439 * loop-init.c (pass_rtl_move_loop_invariants): Add
20440 TODO_do_not_ggc_collect to todo_flags_finish.
20441 (pass_rtl_unswitch): Same.
20442 (pass_rtl_unroll_and_peel_loops): Same.
20443 (pass_rtl_doloop): Same.
20445 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20447 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20448 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20449 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20451 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20452 * tree-streamer.c (record_common_node): Adjust reference in comment.
20454 2013-04-17 Terry Guo <terry.guo@arm.com>
20456 * config/arm/cortex-m4.md: Add a new bypass.
20458 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20460 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20462 (*subs_<optab><mode>_multp2): New pattern.
20463 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20464 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20466 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20468 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20469 (*subs_mul_imm_<mode>): New pattern.
20471 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20474 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20475 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20476 (vsx_movti_32bit): Change j->wa to O->wa.
20478 2013-04-16 Richard Biener <rguenther@suse.de>
20480 PR rtl-optimization/56921
20481 * loop-init.c (pass_rtl_move_loop_invariants): Add
20482 TODO_do_not_ggc_collect to todo_flags_finish.
20483 (pass_rtl_unswitch): Same.
20484 (pass_rtl_unroll_and_peel_loops): Same.
20485 (pass_rtl_doloop): Same.
20487 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20489 * config/arm/arm.c (emit_multi_reg_push): New declaration
20490 for an existing function.
20491 (arm_emit_strd_push): New function.
20492 (arm_expand_prologue): Used here.
20493 (arm_emit_ldrd_pop): New function.
20494 (arm_expand_epilogue): Used here.
20495 (arm_get_frame_offsets): Update condition.
20496 (arm_emit_multi_reg_pop): Add a special case for load of a single
20497 register with writeback.
20499 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20501 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20504 2013-04-16 Richard Biener <rguenther@suse.de>
20506 PR tree-optimization/56756
20507 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20508 (first_mem_ref_loc): New.
20509 (execute_sm): Place the load temporarily before a previous
20510 access instead of in the latch edge to ensure its SSA dependencies
20511 are defined at points dominating the load.
20513 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20515 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20516 correct fix by moving header and footer insn to the footer of
20517 the merged basic block. Clear BB_END of the merged-away block.
20519 PR middle-end/43631
20520 * emit-rtl.c (make_note_raw): New function.
20521 (link_insn_into_chain): New static inline function.
20522 (add_insn): Use it.
20523 (add_insn_before, add_insn_after): Factor insn chain linking code...
20524 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20525 using link_insn_into_chain.
20526 (note_outside_basic_block_p): New helper function for emit_note_after
20527 and emit_note_before.
20528 (emit_note_after): Use nobb variant of add_insn_after if the note
20529 should not be contained in a basic block.
20530 (emit_note_before): Use nobb variant of add_insn_before if the note
20531 should not be contained in a basic block.
20532 (emit_note_copy): Use make_note_raw.
20533 (emit_note): Likewise.
20534 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20535 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20536 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20537 the moved barrier the tail of the basic block it follows.
20538 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20540 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20542 PR tree-optimization/56962
20543 * gimple-ssa-strength-reduction.c (record_increment): Only set
20544 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20545 either rhs1 or rhs2 is equal to c->base_expr.
20547 2013-04-15 Richard Biener <rguenther@suse.de>
20549 PR tree-optimization/56933
20550 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20552 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20553 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20554 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20555 dependence check ...
20556 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20558 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20559 GROUP_READ_WRITE_DEPENDENCE.
20561 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20563 * emit-rtl.c (reset_all_used_flags): New function.
20564 (verify_rtl_sharing): Call reset_all_used_flags before and after
20565 performing the checks.
20567 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20569 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20570 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20571 * config/arm/constraints.md (De): New constraint.
20572 * config/arm/neon.md (anddi3_neon): Delete.
20573 (neon_vand<mode>): Expand to standard anddi3 pattern.
20574 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20575 Move earlier in the file.
20576 (neon_inv_logic_op2): Likewise.
20577 (arm_anddi_operand_neon): New predicate.
20579 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20581 * configure.ac (gcc_cv_ld_as_needed): Set
20582 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20583 Use -z ignore, -z record on *-*-solaris2*.
20584 (HAVE_LD_AS_NEEDED): Update comment.
20585 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20586 * configure: Regenerate.
20587 * config.in: Regenerate.
20588 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20589 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20590 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20591 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20592 equivalents. Fix markup.
20593 * doc/tm.texi: Regenerate.
20595 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20597 * config/i386/i386.opt: New option mstack-protector-guard=.
20598 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20599 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20600 TARGET_SSP_TLS_GUARD.
20601 * config/i386/i386.c (ix86_option_override_internal): Set
20602 ix86_stack_protector_guard.
20603 * config/i386/i386.md (stack_protect_set): Enable for
20604 TARGET_SSP_TLS_GUARD only.
20605 (stack_protect_set_<mode>): Ditto.
20606 (stack_protect_test): Ditto.
20607 (stack_protect_test_<mode>): Ditto.
20608 * doc/invoke.texi (i386 Option): Document.
20610 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20613 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20614 (S_MODES): Set H_MODE bit.
20615 (SF_MODES): Set only S_MODE and SF_MODE bits.
20616 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20617 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20618 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20619 <MODE_FLOAT>: Likewise.
20621 2013-04-15 Joey Ye <joey.ye@arm.com>
20623 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20625 2013-04-15 Joey Ye <joey.ye@arm.com>
20627 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20629 (thumb_far_jump_used_p): Count instruction size and set
20632 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20634 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20635 * resource.c (mark_target_live_regs): Reformat conditional block.
20637 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20639 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20640 notes, they are emitted only just before final.
20641 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20643 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20645 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20646 * cfgrtl.c (delete_insn): Call it here instead.
20647 * lra-spills.c (lra_final_code_change): Use delete_insn.
20648 * haifa-sched.c (sched_remove_insn): Likewise.
20649 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20650 returning to the nop pool.
20651 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20652 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20654 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20656 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20657 * doc/tm.texi: Regenerated.
20659 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20661 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20664 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20666 * df-core.c (df_find_def): Compare register numbers.
20667 (df_find_use): Likewise.
20669 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20672 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20673 lra_in_progress for return.
20675 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20677 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20678 define_insn into define_insn_and_split and emit movsicc patterns.
20680 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20682 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20684 2013-04-12 Richard Biener <rguenther@suse.de>
20686 * tree-pass.h (TODO_do_not_ggc_collect): New.
20687 * passes.c (execute_one_ipa_transform_pass): Honor
20688 TODO_do_not_ggc_collect.
20689 (execute_one_pass): Likewise.
20692 2013-04-10 Richard Biener <rguenther@suse.de>
20694 * passes.c (init_optimization_passes): Remove reload pass.
20695 * ira.c (do_reload): Merge into ...
20697 (rest_of_handle_reload): Remove.
20698 (pass_reload): Likewise.
20699 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20700 of reload for vzeroupper pass placement.
20702 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20704 PR tree-optimization/56918
20705 PR tree-optimization/56920
20706 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20707 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20708 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20709 use wide_mul_with_sign method.
20711 2013-04-12 Richard Biener <rguenther@suse.de>
20713 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20714 not be considered a gimple constant.
20716 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20718 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20719 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20722 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20723 Jakub Jelinek <jakub@redhat.com>
20725 * opts.c: Include diagnostic-color.h.
20726 (common_handle_option): Handle OPT_fdiagnostics_color_.
20727 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20728 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20729 (diagnostic-color.o): New.
20730 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20731 (diagnostic_color_rule): New enum.
20732 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20733 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20734 the location string.
20735 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20736 either NULL, or color kind.
20737 * diagnostic-color.c: New file.
20738 * diagnostic-color.h: New file.
20739 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20741 * doc/invoke.texi (-fdiagnostics-color): Document.
20742 * pretty-print.h (pp_show_color): Define.
20743 (struct pretty_print_info): Add show_color field.
20744 * diagnostic.c: Include diagnostic-color.h.
20745 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20746 macros. Colorize error:, warning: etc. strings and also the location
20748 (diagnostic_show_locus): Colorize the caret line.
20749 * pretty-print.c: Include diagnostic-color.h.
20750 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20751 inside of %< %> quotes or quoted through q format modifier.
20753 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20755 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20757 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20759 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20760 code in CC_NZ mode.
20761 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20764 2013-04-11 Marek Polacek <polacek@redhat.com>
20766 PR tree-optimization/48184
20767 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20769 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20771 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20772 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20773 (skip_simple_arithmetic): Tidy up.
20774 * tree.h (skip_simple_constant_arithmetic): Declare.
20776 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20778 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20780 2013-04-11 Richard Biener <rguenther@suse.de>
20782 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20783 generate vector constants.
20785 2013-04-11 Richard Biener <rguenther@suse.de>
20787 PR tree-optimization/56878
20788 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20789 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20791 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20792 Prefer to align the DR with the most invariant base address.
20794 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20796 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20799 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20801 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20802 floating-point vector comparisons against 0.
20804 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20806 PR tree-optimization/56899
20807 * fold-const.c (extract_muldiv_1): Apply distributive law
20808 only if TYPE_OVERFLOW_WRAPS (ctype).
20810 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20813 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20814 loading has side effect.
20816 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20818 * config/sparc/sparc.c: Include tree-pass.h.
20819 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20820 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20821 head of file. Change return type. Split off gate function.
20822 (sparc_gate_work_around_errata): New function.
20823 (pass_work_around_errata): New pass definition.
20824 (insert_pass_work_around_errata) New pass insert definition to
20825 insert pass_work_around_errata just after delayed-branch scheduling.
20826 (sparc_option_override): Insert the pass.
20827 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20829 2013-04-10 David S. Miller <davem@davemloft.net>
20831 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20832 or -mcpu=hypersparc.
20834 * target.def (cstore_mode): New hook.
20835 * target.h: Include insn-codes.h
20836 * targhooks.c: Likewise.
20837 (default_cstore_mode): New function.
20838 * targhooks.h: Declare it.
20839 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20840 * doc/tm.texi: Rebuild.
20841 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20842 target hook, rather than inspecting the insn_data.
20843 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20844 (TARGET_CSTORE_MODE): Redefine.
20845 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20847 * config/sparc/predicates.md (cstore_result_operand): New special
20849 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20850 Use it for operand 0.
20851 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20852 (*snesi_special): Likewise.
20853 (*snesi_zero): Likewise.
20854 (*seqsi_zero): Likewise.
20855 (*sltu_insn): Likewise.
20856 (*sgeu_insn): Likewise.
20857 (*seqdi_special): Make operand 0 and comparison operation be of
20859 (*snedi_special): Likewise.
20860 (*snedi_special_vis3): Likewise.
20861 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20862 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20863 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20864 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20865 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20866 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20867 (*sltu_extend_sp64): Likewise.
20868 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20869 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20870 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20871 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20872 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20874 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20876 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20877 (aarch64_start_file): Use the new function.
20879 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20880 Jason Merrill <jason@redhat.com>
20882 * common.opt: Add -gdwarf.
20883 * opts.c (common_handle_option): Handle it.
20884 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20886 2013-04-10 Richard Biener <rguenther@suse.de>
20888 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20889 (execute_one_ipa_transform_pass): But unconditionally here.
20890 (execute_one_pass): And here.
20891 (init_optimization_passes): Remove reload pass.
20892 * tree-pass.h (TODO_ggc_collect): Remove.
20893 (pass_reload): Likewise.
20894 * ira.c (do_reload): Merge into ...
20896 (rest_of_handle_reload): Remove.
20897 (pass_reload): Likewise.
20898 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20899 of reload for vzeroupper pass placement.
20900 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20901 and todo_flags_finish of all passes.
20903 2013-04-10 Richard Biener <rguenther@suse.de>
20905 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20906 first_const_oprnd field, rename first_def_type to first_op_type.
20907 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20908 (vect_get_and_check_slp_defs): Always use the type of the
20909 operand. Allow mixed vect_external_def, vect_constant_def types.
20910 (vect_get_constant_vectors): Handle mixed vect_external_def,
20911 vect_constant_def types.
20913 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20915 PR tree-optimization/55524
20916 * tree-ssa-math-opts.c
20917 (convert_mult_to_fma): Don't use an fms construct
20918 when we don't have an fms operation, but fnma, and it looks
20919 likely that we'll be able to use the latter.
20921 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20923 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20925 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20926 inline fail caused by overwritable functions.
20928 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20930 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20931 unnecessary bits in the constant power of two case.
20933 2013-04-10 Richard Biener <rguenther@suse.de>
20935 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20936 broken code swapping operands.
20937 (vect_build_slp_tree): Do not compute load permutations here.
20938 (vect_analyze_slp_instance): Compute load permutations here,
20939 after building the SLP tree.
20941 2013-04-09 Christian Bruel <christian.bruel@st.com>
20943 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20944 of next/prev_real_insn.
20946 2013-04-09 Jan Hubicka <jh@suse.cz>
20948 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20949 Drop aliased parameter.
20950 (function_and_variable_visibility): Do not handle alias pairs.
20951 * cgraph.c (varpool_externally_visible_p): Update prototype.
20952 * varpool.c (varpool_add_new_variable): Update.
20954 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20956 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
20958 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
20960 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
20962 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
20964 2013-04-09 Marek Polacek <polacek@redhat.com>
20966 PR tree-optimization/48762
20967 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
20969 2013-04-09 Richard Biener <rguenther@suse.de>
20971 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
20973 (vect_build_slp_tree): Likewise.
20974 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
20975 calculating the cost of a SLP instance.
20976 (vect_analyze_slp_instance): Use it from here, after building
20979 2013-04-09 Jakub Jelinek <jakub@redhat.com>
20981 PR middle-end/56883
20982 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
20983 expand_omp_for_static_chunk): Use simple_p = true in
20984 force_gimple_operand_gsi calls when assigning to addressable decls.
20986 2013-04-09 Jeff Law <law@redhat.com>
20988 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
20989 when the boolean was created by converting a wider object which
20990 had a boolean range.
20992 2013-04-09 Richard Biener <rguenther@suse.de>
20994 * tree-vectorizer.h (slp_void_p): Remove.
20995 (slp_tree): Typedef before _slp_tree declaration.
20996 (struct _slp_tree): Use a vector of slp_tree as children.
20997 (vect_get_place_in_interleaving_chain): Remove.
20998 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
21000 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
21002 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
21003 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
21004 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
21005 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
21006 Use slp_node instead of slp_void_p and adjust.
21008 2013-04-09 Richard Biener <rguenther@suse.de>
21010 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
21011 work that is not necessary.
21013 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21015 PR tree-optimization/56854
21016 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
21017 forward into clobber stmts if it would change MEM_REF lhs into
21020 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
21022 * tree.c (type_hash_lookup, type_hash_add): Make static.
21023 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
21025 2013-04-09 Richard Biener <rguenther@suse.de>
21027 * tree.h (unsave_expr_now): Remove.
21028 * tree-inline.c (mark_local_for_remap_r): Remove.
21029 (unsave_expr_1): Likewise.
21030 (unsave_r): Likewise.
21031 (unsave_expr_now): Likewise.
21032 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
21033 (propagate_tree_value): Likewise.
21035 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
21037 * doc/rtl.texi (sequence): Rewrite documentation to match the
21038 current use of SEQUENCE rtl objects.
21039 * rtl.def (SEQUENCE): Likewise.
21041 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
21042 Update documentation.
21043 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
21044 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
21046 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
21048 2013-04-08 Teresa Johnson <tejohnson@google.com>
21050 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
21051 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
21053 (estimate_edge_size_and_time): Add comment to suggest using rounding
21055 (estimate_node_size_and_time): Ditto.
21056 (remap_edge_change_prob): Use helper rounding divide methods.
21057 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
21058 (gimple_mod_pow2_value_transform): Ditto.
21059 (gimple_mod_subtract_transform): Ditto.
21060 (gimple_ic_transform): Ditto.
21061 (gimple_stringops_transform): Ditto.
21062 * stmt.c (conditional_probability): Ditto.
21063 (emit_case_dispatch_table): Ditto.
21064 * lto-cgraph.c (merge_profile_summaries): Ditto.
21065 * tree-optimize.c (execute_fixup_cfg): Ditto.
21066 * cfgcleanup.c (try_forward_edges): Ditto.
21067 * cfgloopmanip.c (scale_loop_profile): Ditto.
21069 (duplicate_loop_to_header_edge): Ditto.
21070 (lv_adjust_loop_entry_edge): Ditto.
21071 * tree-vect-loop.c (vect_transform_loop): Ditto.
21072 * profile.c (compute_branch_probabilities): Ditto.
21073 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
21074 * lto-streamer-in.c (input_cfg): Ditto.
21075 * gimple-streamer-in.c (input_bb): Ditto.
21076 * ipa-cp.c (update_profiling_info): Ditto.
21077 (update_specialized_profile): Ditto.
21078 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
21079 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
21081 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21082 (compute_trg_info): Ditto.
21083 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
21084 (purge_dead_edges): Ditto.
21085 * loop-unswitch.c (unswitch_loop): Ditto.
21086 * cgraphclones.c (cgraph_clone_edge): Ditto.
21087 (cgraph_clone_node): Ditto.
21088 * tree-inline.c (copy_bb): Ditto.
21089 (copy_edges_for_bb): Ditto.
21090 (initialize_cfun): Ditto.
21091 (copy_cfg_body): Ditto.
21092 (expand_call_inline): Ditto.
21094 2013-04-08 Kai Tietz <ktietz@redhat.com>
21096 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
21097 TARGET_CYGWIN64 by TARGET_64BIT.
21099 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21101 * config/epiphany/epiphany.md (GPR_1): New constant.
21102 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
21103 * config/epiphany/epiphany.c (gen_compare_reg):
21104 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
21105 is already in place.
21106 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
21107 Don't require being called during rtl expansion; If y operlaps r0,
21109 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
21110 (epiphany_expand_epilogue): Likewise.
21112 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
21113 Don't use CC_FPmode for ORDERED / UNORDERED.
21114 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
21116 * config/epiphany/constraints.md (CnL): New constraint.
21117 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
21118 * config/epiphany/predicates.md (add_operand): Allow 1024.
21120 * config/epiphany/epiphany.md (logical_op): New code iterator.
21121 (op_mnc): New code attribute.
21122 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
21123 (mov_f+1, mov_f+2): New peephole2 patterns.
21125 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
21126 (cstoresi4): Also allow re-use of zero result when doing a NE
21127 comparison to a non-zero operand.
21128 Use (clobber (scratch)) for first insn if the gpr output is not needed.
21130 * config/epiphany/epiphany.md (<insn_opname>v2si3):
21131 Use gen_addsi3_i / gen_subsi3_i.
21133 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21137 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21138 contain anything but clobbers, at most one __builtin_stack_restore,
21139 optionally debug stmts and final resx, and if it has at least one
21140 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21141 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21142 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21143 which isn't defaut definition, remove them.
21144 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21145 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21146 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21147 with MEM_REF LHS with SSA_NAME address.
21149 2013-04-08 Jeff Law <law@redhat.com>
21151 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21153 2013-04-08 Richard Biener <rguenther@suse.de>
21155 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21157 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21158 determined vector type.
21159 (vect_analyze_data_refs): Likewise.
21160 (vect_get_new_vect_var): Adjust.
21161 (vect_create_destination_var): Preserve SSA name versions.
21162 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21163 not dump anything here.
21165 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21167 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21168 Add member lr_slot_known.
21169 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21171 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21172 Remove code that sets lr_slot_offset according to what a previous
21173 version of epiphany_emit_save_restore used to do.
21174 (epiphany_emit_save_restore): When doing an lr save or restore,
21175 set/verify lr_slot_known and lr_slot_offset.
21177 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21180 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21183 2013-04-08 Richard Biener <rguenther@suse.de>
21185 * alias.c (find_base_term): Fix thinko in previous change.
21187 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21189 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21190 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21191 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21192 if possible to compute val.
21193 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21194 For QImode integers don't require anything about precision. Use
21195 const_with_all_bytes_same to find out if the constant doesn't have
21196 repeated bytes in it.
21198 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21200 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21203 2013-04-08 Marek Polacek <polacek@redhat.com>
21205 PR rtl-optimization/48182
21206 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21209 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21212 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21213 nuses, make sure we have a label.
21215 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21218 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21219 (rs6000_emit_swdiv_low_precision): Remove.
21220 (rs6000_emit_swdiv): Rewrite to handle between one and four
21221 iterations of Newton-Raphson generally; modify required number of
21222 iterations for some cases.
21223 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21225 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21227 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21228 set-but-unused variable.
21230 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21231 basic blocks of released function bodies garbage-collectable.
21233 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21234 (struct rtl_opt_pass): Add TODO_df_finish.
21236 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21238 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21240 * config/arm/constraints.md (q): New constraint.
21241 * config/arm/ldrdstrd.md: New file.
21242 * config/arm/arm.md (ldrdstrd.md) New include.
21243 (arm_movdi): Use "q" instead of "r" constraint
21244 for double-word memory access.
21245 (movdf_soft_insn): Likewise.
21246 * config/arm/vfp.md (movdi_vfp): Likewise.
21247 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21248 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21249 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21250 (mem_ok_for_ldrd_strd): Likewise.
21251 (output_move_double): Update assertion.
21253 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21255 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21257 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21259 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21260 define_insn_and_split.
21261 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21263 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21265 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21266 define_insn_and_split.
21267 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21268 (shiftsi3_compare): New pattern.
21269 (rrx): New pattern.
21270 * config/arm/unspecs.md (UNSPEC_RRX): New.
21272 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21274 * config/arm/arm.md (negdi_extendsidi): New pattern.
21275 (negdi_zero_extendsidi): Likewise.
21277 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21279 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21280 define_insn_and_split.
21281 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21282 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21284 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21286 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21287 define_insn_and_split.
21288 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21289 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21291 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21293 * config/arm/arm.md (subsi3_carryin): New pattern.
21294 (subsi3_carryin_const): Likewise.
21295 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21296 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21298 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21300 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21302 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21304 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21305 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21307 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21309 * config/arm/arm.c (arm_expand_builtin): Change fcode
21310 type to unsigned int.
21312 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21314 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21316 2013-04-04 Ian Lance Taylor <iant@google.com>
21318 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21321 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21323 PR middle-end/56729
21324 * df-scan.c (df_insn_delete): Disable failing assert.
21326 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21328 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21329 New function prototype.
21330 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21331 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21332 (arm_builtin_vectorized_function): New function.
21334 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21336 * config/arm/arm_neon_builtins.def: New file.
21337 * config/arm/arm.c (neon_builtin_data): Move contents to
21338 arm_neon_builtins.def.
21339 (enum arm_builtins): Include neon builtin definitions.
21340 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21341 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21343 2013-04-04 Marek Polacek <polacek@redhat.com>
21345 PR tree-optimization/48186
21346 * predict.c (maybe_hot_frequency_p): Return false if
21347 HOT_BB_FREQUENCY_FRACTION is 0.
21348 (cgraph_maybe_hot_edge_p): Likewise.
21350 2013-04-04 Richard Biener <rguenther@suse.de>
21352 PR tree-optimization/56826
21353 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21356 2013-04-04 Richard Biener <rguenther@suse.de>
21358 PR tree-optimization/56213
21359 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21360 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21362 2013-04-04 Richard Biener <rguenther@suse.de>
21364 PR tree-optimization/56837
21365 * tree-loop-distribution.c (classify_partition): For non-zero
21366 values require that the value has the same precision as its
21367 mode to be useful as memset value.
21369 2013-04-03 Nick Clifton <nickc@redhat.com>
21371 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21372 (fmssf4): Use fmsf.s on E3V5 architectures.
21373 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21374 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21376 2013-04-03 Jeff Law <law@redhat.com>
21378 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21379 (lra-eliminations.o): Likewise.
21381 2013-04-03 Teresa Johnson <tejohnson@google.com>
21383 * gcov-io.c (compute_working_sets): Moved most of body of old
21384 compute_working_sets here from profile.c.
21385 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21386 (gcov_working_set_t): Moved typedef here from basic-block.h
21387 (compute_working_set): Declare.
21388 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21389 (get_working_sets): Renamed from compute_working_set,
21390 replace most of body with call to new compute_working_sets.
21391 (get_exec_counts): Replace call to compute_working_sets
21392 to get_working_sets.
21393 * profile.h (get_working_sets): Renamed from compute_working_set.
21394 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21395 to get_working_sets.
21396 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21397 * gcov-dump.c (dump_working_sets): New function.
21399 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21401 * hwint.c (sext_hwi, zext_hwi): New functions.
21402 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21403 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21404 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21405 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21406 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21407 (sext_hwi, zext_hwi): New functions.
21409 2013-04-03 Jeff Law <law@redhat.com>
21411 PR tree-optimization/56799
21412 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21413 back test for widening conversion erroneously dropped in prior change.
21415 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21418 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21419 instead of next_real_insn.
21421 2013-04-03 Marek Polacek <polacek@redhat.com>
21424 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21426 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21429 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21431 (thumb1_output_casesi): Likewise.
21432 (thumb2_output_casesi): Likewise.
21434 2013-04-03 Richard Biener <rguenther@suse.de>
21436 PR tree-optimization/56817
21437 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21439 (tree_unroll_loops_completely_1): ... new function to manually
21440 walk the loop tree, properly defering outer loops of unrolled
21441 loops to later iterations.
21443 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21445 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21446 (vectorizable_load): Likewise.
21447 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21448 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21450 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21452 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21455 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21457 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21459 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21461 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21463 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21465 PR tree-optimization/56790
21466 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21469 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21471 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21473 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21474 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21477 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21480 * tree.h (force_folding_builtin_constant_p): New decl.
21481 * builtins.c (force_folding_builtin_constant_p): New variable.
21482 (fold_builtin_constant_p): Fold immediately also if
21483 force_folding_builtin_constant_p.
21485 2013-04-03 Richard Biener <rguenther@suse.de>
21487 PR tree-optimization/56812
21488 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21489 DRs of the same interleaving chain are independent.
21491 2013-04-02 Jason Merrill <jason@redhat.com>
21493 * gdbinit.in (pbb): Use debug fn.
21495 2013-04-02 Lawrence Crowl <crowl@google.com>
21497 * sese.h (struct ivtype_map_elt_s): Remove unused.
21498 (extern debug_ivtype_map): Remove unused.
21499 (extern eq_ivtype_map_elts): Remove unused.
21500 * sese.c (debug_ivtype_map): Removed unused.
21501 (debug_ivtype_map_1): Removed unused.
21502 (debug_ivtype_elt): Remove unused.
21503 (eq_ivtype_map_elts): Remove unused.
21506 2013-04-02 Kai Tietz <ktietz@redhat.com>
21509 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21510 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21511 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21513 (legitimize_pe_coff_symbol): Likewise.
21514 (is_imported_p): New helper-function.
21515 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21516 for Windows x64 targets.
21517 (ix86_expand_prologue): Optimize for pe-coff targets.
21518 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21519 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21520 medium/large code-model.
21521 (legitimize_pic_address): Likewise.
21522 (legitimize_tls_address): Likewise.
21523 (ix86_expand_call): Likewise.
21524 (x86_output_mi_thunk): Likewise.
21525 (get_dllimport_decl): Add new beimport argument.
21526 (construct_plt_address): Don't assert for x64 pe-coff targets.
21527 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21529 (SYMBOL_FLAG_STUBVAR): New macro.
21530 (SYMBOL_REF_STUBVAR_P): Likewise.
21531 * config/i386/winnt.c (stub_list): New structure.
21532 (stub_head): New local variable.
21533 (i386_pe_record_stub): New function.
21534 (i386_pe_file_end): Emit refptr-stubs.
21536 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21538 PR rtl-optimization/56745
21539 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21540 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21543 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21544 and both of them are MEM_REFs, just compare first argument for
21545 equality and attempt to deal even with differing offsets.
21548 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21549 of gimple_clobber_p to be MEM_REF.
21550 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21551 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21552 after gimplification.
21553 * asan.c (get_mem_ref_of_assignment): Don't instrument
21554 gimple_clobber_p stmts.
21555 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21556 gimple_clobber_p stmt if they have MEM_REF lhs and
21557 are dead because of another gimple_clobber_p stmt.
21558 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21559 gimple_clobber_p stmts like debug stmts.
21560 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21561 that refer to unused VAR_DECLs or uninitialized values.
21562 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21563 gimple_clobber_p stmts if they refer to removed parameters.
21564 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21567 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21569 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21570 using SWI48 mode attribute.
21572 2013-04-02 Wei Mi <wmi@google.com>
21574 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21575 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21576 *<rotate_insn><mode>3_mask in i386.md.
21578 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21580 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21582 2013-04-02 Richard Biener <rguenther@suse.de>
21584 PR tree-optimization/56778
21585 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21586 Runtime alias tests are not supported for gather loads.
21587 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21588 stmts referenced from SSA operands before updating SSA form.
21590 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21591 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21593 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21594 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21595 * config/arm/cortex-a53.md: New file.
21596 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21597 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21598 * config/arm/arm.c (arm_issue_rate): Likewise.
21599 * config/arm/arm-tune.md: Regenerate
21600 * config/arm/arm-tables.opt: Regenerate.
21601 * config/arm/arm-cores.def: Add cortex-a53.
21603 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21605 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21608 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21610 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21611 scalar load/store operations using B/H registers.
21612 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21614 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21616 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21618 * config/aarch64/aarch64.c
21619 (aarch64_simd_scalar_immediate_valid_for_move): New.
21620 * config/aarch64/aarch64-protos.h
21621 (aarch64_simd_scalar_immediate_valid_for_move): New.
21622 * config/aarch64/constraints.md (Dh, Dq): New.
21623 * config/aarch64/iterators.md (hq): New.
21625 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21627 * reorg.c (get_branch_condition): Deal with conditional returns.
21628 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21630 2013-04-01 Wei Mi <wmi@google.com>
21632 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21633 Truncate operand 2 using %b asm operand modifier.
21634 (*<shift_insn><mode>3_mask): Ditto.
21635 (*<rotate_insn><mode>3_mask): Ditto.
21637 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21639 PR middle-end/56798
21640 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21642 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21644 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21646 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21648 2013-03-30 Lawrence Crowl <crowl@google.com>
21650 * dse.c (clear_alias_sets): Remove never set.
21651 (disqualified_clear_alias_sets): Remove never set.
21652 (clear_alias_mode_pool): Remove never set.
21653 (dse_step0): Remove condition that is never true.
21654 (canon_address): Remove condition that is never true.
21655 (dse_step7): Remove condition that is never true.
21656 (rest_of_handle_dse): Remove condition that is never true.
21657 (rest_of_handle_dse::did_global): Remove never read from above.
21658 (dse_step2_spill): Remove never called from above.
21659 (dse_step5_spill): Remove never called from above.
21661 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21663 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21664 JUMP_TABLE_DATA changes.
21665 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21666 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21667 (Insns) <jump_table_data>: New entry.
21668 * doc/tm.texi: Regenerate.
21670 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21672 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21673 for table jump at the end of a basic block using tablejump_p.
21674 * targhooks.c (default_invalid_within_doloop): Likewise.
21675 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21676 target hook implementation that is identical to the default hook.
21677 (rs6000_invalid_within_doloop): Remove.
21679 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21680 unused variable from tablejump_p call.
21682 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21683 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21684 (INSN_DELETED_P): Likewise.
21685 (emit_jump_table_data): New prototype.
21686 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21687 after 4th as unused.
21688 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21689 * sched-vis.c (print_insn): Likewise.
21690 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21691 insn for compatibility with back ends that use next_active_insn to
21692 identify jump table data.
21693 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21694 (remove_insn): Likewise.
21695 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21697 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21698 (emit_jump_table_data): New function.
21700 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21701 basic block, a JUMP_TABLE_DATA never is.
21702 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21703 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21704 off from code handling real insns.
21705 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21706 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21707 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21708 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21709 is not a NONDEBUG_INSN_P.
21710 * ira-costs.c (scan_one_insn): Likewise.
21711 * jump.c (mark_all_labels): Likewise.
21712 (mark_jump_label_1): Likewise.
21713 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21714 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21716 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21717 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21718 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21719 JUMP_TABLE_DATA_P insns.
21720 (calculate_elim_costs_all_insns): Likewise.
21721 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21722 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21723 (delete_output_reload): Code style fixups.
21724 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21725 insn flags on this non-insn.
21726 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21727 as scheduling barriers, for pre-change compatibility.
21728 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21729 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21731 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21732 redundant JUMP_TABLE_DATA_P test.
21733 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21734 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21735 (frv_for_each_packet): Likewise.
21736 * config/i386/i386.c (min_insn_size): Likewise.
21737 (ix86_avoid_jump_mispredicts): Likewise.
21738 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21739 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21740 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21741 (mips16_insn_length): Robustify.
21742 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21743 (mips16_split_long_branches): Likewise.
21744 * config/pa/pa.c (pa_combine_instructions): Likewise.
21745 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21746 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21747 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21748 as contributing to pool range lengths.
21749 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21750 Remove redundant JUMP_TABLE_DATA_P test.
21751 (sh_loop_align): Likewise.
21752 (split_branches): Likewise.
21753 (sh_insn_length_adjustment): Likewise.
21754 * config/spu/spu.c (get_branch_target): Likewise.
21756 2013-03-29 Jan Hubicka <jh@suse.cz>
21758 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21759 gcov streaming; stream hot bb threshold to ltrans.
21760 * predict.c (get_hot_bb_threshold): Break out from ....
21761 (maybe_hot_count_p): ... here.
21762 (set_hot_bb_threshold): New function.
21763 * lto-section-in.c (lto_section_name): Add profile.
21764 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21765 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21766 and data-streamer.h
21767 (histogram_entry): New structure.
21768 (histogram, histogram_pool): New global vars.
21769 (histogram_hash): New structure.
21770 (histogram_hash::hash): New method.
21771 (histogram_hash::equal): Likewise.
21772 (account_time_size): New function.
21773 (cmp_counts): New function.
21774 (dump_histogram): New function.
21775 (ipa_profile_generate_summary): New function.
21776 (ipa_profile_write_summary): New function.
21777 (ipa_profile_read_summary): New function.
21778 (ipa_profile): Decide on threshold.
21779 (pass_ipa_profile): Add ipa_profile_write_summary and
21780 ipa_profile_read_summary.
21781 * Makefile.in (ipa.o): Update dependencies.
21782 * lto-streamer.h (LTO_section_ipa_profile): New section.
21784 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21786 * tree.h (VAR_P): New.
21788 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21791 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21793 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21795 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21796 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21797 (control_flow_insn_p): Likewise.
21798 * cfgrtl.c (duplicate_insn_chain): Likewise.
21799 * final.c (get_attr_length_1): Likewise.
21800 (shorten_branches): Likewise.
21801 (final_scan_insn): Likewise.
21802 * function.c (instantiate_virtual_regs): Likewise.
21803 * gcse.c (insert_insn_end_basic_block): Likewise.
21804 * ira-costs.c (scan_one_insn): Likewise.
21805 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21806 * lra.c (check_rtl): Likewise.
21807 * reload1.c (elimination_costs_in_insn): Likewise.
21808 * reorg.c (follow_jumps): Likewise.
21810 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21811 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21812 (thumb_far_jump_used_p): Likewise.
21813 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21814 (workaround_speculation): Likewise.
21815 (add_sched_insns_for_speculation): Likewise.
21816 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21817 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21818 (frv_for_each_packet): Likewise.
21819 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21820 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21821 (final_emit_insn_group_barriers): Likewise.
21822 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21823 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21824 (mips16_insn_length): Likewise.
21825 * config/pa/pa.c (pa_reorg): Likewise.
21826 (pa_combine_instructions): Likewise.
21827 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21828 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21829 (sh_reorg): Likewise.
21830 (split_branches): Likewise.
21831 * config/spu/spu.c (get_branch_target): Likewise.
21833 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21836 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21838 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21839 Fix declaration name.
21841 2013-03-28 Lawrence Crowl <crowl@google.com>
21843 * graphds.h (struct graph.indicies): Remove unused.
21844 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21845 (SCOP_ORIGINAL_PDDRS): Remove unused.
21846 * sese.h (extern insert_loop_close_phis): Removed unused.
21847 (extern insert_guard_phis): Removed unused.
21848 (extern ivtype_map_elt_info): Removed unused.
21849 (new_ivtype_map_elt): Removed unused.
21850 * sese.c (ivtype_map_elt_info): Removed unused.
21852 2013-03-28 Lawrence Crowl <crowl@google.com>
21854 * Makefile.in: Add several missing include dependences.
21856 (test-dump.o): New. This object is not added to any executable,
21857 but is present for ad-hoc testing.
21859 (debug (const bitmap_head_def &)): New.
21860 (debug (const bitmap_head_def *)): New.
21862 (extern debug (const bitmap_head_def &)): New.
21863 (extern debug (const bitmap_head_def *)): New.
21865 (debug (edge_def &)): New.
21866 (debug (edge_def *)): New.
21868 (debug (basic_block_def &)): New.
21869 (debug (basic_block_def *)): New.
21871 (dump_node (const_tree, int, FILE *)): Correct source file.
21873 (debug (die_struct &)): New.
21874 (debug (die_struct *)): New.
21876 (extern debug (die_struct &)): New.
21877 (extern debug (die_struct *)): New.
21878 * gimple-pretty-print.c
21879 (debug (gimple_statement_d &)): New.
21880 (debug (gimple_statement_d *)): New.
21881 * gimple-pretty-print.h
21882 (extern debug (gimple_statement_d &)): New.
21883 (extern debug (gimple_statement_d *)): New.
21885 (debug (ira_allocno_copy &)): New.
21886 (debug (ira_allocno_copy *)): New.
21887 (debug (ira_allocno &)): New.
21888 (debug (ira_allocno *)): New.
21890 (extern debug (ira_allocno_copy &)): New.
21891 (extern debug (ira_allocno_copy *)): New.
21892 (extern debug (ira_allocno &)): New.
21893 (extern debug (ira_allocno *)): New.
21895 (debug (live_range &)): New.
21896 (debug (live_range *)): New.
21898 (debug (lra_live_range &)): New.
21899 (debug (lra_live_range *)): New.
21901 (debug (lra_live_range &)): New.
21902 (debug (lra_live_range *)): New.
21904 (debug (omega_pb_d &)): New.
21905 (debug (omega_pb_d *)): New.
21907 (extern debug (omega_pb_d &)): New.
21908 (extern debug (omega_pb_d *)): New.
21910 (debug (const rtx_def &)): New.
21911 (debug (const rtx_def *)): New.
21913 (debug_tree (tree): Move within file.
21914 (debug_raw (const tree_node &)): New.
21915 (debug_raw (const tree_node *)): New.
21916 (dump_tree_via_hooks (const tree_node *, int)): New.
21917 (debug (const tree_node &)): New.
21918 (debug (const tree_node *)): New.
21919 (debug_verbose (const tree_node &)): New.
21920 (debug_verbose (const tree_node *)): New.
21921 (debug_head (const tree_node &)): New.
21922 (debug_head (const tree_node *)): New.
21923 (debug_body (const tree_node &)): New.
21924 (debug_body (const tree_node *)): New.
21925 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21926 (debug (vec<tree, va_gc> &)): New.
21927 (debug (vec<tree, va_gc> *)): New.
21929 (extern debug (const rtx_def &)): New.
21930 (extern debug (const rtx_def *)): New.
21932 (debug_raw (simple_bitmap_def &)): New.
21933 (debug_raw (simple_bitmap_def *)): New.
21934 (debug (simple_bitmap_def &)): New.
21935 (debug (simple_bitmap_def *)): New.
21937 (extern debug (simple_bitmap_def &)): New.
21938 (extern debug (simple_bitmap_def *)): New.
21939 (extern debug_raw (simple_bitmap_def &)): New.
21940 (extern debug_raw (simple_bitmap_def *)): New.
21942 (debug (vinsn_def &)): New.
21943 (debug (vinsn_def *)): New.
21944 (debug_verbose (vinsn_def &)): New.
21945 (debug_verbose (vinsn_def *)): New.
21946 (debug (expr_def &)): New.
21947 (debug (expr_def *)): New.
21948 (debug_verbose (expr_def &)): New.
21949 (debug_verbose (expr_def *)): New.
21950 (debug (vec<rtx> &)): New.
21951 (debug (vec<rtx> *)): New.
21953 (extern debug (vinsn_def &)): New.
21954 (extern debug (vinsn_def *)): New.
21955 (extern debug_verbose (vinsn_def &)): New.
21956 (extern debug_verbose (vinsn_def *)): New.
21957 (extern debug (expr_def &)): New.
21958 (extern debug (expr_def *)): New.
21959 (extern debug_verbose (expr_def &)): New.
21960 (extern debug_verbose (expr_def *)): New.
21961 (extern debug (vec<rtx> &)): New.
21962 (extern debug (vec<rtx> *)): New.
21964 (_list_iter_cond_expr): Make inline instead of static.
21966 (debug (sreal &)): New.
21967 (debug (sreal *)): New.
21969 (extern debug (sreal &)): New.
21970 (extern debug (sreal *)): New.
21972 (extern debug_raw (const tree_node &)): New.
21973 (extern debug_raw (const tree_node *)): New.
21974 (extern debug (const tree_node &)): New.
21975 (extern debug (const tree_node *)): New.
21976 (extern debug_verbose (const tree_node &)): New.
21977 (extern debug_verbose (const tree_node *)): New.
21978 (extern debug_head (const tree_node &)): New.
21979 (extern debug_head (const tree_node *)): New.
21980 (extern debug_body (const tree_node &)): New.
21981 (extern debug_body (const tree_node *)): New.
21982 (extern debug (vec<tree, va_gc> &)): New.
21983 (extern debug (vec<tree, va_gc> *)): New.
21985 (debug (struct loop &)): New.
21986 (debug (struct loop *)): New.
21987 (debug_verbose (struct loop &)): New.
21988 (debug_verbose (struct loop *)): New.
21989 * tree-dump.c: Add header dependence.
21991 (extern debug (struct loop &)): New.
21992 (extern debug (struct loop *)): New.
21993 (extern debug_verbose (struct loop &)): New.
21994 (extern debug_verbose (struct loop *)): New.
21996 (debug (data_reference &)): New.
21997 (debug (data_reference *)): New.
21998 (debug (vec<data_reference_p> &)): New.
21999 (debug (vec<data_reference_p> *)): New.
22000 (debug (vec<ddr_p> &)): New.
22001 (debug (vec<ddr_p> *)): New.
22003 (extern debug (data_reference &)): New.
22004 (extern debug (data_reference *)): New.
22005 (extern debug (vec<data_reference_p> &)): New.
22006 (extern debug (vec<data_reference_p> *)): New.
22007 (extern debug (vec<ddr_p> &)): New.
22008 (extern debug (vec<ddr_p> *)): New.
22010 (debug (pt_solution &)): New.
22011 (debug (pt_solution *)): New.
22013 (extern debug (pt_solution &)): New.
22014 (extern debug (pt_solution *)): New.
22016 (debug (_var_map &)): New.
22017 (debug (_var_map *)): New.
22018 (debug (tree_live_info_d &)): New.
22019 (debug (tree_live_info_d *)): New.
22021 (extern debug (_var_map &)): New.
22022 (extern debug (_var_map *)): New.
22023 (extern debug (tree_live_info_d &)): New.
22024 (extern debug (tree_live_info_d *)): New.
22026 2013-03-28 Jan Hubicka <jh@suse.cz>
22028 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
22030 2013-03-28 Ian Bolton <ian.bolton@arm.com>
22032 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
22033 record only when desired or required.
22035 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
22037 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
22038 *vec_extractv2di_1_rex64. Use x64 isa attribute.
22040 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22042 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
22043 (*andsi3_compare0_uxtw): New pattern.
22044 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
22045 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
22047 2013-03-28 Jan Hubicka <jh@suse.cz>
22049 * data-streamer-in.c (streamer_read_gcov_count): New function.
22050 * gimple-streamer-out.c: Include value-prof.h.
22051 (output_gimple_stmt): Output histogram.
22052 (output_bb): Use streamer_write_gcov_count.
22053 * value-prof.c: Include data-streamer.h
22054 (dump_histogram_value): Add HIST_TYPE_MAX.
22055 (stream_out_histogram_value): New function.
22056 (stream_in_histogram_value): New function.
22057 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
22058 (stream_out_histogram_value, stream_in_histogram_value): Declare.
22059 * data-streamer-out.c (streamer_write_gcov_count): New function.
22060 (streamer_write_gcov_count_stream): New function.
22061 * lto-cgraph.c (lto_output_edge): Update counter streaming.
22062 (lto_output_node): Likewise.
22063 (input_node, input_edge): Likewise.
22064 * lto-streamer-out.c (output_cfg): Update streaming.
22065 * lto-streamer-in.c (input_cfg): Likewise.
22066 * data-streamer.h (streamer_write_gcov_count,
22067 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
22068 * gimple-streamer-in.c: Include value-prof.h
22069 (input_gimple_stmt): Input histograms.
22070 (input_bb): Update profile streaming.
22072 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22074 * genmodes.c (emit_max_int): New function.
22075 (emit_insn_modes_h): Added call to emit_max_function.
22076 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
22078 * machmode.def: Fixed comment.
22080 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22082 * combine.c (try_combine): Removed useless assert.
22083 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
22085 2013-03-28 Marek Polacek <polacek@redhat.com>
22086 Richard Biener <rguenther@suse.de>
22088 PR tree-optimization/56695
22089 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
22090 build signed result of a vector comparison.
22091 * tree-cfg.c (verify_gimple_comparison): Check that a result
22092 of a vector comparison has signed type.
22094 2013-03-28 Richard Biener <rguenther@suse.de>
22096 PR tree-optimization/37021
22097 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
22098 do not restrict gaps between groups.
22099 * tree-vect-stmts.c (vectorizable_load): Properly account for
22100 a gap between groups.
22102 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
22104 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
22105 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
22108 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
22110 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
22111 * doc/extend.texi (Named Address Spaces): Ditto.
22112 (Variable Attributes): Ditto.
22114 2013-03-27 Kai Tietz <ktietz@redhat.com>
22116 * config.build: Add support for cygwin x64 target.
22117 * config.gcc: Likewise.
22118 * config.host: Likewise.
22119 * configure.ac: Likewise
22120 * configure: Regenerated.
22122 2013-03-27 Kai Tietz <ktietz@redhat.com>
22124 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
22125 * config/i386/t-cygwin-w64: New file.
22126 * config/i386/cygwin-w64.h: New file.
22127 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
22128 and add support for x64-cygwin target.
22129 (CPP_SPEC): Likewise.
22130 (CXX_WRAP_SPEC_LIST): Undefine before define.
22131 (LIBGCJ_SONAME): Use 15 as version.
22133 2013-03-27 Richard Biener <rguenther@suse.de>
22135 PR tree-optimization/56716
22136 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22137 dumping for ref nodes.
22139 2013-03-27 Martin Jambor <mjambor@suse.cz>
22141 PR tree-optimization/55334
22142 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22143 restricted pointers to arrays.
22145 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22147 * Makefile.in (.SUFFIXES): Add .cc.
22148 (.c.o): Apply same recipe for implicit rule .cc.o.
22150 2013-03-27 Richard Biener <rguenther@suse.de>
22152 PR tree-optimization/37021
22153 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22154 REALPART/IMAGPART_EXPRs around the supported refs.
22155 * tree-ssa-structalias.c (find_func_aliases): Assume that
22156 floating-point values are not used to transfer pointers.
22158 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22160 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22161 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22162 * doc/tm.texi: Regenerate.
22163 * targhooks.h (default_has_ifunc_p): New.
22164 * targhooks.c (default_has_ifunc_p): Ditto.
22165 * config/linux-protos.h: New file.
22166 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22167 hook for linux which disables support of indirect functions in android.
22168 * config/linux-android.c: New file.
22169 * config/t-linux-android.c: Ditto.
22170 * config.gcc: Added new object file linux-android.o.
22171 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22172 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22173 * varasm.c (do_assemble_alias): Likewise.
22174 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22175 doesn't support indirect functions.
22176 * configure: Regenerate.
22178 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22181 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22182 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22185 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22187 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22189 2013-03-27 Terry Guo <terry.guo@arm.com>
22191 * config/arm/arm-cores.def: Added core cortex-r7.
22192 * config/arm/arm-tune.md: Regenerated.
22193 * config/arm/arm-tables.opt: Regenerated.
22194 * doc/invoke.texi: Added entry for core cortex-r7.
22196 2013-03-27 Walter Lee <walt@tilera.com>
22198 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22199 double-decrement of next_scratch_regno.
22201 2013-03-27 Walter Lee <walt@tilera.com>
22203 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22205 (insn_v1mulus): Ditto.
22206 (insn_v2muls): Ditto.
22208 2013-03-27 Walter Lee <walt@tilera.com>
22210 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22211 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22213 2013-03-27 Walter Lee <walt@tilera.com>
22215 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22216 (*sibcall_value): Ditto.
22218 2013-03-27 Walter Lee <walt@tilera.com>
22220 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22221 (insn_mnz_v8qi): ... this ...
22222 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22224 (insn_v<n>mnz): Replaced by ...
22225 (insn_v1mnz): ... this ...
22226 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22228 (insn_mz_<mode>): Replaced by ...
22229 (insn_mz_v8qi): ... this ...
22230 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22232 (insn_v<n>mz): Replaced by ...
22233 (insn_v1mz): ... this ...
22234 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22237 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22239 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22241 2013-03-26 Roland McGrath <mcgrathr@google.com>
22243 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22244 than fprintf with a non-constant, non-format string.
22246 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22248 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22249 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22250 operand 0 predicate.
22251 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22252 attribute. Use general_x64nomem_operand as operand 1 predicate.
22253 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22254 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22255 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22256 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22257 (mov<mode>_insv_1): Remove expander. Merge insn with
22258 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22259 Use general_x64nomem_operand as operand 1 predicate.
22260 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22261 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22262 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22263 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22264 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22265 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22266 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22267 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22268 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22269 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22270 (general_x64nomem_operand): Ditto.
22272 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22274 * config/rtems.opt: Add -pthread option.
22276 2013-03-26 Richard Biener <rguenther@suse.de>
22278 * alias.c (find_base_term): Avoid redundant and not used recursion.
22279 (base_alias_check): Get the initial base term from the caller.
22280 (true_dependence_1): Compute and pass base terms to base_alias_check.
22281 (write_dependence_p): Likewise.
22282 (may_alias_p): Likewise.
22284 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22286 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22287 PC-relative load in SI modes and above only.
22289 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22291 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22292 * config/arm/iwmmxt.md (WCGR0): Update.
22293 (WCGR1, WCGR2, WCGR3): Likewise.
22295 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22297 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22298 Use x64 and nox64 isa attributes.
22300 2013-03-26 Richard Biener <rguenther@suse.de>
22302 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22303 alignment computations and rely on get_object_alignment_1
22304 for the !TYPE_P case.
22305 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22307 2013-03-26 Walter Lee <walt@tilera.com>
22309 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22310 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22312 2013-03-25 Jeff Law <law@redhat.com>
22314 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22315 check for INTEGRAL_TYPE_P that was missing due to checking in
22316 wrong version of prior patch.
22318 2013-03-25 Walter Lee <walt@tilera.com>
22320 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22321 TILEGX_INSN_SHUFFLEBYTES1.
22322 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22324 (tilegx_builtins): Ditto.
22325 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22327 2013-03-25 Walter Lee <walt@tilera.com>
22329 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22330 (floatunssisf2): New pattern.
22331 (floatsidf2): New pattern.
22332 (floatunssidf2): New pattern.
22334 2013-03-25 Walter Lee <walt@tilera.com>
22336 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22337 tests for constraint J, K, N, P.
22339 2013-03-25 Walter Lee <walt@tilera.com>
22341 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22342 Use indirect/pcrel encoding.
22343 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22346 2013-03-25 Steve Ellcey <sellcey@mips.com>
22348 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22349 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22350 * config/mips/mips.c (mips_option_override): Set IMADD default.
22351 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22352 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22353 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22354 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22355 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22357 2013-03-25 Jeff Law <law@redhat.com>
22359 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22360 slightly to avoid creating and folding useless trees. Simplify
22361 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22363 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22365 * config/i386/i386.md (*zero_extendsidi2): Merge with
22366 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22367 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22368 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22369 targets. Clarify comment.
22371 2013-03-25 Martin Jambor <mjambor@suse.cz>
22373 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22374 pass-through jump functions differently.
22375 (ipa_read_jump_function): Likewise. Also use setter functions to set
22378 2013-03-25 Martin Jambor <mjambor@suse.cz>
22380 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22381 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22383 (ipa_get_indirect_edge_target): New function.
22384 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22385 ipa_get_indirect_edge_target. Update all callers.
22386 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22387 ipa_get_indirect_edge_target_1 instead of calling
22388 ipa_get_indirect_edge_target.
22389 (create_specialized_node): Pass aggvlas to
22390 ipcp_discover_new_direct_edges.
22392 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22394 * config/arm/arm.md (f_sels, f_seld): New types.
22395 (*cmov<mode>): New pattern.
22396 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22399 2013-03-25 Kai Tietz <ktietz@redhat.com>
22401 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22402 POSIX-printf for mingw-hosted builds.
22404 2013-03-25 Richard Biener <rguenther@suse.de>
22406 PR middle-end/56694
22407 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22408 must-not-throw stmt location.
22410 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22412 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22413 Emit load-acquire versions when acq is true.
22414 (arm_emit_store_exclusive): Add rel parameter.
22415 Emit store-release versions when rel is true.
22416 (arm_split_compare_and_swap): Use acquire-release instructions
22418 of barriers when appropriate.
22419 (arm_split_atomic_op): Likewise.
22420 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22421 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22422 (VUNSPEC_SLX): Likewise.
22423 (VUNSPEC_LDA): Likewise.
22424 (VUNSPEC_STL): Likewise.
22425 * config/arm/sync.md (atomic_load<mode>): New pattern.
22426 (atomic_store<mode>): Likewise.
22427 (arm_load_acquire_exclusive<mode>): Likewise.
22428 (arm_load_acquire_exclusivesi): Likewise.
22429 (arm_load_acquire_exclusivedi): Likewise.
22430 (arm_store_release_exclusive<mode>): Likewise.
22432 2013-03-25 Catherine Moore <clm@codesourcery.com>
22434 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22435 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22436 * config/mip/predicates.md (lwsp_swsp_operand,
22437 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22438 sb16_operand, db4_operand, db7_operand, ib3_operand,
22439 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22440 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22441 andi16_operand): New predicates.
22442 * config/mips/mips.md (compression): New attribute.
22443 (enabled): New attribute.
22444 (length): Consider compression in computing length.
22445 (shift_compression): New code attribute.
22446 (*add<mode>3): New operands. Record compression.
22447 (sub<mode>3): Likewise.
22448 (one_cmpl<mode>2): Likewise.
22449 (*and<mode>3): Likewise.
22450 (*ior<mode>3): Likewise.
22451 (unnamed pattern for xor): Likewise.
22452 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22453 (*<optab><mode>3): Likewise.
22454 (*mov<mode>_internal: Likewise.
22455 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22456 (mips_unsigned_immediate_p): New.
22457 (umips_lwsp_swsp_address_p): New.
22458 (m16_based_address_p): New.
22459 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22460 (mips_unsigned_immediate_p): New prototype.
22461 (lwsp_swsp_address_p): New prototype.
22462 (m16_based_address_p): New prototype.
22463 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22464 (mips_signed_immediate_p): New function.
22465 (m16_based_address_p): New function.
22466 (lwsp_swsp_address_p): New function.
22467 (mips_print_operand_punctuation): Recognize short delay slot insns
22468 for microMIPS.add<mode>3"
22470 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22473 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22474 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22476 2013-03-25 Richard Biener <rguenther@suse.de>
22478 PR tree-optimization/56689
22479 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22482 2013-03-25 Richard Biener <rguenther@suse.de>
22484 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22486 (memory_references): Likewise.
22487 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22488 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22489 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22490 (gather_mem_refs_in_loops): Fold into ...
22491 (analyze_memory_references): ... this. Move initialization
22492 to tree_ssa_lim_initialize.
22493 (fill_always_executed_in): Rename to ...
22494 (fill_always_executed_in_1): ... this.
22495 (fill_always_executed_in): Move contains_call computation to
22496 this new function from ...
22497 (tree_ssa_lim_initialize): ... here.
22498 (tree_ssa_lim): Call fill_always_executed_in.
22500 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22502 * postreload.c (reload_combine): Fix code detecting returns.
22504 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22506 * function.c (emit_use_return_register_into_block): On cc0 targets,
22507 do not emit the sequence between cc0 setter and user.
22509 2013-03-25 Kai Tietz <ktietz@redhat.com>
22511 * config/i386/predicates.md (local_symbolic_operand): Interpret
22512 dll-imported symbols as none-local.
22514 2013-03-25 Richard Biener <rguenther@suse.de>
22516 * tree-ssa-loop-im.c (struct depend): Remove.
22517 (struct lim_aux_data): Make depends a vec of gimples.
22518 (free_lim_aux_data): Adjust.
22519 (add_dependency): Likewise.
22520 (set_level): Likewise.
22522 2013-03-25 Richard Biener <rguenther@suse.de>
22524 PR middle-end/56434
22525 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22526 the pointer returned by calls with ECF_MALLOC set.
22528 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22530 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22532 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22534 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22535 using MMXMODE mode iterator.
22536 (*move<mode>_internal): Merge with *movv2sf_internal and
22537 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22539 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22541 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22542 (record_last_mem_set_info): Likewise.
22544 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22545 of XNEWVEC followed by memset.
22546 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22548 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22550 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22551 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22552 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22553 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22554 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22555 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22556 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22557 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22558 BARRIER_P instead of GET_CODE.
22560 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22562 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22563 inaccuracy in the probing code.
22565 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22566 (ctrapdi4): Likewise.
22568 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22570 * calls.c (expand_call): Add missing guard to code handling return
22571 of non-BLKmode structures in MSB.
22572 * function.c (expand_function_end): Likewise.
22574 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22576 * combine.c (try_combine): Adjust comment. Do not add the set of
22577 insn #0 if the destination indirectly is set or dies in insn #2.
22578 Tidy up code to distribute a new note.
22580 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22582 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22583 also for alternatives 16 and 17.
22585 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22587 * config/i386/sse.md (*mov<mode>_internal): Merge with
22588 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22589 Emit insn template depending on type attribute. Use
22590 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22591 movd instead of movq mnemonic for interunit moves. Rewrite mode
22592 attribute calculation. Remove unit attribute calculation.
22593 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22594 Set prefix_data16 attribute for DImode ssemov types.
22595 Use Ym instead of y for SSE-MMX conversion alternatives.
22596 Reorder operand constraints.
22598 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22600 * df.h (df_insn_delete): Adjust prototype.
22601 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22602 and let it decide whether mark the basic block dirty.
22603 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22604 * df-scan.c (df_insn_info_delete): New helper function, split
22605 off from df_insn_delete.
22606 (df_scan_free_bb_info): Use it.
22607 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22609 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22610 that the insn is actually an insn and it has a non-NULL basic block.
22611 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22613 2013-03-22 Richard Biener <rguenther@suse.de>
22615 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22617 (mem_ref_alloc): Do not allocate them.
22618 (refs_independent_p): Do not query or maintain a cache.
22620 2013-03-22 Richard Biener <rguenther@suse.de>
22622 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22623 (gather_mem_refs_in_loops): Do not compute it.
22624 (analyze_memory_references): Do not allocate it.
22625 (tree_ssa_lim_finalize): Do not free it.
22626 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22628 2013-03-22 Richard Biener <rguenther@suse.de>
22630 * is-a.h (as_a): Use gcc_checking_assert.
22632 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22634 * config/aarch64/aarch64.c (aarch64_print_operand): New
22635 format specifier for printing a constant in hex.
22636 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22637 format specifier for printing second operand.
22639 2013-03-22 Richard Biener <rguenther@suse.de>
22641 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22643 (gather_mem_refs_in_loops): Perform store accumulation here.
22644 (create_vop_ref_mapping_loop): Remove.
22645 (create_vop_ref_mapping): Likewise.
22646 (analyze_memory_references): Initialize refs_stored_in_loop.
22647 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22648 (record_indep_loop): Remove.
22649 (record_dep_loop): New function.
22650 (ref_indep_loop_p_1): Adjust to only walk over references
22651 in the loop, not its subloops.
22652 (ref_indep_loop_p): Rename to ...
22653 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22654 maintaining a more fine-grained cache.
22655 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22656 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22658 2013-03-22 Richard Biener <rguenther@suse.de>
22660 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22661 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22662 aggregate mem_ref_loc.
22663 (free_mem_ref_locs): Inline into ...
22664 (memref_free): ... this and adjust.
22665 (mem_ref_alloc): Adjust.
22666 (mem_ref_locs_alloc): Remove.
22667 (record_mem_ref_loc): Adjust.
22668 (get_all_locs_in_loop): Rewrite into ...
22669 (for_all_locs_in_loop): ... this iterator.
22670 (rewrite_mem_ref_loc): New functor.
22671 (rewrite_mem_refs): Use for_all_locs_in_loop.
22672 (sm_set_flag_if_changed): New functor.
22673 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22674 (ref_always_accessed): New functor.
22675 (ref_always_accessed_p): Use for_all_locs_in_loop.
22677 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22679 * tree-pass.h (PROP_gimple_lvec): New.
22680 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22681 (init_optimization_passes): Move pass_lower_vector.
22682 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22684 (pass_lower_vector): Provide PROP_gimple_lvec.
22685 (pass_lower_vector_ssa): Likewise.
22686 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22688 2013-03-21 Mark Wielaard <mjw@redhat.com>
22690 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22692 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22694 * config/i386/i386.md (*movdi_internal): Disparage slightly
22695 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22696 conversion alternatives.
22698 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22700 PR middle-end/48087
22701 * diagnostic.def (DK_WERROR): New kind.
22702 * diagnostic.h (werrorcount): Define.
22703 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22704 promoted to DK_ERROR, increment DK_WERROR counter instead of
22706 * toplev.c (toplev_main): Call print_ignored_options even if
22707 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22708 even if just werrorcount is non-zero.
22711 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22713 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22714 (string_cst_pool_decl): New function.
22715 (optimize_one_addr_into_implicit_ptr): New function.
22716 (resolve_addr_in_expr): Optimize DWARF location expression
22717 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22718 which doesn't live in memory, but has DW_AT_location or
22719 DW_AT_const_value, or refers to a string literal, into
22720 DW_OP_GNU_implicit_pointer.
22721 (optimize_location_into_implicit_ptr): New function.
22722 (resolve_addr): If removing DW_AT_location of a variable because
22723 it was DW_OP_addr of address of the variable, but the variable doesn't
22724 live in memory, try to emit const value attribute for the initializer.
22726 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22728 * tree.h (VECTOR_TYPE_P): New macro.
22729 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22730 TYPE_MODE): Use it.
22731 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22732 VEC_COND_EXPR cannot be lvalues.
22733 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22735 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22737 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22738 Restrict the transformation to equal modes.
22740 2013-03-21 Richard Biener <rguenther@suse.de>
22742 PR tree-optimization/39326
22743 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22744 (MEM_ANALYZABLE): Adjust.
22745 (record_mem_ref_loc): Move bitmap ops ...
22746 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22747 unanalyzable refs, do not record locations for it.
22748 (analyze_memory_references): Allocate ref zero as shared
22750 (refs_independent_p): Do not test for unanalyzed mems here.
22751 (ref_indep_loop_p_1): Special-case disambiguation against
22752 the unanalyzed ref.
22753 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22755 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22757 * config/arm/arm-protos.h (tune_params): Add
22758 prefer_neon_for_64bits field.
22759 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22760 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22761 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22762 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22763 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22764 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22765 (arm_option_override): Handle -mneon-for-64bits new option.
22766 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22767 (prefer_neon_for_64bits): Declare new variable.
22768 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22769 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22770 (arch_enabled): Handle new arch types. Remove support for onlya8
22772 (one_cmpldi2): Use new arch names.
22773 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22774 * config/arm/arm.opt (mneon-for-64bits): Add option.
22775 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22776 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22777 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22779 * doc/invoke.texi (-mneon-for-64bits): Document.
22781 2013-03-21 Richard Biener <rguenther@suse.de>
22783 PR tree-optimization/39326
22784 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22785 (sort_bbs_in_loop_postorder_cmp): New function.
22786 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22788 2013-03-21 Richard Biener <rguenther@suse.de>
22790 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22791 (vect_insert_into_interleaving_chain): Likewise.
22792 (vect_drs_dependent_in_basic_block): Inline ...
22793 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22795 (vect_analyze_data_ref_dependence): ... here. Simplify.
22796 (vect_check_interleaving): Simplify.
22797 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22798 (vect_slp_analyze_data_ref_dependences): ... this new function.
22799 (dr_group_sort_cmp): New function.
22800 (vect_analyze_data_ref_accesses): Compute data-reference groups
22801 here instead of in vect_analyze_data_ref_dependence. Use
22802 a more efficient algorithm.
22803 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22804 vect_slp_analyze_data_ref_dependences. Call
22805 vect_analyze_data_ref_accesses earlier.
22806 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22807 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22808 (vect_slp_analyze_data_ref_dependences): New prototype.
22810 2013-03-21 Richard Biener <rguenther@suse.de>
22812 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22813 ref is stored in the loop.
22814 (find_refs_for_sm): Walk only over all stores.
22815 (store_motion_loop): Allocate from lim_bitmap_obstack.
22816 (store_motion): Likewise.
22818 2013-03-21 Richard Biener <rguenther@suse.de>
22820 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22821 Update virtual SSA form.
22823 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22825 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22826 * configure: Regenerate.
22827 * config.in: Regenerate.
22828 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22829 if !HAVE_LD_EH_FRAME_CIEV3.
22831 2013-03-21 Richard Biener <rguenther@suse.de>
22833 * tree-cfg.c (verify_expr_no_block): New function.
22834 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22835 nor DECL_VALUE_EXPR have locations with associated blocks.
22836 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22837 (clear_unused_block_pointer): Remove code dealing with
22838 blocks in DECL_DEBUG_EXPR locations.
22840 2013-03-21 Richard Biener <rguenther@suse.de>
22842 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22843 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22844 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22845 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22846 instead of DECL_DEBUG_EXPR_IS_FROM.
22847 * gimplify.c (gimplify_modify_expr): Likewise.
22848 * tree-cfg.c (verify_expr_location_1): Likewise.
22849 * tree-complex.c (create_one_component_var): Likewise.
22850 * tree-sra.c (create_access_replacement): Likewise.
22851 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22852 (clear_unused_block_pointer): Likewise.
22853 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22854 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22855 * var-tracking.c (var_debug_decl): Likewise.
22856 (track_expr_p): Likewise.
22857 * tree-inline.c (add_local_variables): Likewise. Set
22858 DECL_HAS_DEBUG_EXPR_P after copying it.
22859 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22860 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22862 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22865 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22866 * configure: Regenerate.
22867 * config.in: Regenerate.
22868 * config/i386/i386.md (*movdf_internal): Use
22869 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22870 movd instead of movq mnemonic for interunit moves.
22871 (*movdi_internal): Ditto.
22873 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22875 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22876 (abd<mode>_3): New pattern.
22877 (aba<mode>_3): New pattern.
22878 (fabd<mode>_3): New pattern.
22880 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22882 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22883 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22884 occurrence of REGISTER_PREFIX as its empty string.
22886 2013-03-20 Jeff Law <law@redhat.com>
22888 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22889 addititional equivalences for equality comparisons between an SSA_NAME
22890 and a constant where the SSA_NAME was set from a widening conversion.
22892 2013-03-20 Walter Lee <walt@tilera.com>
22894 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22896 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22898 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22899 depending on type attribute.
22900 (*movti_internal): Ditto.
22901 (*movtf_internal): Ditto.
22902 (*movxf_internal): Ditto.
22903 (*movdf_internal): Ditto.
22904 (*movsf_internal): Ditto.
22906 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22908 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22909 maybe_vex for sselog1 and ssemov types.
22910 (*movdi_internal): Reorder operand constraints.
22911 (*movsi_internal): Ditto. Set prefix attribute to
22912 maybe_vex for sselog1 and ssemov types.
22913 (*movtf_internal): Set prefix attribute to maybe_vex
22914 for sselog1 and ssemov types.
22915 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22916 DImode ssemov types. Reorder operand constraints.
22917 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22918 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22919 attribute for SImode ssemov types. Reorder operand constraints.
22921 2013-03-20 Martin Jambor <mjambor@suse.cz>
22923 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22924 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22926 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22928 * config/rs6000/predicates.md (indexed_address, update_address_mem
22929 update_indexed_address_mem): New predicates.
22930 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22931 attribute for load/store instructions.
22932 * config/rs6000/dfp.md (movsd_store): Likewise.
22933 (movsd_load): Likewise.
22934 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22935 (unnamed HI->DI extend define_insn): Likewise.
22936 (unnamed SI->DI extend define_insn): Likewise.
22937 (unnamed QI->SI extend define_insn): Likewise.
22938 (unnamed QI->HI extend define_insn): Likewise.
22939 (unnamed HI->SI extend define_insn): Likewise.
22940 (unnamed HI->SI extend define_insn): Likewise.
22941 (extendsfdf2_fpr): Likewise.
22942 (movsi_internal1): Likewise.
22943 (movsi_internal1_single): Likewise.
22944 (movhi_internal): Likewise.
22945 (movqi_internal): Likewise.
22946 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22947 attribute for load/store instructions.
22948 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22950 (mov<mode>_softfloat): Likewise.
22951 (mov<mode>_hardfloat32): Likewise.
22952 (mov<mode>_hardfloat64): Likewise.
22953 (mov<mode>_softfloat64): Likewise.
22954 (movdi_internal32): Likewise.
22955 (movdi_internal64): Likewise.
22956 (probe_stack_<mode>): Likewise.
22958 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
22960 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
22961 floating point, and decimal floating point to reload iterator.
22963 * config/rs6000/constraints.md (wl constraint): New constraints to
22964 return FLOAT_REGS if certain options are used to reduce the number
22965 of separate patterns that exist in the file.
22966 (wx constraint): Likewise.
22967 (wz constraint): Likewise.
22969 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
22970 -mdebug=reg, print wg, wl, wx, and wz constraints.
22971 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
22972 Initialize the reload functions for 64-bit binary/decimal floating
22974 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
22975 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
22976 create the buffer on the stack to overcome not having a 32-bit
22978 (rs6000_emit_move): Likewise.
22979 (rs6000_secondary_memory_needed_rtx): Likewise.
22980 (rs6000_alloc_sdmode_stack_slot): Likewise.
22981 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
22982 via xxlxor, just like DFmode 0.0.
22984 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
22985 define as 1 if we are running on a power7 or newer.
22986 (enum r6000_reg_class_enum): Add new constraints.
22988 * config/rs6000/dfp.md (movsd): Delete, combine with binary
22989 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
22990 with other moves by using conditional constraits (wg). Use LFIWZX
22991 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
22992 (movsd splitter): Likewise.
22993 (movsd_hardfloat): Likewise.
22994 (movsd_softfloat): Likewise.
22996 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
22997 binary and decimal floating point moves.
22998 (fmove_ok): New attributes to combine binary and decimal floating
22999 point moves, and to combine power6x (mfpgpr) moves along normal
23001 (real_value_to_target): Likewise.
23002 (f32_lr): Likewise.
23003 (f32_lm): Likewise.
23004 (f32_li): Likewise.
23005 (f32_sr): Likewise.
23006 (f32_sm): Likewise.
23007 (f32_si): Likewise.
23008 (movsf): Combine binary and decimal floating point moves. Combine
23009 power6x (mfpgpr) moves with other moves by using conditional
23010 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
23011 (mov<mode> for SFmode/SDmode); Likewise.
23012 (SFmode/SDmode splitters): Likewise.
23013 (movsf_hardfloat): Likewise.
23014 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
23015 (movsf_softfloat): Likewise.
23016 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
23018 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
23019 wx and wz constraints.
23021 * config/rs6000/constraints.md (wg constraint): New constraint to
23022 return FLOAT_REGS if -mmfpgpr (power6x) was used.
23024 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
23027 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23028 -mdebug=reg, print wg, wl, wx, and wz constraints.
23029 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23030 Initialize the reload functions for 64-bit binary/decimal floating
23032 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23033 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23034 create the buffer on the stack to overcome not having a 32-bit
23036 (rs6000_emit_move): Likewise.
23037 (rs6000_secondary_memory_needed_rtx): Likewise.
23038 (rs6000_alloc_sdmode_stack_slot): Likewise.
23039 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23040 via xxlxor, just like DFmode 0.0.
23042 * config/rs6000/dfp.md (movdd): Delete, combine with binary
23043 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23044 with other moves by using conditional constraits (wg). Use LFIWZX
23045 and STFIWX for loading SDmode on power7.
23046 (movdd splitters): Likewise.
23047 (movdd_hardfloat32): Likewise.
23048 (movdd_softfloat32): Likewise.
23049 (movdd_hardfloat64_mfpgpr): Likewise.
23050 (movdd_hardfloat64): Likewise.
23051 (movdd_softfloat64): Likewise.
23053 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
23054 64-bit binary and decimal floating point moves.
23055 (FMOVE64X): Likewise.
23056 (movdf): Combine 64-bit binary and decimal floating point moves.
23057 Combine power6x (mfpgpr) moves with other moves by using
23058 conditional constraits (wg).
23059 (mov<mode> for DFmode/DDmode): Likewise.
23060 (DFmode/DDmode splitters): Likewise.
23061 (movdf_hardfloat32): Likewise.
23062 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
23063 (movdf_softfloat32): Likewise.
23064 (movdf_hardfloat64_mfpgpr): Likewise.
23065 (movdf_hardfloat64): Likewise.
23066 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
23067 (movdf_softfloat64): Likewise.
23068 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
23069 (reload_<mode>_load): Move to later in the file so they aren't in
23070 the middle of the floating point move insns.
23071 (reload_<mode>_store): Likewise.
23073 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
23076 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
23077 constraint if -mdebug=reg.
23078 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
23079 Enable using dd reload support if needed.
23081 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
23082 binary and decimal floating point moves in rs6000.md.
23083 (movtd_internal): Likewise.
23085 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
23086 decimal floating point moves.
23088 (movtf_internal): Likewise.
23089 (mov<mode>_internal, TDmode/TFmode): Likewise.
23090 (movtf_softfloat): Likewise.
23091 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
23093 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
23094 movdi_internal64, using wg constraint for move direct operations.
23095 (movdi_internal64): Likewise.
23097 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
23098 MODES_TIEABLE_P for selected modes. Print the numerical value of
23099 the various virtual registers. Use GPR/FPR first/last values,
23100 instead of hard coding the register numbers. Print which modes
23101 have reload functions registered.
23102 (rs6000_option_override_internal): If -mdebug=reg, trace the options
23103 settings before/after setting cpu, target and subtarget settings.
23104 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
23105 and for secondary reload failures in rs6000_secondary_reload_inner.
23106 (rs6000_secondary_reload_fail): Likewise.
23107 (rs6000_secondary_reload_inner): Likewise.
23109 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
23110 macros for first/last GPR and FPR registers.
23111 (LAST_GPR_REGNO): Likewise.
23112 (FIRST_FPR_REGNO): Likewise.
23113 (LAST_FPR_REGNO): Likewise.
23115 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
23116 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
23117 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
23118 (vcond<mode><mode>): Likewise.
23119 (vcondu<mode><mode>): Likewise.
23120 (vector_gtu<mode>): Likewise.
23121 (vector_gte<mode>): Likewise.
23122 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
23123 to prevent the compiler from converting DImode operations to TImode.
23124 (ior<mode>3): Likewise.
23125 (and<mode>3): Likewise.
23126 (one_cmpl<mode>2): Likewise.
23127 (nor<mode>3): Likewise.
23128 (andc<mode>3): Likewise.
23130 * config/rs6000/constraints.md (wt constraint): New constraint
23131 that returns VSX_REGS if TImode is allowed in VSX registers.
23133 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23134 constant under VSX.
23136 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23137 similar to TImode, but it is restricted to being in the GPRs.
23139 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23140 TImode to occupy a single VSX register.
23142 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23143 -mvsx-timode for power7/power8.
23144 (power7 cpu): Likewise.
23145 (power8 cpu): Likewise.
23147 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23148 sure that TFmode/TDmode take up two registers if they are ever
23149 allowed in the upper VSX registers.
23150 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23152 (rs6000_init_hard_regno_mode_ok): Likewise.
23153 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23154 constraint. Print if LRA is turned on.
23155 (rs6000_option_override_internal): Give an error if -mvsx-timode
23156 and VSX is not enabled.
23157 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23158 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23159 to reg+offset addressing. Use PTImode when checking offset
23160 addresses for validity.
23161 (reg_offset_addressing_ok_p): Likewise.
23162 (rs6000_legitimate_offset_address_p): Likewise.
23163 (rs6000_legitimize_address): Likewise.
23164 (rs6000_legitimize_reload_address): Likewise.
23165 (rs6000_legitimate_address_p): Likewise.
23166 (rs6000_eliminate_indexed_memrefs): Likewise.
23167 (rs6000_emit_move): Likewise.
23168 (rs6000_secondary_reload): Likewise.
23169 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23170 reloads to fpr registers to continue to use reg+offset addressing,
23171 but 64-bit reloads to altivec registers need reg+reg addressing.
23172 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23173 it. Treat LO_SUM like a PLUS operation.
23174 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23175 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23176 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23177 registers to share a register with a smaller sized type, since VSX
23178 puts scalars in the upper 64-bits.
23179 (print_operand): Add support for PTImode.
23180 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23181 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23182 registers, but don't have arithmetic support.
23183 (rs6000_memory_move_cost): Add test for VSX.
23184 (rs6000_opt_masks): Add -mvsx-timode.
23186 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23189 (VSr): Use wt constraint for TImode.
23190 (VSv): Drop TImode support.
23191 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23192 (vsx_movti_64bit): Likewise.
23193 (vsx_movti_32bit): Likewise.
23194 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23195 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23196 one '?' on the appropriate output constraint. Do not allow TImode
23197 logical operations on 32-bit systems.
23198 (vsx_ior<mode>3): Likewise.
23199 (vsx_xor<mode>3): Likewise.
23200 (vsx_one_cmpl<mode>2): Likewise.
23201 (vsx_nor<mode>3): Likewise.
23202 (vsx_andc<mode>3): Likewise.
23203 (vsx_concat_<mode>): Likewise.
23204 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23206 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23207 OPTION_MASK_VSX_TIMODE.
23208 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23209 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23211 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23212 (TI2 iterator): New iterator for TImode, PTImode.
23213 (wd mode attribute): Add values for vector types.
23214 (movti_string): Replace TI move operations with operations for TImode
23215 and PTImode. Add support for TImode being allowed in VSX registers.
23216 (mov<mode>_string, TImode/PTImode): Likewise.
23217 (movti_ppc64): Likewise.
23218 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23219 (TI mode splitters): Likewise.
23221 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23224 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23226 PR tree-optimization/56355
23227 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23228 Also handle integers with undefined overflow.
23230 2013-03-20 Catherine Moore <clm@codesourcery.com>
23231 Maciej W. Rozycki <macro@codesourcery.com>
23232 Tom de Vries <tom@codesourcery.com>
23233 Nathan Sidwell <nathan@codesourcery.com>
23234 Iain Sandoe <iain@codesourcery.com>
23235 Nathan Froyd <froydnj@codesourcery.com>
23236 Chao-ying Fu <fu@mips.com>
23238 * doc/extend.texi: (micromips, nomicromips, nocompression):
23239 Document new function attributes.
23240 * doc/invoke.texi (minterlink-compressed, mmicromips,
23241 m14k, m14ke, m14kec): Document new options.
23242 (minterlink-mips16): Update documentation.
23243 * doc/md.texi (ZC, ZD): Document new constraints.
23244 * configure.ac (gcc_cv_as_micromips): Check if linker
23245 supports the .set micromips directive.
23246 * configure: Regenerate.
23247 * config.in: Regenerate.
23248 * config/mips/mips-tables.opt: Regenerate.
23249 * config/mips/micromips.md: New file.
23250 * constraints.md (ZC, ZD): New constraints.
23251 * config/mips/predicates.md (movep_src_register): New predicate.
23252 (movep_src_operand): New predicate.
23253 (non_volatile_mem_operand): New predicate.
23254 * config/mips/mips.md (multimem): New type.
23255 (length): Differentiate between 17-bit and 18-bit branch offsets.
23256 (MOVEP1, MOVEP2): New mode iterator.
23257 (mov_<load>l): Use ZC constraint.
23258 (mov_<load>r): Likewise.
23259 (mov_<store>l): Likewise.
23260 (mov_<store>r): Likewise.
23261 (*branch_equality<mode>_inverted): Add microMIPS support.
23262 (*branch_equality<mode>): Likewise.
23263 (*jump_absolute): Likewise.
23264 (indirect_jump_<mode>): Likewise.
23265 (tablejump_<mode>): Likewise.
23266 (<optab>_internal): Likewise.
23267 (sibcall_internal): Likewise.
23268 (sibcall_value_internal): Likewise.
23269 (prefetch): Use constraint ZD.
23270 * config/mips/mips.opt (minterlink-compressed): New option.
23271 (minterlink-mips16): Now an alias for minterlink-compressed.
23272 (mmicromips): New option.
23273 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23274 (compare_and_swap_12): Likewise.
23275 (sync_add<mode>): Likewise.
23276 (sync_<optab>_12): Likewise.
23277 (sync_old_<optab>_12): Likewise.
23278 (sync_new_<optab>_12): Likewise.
23279 (sync_nand_12): Likewise.
23280 (sync_old_nand_12): Likewise.
23281 (sync_new_nand_12): Likewise.
23282 (sync_sub<mode>): Likewise.
23283 (sync_old_add<mode>): Likewise.
23284 (sync_old_sub<mode>): Likewise.
23285 (sync_new_add<mode>): Likewise.
23286 (sync_new_sub<mode>): Likewise.
23287 (sync_<optab><mode>): Likewise.
23288 (sync_old_<optab><mode>): Likewise.
23289 (sync_new_<optab><mode>): Likewise.
23290 (sync_nand<mode>): Likewise.
23291 (sync_old_nand<mode>): Likewise.
23292 (sync_new_nand<mode>): Likewise.
23293 (sync_lock_test_and_set<mode>): Likewise.
23294 (test_and_set_12): Likewise.
23295 (atomic_compare_and_swap<mode>): Likewise.
23296 (atomic_exchange<mode>_llsc): Likewise.
23297 (atomic_fetch_add<mode>_llsc): Likewise.
23298 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23299 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23300 (umips_save_restore_pattern_p): Likewise.
23301 (umips_load_store_pair_p): Likewise.
23302 (umips_output_load_store_pair): Likewise.
23303 (umips_movep_target_p): Likewise.
23304 (umips_12bit_offset_address_p): Likewise.
23305 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23306 (mips_base_mips16): Rename this...
23307 (mips_base_compression_flags): ...to this. Update all uses.
23308 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23309 (mips_mips16_decl_p): Delete.
23310 (mips_nomips16_decl_p): Delete.
23311 (mips_get_compress_on_flags): New function.
23312 (mips_get_compress_off_flags): New function.
23313 (mips_get_compress_mode): New function.
23314 (mips_get_compress_on_name): New function.
23315 (mips_get_compress_off_name): New function.
23316 (mips_insert_attributes): Support multiple compression types.
23317 (mips_merge_decl_attributes): Likewise.
23318 (umips_12bit_offset_address_p): New function.
23319 (mips_start_function_definition): Emit .set micromips directive.
23320 (mips_call_may_need_jalx_p): New function.
23321 (mips_function_ok_for_sibcall): Add microMIPS support.
23322 (mips_print_operand_punctuation): Support short delay slots and
23324 (umips_swm_mask, umips_swm_encoding): New.
23325 (umips_build_save_restore): New function.
23326 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23327 (was_mips16_p): Remove.
23328 (old_compression_mode): New.
23329 (mips_set_compression_mode): New function.
23330 (mips_set_current_function): Add microMIPS support.
23331 (mips_option_override): Likewise.
23332 (umips_save_restore_pattern_p): New function.
23333 (umips_output_save_restore): New function.
23334 (umips_load_store_pair_p_1): New function.
23335 (umips_load_store_pair_p): New function.
23336 (umips_output_load_store_pair_1): New function.
23337 (umips_output_load_store_pair): New function.
23338 (umips_movep_target_p) New function.
23339 (mips_prepare_pch_save): Add microMIPS support.
23340 * config/mips/mips.h (TARGET_COMPRESSION): New.
23341 (TARGET_CPU_CPP_BUILTINS): Update macro
23342 to use new compression flags and to support microMIPS.
23343 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23344 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23345 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23346 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23347 (ASM_SPEC): Support mmicromips and mno-micromips.
23348 (M16STORE_REG_P): New macro.
23349 (MIPS_CALL): Support TARGET_MICROMIPS.
23350 (MICROMIPS_J): New macro.
23351 (mips_base_mips16): Rename this...
23352 (mips_base_compression_flags): ...to this.
23353 (UMIPS_12BIT_OFFSET_P): New macro.
23354 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23355 (MULTILIB_DIRNAMES): Likewise.
23356 2013-03-20 Richard Biener <rguenther@suse.de>
23358 PR tree-optimization/56661
23359 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23360 the result does not have to be distinct.
23362 2013-03-20 Richard Biener <rguenther@suse.de>
23364 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23367 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23368 Steven Bosscher <steven@gcc.gnu.org>
23370 PR rtl-optimization/56605
23371 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23373 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23376 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23377 that require movd instead of movq.
23379 2013-03-20 Richard Biener <rguenther@suse.de>
23381 * tree-ssa-structalias.c (struct variable_info): Add pointer
23382 to the first field of an aggregate with sub-vars. Make
23383 this and the pointer to the next subfield its ID.
23384 (vi_next): New function.
23385 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23386 storedanything_id, integer_id): Increment by one.
23387 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23388 get_call_clobber_vi): Adjust.
23389 (solution_set_expand): Simplify and speedup.
23390 (solution_set_add): Inline into ...
23391 (set_union_with_increment): ... this. Adjust accordingly.
23392 (do_sd_constraint): Likewise.
23393 (do_ds_constraint): Likewise.
23394 (do_complex_constraint): Simplify.
23395 (build_pred_graph): Adjust.
23396 (solve_graph): Likewise. Simplify and speedup.
23397 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23398 get_constraint_for_component_ref, get_constraint_for_1,
23399 first_vi_for_offset, first_or_preceding_vi_for_offset,
23400 create_function_info_for, create_variable_info_for_1,
23401 create_variable_info_for, intra_create_variable_infos): Adjust.
23402 (init_base_vars): Push NULL for ID zero.
23403 (compute_points_to_sets): Adjust.
23405 2013-03-20 Richard Biener <rguenther@suse.de>
23407 * cfgloop.c (verify_loop_structure): Streamline and avoid
23408 ICEing on corrupt loop tree.
23409 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23412 2013-03-20 Richard Biener <rguenther@suse.de>
23414 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23415 check whether an SSA update is needed.
23417 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23419 * config/mips/constraints.md (T): Rename to...
23423 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23424 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23426 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23428 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23429 (*subsi3_carryin_uxtw): Likewise.
23431 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23433 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23434 (*rorsi3_insn_uxtw): Likewise.
23436 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23438 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23439 (*extrsi5_insn_uxtw): Likewise.
23441 2013-03-19 Richard Biener <rguenther@suse.de>
23443 PR tree-optimization/56273
23444 * passes.c (init_optimization_passes): Move second VRP after DOM.
23446 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23448 * config/i386/i386.md (*movti_internal): Merge from
23449 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23450 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23451 nox64 isa attributes.
23453 2013-03-18 Richard Biener <rguenther@suse.de>
23455 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23457 (merge_node_constraints): Likewise.
23458 (build_succ_graph): Likewise.
23459 (valid_graph_edge): Inline into single caller.
23460 (unify_nodes): Likewise. Use bitmap_set_bit return value
23462 (scc_visit): Fix formatting and variable use.
23463 (do_sd_constraint): Use gcc_checking_assert.
23464 (do_ds_constraint): Likewise.
23465 (do_complex_constraint): Likewise.
23466 (condense_visit): Likewise. Cleanup.
23467 (dump_pred_graph): New function.
23468 (perform_var_substitution): Dump the pred-graph before
23469 variable substitution.
23470 (find_equivalent_node): Use gcc_checking_assert.
23471 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23473 2013-03-18 Richard Biener <rguenther@suse.de>
23475 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23476 Remove cond_expr_stmt_list argument and do not gimplify the
23478 (vect_loop_versioning): Adjust.
23479 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23480 Cleanup to use less temporaries.
23481 (vect_create_data_ref_ptr): Cleanup.
23483 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23485 PR tree-optimization/56635
23486 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23487 require types_compatible_p types.
23489 2013-03-18 Nick Clifton <nickc@redhat.com>
23491 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23492 spurious backslash.
23494 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23495 Add missing line to comment describing function.
23497 2013-03-18 Richard Biener <rguenther@suse.de>
23499 PR tree-optimization/56210
23500 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23501 Handle string / character search functions.
23502 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23504 2013-03-18 Richard Biener <rguenther@suse.de>
23506 PR middle-end/56483
23507 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23508 and implement properly.
23509 * gimple.h (gimple_cond_single_var_p): Remove.
23511 2013-03-18 Richard Biener <rguenther@suse.de>
23513 * tree-data-ref.h (find_data_references_in_loop): Declare.
23514 * tree-data-ref.c (get_references_in_stmt): Use a stack
23515 vector pre-allocated in the callers.
23516 (find_data_references_in_stmt): Adjust.
23517 (graphite_find_data_references_in_stmt): Likewise.
23518 (create_rdg_vertices): Likewise.
23519 (find_data_references_in_loop): Export.
23520 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23521 Compute dependences here...
23522 (vect_analyze_data_refs): ...not here. When we encounter
23523 a non-vectorizable data reference in basic-block vectorization
23524 truncate the data reference vector. Do not bother to
23525 fixup data-dependence information for gather loads.
23526 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23527 of data references, as reported.
23529 2013-03-18 Richard Biener <rguenther@suse.de>
23531 PR tree-optimization/3713
23532 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23533 has_constants and expr.
23534 (stmt_has_constants): Properly valueize SSA names when deciding
23535 whether the stmt has constants.
23537 2013-03-18 Richard Biener <rguenther@suse.de>
23539 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23540 whole function when there is nothing to do.
23541 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23542 * tree-vectorizer.c (vectorize_loops): Update virtual and
23543 loop-closed SSA once.
23544 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23546 2013-03-18 Richard Biener <rguenther@suse.de>
23548 PR middle-end/56113
23549 * domwalk.c (bb_postorder): New global static.
23550 (cmp_bb_postorder): New function.
23551 (walk_dominator_tree): Replace scheme imposing an order for
23552 visiting dominator sons by one sorting them at the time they
23553 are pushed on the stack.
23555 2013-03-18 Richard Biener <rguenther@suse.de>
23557 PR tree-optimization/39326
23558 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23559 (struct mem_ref): Replace mem member with ao_ref typed member.
23560 (MEM_ANALYZABLE): Adjust.
23561 (memref_eq): Likewise.
23562 (mem_ref_alloc): Likewise.
23563 (gather_mem_refs_stmt): Likewise.
23564 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23565 (execute_sm_if_changed_flag_set): Adjust.
23566 (execute_sm): Likewise.
23567 (ref_always_accessed_p): Likewise.
23568 (refs_independent_p): Likewise.
23569 (can_sm_ref_p): Likewise.
23571 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23574 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23575 return 1 even for !unsignedp.
23577 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23579 * config/i386/i386.md (isa): Add x64 and nox64.
23580 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23581 (*pushtf): Enable *roF alternative for x64 isa only.
23582 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23583 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23584 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23585 (*movtf_internal): Merge from *movtf_internal_rex64 and
23586 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23587 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23588 nox64 isa attributes.
23589 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23590 nox64 isa attributes.
23591 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23593 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23595 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23597 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23600 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23601 class names. Remove trailing comma after "ALL_REGS".
23603 2013-03-16 Jan Hubicka <jh@suse.cz>
23605 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23606 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23607 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23608 of cgraph_get_create_node.
23609 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23611 2013-03-16 Jason Merrill <jason@redhat.com>
23614 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23615 with DW_AT_default_value.
23617 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23619 * BASE-VER: Set to 4.9.0.
23621 2013-03-14 Andi Kleen <ak@linux.intel.com>
23624 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23625 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23626 Document _x* TSX intrinsics.
23628 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23629 David Holsgrove <david.holsgrove@xilinx.com>
23631 * configure.ac: Add MicroBlaze TLS support detection.
23632 * configure: Regenerate.
23633 * config/microblaze/microblaze-protos.h
23634 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23635 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23636 * config/microblaze/microblaze.c (microblaze_address_type): Add
23637 ADDRESS_TLS and tls_reloc address types.
23638 (microblaze_address_info): Add tls_reloc.
23639 (TARGET_HAVE_TLS): Define.
23640 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23641 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23642 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23643 load_tls_operand, microblaze_call_tls_get_addr,
23644 microblaze_legitimize_tls_address): New functions.
23645 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23646 (get_base_reg): Use microblaze_tls_symbol_p.
23647 (microblaze_classify_address): Handle TLS.
23648 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23649 label_mentioned_p and microblaze_tls_referenced_p.
23650 (microblaze_legitimize_address): Handle TLS.
23651 (microblaze_address_insns): Handle ADDRESS_TLS.
23652 (pic_address_needs_scratch): Handle TLS.
23653 (print_operand_address): Handle TLS.
23654 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23655 (microblaze_expand_move): Handle TLS.
23656 (microblaze_legitimate_constant_p): Check
23657 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23658 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23659 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23660 (PIC_OFFSET_TABLE_REGNUM): Set.
23661 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23662 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23663 (addsi3, movsi_internal2, movdf_internal): Update constraints
23664 * config/microblaze/predicates.md (arith_plus_operand): Define
23665 (move_operand): Redefine as move_src_operand,
23666 check microblaze_tls_referenced_p.
23668 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23670 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23671 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23673 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23675 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23678 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23680 PR tree-optimization/53265
23681 * common.opt (Waggressive-loop-optimizations): New option.
23682 * tree-ssa-loop-niter.c: Include tree-pass.h.
23683 (do_warn_aggressive_loop_optimizations): New function.
23684 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23685 if number_of_latch_executions returned constant.
23686 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23687 early. If number_of_latch_executions returned constant, set
23688 nb_iterations_upper_bound back to it.
23689 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23691 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23692 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23694 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23695 (MULTILIB_OSDIRNAMES): Set.
23696 * genmultilib: If defaultosdirname doesn't start with :: , set
23697 defaultosdirname2 instead, clear it and emit two . multilib_raw
23698 entries instead of just one.
23700 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23702 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23703 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23704 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23705 (SUBTARGET_OVERRIDE_OPTIONS): New.
23707 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23710 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23711 (musermode): Convert to Var(TARGET_USERMODE).
23712 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23713 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23714 * config/sh/sh.c (sh_option_override): Use
23715 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23716 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23718 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23720 (udivsi3_i4_single, divsi3_i4_single): Use
23721 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23723 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
23725 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23728 2013-03-13 Richard Biener <rguenther@suse.de>
23730 PR tree-optimization/56608
23731 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23732 calls when vectorizing basic-blocks.
23734 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23737 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23740 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23742 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23744 2013-03-11 Jan Hubicka <jh@suse.cz>
23747 * lto-streamer-out.c (output_symbol_p): Skip references from
23748 constructors of external variables.
23750 2013-03-11 Jan Hubicka <jh@suse.cz>
23752 PR middle-end/56571
23753 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23755 * emit-rtl.c (verify_rtx_sharing): Likewise.
23756 (copy_insn_1): Likewise.
23757 * rtl.c (copy_rtx): Likewise.
23759 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23762 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23763 output_operand_lossage message.
23765 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23768 * arm.c (shift_op): Validate RTL pattern on the fly.
23769 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23772 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23775 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23776 new call patterns that clobber %fr12.
23777 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23778 split and postreload patterns.
23779 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23780 registers %fr12 and %fr12R as call used.
23782 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23784 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23785 (canon_address, record_store, replace_read, check_mem_read_rtx,
23786 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23787 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23788 rest_of_handle_dse): Likewise.
23790 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23792 PR middle-end/56524
23793 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23795 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23796 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23797 (save_optabs_if_changed): Replace with...
23798 (init_tree_optimization_optabs): ...this.
23799 * optabs.c (save_optabs_if_changed): Rename to...
23800 (init_tree_optimization_optabs): ...this. Take the optimization node
23801 as argument. Do nothing if the base optabs are already correct.
23802 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23803 to recompute optabs.
23804 * function.h (function): Remove optabs field.
23805 * function.c (invoke_set_current_function_hook): Call
23806 init_tree_optimization_optabs. Use the result to initialize
23809 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23811 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23812 if GTMA_HAS_NO_INSTRUMENTATION.
23813 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23814 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23815 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23816 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23817 GTMA_HAS_NO_INSTRUMENTATION.
23819 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23821 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23824 2013-03-08 Marek Polacek <polacek@redhat.com>
23825 Jakub Jelinek <jakub@redhat.com>
23827 PR tree-optimization/56478
23828 * predict.c (is_comparison_with_loop_invariant_p): Change the
23829 type of loop_step to tree.
23830 (predict_loops): Adjust.
23831 (predict_iv_comparison): Perform the computations on double_ints.
23833 2013-03-08 Richard Biener <rguenther@suse.de>
23835 PR tree-optimization/56570
23836 * tree-cfg.c (verify_expr_location_1): Verify locations for
23838 * tree-sra.c (create_access_replacement): Strip locations
23839 from DECL_DEBUG_EXPRs.
23841 2013-03-08 Richard Biener <rguenther@suse.de>
23843 * tree-inline.c (expand_call_inline): Do not associate
23844 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23845 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23847 2013-03-08 Richard Biener <rguenther@suse.de>
23849 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23850 or block changes with -Og. Fix for location / block encoding
23851 changes and PHI arguments with locations.
23853 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23855 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23857 (struct output_info): Likewise.
23858 (register_overhead): Remove bad gcc_assert.
23859 (bitmap_find_bit): If there is only a single bitmap element, do not
23860 count a miss as a search.
23861 (print_statistics): Update for counter type changes.
23862 (dump_bitmap_statistics): Likewise. Print headers such that they
23863 are properly lined up with the printed counters.
23865 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23867 PR tree-optimization/56559
23868 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23869 check that it has only a single use.
23871 2013-03-07 Richard Biener <rguenther@suse.de>
23873 * doc/invoke.texi (fwhole-program): Discourage use in combination
23876 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23878 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23880 PR tree-optimization/56539
23881 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23882 instead of GSI_CONTINUE_LINKING as last argument to
23883 force_gimple_operand_gsi. Adjust function comment.
23885 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23888 PR middle-end/56548
23889 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23890 promoted mode, convert the result back to the original mode.
23892 2013-03-06 Richard Biener <rguenther@suse.de>
23894 PR middle-end/56294
23895 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23896 (insert_updated_phi_nodes_compare_uids): New function.
23897 (update_ssa): Sort symbols_to_rename after UID before
23898 traversing it to insert PHI nodes.
23900 2013-03-06 Richard Biener <rguenther@suse.de>
23902 PR middle-end/50494
23903 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23904 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23907 2013-02-13 Richard Biener <rguenther@suse.de>
23910 * varasm.c (output_constant_def_1): Get the decl representing
23911 the constant as argument.
23912 (output_constant_def): Wrap output_constant_def_1.
23913 (make_decl_rtl): Use output_constant_def_1 with the decl
23914 representing the constant.
23915 (build_constant_desc): Optionally re-use a decl already
23916 representing the constant.
23917 (tree_output_constant_def): Adjust.
23919 2013-03-06 Joey Ye <joey.ye@arm.com>
23922 * gcc.c (convert_white_space): New function.
23923 (main): Handles white space in function name.
23925 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23928 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23929 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23930 to SH_DIV_CALL_TABLE for TARGET_SH2.
23931 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23933 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23934 call-table options.
23936 2013-03-05 Sterling Augustine <saugustine@google.com>
23937 Cary Coutant <ccoutant@google.com>
23940 * dwarf2out.c (resolve_addr): Don't call
23941 remove_loc_list_addr_table_entries a second time for the same
23944 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23947 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23948 (avoid_complex_debug_insns): New function.
23949 (expand_debug_locations): Call it.
23951 PR rtl-optimization/56484
23952 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23953 lifetimes of hard registers on small register class machines.
23955 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23957 * config/microblaze/microblaze-protos.h: Rename
23958 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
23959 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
23961 (microblaze_fast_interrupt_function_p): New function.
23962 (microblaze_is_interrupt_handler): Rename to
23963 microblaze_is_interrupt_variant and add fast_interrupt check.
23964 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
23965 (save_restore_insns): Likewise.
23966 (compute_frame_size): Likewise.
23967 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
23968 (microblaze_globalize_label): Likewise.
23969 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
23970 * config/microblaze/microblaze.md: Use wrapper
23971 microblaze_is_interrupt_variant.
23973 2013-03-05 Kai Tietz <ktietz@redhat.com>
23975 * sdbout.c (sdbout_one_type): Switch to current function's section
23976 supporting cold/hot.
23978 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
23980 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
23983 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23985 PR middle-end/56461
23986 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
23987 if VALGRIND_GET_VBITS is defined, temporarily make object
23988 memory all defined, and restore previous valgrind addressability
23989 and definability afterwards. Free this_object at the end.
23991 PR middle-end/56461
23992 * lra.c (lra): Call lra_clear_live_ranges if live_p,
23993 right before calling lra_create_live_ranges, also call it
23994 when clearing live_p. Only call lra_clear_live_ranges
23995 at the end if live_p.
23997 PR middle-end/56461
23998 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
24000 2013-03-05 Richard Biener <rguenther@suse.de>
24002 PR tree-optimization/56521
24003 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
24006 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
24009 * except.h (remove_unreachable_eh_regions): New prototype.
24010 * except.c (remove_eh_handler_splicer): New function, split out
24011 of remove_eh_handler.
24012 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
24013 warning about running it on many EH regions one at a time.
24014 (remove_unreachable_eh_regions_worker): New function, walk the
24015 EH tree in depth-first order and remove non-marked regions.
24016 (remove_unreachable_eh_regions): New function.
24017 * tree-eh.c (mark_reachable_handlers): New function, split out
24018 from remove_unreachable_handlers.
24019 (remove_unreachable_handlers): Use mark_reachable_handlers and
24020 remove_unreachable_eh_regions.
24021 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
24022 and remove_unreachable_eh_regions.
24024 2013-03-05 Richard Biener <rguenther@suse.de>
24026 PR middle-end/56525
24027 * loop-init.c (fix_loop_structure): Remove loops in two stages,
24028 not freeing them until the end.
24030 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24032 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
24034 2013-03-05 Richard Biener <rguenther@suse.de>
24036 PR tree-optimization/56270
24037 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
24038 of loads after scheduling an SLP instance.
24040 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24042 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
24044 (check_gcc_parallelize): Run guality.exp as a separate job from
24045 vect.exp with unsorted.exp and $(dg_target_exps) separately from
24046 struct-layout-1.exp with stackalign.exp.
24048 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
24050 PR middle-end/56461
24051 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
24052 load_index sbitmap even if some bit in it isn't set.
24054 PR middle-end/56461
24055 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
24056 (discover_iteration_bound_by_body_walk): Change queues to
24057 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
24058 spelling in comment. Call safe_push on queues[bound_index] directly.
24059 Release queues[queue_index] in every iteration unconditionally.
24060 Release bounds vector.
24062 PR middle-end/56461
24063 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
24064 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
24065 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
24068 2013-03-05 Richard Biener <rguenther@suse.de>
24071 * tree-inline.c (remap_blocks_to_null): New function.
24072 (expand_call_inline): When expanding a call stmt without
24073 an associated block inline remap all callee blocks to NULL.
24075 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24077 PR rtl-optimization/56494
24078 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
24079 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
24080 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
24082 PR middle-end/56461
24083 * sel-sched-ir.c (free_sched_pools): Release
24084 succs_info_pool.stack[succs_info_pool.max_top] vectors too
24085 if succs_info_pool.max_top isn't -1.
24088 * opts.c (opts_obstack, opts_concat): Moved to...
24089 * opts-common.c (opts_obstack, opts_concat): ... here.
24091 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24093 PR middle-end/56461
24094 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
24096 2013-03-04 Martin Jambor <mjambor@suse.cz>
24098 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
24099 all appropriate places.
24101 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
24103 PR tree-optimization/56424
24104 * ipa-split.c (split_function): Do not set the RSO flag if result is
24105 not by reference and its type is a register type.
24107 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
24109 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
24110 (microblaze_legitimate_pic_operand): Likewise
24111 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
24112 new function microblaze_legitimate_pic_operand
24113 * config/microblaze/microblaze-protos.h
24114 (microblaze_legitimate_pic_operand): Declare.
24116 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24118 * config/microblaze/predicates.md (call_insn_simple_operand):
24119 New predicate for supported rtx code types.
24120 * config/microblaze/microblaze.md (call_internal1): Use
24121 call_insn_simple_operand predicate.
24123 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24125 PR middle-end/56461
24126 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
24127 partitions.ordered_remove.
24129 PR middle-end/56461
24130 * tree-vect-stmts.c (vectorizable_conversion): Don't call
24131 vec_oprnds0.create (1) for modifier == NONE.
24133 PR middle-end/56461
24134 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24135 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24136 vec_oprnds1 right before pushing anything to it for
24139 PR middle-end/56461
24140 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24141 set nbbs to 0 instead of having separate code path.
24142 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24143 instead of false as last argument if returning NULL.
24145 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24147 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24148 the attribute is now called "target" instead of "option".
24149 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24150 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24151 attribute/pragma name for TARGET_OPTION_VALID_P and
24152 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24153 * doc/tm.texi: Regenerated.
24155 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24157 * config/microblaze/microblaze.c:
24158 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24159 * config/microblaze/microblaze.h: Add -mxl-reorder to
24161 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24162 instructions emitted if TARGET_REORDER.
24163 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24164 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24167 2013-03-01 Xinliang David Li <davidxl@google.com>
24169 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24172 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24174 PR middle-end/56461
24175 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24176 vector even when returning true. Fix up function comment formatting.
24178 PR middle-end/56461
24179 * ira-build.c (ira_loop_nodes_count): New variable.
24180 (create_loop_tree_nodes): Initialize it.
24181 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24183 PR middle-end/56461
24184 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24185 method on dr_chain and result_chain.
24186 * tree-vect-stmts.c (vectorizable_store): Only call
24187 result_chain.create if j == 0.
24189 PR middle-end/56461
24190 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24191 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24192 before overwriting it.
24194 2013-03-01 Tobias Burnus <burnus@net-b.de>
24196 * doc/extended.texi (C Extensions): Change order in @menu
24198 (Other MIPS Built-in Functions): Move last MIPS entry before
24199 "picoChip Built-in Functions".
24200 (SH Built-in Functions): Move after RX Built-in Functions.
24201 * doc/gcc.texi (Introduction): Change order in @menu
24203 * doc/md.texi (Constraints): Ditto.
24204 * gty.texi (Type Information): Ditto.
24205 (User-provided marking routines for template types): Make
24207 * doc/invoke.texi (AArch64 Options): Move before
24208 "Adapteva Epiphany Options".
24210 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24211 Jakub Jelinek <jakub@redhat.com>
24214 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24215 no_address_safety_analysis attribute.
24216 * doc/extend.texi (no_address_safety_attribute): Rename to
24217 no_sanitize_address attribute, mention no_address_safety_analysis
24218 attribute as deprecated alias.
24220 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24222 PR middle-end/56461
24223 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24224 type to vec<vec<tree> > *.
24225 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24226 to be vec<tree> instead of vec<tree> *, set vec_defs
24227 to vNULL and call vec_defs.create (number_of_vects), adjust other
24229 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24230 vectorizable_condition): Adjust vect_get_slp_defs callers.
24232 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24234 * config/aarch64/aarch64.c
24235 (aarch64_float_const_representable): Remove unused variable.
24237 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24239 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24241 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24243 * config/aarch64/aarch64-builtins.c
24244 (aarch64_init_simd_builtins): Make static.
24246 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24248 * config/aarch64/aarch64.c
24249 (aarch64_simd_make_constant): Make static.
24251 2013-02-28 Martin Jambor <mjambor@suse.cz>
24253 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24254 with no initialization to the RHS of debug statements.
24256 2013-02-28 Martin Jambor <mjambor@suse.cz>
24258 PR tree-optimization/56294
24259 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24261 (get_access_replacement): Do not call create_access_replacement.
24262 Assert a replacement exists.
24263 (get_repl_default_def_ssa_name): Create the replacement declaration
24266 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24268 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24269 final_end_function.
24271 2013-02-28 Marek Polacek <polacek@redhat.com>
24273 PR rtl-optimization/56466
24274 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24275 if we're changing a loop.
24276 (peel_loops_completely): Likewise.
24278 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24281 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24283 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24286 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24287 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24288 INTX_FTYPE_FX, FX_FTYPE_INTX.
24289 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24291 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24293 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24294 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24295 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24296 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24297 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24298 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24299 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24300 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24301 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24302 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24303 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24304 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24305 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24306 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24307 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24308 (avrxmega6): Increase max flash segments from 5 to 6.
24309 * config/avr/t-multilib: Regenerate.
24310 * config/avr/avr-tables.opt: Regenerate.
24311 * doc/avr-mmcu.texi: Regenerate.
24313 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24315 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24316 (avr_device_to_arch): Rename to avr_device_to_ld.
24317 (avr_device_to_as): New prototype.
24318 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24319 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24320 * config/avr/driver-avr.c (avr_device_to_as): New.
24321 (avr_device_to_arch): Rename to avr_device_to_ld.
24323 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24325 PR middle-end/56461
24326 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24327 method on dr_chain and result_chain.
24329 PR middle-end/56461
24330 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24331 pointer_set_destroy on not_executed_last_iteration.
24333 PR middle-end/56461
24334 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24336 PR middle-end/56461
24337 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24338 FOR_EACH_DEFINED_FUNCTION when freeing state.
24340 PR middle-end/56461
24341 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24343 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24346 PR middle-end/56461
24347 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24348 known_aggs[i].items and release known_aggs vector.
24350 PR middle-end/56461
24351 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24353 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24355 * config/microblaze/microblaze.c (microblaze_emit_compare):
24356 Use xor for EQ/NE comparisions.
24357 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24358 (cbranchsf4): Adjust operator to comparison_operator.
24360 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24362 PR middle-end/56461
24363 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24365 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24366 vec_safe_push, always update *slot.
24367 (redirect_edge_var_map_clear): Use vec_free.
24368 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24369 (free_var_map_entry): Use vec_free.
24370 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24371 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24373 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24375 PR middle-end/45472
24376 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24377 when the may_trap_p bit of the exprs being merged differs.
24378 Reorder tests for speculativeness in the logical and operator.
24380 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24382 * incpath.c (add_standard_paths): Use reconcat instead of concat
24383 where appropriate and avoid leaking memory.
24385 * opts.h: Include obstack.h.
24386 (opts_concat): New prototype.
24387 (opts_obstack): New declaration.
24388 * opts.c (opts_concat): New function.
24389 (opts_obstack): New variable.
24390 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24391 (finish_options): Use opts_concat instead of concat
24392 and XOBNEWVEC instead of XNEWVEC.
24393 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24394 generate_option): Likewise.
24395 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24396 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24399 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24400 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24402 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24404 PR middle-end/56461
24405 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24407 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24409 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24410 (arm_block_move_unaligned_straight): Likewise.
24411 (arm_adjust_block_mem): Likewise.
24413 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24416 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24417 temp, cond and label.
24418 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24421 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24422 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24425 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24426 of prologue/epilogue functions.
24429 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24430 Remove unused variable cfa_offset.
24431 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24434 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24435 type promotion to unsigned.
24438 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24439 for HOST_WIDE_INT of 32 bit / same size as int.
24440 (arm_block_move_unaligned_straight): Likewise.
24441 (arm_adjust_block_mem): Likewise.
24444 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24447 2013-02-26 Marek Polacek <polacek@redhat.com>
24449 PR tree-optimization/56426
24450 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24452 2013-02-26 Richard Biener <rguenther@suse.de>
24455 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24456 unused variable loops.
24458 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24460 PR tree-optimization/56448
24461 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24462 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24463 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24464 later operands of the references, or even first operand for
24465 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24467 PR tree-optimization/56443
24468 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24469 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24470 to type_for_mode langhook.
24472 2013-02-25 Matt Turner <mattst88@gmail.com>
24474 * doc/invoke.texi: Document r4700.
24476 2013-02-25 Richard Biener <rguenther@suse.de>
24478 PR tree-optimization/56175
24479 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24481 (simplify_bitwise_binary): ... here. Also guard the conversion
24482 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24484 2013-02-25 Catherine Moore <clm@codesourcery.com>
24487 2013-02-24 Catherine Moore <clm@codesourcery.com>
24488 Maciej W. Rozycki <macro@codesourcery.com>
24489 Tom de Vries <tom@codesourcery.com>
24490 Nathan Sidwell <nathan@codesourcery.com>
24491 Iain Sandoe <iain@codesourcery.com>
24492 Nathan Froyd <froydnj@codesourcery.com>
24493 Chao-ying Fu <fu@mips.com>
24495 * doc/extend.texi: (micromips, nomicromips, nocompression):
24496 Document new function attributes.
24497 * doc/invoke.texi (minterlink-compressed, mmicromips,
24498 m14k, m14ke, m14kec): Document new options.
24499 (minterlink-mips16): Update documentation.
24500 * doc/md.texi (ZC, ZD): Document new constraints.
24501 * configure.ac (gcc_cv_as_micromips): Check if linker
24502 supports the .set micromips directive.
24503 * configure: Regenerate.
24504 * config.in: Regenerate.
24505 * config/mips/mips-tables.opt: Regenerate.
24506 * config/mips/micromips.md: New file.
24507 * constraints.md (ZC, AD): New constraints.
24508 * config/mips/predicates.md (movep_src_register): New predicate.
24509 (movep_src_operand): New predicate.
24510 (non_volatile_mem_operand): New predicate.
24511 * config/mips/mips.md (multimem): New type.
24512 (length): Differentiate between 17-bit and 18-bit branch offsets.
24513 (MOVEP1, MOVEP2): New mode iterator.
24514 (mov_<load>l): Use ZC constraint.
24515 (mov_<load>r): Likewise.
24516 (mov_<store>l): Likewise.
24517 (mov_<store>r): Likewise.
24518 (*branch_equality<mode>_inverted): Add microMIPS support.
24519 (*branch_equality<mode>): Likewise.
24520 (*jump_absolute): Likewise.
24521 (indirect_jump_<mode>): Likewise.
24522 (tablejump_<mode>): Likewise.
24523 (<optab>_internal): Likewise.
24524 (sibcall_internal): Likewise.
24525 (sibcall_value_internal): Likewise.
24526 (prefetch): Use constraint ZD.
24527 * config/mips/mips.opt (minterlink-compressed): New option.
24528 (minterlink-mips16): Now an alias for minterlink-compressed.
24529 (mmicromips): New option.
24530 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24531 (compare_and_swap_12): Likewise.
24532 (sync_add<mode>): Likewise.
24533 (sync_<optab>_12): Likewise.
24534 (sync_old_<optab>_12): Likewise.
24535 (sync_new_<optab>_12): Likewise.
24536 (sync_nand_12): Likewise.
24537 (sync_old_nand_12): Likewise.
24538 (sync_new_nand_12): Likewise.
24539 (sync_sub<mode>): Likewise.
24540 (sync_old_add<mode>): Likewise.
24541 (sync_old_sub<mode>): Likewise.
24542 (sync_new_add<mode>): Likewise.
24543 (sync_new_sub<mode>): Likewise.
24544 (sync_<optab><mode>): Likewise.
24545 (sync_old_<optab><mode>): Likewise.
24546 (sync_new_<optab><mode>): Likewise.
24547 (sync_nand<mode>): Likewise.
24548 (sync_old_nand<mode>): Likewise.
24549 (sync_new_nand<mode>): Likewise.
24550 (sync_lock_test_and_set<mode>): Likewise.
24551 (test_and_set_12): Likewise.
24552 (atomic_compare_and_swap<mode>): Likewise.
24553 (atomic_exchange<mode>_llsc): Likewise.
24554 (atomic_fetch_add<mode>_llsc): Likewise.
24555 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24556 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24557 (umips_save_restore_pattern_p): Likewise.
24558 (umips_load_store_pair_p): Likewise.
24559 (umips_output_load_store_pair): Likewise.
24560 (umips_movep_target_p): Likewise.
24561 (umips_12bit_offset_address_p): Likewise.
24562 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24563 (mips_base_mips16): Rename this...
24564 (mips_base_compression_flags): ...to this. Update all uses.
24565 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24566 (mips_mips16_decl_p): Delete.
24567 (mips_nomips16_decl_p): Delete.
24568 (mips_get_compress_on_flags): New function.
24569 (mips_get_compress_off_flags): New function.
24570 (mips_get_compress_mode): New function.
24571 (mips_get_compress_on_name): New function.
24572 (mips_get_compress_off_name): New function.
24573 (mips_insert_attributes): Support multiple compression types.
24574 (mips_merge_decl_attributes): Likewise.
24575 (umips_12bit_offset_address_p): New function.
24576 (mips_start_function_definition): Emit .set micromips directive.
24577 (mips_call_may_need_jalx_p): New function.
24578 (mips_function_ok_for_sibcall): Add microMIPS support.
24579 (mips_print_operand_punctuation): Support short delay slots and
24581 (umips_swm_mask, umips_swm_encoding): New.
24582 (umips_build_save_restore): New function.
24583 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24584 (was_mips16_p): Remove.
24585 (old_compression_mode): New.
24586 (mips_set_compression_mode): New function.
24587 (mips_set_current_function): Add microMIPS support.
24588 (mips_option_override): Likewise.
24589 (umips_save_restore_pattern_p): New function.
24590 (umips_output_save_restore): New function.
24591 (umips_load_store_pair_p_1): New function.
24592 (umips_load_store_pair_p): New function.
24593 (umips_output_load_store_pair_1): New function.
24594 (umips_output_load_store_pair): New function.
24595 (umips_movep_target_p) New function.
24596 (mips_prepare_pch_save): Add microMIPS support.
24597 * config/mips/mips.h (TARGET_COMPRESSION): New.
24598 (TARGET_CPU_CPP_BUILTINS): Update macro
24599 to use new compression flags and to support microMIPS.
24600 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24601 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24602 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24603 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24604 (ASM_SPEC): Support mmicromips and mno-micromips.
24605 (M16STORE_REG_P): New macro.
24606 (MIPS_CALL): Support TARGET_MICROMIPS.
24607 (MICROMIPS_J): New macro.
24608 (mips_base_mips16): Rename this...
24609 (mips_base_compression_flags): ...to this.
24610 (UMIPS_12BIT_OFFSET_P): New macro.
24611 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24612 (MULTILIB_DIRNAMES): Likewise.
24614 2013-02-25 Tom de Vries <tom@codesourcery.com>
24616 PR rtl-optimization/56131
24617 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24618 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24619 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24621 2013-02-25 Tobias Burnus <burnus@net-b.de>
24623 * doc/invoke.texi (-fsanitize=): Move from optimization
24624 to debugging options.
24626 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24628 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24630 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24631 Alexander Monakov <amonakov@ispras.ru>
24633 PR middle-end/56077
24634 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24635 flush pending lists also on non-jumps. Adjust comment.
24637 2013-02-24 Catherine Moore <clm@codesourcery.com>
24638 Maciej W. Rozycki <macro@codesourcery.com>
24639 Tom de Vries <tom@codesourcery.com>
24640 Nathan Sidwell <nathan@codesourcery.com>
24641 Iain Sandoe <iain@codesourcery.com>
24642 Nathan Froyd <froydnj@codesourcery.com>
24643 Chao-ying Fu <fu@mips.com>
24645 * doc/extend.texi: (micromips, nomicromips, nocompression):
24646 Document new function attributes.
24647 * doc/invoke.texi (minterlink-compressed, mmicromips,
24648 m14k, m14ke, m14kec): Document new options.
24649 (minterlink-mips16): Update documentation.
24650 * doc/md.texi (ZC, ZD): Document new constraints.
24651 * configure.ac (gcc_cv_as_micromips): Check if linker
24652 supports the .set micromips directive.
24653 * configure: Regenerate.
24654 * config.in: Regenerate.
24655 * config/mips/mips-tables.opt: Regenerate.
24656 * config/mips/micromips.md: New file.
24657 * constraints.md (ZC, AD): New constraints.
24658 * config/mips/predicates.md (movep_src_register): New predicate.
24659 (movep_src_operand): New predicate.
24660 (non_volatile_mem_operand): New predicate.
24661 * config/mips/mips.md (multimem): New type.
24662 (length): Differentiate between 17-bit and 18-bit branch offsets.
24663 (MOVEP1, MOVEP2): New mode iterator.
24664 (mov_<load>l): Use ZC constraint.
24665 (mov_<load>r): Likewise.
24666 (mov_<store>l): Likewise.
24667 (mov_<store>r): Likewise.
24668 (*branch_equality<mode>_inverted): Add microMIPS support.
24669 (*branch_equality<mode>): Likewise.
24670 (*jump_absolute): Likewise.
24671 (indirect_jump_<mode>): Likewise.
24672 (tablejump_<mode>): Likewise.
24673 (<optab>_internal): Likewise.
24674 (sibcall_internal): Likewise.
24675 (sibcall_value_internal): Likewise.
24676 (prefetch): Use constraint ZD.
24677 * config/mips/mips.opt (minterlink-compressed): New option.
24678 (minterlink-mips16): Now an alias for minterlink-compressed.
24679 (mmicromips): New option.
24680 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24681 (compare_and_swap_12): Likewise.
24682 (sync_add<mode>): Likewise.
24683 (sync_<optab>_12): Likewise.
24684 (sync_old_<optab>_12): Likewise.
24685 (sync_new_<optab>_12): Likewise.
24686 (sync_nand_12): Likewise.
24687 (sync_old_nand_12): Likewise.
24688 (sync_new_nand_12): Likewise.
24689 (sync_sub<mode>): Likewise.
24690 (sync_old_add<mode>): Likewise.
24691 (sync_old_sub<mode>): Likewise.
24692 (sync_new_add<mode>): Likewise.
24693 (sync_new_sub<mode>): Likewise.
24694 (sync_<optab><mode>): Likewise.
24695 (sync_old_<optab><mode>): Likewise.
24696 (sync_new_<optab><mode>): Likewise.
24697 (sync_nand<mode>): Likewise.
24698 (sync_old_nand<mode>): Likewise.
24699 (sync_new_nand<mode>): Likewise.
24700 (sync_lock_test_and_set<mode>): Likewise.
24701 (test_and_set_12): Likewise.
24702 (atomic_compare_and_swap<mode>): Likewise.
24703 (atomic_exchange<mode>_llsc): Likewise.
24704 (atomic_fetch_add<mode>_llsc): Likewise.
24705 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24706 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24707 (umips_save_restore_pattern_p): Likewise.
24708 (umips_load_store_pair_p): Likewise.
24709 (umips_output_load_store_pair): Likewise.
24710 (umips_movep_target_p): Likewise.
24711 (umips_12bit_offset_address_p): Likewise.
24712 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24713 (mips_base_mips16): Rename this...
24714 (mips_base_compression_flags): ...to this. Update all uses.
24715 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24716 (mips_mips16_decl_p): Delete.
24717 (mips_nomips16_decl_p): Delete.
24718 (mips_get_compress_on_flags): New function.
24719 (mips_get_compress_off_flags): New function.
24720 (mips_get_compress_mode): New function.
24721 (mips_get_compress_on_name): New function.
24722 (mips_get_compress_off_name): New function.
24723 (mips_insert_attributes): Support multiple compression types.
24724 (mips_merge_decl_attributes): Likewise.
24725 (umips_12bit_offset_address_p): New function.
24726 (mips_start_function_definition): Emit .set micromips directive.
24727 (mips_call_may_need_jalx_p): New function.
24728 (mips_function_ok_for_sibcall): Add microMIPS support.
24729 (mips_print_operand_punctuation): Support short delay slots and
24731 (umips_swm_mask, umips_swm_encoding): New.
24732 (umips_build_save_restore): New function.
24733 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24734 (was_mips16_p): Remove.
24735 (old_compression_mode): New.
24736 (mips_set_compression_mode): New function.
24737 (mips_set_current_function): Add microMIPS support.
24738 (mips_option_override): Likewise.
24739 (umips_save_restore_pattern_p): New function.
24740 (umips_output_save_restore): New function.
24741 (umips_load_store_pair_p_1): New function.
24742 (umips_load_store_pair_p): New function.
24743 (umips_output_load_store_pair_1): New function.
24744 (umips_output_load_store_pair): New function.
24745 (umips_movep_target_p) New function.
24746 (mips_prepare_pch_save): Add microMIPS support.
24747 * config/mips/mips.h (TARGET_COMPRESSION): New.
24748 (TARGET_CPU_CPP_BUILTINS): Update macro
24749 to use new compression flags and to support microMIPS.
24750 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24751 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24752 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24753 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24754 (ASM_SPEC): Support mmicromips and mno-micromips.
24755 (M16STORE_REG_P): New macro.
24756 (MIPS_CALL): Support TARGET_MICROMIPS.
24757 (MICROMIPS_J): New macro.
24758 (mips_base_mips16): Rename this...
24759 (mips_base_compression_flags): ...to this.
24760 (UMIPS_12BIT_OFFSET_P): New macro.
24761 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24762 (MULTILIB_DIRNAMES): Likewise.
24764 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24767 * target-globals.c (save_target_globals): For init_reg_sets and
24768 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24770 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
24772 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24773 * config/aarch64/t-aarch64
24774 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24776 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24778 PR inline-asm/56148
24779 * lra-constraints.c (process_alt_operands): Reload operand
24780 conflicting with earlier clobber only if no more other conflicting
24783 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24786 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24787 if not linking a shared library.
24789 2013-02-22 Seth LaForge <sethml@google.com>
24791 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24793 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24795 * config/arm/arm.md (split for extendsidi): Update condition.
24796 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24797 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24798 (qhs_zextenddi_cstr): Likewise.
24800 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24802 PR middle-end/56420
24803 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24804 avoid signed wrapping.
24805 (expand_mult): Handle properly multiplication by
24806 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24807 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24808 in the compiler if coeff is HOST_WIDE_INT_MIN.
24809 (expand_divmod): Don't make ext_op1 static, change it's type to
24810 uhwi. Avoid undefined behavior in -INTVAL (op1).
24812 PR rtl-optimization/50339
24813 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24815 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24816 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24817 into splitting_ashiftrt field.
24818 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24820 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24823 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24825 PR middle-end/56108
24826 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24827 are sure to go irrevocable.
24829 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24831 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24832 scalars are valid operands.
24834 2013-02-21 Martin Jambor <mjambor@suse.cz>
24836 PR tree-optimization/56310
24837 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24838 only matching indices and non-negative final offsets.
24839 (intersect_aggregates_with_edge): Pass src_idx to
24840 agg_replacements_to_vector. Pass src_idx insstead of index to
24841 intersect_with_agg_replacements.
24843 2013-02-21 Martin Jambor <mjambor@suse.cz>
24845 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24846 instead of hard-wired defaults.
24848 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24850 * doc/invoke.texi (MIPS Options): Update documentation of the
24851 floating-point multiply-accumulate instruction restrictions.
24853 2013-02-21 Kostya Serebryany <kcc@google.com>
24855 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24856 asan_shadow_offset on x86_64 linux.
24858 2013-02-21 Richard Biener <rguenther@suse.de>
24860 PR tree-optimization/56415
24862 2013-02-11 Richard Biener <rguenther@suse.de>
24864 PR tree-optimization/56273
24865 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24868 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24871 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24874 PR inline-asm/56405
24875 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24876 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24878 2013-02-20 Jan Hubicka <jh@suse.cz>
24880 PR tree-optimization/56265
24881 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24882 when target is referenced for first time.
24884 2013-02-20 Richard Biener <rguenther@suse.de>
24886 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24887 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24888 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24889 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24890 not return anything.
24891 (rename_ssa_copies): Do not remove unused locals.
24892 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24893 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24894 * passes.c (execute_function_todo): Do not schedule unused locals
24895 removal if cleanup_tree_cfg did something.
24896 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24897 about the number of removed locals.
24899 2013-02-20 Richard Biener <rguenther@suse.de>
24901 PR tree-optimization/56398
24902 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24904 2013-02-20 Martin Jambor <mjambor@suse.cz>
24906 PR tree-optimization/55334
24907 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24908 restricted pointers to arrays.
24910 2013-02-20 Richard Biener <rguenther@suse.de>
24911 Jakub Jelinek <jakub@redhat.com>
24913 PR tree-optimization/56396
24914 * tree-ssa-ccp.c (n_const_val): New static variable.
24915 (get_value): Return NULL for SSA names we don't have a lattice
24917 (ccp_initialize): Initialize n_const_val.
24918 * tree-ssa-copy.c (n_copy_of): New static variable.
24919 (init_copy_prop): Initialize n_copy_of.
24920 (get_value): Return NULL_TREE for SSA names we don't have a
24923 2013-02-20 Martin Jambor <mjambor@suse.cz>
24925 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24927 2013-02-20 Richard Biener <rguenther@suse.de>
24929 * genpreds.c (write_lookup_constraint): Do not compare first
24930 letter of the constraint again.
24932 2013-02-20 Richard Biener <rguenther@suse.de>
24934 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24936 (get_use_iv_cost): Terminate hashtable walk when coming across
24939 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24941 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24942 reassociation for avx2 targets.
24944 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24946 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24947 Add version check for v8.10.a to enable microblaze_has_clz
24948 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24949 version and TARGET_PATTERN_COMPARE check
24950 * config/microblaze/microblaze.md: New clzsi2 instruction
24952 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24954 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24955 function before branching.
24957 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24959 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
24961 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
24963 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
24965 PR middle-end/55889
24966 * sel-sched.c: Include ira.h.
24967 (implicit_clobber_conflict_p): New function.
24968 (moveup_expr): Use it.
24969 * Makefile.in (sel-sched.o): Depend on ira.h.
24971 2013-02-19 Richard Biener <rguenther@suse.de>
24973 PR tree-optimization/56384
24974 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
24975 (vn_hash_type): Split out from ...
24976 (vn_hash_constant_with_type): ... here.
24977 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
24978 (vn_phi_eq): Compare types from vn_phi_s structure.
24979 (vn_phi_lookup): Populate vn_phi_s type.
24980 (vn_phi_insert): Likewise.
24982 2013-02-19 Jakub Jelinek <jakub@redhat.com>
24984 PR tree-optimization/56350
24985 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
24986 if haven't found reduction or nested cycle operand, rather than
24987 asserting we must find it.
24989 PR tree-optimization/56381
24990 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
24993 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
24994 Jakub Jelinek <jakub@redhat.com>
24997 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
24998 (swap_optab_enable): Same.
24999 (init_all_optabs): Use argument instead of global.
25000 * tree.h (struct tree_optimization_option): New field target_optabs.
25001 * expr.h (init_all_optabs): Add argument to prototype.
25002 (TREE_OPTIMIZATION_OPTABS): New.
25003 (save_optabs_if_changed): Protoize.
25004 * optabs.h: Declare this_fn_optabs.
25005 * optabs.c (save_optabs_if_changed): New.
25006 Declare this_fn_optabs.
25007 (init_optabs): Add argument to init_all_optabs() call.
25008 * function.c (invoke_set_current_function_hook): Handle per
25010 * function.h (struct function): New field optabs.
25011 * config/mips/mips.c (mips_set_mips16_mode): Handle when
25012 optimization_current_node has changed.
25013 * target-globals.h (save_target_globals_default_opts): Protoize.
25014 * target-globals.c (save_target_globals_default_opts): New.
25016 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25019 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
25020 registers %fr12 and %fr12R as call used.
25023 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
25024 and HImode, require all displacements to be an integer multiple of
25026 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
25027 only allow QImode and HImode when reload is in progress and strict is
25028 true. Likewise for symbolic addresses. Use base14_operand to check
25029 displacements in REG+BASE addresses.
25031 2013-02-18 Richard Biener <rguenther@suse.de>
25033 PR tree-optimization/56366
25034 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25035 handle sign-conversion of outer-loop initial induction value.
25037 2013-02-18 Richard Biener <rguenther@suse.de>
25039 PR middle-end/56349
25040 * cfghooks.c (merge_blocks): If we merge a latch into another
25041 block adjust references to it.
25042 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
25043 (verify_loop_structure): Verify that a recorded latch is in fact
25046 2013-02-18 Richard Biener <rguenther@suse.de>
25048 PR tree-optimization/56321
25049 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
25050 order SSA name release and virtual operand unlinking.
25052 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25054 * config/microblaze/microblaze.md (save_stack_block): Define.
25055 (restore_stack_block): Likewise.
25057 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25059 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
25060 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
25061 * config/microblaze/microblaze.c (microblaze_option_override):
25062 Bail out early for PIC modes when target does not support PIC.
25064 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25066 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
25067 Replace with a microblaze version.
25068 (microblaze_trampoline_init): Adapt for microblaze.
25069 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
25072 2013-02-16 Jakub Jelinek <jakub@redhat.com>
25073 Dodji Seketeli <dodji@redhat.com>
25076 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
25077 (instrument_mem_region_access): Do not forget to always put
25078 instrumentation of the of 'base' and 'base + len' in a "if (len !=
25079 0) statement, even for cases where either 'base' or 'base + len'
25080 are not instrumented -- because they have been previously
25081 instrumented. Simplify the logic by putting all the statements
25082 instrument 'base + len' inside a sequence, and then insert that
25083 sequence right before the current insertion point. Then, to
25084 instrument 'base + len', just get an iterator on that statement.
25085 And do not forget to update the pointer to iterator the function
25086 received as argument.
25088 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
25090 PR rtl-optimization/56348
25091 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
25093 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
25095 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
25096 (clean_graph_dump_file): Pass base to start_graph_dump.
25098 2013-02-14 Richard Henderson <rth@redhat.com>
25101 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
25103 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
25105 * collect2-aix.h: Define F_LOADONLY.
25107 2013-02-14 Richard Biener <rguenther@suse.de>
25110 * varasm.c (output_constant_def_1): Get the decl representing
25111 the constant as argument.
25112 (output_constant_def): Wrap output_constant_def_1.
25113 (make_decl_rtl): Use output_constant_def_1 with the decl
25114 representing the constant.
25115 (build_constant_desc): Optionally re-use a decl already
25116 representing the constant.
25117 (tree_output_constant_def): Adjust.
25119 2013-02-14 Dodji Seketeli <dodji@redhat.com>
25122 * asan.c (instrument_builtin_call): Really put the length of the
25123 second source argument into src1_len.
25125 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25127 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
25128 argument. If it is false, don't create edge from then_bb to
25130 (insert_if_then_before_iter): Pass true to it.
25131 (build_check_stmt): Pass false to it.
25132 (transform_statements): Flush hash table only on extended basic
25133 block boundaries, rather than at the beginning of every bb.
25134 Don't flush hash table on nonfreeing_call_p calls.
25135 * tree-flow.h (nonfreeing_call_p): New prototype.
25136 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25138 2013-02-13 David S. Miller <davem@davemloft.net>
25140 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25142 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25145 * ira.c (max_regno_before_ira): Move from ...
25147 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25150 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25152 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25154 2013-02-13 Richard Biener <rguenther@suse.de>
25157 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25158 globals in MEM_REFs.
25160 2013-02-13 Richard Biener <rguenther@suse.de>
25162 * loop-init.c (loop_optimizer_init): Clear loop state when
25163 re-initializing preserved loops.
25164 * loop-unswitch.c (unswitch_single_loop): Return whether
25165 we unswitched the loop. Do not verify loop state here.
25166 (unswitch_loops): When we unswitched a loop discover new loops.
25168 2013-02-13 Kostya Serebryany <kcc@google.com>
25170 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25172 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25174 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25176 Avoid instrumenting duplicated memory access in the same basic block
25177 * Makefile.in (asan.o): Add new dependency on hash-table.h
25178 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25179 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25180 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25181 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25182 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25183 (get_mem_ref_of_assignment): New functions.
25184 (get_mem_refs_of_builtin_call): Extract from
25185 instrument_builtin_call and tweak a little bit to make it fit with
25187 (instrument_builtin_call): Use the new
25188 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25189 of is_gimple_builtin_call.
25190 (instrument_derefs, instrument_mem_region_access): Insert the
25191 instrumented memory reference into the hash table.
25192 (maybe_instrument_assignment): Renamed instrument_assignment into
25193 this, and change it to advance the iterator when instrumentation
25194 actually happened and return true in that case. This makes it
25195 homogeneous with maybe_instrument_assignment, and thus give a
25196 chance to callers to be more 'regular'.
25197 (transform_statements): Clear the memory reference hash table
25198 whenever we enter a new BB, when we cross a function call, or when
25199 we are done transforming statements. Use
25200 maybe_instrument_assignment instead of instrumentation. No more
25201 need to special case maybe_instrument_assignment and advance the
25202 iterator after calling it; it's now handled just like
25203 maybe_instrument_call. Update comment.
25205 2013-02-13 Richard Biener <rguenther@suse.de>
25207 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25208 Fix loop discovery code.
25210 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25212 PR inline-asm/56148
25213 * lra-constraints.c (process_alt_operands): Match early clobber
25214 operand with itself. Check conflicts with earlyclobber only if
25215 the operand is not reloaded. Prefer to reload conflicting operand
25216 if earlyclobber and matching operands are the same.
25218 2013-02-12 Richard Biener <rguenther@suse.de>
25221 * lto-streamer-out.c (write_symbol): Do not output symbols
25222 for hard register variables.
25224 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25227 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25228 (umulsidi3_insn, mulsidi3_insn): New insns.
25230 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25232 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25233 (struct tune_params): Add vec_costs field.
25234 * config/arm/arm.c (arm_builtin_vectorization_cost)
25235 (arm_add_stmt_cost): New functions.
25236 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25237 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25238 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25239 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25240 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25241 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25242 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25244 2013-02-12 Richard Biener <rguenther@suse.de>
25247 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25248 decls again if possible.
25250 2013-02-12 Richard Biener <rguenther@suse.de>
25252 PR middle-end/56288
25253 * tree-ssa.c (verify_ssa_name): Fix check, move
25254 SSA_NAME_IN_FREE_LIST check up.
25256 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25257 Steven Bosscher <steven@gcc.gnu.org>
25259 PR rtl-optimization/56151
25260 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25261 equal to op0 or op1, and last_insn pattern is CODE operation
25262 with MEM dest and one of the operands matches that MEM.
25264 2013-02-11 Sriraman Tallam <tmsriram@google.com>
25266 * doc/extend.texi: Document Function Multiversioning and "default"
25267 parameter string to target attribute.
25268 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25269 target attribute parameter is "default".
25270 (ix86_compare_version_priority): Remove checks for target attribute.
25271 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25272 Remove check for target attribute equal to NULL. Add assert.
25273 (ix86_generate_version_dispatcher_body): Change error to sorry.
25275 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25276 Jack Howarth <howarth@bromo.med.uc.edu>
25277 Patrick Marlier <patrick.marlier@gmail.com>
25280 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25281 define ENDFILE_SPEC as TM_DESTRUCTOR.
25282 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25284 2013-02-11 Alexander Potapenko <glider@google.com>
25285 Jack Howarth <howarth@bromo.med.uc.edu>
25286 Jakub Jelinek <jakub@redhat.com>
25289 * config/darwin.c (cdtor_record): Rename ctor_record.
25290 (sort_cdtor_records): Rename sort_ctor_records.
25291 (finalize_dtors): New routine to sort destructors by
25292 priority before use in assemble_integer.
25293 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25295 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25297 PR rtl-optimization/56275
25298 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25299 offset is non-negative and less than cmode size before
25300 calling simplify_subreg.
25302 2013-02-11 Richard Biener <rguenther@suse.de>
25304 PR tree-optimization/56264
25305 * cfgloop.h (fix_loop_structure): Adjust prototype.
25306 * loop-init.c (fix_loop_structure): Return the number of
25307 newly discovered loops.
25308 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25309 are discovered, do a full loop-closed SSA rewrite.
25311 2013-02-11 Richard Biener <rguenther@suse.de>
25313 PR tree-optimization/56273
25314 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25316 (check_array_ref): Fix missing newline in dumps.
25317 (search_for_addr_array): Likewise.
25319 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25321 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25323 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25326 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25328 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25330 PR rtl-optimization/56246
25331 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25333 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25334 constraints are satisfied.
25336 2013-02-08 Jeff Law <law@redhat.com>
25339 * emit-rtl.c (reg_is_parm_p): New function.
25340 * regs.h (reg_is_parm_p): New prototype.
25341 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25342 callee-clobbered registers.
25344 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25347 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25348 If there is no implicit builtin declaration, just return NULL.
25350 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25352 * config/i386/sse.md (FMAMODEM): New mode iterator.
25353 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25354 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25356 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25358 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25359 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25360 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25362 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25364 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25365 (microblaze*-*-elf): Likewise.
25366 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25368 * config/microblaze/microblaze-c.c: Add builtin defines for
25369 _LITTLE_ENDIAN and _BIG_ENDIAN.
25370 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25371 add to TARGET_DEFAULT flags.
25372 Expand ASM_SPEC and LINK_SPEC.
25373 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25374 * config/microblaze/microblaze.md: Update extendsidi2 and
25375 movdi_internal instructions to use low-order / high-order reg
25377 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25378 options and inversemask / mask of LITTLE_ENDIAN.
25379 * config/microblaze/t-microblaze: Expand multilib options to
25380 include mlittle-endian (le) and update exceptions patterns.
25382 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25384 PR rtl-optimization/56195
25385 * lra-constraints.c (get_reload_reg): Don't reuse regs
25386 if they have smaller mode than requested, if they have
25387 wider mode than requested, try to return a SUBREG.
25389 PR tree-optimization/56250
25390 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25391 if type is unsigned and code isn't MULT_EXPR.
25393 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25395 PR tree-optimization/56064
25396 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25397 bits according to mode.
25398 * fixed-value.h (fixed_from_double_int)
25399 (const_fixed_from_double_int): Adjust comments.
25401 2013-02-08 Richard Biener <rguenther@suse.de>
25404 * lto-streamer.h (struct data_in): Remove current_file, current_line
25405 and current_col members.
25406 * lto-streamer-out.c (lto_output_location): Stream changed bits
25407 en-block for efficiency.
25408 * lto-streamer-in.c (clear_line_info): Remove.
25409 (lto_input_location): Cache current file, line and column
25410 globally via local statics. Read changed bits en-block.
25411 (input_function): Do not call clear_line_info.
25412 (lto_read_body): Likewise.
25413 (lto_input_toplevel_asms): Likewise.
25415 2013-02-08 Michael Matz <matz@suse.de>
25417 PR tree-optimization/52448
25418 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25419 (nt_call_phase): New static.
25420 (add_or_mark_expr): Only mark accesses with newer phase than any
25422 (nonfreeing_call_p): New.
25423 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25424 (nt_fini_block): Keep blocks marked as visited.
25425 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25427 2013-02-08 Richard Biener <rguenther@suse.de>
25429 * ira.c (ira): Free broken dominator information.
25431 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25433 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25435 2013-02-08 Marek Polacek <polacek@redhat.com>
25437 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25439 2013-02-08 Richard Biener <rguenther@suse.de>
25441 PR middle-end/56181
25442 * cfgloop.h (flow_loops_find): Adjust.
25443 (bb_loop_header_p): Declare.
25444 * cfgloop.c (bb_loop_header_p): New function split out from ...
25445 (flow_loops_find): ... here. Adjust function signature,
25446 support incremental loop structure update.
25447 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25448 * cfgloopmanip.c (fix_loop_structure): Move ...
25449 * loop-init.c (fix_loop_structure): ... here.
25450 (apply_loop_flags): Split out from ...
25451 (loop_optimizer_init): ... here.
25452 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25453 in incremental mode, only remove dead loops here.
25455 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25458 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25459 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25460 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25461 (*round<mode>3.libgcc): New insns for fixed-modes.
25462 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25463 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25464 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25465 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25466 implementations. Define to __builtin_avr_absFX,
25467 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25468 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25469 __builtin_avr_countlsFX, respectively.
25470 * config/avr/avr-c.c (target.h): Include it.
25471 (enum avr_builtin_id): New enum.
25472 (avr_resolve_overloaded_builtin): New static function.
25473 (avr_register_target_pragmas): Use it to set
25474 targetm.resolve_overloaded_builtin.
25475 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25476 tree nodes used by DEF_BUILTIN.
25477 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25478 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25479 <AVR_BUILTIN_xxBITS>: Same.
25481 2013-02-08 Richard Biener <rguenther@suse.de>
25483 * cfgloop.c (verify_loop_structure): Properly handle
25484 a loop exiting to another loop header.
25485 * ira-int.h (ira_loops): Remove.
25486 * ira.c (ira_loops): Remove.
25487 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25488 (do_reload): Use loop_optimizer_finalize.
25489 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25490 number_of_loops to access the loop tree.
25491 (more_one_region_p): Likewise.
25492 (finish_loop_tree_nodes): Likewise.
25493 (rebuild_regno_allocno_maps): Likewise.
25494 (mark_loops_for_removal): Likewise.
25495 (mark_all_loops_for_removal): Likewise.
25496 (remove_unnecessary_regions): Likewise.
25497 (ira_build): Likewise.
25498 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25500 2013-02-08 Richard Biener <rguenther@suse.de>
25502 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25503 * ipa-pure-const.c (analyze_function): Avoid calling
25504 mark_irreducible_loops twice.
25505 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25507 2013-02-07 David S. Miller <davem@davemloft.net>
25509 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25511 * var-tracking.c (vt_add_function_parameter): Test the presence of
25512 HAVE_window_save properly and do not remap argument registers when
25513 we have a leaf function.
25515 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25518 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25520 * config/i386/i386.c (ix86_print_operand): Ditto.
25522 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25524 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25526 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25528 PR rtl-optimization/56225
25529 * lra-constraints.c (process_alt_operands): Check that reload hard
25530 reg can hold value for strict_low_part.
25532 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25535 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25536 dwarf2out_end_function.
25537 (in_first_function_p, maybe_at_text_label_p,
25538 first_loclabel_num_not_at_text_label): New variables.
25539 (dwarf2out_var_location): In the first function find out
25540 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25541 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25544 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25546 PR rtl-optimization/56178
25547 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25548 SUBREG of a register. Tidy up related block of code.
25549 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25550 note if the source is a register or a SUBREG of a register.
25552 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25555 * config/rs6000/rs6000.md (ptrm): New mode attr.
25556 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25557 call_value_indirect_aix<pttrsize>,
25558 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25561 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25563 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25564 if -bnortl. Convert to strcmp and strncmp.
25566 2013-02-07 Alan Modra <amodra@gmail.com>
25569 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25570 addresses won't wrap when offsetting.
25571 (rs6000_secondary_reload): Provide secondary reloads needed for
25572 wrapping LO_SUM addresses.
25574 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25576 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25577 MACH, just __MACH__.
25579 2013-02-06 Richard Biener <rguenther@suse.de>
25581 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25582 instead of calling fix_loop_structure.
25584 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25586 PR middle-end/56217
25587 * omp-low.c (use_pointer_for_field): Return false if
25588 lower_send_shared_vars doesn't generate any copy-out code.
25590 2013-02-06 Tom de Vries <tom@codesourcery.com>
25592 PR rtl-optimization/56131
25593 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25594 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25595 of the label is NULL. Add comment.
25597 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25599 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25602 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25603 (STATIC_LIBTSAN_LIBS): Likewise.
25604 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25605 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25606 is defined, don't add anything else beyond that.
25607 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25608 (LINK_COMMAND_SPEC): Use them.
25610 PR tree-optimization/56205
25611 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25612 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25613 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25615 2013-02-05 Richard Biener <rguenther@suse.de>
25617 PR tree-optimization/53342
25618 PR tree-optimization/53185
25619 * tree-vectorizer.h (vect_check_strided_load): Remove.
25620 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25621 not disallow peeling for vectorized strided loads.
25622 (vect_check_strided_load): Make static and simplify.
25623 (vect_analyze_data_refs): Adjust.
25624 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25625 correctly when vectorizing strided loads.
25627 2013-02-05 Richard Biener <rguenther@suse.de>
25629 * doc/install.texi: Refer to ISL, not PPL.
25631 2013-02-05 Jan Hubicka <jh@suse.cz>
25633 PR tree-optimization/55789
25634 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25636 2013-02-05 Jan Hubicka <jh@suse.cz>
25638 PR tree-optimization/55789
25639 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25640 the dead call anyway.
25642 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25645 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25647 2013-02-04 Alexander Potapenko <glider@google.com>
25648 Jack Howarth <howarth@bromo.med.uc.edu>
25649 Jakub Jelinek <jakub@redhat.com>
25652 * config/darwin.c (sort_ctor_records): Stabilized qsort
25653 on constructor priority by using original position.
25654 (finalize_ctors): New routine to sort constructors by
25655 priority before use in assemble_integer.
25656 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25658 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25661 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25662 about visibility on artificial decls.
25663 * config/sol2.c (solaris_assemble_visibility): Likewise.
25665 2013-02-04 Kai Tietz <ktietz@redhat.com>
25668 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25669 argument and improve checking of return-argument types for 16-byte
25671 (ix86_function_value_1): Add additional valtype argument on call
25672 of function_value_64.
25673 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25674 handling infunction_value_64 function.
25676 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25678 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25680 2013-02-04 Richard Biener <rguenther@suse.de>
25682 PR tree-optimization/56188
25683 * tree-ssa-structalias.c (label_visit): Consider case with
25684 initially non-empty points-to set.
25685 (perform_var_substitution): Dump node mapping and clean up.
25687 2013-02-04 Richard Guenther <rguenther@suse.de>
25690 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25691 node prevail as last resort.
25692 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25693 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25695 2013-02-04 Richard Biener <rguenther@suse.de>
25697 PR tree-optimization/56113
25698 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25700 (equiv_class_lookup_or_add): ... this.
25701 (label_visit): Adjust and fix error in previous patch.
25702 (perform_var_substitution): Adjust.
25704 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25706 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25707 * config/sh/sh4-300.md: Likewise.
25708 * config/sh/sh4a.md: Likewise.
25709 * config/sh/constraints.md: Likewise.
25710 * config/sh/sh.md: Likewise.
25711 * config/sh/netbsd-elf.h: Likewise.
25712 * config/sh/predicates.md: Likewise.
25713 * config/sh/sh-protos.h: Likewise.
25714 * config/sh/ushmedia.h: Likewise.
25715 * config/sh/linux.h: Likewise.
25716 * config/sh/sh.c: Likewise.
25717 * config/sh/superh.h: Likewise.
25718 * config/sh/elf.h: Likewise.
25719 * config/sh/sh4.md: Likewise.
25720 * config/sh/sh.h: Likewise.
25722 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25724 * config/pa/constraints.md: Adjust unused letters. Change "T"
25725 constraint to match_test floating_point_store_memory_operand().
25726 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25727 (base14_operand): New.
25728 (floating_point_store_memory_operand): New.
25729 (integer_store_memory_operand): Revise to use base14_operand and
25730 reg_plus_base_memory_operand.
25731 (move_dest_operand): Allow symbolic_memory_operands.
25732 (symbolic_memory_operand): Check for LO_SOM.
25733 (symbolic_operand): Change default case to break.
25734 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25735 CONST_DOUBLE values to be reloaded by putting them into memory when
25736 the destination is a floating point register.
25737 (movdf): Remove code to handle CONST_DOUBLE.
25739 (reload_indf_r1): New.
25740 (reload_insf_r1): New.
25741 Consistently use "Q" and "T" constraints with integer and floating
25742 point move instructions, respectively.
25743 (movdi): Remove FAIL.
25744 Change predicate for source operand unamed DImode move from
25745 general_operand to move_src_operand.
25746 (umulsidi3): Change predicate for destination operand to
25748 Likewise for similar unamed patterns.
25749 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25750 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25751 (hppa_legitimize_address): Simplify mask calculation.
25752 (pa_emit_move_sequence): Revised handling of secondary reloads from
25753 REG+D addresses for floating point loads and stores. Directly handle
25754 loading CONST0_RTX (mode) to a floating point register.
25755 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25756 to floating point registers. Don't restrict secondary reloads to
25757 floating point registers to integer modes. Revise some comments and
25759 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25760 (pa_legitimate_address_p): New.
25761 (pa_legitimize_reload_address): New.
25762 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25763 (STRICT_REG_OK_FOR_BASE_P): New.
25764 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25765 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25767 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25768 Andrew Dixie <andrewd@gentrack.com>
25770 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25773 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25775 * expmed.c (extract_bit_field_1): Pass the full width of the
25776 structure to get_best_reg_extraction_insn.
25778 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25781 * configure.ac (use_cxa_atexit): Add AIX.
25782 * configure: Regenerate.
25784 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25786 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25789 * final.c (need_profile_function): New variable.
25790 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25791 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25792 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25793 notes, targetm.asm_out.function_prologue doesn't emit anything,
25794 HAVE_prologue and profiler should be emitted before prologue,
25795 set need_profile_function instead of emitting it.
25796 (final_scan_insn): If need_profile_function, emit
25797 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25798 NOTE_INSN_FUNCTION_BEG note.
25800 2013-02-01 Richard Henderson <rth@redhat.com>
25802 * config/rs6000/rs6000.md (smulditi3): New.
25805 * config/alpha/alpha.md (umulditi3): New.
25807 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25809 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25810 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25812 2013-02-01 Richard Biener <rguenther@suse.de>
25814 PR tree-optimization/56113
25815 * tree-ssa-structalias.c (label_visit): Reduce work for
25816 single-predecessor nodes.
25818 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25820 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25821 range isn't testing for zero.
25823 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25825 PR middle-end/56113
25826 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25828 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25829 Nick Clifton <nickc@redhat.com>
25831 * config/v850/constraints.md (Q): Define as a memory constraint.
25832 * config/v850/predicates.md (label_ref_operand): New predicate.
25833 (e3v5_shift_operand): New predicate.
25834 (ior_operator): New predicate.
25835 * config/v850/t-v850: Add e3v5 multilib.
25836 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25837 (v850_gen_movdi): Prototype.
25838 * config/v850/v850.c: Add support for e3v5 architecture.
25839 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25841 (construct_save_jarl): Add e3v5 long JARL support.
25842 (v850_adjust_insn_length): New function. Adjust length of call
25843 insns when using e3v5 instructions.
25844 (v850_gen_movdi): New function: Generate instructions to move a
25846 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25847 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25848 (TARGET_USE_FPU): Enable for e3v5.
25849 (CONST_OK_FOR_W): New macro.
25850 (ADJUST_INSN_LENGTH): Define.
25851 * config/v850/v850.md (UNSPEC_LOOP): Define.
25852 (attr cpu): Add v850e3v5.
25853 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25854 (movdi): New pattern.
25855 (movdi_internal): New pattern.
25856 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25857 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25858 (cstoresf4): Likewise.
25859 (cstoredf4): Likewise.
25860 (insv): New pattern.
25861 (rotlso3_a): New pattern.
25862 (rotlsi3_b): New pattern
25863 (rotlsi3_v850e3v5): New pattern.
25864 (doloop_begin): New pattern.
25865 (fix_loop_counter): New pattern.
25866 (doloop_end): New pattern.
25867 (branch_normal): Add e3v5 long branch support.
25868 (branch_invert): Likewise.
25869 (branch_z_normal): Likewise.
25870 (branch_z_invert): Likewise.
25871 (branch_nz_normal): Likewise.
25872 (branch_nz_invert): Likewise.
25873 (call_internal_short): Add e3v5 register-indirect JARL support.
25874 (call_internal_long): Likewise.
25875 (call_value_internal_short): Likewise.
25876 (call_value_internal_long): Likewise.
25877 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25878 (mloop): New option.
25879 * config.gcc: Add support for configuring v840e3v5 target.
25880 * doc/invoke.texi: Document new v850 specific command line options.
25882 2013-01-31 Paul Koning <ni1d@arrl.net>
25886 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25887 children if parent is a class.
25888 (prune_unused_types_prune): Don't add DW_AT_declaration.
25890 2013-01-31 Richard Biener <rguenther@suse.de>
25892 PR tree-optimization/56157
25893 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25894 match up operand with SLP child.
25896 2013-01-31 Jason Merrill <jason@redhat.com>
25899 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25900 parameters the first time.
25901 (gen_scheduled_generic_parms_dies): Check completeness here.
25903 2013-01-31 Richard Biener <rguenther@suse.de>
25905 PR middle-end/53073
25906 * common.opt (faggressive-loop-optimizations): New flag,
25907 enabled by default.
25908 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25909 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25910 infer_loop_bounds_from_undefined by it.
25912 2013-01-31 Richard Biener <rguenther@suse.de>
25914 PR tree-optimization/56150
25915 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25916 visit virtual operands.
25917 (find_uses_to_rename_bb): Likewise.
25919 2013-01-31 Richard Biener <rguenther@suse.de>
25921 PR tree-optimization/56150
25922 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25923 mixed store non-store stmts.
25925 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25928 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25929 LIBASAN_EARLY_SPEC is defined.
25930 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25931 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25933 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25936 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25937 invalid args instead of ICEing on it.
25938 (ix86_valid_target_attribute_tree): Return error_mark_node if
25939 ix86_valid_target_attribute_inner_p failed.
25940 (ix86_valid_target_attribute_p): Return false only if
25941 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25942 target("default") attribute.
25943 (sorted_attr_string): Change argument from const char * to tree,
25944 merge in all target attribute arguments rather than just one.
25945 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25946 instead of free. Avoid using strcat.
25947 (ix86_mangle_function_version_assembler_name): Mangle
25948 target("default") as if no target attribute is present. Adjust
25949 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25950 instead of xmalloc and XDELETEVEC instead of free.
25951 (ix86_function_versions): Don't return true if one of the decls
25952 doesn't have target attribute. If they don't and one of the decls
25953 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25954 sorted_attr_string caller. Use XDELETEVEC instead of free.
25955 (ix86_supports_function_versions): Remove.
25956 (make_name): Fix up formatting.
25957 (make_dispatcher_decl): Remove resolver_name and its initialization.
25958 Avoid leaking memory.
25959 (is_function_default_version): Return true if there is
25960 target("default") attribute rather than no target attribute at all.
25961 (make_resolver_func): Avoid leaking memory.
25962 (ix86_generate_version_dispatcher_body): Likewise.
25963 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
25964 * target.def (supports_function_versions): Remove.
25965 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
25966 * doc/tm.texi: Regenerated.
25968 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
25970 PR rtl-optimization/56144
25971 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
25972 for values with side effects.
25974 2013-01-30 Richard Biener <rguenther@suse.de>
25976 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
25977 (sparseset_pop): Likewise.
25978 * cfganal.c (compute_idf): Likewise. Increase work-stack size
25979 to be able to use quick_push in the worker loop.
25981 2013-01-30 Marek Polacek <polacek@redhat.com>
25983 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
25985 2013-01-30 Richard Biener <rguenther@suse.de>
25988 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
25990 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
25992 PR tree-optimization/56064
25993 * fixed-value.c (fixed_from_double_int): New function.
25994 * fixed-value.h (fixed_from_double_int): New prototype.
25995 (const_fixed_from_double_int): New static inline function.
25996 * fold-const.c (native_interpret_fixed): New static function.
25997 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
25998 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
25999 (native_encode_fixed): New static function.
26000 (native_encode_expr) <FIXED_CST>: Use it.
26001 (native_interpret_int): Move double_int worker code to...
26002 * double-int.c (double_int::from_buffer): ...this new static method.
26003 * double-int.h (double_int::from_buffer): Prototype it.
26005 2013-01-30 Richard Biener <rguenther@suse.de>
26007 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
26008 New pointer-map and obstack.
26009 (init_alias_vars): Allocate pointer-map and obstack.
26010 (delete_points_to_sets): Free them.
26011 (find_what_var_points_to): Cache result.
26012 (find_what_p_points_to): Adjust for changed interface of
26013 find_what_var_points_to.
26014 (compute_points_to_sets): Likewise.
26015 (ipa_pta_execute): Likewise.
26017 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26019 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
26020 * configure: Regenerate.
26021 * config.in: Regenerate.
26022 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
26023 #nobits/#progbits if supported.
26025 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
26028 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
26029 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
26030 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
26032 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26034 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
26035 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
26037 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26039 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
26040 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
26042 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26044 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
26046 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
26047 * config/arm/cortex-a7.md: New bypasses using
26048 arm_mac_accumulator_is_result.
26050 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26052 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
26053 (cortex_a7_neon_mla): Likewise.
26054 (cortex_a7_fpfmad): New reservation.
26055 (cortex_a7_fpmacs): Use ffmas and update required units.
26056 (cortex_a7_fpmuld): Update required units and latency.
26057 (cortex_a7_fpmacd): Likewise.
26058 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
26059 (cortex_a7_neon). Likewise.
26060 (bypass) Update participating units.
26062 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26064 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
26065 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
26067 * config/arm/vfp11.md (vfp_farith): Use ffmas.
26068 (vfp_fmul): Use ffmad.
26069 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
26070 (cortex_r4_fmacd): Use ffmad.
26071 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
26072 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
26073 (cortex_a9_fmacd): Use ffmad.
26074 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
26075 (cortex_a8_vfp_macd): Use ffmad.
26076 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
26077 (cortex_a5_fpmacd): Use ffmad.
26078 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
26079 (cortex_a15_vfp_macd): Use ffmad.
26080 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
26082 2013-01-29 Jason Merrill <jason@redhat.com>
26085 * varasm.c (default_assemble_visibility): Don't warn about
26086 visibility on artificial decls.
26088 2013-01-29 Richard Biener <rguenther@suse.de>
26090 PR tree-optimization/56113
26091 * tree-ssa-structalias.c (equiv_class_lookup): Also return
26093 (label_visit): Free duplicate bitmaps and record the leader instead.
26094 (perform_var_substitution): Adjust.
26096 2013-01-29 Richard Biener <rguenther@suse.de>
26098 PR tree-optimization/55270
26099 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
26100 the CFG, schedule loops for fixup.
26102 2013-01-29 Nick Clifton <nickc@redhat.com>
26104 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
26107 2013-01-28 Leif Ekblad <leif@rdos.net>
26109 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
26110 * config/i386/i386.h (TARGET_RDOS): New macro.
26111 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
26112 * config/i386/i386.c (ix86_option_override_internal): For 64bit
26113 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
26114 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
26115 DEFAULT_LARGE_SECTION_THRESHOLD.
26116 * config/i386/i386.md (R14_REG, R15_REG): New constants.
26117 * config/i386/rdos.h: New file.
26118 * config/i386/rdos64.h: New file.
26120 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
26123 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
26126 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26128 PR rtl-optimization/56117
26129 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
26130 call cselib_lookup_from_insn on the MEM before calling
26131 add_insn_mem_dependence.
26133 2013-01-28 Richard Biener <rguenther@suse.de>
26135 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26136 to a stmt that didn't have one.
26137 (copy_phis_for_bb): Likewise for PHI arguments.
26138 (copy_debug_stmt): Likewise for debug stmts.
26140 2013-01-28 Richard Biener <rguenther@suse.de>
26142 PR tree-optimization/56034
26143 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26144 (partition_builtin_p): Adjust.
26145 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26146 it is the last partition.
26147 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26148 up the vertex for the definition.
26149 (classify_partition): Classify whether a partition is a
26150 PKIND_REDUCTION, thus has uses outside of the loop.
26151 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26152 Merge all PKIND_REDUCTION partitions into the last partition.
26153 (tree_loop_distribution): Seed partitions from reductions as well.
26155 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26157 PR tree-optimization/56125
26158 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26159 pow(x,c) into sqrt(x) * powi(x, n/2) or
26160 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26161 optimizing for size.
26162 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26163 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26166 PR tree-optimization/56094
26167 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26168 to UNKNOWN_LOCATION while gimplifying expr.
26170 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26173 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26174 operand 0 in movabs insn template for -masm=intel asm alternative.
26175 (*movabs<mode>_2): Ditto for operand 1.
26177 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26180 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26183 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26185 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26188 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26190 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26191 Undef to avoid warning.
26193 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26195 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26196 * configure: Regenerate.
26198 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26200 PR tree-optimization/56098
26201 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26202 for stmts with volatile ops.
26203 (cond_store_replacement): Don't optimize if assign has volatile ops.
26204 (cond_if_else_store_replacement_1): Don't optimize if either
26205 then_assign or else_assign have volatile ops.
26206 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26209 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26211 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26213 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26215 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26216 missing ':' in asm example.
26218 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26220 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26221 entries into lane and laneq entries.
26222 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26223 Remove AdvSIMD scalar modes.
26224 (aarch64_sq<r>dmulh_laneq<mode>): New.
26225 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26227 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26228 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26229 * config/aarch64/iterators.md (VCOND): New.
26232 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26235 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26236 Add NULL LIBNAME argument to existing definitions.
26237 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26238 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26239 * config/avr/avr.c (DEF_BUILTIN): Same.
26240 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26241 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26242 implementation is available (DECL_ASSEMBLER_NAME is set).
26243 (avr_fold_absfx): New static function.
26244 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26245 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26246 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26247 AVR_BUILTIN_ABSLLK.
26248 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26249 (abshk, absk, abslk, absllk): Provide as static inline functions.
26251 2013-01-25 Marek Polacek <polacek@redhat.com>
26253 PR tree-optimization/56035
26254 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26256 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26258 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26259 (*movtf_internal_rex64): Add (!o,C) alternative
26260 (*movxf_internal_rex64): Ditto.
26261 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26263 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26265 * doc/invoke.texi: fix typo.
26266 * doc/objc.texi: fix typo.
26268 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26270 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26271 for the first two alternatives.
26273 2013-01-24 Diego Novillo <dnovillo@google.com>
26275 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26276 (ggc-zone.o): Remove.
26277 * configure.ac: Remove option --with-gc.
26278 * configure: Re-generate.
26279 * doc/install.texi: Remove documentation for --with-gc.
26280 * gengtype.c (write_enum_defn): Remove. Update all users.
26281 (write_Types_process_field): Remove generation of gt_e_* argument.
26282 (output_type_enum): Remove. Update all users.
26283 (write_enum_defn): Remove. Update all users.
26284 (enum alloc_zone): Remove. Update all users.
26285 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26286 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26287 Update all callers.
26288 (struct ptr_data): Remove field TYPE. Update all users.
26289 (gt_pch_note_object): Remove argument TYPE. Update all users.
26290 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26292 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26293 (struct alloc_zone): Remove.
26294 (ggc_internal_alloc_zone_stat): Remove.
26295 (ggc_internal_cleared_alloc_zone_stat): Remove.
26296 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26297 (ggc_pch_count_object): Remove last argument. Update all users.
26298 (ggc_pch_alloc_object): Remove last argument. Update all users.
26299 (struct alloc_zone): Remove.
26300 * ggc-zone.c: Remove.
26301 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26302 (struct alloc_zone): Remove.
26303 (ggc_alloc_typed_stat): Remove.
26304 (ggc_alloc_typed): Remove.
26305 (ggc_splay_alloc): Remove first argument.
26306 (rtl_zone): Remove. Update all users.
26307 (tree_zone): Remove. Update all users.
26308 (tree_id_zone): Remove. Update all users.
26309 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26310 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26311 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26312 * tree-ssanames.c: Remove references to zone allocator in comments.
26314 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26316 * config/avr/avr.c (avr_out_fract): Make register numbers that
26317 might be outside of source operand signed.
26319 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26321 * config/i386/constraints.md (Yf): New constraint.
26322 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26323 of f constraint to conditionaly disable x87 register preferences.
26324 (*movdf_internal): Ditto.
26325 (*movsf_internal): Ditto.
26327 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26329 PR inline-asm/55934
26330 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26331 that have operands with impossible constraints.
26332 Add a FIXME for a speed-up opportunity.
26333 * lra-constraints.c (process_alt_operands): Verify that a class
26334 selected from constraints on asms is valid for the operand mode.
26335 (curr_insn_transform): Remove incorrect comment.
26337 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26339 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26340 TOC operand is a valid symbol ref in the constant pool.
26342 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26344 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26346 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26349 * config/avr/stdfix.h: New file.
26350 * t-avr (stdfix-gcc.h): New rule to build it.
26351 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26353 2013-01-23 Kostya Serebryany <kcc@google.com>
26355 * config/darwin.h: remove dependency on
26356 CoreFoundation (asan on Mac OS).
26358 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26361 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26362 instead of cmpdi_operand for first comparison operand.
26363 Don't assert that comparison operands aren't both constants.
26365 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26367 * doc/install.texi (Downloading the Source): Update references to
26368 downloading separate components.
26370 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26372 * doc/extend.texi (__int128): Improve grammar.
26374 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26377 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26378 alternative to (o,r).
26379 (*movdi_internal_rex64): Remove (!o,n) alternative.
26380 (DImode immediate->memory splitter): Remove.
26381 (DImode immediate->memory peephole2): Remove.
26382 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26383 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26384 alternative to (!o,*r).
26385 (*movtf_internal_sse): New pattern.
26386 (*movxf_internal_rex64): New pattern.
26387 (*movxf_internal): Disable for TARGET_64BIT.
26388 (*movdf_internal_rex64): Remove (!o,F) alternative.
26390 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26392 PR middle-end/56074
26393 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26394 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26395 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26396 stmt locations where LOCATION_LOCUS of the stmt location is
26397 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26400 * config/i386/i386.md (UNSPEC_STOS): New.
26401 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26402 *strsetqi_1): Add UNSPEC_STOS.
26404 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26407 * doc/invoke.texi: Remove left over -Wsynth example.
26409 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26411 PR tree-optimization/56051
26412 * fold-const.c (fold_binary_loc): Don't fold
26413 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26414 a narrowing conversion, or widening conversion from signed
26417 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26419 PR rtl-optimization/56023
26420 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26421 dependent on debug instruction.
26423 2013-01-21 Martin Jambor <mjambor@suse.cz>
26425 PR middle-end/56022
26426 * function.c (allocate_struct_function): Call
26427 invoke_set_current_function_hook earlier.
26429 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26431 * reload1.c (init_reload): Only initialize reload_obstack
26432 during the first call.
26434 2013-01-21 Marek Polacek <polacek@redhat.com>
26436 * cfgloop.c (verify_loop_structure): Fix up grammar.
26438 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26440 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26441 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26443 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26446 * config/arm/marvell-pj4.md: Update copyright year.
26447 Fix up use of alu to alu_reg and simple_alu_imm.
26449 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26451 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26453 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26456 * lra-constraints.c (curr_insn_transform): Don't reuse original
26457 insn for secondary memory move when memory mode should be different.
26459 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26461 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26462 atomic_storedi_1): New patterns.
26464 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26466 btver2 pipeline descriptions.
26467 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26469 * config/i386/i386.md (btver2_decode): New type attributes.
26470 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26472 * config/i386/btver2.md: New file describing btver2 pipelines.
26474 2013-01-19 Andrew Pinski <apinski@cavium.com>
26476 PR tree-optimization/52631
26477 * tree-ssa-sccvn (visit_use): Before looking up the original
26478 statement, try looking up the simplified expression.
26480 2013-01-19 Anthony Green <green@moxielogic.com>
26482 * config/moxie/moxie.c (moxie_expand_prologue): Set
26483 current_function_static_stack_size.
26485 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26487 PR tree-optimization/56029
26488 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26489 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26491 2013-01-18 Sharad Singhai <singhai@google.com>
26493 PR tree-optimization/55995
26494 * dumpfile.c (dump_loc): Print location only if available.
26495 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26497 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26500 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26501 secondary memory move.
26502 (inherit_reload_reg): Use rclass instead of cl for
26503 check_secondary_memory_needed_p.
26505 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26507 PR middle-end/56015
26508 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26509 the case where writing real complex part of target modifies op1.
26511 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26513 * config/aarch64/aarch64-simd.md
26514 (aarch64_vcond_internal<mode>): Handle unordered cases.
26515 * config/aarch64/iterators.md (v_cmp_result): New.
26517 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26518 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26520 * config/arm/marvell-pj4.md: New file.
26521 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26522 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26523 (generic_vfp): Likewise.
26524 * config/arm/arm-cores.def: Add marvell-pj4.
26525 * config/arm/arm-tune.md: Regenerate.
26526 * config/arm/arm-tables.opt: Regenerate.
26527 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26528 * doc/invoke.texi: Document marvell-pj4.
26530 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26532 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26534 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26539 * var-tracking.c (negative_power_of_two_p): New.
26540 (global_get_addr_cache, local_get_addr_cache): New.
26541 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26542 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26544 (vt_stack_offset_p): Remove.
26545 (vt_canon_true_dep): Always canonicalize loc's address.
26546 (clobber_overlapping_mems): Make sure we have a MEM.
26547 (local_get_addr_clear_given_value): New.
26548 (val_reset): Clear local cached entries.
26549 (compute_bb_dataflow): Create and release the local cache.
26550 Disable duplicate MEMs clobbering.
26551 (emit_notes_in_bb): Clobber MEMs likewise.
26552 (vt_emit_notes): Create and release the local cache.
26553 (vt_initialize, vt_finalize): Create and release the global
26554 cache, respectively.
26555 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26557 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26559 PR libmudflap/53359
26560 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26561 not found in the symtab.
26563 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26566 PR rtl-optimization/55547
26567 PR rtl-optimization/53827
26570 * alias.c (offset_overlap_p): New, factored out of...
26571 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26572 the conservative special case for symbolic constants. Don't
26573 adjust zero sizes on alignment.
26575 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26577 PR rtl-optimization/52573
26578 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26579 REG_UNUSED for the same register.
26581 2013-01-17 Richard Biener <rguenther@suse.de>
26582 Marek Polacek <polacek@redhat.com>
26584 PR rtl-optimization/55833
26585 * loop-unswitch.c (unswitch_loops): Move loop verification...
26586 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26587 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26588 Set it to true when we're removing a loop from hierarchy tree in
26589 an irreducible region.
26590 (fix_bb_placements): Adjust caller.
26591 (fix_loop_placements): Likewise.
26593 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26595 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26596 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26597 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26598 Remove ID. Adjust comments.
26599 * config/avr/avr-c.c (avr_builtin_name): Remove.
26600 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26601 * config/avr/avr.c (avr_tolower): New static function.
26602 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26603 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26604 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26607 2013-01-17 Jan Hubicka <jh@suse.cz>
26609 PR tree-optimization/55273
26610 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26612 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26615 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26616 store through atomic_store<mode>_1.
26617 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26619 2013-01-17 Martin Jambor <mjambor@suse.cz>
26621 PR tree-optimizations/55264
26622 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26623 for virtual methods.
26624 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26625 virtual methods before inlining is over.
26626 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26628 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26631 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26633 PR rtl-optimization/56005
26634 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26635 pending reads for prefetch.
26637 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26639 * config/aarch64/aarch64.md
26640 (*cstoresi_neg_uxtw): New pattern.
26641 (*cmovsi_insn_uxtw): New pattern.
26642 (*<optab>si3_uxtw): New pattern.
26643 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26644 (*<optab>si3_insn_uxtw): New pattern.
26645 (*bswapsi2_uxtw): New pattern.
26647 2013-01-16 Richard Biener <rguenther@suse.de>
26649 * tree-inline.c (tree_function_versioning): Remove set but
26650 never used variable.
26652 2013-01-16 Richard Biener <rguenther@suse.de>
26654 PR tree-optimization/55964
26655 * tree-flow.h (rename_variables_in_loop): Remove.
26656 (rename_variables_in_bb): Likewise.
26657 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26658 (copy_loop_before): Adjust and delete update-ssa status.
26659 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26660 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26661 (rename_variables_in_loop): Remove.
26662 (slpeel_update_phis_for_duplicate_loop): Likewise.
26663 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26664 use available cfg machinery instead of duplicating it.
26665 Update PHI nodes and perform poor-mans SSA update here.
26666 (slpeel_tree_peel_loop_to_edge): Adjust.
26668 2013-01-16 Richard Biener <rguenther@suse.de>
26670 PR tree-optimization/54767
26671 PR tree-optimization/53465
26672 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26673 (vrp_visit_phi_node): For PHI arguments coming via backedges
26674 drop all symbolical range information.
26675 (execute_vrp): Compute backedges.
26677 2013-01-16 Richard Biener <rguenther@suse.de>
26679 * doc/install.texi: Update CLooG and ISL requirements to
26682 2013-01-16 Christian Bruel <christian.bruel@st.com>
26685 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26686 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26687 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26689 2013-01-16 DJ Delorie <dj@redhat.com>
26691 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26692 (UNSPECV_SP_SWITCH_E): New.
26693 (sp_switch_1): Change to an unspec.
26694 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26697 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26699 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26700 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26701 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26702 (expand_mem_thread_fence): Ditto.
26703 (expand_mem_signal_fence): Ditto.
26704 (expand_atomic_load): Ditto.
26705 (expand_atomic_store): Ditto.
26707 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26709 PR rtl-optimization/55547
26710 PR rtl-optimization/53827
26713 * alias.c (memrefs_conflict_p): Set sizes to negative after
26716 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26719 * function.c (thread_prologue_and_epilogue_insns): Always
26720 add crtl->drap_reg to set_up_by_prologue.set, even if
26721 stack_realign_drap is false.
26723 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26725 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26726 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26727 *call): Fix indention.
26729 2013-01-15 Tom de Vries <tom@codesourcery.com>
26732 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26735 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26737 PR rtl-optimization/55153
26738 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26740 2013-01-15 Martin Jambor <mjambor@suse.cz>
26742 PR tree-optimization/55920
26743 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26744 accesses as grp_to_be_debug_replaced.
26746 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26748 PR tree-optimization/55920
26749 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26750 there is non-useless type conversion needed from debug rhs to lhs,
26751 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26753 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26754 Mikael Pettersson <mikpe@it.uu.se>
26757 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26759 (ASM_OUTPUT_CASE_LABEL): Remove.
26760 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26761 * final.c (shorten_branches): Update alignment of labels before
26762 jump tables if CASE_VECTOR_SHORTEN_MODE.
26764 2013-01-15 Richard Biener <rguenther@suse.de>
26767 * system.h: Do not include gmp.h for building host tools.
26769 2013-01-15 Richard Biener <rguenther@suse.de>
26771 PR middle-end/55882
26772 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26773 account for bitpos when computing alignment.
26775 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26777 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26778 (ix86_target_macros_internal): Likewise.
26780 * config/i386/i386.c (m_CORE2I7): Removed.
26781 (m_CORE_HASWELL): New macro.
26782 (m_CORE_ALL): Likewise.
26783 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26784 (initial_ix86_arch_features): Likewise.
26785 (processor_target_table): Initializations for Core avx2.
26786 (cpu_names): New names "core-avx2".
26787 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26788 PROCESSOR_CORE_HASWELL.
26789 (ix86_issue_rate): New case.
26790 (ia32_multipass_dfa_lookahead): Likewise.
26791 (ix86_sched_init_global): Likewise.
26793 * config/i386/i386.h (TARGET_HASWELL): New macro.
26794 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26795 (processor_type): New PROCESSOR_HASWELL.
26797 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26799 PR tree-optimization/55955
26800 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26801 *SHIFT_EXPR and *ROTATE_EXPR codes.
26803 PR tree-optimization/48766
26804 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26805 -ftrapv disable -fwrapv.
26807 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26810 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26811 etc. to 1 and not to __flash.
26812 Use LL suffix for __INT24_MAX__ with -mint8.
26813 Use ULL suffix for __UINT24_MAX__ with -mint8.
26815 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26817 * config/avr/avr-arch.h
26818 (struct base_arch_s): Use typedef avr_arch_t instead.
26819 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26820 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26821 * config/avr/avr.c: Same.
26822 * config/avr/avr-devices.c: Same.
26823 * config/avr/driver-avr.c: Same.
26824 * config/avr/gen-avr-mmcu-texi.c: Same.
26825 * config/avr/avr-mcus.def: Adjust comment.
26827 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26829 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26830 * config/aarch64/iterators.md (VALLDI): New.
26832 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26833 Andi Kleen <ak@linux.intel.com>
26836 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26837 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26840 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26842 * config/avr/avr-stdint.h: Remove trailing blanks.
26843 * config/avr/avr-log.h: Same.
26844 * config/avr/avr-arch.h: Same.
26845 * config/avr/avr-devices.c: Same.
26846 * config/avr/avr-dimode.md: Same.
26847 * config/avr/predicates.md: Same.
26848 * config/avr/avr-c.c: Same. And fix typo.
26850 * config/avr/avr-protos.h: Same. And:
26851 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26852 (init_cumulative_args): Rename to avr_init_cumulative_args.
26853 (expand_prologue): Rename to avr_expand_prologue.
26854 (expand_epilogue): Rename to avr_expand_epilogue.
26855 (adjust_insn_length): Rename to avr_adjust_insn_length.
26856 (notice_update_cc): Rename to avr_notice_update_cc.
26857 (final_prescan_insn): Rename to avr_final_prescan_insn.
26858 * config/avr/avr.c: Same.
26859 * config/avr/avr.h: Same.
26860 * config/avr/avr.md: Remove trailing blanks.
26861 (prologue): Use avr_expand_prologue.
26862 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26864 2013-01-14 Richard Biener <rguenther@suse.de>
26866 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26867 verify_location, collect_subblocks): New functions.
26868 (verify_gimple_in_cfg): Verify that locations only reference
26869 BLOCKs in the functions BLOCK tree.
26871 2013-01-14 Richard Biener <rguenther@suse.de>
26873 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26875 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26877 (insert_out_of_ssa_copy_on_edge): Likewise.
26878 (rewrite_close_phi_out_of_ssa): Likewise.
26879 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26881 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26882 propagated constants.
26883 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26886 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26888 * config/avr/avr-modes.def: Add GPL copyright notice.
26890 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26892 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26893 MEMMODEL_MASK to determine memory model.
26894 (atomic_store<mode>): Ditto from operands[2].
26895 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26897 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26900 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26901 (fold_gimple_assign): Don't call unshare_expr here.
26902 (fold_ctor_reference): Call unshare_expr.
26904 2013-01-13 Terry Guo <terry.guo@arm.com>
26906 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26907 * doc/fragments.texi: Document MULTILIB_REUSE.
26908 * gcc.c (multilib_reuse): New internal spec.
26909 (set_multilib_dir): Also search multilib from multilib_reuse.
26910 * genmultilib (tmpmultilib3): Refactor code.
26911 (tmpmultilib4): Ditto.
26912 (multilib_reuse): New multilib argument.
26914 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26916 * Makefile.in: Update copyright.
26918 2013-01-12 Tom de Vries <tom@codesourcery.com>
26920 PR middle-end/55890
26921 * calls.c (expand_call): Check if arg_nr is valid.
26923 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26925 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26926 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26927 documentation. Add missing '__' in front of
26928 __builtin_ia32_packssdw256.
26930 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26933 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26934 NO_REGS for larl operands.
26935 (s390_reload_larl_operand): Use s390_load_address instead of
26938 2013-01-11 Richard Biener <rguenther@suse.de>
26940 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26941 (verify_node_sharing): ... here.
26942 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26944 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26946 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26947 Substitute TREECHECKING.
26948 * configure: Regenerate.
26949 * Makefile.in (TREECHECKING): New.
26951 2013-01-11 Richard Guenther <rguenther@suse.de>
26953 PR tree-optimization/44061
26954 * tree-vrp.c (extract_range_basic): Compute zero as
26955 value-range for __builtin_constant_p of function parameters.
26957 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
26959 Update copyright years.
26961 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
26963 PR rtl-optimization/55672
26964 * lra-eliminations.c (mark_not_eliminable): Permit addition with
26965 const to be eliminable.
26967 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
26969 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
26970 * configure: Regenerate.
26972 2013-01-10 Richard Biener <rguenther@suse.de>
26974 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
26976 2013-01-10 Richard Biener <rguenther@suse.de>
26979 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
26980 locations for virtual PHI arguments.
26981 (rewrite_update_phi_arguments): Likewise.
26983 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
26985 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
26988 2013-01-10 Jakub Jelinek <jakub@redhat.com>
26990 PR tree-optimization/55921
26991 * tree-complex.c (expand_complex_asm): New function.
26992 (expand_complex_operations_1): Call it for GIMPLE_ASM.
26994 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26997 * config/s390/s390.c (s390_symref_operand_p)
26998 (s390_loadrelative_operand_p): Merge the two functions.
26999 (s390_check_qrst_address, print_operand_address): Add parameters
27000 to s390_loadrelative_operand_p invokation.
27001 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
27002 (s390_reload_larl_operand, s390_secondary_reload): Use
27003 s390_loadrelative_operand_p instead of s390_symref_operand_p.
27004 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
27006 2013-01-09 Mike Stump <mikestump@comcast.net>
27008 * dse.c (record_store): Remove unnecessary assert.
27010 2013-01-09 Jan Hubicka <jh@suse.cz>
27012 PR tree-optimization/55569
27013 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
27014 * cfgloop.h (scale_loop_profile): Likewise.
27016 2013-01-09 Jan Hubicka <jh@suse.cz>
27019 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
27021 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
27023 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
27025 PR middle-end/55114
27026 * expr.h (maybe_emit_group_store): Declare.
27027 * expr.c (maybe_emit_group_store): New function.
27028 * builtins.c (expand_builtin_int_roundingfn): Call it.
27029 (expand_builtin_int_roundingfn_2): Likewise.
27031 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
27033 PR rtl-optimization/55829
27034 * lra-constraints.c (match_reload): Add code for absent output.
27035 (curr_insn_transform): Add code for reloads of matched inputs
27038 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
27040 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
27041 attribute of movddup insn to DF.
27042 (*vec_interleave_lowv2df): Ditto.
27043 (vec_dupv2df): Ditto.
27045 2013-01-09 Jan Hubicka <jh@suse.cz>
27047 PR tree-optimiation/55875
27048 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
27049 EVERY_ITERATION parameter.
27050 (number_of_iterations_exit): Check if exit is executed every iteration.
27051 (idx_infer_loop_bounds): Similarly here.
27052 (n_of_executions_at_most): Simplify
27053 to only test for cases where statement is dominated by the
27054 particular bound; handle correctly the "postdominance" test.
27055 (scev_probably_wraps_p): Use max loop iterations info
27056 as a global bound first.
27058 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
27059 Nick Clifton <nickc@redhat.com>
27061 * config/v850/v850.md (cbranchsf4): New pattern.
27062 (cstoresf4): New pattern.
27063 (cbranchdf4): New pattern.
27064 (cstoredf4): New pattern.
27065 (movsicc): Disallow floating point comparisons.
27066 (cmpsf_le_insn): Fix order of operators.
27067 (cmpsf_lt_insn): Likewise.
27068 (cmpsf_eq_insn): Likewise.
27069 (cmpdf_le_insn): Likewise.
27070 (cmpdf_lt_insn): Likewise.
27071 (cmpdf_eq_insn): Likewise.
27072 (cmpsf_ge_insn): Use LE comparison.
27073 (cmpdf_ge_insn): Likewise.
27074 (cmpsf_gt_insn): Use LT comparison.
27075 (cmpdf_gt_insn): Likewise.
27076 (cmpsf_ne_insn): Delete pattern.
27077 (cmpdf_ne_insn): Delete pattern.
27078 * config/v850/v850.c (v850_gen_float_compare): Use
27079 gen_cmpdf_eq_insn for NE comparison.
27080 (v850_float_z_comparison_operator)
27081 (v850_float_nz_comparison_operator): Move from here ...
27082 * config/v850/predicates.md: ... to here. Move GT and GE
27083 comparisons into v850_float_z_comparison_operator.
27084 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
27086 (v850_float_nz_comparison_operator): Likewise.
27088 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27090 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
27091 with calls to gen_insvsi/gen_insvdi.
27093 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27095 * config/i386/i386.c (initial_ix86_tune_features): Set up
27096 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
27098 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
27099 Jakub Jelinek <jakub@redhat.com>
27101 PR tree-optimization/48189
27102 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
27103 If nitercst is 0, don't predict the exit edge.
27105 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27107 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
27108 in asm_fprintf with reg_names.
27109 (aarch64_print_operand_address): Likewise.
27110 (aarch64_return_addr): Likewise.
27111 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
27113 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27115 * config/pa/pa.h (VAL_U6_BITS_P): Define.
27116 (INT_U6_BITS): Likewise.
27117 * config/pa/predicates.md (uint6_operand): New predicate.
27118 (shift5_operand, shift6_operand): Likewise.
27119 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
27121 (lshrdi3): Use shift6_operand.
27122 (shrpsi4, shrpdi4): New insn patterns.
27123 (extzv): Delete expander.
27124 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
27125 predicates in unamed zero extract patterns. Tighten common constraint.
27126 (extv): Delete expander.
27127 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
27128 predicates in unamed sign extract patterns. Tighten common constraint.
27129 (insv): Delete expander.
27130 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
27131 predicates in unamed insert patterns. Tighten common constraint.
27132 Change uint32_operand predicate to uint6_operand predicate in unamed
27133 DImode pattern to insert constant values of type 1...1xxxx.
27135 2013-01-04 Jan Hubicka <jh@suse.cz>
27137 PR tree-optimization/55823
27138 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27141 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27142 Uros Bizjak <ubizjak@gmail.com>
27144 PR rtl-optimization/55845
27145 * df-problems.c (can_move_insns_across): Stop scanning at
27146 volatile_insn_p source instruction or give up if
27147 across_from .. across_to range contains any volatile_insn_p
27150 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27152 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27153 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27155 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27156 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27158 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27161 * asan.c (asan_clear_shadow): New function.
27162 (asan_emit_stack_protection): Use it.
27164 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27166 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27167 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27168 with tab instead of space.
27170 2013-01-08 Nick Clifton <nickc@redhat.com>
27172 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27173 register bank 0 at the start of an interrupt handler.
27174 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27177 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27179 * config/aarch64/aarch64-simd.md
27180 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27181 (aarch64_simd_bsl): Likewise.
27182 (aarch64_vcond_internal<mode>): Likewise.
27183 (vcond<mode><mode>): Likewise.
27184 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27185 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27187 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27189 * config/aarch64/aarch64-builtins.c
27190 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27192 2013-01-08 Martin Jambor <mjambor@suse.cz>
27195 * tree-sra.c (analyze_access_subtree): Return true also after
27196 potentially creating a debug-only replacement.
27198 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27200 PR middle-end/55890
27201 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27203 PR tree-optimization/54120
27204 * tree-vrp.c (range_fits_type_p): Don't allow
27205 src_precision < precision from signed vr to unsigned_p
27206 if vr->min or vr->max is negative.
27207 (simplify_float_conversion_using_ranges): Test can_float_p
27208 against CODE_FOR_nothing.
27210 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27211 Richard Biener <rguenther@suse.de>
27213 PR middle-end/55851
27214 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27215 types instead of just INTEGER_TYPE types.
27217 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27219 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27222 2013-01-07 Steve Ellcey <sellcey@mips.com>
27225 * config/mips/mips.opt: Change mad to mmad to match documentation.
27227 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27230 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27231 .progmemx.data now.
27233 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27236 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27237 (avr_addrspace_t): Add .section_name field.
27238 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27240 (avr_addrspace): Same. Initialize .section_name. Remove last
27241 NULL entry. Put __memx into .progmemx.data.
27242 (progmem_section_prefix): Remove.
27243 (avr_asm_init_sections): No need to initialize progmem_section.
27244 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27245 section name prefix.
27246 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27247 retrieve the progmem section.
27248 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27249 boundary to run over avr_addrspace[].
27250 (avr_register_target_pragmas): Ditto.
27252 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27254 * varasm.c (output_constant_def_contents): For asan_protect_global
27255 protected strings, adjust DECL_ALIGN if needed, before testing for
27257 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27258 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27260 (output_object_block): For asan protected decls, emit asan padding
27261 after their contents.
27262 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27263 (asan_finish_file): Test it here instead.
27265 2013-01-07 Nick Clifton <nickc@redhat.com>
27266 Matthias Klose <doko@debian.org>
27267 Doug Kwan <dougkwan@google.com>
27268 H.J. Lu <hongjiu.lu@intel.com>
27271 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27273 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27275 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27277 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27279 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27281 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27284 * doc/install.texi (Cross-Compiler-Specific Options): Document
27287 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27289 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27290 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27291 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27292 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27293 vqmovun_high_s64): Fix source operand number and update copyright.
27295 2013-01-07 Richard Biener <rguenther@suse.de>
27297 PR middle-end/55890
27298 * gimple.h (gimple_call_builtin_p): New overload.
27299 * gimple.c (validate_call): New function.
27300 (gimple_call_builtin_p): Likewise.
27301 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27302 Use gimple_call_builtin_p.
27303 (find_func_clobbers): Likewise.
27304 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27305 (strlen_optimize_stmt): Likewise.
27307 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27309 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27310 (vld1q_dup_*): Likewise.
27311 (vld1_*): Likewise.
27312 (vld1q_*): Likewise.
27313 (vld1_lane_*): Likewise.
27314 (vld1q_lane_*): Likewise.
27316 2013-01-07 Richard Biener <rguenther@suse.de>
27318 * lto-streamer.h (LTO_minor_version): Bump to 2.
27320 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27322 * config/aarch64/aarch64-protos.h
27323 (aarch64_const_double_zero_rtx_p): Rename to...
27324 (aarch64_float_const_zero_rtx_p): ...this.
27325 (aarch64_float_const_representable_p): New.
27326 (aarch64_output_simd_mov_immediate): Likewise.
27327 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27328 move immediate case.
27329 * config/aarch64/aarch64.c
27330 (aarch64_const_double_zero_rtx_p): Rename to...
27331 (aarch64_float_const_zero_rtx_p): ...this.
27332 (aarch64_print_operand): Allow printing of new constants.
27333 (aarch64_valid_floating_const): New.
27334 (aarch64_legitimate_constant_p): Check for valid floating-point
27336 (aarch64_simd_valid_immediate): Likewise.
27337 (aarch64_vect_float_const_representable_p): New.
27338 (aarch64_float_const_representable_p): Likewise.
27339 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27340 (aarch64_output_simd_mov_immediate): New.
27341 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27342 (*movdf_aarch64): Likewise.
27343 * config/aarch64/constraints.md (Ufc): New.
27344 (Y): call aarch64_float_const_zero_rtx.
27345 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27347 2013-01-07 Richard Biener <rguenther@suse.de>
27349 PR tree-optimization/55888
27350 PR tree-optimization/55862
27351 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27352 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27353 not if it is contained therein.
27355 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27357 * config/avr/t-avr: Typo.
27359 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27362 * config/avr/t-avr: Don't automatically rebuild
27363 $(srcdir)/config/avr/t-multilib
27364 $(srcdir)/config/avr/avr-tables.opt
27365 $(srcdir)/doc/avr-mmcu.texi
27366 (avr-mcus): New phony target to build them on request.
27367 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27368 * avr/avr-mcus.def: Adjust comments.
27370 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27372 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27374 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27376 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27378 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27380 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27382 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27384 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27385 to generate profiling.
27386 * config/rs6000/aix64.h (LIB_SPEC): Same.
27388 2013-01-04 Andrew Pinski <apinski@cavium.com>
27390 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27392 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27394 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27396 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27398 (ix86_expand_move): Ditto.
27399 (ix86_zero_extend_to_Pmode): Ditto.
27400 (ix86_expand_call): Ditto.
27401 (ix86_expand_special_args_builtin): Ditto.
27402 (ix86_expand_builtin): Ditto.
27404 2013-01-04 Richard Biener <rguenther@suse.de>
27406 PR tree-optimization/55862
27407 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27408 translating them through PHI nodes.
27410 2013-01-04 Martin Jambor <mjambor@suse.cz>
27412 PR tree-optimization/55755
27413 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27414 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27416 2013-01-04 Marek Polacek <polacek@redhat.com>
27418 PR middle-end/55859
27419 * opts.c (default_options_optimization): Clarify error message.
27421 2013-01-04 Richard Biener <rguenther@suse.de>
27423 PR middle-end/55863
27424 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27427 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27430 * config/pa/pa.md (movsi): Revert previous change.
27431 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27434 2013-01-03 Richard Henderson <rth@redhat.com>
27436 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27437 after eliminating TLS symbols.
27439 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27442 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27443 * graphite-poly.c (debug_gmp_value): Likewise.
27445 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27448 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27449 selected code model, define __code_mode_small__, __code_model_medium__,
27450 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27451 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27452 xchg temporary register with %k. Declare temporary register as
27454 [__x86_64__]: For medium and large code models, preserve %rbx register.
27456 2013-01-03 Richard Biener <rguenther@suse.de>
27458 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27459 (dump_subscript): Adjust.
27460 (finalize_ddr_dependent): Do not dump redundant info.
27461 (analyze_siv_subscript): Adjust.
27462 (subscript_dependence_tester): Likewise.
27463 (compute_affine_dependence): Likewise.
27465 2013-01-03 Richard Biener <rguenther@suse.de>
27468 2013-01-03 Richard Biener <rguenther@suse.de>
27470 PR tree-optimization/55857
27471 * tree-vect-stmts.c (vectorizable_load): Do not setup
27472 re-alignment for invariant loads.
27474 2013-01-02 Richard Biener <rguenther@suse.de>
27476 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27477 invariant load do not generate a vector load from the scalar location.
27479 2013-01-03 Richard Biener <rguenther@suse.de>
27481 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27482 for not vectorizing.
27483 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27484 not build INDIRECT_REFs, call get_name once only.
27485 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27486 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27488 2013-01-03 Richard Biener <rguenther@suse.de>
27490 PR tree-optimization/55857
27491 * tree-vect-stmts.c (vectorizable_load): Do not setup
27492 re-alignment for invariant loads.
27494 2013-01-03 Richard Biener <rguenther@suse.de>
27497 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27498 prefer a built-in decl.
27500 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27502 * gcc.c (process_command): Update copyright notice dates.
27503 * gcov.c (print_version): Likewise.
27504 * gcov-dump.c (print_version): Likewise.
27506 PR rtl-optimization/55838
27507 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27508 iv0.step, iv1.step and step.
27510 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27511 Marc Glisse <marc.glisse@inria.fr>
27513 PR tree-optimization/55832
27514 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27515 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27516 integer_{one,zero}_node.
27518 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27521 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27522 * var-tracking.c (reverse_op): Don't add reverse ops to
27523 VALUEs that have already
27524 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27526 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27528 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27530 2013-01-02 Teresa Johnson <tejohnson@google.com>
27532 * dumpfile.c (dump_loc): Print filename with location.
27533 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27534 new location_t parameter to emit complete unroll message with
27535 new dump framework.
27536 (canonicalize_loop_induction_variables): Compute loops location
27537 and pass to try_unroll_loop_completely.
27538 * loop-unroll.c (report_unroll_peel): New function.
27539 (peel_loops_completely): Use new dump format with location
27540 for main dumpfile message, and invoke report_unroll_peel on success.
27541 (decide_unrolling_and_peeling): Ditto.
27542 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27543 by report_unroll_peel.
27544 (decide_peel_completely): Ditto.
27545 (decide_unroll_constant_iterations): Ditto.
27546 (decide_unroll_runtime_iterations): Ditto.
27547 (decide_peel_simple): Ditto.
27548 (decide_unroll_stupid): Ditto.
27549 * cfgloop.c (get_loop_location): New function.
27550 * cfgloop.h (get_loop_location): Declare.
27552 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27554 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27557 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27559 PR middle-end/55198
27560 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27561 BLKmode objects when EXPAND_MEMORY is specified.
27563 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27565 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27567 (fold_builtin_cpu): Do not share cpu model decls across statements.
27569 2013-01-02 Jason Merrill <jason@redhat.com>
27572 * tree.c (build_array_type_1): Revert earlier change.
27574 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27576 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27578 * config/aarch64/aarch64-tune.md: Re-generate.
27580 2013-01-02 Richard Biener <rguenther@suse.de>
27582 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27583 invariant load do not generate a vector load from the scalar location.
27585 2013-01-02 Richard Biener <rguenther@suse.de>
27588 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27589 * configure: Regenerate.
27591 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27593 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27594 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27595 (expand_builtin_int_roundingfn_2): Keep the original target around
27596 for the fallback case.
27598 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27600 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27601 to be clear for sign changes.
27603 2013-01-01 Jan Hubicka <jh@suse.cz>
27605 * ipa-inline-analysis.c: Fix formatting.
27607 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27609 PR tree-optimization/55831
27610 * tree-vect-loop.c (get_initial_def_for_induction): Use
27611 gsi_after_labels instead of gsi_start_bb.
27613 Copyright (C) 2013 Free Software Foundation, Inc.
27615 Copying and distribution of this file, with or without modification,
27616 are permitted in any medium without royalty provided the copyright
27617 notice and this notice are preserved.