1 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
3 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
5 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
8 2013-08-24 Tobias Burnus <burnus@net-b.de>
11 * tree-cfg.c (replace_loop_annotate): New function.
12 (execute_build_cfg): Call it.
13 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
14 * internal-fn.c (expand_ANNOTATE): New function.
15 * internal-fn.def (ANNOTATE): Define as new internal function.
16 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
17 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
18 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
19 * doc/extend.texi (Pragmas): Document #pragma ivdep.
20 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
22 2013-10-17 Ian Bolton <ian.bolton@arm.com>
23 Marcus Shawcroft <marcus.shawcroft@arm.com>
25 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
26 Special case reload SP+C into none GENERAL_REGS.
28 2013-10-24 Michael Matz <matz@suse.de>
30 * gengtype.c (is_file_equal): Check that files will be same
33 2013-10-25 Christian Bruel <christian.bruel@st.com>
35 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
36 * gcc/config/sh/t-sh (sh-mem.o): New rule.
37 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
38 (sh4_expand_cmpstr): New function.
39 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c
40 * gcc/config/sh/sh-protos.h (sh4_expand_cmpstr): Declare.
41 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
44 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
46 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
47 * Makefile.in (CXXDEPMODE): Assign and change users.
49 * configure: Regenerate.
51 2013-10-23 David Malcolm <dmalcolm@redhat.com>
53 * gengtype-parse.c (require_without_advance): New.
54 (type): For GTY-marked types that are not GTY((user)), parse any
55 base classes, requiring them to be single-inheritance, and not
56 be templates. For non-GTY-marked types and GTY((user)),
57 continue to skip over any C++ inheritance specification.
58 * gengtype-state.c (state_writer::write_state_struct_type):
59 Write base class of type (if any).
60 (read_state_struct_type): Read base class of type (if any).
61 * gengtype.c (new_structure): Add a "base_class" parameter.
62 (create_optional_field_): Update for new parameter to
64 (adjust_field_rtx_def): Likewise.
65 (adjust_field_tree_exp): Likewise.
66 * gengtype.h (struct type): Add "base_class" field to the s
68 (new_structure): Add "base" parameter.
70 2013-10-23 Sriraman Tallam <tmsriram@google.com>
73 * config/i386/i386.c (ix86_option_override_internal):
74 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
75 (ix86_valid_target_attribute_tree):
76 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
77 Change TARGET_SSE to TARGET_SSE_P (opts->...)
79 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
81 * tree-ssa-loop.h: Remove include files.
82 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
83 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
84 * graphite-clast-to-gimple.c: Likewise.
85 * graphite-scop-detection.c: Likewise.
86 * graphite-sese-to-poly.c: Likewise.
87 * ipa-inline-analysis.c: Likewise.
88 * ipa-pure-const.c: Likewise.
89 * loop-init.c: Likewise.
91 * predict.c: Likewise.
92 * tree-cfg.c: Likewise.
93 * tree-cfgcleanup.c: Likewise.
94 * tree-chrec.c: Likewise.
95 * tree-data-ref.c: Likewise.
96 * tree-loop-distribution.c: Likewise.
97 * tree-parloops.c: Likewise.
98 * tree-predcom.c: Likewise.
99 * tree-scalar-evolution.c: Likewise.
100 * tree-ssa-address.c: Likewise.
101 * tree-ssa.c: Likewise.
102 * tree-ssa-dce.c: Likewise.
103 * tree-ssa-loop.c: Likewise.
104 * tree-ssa-loop-im.c: Likewise.
105 * tree-ssa-loop-ivcanon.c: Likewise.
106 * tree-ssa-loop-ivopts.c: Likewise.
107 * tree-ssa-loop-manip.c: Likewise.
108 * tree-ssa-loop-niter.c: Likewise.
109 * tree-ssa-loop-prefetch.c: Likewise.
110 * tree-ssa-loop-unswitch.c: Likewise.
111 * tree-ssa-reassoc.c: Likewise.
112 * tree-vect-data-refs.c: Likewise.
113 * tree-vect-loop.c: Likewise.
114 * tree-vect-loop-manip.c: Likewise.
115 * tree-vectorizer.c: Likewise.
116 * tree-vect-stmts.c: Likewise.
117 * tree-vrp.c: Likewise.
119 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
121 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
123 2013-10-23 Jakub Jelinek <jakub@redhat.com>
125 PR tree-optimization/58775
126 PR tree-optimization/58791
127 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
128 (insert_stmt_after): Rewritten, don't move the stmt, but really
130 (get_stmt_uid_with_default): Remove.
131 (build_and_add_sum): Use insert_stmt_after and
132 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only
134 (update_range_test): Set uid on stmts added by
135 force_gimple_operand_gsi. Don't immediately modify statements
136 in inter-bb optimization, just update oe->op values.
137 (optimize_range_tests): Return bool whether any changed have
139 (update_ops): New function.
140 (struct inter_bb_range_test_entry): New type.
141 (maybe_optimize_range_tests): Perform statement changes here.
142 (not_dominated_by, appears_later_in_bb, get_def_stmt,
143 ensure_ops_are_available): Remove.
144 (find_insert_point): Rewritten.
145 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
146 return LHS of the (new resp. old) stmt. Don't call
147 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
148 instead of last, move new stmt at the right place.
149 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
150 (negate_value): Likewise. Set uids.
151 (break_up_subtract_bb): Initialize uids.
152 (reassociate_bb): Adjust rewrite_expr_tree caller.
153 (do_reassoc): Don't call renumber_gimple_stmt_uids.
155 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
158 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
159 TARGET_32BIT final condition.
160 (mulsi3_internal2 and splitter): Same.
162 2013-10-23 Jeff Law <law@redhat.com>
164 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
165 through joiner blocks with abnormal outgoing edges.
167 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
168 Add parameter JOINERS, to allow/disallow threading through joiner
170 (thread_block): New. Call thread_block_1.
171 (mark_threaded_blocks): Remove code to filter out certain cases
172 of threading through joiner blocks.
173 (thread_through_all_blocks): Document how we can have a dangling
174 edge AUX field and clear it.
176 2013-10-23 Ian Lance Taylor <iant@google.com>
178 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
179 (C++ Dialect Options): Likewise.
180 (Optimize Options): Likewise.
182 2013-10-23 Tom de Vries <tom@codesourcery.com>
184 PR tree-optimization/58805
185 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
187 2013-10-23 Jakub Jelinek <jakub@redhat.com>
189 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
190 sequence based on get_range_info returned range.
192 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
194 * tree-ssa.h: Remove all #include's
195 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
196 * alias.c: Move required includes from tree-ssa.h.
198 * builtins.c: Likewise.
200 * cfgexpand.c: Likewise.
201 * cfghooks.c: Likewise.
202 * cfgloop.c: Likewise.
203 * cfgloopmanip.c: Likewise.
204 * cgraph.c: Likewise.
205 * cgraphbuild.c: Likewise.
206 * cgraphclones.c: Likewise.
207 * cgraphunit.c: Likewise.
209 * except.c: Likewise.
212 * fold-const.c: Likewise.
213 * ggc-page.c: Likewise.
214 * gimple-builder.c: Likewise.
215 * gimple-fold.c: Likewise.
216 * gimple-iterator.c: Likewise.
217 * gimple-low.c: Likewise.
218 * gimple-pretty-print.c: Likewise.
219 * gimple-ssa-strength-reduction.c: Likewise.
220 * gimple-streamer-in.c: Likewise.
221 * gimple-streamer-out.c: Likewise.
222 * gimplify.c: Likewise.
223 * graphite-blocking.c: Likewise.
224 * graphite-clast-to-gimple.c: Likewise.
225 * graphite-dependences.c: Likewise.
226 * graphite-interchange.c: Likewise.
227 * graphite-optimize-isl.c: Likewise.
228 * graphite-poly.c: Likewise.
229 * graphite-scop-detection.c: Likewise.
230 * graphite-sese-to-poly.c: Likewise.
231 * graphite.c: Likewise.
232 * ipa-cp.c: Likewise.
233 * ipa-inline-analysis.c: Likewise.
234 * ipa-inline-transform.c: Likewise.
235 * ipa-inline.c: Likewise.
236 * ipa-prop.c: Likewise.
237 * ipa-pure-const.c: Likewise.
238 * ipa-reference.c: Likewise.
239 * ipa-split.c: Likewise.
240 * ipa-utils.c: Likewise.
241 * loop-init.c: Likewise.
242 * lto-cgraph.c: Likewise.
243 * lto-section-in.c: Likewise.
244 * lto-section-out.c: Likewise.
245 * lto-streamer-in.c: Likewise.
246 * lto-streamer-out.c: Likewise.
247 * lto-streamer.c: Likewise.
248 * omp-low.c: Likewise.
249 * passes.c: Likewise.
250 * predict.c: Likewise.
251 * print-tree.c: Likewise.
252 * profile.c: Likewise.
254 * targhooks.c: Likewise.
255 * tracer.c: Likewise.
256 * trans-mem.c: Likewise.
257 * tree-call-cdce.c: Likewise.
258 * tree-cfg.c: Likewise.
259 * tree-cfgcleanup.c: Likewise.
260 * tree-chrec.c: Likewise.
261 * tree-complex.c: Likewise.
262 * tree-data-ref.c: Likewise.
263 * tree-dfa.c: Likewise.
264 * tree-eh.c: Likewise.
265 * tree-emutls.c: Likewise.
266 * tree-if-conv.c: Likewise.
267 * tree-inline.c: Likewise.
268 * tree-into-ssa.c: Likewise.
269 * tree-loop-distribution.c: Likewise.
270 * tree-mudflap.c: Likewise.
271 * tree-nested.c: Likewise.
272 * tree-nrv.c: Likewise.
273 * tree-object-size.c: Likewise.
274 * tree-outof-ssa.c: Likewise.
275 * tree-parloops.c: Likewise.
276 * tree-phinodes.c: Likewise.
277 * tree-predcom.c: Likewise.
278 * tree-pretty-print.c: Likewise.
279 * tree-profile.c: Likewise.
280 * tree-scalar-evolution.c: Likewise.
281 * tree-sra.c: Likewise.
282 * tree-ssa-address.c: Likewise.
283 * tree-ssa-alias.c: Likewise.
284 * tree-ssa-ccp.c: Likewise.
285 * tree-ssa-coalesce.c: Likewise.
286 * tree-ssa-copy.c: Likewise.
287 * tree-ssa-copyrename.c: Likewise.
288 * tree-ssa-dce.c: Likewise.
289 * tree-ssa-dom.c: Likewise.
290 * tree-ssa-dse.c: Likewise.
291 * tree-ssa-forwprop.c: Likewise.
292 * tree-ssa-ifcombine.c: Likewise.
293 * tree-ssa-live.c: Likewise.
294 * tree-ssa-loop-ch.c: Likewise.
295 * tree-ssa-loop-im.c: Likewise.
296 * tree-ssa-loop-ivcanon.c: Likewise.
297 * tree-ssa-loop-ivopts.c: Likewise.
298 * tree-ssa-loop-manip.c: Likewise.
299 * tree-ssa-loop-niter.c: Likewise.
300 * tree-ssa-loop-prefetch.c: Likewise.
301 * tree-ssa-loop-unswitch.c: Likewise.
302 * tree-ssa-loop.c: Likewise.
303 * tree-ssa-math-opts.c: Likewise.
304 * tree-ssa-operands.c: Likewise.
305 * tree-ssa-phiopt.c: Likewise.
306 * tree-ssa-phiprop.c: Likewise.
307 * tree-ssa-pre.c: Likewise.
308 * tree-ssa-propagate.c: Likewise.
309 * tree-ssa-reassoc.c: Likewise.
310 * tree-ssa-sccvn.c: Likewise.
311 * tree-ssa-sink.c: Likewise.
312 * tree-ssa-strlen.c: Likewise.
313 * tree-ssa-structalias.c: Likewise.
314 * tree-ssa-tail-merge.c: Likewise.
315 * tree-ssa-ter.c: Likewise.
316 * tree-ssa-threadedge.c: Likewise.
317 * tree-ssa-threadupdate.c: Likewise.
318 * tree-ssa-uncprop.c: Likewise.
319 * tree-ssa-uninit.c: Likewise.
320 * tree-ssa.c: Likewise.
321 * tree-ssanames.c: Likewise.
322 * tree-stdarg.c: Likewise.
323 * tree-streamer-in.c: Likewise.
324 * tree-switch-conversion.c: Likewise.
325 * tree-tailcall.c: Likewise.
326 * tree-vect-data-refs.c: Likewise.
327 * tree-vect-generic.c: Likewise.
328 * tree-vect-loop-manip.c: Likewise.
329 * tree-vect-loop.c: Likewise.
330 * tree-vect-patterns.c: Likewise.
331 * tree-vect-slp.c: Likewise.
332 * tree-vect-stmts.c: Likewise.
333 * tree-vectorizer.c: Likewise.
334 * tree-vrp.c: Likewise.
337 * value-prof.c: Likewise.
338 * var-tracking.c: Likewise.
339 * varpool.c: Likewise.
340 * vtable-verify.c: Likewise.
342 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
344 * config/tilegx/tilegx.c: Include "tree.h".
346 2013-10-23 Jakub Jelinek <jakub@redhat.com>
348 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
349 the info, not after it.
350 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
351 dump_ssaname_info call.
352 (pp_gimple_stmt_1): Adjust caller.
353 (dump_phi_nodes): Likewise. Don't print "# " here.
355 2013-10-22 Jan Hubicka <jh@suse.cz>
357 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
358 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
359 * i386.c (expand_small_movmem_or_setmem): New function.
360 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
361 (alg_usable_p): Add support for value ranges; cleanup.
362 (ix86_expand_set_or_movmem): Add support for misaligned moves.
364 2013-10-22 Sterling Augustine <saugustine@google.com>
366 * doc/invoke.texi: Document -ggnu-pubnames.
367 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
369 * dwarf2out.c: Include gdb/gdb-index.h.
370 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
371 debug_generate_pub_sections.
372 (is_java, output_pubtables, output_pubname): New functions.
373 (include_pubname_in_output): Handle debug_generate_pub_sections at
375 (size_of_pubnames): Use new local space_for_flags based on
376 debug_generate_pub_sections.
377 (output_pubnames): Unify pubnames and pubtypes output logic.
378 Genericize comments. Call output_pubname.
379 (dwarf2out_finish): Move logic to output_pubnames and call it.
381 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
384 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
385 Remove CCCmode handling.
386 <case LTU>: Return 'c' suffix for CCCmode.
387 <case GEU>: Return 'nc' suffix for CCCmode.
388 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
389 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
390 (*sub<mode>3_cc_overflow): Ditto.
391 (*subsi3_zext_cc_overflow): Ditto.
393 2013-10-22 Steve Ellcey <sellcey@mips.com>
395 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
396 (AND (NOT OP1) (NOT OP2)).
398 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
400 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
401 meaning of merge-high and merge-low masks for little endian; avoid
402 use of vector-pack masks for little endian for mismatched modes.
404 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
406 * config/tilepro/tilepro.c: Include "tree.h".
408 2013-10-22 Andreas Schwab <schwab@suse.de>
410 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
413 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
415 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
416 [Ed Smith-Rowland]): New entries.
417 ([Stephen M. Webb]): Update.
419 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
421 * tree-ssa-ter.h: Remove duplicate copy of file contents.
423 2013-10-21 Marek Polacek <polacek@redhat.com>
426 * fold-const.c (fold_range_test): Return 0 if the type is not
429 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
431 * system.h: Move hwint.h include further down.
432 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
433 gcc_checking_asserts.
434 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
436 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
438 Fix volatile issues in optimize_bit_field_compare.
439 * fold-const.c (optimize_bit_field_compare): Bail out if
440 lvolatilep or rvolatilep.
442 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
444 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
445 and get_inner_reference returning different pmode for non-volatile
446 bit-field members dependent on flag_strict_volatile_bitfields.
447 * stor-layout.c (layout_decl): Remove special handling of
448 flag_strict_volatile_bitfields.
449 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
450 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
452 2013-10-21 Paulo Matos <pmatos@broadcom.com>
454 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
457 2013-10-21 Jeff Law <law@redhat.com>
459 * tree-ssa-threadedge.c (thread_through_normal_block): New
460 argument VISITED. Remove VISISTED as a local variable. When we
461 have a threadable jump, verify the destination of the jump has not
463 (thread_across_edge): Allocate VISITED bitmap once at function
464 scope and use it throughout. Make sure to set appropriate bits in
465 VISITED for E (start of jump thread path).
466 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
467 through a joiner if any edge on the path has a recorded jump thread.
469 2013-10-21 Ian Lance Taylor <iant@google.com>
471 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
472 don't imply that attributes can solve all problems.
473 (Directory Options): Fix typo.
475 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
477 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
478 extend and extend_add.
480 2013-10-21 Richard Biener <rguenther@suse.de>
482 PR tree-optimization/58794
483 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
484 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
486 2013-10-21 Richard Biener <rguenther@suse.de>
489 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
490 to (T) X for sign-changing conversions (or no conversion).
492 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
494 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
496 2013-10-20 Jan Hubicka <jh@suse.cz>
498 * config/i386/i386-tune.def: Add comment; organize into categories
500 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
502 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
503 argument. Update function comment.
504 (expand_set_or_movmem_via_rep): New function combining
505 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
506 (expand_movmem_via_rep_mov): Remove.
507 expand_setmem_via_rep_stos): Remove.
508 (expand_movmem_epilogue): Update calls correspondingly.
509 (expand_setmem_epilogue_via_loop): Likewise.
511 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
512 (expand_set_or_movmem_prologue): New function combining
513 expand_movmem_prologue and expand_setmem_prologue.
514 (expand_movmem_prologue): Remove.
515 (expand_setmem_prologue): Remove.
516 (expand_set_or_movmem_constant_prologue): New function combining
517 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
518 (expand_constant_movmem_prologue): Remove.
519 (expand_constant_setmem_prologue): Remove.
520 (promote_duplicated_reg): Allow vector-const0 value.
521 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
522 and ix86_expand_setmem.
523 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
524 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
526 2013-10-21 Diego Novillo <dnovillo@google.com>
528 * asan.c: Include tree.h
529 * bb-reorder.c: Likewise.
530 * cfgcleanup.c: Likewise.
531 * cfgloopmanip.c: Likewise.
532 * data-streamer-in.c: Likewise.
533 * data-streamer-out.c: Likewise.
534 * data-streamer.c: Likewise.
535 * dwarf2cfi.c: Likewise.
536 * graphite-blocking.c: Likewise.
537 * graphite-clast-to-gimple.c: Likewise.
538 * graphite-dependences.c: Likewise.
539 * graphite-interchange.c: Likewise.
540 * graphite-optimize-isl.c: Likewise.
541 * graphite-poly.c: Likewise.
542 * graphite-scop-detection.c: Likewise.
543 * graphite-sese-to-poly.c: Likewise.
544 * graphite.c: Likewise.
545 * ipa-devirt.c: Likewise.
546 * ipa-profile.c: Likewise.
549 * loop-init.c: Likewise.
550 * loop-unroll.c: Likewise.
551 * lower-subreg.c: Likewise.
552 * lto/lto-object.c: Likewise.
554 * reginfo.c: Likewise.
555 * tree-loop-distribution.c: Likewise.
556 * tree-parloops.c: Likewise.
557 * tree-ssa-strlen.c: Likewise.
558 * tree-streamer.c: Likewise.
559 * value-prof.c: Likewise.
560 * target-globals.c: Likewise.
561 * expr.h: Include tree-core.h instead of tree.h.
562 * gimple.h: Likewise.
563 * ipa-prop.h: Likewise.
564 * ipa-utils.h: Likewise.
565 * lto-streamer.h: Likewise.
566 * streamer-hooks.h: Likewise.
567 * ipa-reference.h: Include cgraph.h instead of tree.h.
568 * cgraph.h: Include basic-block.h instead of tree.h.
569 * tree-streamer.h: Do not include tree.h.
570 * genattrtab.c (write_header): Generate inclusion of
572 * genautomata.c (main): Likewise.
573 * genemit.c: Likewise.
574 * genopinit.c: Likewise.
575 * genoutput.c (output_prologue): Likewise.
576 * genpeep.c: Likewise.
578 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
580 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
582 (vec_unpacku_hi_v8hi): Likewise.
583 (vec_unpacku_lo_v16qi): Likewise.
584 (vec_unpacku_lo_v8hi): Likewise.
586 2013-10-20 Jan Hubicka <jh@suse.cz>
588 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
589 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
590 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
592 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
594 * config/mips/mips.h (ISA_HAS_WSBH): Define.
595 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
597 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
599 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
602 * system.h: Undef m_slot.
604 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
606 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
607 all elements for both endian flavors.
609 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
612 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
613 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
614 and SI_REG for 64bit SYSV ABI targets.
616 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
618 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
619 to multi_reg_return. Clarify that we are skipping USEs of multiple
620 return registers. Use bool type where appropriate.
622 2013-10-18 Jan Hubicka <jh@suse.cz>
624 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
626 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
627 (X86_TUNE_PUSH_MEMORY): Likewise.
628 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
629 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
630 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
632 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
633 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
635 (ix86_option_override_internal): Update to use tune features instead
638 2013-10-18 Cong Hou <congh@google.com>
640 PR tree-optimization/58508
641 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
642 statement that contains data refs with zero-step.
644 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
646 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
647 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
648 * gimple-low.c (gimple_check_call_arg,
649 gimple_check_call_matching_types): Move to cgraph.c.
650 * gimple-low.h: Remove prototype.
651 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
652 Relocate from gimple-low.c.
653 * cgraph.h: Add prototype. Don't include basic-block.h.
654 * gimplify.c: Add gimple-low to include list.
655 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
656 * tree-eh.c: Add gimple-low to include list.
657 * tree-nested.c: Likewise.
658 * cfgexpand.c: Add tree-ssa-address.h to include list.
660 * gimple-fold.c: Likewise.
661 * gimple-ssa-strength-reduction.c: Likewise.
662 * trans-mem.c: Likewise.
663 * tree-mudflap.c: Likewise.
664 * tree-ssa-loop-ivopts.c: Likewise.
665 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
666 (degenerate_phi_result): Move to tree-phinodes.c.
667 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
668 * tree-ssa-threadedge.c: Likewise.
669 * tree-vrp.c: Likewise.
670 * tree-phinodes.c (degenerate_phi_result): Relocate here.
671 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
672 * tree-phinodes.h (degenerate_phi_result): Add prototype.
673 * tree-ssa-copy.c: Include tree-ssa-dom.h.
674 * tree-ssa-forwprop.c: Likewise.
675 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
676 pass_data_cleanup_cfg_post_optimizing,
677 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
678 * tree-optimize.c: Delete File.
679 * graphite.c: Include tree-cfgcleanup.h.
680 * passes.c: Likewise.
681 * tree-cfg.c: Likewise.
682 * tree-profile.c: Likewise.
683 * tree-ssa-dse.c: Likewise.
684 * tree-ssa-loop-ivcanon.c: Likewise.
685 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
686 * tree-outof-ssa.c: Include sbitmap.h.
687 * tree-ssa-live.c: Likewise.
688 * tree-ssa-propagate.c: Likewise.
689 * tree-ssa-structalias.c: Likewise.
690 * tree-stdarg.c: Likewise.
691 * Makefile.in (OBJS): Delete tree-optimize.o.
692 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
693 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
694 * varasm.c: Include basic-block.h.
695 * cfgloop.h: Include function.h instead of basic-block.h
696 (bb_loop_depth): Move to cfgloop.c.
697 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
699 2013-10-18 Teresa Johnson <tejohnson@google.com>
701 * predict.c (probably_never_executed): Compare frequency-based
702 count to number of training runs.
703 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
705 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
707 * config/arm/arm.c (cortexa9_extra_costs): New table.
708 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
710 2013-10-18 Jeff Law <law@redhat.com>
712 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
714 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
715 (register_jump_thread): Add "registered_jump_thread" debug
717 * dbgcnt.def (registered_jump_thread): New debug counter.
719 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
721 * config/rs6000/rs6000.c: Include cgraph.h.
723 2013-10-18 Teresa Johnson <tejohnson@google.com>
725 * tree-ssa-tail-merge.c (replace_block_by): Update edge
726 weights during merging.
728 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
730 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
731 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
732 * tree-ssa-operands.h: Remove prototype.
733 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
734 * gimple.c (virtual_operand_p): Relocate from gimple.c.
735 * gimple.h: Add prototype.
736 * gimple-ssa.h: Include tree-ssa-operands.h.
737 * tree-dump.c: Add tree-cfg.h to include list.
738 * tree-ssa-alias.c: Add ipa-reference.h to include list.
739 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
740 * config/i386/i386.c: Don't include tree-flow.h.
741 * config/rs6000/rs6000.c: Likewise.
743 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
745 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
747 2013-10-18 Richard Biener <rguenther@suse.de>
749 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
750 or TYPE_UNSIGNED of integral types.
751 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
752 NULL_TREE for zero-precision integral types.
754 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
756 * config/aarch64/arm_neon.h
757 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
759 2013-10-17 Sriraman Tallam <tmsriram@google.com>
762 * opth-gen.awk: Define target_flags_explicit.
764 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
766 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
767 fields to the reg_addr array that describes the valid addressing
768 mode for any register, general purpose registers, floating point
769 registers, and Altivec registers.
770 (FIRST_RELOAD_REG_CLASS): Likewise.
771 (LAST_RELOAD_REG_CLASS): Likewise.
772 (struct reload_reg_map_type): Likewise.
773 (reload_reg_map_type): Likewise.
774 (RELOAD_REG_VALID): Likewise.
775 (RELOAD_REG_MULTIPLE): Likewise.
776 (RELOAD_REG_INDEXED): Likewise.
777 (RELOAD_REG_OFFSET): Likewise.
778 (RELOAD_REG_PRE_INCDEC): Likewise.
779 (RELOAD_REG_PRE_MODIFY): Likewise.
780 (reg_addr): Likewise.
781 (mode_supports_pre_incdec_p): New helper functions to say whether
782 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
783 (mode_supports_pre_modify_p): Likewise.
784 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
785 print the valid address mode bits for each mode.
786 (rs6000_debug_print_mode): Likewise.
787 (rs6000_debug_reg_global): Likewise.
788 (rs6000_setup_reg_addr_masks): New function to set up the address
789 mask bits for each type.
790 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
791 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
792 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
793 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
794 PRE_MODIFY are supported.
795 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
796 registers, instead of {src,dest}_av_p.
797 (rs6000_print_options_internal): Tweak the debug output slightly.
799 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
801 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
804 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
806 * tree-flow.h (struct omp_region): Move to omp-low.c.
807 Remove omp_ prototypes and variables.
808 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
809 (copy_var_decl): Relocate prototype from tree-flow.h.
810 * gimple.c (copy_var_decl): Relocate from omp-low.c.
811 * tree.h: Move prototype to omp-low.h.
812 * omp-low.h: New File. Relocate prototypes here.
813 * omp-low.c (struct omp_region): Make local here.
814 (root_omp_region): Make static.
815 (copy_var_decl) Move to gimple.c.
816 (new_omp_region): Make static.
817 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
818 * tree-cfg.c: Include omp-low.h.
819 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
820 * gimplify.c: Include omp-low.h.
821 * tree-parloops.c: Likewise.
823 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
825 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
826 operands are in memory, prefer to force non-matched operand 1 to
829 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
832 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
833 restrict TImode addresses to single indirect registers if both
834 -mquad-memory and -mvsx-timode are used.
835 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
836 we should emit load/store quad. Remove using %y for quad memory
839 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
840 constraints to allow load/store quad on machines where TImode is
841 not allowed in VSX registers. Use 'n' instead of 'F' constraint
842 for TImode to load integer constants.
844 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
846 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
848 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
850 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
851 handling of STACK_REG.
853 2013-10-17 Richard Biener <rguenther@suse.de>
855 PR tree-optimization/58143
856 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
858 (rewrite_to_defined_overflow): Likewise.
859 (move_computations_dom_walker::before_dom): Rewrite stmts
860 with undefined signed overflow that are not always executed
861 into unsigned arithmetic.
863 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
866 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
867 explicit isa flag to be an options variable, instead of using
868 global_options_set. Remove define from rs6000.h.
869 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
871 * config/rs6000/rs6000.c (rs6000_option_override_internal):
872 Initialize rs6000_isa_flags_explicit.
873 (rs6000_function_specific_save): Add gcc_options* parameter, so
874 that the powerpc builds after the 2013-10-15 changes.
875 (rs6000_function_specific_restore): Likewise.
877 2013-10-16 DJ Delorie <dj@redhat.com>
879 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
880 op is a REG before checking REGNO.
881 (rl78_alloc_physical_registers): Verify pattern is a SET before
884 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
886 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
888 (vec_unpacks_lo_v4sf): Likewise.
889 (vec_unpacks_float_hi_v4si): Likewise.
890 (vec_unpacks_float_lo_v4si): Likewise.
891 (vec_unpacku_float_hi_v4si): Likewise.
892 (vec_unpacku_float_lo_v4si): Likewise.
894 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
896 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
897 (vsx_concat_v2sf): Likewise.
899 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
901 * config/aarch64/aarch64.md
902 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
904 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
906 PR tree-optimization/58697
907 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
908 estimated_loop_iterations_int.
909 (max_stmt_executions_int): Call get_max_loop_iterations_int.
910 (get_max_loop_iterations_int): New. HWINT version of
911 get_max_loop_iterations.
912 * cfgloop.h: Add prototypes.
913 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
914 * loop-unroll.c (decide_peel_once_rolling): Call
915 get_estimated_loop_iterations_int.
916 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
917 * tree-ssa-loop-niter.h: Tweak prototypes.
919 2013-10-16 David Malcolm <dmalcolm@redhat.com>
921 * gengtype-parse.c (struct_field_seq): Ignore access-control
922 keywords ("public:" etc).
924 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
926 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
927 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
929 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
931 * config/arm/arm.opt (mlra): New option.
932 * config/arm/arm.c (arm_lra_p): New function.
933 (TARGET_LRA_P): Define.
935 2013-10-16 Paulo Matos <pmatos@broadcom.com>
937 * tree-core.h (tree_code_name): Remove.
938 * tree.h (get_tree_code_name): New prototype.
939 * tree.c (tree_code_name): Make static.
940 (get_tree_code_name): New function.
941 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
942 tree_class_check_failed, tree_range_check_failed,
943 tree_not_class_check_failed, omp_clause_check_failed,
944 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
945 wrapper get_tree_code_name instead of calling tree_code_name directly.
946 * tree-vrp.c (dump_asserts_for): Likewise.
947 * tree-dump.c (dequeue_and_dump): Likewise.
948 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
949 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
950 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
951 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
952 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
953 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
954 dump_gimple_omp_for): Likewise.
955 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
956 * tree-ssa-pre.c (print_pre_expr): Likewise.
957 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
958 * print-tree.c (print_node_brief, print_node): Likewise.
959 * gimple.c (gimple_check_failed): Likewise.
960 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
961 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
962 * config/mep/mep.c (mep_validate_vliw): Likewise.
963 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
964 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
966 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
968 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
971 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
973 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
974 (MULTILIB_MATCHES): Add multilib for -march=v8.
976 2013-10-15 Sriraman Tallam <tmsriram@google.com>
979 * optc-save-gen.awk: Add extra parameter to the save and restore
981 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
982 * tree.c (build_optimization_node): New parameter. Add extra parameter
983 to call to cl_optimization_save.
984 (build_target_option_node): New parameter. Add extra parameter
985 to call to cl_target_option_save.
986 * tree.h (build_optimization_node): New parameter.
987 (build_target_option_node): New parameter.
988 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
989 build_optimization_node and build_target_option_node.
990 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
991 (handle_pragma_push_options): Ditto.
992 * toplev.c (process_options): Ditto.
993 * opts.c (init_options_struct): Check for opts_set non-null.
994 * target.def (target_option.save): New parameter.
995 (target_option.restore): New parameter.
997 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
998 (ix86_pragma_target_parse): Ditto.
999 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
1001 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
1002 to build_optimization_node and build_target_option_node.
1003 (rs6000_valid_attribute_p): Ditto.
1004 (rs6000_pragma_target_parse): Ditto.
1005 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
1007 * config/i386/i386.h:
1008 TARGET_64BIT_P: New Macro
1009 TARGET_MMX_P: New Macro.
1010 TARGET_3DNOW_P: New Macro.
1011 TARGET_3DNOW_A_P: New Macro.
1012 TARGET_SSE_P: New Macro.
1013 TARGET_SSE2_P: New Macro.
1014 TARGET_SSE3_P: New Macro.
1015 TARGET_SSSE3_P: New Macro.
1016 TARGET_SSE4_1_P: New Macro.
1017 TARGET_SSE4_2_P: New Macro.
1018 TARGET_AVX_P: New Macro.
1019 TARGET_AVX2_P: New Macro.
1020 TARGET_AVX512F_P: New Macro.
1021 TARGET_AVX512PF_P: New Macro.
1022 TARGET_AVX512ER_P: New Macro.
1023 TARGET_AVX512CD_P: New Macro.
1024 TARGET_FMA_P: New Macro.
1025 TARGET_SSE4A_P: New Macro.
1026 TARGET_FMA4_P: New Macro.
1027 TARGET_XOP_P: New Macro.
1028 TARGET_LWP_P: New Macro.
1029 TARGET_ABM_P: New Macro.
1030 TARGET_BMI_P: New Macro.
1031 TARGET_BMI2_P: New Macro.
1032 TARGET_LZCNT_P: New Macro.
1033 TARGET_TBM_P: New Macro.
1034 TARGET_POPCNT_P: New Macro.
1035 TARGET_SAHF_P: New Macro.
1036 TARGET_MOVBE_P: New Macro.
1037 TARGET_CRC32_P: New Macro.
1038 TARGET_AES_P: New Macro.
1039 TARGET_PCLMUL_P: New Macro.
1040 TARGET_CMPXCHG16B_P: New Macro.
1041 TARGET_FSGSBASE_P: New Macro.
1042 TARGET_RDRND_P: New Macro.
1043 TARGET_F16C_P: New Macro.
1044 TARGET_RTM_P: New Macro.
1045 TARGET_HLE_P: New Macro.
1046 TARGET_RDSEED_P: New Macro.
1047 TARGET_PRFCHW_P: New Macro.
1048 TARGET_ADX_P: New Macro.
1049 TARGET_FXSR_P: New Macro.
1050 TARGET_XSAVE_P: New Macro.
1051 TARGET_XSAVEOPT_P: New Macro.
1052 TARGET_LP64_P: New Macro.
1053 TARGET_X32_P: New Macro.
1054 TARGET_FPMATH_DEFAULT_P: New Macro.
1055 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
1056 * config/i386/i386.c (ix86_option_override_internal): New parameters.
1058 Change ix86_tune_string to access opts->x_ix86_tune_string.
1059 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
1060 Change ix86_arch_string to access opts->x_ix86_arch_string.
1061 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
1062 Change ix86_pmode to access opts->x_ix86_pmode.
1063 Change ix86_abi to access opts->x_ix86_abi.
1064 Change ix86_cmodel to access opts->x_ix86_cmodel.
1065 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
1066 Change ix86_isa_flags_explicit to access
1067 opts->x_ix86_isa_flags_explicit.
1068 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
1069 Change ix86_regparm to access opts->x_ix86_regparm.
1070 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
1071 Change ix86_preferred_stack_boundary_arg to access
1072 opts->x_ix86_preferred_stack_boundary_arg.
1073 Change ix86_force_align_arg_pointer to access
1074 opts->x_ix86_force_align_arg_pointer.
1075 Change ix86_incoming_stack_boundar_arg to access
1076 opts->x_ix86_incoming_stack_boundar_arg.
1077 Change ix86_fpmath to access opts->x_ix86_fpmath.
1078 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
1079 Change ix86_recip_name to access opts->x_ix86_recip_name.
1080 Change ix86_stack_protector_guard to access
1081 opts->x_ix86_stack_protector_guard.
1082 Change ix86_tune_memcpy_strategy to access
1083 opts->x_ix86_tune_memcpy_strategy.
1084 Change ix86_tune_memset_strategy to access
1085 opts->x_ix86_tune_memset_strategy.
1086 Change global_options to access opts.
1087 Change global_options_set to access opts_set.
1088 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
1089 Change TARGET_MMX to TARGET_MMX_P (opts->...).
1090 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
1091 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
1092 Change TARGET_SSE to TARGET_SSE_P (opts->...).
1093 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
1094 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
1095 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
1096 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
1097 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
1098 Change TARGET_AVX to TARGET_AVX_P (opts->...).
1099 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
1100 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
1101 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
1102 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
1103 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
1104 Change TARGET_FMA to TARGET_FMA_P (opts->...).
1105 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
1106 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
1107 Change TARGET_XOP to TARGET_XOP_P (opts->...).
1108 Change TARGET_LWP to TARGET_LWP_P (opts->...).
1109 Change TARGET_ABM to TARGET_ABM_P (opts->...).
1110 Change TARGET_BMI to TARGET_BMI_P (opts->...).
1111 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
1112 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
1113 Change TARGET_TBM to TARGET_TBM_P (opts->...).
1114 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
1115 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
1116 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
1117 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
1118 Change TARGET_AES to TARGET_AES_P (opts->...).
1119 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
1120 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
1121 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
1122 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
1123 Change TARGET_F16C to TARGET_F16C_P (opts->...).
1124 Change TARGET_RTM to TARGET_RTM_P (opts->...).
1125 Change TARGET_HLE to TARGET_HLE_P (opts->...).
1126 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
1127 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
1128 Change TARGET_ADX to TARGET_ADX_P (opts->...).
1129 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
1130 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
1131 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
1132 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
1133 Change TARGET_X32 to TARGET_X32_P (opts->...).
1134 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
1135 Change TARGET_FLOAT_RETURNS_IN_80387 to
1136 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
1137 (ix86_function_specific_save): New parameter. Use opts-> fields
1138 to replace global fields.
1139 (ix86_function_specific_restore): Ditto.
1140 (ix86_valid_target_attribute_inner_p): New parameters.
1142 Fix call to ix86_handle_option and set_option.
1143 (ix86_valid_target_attribute_tree): New parameters.
1144 Change global_options to access opts.
1145 Change global_options_set to access opts_set.
1146 Fix call to ix86_valid_target_attribute_inner_p.
1147 Change ix86_tune_string to access opts->x_ix86_tune_string.
1148 Change ix86_arch_string to access opts->x_ix86_arch_string.
1149 Change ix86_fpmath to access opts->x_ix86_fpmath
1150 Fix call to ix86_option_override_internal.
1151 Fix call to ix86_add_new_builtins.
1152 Fix calls to build_optimization_node and build_target_option_node.
1153 (ix86_valid_target_attribute_p): Remove access to global_options.
1154 Use new gcc_options structure func_options.
1155 Fix call to ix86_valid_target_attribute_tree.
1156 Fix call to build_optimization_node.
1157 (get_builtin_code_for_version): Fix call to
1158 ix86_valid_target_attribute_tree.
1160 2013-10-15 David Malcolm <dmalcolm@redhat.com>
1162 * Makefile.in (PICFLAG): New.
1163 (enable_host_shared): New.
1164 (INTERNAL_CFLAGS): Use PICFLAG.
1165 (LIBIBERTY): Use pic build of libiberty.a if configured with
1166 --enable-host-shared.
1167 * configure.ac: Add --enable-host-shared, setting up new
1169 * configure: Regenerate.
1170 * doc/install.texi (--enable-shared): Add note contrasting it with ...
1171 (--enable-host-shared): New option.
1173 2013-10-15 Richard Biener <rguenther@suse.de>
1175 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1176 for built-in functions.
1178 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1180 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1181 (optimize_range_tests_1): New function,
1182 extracted from optimize_range_tests.
1183 (optimize_range_tests_xor): Similarly.
1184 (optimize_range_tests_diff): New function.
1185 (optimize_range_tests): Use optimize_range_tests_1.
1187 2013-10-15 Cong Hou <congh@google.com>
1189 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1190 requirement of the reduction pattern so that one operand of the
1191 reduction operation can come from outside of the loop.
1193 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1195 * config/arm/neon-schedgen.ml: Remove.
1196 * config/arm/cortex-a9-neon.md: Remove comment regarding
1199 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1201 * config/arm/types: Remove old neon types.
1203 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1205 * config/arm/cortex-a7.md
1206 (cortex_a7_neon_type): New.
1207 (cortex_a7_neon_mul): Update for new types.
1208 (cortex_a7_neon_mla): Likewise.
1209 (cortex_a7_neon): Likewise.
1211 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1213 * config/arm/cortex-a15-neon.md
1214 (cortex_a15_neon_type): New,
1216 (cortex_a15_neon_int_1): Remove.
1217 (cortex_a15_neon_int_2): Likewise.
1218 (cortex_a15_neon_int_3): Likewise.
1219 (cortex_a15_neon_int_4): Likewise.
1220 (cortex_a15_neon_int_5): Likewise.
1221 (cortex_a15_neon_vqneg_vqabs): Likewise.
1222 (cortex_a15_neon_vmov): Likewise.
1223 (cortex_a15_neon_vaba): Likewise.
1224 (cortex_a15_neon_vaba_qqq): Likewise.
1225 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1226 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1227 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1229 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1230 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1231 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1232 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1233 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1234 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1235 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1236 (cortex_a15_neon_shift_1): Likewise.
1237 (cortex_a15_neon_shift_2): Likewise.
1238 (cortex_a15_neon_shift_3): Likewise.
1239 (cortex_a15_neon_vshl_ddd): Likewise.
1240 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1241 (cortex_a15_neon_vsra_vrsra): Likewise.
1242 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1243 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1244 (cortex_a15_neon_bp_3cycle): Likewise.
1245 (cortex_a15_neon_ldm_2): Likewise.
1246 (cortex_a15_neon_stm_2): Likewise.
1247 (cortex_a15_neon_mcr): Likewise.
1248 (cortex_a15_neon_mrc): Likewise.
1249 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1250 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1251 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1252 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1253 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1254 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1255 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1256 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1257 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1258 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1259 (cortex_a15_neon_bp_simple): Likewise.
1260 (cortex_a15_neon_bp_2cycle): Likewise.
1261 (cortex_a15_neon_bp_3cycle): Likewise.
1262 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1263 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1264 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1265 (cortex_a15_neon_vld2_4_regs): Likewise.
1266 (cortex_a15_neon_vld3_vld4): Likewise.
1267 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1268 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1269 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1270 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1271 (cortex_a15_neon_vst3_vst4): Rename to...
1272 (cortex_a15_neon_vst4): This, update for new attributes.
1273 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1274 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1275 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1276 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1277 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1278 (cortex_a15_neon_ldm_2): Likewise.
1279 (cortex_a15_neon_stm_2): Likewise.
1280 (cortex_a15_neon_mcr): Likewise.
1281 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1282 (cortex_a15_neon_mrc): Likewise.
1283 (cortex_a15_neon_mrrc): Likewise.
1285 (cortex_a15_neon_abd): New.
1286 (cortex_a15_neon_abd_q): Likewise.
1287 (cortex_a15_neon_aba): Likewise.
1288 (cortex_a15_neon_aba_q): Likewise.
1289 (cortex_a15_neon_acc): Likewise.
1290 (cortex_a15_neon_acc_q): Likewise.
1291 (cortex_a15_neon_arith_basic): Likewise.
1292 (cortex_a15_neon_arith_complex): Likewise.
1293 (cortex_a15_neon_multiply): Likewise.
1294 (cortex_a15_neon_multiply_q): Likewise.
1295 (cortex_a15_neon_mla): Likewise.
1296 (cortex_a15_neon_mla_q): Likewise.
1297 (cortex_a15_neon_sat_mla_long): Likewise.
1298 (cortex_a15_neon_shift_acc): Likewise.
1299 (cortex_a15_neon_shift_imm_basic): Likewise.
1300 (cortex_a15_neon_shift_imm_complex): Likewise.
1301 (cortex_a15_neon_shift_reg_basic): Likewise.
1302 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1303 (cortex_a15_neon_shift_reg_complex): Likewise.
1304 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1305 (cortex_a15_neon_fp_negabs): Likewise
1306 (cortex_a15_neon_fp_arith): Likewise
1307 (cortex_a15_neon_fp_arith_q): Likewise
1308 (cortex_a15_neon_fp_cvt_int): Likewise
1309 (cortex_a15_neon_fp_cvt_int_q): Likewise
1310 (cortex_a15_neon_fp_cvt_16): Likewise
1311 (cortex_a15_neon_fp_mul): Likewise
1312 (cortex_a15_neon_fp_mul_q): Likewise
1313 (cortex_a15_neon_fp_mla): Likewise
1314 (cortex_a15_neon_fp_mla_q): Likewise
1315 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1316 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1317 (cortex_a15_neon_bitops): Likewise.
1318 (cortex_a15_neon_bitops_q): Likewise.
1319 (cortex_a15_neon_from_gp): Likewise.
1320 (cortex_a15_neon_from_gp_q): Likewise.
1321 (cortex_a15_neon_tbl3_tbl4): Likewise.
1322 (cortex_a15_neon_zip_q): Likewise.
1323 (cortex_a15_neon_to_gp): Likewise.
1324 (cortex_a15_neon_load_a): Likewise.
1325 (cortex_a15_neon_load_b): Likewise.
1326 (cortex_a15_neon_load_c): Likewise.
1327 (cortex_a15_neon_load_d): Likewise.
1328 (cortex_a15_neon_load_e): Likewise.
1329 (cortex_a15_neon_load_f): Likewise.
1330 (cortex_a15_neon_store_a): Likewise.
1331 (cortex_a15_neon_store_b): Likewise.
1332 (cortex_a15_neon_store_c): Likewise.
1333 (cortex_a15_neon_store_d): Likewise.
1334 (cortex_a15_neon_store_e): Likewise.
1335 (cortex_a15_neon_store_f): Likewise.
1336 (cortex_a15_neon_store_g): Likewise.
1337 (cortex_a15_neon_store_h): Likewise.
1338 (cortex_a15_vfp_to_from_gp): Likewise.
1340 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1342 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1344 (cortex_a9_neon_vshl_ddd): Remove.
1345 (cortex_a9_neon_vst3_vst4): Likewise.
1346 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1348 (cortex_a9_neon_bit_ops_q): New.
1350 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1351 (cortex_a9_neon_int_2): Likewise.
1352 (cortex_a9_neon_int_3): Likewise.
1353 (cortex_a9_neon_int_4): Likewise.
1354 (cortex_a9_neon_int_5): Likewise.
1355 (cortex_a9_neon_vqneg_vqabs): Likewise.
1356 (cortex_a9_neon_vmov): Likewise.
1357 (cortex_a9_neon_vaba): Likewise.
1358 (cortex_a9_neon_vaba_qqq): Likewise.
1359 (cortex_a9_neon_shift_1): Likewise.
1360 (cortex_a9_neon_shift_2): Likewise.
1361 (cortex_a9_neon_shift_3): Likewise.
1362 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1363 (cortex_a9_neon_vsra_vrsra): Likewise.
1364 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1365 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1366 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1368 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1369 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1370 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1372 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1373 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1374 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1375 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1376 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1377 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1378 (cortex_a9_neon_fp_vsum): Likewise.
1379 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1380 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1381 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1382 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1383 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1384 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1385 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1386 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1387 (cortex_a9_neon_bp_simple): Likewise.
1388 (cortex_a9_neon_bp_2cycle): Likewise.
1389 (cortex_a9_neon_bp_3cycle): Likewise.
1390 (cortex_a9_neon_ldr): Likewise.
1391 (cortex_a9_neon_str): Likewise.
1392 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1393 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1394 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1395 (cortex_a9_neon_vld2_4_regs): Likewise.
1396 (cortex_a9_neon_vld3_vld4): Likewise.
1397 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1398 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1399 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1400 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1401 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1402 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1403 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1404 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1405 (cortex_a9_neon_mcr): Likewise.
1406 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1407 (cortex_a9_neon_mrc): Likewise.
1408 (cortex_a9_neon_mrrc): Likewise.
1410 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1412 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1414 (cortex_a8_neon_vshl_ddd): Remove.
1415 (cortex_a8_neon_vst3_vst4): Likewise.
1416 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1418 (cortex_a8_neon_bit_ops_q): New.
1420 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1421 (cortex_a8_neon_int_2): Likewise..
1422 (cortex_a8_neon_int_3): Likewise.
1423 (cortex_a8_neon_int_5): Likewise.
1424 (cortex_a8_neon_vqneg_vqabs): Likewise.
1425 (cortex_a8_neon_int_4): Likewise.
1426 (cortex_a8_neon_vaba): Likewise.
1427 (cortex_a8_neon_vaba_qqq): Likewise.
1428 (cortex_a8_neon_shift_1): Likewise.
1429 (cortex_a8_neon_shift_2): Likewise.
1430 (cortex_a8_neon_shift_3): Likewise.
1431 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1432 (cortex_a8_neon_vsra_vrsra): Likewise.
1433 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1434 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1435 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1437 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1438 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1439 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1441 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1442 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1443 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1444 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1445 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1446 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1447 (cortex_a8_neon_fp_vsum): Likewise.
1448 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1449 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1450 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1451 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1452 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1453 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1454 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1455 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1456 (cortex_a8_neon_bp_simple): Likewise.
1457 (cortex_a8_neon_bp_2cycle): Likewise.
1458 (cortex_a8_neon_bp_3cycle): Likewise.
1459 (cortex_a8_neon_ldr): Likewise.
1460 (cortex_a8_neon_str): Likewise.
1461 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1462 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1463 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1464 (cortex_a8_neon_vld2_4_regs): Likewise.
1465 (cortex_a8_neon_vld3_vld4): Likewise.
1466 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1467 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1468 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1469 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1470 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1471 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1472 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1473 (cortex_a8_neon_mcr): Likewise.
1474 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1475 (cortex_a8_neon_mrc): Likewise.
1476 (cortex_a8_neon_mrrc): Likewise.
1478 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1480 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1482 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1483 (aarch64_simd_dup<mode>): Add "type" attribute.
1484 (aarch64_dup_lane<mode>): Likewise.
1485 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1486 (*aarch64_simd_mov<mode>): Likewise.
1487 (aarch64_simd_mov_from_<mode>low): Likewise.
1488 (aarch64_simd_mov_from_<mode>high): Likewise.
1489 (orn<mode>3): Likewise.
1490 (bic<mode>3): Likewise.
1491 (add<mode>3): Likewise.
1492 (sub<mode>3): Likewise.
1493 (mul<mode>3): Likewise.
1494 (*aarch64_mul3_elt<mode>): Likewise.
1495 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1496 (*aarch64_mul3_elt_to_128df): Likewise.
1497 (*aarch64_mul3_elt_to_64v2df): Likewise.
1498 (neg<mode>2): Likewise.
1499 (abs<mode>2): Likewise.
1500 (abd<mode>_3): Likewise.
1501 (aba<mode>_3): Likewise.
1502 (fabd<mode>_3): Likewise.
1503 (*fabd_scalar<mode>3): Likewise.
1504 (and<mode>3): Likewise.
1505 (ior<mode>3): Likewise.
1506 (xor<mode>3): Likewise.
1507 (one_cmpl<mode>2): Likewise.
1508 (aarch64_simd_vec_set<mode>): Likewise.
1509 (aarch64_simd_lshr<mode>): Likewise.
1510 (aarch64_simd_ashr<mode>): Likewise.
1511 (aarch64_simd_imm_shl<mode>): Likewise.
1512 (aarch64_simd_reg_sshl<mode): Likewise.
1513 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1514 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1515 (aarch64_simd_vec_setv2di): Likewise.
1516 (aarch64_simd_vec_set<mode>): Likewise.
1517 (aarch64_mla<mode>): Likewise.
1518 (*aarch64_mla_elt<mode>): Likewise.
1519 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1520 (aarch64_mls<mode>): Likewise.
1521 (*aarch64_mls_elt<mode>): Likewise.
1522 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1523 (<su><maxmin><mode>3): Likewise.
1524 (move_lo_quad_<mode>): Likewise.
1525 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1526 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1527 (vec_pack_trunc_<mode>): Likewise.
1528 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1529 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1530 (*aarch64_<su>mlal_lo<mode>): Likewise.
1531 (*aarch64_<su>mlal_hi<mode>): Likewise.
1532 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1533 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1534 (*aarch64_<su>mlal<mode>): Likewise.
1535 (*aarch64_<su>mlsl<mode>): Likewise.
1536 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1537 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1538 (add<mode>3): Likewise.
1539 (sub<mode>3): Likewise.
1540 (mul<mode>3): Likewise.
1541 (div<mode>3): Likewise.
1542 (neg<mode>2): Likewise.
1543 (abs<mode>2): Likewise.
1544 (fma<mode>4): Likewise.
1545 (*aarch64_fma4_elt<mode>): Likewise.
1546 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1547 (*aarch64_fma4_elt_to_128df): Likewise.
1548 (*aarch64_fma4_elt_to_64v2df): Likewise.
1549 (fnma<mode>4): Likewise.
1550 (*aarch64_fnma4_elt<mode>): Likewise.
1551 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1552 (*aarch64_fnma4_elt_to_128df): Likewise.
1553 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1554 (<frint_pattern><mode>2): Likewise.
1555 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1556 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1557 (vec_unpacks_lo_v4sf): Likewise.
1558 (aarch64_float_extend_lo_v2df): Likewise.
1559 (vec_unpacks_hi_v4sf): Likewise.
1560 (aarch64_float_truncate_lo_v2sf): Likewise.
1561 (aarch64_float_truncate_hi_v4sf): Likewise.
1562 (aarch64_vmls<mode>): Likewise.
1563 (<su><maxmin><mode>3): Likewise.
1564 (<maxmin_uns><mode>3): Likewise.
1565 (reduc_<sur>plus_<mode>): Likewise.
1566 (reduc_<sur>plus_v2di): Likewise.
1567 (reduc_<sur>plus_v2si): Likewise.
1568 (reduc_<sur>plus_<mode>): Likewise.
1569 (aarch64_addpv4sf): Likewise.
1570 (clz<mode>2): Likewise.
1571 (reduc_<maxmin_uns>_<mode>): Likewise.
1572 (reduc_<maxmin_uns>_v2di): Likewise.
1573 (reduc_<maxmin_uns>_v2si): Likewise.
1574 (reduc_<maxmin_uns>_<mode>): Likewise.
1575 (reduc_<maxmin_uns>_v4sf): Likewise.
1576 (aarch64_simd_bsl<mode>_internal): Likewise.
1577 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1578 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1579 (aarch64_get_lane<mode>): Likewise.
1580 (*aarch64_combinez<mode>): Likewise.
1581 (aarch64_combine<mode>): Likewise.
1582 (aarch64_simd_combine<mode>): Likewise.
1583 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1584 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1585 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1586 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1587 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1588 (aarch64_<sur>h<addsub><mode>): Likewise.
1589 (aarch64_<sur><addsub>hn<mode>): Likewise.
1590 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1591 (aarch64_pmul<mode>): Likewise.
1592 (aarch64_<su_optab><optab><mode>): Likewise.
1593 (aarch64_<sur>qadd<mode>): Likewise.
1594 (aarch64_sqmovun<mode>): Likewise.
1595 (aarch64_<sur>qmovn<mode>): Likewise.
1596 (aarch64_s<optab><mode>): Likewise.
1597 (aarch64_sq<r>dmulh<mode>): Likewise.
1598 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1599 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1600 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1601 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1602 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1603 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1604 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1605 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1606 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1607 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1608 (aarch64_sqdmull<mode>): Likewise.
1609 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1610 (aarch64_sqdmull_n<mode>): Likewise.
1611 (aarch64_sqdmull2<mode>_internal): Likewise.
1612 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1613 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1614 (aarch64_<sur>shl<mode>): Likewise.
1615 (aarch64_<sur>q<r>shl<mode>
1616 (aarch64_<sur>shll_n<mode>): Likewise.
1617 (aarch64_<sur>shll2_n<mode>): Likewise.
1618 (aarch64_<sur>shr_n<mode>): Likewise.
1619 (aarch64_<sur>sra_n<mode>): Likewise.
1620 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1621 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1622 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1623 (aarch64_cm<optab><mode>): Likewise.
1624 (aarch64_cm<optab>di): Likewise.
1625 (aarch64_cm<optab><mode>): Likewise.
1626 (aarch64_cm<optab>di): Likewise.
1627 (aarch64_cmtst<mode>): Likewise.
1628 (aarch64_cmtstdi): Likewise.
1629 (aarch64_cm<optab><mode>): Likewise.
1630 (*aarch64_fac<optab><mode>): Likewise.
1631 (aarch64_addp<mode>): Likewise.
1632 (aarch64_addpdi): Likewise.
1633 (sqrt<mode>2): Likewise.
1634 (vec_load_lanesoi<mode>): Likewise.
1635 (vec_store_lanesoi<mode>): Likewise.
1636 (vec_load_lanesci<mode>): Likewise.
1637 (vec_store_lanesci<mode>): Likewise.
1638 (vec_load_lanesxi<mode>): Likewise.
1639 (vec_store_lanesxi<mode>): Likewise.
1640 (*aarch64_mov<mode>): Likewise.
1641 (aarch64_ld2<mode>_dreg): Likewise.
1642 (aarch64_ld2<mode>_dreg): Likewise.
1643 (aarch64_ld3<mode>_dreg): Likewise.
1644 (aarch64_ld3<mode>_dreg): Likewise.
1645 (aarch64_ld4<mode>_dreg): Likewise.
1646 (aarch64_ld4<mode>_dreg): Likewise.
1647 (aarch64_tbl1<mode>): Likewise.
1648 (aarch64_tbl2v16qi): Likewise.
1649 (aarch64_combinev16qi): Likewise.
1650 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1651 (aarch64_st2<mode>_dreg): Likewise.
1652 (aarch64_st2<mode>_dreg): Likewise.
1653 (aarch64_st3<mode>_dreg): Likewise.
1654 (aarch64_st3<mode>_dreg): Likewise.
1655 (aarch64_st4<mode>_dreg): Likewise.
1656 (aarch64_st4<mode>_dreg): Likewise.
1657 (*aarch64_simd_ld1r<mode>): Likewise.
1658 (aarch64_frecpe<mode>): Likewise.
1659 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1660 (aarch64_frecps<mode>): Likewise.
1662 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1664 * config/arm/iterators.md (V_elem_ch): New.
1666 (VQH_type): Likewise.
1667 * config/arm/arm.md (is_neon_type): New.
1668 (conds): Use is_neon_type.
1669 (anddi3_insn): Update type attribute.
1670 (xordi3_insn): Likewise.
1671 (one_cmpldi2): Likewise.
1672 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1673 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1674 (*movmisalign<mode>_neon_store): Likewise.
1675 (*movmisalign<mode>_neon_load): Likewise.
1676 (vec_set<mode>_internal): Likewise.
1677 (vec_set<mode>_internal): Likewise.
1678 (vec_setv2di_internal): Likewise.
1679 (vec_extract<mode>): Likewise.
1680 (vec_extract<mode>): Likewise.
1681 (vec_extractv2di): Likewise.
1682 (*add<mode>3_neon): Likewise.
1683 (adddi3_neon): Likewise.
1684 (*sub<mode>3_neon): Likewise.
1685 (subdi3_neon): Likewise.
1686 (fma<VCVTF:mode>4): Likewise.
1687 (fma<VCVTF:mode>4_intrinsic): Likewise.
1688 (*fmsub<VCVTF:mode>4): Likewise.
1689 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1690 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1691 (ior<mode>3): Likewise.
1692 (and<mode>3): Likewise.
1693 (orn<mode>3_neon): Likewise.
1694 (orndi3_neon): Likewise.
1695 (bic<mode>3_neon): Likewise.
1696 (bicdi3_neon): Likewise.
1697 (xor<mode>3): Likewise.
1698 (one_cmpl<mode>2): Likewise.
1699 (abs<mode>2): Likewise.
1700 (neg<mode>2): Likewise.
1701 (negdi2_neon): Likewise.
1702 (*umin<mode>3_neon): Likewise.
1703 (*umax<mode>3_neon): Likewise.
1704 (*smin<mode>3_neon): Likewise.
1705 (*smax<mode>3_neon): Likewise.
1706 (vashl<mode>3): Likewise.
1707 (vashr<mode>3_imm): Likewise.
1708 (vlshr<mode>3_imm): Likewise.
1709 (ashl<mode>3_signed): Likewise.
1710 (ashl<mode>3_unsigned): Likewise.
1711 (neon_load_count): Likewise.
1712 (ashldi3_neon_noclobber): Likewise.
1713 (ashldi3_neon): Likewise.
1714 (signed_shift_di3_neon): Likewise.
1715 (unsigned_shift_di3_neon): Likewise.
1716 (ashrdi3_neon_imm_noclobber): Likewise.
1717 (lshrdi3_neon_imm_noclobber): Likewise.
1718 (<shift>di3_neon): Likewise.
1719 (widen_ssum<mode>3): Likewise.
1720 (widen_usum<mode>3): Likewise.
1721 (quad_halves_<code>v4si): Likewise.
1722 (quad_halves_<code>v4sf): Likewise.
1723 (quad_halves_<code>v8hi): Likewise.
1724 (quad_halves_<code>v16qi): Likewise.
1725 (reduc_splus_v2di): Likewise.
1726 (neon_vpadd_internal<mode>): Likewise.
1727 (neon_vpsmin<mode>): Likewise.
1728 (neon_vpsmax<mode>): Likewise.
1729 (neon_vpumin<mode>): Likewise.
1730 (neon_vpumax<mode>): Likewise.
1731 (*ss_add<mode>_neon): Likewise.
1732 (*us_add<mode>_neon): Likewise.
1733 (*ss_sub<mode>_neon): Likewise.
1734 (*us_sub<mode>_neon): Likewise.
1735 (neon_vadd<mode>_unspec): Likewise.
1736 (neon_vaddl<mode>): Likewise.
1737 (neon_vaddw<mode>): Likewise.
1738 (neon_vhadd<mode>): Likewise.
1739 (neon_vqadd<mode>): Likewise.
1740 (neon_vaddhn<mode>): Likewise.
1741 (neon_vmul<mode>): Likewise.
1742 (neon_vfms<VCVTF:mode>): Likewise.
1743 (neon_vmlal<mode>): Likewise.
1744 (neon_vmls<mode>): Likewise.
1745 (neon_vmlsl<mode>): Likewise.
1746 (neon_vqdmulh<mode>): Likewise.
1747 (neon_vqdmlal<mode>): Likewise.
1748 (neon_vqdmlsl<mode>): Likewise.
1749 (neon_vmull<mode>): Likewise.
1750 (neon_vqdmull<mode>): Likewise.
1751 (neon_vsub<mode>_unspec): Likewise.
1752 (neon_vsubl<mode>): Likewise.
1753 (neon_vsubw<mode>): Likewise.
1754 (neon_vqsub<mode>): Likewise.
1755 (neon_vhsub<mode>): Likewise.
1756 (neon_vsubhn<mode>): Likewise.
1757 (neon_vceq<mode>): Likewise.
1758 (neon_vcge<mode>): Likewise.
1759 (neon_vcgeu<mode>): Likewise.
1760 (neon_vcgt<mode>): Likewise.
1761 (neon_vcgtu<mode>): Likewise.
1762 (neon_vcle<mode>): Likewise.
1763 (neon_vclt<mode>): Likewise.
1764 (neon_vcage<mode>): Likewise.
1765 (neon_vcagt<mode>): Likewise.
1766 (neon_vtst<mode>): Likewise.
1767 (neon_vabd<mode>): Likewise.
1768 (neon_vabdl<mode>): Likewise.
1769 (neon_vaba<mode>): Likewise.
1770 (neon_vabal<mode>): Likewise.
1771 (neon_vmax<mode>): Likewise.
1772 (neon_vmin<mode>): Likewise.
1773 (neon_vpaddl<mode>): Likewise.
1774 (neon_vpadal<mode>): Likewise.
1775 (neon_vpmax<mode>): Likewise.
1776 (neon_vpmin<mode>): Likewise.
1777 (neon_vrecps<mode>): Likewise.
1778 (neon_vrsqrts<mode>): Likewise.
1779 (neon_vqabs<mode>): Likewise.
1780 (neon_vqneg<mode>): Likewise.
1781 (neon_vcls<mode>): Likewise.
1782 (clz<mode>2): Likewise.
1783 (popcount<mode>2): Likewise.
1784 (neon_vrecpe<mode>): Likewise.
1785 (neon_vrsqrte<mode>): Likewise.
1786 (neon_vget_lane<mode>_sext_internal): Likewise.
1787 (neon_vget_lane<mode>_zext_internal): Likewise.
1788 (neon_vdup_n<mode>): Likewise.
1789 (neon_vdup_n<mode>): Likewise.
1790 (neon_vdup_nv2di): Likewise.
1791 (neon_vdup_lane<mode>_interal): Likewise.
1792 (*neon_vswp<mode>): Likewise.
1793 (neon_vcombine<mode>): Likewise.
1794 (float<mode><V_cvtto>2): Likewise.
1795 (floatuns<mode><V_cvtto>2): Likewise.
1796 (fix_trunc<mode><V_cvtto>2): Likewise.
1797 (fixuns_trunc<mode><V_cvtto>2
1798 (neon_vcvt<mode>): Likewise.
1799 (neon_vcvt<mode>): Likewise.
1800 (neon_vcvtv4sfv4hf): Likewise.
1801 (neon_vcvtv4hfv4sf): Likewise.
1802 (neon_vcvt_n<mode>): Likewise.
1803 (neon_vcvt_n<mode>): Likewise.
1804 (neon_vmovn<mode>): Likewise.
1805 (neon_vqmovn<mode>): Likewise.
1806 (neon_vqmovun<mode>): Likewise.
1807 (neon_vmovl<mode>): Likewise.
1808 (neon_vmul_lane<mode>): Likewise.
1809 (neon_vmul_lane<mode>): Likewise.
1810 (neon_vmull_lane<mode>): Likewise.
1811 (neon_vqdmull_lane<mode>): Likewise.
1812 (neon_vqdmulh_lane<mode>): Likewise.
1813 (neon_vqdmulh_lane<mode>): Likewise.
1814 (neon_vmla_lane<mode>): Likewise.
1815 (neon_vmla_lane<mode>): Likewise.
1816 (neon_vmlal_lane<mode>): Likewise.
1817 (neon_vqdmlal_lane<mode>): Likewise.
1818 (neon_vmls_lane<mode>): Likewise.
1819 (neon_vmls_lane<mode>): Likewise.
1820 (neon_vmlsl_lane<mode>): Likewise.
1821 (neon_vqdmlsl_lane<mode>): Likewise.
1822 (neon_vext<mode>): Likewise.
1823 (neon_vrev64<mode>): Likewise.
1824 (neon_vrev32<mode>): Likewise.
1825 (neon_vrev16<mode>): Likewise.
1826 (neon_vbsl<mode>_internal): Likewise.
1827 (neon_vshl<mode>): Likewise.
1828 (neon_vqshl<mode>): Likewise.
1829 (neon_vshr_n<mode>): Likewise.
1830 (neon_vshrn_n<mode>): Likewise.
1831 (neon_vqshrn_n<mode>): Likewise.
1832 (neon_vqshrun_n<mode>): Likewise.
1833 (neon_vshl_n<mode>): Likewise.
1834 (neon_vqshl_n<mode>): Likewise.
1835 (neon_vqshlu_n<mode>): Likewise.
1836 (neon_vshll_n<mode>): Likewise.
1837 (neon_vsra_n<mode>): Likewise.
1838 (neon_vsri_n<mode>): Likewise.
1839 (neon_vsli_n<mode>): Likewise.
1840 (neon_vtbl1v8qi): Likewise.
1841 (neon_vtbl2v8qi): Likewise.
1842 (neon_vtbl3v8qi): Likewise.
1843 (neon_vtbl4v8qi): Likewise.
1844 (neon_vtbl1v16qi): Likewise.
1845 (neon_vtbl2v16qi): Likewise.
1846 (neon_vcombinev16qi): Likewise.
1847 (neon_vtbx1v8qi): Likewise.
1848 (neon_vtbx2v8qi): Likewise.
1849 (neon_vtbx3v8qi): Likewise.
1850 (neon_vtbx4v8qi): Likewise.
1851 (*neon_vtrn<mode>_insn): Likewise.
1852 (*neon_vzip<mode>_insn): Likewise.
1853 (*neon_vuzp<mode>_insn): Likewise.
1854 (neon_vld1<mode>): Likewise.
1855 (neon_vld1_lane<mode>): Likewise.
1856 (neon_vld1_lane<mode>): Likewise.
1857 (neon_vld1_dup<mode>): Likewise.
1858 (neon_vld1_dup<mode>): Likewise.
1859 (neon_vld1_dupv2di): Likewise.
1860 (neon_vst1<mode>): Likewise.
1861 (neon_vst1_lane<mode>): Likewise.
1862 (neon_vst1_lane<mode>): Likewise.
1863 (neon_vld2<mode>): Likewise.
1864 (neon_vld2<mode>): Likewise.
1865 (neon_vld2_lane<mode>): Likewise.
1866 (neon_vld2_lane<mode>): Likewise.
1867 (neon_vld2_dup<mode>): Likewise.
1868 (neon_vst2<mode>): Likewise.
1869 (neon_vst2<mode>): Likewise.
1870 (neon_vst2_lane<mode>): Likewise.
1871 (neon_vst2_lane<mode>): Likewise.
1872 (neon_vld3<mode>): Likewise.
1873 (neon_vld3qa<mode>): Likewise.
1874 (neon_vld3qb<mode>): Likewise.
1875 (neon_vld3_lane<mode>): Likewise.
1876 (neon_vld3_lane<mode>): Likewise.
1877 (neon_vld3_dup<mode>): Likewise.
1878 (neon_vst3<mode>): Likewise.
1879 (neon_vst3qa<mode>): Likewise.
1880 (neon_vst3qb<mode>): Likewise.
1881 (neon_vst3_lane<mode>): Likewise.
1882 (neon_vst3_lane<mode>): Likewise.
1883 (neon_vld4<mode>): Likewise.
1884 (neon_vld4qa<mode>): Likewise.
1885 (neon_vld4qb<mode>): Likewise.
1886 (neon_vld4_lane<mode>): Likewise.
1887 (neon_vld4_lane<mode>): Likewise.
1888 (neon_vld4_dup<mode>): Likewise.
1889 (neon_vst4<mode>): Likewise.
1890 (neon_vst4qa<mode>): Likewise.
1891 (neon_vst4qb<mode>): Likewise.
1892 (neon_vst4_lane<mode>): Likewise.
1893 (neon_vst4_lane<mode>): Likewise.
1894 (neon_vec_unpack<US>_lo_<mode>): Likewise.
1895 (neon_vec_unpack<US>_hi_<mode>): Likewise.
1896 (neon_vec_<US>mult_lo_<mode>): Likewise.
1897 (neon_vec_<US>mult_hi_<mode>): Likewise.
1898 (neon_vec_<US>shiftl_<mode>): Likewise.
1899 (neon_unpack<US>_<mode>): Likewise.
1900 (neon_vec_<US>mult_<mode>): Likewise.
1901 (vec_pack_trunc_<mode>): Likewise.
1902 (neon_vec_pack_trunc_<mode>): Likewise.
1903 (neon_vabd<mode>_2): Likewise.
1904 (neon_vabd<mode>_3): Likewise.
1906 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1908 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
1909 (load_pair): Update type attribute.
1910 (store_pair): Update type attribute.
1911 * config/aarch64/iterators.md (q): New.
1913 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1915 * config/arm/types.md: Add new types for Neon insns.
1917 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1918 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1919 Sergey Lega <sergey.s.lega@intel.com>
1920 Anna Tikhonova <anna.tikhonova@intel.com>
1921 Ilya Tocar <ilya.tocar@intel.com>
1922 Andrey Turetskiy <andrey.turetskiy@intel.com>
1923 Ilya Verbin <ilya.verbin@intel.com>
1924 Kirill Yukhin <kirill.yukhin@intel.com>
1925 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1927 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
1928 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
1929 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
1931 (srcp14<mode>): Ditto.
1932 (rsqrt14<mode>): Ditto.
1933 (rsqrt14<mode>): Ditto.
1934 (avx512f_vmscalef<mode>): Ditto.
1935 (avx512f_scalef<mode>): Ditto.
1936 (avx512f_getexp<mode>): Ditto.
1937 (avx512f_sgetexp<mode>): Ditto.
1938 (avx512f_fixupimm<mode>): Ditto.
1939 (avx512f_sfixupimm<mode>): Ditto.
1940 (avx512f_rndscale<mode>): Ditto.
1941 (*avx512er_exp2<mode>): Ditto.
1942 (*avx512er_rcp28<mode>): Ditto.
1943 (avx512er_rsqrt28<mode>): Ditto.
1944 (avx512f_getmant<mode>): Ditto.
1945 (avx512f_getmant<mode>): Ditto.
1946 (avx512f_rndscale<mode>): Fix formatting.
1948 2013-10-15 Martin Jambor <mjambor@suse.cz>
1950 * ipa-utils.h (ipa_edge_within_scc): Declare.
1951 * ipa-cp.c (edge_within_scc): Moved...
1952 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
1954 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1955 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1956 Sergey Lega <sergey.s.lega@intel.com>
1957 Anna Tikhonova <anna.tikhonova@intel.com>
1958 Ilya Tocar <ilya.tocar@intel.com>
1959 Andrey Turetskiy <andrey.turetskiy@intel.com>
1960 Ilya Verbin <ilya.verbin@intel.com>
1961 Kirill Yukhin <kirill.yukhin@intel.com>
1962 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1964 * config/i386/predicates.md (const_8_to_15_operand): New.
1965 (const_16_to_31_operand): Ditto.
1966 * config/i386/sse.md (V8FI): New.
1968 (reduc_splus_v8df): Ditto.
1969 (reduc_splus_v16sf): Ditto.
1970 (avx512f_vextract<shuffletype>32x4_1): Ditto.
1971 (vec_extract_hi_<mode>): Ditto.
1972 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
1973 (vec_set_lo_<mode>): Ditto.
1974 (vec_set_hi_<mode>): Ditto.
1975 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
1976 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
1977 (avx512f_pshufd_1): Ditto.
1978 (avx512f_broadcast<mode>): Ditto.
1979 (avx512f_broadcast<mode>): Ditto.
1980 (define_split): Split vec_extract_lo into move.
1981 (ssequartermode): Ditto.
1982 (ssedoublemode): Extened with wider modes.
1983 (vec_extract_lo_<mode>): Ditto.
1985 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
1986 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1987 Sergey Lega <sergey.s.lega@intel.com>
1988 Anna Tikhonova <anna.tikhonova@intel.com>
1989 Ilya Tocar <ilya.tocar@intel.com>
1990 Andrey Turetskiy <andrey.turetskiy@intel.com>
1991 Ilya Verbin <ilya.verbin@intel.com>
1992 Kirill Yukhin <kirill.yukhin@intel.com>
1993 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1995 * config/i386/predicates.md (register_or_constm1_operand): New.
1996 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
1997 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
1998 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
1999 UNSPEC_SCATTER_PREFETCH
2001 (avx512f_ucmp<mode>3): Ditto.
2002 (avx512f_vternlog<mode>): Ditto.
2003 (avx512f_align<mode>): Ditto.
2004 (<shift_insn><mode>3): Ditto.
2005 (avx512f_<rotate>v<mode>): Ditto.
2006 (avx512f_<rotate><mode>): Ditto.
2007 (avx512f_eq<mode>3): Ditto.
2008 (avx512f_eq<mode>3_1): Ditto.
2009 (avx512f_gt<mode>3): Ditto.
2010 (avx512f_testm<mode>3): Ditto.
2011 (avx512f_testnm<mode>3): Ditto.
2012 (avx512pf_gatherpf<mode>): Ditto.
2013 (*avx512pf_gatherpf<mode>_mask): Ditto.
2014 (*avx512pf_gatherpf<mode>): Ditto.
2015 (avx512pf_scatterpf<mode>): Ditto.
2016 (*avx512pf_scatterpf<mode>_mask): Ditto.
2017 (*avx512pf_scatterpf<mode>): Ditto.
2018 (avx512f_vec_dup_gpr<mode>): Ditto.
2019 (clz<mode>2): Ditto.
2020 (conflict<mode>): Ditto.
2021 (REDUC_SMINMAX_MODE): Extened with wider modes.
2022 (reduc_<code>_<mode>): Ditto.
2023 (vlshr<mode>3): Ditto.
2024 (vashl<mode>3): Ditto.
2026 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2027 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2028 Sergey Lega <sergey.s.lega@intel.com>
2029 Anna Tikhonova <anna.tikhonova@intel.com>
2030 Ilya Tocar <ilya.tocar@intel.com>
2031 Andrey Turetskiy <andrey.turetskiy@intel.com>
2032 Ilya Verbin <ilya.verbin@intel.com>
2033 Kirill Yukhin <kirill.yukhin@intel.com>
2034 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2036 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
2039 (avx512fmaskmode): Ditto.
2040 (bcstscalarsuff): Ditto.
2041 (avx512f_blendm<mode>): Ditto.
2042 (cmp_imm_predicate): Ditto.
2043 (avx512f_cmp<mode>3): Ditto.
2044 (avx512f_vec_dup<mode>): Ditto.
2045 (avx512f_vec_dup_mem<mode>): Ditto.
2046 (avx512f_vpermi2var<mode>3): Ditto.
2047 (avx512f_vpermt2var<mode>3): Ditto.
2048 (vec_init<mode>): Ditto.
2049 (avx512f_gathersi<mode>): Ditto.
2050 (*avx512f_gathersi<mode>): Ditto.
2051 (*avx512f_gathersi<mode>_2): Ditto.
2052 (avx512f_gatherdi<mode>): Ditto.
2053 (*avx512f_gatherdi<mode>): Ditto.
2054 (*avx512f_gatherdi<mode>_2): Ditto.
2055 (avx512f_scattersi<mode>): Ditto.
2056 (*avx512f_scattersi<mode>): Ditto.
2057 (avx512f_scatterdi<mode>): Ditto.
2058 (*avx512f_scatterdi<mode>): Ditto.
2059 (sseintprefix): Extened with wider modes.
2060 (VEC_GATHER_IDXSI): Ditto.
2061 (VEC_GATHER_IDXDI): Ditto.
2062 (VEC_GATHER_SRCDI): Ditto.
2064 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
2065 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2067 * config/arm/t-aprofile: New file.
2068 * config.gcc: Handle --with-multilib-list option.
2070 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
2072 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
2075 2013-10-15 Richard Biener <rguenther@suse.de>
2077 * tree-loop-distribution.c (build_empty_rdg): Inline into
2079 (rdg_flag_vertex): Inline into single user.
2080 (rdg_flag_vertex_and_dependent): Likewise.
2081 (build_rdg_partition_for_vertex): Remove processed bitmap.
2082 (rdg_build_partitions): Simplify.
2084 2013-10-15 Richard Biener <rguenther@suse.de>
2086 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
2087 Restructure forwarding through conversions and copies to
2088 avoid performing copy-propagation the wrong way. Adjust
2089 recursion invocations.
2090 (forward_propagate_addr_expr): Add argument stating if we
2091 are recursing from a single-use.
2092 (ssa_forward_propagate_and_combine): Adjust.
2094 2013-10-14 David Malcolm <dmalcolm@redhat.com>
2096 * dumpfile.h (gcc::dump_manager): New class, to hold state
2097 relating to dumpfile management.
2098 (get_dump_file_name): Remove in favor of method of dump_manager.
2099 (dump_initialized_p): Likewise.
2100 (dump_start): Likewise.
2101 (dump_finish): Likewise.
2102 (dump_switch_p): Likewise.
2103 (dump_register): Likewise.
2104 (get_dump_file_info): Likewise.
2105 * context.c (gcc::context::context): Construct the dump_manager
2107 * context.h (gcc::context::get_dumps): New.
2108 (gcc::context::m_dumps): New.
2109 * coverage.c (coverage_init): Port to dump_manager API.
2110 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
2111 (extra_dump_files_in_use): Likewise.
2112 (extra_dump_files_alloced): Likewise.
2113 (gcc::dump_manager::dump_manager): New.
2114 (dump_register): Convert to...
2115 (gcc::dump_manager::dump_register): ...method, replacing
2116 function-static next_dump with m_next_dump field.
2117 (get_dump_file_info): Convert to...
2118 (gcc::dump_manager::get_dump_file_info): ...method.
2119 (get_dump_file_name): Convert to...
2120 (gcc::dump_manager::get_dump_file_name): ...method.
2121 (dump_start): Convert to...
2122 (gcc::dump_manager::dump_start): ...method.
2123 (dump_finish): Convert to...
2124 (gcc::dump_manager::dump_finish): ...method.
2125 (dump_begin): Replace body with...
2126 (gcc::dump_manager::dump_begin): ...new method.
2127 (dump_phase_enabled_p): Convert to...
2128 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2129 (dump_phase_enabled_p): Convert to...
2130 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2131 (dump_initialized_p): Convert to...
2132 (gcc::dump_manager::dump_initialized_p): ...method.
2133 (dump_flag_name): Replace body with...
2134 (gcc::dump_manager::dump_flag_name): ...new method.
2135 (dump_enable_all): Convert to...
2136 (gcc::dump_manager::dump_enable_all): ...new method.
2137 (opt_info_enable_passes): Convert to...
2138 (gcc::dump_manager::opt_info_enable_passes): ...new method.
2139 (dump_switch_p_1): Convert to...
2140 (gcc::dump_manager::dump_switch_p_1): ...new method.
2141 (dump_switch_p): Convert to...
2142 (gcc::dump_manager::dump_switch_p): ...new method.
2143 (opt_info_switch_p): Port to dump_manager API.
2144 (enable_rtl_dump_file): Likewise.
2145 * opts-global.c (handle_common_deferred_options): Port to new
2147 * passes.c (pass_manager::finish_optimization_passes): Likewise.
2148 (pass_manager::register_one_dump_file): Likewise.
2149 (pass_manager::register_pass): Likewise.
2150 (pass_init_dump_file): Likewise.
2151 (pass_fini_dump_file): Likewise.
2152 * statistics.c (statistics_early_init): Likewise.
2154 2013-10-14 Richard Biener <rguenther@suse.de>
2156 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
2157 iterative_hash_canonical_type, gimple_canonical_type_hash,
2158 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
2159 gimple_register_canonical_type, print_gimple_types_stats,
2160 free_gimple_type_tables): Move to lto/lto.c
2161 (gt-gimple.h): Do not include.
2162 * gimple.h (gimple_register_canonical_type,
2163 print_gimple_types_stats, free_gimple_type_tables): Remove.
2164 * Makefile.in (GTFILES): Remove gimple.c.
2166 2013-10-14 Travis Snoozy <quandary@remstate.com>
2169 * config/msp430/msp430.c (msp430_option_override): Correct thinko
2170 scanning for msp430x targets.
2172 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
2175 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2176 (registers_ok_for_ldd_peep): Move around.
2177 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2178 * config/sparc/sparc.md (widening peepholes): Use it.
2180 2013-10-14 Richard Biener <rguenther@suse.de>
2184 * gimple.c (iterative_hash_canonical_type): Make sure to
2185 record the hash into the correct hashtable slot.
2187 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2189 PR rtl-optimization/58662
2190 * combine.c (try_combine): Take into account death nodes on I2 when
2191 splitting a PARALLEL of two independent SETs. Fix dump message.
2193 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2196 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2197 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2198 * config/sh/t-sh (sh_treg_combine.o): New entry.
2199 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2200 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2201 (register_sh_passes): New function. Register sh_treg_combine pass.
2202 (sh_option_override): Invoke it.
2203 (sh_canonicalize_comparison): Handle op0_preserve_value.
2204 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2205 opportunities. Canonicalize branch condition.
2206 (nott): Allow only if pseudos can be created for non-SH2A.
2208 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2211 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2212 (ix86_expand_movmem): Replace copy_addr_to_reg with
2213 ix86_copy_addr_to_reg.
2214 (ix86_expand_setmem): Likewise.
2216 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2218 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2219 provided by ix86_fp_compare_mode instead of CCFPUmode.
2221 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2223 * config/aarch64/arm_neon.h
2224 (vtbx<1,3>_<psu>8): Fix register constriants.
2226 2013-10-11 Jeff Law <law@redhat.com>
2228 PR tree-optimization/58640
2229 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2230 threading paths that cross over two loop entry points.
2232 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2234 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2235 handle vector float as well.
2236 (*vsx_le_perm_load_v4si): Likewise.
2237 (*vsx_le_perm_store_v2di): Likewise.
2238 (*vsx_le_perm_store_v4si): Likewise.
2240 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2242 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2243 directly to circumvent subtract from splat{31} workaround.
2244 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2246 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2247 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2248 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2249 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2250 separate big and little endian logic.
2251 (*altivec_vperm_<mode>_internal): New define_insn.
2252 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2253 separate big and little endian logic.
2254 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2255 (vec_permv16qi): Add little endian logic.
2257 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2259 * doc/extend.texi (returns_nonnull): Remove arguments.
2261 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2262 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2263 Sergey Lega <sergey.s.lega@intel.com>
2264 Anna Tikhonova <anna.tikhonova@intel.com>
2265 Ilya Tocar <ilya.tocar@intel.com>
2266 Andrey Turetskiy <andrey.turetskiy@intel.com>
2267 Ilya Verbin <ilya.verbin@intel.com>
2268 Kirill Yukhin <kirill.yukhin@intel.com>
2269 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2271 * config/i386/sse.md (VI48F_256_512): New.
2272 (avx2_permvar<mode>): Change to ...
2273 (<avx2_avx512f>_permvar<mode>): This.
2275 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2276 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2277 Sergey Lega <sergey.s.lega@intel.com>
2278 Anna Tikhonova <anna.tikhonova@intel.com>
2279 Ilya Tocar <ilya.tocar@intel.com>
2280 Andrey Turetskiy <andrey.turetskiy@intel.com>
2281 Ilya Verbin <ilya.verbin@intel.com>
2282 Kirill Yukhin <kirill.yukhin@intel.com>
2283 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2285 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2286 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2287 * config/i386/sse.md (VI4_AVX): New.
2288 (sf2simodelower): Ditto.
2289 (sse2_cvtps2dq): Change to ...
2290 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2292 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2293 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2294 Sergey Lega <sergey.s.lega@intel.com>
2295 Anna Tikhonova <anna.tikhonova@intel.com>
2296 Ilya Tocar <ilya.tocar@intel.com>
2297 Andrey Turetskiy <andrey.turetskiy@intel.com>
2298 Ilya Verbin <ilya.verbin@intel.com>
2299 Kirill Yukhin <kirill.yukhin@intel.com>
2300 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2302 * config/i386/sse.md (V_512): New.
2304 (vcond<V_512:mode><VF_512:mode>): Ditto.
2305 (vcond<V_512:mode><VI_512:mode>): Ditto.
2306 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2308 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2309 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2310 Sergey Lega <sergey.s.lega@intel.com>
2311 Anna Tikhonova <anna.tikhonova@intel.com>
2312 Ilya Tocar <ilya.tocar@intel.com>
2313 Andrey Turetskiy <andrey.turetskiy@intel.com>
2314 Ilya Verbin <ilya.verbin@intel.com>
2315 Kirill Yukhin <kirill.yukhin@intel.com>
2316 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2318 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2319 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2321 (fma<mode>4): Removed condition.
2322 (fms<mode>4): Ditto.
2323 (fnma<mode>4): Ditto.
2324 (fnms<mode>4): Ditto.
2325 (fma4i_fmadd_<mode>): Ditto.
2326 (*fma_fmadd_<mode>): Ditto.
2327 (*fma_fmsub_<mode>): Ditto.
2328 (*fma_fnmadd_<mode>): Ditto.
2329 (*fma_fnmsub_<mode>): Ditto.
2330 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2331 (*fma_fmaddsub_<mode>): Ditto.
2332 (*fma_fmsubadd_<mode>): Ditto.
2333 (*fmai_fmadd_<mode>): Ditto.
2334 (*fmai_fmsub_<mode>): Ditto.
2335 (*fmai_fnmadd_<mode>): Ditto.
2336 (*fmai_fnmsub_<mode>): Ditto.
2338 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2339 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2340 Sergey Lega <sergey.s.lega@intel.com>
2341 Anna Tikhonova <anna.tikhonova@intel.com>
2342 Ilya Tocar <ilya.tocar@intel.com>
2343 Andrey Turetskiy <andrey.turetskiy@intel.com>
2344 Ilya Verbin <ilya.verbin@intel.com>
2345 Kirill Yukhin <kirill.yukhin@intel.com>
2346 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2348 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2349 (VI124_256): Changed to ...
2350 (VI124_256_48_512): This.
2351 (ssepackmode): Extended with wider modes.
2352 (<code><mode>3): Changed iterator.
2353 (*avx2_<code><mode>3): Ditto.
2354 (vec_pack_trunc_<mode>): Ditto.
2356 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2357 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2358 Sergey Lega <sergey.s.lega@intel.com>
2359 Anna Tikhonova <anna.tikhonova@intel.com>
2360 Ilya Tocar <ilya.tocar@intel.com>
2361 Andrey Turetskiy <andrey.turetskiy@intel.com>
2362 Ilya Verbin <ilya.verbin@intel.com>
2363 Kirill Yukhin <kirill.yukhin@intel.com>
2364 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2366 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2367 (VI8F_256_512): Ditto.
2368 (abs<mode>2): Changed iterator.
2369 (avx2_perm<mode>): Changed to ...
2370 (<avx2_avx512f>_perm<mode>): This.
2371 (avx2_perm<mode>_1): Changed to ...
2372 (<avx2_avx512f>_perm<mode>_1): This.
2374 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2375 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2376 Sergey Lega <sergey.s.lega@intel.com>
2377 Anna Tikhonova <anna.tikhonova@intel.com>
2378 Ilya Tocar <ilya.tocar@intel.com>
2379 Andrey Turetskiy <andrey.turetskiy@intel.com>
2380 Ilya Verbin <ilya.verbin@intel.com>
2381 Kirill Yukhin <kirill.yukhin@intel.com>
2382 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2384 * config/i386/sse.md (VI48_AVX512F): New.
2385 (VI48_AVX2): Changed to ...
2386 (VI48_AVX2_48_AVX512F): This.
2387 (avx2_ashrv<mode>): Changed to ...
2388 (<avx2_avx512f>_ashrv<mode>): This.
2389 (avx2_<shift_insn>v<mode>): Changed to ...
2390 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2392 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2393 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2394 Sergey Lega <sergey.s.lega@intel.com>
2395 Anna Tikhonova <anna.tikhonova@intel.com>
2396 Ilya Tocar <ilya.tocar@intel.com>
2397 Andrey Turetskiy <andrey.turetskiy@intel.com>
2398 Ilya Verbin <ilya.verbin@intel.com>
2399 Kirill Yukhin <kirill.yukhin@intel.com>
2400 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2402 * config/i386/sse.md (VI4_AVX512F): New.
2403 (VI8_AVX2_AVX512F): Ditto.
2404 (mul<mode>3): Extended with wider modes.
2405 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2406 (mul<mode>3): Ditto.
2407 (vec_widen_<s>mult_odd_<mode>): Ditto.
2409 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2410 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2411 Sergey Lega <sergey.s.lega@intel.com>
2412 Anna Tikhonova <anna.tikhonova@intel.com>
2413 Ilya Tocar <ilya.tocar@intel.com>
2414 Andrey Turetskiy <andrey.turetskiy@intel.com>
2415 Ilya Verbin <ilya.verbin@intel.com>
2416 Kirill Yukhin <kirill.yukhin@intel.com>
2417 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2419 * config/i386/sse.md (VI2_AVX512F): New.
2420 (VI124_AVX512F): Ditto.
2421 (sseunpackmode): Extended with wider modes.
2422 (sseunpackfltmode): Ditto.
2423 (vec_unpacks_float_hi_<mode>): Ditto.
2424 (vec_unpacks_float_lo_<mode>): Ditto.
2425 (vec_unpacku_float_hi_<mode>): Ditto.
2426 (vec_unpacku_float_lo_<mode>): Ditto.
2427 (vec_unpacks_lo_<mode>): Ditto.
2428 (vec_unpacks_hi_<mode>): Ditto.
2429 (vec_unpacku_lo_<mode>): Ditto.
2430 (vec_unpacku_hi_<mode>): Ditto.
2432 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2433 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2434 Sergey Lega <sergey.s.lega@intel.com>
2435 Anna Tikhonova <anna.tikhonova@intel.com>
2436 Ilya Tocar <ilya.tocar@intel.com>
2437 Andrey Turetskiy <andrey.turetskiy@intel.com>
2438 Ilya Verbin <ilya.verbin@intel.com>
2439 Kirill Yukhin <kirill.yukhin@intel.com>
2440 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2442 * config/i386/i386.md (multdiv): New.
2443 (multdiv_mnemonic): Ditto.
2444 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2445 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2446 (<sse>_vmdiv<mode>3): Removed.
2448 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2449 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2450 Sergey Lega <sergey.s.lega@intel.com>
2451 Anna Tikhonova <anna.tikhonova@intel.com>
2452 Ilya Tocar <ilya.tocar@intel.com>
2453 Andrey Turetskiy <andrey.turetskiy@intel.com>
2454 Ilya Verbin <ilya.verbin@intel.com>
2455 Kirill Yukhin <kirill.yukhin@intel.com>
2456 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2458 * config/i386/sse.md (V): Extended with wider modes.
2460 (ssehalfvecmode): Ditto.
2462 (ssepackfltmode): Ditto.
2463 (avx_vec_concat<mode>): Ditto.
2464 (V_256_512): New iterator.
2465 (VF2_512_256): Ditto.
2466 (si2dfmode): New attribute.
2467 (si2dfmodelower): Ditto.
2469 (concat_tg_mode): Ditto.
2470 (floatv4siv4df2): Changed to ...
2471 (float<si2dfmodelower><mode>2): This.
2472 (avx_cvtps2pd256): Changed to ...
2473 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2474 (vec_pack_trunc_v4df): Changed to ...
2475 (vec_pack_trunc_<mode>): This.
2476 (avx_vpermil<mode>): Changed to ...
2477 (<sse2_avx_avx512f>_vpermil<mode>): This.
2478 (<fixsuffix>fix_truncv8dfv8si2): New.
2479 (vec_pack_sfix_trunc_v8df): Ditto.
2480 (avx512f_rndscale<mode>): Ditto.
2481 (avx512f_roundpd512): Ditto.
2482 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2484 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2485 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2486 Sergey Lega <sergey.s.lega@intel.com>
2487 Anna Tikhonova <anna.tikhonova@intel.com>
2488 Ilya Tocar <ilya.tocar@intel.com>
2489 Andrey Turetskiy <andrey.turetskiy@intel.com>
2490 Ilya Verbin <ilya.verbin@intel.com>
2491 Kirill Yukhin <kirill.yukhin@intel.com>
2492 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2494 * config/i386/i386.md (any_fix): New iterator.
2495 (fixsuffix): New attribute.
2496 * config/i386/sse.md (VF1): Extened with wider modes.
2500 (sseintvecmodelower): Ditto.
2501 (ssescalarmode): Ditto.
2502 (ssescalarnum): Ditto.
2504 (ssexmmmode): Ditto.
2505 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2506 (<sse>_rcp<mode>2): Change iterator.
2507 (rsqrt<mode>2): Ditto.
2508 (<sse>_rsqrt<mode>2): Ditto.
2509 (avx2_vec_dup<mode>): Ditto.
2510 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2511 (round<mode>2_sfix): Ditto.
2512 (avx2_pbroadcast<mode>): Ditto.
2513 (*andnot<mode>3): Handle XI mode.
2514 (*<code><mode>3): Ditto.
2515 (AVXTOSSEMODE): Removed.
2516 (avx_vpermil<mode>): Changed to ...
2517 (<sse2_avx_avx512f>_vpermil<mode>): This.
2519 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2520 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2521 Sergey Lega <sergey.s.lega@intel.com>
2522 Anna Tikhonova <anna.tikhonova@intel.com>
2523 Ilya Tocar <ilya.tocar@intel.com>
2524 Andrey Turetskiy <andrey.turetskiy@intel.com>
2525 Ilya Verbin <ilya.verbin@intel.com>
2526 Kirill Yukhin <kirill.yukhin@intel.com>
2527 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2529 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2530 (<sse>_comi): Ditto.
2531 (<sse>_ucomi): Ditto.
2532 (sse_cvtss2siq_2): Ditto.
2533 (sse2_cvtsd2si): Ditto.
2534 (sse2_cvtsd2siq): Ditto.
2535 (sse2_cvttsd2si): Ditto.
2536 (sse2_cvttsd2siq): Ditto.
2537 (<shift_insn><mode>3): Ditto.
2538 (sse2_cvtsi2sdq): Update constraint and prefix.
2539 (sse_cvtsi2ss): Update prefix.
2540 (sse_cvtsi2ssq): Ditto.
2542 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2544 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2545 ignore internal calls.
2547 2013-10-11 Richard Biener <rguenther@suse.de>
2549 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2550 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2551 TREE_OVERFLOW is set.
2553 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2555 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2557 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2558 * gimple.h: Likewise.
2560 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2562 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2565 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2567 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2568 * configure: Regenerate.
2570 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2572 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2573 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2574 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2575 similarly pp_right_paren.
2576 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2577 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2578 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2579 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2581 (omp_declare_simd_clauses_equal,
2582 omp_remove_redundant_declare_simd_attrs): New functions.
2583 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2584 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2585 * tree.h (OMP_LOOP_CHECK): Define.
2586 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2587 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2588 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2589 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2590 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2591 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2592 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2593 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2594 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2595 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2596 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2597 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2598 (omp_remove_redundant_declare_simd_attrs): New prototype.
2599 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2600 GIMPLE_OMP_TEAMS): New codes.
2601 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2602 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2604 (target_nesting_level): New variable.
2605 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2606 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2607 for collapse > 1 static schedule unless ordered.
2608 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2609 (determine_parallel_type): Adjust get_ws_args_for caller.
2610 (install_var_field): Handle mask & 4 for double indirection.
2611 (scan_sharing_clauses): Ignore shared clause on teams construct.
2612 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2613 (create_omp_child_function): If inside target or declare target
2614 constructs, set "omp declare target" attribute on the child function.
2615 (find_combined_for): New function.
2616 (scan_omp_parallel): Handle combined loops.
2617 (scan_omp_target, scan_omp_teams): New functions.
2618 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2619 restrictions and set ctx->cancellable for cancellable constructs.
2620 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2621 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2622 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2623 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2624 (omp_clause_aligned_alignment): New function.
2625 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2626 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2627 on teams constructs. Handle user defined reductions and new
2629 (lower_reduction_clauses): Don't set placeholder to address of ref
2630 if it has already the right type.
2631 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2632 (expand_parallel_call): Use the new non-_start suffixed builtins,
2633 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2634 and GOMP_parallel_end after the call.
2635 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2636 (expand_omp_for_init_counts): Handle combined loops.
2637 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2639 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2640 end of cancellable loops.
2641 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2642 Likewise. Handle collapse > 1 loops.
2643 (expand_omp_simd): Handle combined loops.
2644 (expand_omp_for): Add inner_stmt argument, adjust callers of
2645 expand_omp_for* functions, use expand_omp_for_static*chunk even
2646 for collapse > 1 unless ordered.
2647 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2648 of cancellable sections.
2649 (expand_omp_single): Remove need_barrier variable, just rely on
2650 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2651 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2652 (expand_omp_atomic_load, expand_omp_atomic_store,
2653 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2654 (expand_omp_target): New function.
2655 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2656 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2657 (build_omp_regions_1): Immediately close region for
2658 GF_OMP_TARGET_KIND_UPDATE.
2659 (maybe_add_implicit_barrier_cancel): New function.
2660 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2662 (lower_omp_single): Likewise. Add clobber after the barrier.
2663 (lower_omp_taskgroup): New function.
2664 (lower_omp_for): Handle combined loops. Adjust
2665 lower_rec_input_clauses caller. Handle cancellation.
2666 (lower_depend_clauses): New function.
2667 (lower_omp_taskreg): Lower depend clauses. Adjust
2668 lower_rec_input_clauses caller. Add clobber after the call. Handle
2670 (lower_omp_target, lower_omp_teams): New functions.
2671 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2672 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2673 and GOMP_cancellation_point calls.
2674 (lower_omp): Fold stmts inside of target region.
2675 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2676 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2677 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2678 (BT_FN_VOID_OMPFN_PTR_UINT,
2679 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2680 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2681 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2682 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2683 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2684 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2685 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2686 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2687 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2688 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2689 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2690 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2691 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2692 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2693 BUILT_IN_GOMP_PARALLEL_END.
2694 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2695 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2696 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2697 GF_OMP_FOR_KIND_DISTRIBUTE.
2698 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2699 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2700 (dump_gimple_omp_return): Print lhs if it has any.
2701 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2702 gimple_omp_atomic_seq_cst_p.
2703 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2704 and GIMPLE_OMP_TEAMS.
2705 * langhooks.c (lhd_omp_mappable_type): New function.
2706 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2707 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2709 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2710 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2711 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2712 (struct gimplify_omp_ctx): Add combined_loop field.
2713 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2714 on stmts inside of target region.
2715 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2716 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2717 ORT_TARGET and ORT_TARGET_DATA.
2718 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2720 (omp_notice_threadprivate_variable): Complain about threadprivate
2721 variables in target region.
2722 (omp_notice_variable): Complain about vars with non-mappable type
2723 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2724 (omp_check_private): Ignore ORT_TARGET* regions.
2725 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2726 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2727 (find_combined_omp_for): New function.
2728 (gimplify_omp_for): Handle gimplification of combined loops.
2729 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2731 (gimplify_omp_target_update): New function.
2732 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2733 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2734 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2735 (gimplify_body): If fndecl has "omp declare target" attribute, add
2736 implicit ORT_TARGET context around it.
2737 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2738 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2739 * tree-nested.c (convert_nonlocal_reference_stmt,
2740 convert_local_reference_stmt, convert_gimple_call): Handle
2741 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2742 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2743 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2744 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2745 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2746 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2747 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2748 BUILT_IN_GOMP_LOOP_END_CANCEL,
2749 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2750 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2751 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2752 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2753 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2754 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2755 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2756 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2757 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2758 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2759 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2760 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2761 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2762 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2763 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2764 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2765 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2766 gimple_build_omp_teams): New functions.
2767 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2768 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2769 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2770 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2771 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2772 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2773 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2774 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2775 GF_OMP_ATOMIC_SEQ_CST): New.
2776 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2777 gimple_build_omp_teams): New prototypes.
2778 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2779 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2780 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2781 GIMPLE_OMP_SINGLE as end of range.
2782 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2783 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2784 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2785 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2786 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2787 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2788 gimple_omp_target_kind, gimple_omp_target_set_kind,
2789 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2790 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2791 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2792 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2793 gimple_omp_teams_set_clauses): New inlines.
2794 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2795 and GIMPLE_OMP_TASKGROUP.
2796 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2797 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2798 enum omp_clause_proc_bind_kind): New.
2799 (union omp_clause_subcode): Add depend_kind, map_kind and
2800 proc_bind_kind fields.
2801 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2802 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2803 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2804 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2805 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2807 2013-10-10 Teresa Johnson <tejohnson@google.com>
2809 * predict.c (tree_estimate_probability): Add new parameter
2810 for estimate_bb_frequencies.
2811 (estimate_bb_frequencies): Add new parameter to force estimation.
2812 (rebuild_frequencies): When max frequency in function is small,
2813 recompute counts from frequencies.
2814 * predict.h (estimate_bb_frequencies): New parameter.
2816 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2818 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2819 optimizations are disabled.
2821 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2823 * coverage.c (coverage_finish): Fix leak of da_file_name.
2825 2013-10-10 Jan Hubicka <jh@suse.cz>
2827 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2828 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2831 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2834 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2835 if any labels are in FALLTHRU_BB, use a special label emitted
2836 immediately after the asm goto insn rather than label_rtx
2838 (expand_asm_stmt): Adjust caller.
2839 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
2840 edge if the last insn in predecessor is a jump with single successor,
2841 but it isn't simplejump_p.
2843 2013-10-10 Richard Biener <rguenther@suse.de>
2845 PR tree-optimization/58656
2846 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
2848 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2850 * gimplify.c: Include expr.h and tm_p.h for targets with special
2851 va-arg padding requirements.
2853 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2855 * tree-flow.h: Move some prototypes to gimple.h.
2856 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
2857 * gimple.h: Relocate some prototypes from tree-flow.h
2858 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
2860 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
2861 (build_va_arg_indirect_ref): Relocate and make static.
2862 (std_gimplify_va_arg_expr): Relocate here.
2863 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
2864 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
2866 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2868 * doc/md.texi: Document the mnemonic attribute.
2870 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2873 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
2874 (set_attr_alternative x ...) when searching for user defined
2877 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
2879 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
2881 2013-10-09 Easwaran Raman <eraman@google.com>
2883 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
2884 * cfgexpand.c (defer_stack_allocation): ...use here
2885 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
2887 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
2889 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
2890 (operand_equal_for_value_replacement): New function, extracted from
2891 value_replacement and enhanced to catch more cases.
2892 (value_replacement): Use operand_equal_for_value_replacement.
2894 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2896 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2897 get_max_loop_iterations.
2899 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2901 * config/arm/aarch-common.c (arm_early_load_addr_dep):
2902 Place comment above function.
2904 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2906 * tree-flow.h: Remove all remaining prototypes, enums and structs that
2907 are not related to tree-cfg.c.
2908 * tree-ssa-address.h: New file. Relocate prototypes.
2909 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
2910 (addr_for_mem_ref): New. Combine call to get_address_description and
2911 return addr_for_mem_ref.
2912 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
2913 * tree-ssa-live.h: Adjust prototypes.
2914 * passes.c: Include tree-ssa-live.h.
2915 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
2916 * graphite.c (graphite_transform_loops): Make static.
2917 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2918 make_pass_graphite, pass_data_graphite_transforms,
2919 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
2920 * ipa-pure-const.c (warn_function_noreturn): Make static.
2921 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2922 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2923 Relocate from tree-cfg.c
2924 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
2926 (execute_warn_function_noreturn, gate_warn_function_noreturn,
2927 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
2928 Move to ipa-pure-const.c.
2929 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
2930 Relocate from tree-optimize.c.
2931 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
2932 make_pass_fixup_cfg): Move to tree-cfg.c.
2933 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
2934 Relocate some prototypes.
2935 * tree-data-ref.h (tree_check_data_deps) Add prototype.
2936 * tree-dump.c (dump_function_to_file): Remove prototype.
2937 Add tree-flow.h to the include file.
2938 * tree-dump.h: Remove prototype.
2939 * tree-parloops.h: New File. Add prototypes.
2940 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
2941 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
2942 from tree-ssa-loop.c.
2943 * tree-predcom.c (run_tree_predictive_commoning,
2944 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
2945 Relocate here from tree-ssa-loop.c.
2946 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
2947 ssa_name_values.release ().
2948 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
2949 (ssa_name_values): Relocate from tree-flow.h.
2950 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
2951 * tree-ssa-loop.c (run_tree_predictive_commoning,
2952 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
2953 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
2954 make_pass_graphite, pass_data_graphite_transforms,
2955 make_pass_graphite_transforms, gate_tree_parallelize_loops,
2956 tree_parallelize_loops, pass_data_parallelize_loops,
2957 make_pass_parallelize_loops): Move to other files.
2958 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
2960 * tree.h: Remove prototypes from tree-address.c.
2962 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2964 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
2965 (struct int_tree_map): Move to tree-hasher.h
2966 (SCALE, LABEL, PERCENT): Move to gimple.h
2967 * tree-flow-inline.h: Delete. Move functions to other files.
2968 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
2969 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
2970 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
2971 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
2972 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
2973 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
2974 * tree-hasher.h: Don't include tree-flow.h.
2975 (struct int_tree_map): Relocate from tree-flow.h.
2976 * tree-sra.c (contains_view_convert_expr_p): Relocate from
2977 tree-flow-inline.h and make static.
2978 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
2980 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
2981 tree-flow-inline.h and make static.
2982 * tree.h (is_global_var, may_be_aliased): Relocate from
2984 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
2985 * value-prof.c: No longer include tree-flow-inline.h.
2986 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
2988 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
2990 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
2991 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
2992 (enum move_pos): Move to tree-ssa-loop-im.h
2993 * cfgloop.h: Move some prototypes.
2994 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
2995 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
2996 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
2997 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
2998 (loop_containing_stmt): Relocate from tree-flow-inline.h.
2999 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
3000 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
3001 (enum move_pos): Relocate here.
3002 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
3004 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
3005 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
3006 make_pass_lim): Relocate here from tree-ssa-loop.c.
3007 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
3009 (loop_edge_to_cancel, unloop_loops): Make static.
3010 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3011 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
3012 (tree_complete_unroll, gate_tree_complete_unroll,
3013 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
3014 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3015 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
3016 * tree-ssa-loop-ivopts.c: Remove local prototypes.
3017 (stmt_invariant_in_loop_p): Remove unused function.
3018 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
3019 * tree-ssa-loop-manip.h: New file. Add prototypes.
3020 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
3021 (gcov_type_to_double_int): Move to cfgloop.h.
3022 (double_int_cmp, bound_index,
3023 estimate_numbers_of_iterations_loop): Make static.
3024 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
3025 (max_loop_iterations): Factor out get_max_loop_iterations.
3026 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
3028 * tree-ssa-loop-niter.h: New file. Add prototypes.
3029 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
3030 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
3031 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
3032 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
3033 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
3034 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
3035 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
3036 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
3037 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
3038 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
3039 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3040 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
3041 pass_data_complete_unroll, make_pass_complete_unroll,
3042 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3043 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
3044 tree-ssa-loop-ivcanon.c.
3045 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
3046 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
3047 tree-ssa-loop-prefetch.c.
3048 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
3050 (get_lsm_tmp_name): Relocate and add suffix parameter.
3051 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
3052 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
3053 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
3054 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
3055 (get_estimated_loop_iterations): Factor out accessor from
3056 estimated_loop_iterations in tree-ssa-loop-niter.c.
3057 (get_max_loop_iterations): Factor out accessor from
3058 _max_loop_iterations in tree-ssa-niter.c.
3059 * loop-unroll.c (decide_unroll_constant_iterations,
3060 decide_unroll_runtime_iterations, decide_peel_simple,
3061 decide_unroll_stupid): Use new get_* accessors.
3063 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
3065 PR tree-optimization/20318
3066 * doc/extend.texi (returns_nonnull): New function attribute.
3067 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
3069 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
3070 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
3072 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
3075 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
3076 false if both components are bitfields.
3078 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3080 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
3081 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
3082 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
3083 * config/aarch64/aarch64.h
3084 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
3085 * config/aarch64/aarch64-simd-builtins.def
3086 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
3088 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3090 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
3092 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3094 * config/aarch64/arm_neon.h (vdiv_f64): Added.
3096 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3098 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
3099 (vneg_f64): New intrinsic.
3100 (vneg_s8): Asm replaced with C.
3101 (vneg_s16): Likewise.
3102 (vneg_s32): Likewise.
3103 (vneg_s64): New intrinsic.
3104 (vnegq_f32): Asm replaced with C.
3105 (vnegq_f64): Likewise.
3106 (vnegq_s8): Likewise.
3107 (vnegq_s16): Likewise.
3108 (vnegq_s32): Likewise.
3109 (vnegq_s64): Likewise.
3111 2013-10-09 Renlin Li <Renlin.Li@arm.com>
3113 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
3115 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3117 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
3118 packed stack special handling.
3119 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
3120 back to fixed stack slots for FPRs saved due to stdarg.
3122 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3124 * config/s390/s390.c (s390_frame_info): Restructure function.
3126 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3128 * config/s390/s390.c (struct s390_frame_layout): New field
3130 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
3131 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
3132 regs_ever_clobbered to char*.
3133 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
3134 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
3135 (s390_register_info): Enable FPR save slots. Move/Copy some
3136 functionality into ...
3137 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
3138 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
3140 (s390_frame_info): Do gpr slot allocation here now. stdarg does
3141 not imply a stack frame.
3142 (s390_init_frame_layout): Remove variable clobbered_regs.
3143 (s390_update_register_info): Remove function.
3144 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
3145 cannot be used for register renaming.
3146 (s390_hard_regno_scratch_ok): New function.
3147 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
3148 (s390_initial_elimination_offset): Change offset calculation of
3149 the return address pointer.
3150 (save_gprs): Deal with only r6 being saved from the call-saved regs.
3151 (restore_gprs): Set frame related flag.
3152 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
3153 (s390_emit_prologue): Call s390_register_info instead of
3154 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
3155 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
3156 (s390_optimize_prologue): Call s390_optimize_register_info.
3157 Try to remove also FPR slot save/restore INSNs. Remove frame
3158 related flags from restore INSNs.
3160 2013-10-08 DJ Delorie <dj@redhat.com>
3162 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
3165 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
3166 avoid conflict with the MI use of %c.
3167 * config/rl78/rl78-real.md: change %c to %C throughout.
3168 * config/rl78/rl78-virt.md: Likewise.
3170 2013-10-08 Jan Hubicka <jh@suse.cz>
3172 * config/i386/i386.c (ix86_option_override_internal): Switch
3173 to SSE math for -ffast-math when target ISA supports SSE2.
3175 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3177 * tree-flow.h: Remove some prototypes.
3178 * tree.h: Remove some protypes, add a couple.
3179 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3180 using_eh_for_cleanups_p): Add interface routines for front ends.
3181 * tree-eh.h: New file. Add protoptyes.
3182 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3183 (add_stmt_to_eh_lp_fn): Make static.
3184 (lower_try_finally): Use new using_eh_for_cleanups_p.
3185 * emit-rtl.c: Include tree-eh.h.
3186 * gimple.h: Include tree-eh.h.
3188 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3190 PR tree-optimization/58480
3191 * tree-vrp.c (infer_nonnull_range): New function.
3192 (infer_value_range): Call infer_nonnull_range.
3194 2013-10-08 Dehao Chen <dehao@google.com>
3196 PR tree-optimization/58619
3197 * tree-inline.c (copy_phis_for_bb): Combine location data
3200 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3203 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3204 RTX_FRAME_RELATED_P on INSN.
3206 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3208 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3209 (altivec_expand_vec_perm_const): Call it.
3211 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3213 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3214 sequences for LE VSX loads and stores at expand time.
3215 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3217 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3218 (rs6000_gen_le_vsx_permute): New.
3219 (rs6000_gen_le_vsx_load): New.
3220 (rs6000_gen_le_vsx_store): New.
3221 (rs6000_gen_le_vsx_move): New.
3222 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3223 (*vsx_le_perm_load_v4si): New.
3224 (*vsx_le_perm_load_v8hi): New.
3225 (*vsx_le_perm_load_v16qi): New.
3226 (*vsx_le_perm_store_v2di): New.
3227 (*vsx_le_perm_store_v4si): New.
3228 (*vsx_le_perm_store_v8hi): New.
3229 (*vsx_le_perm_store_v16qi): New.
3230 (*vsx_xxpermdi2_le_<mode>): New.
3231 (*vsx_xxpermdi4_le_<mode>): New.
3232 (*vsx_xxpermdi8_le_V8HI): New.
3233 (*vsx_xxpermdi16_le_V16QI): New.
3234 (*vsx_lxvd2x2_le_<mode>): New.
3235 (*vsx_lxvd2x4_le_<mode>): New.
3236 (*vsx_lxvd2x8_le_V8HI): New.
3237 (*vsx_lxvd2x16_le_V16QI): New.
3238 (*vsx_stxvd2x2_le_<mode>): New.
3239 (*vsx_stxvd2x4_le_<mode>): New.
3240 (*vsx_stxvd2x8_le_V8HI): New.
3241 (*vsx_stxvd2x16_le_V16QI): New.
3243 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3245 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3247 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3249 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3250 loop to work also for 31bit ABI.
3251 Save the stack pointer for frame_size > 0.
3253 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3255 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3256 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3257 constraint letters from expanders.
3258 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3259 retry count to general_operand.
3260 ("tabort"): Give operand 0 a mode.
3261 ("tabort_1"): Add mode and constraint letter for operand 0.
3262 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3264 2013-10-04 Jeff Law <law@redhat.com>
3266 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3268 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3270 (thread_across_edge): Here. Call it.
3272 2013-10-04 Cary Coutant <ccoutant@google.com>
3274 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3275 discarding a location list expression (or a piece of one).
3277 2013-10-03 Jan Hubicka <jh@suse.cz>
3279 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3280 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3281 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3283 2013-10-03 Jan Hubicka <jh@suse.cz>
3285 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3286 accumulate-outgoing-args when producing unwind info.
3288 2013-10-03 Wei Mi <wmi@google.com>
3290 * lra-constraints.c (insert_move_for_subreg): New function
3291 extracted from simplify_operand_subreg.
3292 (simplify_operand_subreg): Add reload for paradoxical subreg.
3294 2013-10-03 Rong Xu <xur@google.com>
3296 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3297 the candidate of builtin_expect such that we should fix the
3298 size/time estimation.
3299 (estimate_function_body_sizes): Do the acutally size/time fix-up
3302 2013-10-03 Rong Xu <xur@google.com>
3304 * predict.c (tree_predict_by_opcode): Get the probability
3305 for builtin_expect from param builtin_expect_probability.
3306 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3307 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3308 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3310 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3313 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3314 * calls.c (alloca_call_p): Use get_callee_fndecl.
3315 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3316 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3318 (vrp_visit_stmt): Remove duplicated code.
3320 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3322 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3323 ceildf2, btruncdf2, instead of vsx_* name.
3325 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3326 iterators to only do V2DF and V4SF here. Move the DF code to
3327 rs6000.md where it is combined with SF mode. Replace <VSv> with
3328 just 'v' since only vector operations are handled with these insns
3329 after moving the DF support to rs6000.md.
3330 (vsx_sub<mode>3): Likewise.
3331 (vsx_mul<mode>3): Likewise.
3332 (vsx_div<mode>3): Likewise.
3333 (vsx_fre<mode>2): Likewise.
3334 (vsx_neg<mode>2): Likewise.
3335 (vsx_abs<mode>2): Likewise.
3336 (vsx_nabs<mode>2): Likewise.
3337 (vsx_smax<mode>3): Likewise.
3338 (vsx_smin<mode>3): Likewise.
3339 (vsx_sqrt<mode>2): Likewise.
3340 (vsx_rsqrte<mode>2): Likewise.
3341 (vsx_fms<mode>4): Likewise.
3342 (vsx_nfma<mode>4): Likewise.
3343 (vsx_copysign<mode>3): Likewise.
3344 (vsx_btrunc<mode>2): Likewise.
3345 (vsx_floor<mode>2): Likewise.
3346 (vsx_ceil<mode>2): Likewise.
3347 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3348 (vsx_sminsf3): Likewise.
3349 (vsx_fmadf4): Likewise.
3350 (vsx_fmsdf4): Likewise.
3351 (vsx_nfmadf4): Likewise.
3352 (vsx_nfmsdf4): Likewise.
3353 (vsx_cmpdf_internal1): Likewise.
3355 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3356 simpler to select whether a target has SPE or traditional floating
3357 point support in iterators.
3358 (TARGET_DF_SPE): Likewise.
3359 (TARGET_SF_FPR): Likewise.
3360 (TARGET_DF_FPR): Likewise.
3361 (TARGET_SF_INSN): Macros to say whether floating point support
3362 exists for a given operation for expanders.
3363 (TARGET_DF_INSN): Likewise.
3365 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3366 combining of SF/DF mode operations, using both traditional and VSX
3374 (abs<mode>2): Combine SF/DF modes using traditional floating point
3375 instructions. Add support for using the upper DF registers with
3376 VSX support, and SF registers with power8-vector support. Update
3377 expanders for operations supported by both the SPE and traditional
3378 floating point units.
3379 (abs<mode>2_fpr): Likewise.
3380 (nabs<mode>2): Likewise.
3381 (nabs<mode>2_fpr): Likewise.
3382 (neg<mode>2): Likewise.
3383 (neg<mode>2_fpr): Likewise.
3384 (add<mode>3): Likewise.
3385 (add<mode>3_fpr): Likewise.
3386 (sub<mode>3): Likewise.
3387 (sub<mode>3_fpr): Likewise.
3388 (mul<mode>3): Likewise.
3389 (mul<mode>3_fpr): Likewise.
3390 (div<mode>3): Likewise.
3391 (div<mode>3_fpr): Likewise.
3392 (sqrt<mode>3): Likewise.
3393 (sqrt<mode>3_fpr): Likewise.
3394 (fre<Fs>): Likewise.
3395 (rsqrt<mode>2): Likewise.
3396 (cmp<mode>_fpr): Likewise.
3397 (smax<mode>3): Likewise.
3398 (smin<mode>3): Likewise.
3399 (smax<mode>3_vsx): Likewise.
3400 (smin<mode>3_vsx): Likewise.
3401 (negsf2): Delete SF operations that are merged with DF.
3408 (fmasf4_fpr): Likewise.
3409 (fmssf4_fpr): Likewise.
3410 (nfmasf4_fpr): Likewise.
3411 (nfmssf4_fpr): Likewise.
3412 (sqrtsf2): Likewise.
3413 (rsqrtsf_internal1): Likewise.
3414 (smaxsf3): Likewise.
3415 (sminsf3): Likewise.
3416 (cmpsf_internal1): Likewise.
3417 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3418 (negdf2): Delete DF operations that are merged with SF.
3420 (nabsdf2): Likewise.
3426 (rsqrtdf_internal1): Likewise.
3427 (fmadf4_fpr): Likewise.
3428 (fmsdf4_fpr): Likewise.
3429 (nfmadf4_fpr): Likewise.
3430 (nfmsdf4_fpr): Likewise.
3431 (sqrtdf2): Likewise.
3432 (smaxdf3): Likewise.
3433 (smindf3): Likewise.
3434 (cmpdf_internal1): Likewise.
3435 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3436 (btrunc<mode>2): Delete separate expander, and combine with the
3437 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3438 (btrunc<mode>2_fpr): Likewise.
3439 (ceil<mode>2): Likewise.
3440 (ceil<mode>2_fpr): Likewise.
3441 (floor<mode>2): Likewise.
3442 (floor<mode>2_fpr): Likewise.
3443 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3444 Add support for using the upper registers with VSX and
3445 power8-vector. Move insns to be closer to the define_expands. On
3446 VSX systems, prefer the traditional form of FMA over the VSX
3447 version, since the traditional form allows the target not to
3448 overlap with the inputs.
3449 (fms<mode>4_fpr): Likewise.
3450 (nfma<mode>4_fpr): Likewise.
3451 (nfms<mode>4_fpr): Likewise.
3453 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3454 Richard Earnshaw <richard.earnshaw@arm.com>
3456 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3457 (struct mult_cost_table): Likewise.
3458 (struct mem_cost_table): Likewise.
3459 (struct fp_cost_table): Likewise.
3460 (struct vector_cost_table): Likewise.
3461 (cpu_cost_table): Likewise.
3462 * config/arm/arm.opt (mold-rts-costs): New option.
3463 (mnew-generic-costs): Likewise.
3464 * config/arm/arm.c (generic_extra_costs): New table.
3465 (cortexa15_extra_costs): Likewise.
3466 (arm_slowmul_tune): Use NULL as new costs.
3467 (arm_fastmul_tune): Likewise.
3468 (arm_strongarm_tune): Likewise.
3469 (arm_xscale_tune): Likewise.
3470 (arm_9e_tune): Likewise.
3471 (arm_v6t2_tune): Likewise.
3472 (arm_cortex_a5_tune): Likewise.
3473 (arm_cortex_a9_tune): Likewise.
3474 (arm_v6m_tune): Likewise.
3475 (arm_fa726te_tune): Likewise.
3476 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3477 (arm_cortex_tune): Use generict_extra_costs.
3478 (shifter_op_p): New function.
3479 (arm_unspec_cost): Likewise.
3480 (LIBCALL_COST): Define.
3481 (arm_new_rtx_costs): New function.
3482 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3483 table is available. Use old costs otherwise unless mnew-generic-costs
3485 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3486 (cpu_cost_table): Declare.
3488 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3491 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3492 (*subs_mul_imm_<mode>)
3493 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3494 (*sub_<shift>_<mode>)
3495 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3496 Remove k constraint.
3498 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3500 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3502 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3504 2013-10-02 Teresa Johnson <tejohnson@google.com>
3506 * predict.c (probably_never_executed): New function.
3507 (probably_never_executed_bb_p): Invoke probably_never_executed.
3508 (probably_never_executed_edge_p): Ditto.
3509 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3510 Treat profile insanities conservatively.
3512 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3514 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3516 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3518 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3519 use it. Use smaller increase for scratch. Don't increase reject
3520 for early clobber scratch.
3521 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3522 setting eliminated regs except setting fp from hfp.
3523 (lra_eliminate): Check lra_insn_recog_data on NULL.
3525 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3528 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3529 setting -mvsx-timode by default until the underlying problem is fixed.
3530 (RS6000_CPU, power7 defaults): Likewise.
3532 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3534 * config/x-linux (host-linux.o): Remove header dependencies.
3535 Use $(COMPILE) and $(POSTCOMPILE).
3536 * config/t-linux-android (linux-android.o): Ditto.
3538 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3540 * Makefile.in (expmed.o-warn): Remove.
3542 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3544 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3545 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3547 2013-10-02 Teresa Johnson <tejohnson@google.com>
3549 * dojump.c (do_jump_1): Divide probability between
3550 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3552 2013-10-02 Tom Tromey <tromey@redhat.com>
3554 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3556 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3558 * tree-flow.h: Remove some prototypes.
3559 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3560 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3561 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3562 mark_virtual_phi_result_for_renaming): Relocate here.
3563 * tree-into-ssa.h: Add prototypes.
3564 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3565 single_pred_before_succ_order.
3566 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3567 (nonfreeing_call_p) Move to gimple.c.
3568 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3570 * basic-block.h (single_pred_before_succ_order): Add prototype.
3571 * gimple.c (nonfreeing_call_p): Relocate here.
3572 * gimple.h: Add prototype.
3573 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3574 * tree-ssa-dom.h: New file. Relocate prototypes here.
3575 * tree-ssa.h: Include tree-ssa-dom.h.
3577 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3579 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3580 Use $(COMPILE) and $(POSTCOMPILE).
3582 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3584 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3586 * tree-flow.h: Remove some prototypes.
3587 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3588 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3589 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3590 propagate_tree_value*): Move from here to...
3591 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3592 propagate_tree_value*): Relocate here.
3593 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3594 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3595 * gimple-fold.c: Remove gimple-fold.h from include list.
3596 * tree-vrp.c: Remove gimple-fold.h from include list.
3597 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3598 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3599 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3600 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3601 * sese.c: Add tree-ssa-propagate.h to include list.
3603 2013-10-02 Richard Biener <rguenther@suse.de>
3605 * tree-loop-distribution.c: Include tree-vectorizer.h for
3607 (enum partition_kind): Remove PKIND_REDUCTION.
3608 (struct partition_s): Remove has_writes member, add reduction_p member.
3609 (partition_alloc): Adjust.
3610 (partition_builtin_p): Likewise.
3611 (partition_has_writes): Remove.
3612 (partition_reduction_p): New function.
3613 (partition_merge_into): Likewise.
3614 (generate_code_for_partition): Commonize builtin partition
3616 (rdg_cannot_recompute_vertex_p): Remove.
3617 (already_processed_vertex_p): Likewise.
3618 (rdg_flag_vertex): Do not set has_writes.
3619 (classify_partition): Adjust.
3620 (rdg_build_partitions): Do not set has_writes, treat all
3621 partitions as useful.
3622 (distribute_loop): Record number of library calls generated. Adjust.
3623 (tree_loop_distribution): Report number of loops and library
3624 calls generated as opt-info.
3626 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3628 * tree-flow.h: Include new .h files. Move prototypes.
3629 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3630 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3631 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3632 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3633 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3634 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3635 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3636 * tree.h (get_ref_base_and_extent): Move prototype out.
3637 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3639 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3640 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3641 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3642 * tree-scalar-evolution.c: Include tree.h.
3643 * sese.c: Include tree.h.
3644 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3645 * dwarf2out.c: Include tree-dfa.h.
3646 * tree-chrec.c: Include tree.h.
3647 * tree-data-ref.c: Include tree.h.
3649 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3651 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3654 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3656 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3657 Use $(COMPILE) and $(POSTCOMPILE).
3658 (sol2-cxx.o): Likewise.
3659 (sol2-stubs.o): Likewise.
3661 * config/x-solaris (host-solaris.o): Likewise.
3663 * config/sparc/t-sparc (sparc.o): Remove.
3664 (sparc-c.o): Remove header dependencies.
3665 Use $(COMPILE) and $(POSTCOMPILE).
3666 * config/sparc/x-sparc: Likewise.
3668 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3670 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3671 * config/arc/arc700.md: Likewise.
3672 * config/arc/arc-modes.def: Likewise.
3673 * config/arc/arc-simd.h: Likewise.
3674 * config/arc/t-arc-uClibc: Likewise.
3675 * config/arc/t-arc-newlib: Likewise.
3677 2013-10-02 Renlin Li <renlin.li@arm.com>
3679 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3681 (aarch64_expand_epilogue): Likewise.
3683 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3684 Yufeng Zhang <yufeng.zhang@arm.com>
3686 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3688 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3689 'base_in' represent a conversion and legal_cast_p_1 holds; set
3690 'base_in' with the returned value from get_unwidened.
3692 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3694 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3695 plus_constant is not used.
3697 2013-10-01 Wei Mi <wmi@google.com>
3699 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3700 * config/i386/i386.md: Add define_peephole2 to
3701 break partial reg stall for cvtss2sd/cvtsd2ss.
3703 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3705 * config/arc/arc.c (pass_arc_ifcvt::clone):
3706 Update for ctxt_ -> m_ctxt change.
3708 2013-10-01 Jeff Law <law@redhat.com>
3710 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3711 outgoing_edge and intermediate_edge fields. Instead store the path.
3712 (redirection_data::hash): Hash on the last edge's destination index.
3713 (redirection_data::equal): Check the entire thread path.
3714 (lookup_redirectio_data): Corresponding changes.
3715 (create_edge_and_update_destination_phis): Likewise.
3716 (thread_single_edge): Likewise.
3718 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3719 Diego Novillo <dnovillo@google.com>
3721 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3722 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3723 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3724 versions of these patterns.
3726 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3727 (__builtin_arc_aligned): Likewise.
3729 * config/arc/arc.md: Expand adc_0 comment stating the intended
3730 purpose and why it isn't ready.
3731 Replace commented out call_value_via_label_mixed with a
3732 plain comment about bl_s.
3734 * config/arc/arc.c (stdio.h): Don't include directly.
3735 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3736 Remove if (1) condition.
3737 (arc_encode_section_info): Fix comment.
3739 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3741 * config/arc/arc.c (arc_conditional_register_usage):
3742 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3743 Also set reg_alloc_order for DMA config regs.
3745 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3746 Jeremy Bennett <jeremy.bennett@embecosm.com>
3748 * doc/install.texi (--with-cpu): Mention ARC.
3749 (arc-*-elf32): New paragraph.
3750 (arc-linux-uclibc): Likewise.
3751 * doc/md.texi (Machine Constraints): Add ARC part.
3752 * doc/invoke.texi: (menu): Add ARC Options.
3753 (Machine Dependent Options) <ARC Options>: Add synopsis.
3754 (node ARC Options): Add.
3755 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3756 (ARC Built-in Functions): New section defining
3757 generic ARC built-in functions.
3758 (ARC SIMD Built-in Functions): New section defining SIMD specific
3760 (Declaring Attributes of Functions): Extended
3761 description of short_call and long_call attributes for ARC and
3762 added index entries.
3764 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3765 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3766 Joern Rennecke <joern.rennecke@embecosm.com>
3767 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3768 Brendan Kehoe <brendan@zen.org>
3769 Michael Eager <eager@eagercon.com>
3770 Simon Cook <simon.cook@embecosm.com>
3771 Jeremy Bennett <jeremy.bennett@embecosm.com>
3773 * config/arc, common/config/arc: New directories.
3775 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3776 Brendan Kehoe <brendan@zen.org>
3777 Simon Cook <simon.cook@embecosm.com>
3779 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3781 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3783 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3784 * tree-ssa-coalesce.h: New. Move prototype to here.
3785 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3786 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3787 (gimple_can_coalesce_p): Move to...
3788 * gimple.c (gimple_can_coalesce_p): Here.
3790 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3792 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3793 (dump_decl_set): Move to gimple.c.
3794 * gimple.h: Don't include tree-ssa-operands.h.
3795 (dump_decl_set): Add prototype.
3796 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3797 Move to gimple-ssa.h.
3798 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3799 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3800 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3801 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3802 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3804 (dump_decl_set): Relocate here.
3805 * gimple-ssa.h: New file.
3806 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3807 Relocate from gimple.h.
3808 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3809 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3810 swap_tree_operands and remove non-ssa path.
3811 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3812 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3814 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3815 vect_is_simple_reduction_1): Use swap_ssa_operands.
3816 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3817 (enum need_phi_state): Move to tree-into-ssa.c.
3818 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3819 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3820 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3821 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3822 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3823 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3824 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3825 num_imm_uses): Move to ssa-iterators.h.
3826 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3827 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3829 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3830 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3831 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3832 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3833 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3834 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3835 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3836 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3837 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3838 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
3839 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
3840 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
3841 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
3842 (set_phi_nodes): Move to tree-phinodes.h.
3843 * tree-ssa-operands.h (enum ssa_op_iter_type,
3844 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
3845 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
3846 (dump_decl_set): Remove prototype.
3847 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
3848 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
3849 (set_phi_nodes): Relocate from tree-flow-inline.h.
3850 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
3852 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
3853 include list. Temporarily add gimple.h to include list.
3854 * ssa-iterators.h: New file.
3855 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3856 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
3857 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
3858 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
3859 Relocate from tree-ssa-operands.h.
3860 (delink_imm_use, link_imm_use_to_list, link_imm_use,
3861 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3862 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3863 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3864 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
3865 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
3866 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
3867 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
3868 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
3869 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
3870 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
3871 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
3872 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
3873 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
3874 Relocate from tree-flow-inline.h.
3875 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
3877 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
3880 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
3881 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
3882 Insert '\t' to output template.
3883 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
3884 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
3885 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
3886 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
3888 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
3890 * doc/install.texi (Host/target specific installation notes for GCC):
3891 Put @anchor before @heading.
3892 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
3893 Use @email for email addresses.
3895 2013-10-01 Jeff Law <law@redhat.com>
3897 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
3898 a vec. Only delete the path if we create one without successfully
3899 registering a jump thread.
3900 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
3902 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
3903 (paths): New vector of jump threading paths.
3904 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
3905 (THREAD_PATH): New accessor macro for the entire thread path.
3906 (lookup_redirection_data): Get intermediate and final outgoing edge
3907 from the thread path.
3908 (create_edge_and_update_destination_phis): Copy the threading path.
3909 (ssa_fix_duplicate_block_edges): Get edges and block types from the
3910 jump threading path.
3911 (ssa_redirect_edges): Get edges and block types from the jump threading
3912 path. Free the path vector.
3913 (thread_block): Get edges from the jump threading path. Look at the
3914 entire path to see if we thread to a loop exit. If we cancel a jump
3915 thread request, then free the path vector.
3916 (thread_single_edge): Get edges and block types from the jump threading
3917 path. Free the path vector.
3918 (thread_through_loop_header): Get edges and block types from the jump
3919 threading path. Free the path vector.
3920 (mark_threaded_blocks): Iterate over the vector of paths and store
3921 the path on the appropriate edge. Get edges and block types from the
3922 jump threading path.
3923 (mark_threaded_blocks): Get edges and block types from the jump
3924 threading path. Free the path vector.
3925 (thread_through_all_blocks): Use the vector of paths rather than
3926 a vector of 3-edge sets.
3927 (register_jump_thread): Accept pointer to a path vector rather
3928 than the path vector itself. Store the path vector for later use.
3931 2013-10-01 Jakub Jelinek <jakub@redhat.com>
3932 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3935 * config/s390/s390.c (s390_split_branches): Modify check for table
3937 (s390_chunkify_start): Rearrange table jump insn check in order to
3938 deal with compare and branch insns correctly.
3940 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
3944 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
3945 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
3946 define_insn_and_split.
3947 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
3948 (shiftsi3_compare): New pattern.
3950 * config/arm/unspecs.md (UNSPEC_RRX): New.
3952 2013-10-01 Alan Modra <amodra@gmail.com>
3954 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
3955 casing inout operands.
3957 2013-10-01 Richard Biener <rguenther@suse.de>
3959 PR tree-optimization/58553
3960 * tree-loop-distribution.c (struct partition_s): Add niter member.
3961 (classify_partition): Populate niter member for the partition
3962 and properly identify whether the relevant store happens before
3963 or after the loop exit.
3964 (generate_memset_builtin): Use niter member from the partition.
3965 (generate_memcpy_builtin): Likewise.
3967 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3969 * vec.h (vec_prefix, vec): Prefix member names with "m_".
3970 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
3972 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
3974 * basic-block.h (edge_list): Prefix member names with "m_".
3975 * context.h (context): Likewise.
3976 * domwalk.h (dom_walker): Likewise.
3977 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
3978 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
3979 * hash-table.h (hash_table): Likewise.
3980 * machmode.h (bit_field_mode_iterator): Likewise.
3981 * pass_manager.h (pass_list): Likewise.
3982 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
3983 * tree-pass.h (pass_data): Likewise.
3984 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
3985 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
3986 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
3987 * asan.c (pass_data_asan): Update accordingly.
3988 * cfganal.c (control_dependences::find_control_dependence): Likewise.
3989 (control_dependences::control_dependences): Likewise.
3990 (control_dependences::~control_dependences): Likewise.
3991 (control_dependences::~control_dependences): Likewise.
3992 (control_dependences::get_edges_dependent_on): Likewise.
3993 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
3994 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
3995 * context.c (gcc::context::context): Likewise.
3996 * cprop.c (pass_rtl_cprop::clone): Likewise.
3997 * domwalk.c (dom_walker::walk): Likewise.
3998 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
3999 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
4000 * mode-switching.c (pass_mode_switching::clone): Likewise.
4001 * passes.c (opt_pass::opt_pass): Likewise.
4002 (pass_manager::pass_manager): Likewise.
4003 * predict.c (pass_strip_predict_hints::clone): Likewise.
4004 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
4005 (pass_split_all_insns::clone): Likewise.
4006 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
4008 (bit_field_mode_iterator::next_mode): Likewise.
4009 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
4010 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
4011 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
4012 * tree-complex.c (pass_lower_complex::clone): Likewise.
4013 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
4014 * tree-object-size.c (pass_object_sizes::clone): Likewise.
4015 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
4016 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
4017 (pass_fold_builtins::clone): Likewise.
4018 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
4019 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
4020 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
4021 (pass_cd_dce::clone): Likewise.
4022 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
4023 (pass_phi_only_cprop::clone): Likewise.
4024 * tree-ssa-dse.c (pass_dse::clone): Likewise.
4025 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
4026 * tree-ssa-loop.c (pass_lim::clone): Likewise.
4027 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
4028 * tree-ssa-pre.c (pass_fre::clone): Likewise.
4029 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
4030 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
4031 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
4032 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
4033 * tree-vrp.c (pass_vrp::clone): Likewise.
4034 * tsan.c (pass_tsan::clone): Likewise.
4036 2013-09-30 Jakub Jelinek <jakub@redhat.com>
4039 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
4040 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
4043 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
4044 optimization, punt if sign_bit_p looked through any zero extension.
4046 2013-09-30 Teresa Johnson <tejohnson@google.com>
4048 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
4049 Update redirected out edge count in joiner case.
4050 (ssa_redirect_edges): Common the joiner and non-joiner cases
4051 so that joiner case gets profile updates.
4053 2013-09-30 Richard Biener <rguenther@suse.de>
4055 PR tree-optimization/58554
4056 * tree-loop-distribution.c (classify_partition): Require
4057 unconditionally executed stores for memcpy and memset recognition.
4058 (tree_loop_distribution): Calculate dominance info.
4060 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4062 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
4063 (NO_PROFILE_COUNTERS): Likewise.
4064 (PROFILE_HOOK): Likewise.
4065 (FUNCTION_PROFILER): Likewise.
4066 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
4068 2013-09-30 Iain Sandoe <iain@codesourcery.com>
4070 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
4071 calls and defines in TARGET_MACHO conditional.
4072 (load_macho_picbase_di): Likewise.
4073 (reload_macho_picbase): Likewise.
4074 (reload_macho_picbase_si): Likewise.
4075 (reload_macho_picbase_di): Likewise.
4076 (nonlocal_goto_receiver): Likewise.
4078 2013-09-30 Nick Clifton <nickc@redhat.com>
4080 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
4081 that use the MSP430X ISA.
4082 (msp430_option_override): Scan -mmcu command line option for any
4083 MCU name that supports the MSP430X ISA.
4084 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
4085 -mmcu options which enable the MSP430X ISA.
4087 2013-09-30 Richard Biener <rguenther@suse.de>
4090 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
4091 before looking for setjmp-like calls.
4093 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4096 * config/darwin-protos.h (machopic_get_function_picbase): New.
4097 * config/darwin.c (machopic_get_function_picbase): New.
4098 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
4099 label for a new func. (load_macho_picbase_di): Likewise.
4100 (reload_macho_picbase): New expand.
4101 (reload_macho_picbase_si): New insn.
4102 (reload_macho_picbase_di): New insn.
4103 (nonlocal_goto_receiver): New define and split.
4104 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
4105 (unspecv enum): Add UNSPECV_NLGR.
4107 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4109 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
4110 that altivec registers are correctly sized on Darwin.
4112 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4114 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
4115 darwin-driver.o): Use COMPILE and POSTCOMPILE.
4116 * config/x-darwin (host-darwin.o): Likewise.
4117 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
4118 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
4119 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
4121 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
4123 * doc/invoke.texi: Fix usage of @tie{} command.
4125 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
4127 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
4129 2013-09-28 Jan Hubicka <jh@suse.cz>
4131 * config/i386/x86-tune.def: Add documentation for each of the options;
4134 2013-09-28 Jan Hubicka <jh@suse.cz>
4136 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
4138 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
4139 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
4140 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
4141 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
4142 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
4144 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
4146 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
4147 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
4148 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
4149 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
4150 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
4151 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
4152 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
4153 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
4154 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
4155 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
4156 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
4157 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
4158 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
4159 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
4160 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
4161 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
4162 gimple.h, godump.c, graphite-clast-to-gimple.c,
4163 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
4164 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
4165 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
4166 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4167 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
4168 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
4169 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
4170 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
4171 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
4172 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
4173 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
4174 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
4175 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4176 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4177 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4178 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4179 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4180 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4181 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4182 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4183 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4184 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4185 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4186 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4187 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4188 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4189 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4190 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4191 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4192 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4193 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4194 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4195 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4196 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4197 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4198 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4199 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4200 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4201 whitespace before "(".
4203 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4205 * expr.h (extract_bit_field): Remove packedp parameter.
4206 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4207 from forward declaration.
4208 (store_split_bit_field): Remove packedp arg from calls to
4209 extract_fixed_bit_field.
4210 (extract_bit_field_1): Remove packedp parameter and packedp
4211 argument from recursive calls and calls to extract_fixed_bit_field.
4212 (extract_bit_field): Remove packedp parameter and corresponding
4213 arg to extract_bit_field_1.
4214 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4216 (extract_split_bit_field): Remove packedp arg from call to
4217 extract_fixed_bit_field.
4218 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4219 (copy_blkmode_from_reg): Likewise.
4220 (copy_blkmode_to_reg): Likewise.
4221 (read_complex_part): Likewise.
4222 (store_field): Likewise.
4223 (expand_expr_real_1): Likewise.
4224 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4225 to extract_bit_field.
4226 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4227 call to extract_bit_field.
4228 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4229 call to extract_bit_field.
4230 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4231 and special packedp behavior from -fstrict-volatile-bitfields
4234 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4236 * lra-eliminations.c (init_elim_table): Guard value_p.
4238 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4240 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4241 DFmode, DImode, and SFmode in the upper VSX registers based on the
4242 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4243 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4244 of the VSX register handling.
4246 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4248 (zero_extendsidi2_lfiwzx): Add support for loading into the
4249 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4250 only do VSX memory options on Altivec registers.
4251 (extendsidi2_lfiwax): Likewise.
4252 (extendsfdf2_fpr): Likewise.
4253 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4254 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4255 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4256 (movdi_internal64): Likewise.
4258 2013-09-27 Xinliang David Li <davidxl@google.com>
4260 * opts.c (finish_options): Adjust parameters
4261 according to vect cost model.
4262 (common_handle_option): Set dynamic vect cost
4264 targhooks.c (default_add_stmt_cost): Compute stmt cost
4266 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4267 Use helper function.
4268 * tree-vectorizer.h (unlimited_cost_model): New function.
4269 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4270 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4272 (vect_enhance_data_refs_alignment): Ditto.
4273 * flag-types.h: New enum.
4274 * common/config/i386/i386-common.c (ix86_option_init_struct):
4275 No need to initialize vect_cost_model flag.
4276 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4279 2013-09-27 Diego Novillo <dnovillo@google.com>
4281 * gimple.h (enum ssa_mode): Remove.
4283 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4285 * cfgloop.h (number_of_loops): Fix typo in check for null.
4287 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4290 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4291 are children of outermost saved_cfun's loop, and set it up to
4292 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4293 if loop != loop0 and SESE region contains bbs that belong to loop0.
4295 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4297 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4298 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4299 (set_address_segment, set_address_base, set_address_index)
4300 (set_address_disp): Accept the argument unconditionally.
4301 (baseness): Remove must_be_base_p and must_be_index_p checks.
4302 (decompose_normal_address): Classify as much as possible in the
4305 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4307 * cse.c (count_reg_usage): Handle INT_LIST.
4308 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4309 * reginfo.c (reg_scan_mark_refs): Likewise.
4310 * reload1.c (eliminate_regs_1): Likewise.
4312 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4315 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4318 2013-09-27 Richard Biener <rguenther@suse.de>
4320 PR tree-optimization/58459
4321 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4322 restriction not propagating into loops.
4324 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4326 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4327 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4328 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4330 2013-09-26 Richard Biener <rguenther@suse.de>
4332 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4334 2013-09-26 Richard Biener <rguenther@suse.de>
4336 * alias.h (component_uses_parent_alias_set): Rename to ...
4337 (component_uses_parent_alias_set_from): ... this.
4338 * alias.c (component_uses_parent_alias_set): Rename to ...
4339 (component_uses_parent_alias_set_from): ... this and return
4341 (reference_alias_ptr_type_1): Use the result from
4342 component_uses_parent_alias_set_from instead of stripping
4343 components one at a time.
4344 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4346 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4348 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4349 Move prototypes to...
4350 * tree-ssa-ter.h: New File. Move prototypes here.
4351 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4352 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4353 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4354 from is_replaceable_p.
4355 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4356 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4357 (process_replaceable): Use ter_is_replaceable_p.
4358 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4359 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4360 newly refactored ssa_is_replaceable_p.
4361 * cfgexpand.c: Include tree-outof-ssa.h.
4362 * ssaexpand.h: Delete.
4364 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4366 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4367 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4369 (create_gimple_tmp): Delete.
4370 (get_expr_type, build_assign, build_type_cast): Move to...
4371 * gimple-builder.c: New File.
4372 (get_expr_type): Relocate from gimple.c.
4373 (build_assign, build_type_cast): Change to only create ssanames.
4374 * gimple.h: Move prototypes to...
4375 * gimple-builder.h: New File. Here.
4376 * tree-ssa.h: And here.
4377 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4378 count_uses_and_derefs): Relocate from gimple.c.
4379 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4380 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4381 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4382 gimple_replace_ssa_lhs.
4383 * asan.c: Include gimple-builder.h.
4384 * Makefile.in: Add gimple-builder.o.
4386 2013-09-26 Richard Biener <rguenther@suse.de>
4388 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4389 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4390 (loe_visit_block): Use gcc_checking_assert.
4391 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4392 gimple_assign_ssa_name_copy_p.
4393 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4396 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4398 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4399 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4401 2013-09-26 Richard Biener <rguenther@suse.de>
4403 PR tree-optimization/58539
4404 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4405 the fact that debug statements are not taking part in loop-closed
4408 2013-09-26 Nick Clifton <nickc@redhat.com>
4410 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4411 time warning message.
4412 (msp430_print_operand_raw): Delete unused letter parameter.
4413 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4414 (msp430_print_operand_address): New function.
4415 (msp430_print_operand): Move address printing code from here to
4417 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4419 (zero_extendpsisi2): Likewise.
4420 (extendpsisi2): New pattern.
4421 (andneghi3): New pattern.
4423 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4425 * config/aarch64/aarch64.opt (mlra): New option.
4426 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4427 (TARGET_LRA_P): Define.
4429 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4431 * expr.c (expand_assignment): Remove obsolete comment.
4433 2013-09-25 Jeff Law <law@redhat.com>
4435 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4436 tree-ssa-threadupdate.h.
4437 (register_jump_thread): Similarly.
4438 * tree-ssa-threadupdate.h: New header file.
4439 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4440 * tree-vrp.c: Likewise.
4441 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4442 (thread_around_empty_blocks): Change type of path vector argument to
4443 an edge,type pair from just an edge. Initialize both elements when
4444 appending to a jump threading path. Tweak references to elements
4446 (thread_across_edge): Similarly. Release memory for the elements
4448 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4449 (dump_jump_thread_path): New function broken out from
4450 register_jump_thread.
4451 (register_jump_thread): Use dump_jump_thread_path. Change type of
4452 path vector entries. Search the path for NULL edges and dump
4453 the path if one is found. Tweak the conversion of path to 3-edge
4454 form to use the block copy type information embedded in the path.
4456 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4458 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4460 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4461 Vladimir Makarov <vmakarov@redhat.com>
4463 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4464 from the least significant bit.
4465 (strip_address_mutations): Add bitfield operations handling.
4466 (must_be_index_p): Add shifting and rotate operations handling.
4467 (set_address_base): Use must_be_base_p predicate.
4468 (set_address_index): Use must_be_index_p predicate.
4470 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4471 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4472 Sergey Lega <sergey.s.lega@intel.com>
4473 Anna Tikhonova <anna.tikhonova@intel.com>
4474 Ilya Tocar <ilya.tocar@intel.com>
4475 Andrey Turetskiy <andrey.turetskiy@intel.com>
4476 Ilya Verbin <ilya.verbin@intel.com>
4477 Kirill Yukhin <kirill.yukhin@intel.com>
4478 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4480 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4482 (ix86_expand_vector_move_misalign): Support new unaligned load and
4483 stores and use new names.
4484 (CODE_FOR_sse2_storedqu): Rename to ...
4485 (CODE_FOR_sse2_storedquv16qi): ... this.
4486 (CODE_FOR_sse2_loaddqu): Rename to ...
4487 (CODE_FOR_sse2_loaddquv16qi): ... this.
4488 (CODE_FOR_avx_loaddqu256): Rename to ...
4489 (CODE_FOR_avx_loaddquv32qi): ... this.
4490 (CODE_FOR_avx_storedqu256): Rename to ...
4491 (CODE_FOR_avx_storedquv32qi): ... this.
4492 * config/i386/i386.md (fpint_logic): New.
4493 * config/i386/sse.md (VMOVE): Extend for AVX512.
4497 (VI_UNALIGNED_LOADSTORE): Ditto.
4498 (sse2_avx_avx512f): Ditto.
4499 (sse2_avx2): Extend for AVX512.
4500 (sse4_1_avx2): Ditto.
4501 (avx2_avx512f): New.
4502 (sse): Extend for AVX512.
4505 (avxsizesuffix): Ditto.
4506 (sseintvecmode): Ditto.
4508 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4509 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4510 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4511 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4512 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4513 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4514 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4515 (STORENT_MODE): Extend for AVX512.
4516 (*absneg<mode>2): Replace constraint "x" with "v".
4517 (*mul<mode>3): Ditto.
4518 (*ieee_smin<mode>3): Ditto.
4519 (*ieee_smax<mode>3): Ditto.
4520 (avx_cmp<mode>3): Replace VF with VF_128_256.
4521 (*<sse>_maskcmp<mode>3_comm): Ditto.
4522 (<sse>_maskcmp<mode>3): Ditto.
4523 (<sse>_andnot<mode>3): Extend for AVX512.
4524 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4525 (<code><mode>3, fpint_logic): New.
4526 (*<code><mode>3): Extend for AVX512.
4527 (avx512flogicsuff): New.
4528 (avx512f_<logic><mode>): Ditto.
4529 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4531 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4532 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4533 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4534 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4535 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4536 (xop_vpermil2<mode>3): Ditto.
4537 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4538 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4539 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4540 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4542 2013-09-25 Tom Tromey <tromey@redhat.com>
4544 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4545 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4546 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4547 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4548 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4549 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4550 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4551 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4552 (GRAPHITE_HTAB_H): Remove.
4554 2013-09-25 Tom Tromey <tromey@redhat.com>
4556 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4558 2013-09-25 Tom Tromey <tromey@redhat.com>
4560 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4562 2013-09-25 Tom Tromey <tromey@redhat.com>
4564 * config/i386/t-i386 (i386.o): Remove.
4565 (i386-c.o): Use COMPILE and POSTCOMPILE.
4567 2013-09-25 Tom Tromey <tromey@redhat.com>
4569 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4571 2013-09-25 Tom Tromey <tromey@redhat.com>
4573 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4574 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4575 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4576 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4577 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4578 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4579 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4580 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4581 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4582 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4583 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4584 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4585 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4586 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4587 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4588 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4589 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4590 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4591 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4592 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4593 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4594 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4595 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4596 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4597 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4598 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4599 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4600 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4601 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4602 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4603 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4604 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4605 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4606 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4607 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4608 (graphite-clast-to-gimple.o, graphite-dependences.o)
4609 (graphite-interchange.o, graphite-poly.o)
4610 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4611 (graphite-optimize-isl.o, tree-vect-loop.o)
4612 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4613 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4614 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4615 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4616 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4617 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4618 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4619 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4620 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4621 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4622 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4623 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4624 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4625 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4626 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4627 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4628 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4629 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4630 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4631 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4632 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4633 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4634 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4635 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4636 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4637 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4638 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4639 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4640 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4641 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4642 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4643 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4644 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4645 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4646 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4647 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4648 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4649 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4650 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4651 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4652 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4653 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4654 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4655 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4656 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4657 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4658 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4659 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4660 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4661 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4662 (gcov-dump.o): Remove.
4663 (default-c.o): Use COMPILE and POSTCOMPILE.
4664 (CFLAGS-gcc.o): New variable.
4665 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4667 2013-09-25 Tom Tromey <tromey@redhat.com>
4669 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4670 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4671 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4672 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4673 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4674 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4675 (c-family/array-notation-common.o, c-family/stub-objc.o)
4676 (c-family/c-ubsan.o): Remove.
4678 2013-09-25 Tom Tromey <tromey@redhat.com>
4680 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4682 2013-09-25 Tom Tromey <tromey@redhat.com>
4684 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4685 to add -DENABLE_SHARED_LIBGCC.
4686 (gcc.o): Don't use subshell.
4688 2013-09-25 Tom Tromey <tromey@redhat.com>
4690 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4691 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4692 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4693 * configure, config.in: Rebuild.
4695 2013-09-25 Tom Tromey <tromey@redhat.com>
4697 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4698 (COMPILE, POSTCOMPILE): New variables.
4699 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4700 (DEPFILES): New variable.
4701 Include ".Po" files.
4702 * configure.ac: Add checks for dependency checking.
4703 * configure, aclocal.m4: Regenerate.
4705 2013-09-25 Tom Tromey <tromey@redhat.com>
4707 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4708 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4710 2013-09-25 Tom Tromey <tromey@redhat.com>
4712 * Makefile.in (generated_files): Add options.h,
4713 target-hooks-def.h, insn-opinit.h,
4714 common/common-target-hooks-def.h, pass-instances.def,
4715 c-family/c-target-hooks-def.h.
4717 2013-09-25 Jeff Law <law@redhat.com>
4719 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4720 than foo[foo.length () - 1] to access last member in a vec.
4721 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4723 2013-09-25 Richard Biener <rguenther@suse.de>
4726 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4728 2013-09-25 Jan Hubicka <jh@suse.cz>
4730 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4733 2013-09-25 Marek Polacek <polacek@redhat.com>
4736 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4737 when determining the type name.
4739 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4741 * config/sh/sh.md: Fix formatting.
4743 2013-09-24 Xinliang David Li <davidxl@google.com>
4745 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4746 max peel iterations parameter.
4747 * param.def: New parameter.
4748 * doc/invoke.texi: Document New parameter.
4750 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4752 * gimple-pretty-print.c: Various whitespace tweaks.
4753 * tree-core.h: Likewise.
4754 * tree-pretty-print.c: Likewise.
4755 * tree-ssa-alias.c: Likewise.
4756 * tree-ssa-copy.c: Likewise.
4757 * tree-ssanames.c: Likewise.
4758 * tree-ssanames.h: Likewise.
4759 * tree-vrp.c: Likewise.
4761 2013-09-24 Alan Modra <amodra@gmail.com>
4765 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4766 for output operands that disallow regs. Don't use EXPAND_WRITE on
4769 2013-09-24 Richard Biener <rguenther@suse.de>
4772 * tree.c (reference_alias_ptr_type): Move ...
4773 * alias.c (reference_alias_ptr_type): ... here and implement
4774 in terms of the new reference_alias_ptr_type_1.
4775 (ref_all_alias_ptr_type_p): New helper.
4776 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4777 use ref_all_alias_ptr_type_p.
4778 (get_deref_alias_set): Add flag_strict_aliasing check here.
4779 (reference_alias_ptr_type_1): New function, split out from ...
4780 (get_alias_set): ... here.
4781 (alias_ptr_types_compatible_p): New function.
4782 * alias.h (reference_alias_ptr_type): Declare.
4783 (alias_ptr_types_compatible_p): Likewise.
4784 * tree.h (reference_alias_ptr_type): Remove.
4785 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4786 to compare MEM_REF alias types.
4788 2013-09-24 Richard Biener <rguenther@suse.de>
4790 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4792 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4794 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4795 reload helper function arrays into a single array reg_addr.
4796 (reload_fpr_gpr): Likewise.
4797 (reload_gpr_vsx): Likewise.
4798 (reload_vsx_gpr): Likewise.
4799 (struct rs6000_reg_addr): Likewise.
4800 (reg_addr): Likewise.
4801 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4802 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4803 (rs6000_init_hard_regno_mode_ok): Likewise.
4804 (rs6000_secondary_reload_direct_move): Likewise.
4805 (rs6000_secondary_reload): Likewise.
4807 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4808 constraints: wu, ww, and wy. Repurpose wv constraint added during
4809 power8 changes. Put wg constraint in alphabetical order.
4811 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4812 for future work to add ISA 2.07 VSX single precision support.
4813 (-mvsx-scalar-double): Change default from -1 to 1, update
4814 documentation comment.
4815 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4816 (-mupper-regs-df): New debug switch to control whether DF values
4817 can go in the traditional Altivec registers.
4818 (-mupper-regs-sf): New debug switch to control whether SF values
4819 can go in the traditional Altivec registers.
4821 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4823 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4824 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4825 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4826 (rs6000_debug_legitimate_address_p): Print if we are running
4827 before, during, or after reload.
4828 (rs6000_secondary_reload): Add a comment.
4829 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4831 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4832 constraints. Update documentation string.
4833 (wd constraint): Likewise.
4834 (wf constraint): Likewise.
4835 (wg constraint): Likewise.
4836 (wn constraint): Likewise.
4837 (ws constraint): Likewise.
4838 (wt constraint): Likewise.
4839 (wx constraint): Likewise.
4840 (wz constraint): Likewise.
4841 (wu constraint): New constraint for ISA 2.07 SFmode scalar
4843 (ww constraint): Likewise.
4844 (wy constraint): Likewise.
4845 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
4846 in the previous submissions.
4847 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
4849 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
4851 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
4852 (REG_BR_PROB): Say that the probability is stored in an int_list.
4853 * reg-notes.def: Update commentary to mention INT_LIST.
4854 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
4855 (INT_LIST): New rtx.
4856 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
4857 * rtlanal.c (int_reg_note_p): New function.
4858 (alloc_reg_note): Assert that the note does not have an int argument.
4859 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
4860 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
4861 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
4862 rather than an INSN_LIST. Handle INT_LIST.
4863 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
4864 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
4865 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
4866 Manipulate them as ints rather than rtxes.
4867 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
4868 * regmove.c (copy_src_to_dest): Likewise.
4869 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
4871 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
4872 into the cases that need it.
4873 * config/arm/arm.c (arm_unwind_emit): Likewise.
4875 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
4876 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
4877 * loop-doloop.c (add_test, doloop_modify): Likewise.
4878 * loop-unswitch.c (compare_and_jump_seq): Likewise.
4879 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
4880 * predict.c (combine_predictions_for_insn): Likewise.
4881 * print-rtl.c (print_rtx): Handle INT_LIST.
4882 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
4883 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
4884 * config/arm/arm.c (emit_unlikely_jump): Likewise.
4885 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
4886 (ix86_split_fp_branch, predict_jump): Likewise.
4887 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
4888 * config/sh/sh.c (expand_cbranchsi4): Likewise.
4889 * config/spu/spu.c (ea_load_store_inline): Likewise.
4891 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
4892 value of a REG_BR_PROB note.
4893 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
4894 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
4895 * emit-rtl.c (try_split): Likewise.
4896 * predict.c (br_prob_note_reliable_p): Likewise.
4897 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
4898 * reorg.c (mostly_true_jump): Likewise.
4899 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
4900 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
4901 * config/i386/i386.c (ix86_print_operand): Likewise.
4902 * config/ia64/ia64.c (ia64_print_operand): Likewise.
4903 * config/mmix/mmix.c (mmix_print_operand): Likewise.
4904 * config/rs6000/rs6000.c (output_cbranch): Likewise.
4905 * config/s390/s390.c (s390_expand_tbegin): Likewise.
4906 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
4907 * config/sparc/sparc.c (output_cbranch): Likewise.
4908 * config/spu/spu.c (get_branch_target): Likewise.
4909 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
4910 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
4912 2013-09-23 Jan Hubicka <jh@suse.cz>
4914 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
4916 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
4917 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
4919 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
4920 (gimple_fold_call): Dump inconsistent devirtualizations; add
4921 sanity check for type based devirtualizations.
4922 * ipa-prop.c: Include ipa-utils.h
4923 (ipa_intraprocedural_devirtualization): Add sanity check.
4924 (try_make_edge_direct_virtual_call): Likewise.
4926 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
4928 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
4929 assignment statements.
4931 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
4933 * gimple-pretty-print.c (dump_ssaname_info): New function.
4934 (dump_gimple_phi): Call it.
4935 (pp_gimple_stmt_1): Likewise.
4936 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
4937 (range_info_def): Declare.
4938 * tree-pretty-print.c (pp_double_int): New function.
4939 (dump_generic_node): Call it.
4940 * tree-pretty-print.h (pp_double_int): Declare.
4941 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
4942 * tree-ssanames.h (range_info_def): New structure.
4943 (value_range_type): Move definition here.
4944 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
4946 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
4948 (set_range_info): New function.
4949 (get_range_info): Likewise.
4950 (duplicate_ssa_name_range_info): Likewise.
4951 (duplicate_ssa_name_fn): Check pointer type and call
4952 duplicate_ssa_name_range_info.
4953 * tree-ssa-copy.c (fini_copy_prop): Likewise.
4954 * tree-vrp.c (value_range_type): Remove definition, now in
4956 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
4957 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
4958 (SSA_NAME_RANGE_INFO): New macro.
4960 2013-09-23 Richard Biener <rguenther@suse.de>
4962 PR tree-optimization/58464
4963 * tree-ssa-pre.c (phi_trans_lookup): Remove.
4964 (phi_trans_add): Change to add conditionally on being not present.
4965 (phi_translate_1): Remove recursion detection here.
4966 (phi_translate): Pre-seed the cache with NULL to catch
4967 recursion here in a more generic way.
4968 (bitmap_find_leader): Adjust comment.
4969 (get_representative_for): Dump value-numbers.
4970 (create_expression_by_pieces): Likewise.
4971 (insert_into_preds_of_block): Likewise.
4973 2013-09-23 Christian Bruel <christian.bruel@st.com>
4976 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
4977 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
4979 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
4982 2013-09-20 Renlin Li <renlin.li@arm.com>
4984 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
4986 (aarch64_expand_epilogue): Likewise.
4987 (aarch64_legitimize_reload_address): Likewise.
4989 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
4991 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
4992 NULL_TREE before pushing them onto the vector. Likewise for labels.
4994 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
4996 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
4997 comparison operators when -fno-trapping-math is in effect.
4998 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
4999 comparison operators in TFmode and assert that unsupported operators
5001 (ia64_print_operand): Likewise.
5003 2013-09-21 Jan Hubicka <jh@suse.cz>
5005 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
5006 (sse_typeless_stores): Enable for core
5007 (sse_load0_by_pxor): Likewise.
5008 (four_jump_limit): Disable for core.
5009 (pad_returns): Likewise.
5010 (avoid_vector_decode): Likewise.
5011 (fuse_cmp_and_branch): Enable for cores.
5012 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
5014 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5017 * config/pa/pa.c (pa_option_override): Disable auto increment and
5018 decrement instructions until reload is completed.
5020 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
5021 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
5022 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
5024 2013-09-20 DJ Delorie <dj@redhat.com>
5025 Nick Clifton <nickc@redhat.com>
5027 * config/rl78/rl78.c: Various whitespace and comment tweaks.
5028 (need_to_save): Save bank 0 on interrupts.
5029 (characterize_address): Strip far address wrappers.
5030 (rl78_as_legitimate_address): Likewise.
5031 (transcode_memory_rtx): Likewise.
5032 (rl78_peep_movhi_p): Disable this peephole after devirt.
5033 (rl78_propogate_register_origins): Forget all origins when a
5035 * config/rl78/rl78-virt.md: Various whitespace tweaks.
5036 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
5038 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
5039 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
5040 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
5042 * config/rl78/constraints.md: Various whitespace and paren tweaks.
5044 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5046 * config/pa/pa.md: In "scc" insn patterns, change output template to
5047 handle const0_rtx in reg_or_0_operand operands.
5049 2013-09-20 Martin Husemann <martin@NetBSD.org>
5052 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
5053 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
5055 2013-09-20 Richard Biener <rguenther@suse.de>
5058 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
5059 remembering SSA name version and block index.
5060 (new_scev_info_str): Adjust.
5061 (hash_scev_info): Likewise. Also hash the block index.
5062 (eq_scev_info): Adjust.
5063 (find_var_scev_info): Likewise.
5064 (struct instantiate_cache_entry): Remove.
5065 (struct instantiate_cache_type): Use a htab to map name, block
5067 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5068 (get_instantiated_value_entry): Likewise.
5069 (hash_idx_scev_info, eq_idx_scev_info): New functions.
5070 (instantiate_scev_name): Adjust.
5072 2013-09-20 Jeff Law <law@redhat.com>
5074 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
5076 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
5078 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
5079 Call aarch64_simd_expand_args to update op[argc].
5081 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5083 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
5086 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5088 * gengtype.c (file_rules): Added rule for *.cc files.
5089 (get_output_file_with_visibility): Give fatal message when no
5092 2013-09-20 Renlin Li <renlin.li@arm.com>
5094 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
5095 (aarch64_expand_epilogue): Likewise.
5096 (aarch64_legitimize_reload_address): Likewise.
5098 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
5101 * expr.c (expand_assignment): Remove misalignp code path.
5103 2013-09-20 Marek Polacek <polacek@redhat.com>
5106 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
5107 TYPE_PRECISION. Add asserts.
5109 2013-09-20 Richard Biener <rguenther@suse.de>
5111 PR tree-optimization/58453
5112 * tree-loop-distribution.c (distribute_loop): Apply the cost
5113 model for -ftree-loop-distribute-patterns, too.
5115 2013-09-20 Richard Biener <rguenther@suse.de>
5118 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
5119 make type comparison less strict.
5121 2013-09-20 Alan Modra <amodra@gmail.com>
5123 * configure: Regenerate.
5124 * aclocal.m4: Regenerate.
5126 2013-09-20 Marek Polacek <polacek@redhat.com>
5129 * doc/extend.texi: Document that attribute used is meant to be used
5130 on variables with static storage duration.
5132 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5134 PR tree-optimization/58472
5135 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
5136 simd_lane_access set inv_p = false.
5137 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
5138 the simduid magic VAR_DECL.
5140 2013-09-19 Jan Hubicka <jh@suse.cz>
5142 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
5144 2013-09-17 Jeff Law <law@redhat.com>
5146 * tree-ssa-dom.c (record_temporary_equivalences): New function
5147 split out of dom_opt_dom_walker::after_dom_children.
5148 (dom_opt_dom_walker::thread_across_edge): Move common code
5149 in here from dom_opt_dom_walker::after_dom_children.
5150 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
5152 2013-09-19 Jan Hubicka <jh@suse.cz>
5154 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
5155 (TARGET_GENERIC): Use PROCESOR_GENERIC
5156 (enum processor_type): Unify generic32 and 64.
5157 * i386.md (cpu): Likewise.
5158 * x86-tune.def (use_leave): Enable for generic32.
5159 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
5160 * athlon.md: Change generic64 to generic in all occurences.
5161 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
5162 (ix86_target_macros_internal): Likewise.
5163 * driver-i386.c (host_detect_local_cpu): Likewise.
5164 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
5166 (generic_memcpy, generic_memset, generic_cost): This one.
5167 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
5168 (m_GENERIC32, m_GENERIC64): Remove.
5169 (m_GENERIC): Turn into one flag.
5170 (processor_target): Unify generic tunnings.
5171 (ix86_option_override_internal): Replace generic32/64 by generic.
5172 (ix86_issue_rate): Likewise.
5173 (ix86_adjust_cost): Likewise.
5175 2013-09-19 Jan Hubicka <jh@suse.cz>
5177 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5178 of speculative flag.
5180 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5182 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5183 GOMP_sections_start, even if !exit_reachable.
5185 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5187 * lra-constraints.c (need_for_all_save_p): Use macro
5188 HARD_REGNO_CALL_PART_CLOBBERED.
5189 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5190 set up pseudo conflict hard regs.
5192 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5195 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5198 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5200 PR rtl-optimization/58438
5201 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5202 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5203 from previous subpasses.
5205 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5207 * arm.c (arm_get_frame_offsets): Validate architecture supports
5208 LDRD/STRD before accepting the tuning preference.
5209 (arm_expand_prologue): Likewise.
5210 (arm_expand_epilogue): Likewise.
5212 2013-09-18 Richard Biener <rguenther@suse.de>
5214 PR tree-optimization/58417
5215 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5216 have chrecs with symbols defined in the loop as operands.
5217 (chrec_fold_multiply): Likewise.
5218 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5219 parameters before folding binary operations.
5220 (struct instantiate_cache_entry_hasher): Remove.
5221 (struct instantiate_cache_type): Use a pointer-map.
5222 (instantiate_cache_type::instantiate_cache_type): New function.
5223 (instantiate_cache_type::get): Likewise.
5224 (instantiate_cache_type::set): Likewise.
5225 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5226 (get_instantiated_value_entry): Likewise.
5227 (global_cache): New global.
5228 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5229 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5230 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5231 (instantiate_scev_name): Adjust.
5232 (instantiate_scev): Construct global instead of local cache.
5233 (resolve_mixers): Likewise.
5235 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5236 Paolo Carlini <paolo.carlini@oracle.com>
5239 * doc/implement-cxx.texi: Fix references to the C++ standards.
5241 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5243 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5244 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5245 simd_lane_access drs, update also DR_ALIGNED_TO.
5247 2013-09-18 Marek Polacek <polacek@redhat.com>
5250 * doc/extend.texi: Document no_sanitize_undefined attribute.
5251 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5252 no_sanitize_undefined attribute.
5254 2013-09-18 Nick Clifton <nickc@redhat.com>
5256 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5257 (ASM_DECLARE_FUNCTION_NAME): Define.
5259 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5261 * compare-elim.c (find_comparison_dom_walker): New class
5262 (find_comparisons_in_bb): Rename to
5263 find_comparison_dom_walker::before_dom_children
5264 (find_comparisons): Adjust
5265 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5267 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5268 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5269 (init_walk_dominator_tree): Remove declaration.
5270 (fini_walk_dominator_tree): Remove declaration.
5271 * fwprop.c (single_def_use_dom_walker): New class
5272 (single_def_use_enter_block): Convert to
5273 single_def_use_dom_walker::before_dom_children.
5274 (single_def_use_leave_block): Convert to
5275 single_def_use_dom_walker::after_dom_children.
5276 (build_single_def_use_links): Adjust.
5277 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5279 (find_candidates_in_block): Convert to
5280 find_candidates_dom_walker::before_dom_children.
5281 (execute_strength_reduction): Adjust.
5282 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5283 (sese_dom_walker): New class.
5284 (sese_dom_walker::sese_dom_walker): New constructor.
5285 (sese_dom_walker::~sese_dom_walker): New destructor.
5286 (build_sese_conditions_before): Convert to
5287 sese_dom_walker::before_dom_children.
5288 (build_sese_conditions_after): Convert to
5289 sese_dom_walker::after_dom_children.
5290 (build_poly_scop): Adjust
5291 * tree-into-ssa.c (rewrite_dom_walker): New class
5292 (rewrite_enter_block): Convert to
5293 rewrite_dom_walker::before_dom_children.
5294 (rewrite_leave_block): Convert to
5295 rewrite_dom_walker::after_dom_children.
5296 (rewrite_update_dom_walker): New class.
5297 (rewrite_update_enter_block): Convert to
5298 rewrite_update_dom_walker::before_dom_children.
5299 (rewrite_update_leave_block): Convert to
5300 rewrite_update_dom_walker::after_dom_children.
5301 (rewrite_blocks, rewrite_into_ssa): Adjust.
5302 (mark_def_dom_walker): New class.
5303 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5304 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5305 (mark_def_sites_blocks): Convert to
5306 mark_def_dom_walker::before_dom_children.
5307 (mark_def_site_blocks): Remove.
5308 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5309 (tree_ssa_dominator_optimize): Adjust.
5310 (dom_thread_across_edge): Convert to method
5311 dom_opt_dom_walker::thread_across_edge.
5312 (dom_opt_enter_block): Convert to member function
5313 dom_opt_dom_walker::before_dom_children.
5314 (dom_opt_leave_block): Convert to member function
5315 dom_opt_dom_walker::after_dom_children.
5316 * tree-ssa-dse.c (dse_dom_walker): New class.
5317 (dse_enter_block): Convert to member function
5318 dse_dom_walker::before_dom_children.
5319 (tree_ssa_dse): Adjust.
5320 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5321 (determine_invariantness_stmt): Convert to method
5322 invariantness_dom_walker::before_dom_children.
5323 (determine_invariantness): Remove
5324 (move_computations_dom_walker): New class.
5325 (move_computations_stmt): Convert to method
5326 move_computations_dom_walker::before_dom_children.
5327 (move_computations, tree_ssa_lim): Adjust.
5328 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
5329 (nt_init_block): Convert to method
5330 notrappping_dom_walker::before_dom_children.
5331 (nt_fini_block): Convert to method
5332 method nontrapping_dom_walker::after_dom_children.
5333 (get_non_trapping): Adjust.
5334 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5335 (eliminate_bb): Convert to method
5336 eliminate_dom_walker::before_dom_children.
5337 (eliminate_leave_block): Convert to method
5338 eliminate_dom_walker::after_dom_children.
5339 (eliminate): Adjust.
5340 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5341 (strlen_enter_block): Convert to method
5342 strlen_dom_walker::before_dom_children.
5343 (strlen_leave_block): Convert to method
5344 method strlen_dom_walker::after_dom_children.
5345 (tree_ssa_strlen): Adjust.
5346 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5347 (tree_ssa_uncprop): Adjust.
5348 (uncprop_leave_block): Convert to method
5349 uncprop_dom_walker::after_dom_children.
5350 (uncprop_leave_block): Convert to method
5351 uncprop_dom_walker::before_dom_children.
5353 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5355 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5356 branch in same basic block. Check both src and dest of the move insn.
5358 2013-09-17 Nick Clifton <nickc@redhat.com>
5360 * config/rl78/rl78-real.md (bf): New pattern.
5362 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5363 (rl78_print_operand): Do not put a # before a %B.
5364 * config/rl78/rl78.opt: Tweak doc strings.
5366 2013-09-17 DJ Delorie <dj@redhat.com>
5368 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5369 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5370 (rl78_option_override): Likewise, if -mallregs.
5371 (is_virtual_register): Likewise.
5372 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5373 (REGNO_OK_FOR_BASE_P): Likewise.
5374 * config/rl78/rl78.opt (-mallregs): New.
5376 2013-09-17 Nick Clifton <nickc@redhat.com>
5378 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5379 For interrupt functions: save all call clobbered registers if the
5380 interrupt handler is not a leaf function.
5381 (rl78_expand_prologue): Always recompute the frame information.
5382 For interrupt functions: only select bank 0 if one of the bank 0
5383 registers is going to be psuhed.
5385 2013-09-17 DJ Delorie <dj@redhat.com>
5387 * config/rl78/constraints.md: For each W* constraint, rename to C*
5388 and create a W* constraint that checks for an optional ES: prefix
5390 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5391 (es_addr): New. Used to wrap far addresses.
5392 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5393 (rl78_es_base): New.
5394 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5395 wrapped far addresses.
5396 (rl78_print_operand_1): Unwrap far addresses before processing.
5397 (rl78_lo16): Wrap far addresses in unspecs.
5398 (rl78_es_addr): New.
5399 (rl78_es_base): New.
5400 (insn_ok_now): Check for not-yet-wrapped far addresses.
5401 (transcode_memory_rtx): Properly re-wrap far addresses.
5403 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5405 * config/sparc/t-rtems: Add leon3 multilibs.
5407 2013-09-17 Cong Hou <congh@google.com>
5409 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5410 when checking the dot production pattern. The type of rhs operand
5411 of multiply is now checked correctly.
5413 2013-09-17 Jeff Law <law@redhat.com>
5415 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5416 edge implied equivalences into successor phis.
5417 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5418 here from tree-ssa-threadedge.c.
5419 (mark_threaded_blocks): When threading through a joiner, if both
5420 successors of the joiner's clone reach the same block, verify the
5421 PHI arguments are equal. If not, cancel the jump threading request.
5422 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5423 tree-ssa-threadupdate.c
5424 (thread_across_edge): Don't check PHI argument equality when
5425 threading through joiner block here.
5427 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5429 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5430 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5431 (warn_uninit, warn_uninitialized_vars,
5432 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5433 Move to tree-ssa-uninit.c.
5434 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5435 (has_undefined_value_p): New. Pass dependant parts of
5436 ssa_undefined_value_p.
5437 (uninit_undefined_value_p): Use has_undefined_value_p.
5438 (warn_uninit, warn_uninitialized_vars,
5439 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5440 Move from tree-ssa.c.
5441 * tree-ssa.h: Adjust prototypes.
5443 2013-09-17 Jan Hubicka <jh@suse.cz>
5446 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5447 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5448 FUNCTION_NOT_OPTIMIZED.
5449 * ipa-inline-analysis.c (compute_inline_parameters): Function
5450 not optimized is not inlinable unless it is alwaysinline.
5451 (inline_analyze_function): Force calls in not optimized
5452 function not inlinable.
5454 2013-09-17 Jan Hubicka <jh@suse.cz>
5457 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5459 * ipa.c (function_and_variable_visibility): Likewise.
5460 * ipa-profile.c (ipa_profile): Likewise.
5462 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5465 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5466 attribute, and don't inline the resolver in this case.
5468 2013-09-17 Teresa Johnson <tejohnson@google.com>
5470 * coverage.c (get_coverage_counts): Add missing newline.
5472 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5474 PR tree-optimization/58088
5475 * fold-const.c (mask_with_trailing_zeros): New function.
5476 (fold_binary_loc): Make sure we don't recurse infinitely
5477 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5478 Use mask_with_trailing_zeros where appropriate.
5480 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5482 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5483 of 'prev' var to get better distance estimation.
5485 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5487 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5488 * tree-inline.c (is_parameter_of): New predicate.
5489 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5490 a parameter has been remapped.
5491 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5492 (optimize_inline_calls): Initialize transform_parameter.
5493 (copy_gimple_seq_and_replace_locals): Likewise.
5494 (tree_function_versioning): Likewise.
5495 (maybe_inline_call_in_expr): Likewise.
5497 2013-09-17 Nick Clifton <nickc@redhat.com>
5499 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5500 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5502 (is_attr_func): New function.
5503 (msp430_is_interrupt_func): New function.
5504 (is_naked_func): New function.
5505 (is_reentrant_func): New function.
5506 (is_critical_func): New function.
5507 (msp430_start_function): Add annotations for function attributes.
5508 (msp430_attr): New function.
5509 (msp430_attribute_table): New.
5510 (msp430_function_section): New function.
5511 (TARGET_ASM_FUNCTION_SECTION): Define.
5512 (msp430_builtin): New enum.
5513 (msp430_init_builtins): New function.
5514 (msp430_builtin_devl): New function.
5515 (msp430_expand_builtin): New function.
5516 (TARGET_INIT_BUILTINS): Define.
5517 (TARGET_EXPAND_BUILTINS): Define.
5518 (TARGET_BUILTIN_DECL): Define.
5519 (msp430_expand_prologue): Add support for naked, interrupt,
5520 critical and reentrant functions.
5521 (msp430_expand_epilogue): Likewise.
5522 (msp430_print_operand): Handle 'O' character.
5523 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5525 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5526 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5527 (pushm): Use a 'n' rather than an 'i' constraint.
5528 (msp_return): Add generation of the interrupt return instruction.
5529 (disable_interrupts): New pattern.
5530 (enable_interrupts): New pattern.
5531 (push_intr_state): New pattern.
5532 (pop_intr_state): New pattern.
5533 (bic_SR): New pattern.
5534 (bis_SR): New pattern.
5535 * doc/extend.texi: Document MSP430 function attributes and builtin
5538 2013-09-17 Richard Biener <rguenther@suse.de>
5540 PR tree-optimization/58432
5541 * tree-loop-distribution.c (tree_loop_distribution): Also
5542 scan PHIs for outside loop uses and seed a partition from them.
5544 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5546 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5547 (restructure_reference): Call backtrace_base_for_ref.
5549 2013-09-17 Alan Modra <amodra@gmail.com>
5552 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5555 2013-09-16 DJ Delorie <dj@redhat.com>
5557 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5558 vregs location for RL78/G10.
5559 (rl78_expand_prologue): Avoid SEL on G10.
5560 (rl78_expand_epilogue): Likewise.
5561 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5562 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5563 __RL78_G10__ when appropriate.
5564 (ASM_SPEC): Pass -mg10 along to the assembler.
5565 * config/rl78/rl78.md (sel_rb): Disable for G10.
5566 * config/rl78/rl78.opt: Add -mg10 option.
5567 * config/rl78/t-rl78: Add -mg10 multilib.
5569 2013-09-16 Xinliang David Li <davidxl@google.com>
5571 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5572 * omp-low.c (omp_max_vf): Ditto.
5573 (expand_omp_simd): Ditto.
5574 * tree-vectorizer.c (vectorize_loops): Ditto.
5575 (gate_vect_slp): Ditto.
5576 (gate_increase_alignment): Ditto.
5577 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5578 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5579 (gate_tree_vectorize): Name change.
5580 (tree_vectorize): Ditto.
5581 (pass_vectorize::gate): Call new function.
5582 (pass_vectorize::execute): Ditto.
5583 * opts.c: O3 default setting change.
5584 (finish_options): Check new flag.
5585 * doc/invoke.texi: Document new flags.
5586 * common.opt: New flags.
5588 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5590 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5591 (Condition Code): Don't reference it.
5592 * doc/tm.texi: Regenerate.
5594 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5597 * lra-constraints.c (undo_optional_reloads): Consider all optional
5598 reload even if it did not get a hard reg.
5600 2013-09-16 Teresa Johnson <tejohnson@google.com>
5602 * dumpfile.c (dump_loc): Remove newline emission.
5603 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5604 emission to dump_printf_loc calls where missing.
5605 (vect_mark_for_runtime_alias_test): Ditto.
5606 (vect_analyze_data_ref_dependence): Ditto.
5607 (vect_analyze_data_ref_dependences): Ditto.
5608 (vect_slp_analyze_data_ref_dependence): Ditto.
5609 (vect_slp_analyze_data_ref_dependences): Ditto.
5610 (vect_compute_data_ref_alignment): Ditto.
5611 (vect_update_misalignment_for_peel): Ditto.
5612 (vect_verify_datarefs_alignment): Ditto.
5613 (vector_alignment_reachable_p): Ditto.
5614 (vect_get_data_access_cost): Ditto.
5615 (vect_enhance_data_refs_alignment): Ditto.
5616 (vect_find_same_alignment_drs): Ditto.
5617 (vect_analyze_data_refs_alignment): Ditto.
5618 (vect_analyze_group_access): Ditto.
5619 (vect_analyze_data_ref_access): Ditto.
5620 (vect_analyze_data_ref_accesses): Ditto.
5621 (vect_prune_runtime_alias_test_list): Ditto.
5622 (vect_analyze_data_refs): Ditto.
5623 (vect_create_addr_base_for_vector_ref): Ditto.
5624 (vect_create_data_ref_ptr): Ditto.
5625 (vect_grouped_store_supported): Ditto.
5626 (vect_grouped_load_supported): Ditto.
5627 * value-prof.c (check_counter): Ditto.
5628 (check_ic_target): Ditto.
5629 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5630 (vect_recog_widen_mult_pattern): Ditto.
5631 (vect_recog_widen_sum_pattern): Ditto.
5632 (vect_recog_over_widening_pattern): Ditto.
5633 (vect_recog_widen_shift_pattern): Ditto.
5634 (vect_recog_rotate_pattern): Ditto.
5635 (vect_recog_vector_vector_shift_pattern): Ditto.
5636 (vect_recog_divmod_pattern): Ditto.
5637 (vect_recog_mixed_size_cond_pattern): Ditto.
5638 (vect_recog_bool_pattern): Ditto.
5639 (vect_pattern_recog_1): Ditto.
5640 (vect_pattern_recog): Ditto.
5641 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5642 (vect_is_simple_iv_evolution): Ditto.
5643 (vect_analyze_scalar_cycles_1): Ditto.
5644 (vect_get_loop_niters): Ditto.
5645 (vect_analyze_loop_1): Ditto.
5646 (vect_analyze_loop_form): Ditto.
5647 (vect_analyze_loop_operations): Ditto.
5648 (vect_analyze_loop_2): Ditto.
5649 (vect_analyze_loop): Ditto.
5650 (report_vect_op): Ditto.
5651 (vect_is_slp_reduction): Ditto.
5652 (vect_is_simple_reduction_1): Ditto.
5653 (vect_get_known_peeling_cost): Ditto.
5654 (vect_estimate_min_profitable_iters): Ditto.
5655 (vect_model_reduction_cost): Ditto.
5656 (vect_model_induction_cost): Ditto.
5657 (get_initial_def_for_induction): Ditto.
5658 (vect_create_epilog_for_reduction): Ditto.
5659 (vectorizable_reduction): Ditto.
5660 (vectorizable_induction): Ditto.
5661 (vectorizable_live_operation): Ditto.
5662 (vect_loop_kill_debug_uses): Ditto.
5663 (vect_transform_loop): Ditto.
5664 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5665 (vect_stmt_relevant_p): Ditto.
5666 (process_use): Ditto.
5667 (vect_mark_stmts_to_be_vectorized): Ditto.
5668 (vect_model_simple_cost): Ditto.
5669 (vect_model_promotion_demotion_cost): Ditto.
5670 (vect_model_store_cost): Ditto.
5671 (vect_get_store_cost): Ditto.
5672 (vect_model_load_cost): Ditto.
5673 (vect_get_load_cost): Ditto.
5674 (vect_init_vector_1): Ditto.
5675 (vect_get_vec_def_for_operand): Ditto.
5676 (vect_finish_stmt_generation): Ditto.
5677 (vectorizable_call): Ditto.
5678 (vectorizable_conversion): Ditto.
5679 (vectorizable_assignment): Ditto.
5680 (vectorizable_shift): Ditto.
5681 (vectorizable_operation): Ditto.
5682 (vectorizable_store): Ditto.
5683 (vectorizable_load): Ditto.
5684 (vectorizable_condition): Ditto.
5685 (vect_analyze_stmt): Ditto.
5686 (vect_transform_stmt): Ditto.
5687 (vect_is_simple_use): Ditto.
5688 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5689 (vect_can_advance_ivs_p): Ditto.
5690 (vect_update_ivs_after_vectorizer): Ditto.
5691 (vect_do_peeling_for_loop_bound): Ditto.
5692 (vect_gen_niters_for_prolog_loop): Ditto.
5693 (vect_update_inits_of_drs): Ditto.
5694 (vect_create_cond_for_alias_checks): Ditto.
5695 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5696 (vect_build_slp_tree_1): Ditto.
5697 (vect_supported_load_permutation_p): Ditto.
5698 (vect_analyze_slp_instance): Ditto.
5699 (vect_analyze_slp): Ditto.
5700 (vect_make_slp_decision): Ditto.
5701 (vect_detect_hybrid_slp): Ditto.
5702 (vect_bb_vectorization_profitable_p): Ditto.
5703 (vect_slp_analyze_bb_1): Ditto.
5704 (vect_update_slp_costs_according_to_vf): Ditto.
5705 (vect_get_mask_element): Ditto.
5706 (vect_transform_slp_perm_load): Ditto.
5707 (vect_schedule_slp_instance): Ditto.
5708 (vect_schedule_slp): Ditto.
5709 (vect_slp_transform_bb): Ditto.
5710 * profile.c (read_profile_edge_counts): Ditto.
5711 (compute_branch_probabilities): Ditto.
5712 * coverage.c (get_coverage_counts): Ditto.
5714 2013-09-16 Diego Novillo <dnovillo@google.com>
5716 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5718 2013-09-16 Jan Hubicka <jh@suse.cz>
5720 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5721 abstract functions; for static functions check the presence of body.
5723 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5725 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5726 * config/aarch64/aarch64-simd.md
5727 (aarch64_mla_elt<mode>): New.
5728 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5729 (aarch64_mls_elt<mode>): Likewise.
5730 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5731 (aarch64_fma4_elt<mode>): Likewise.
5732 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5733 (aarch64_fma4_elt_to_128v2df): Likewise.
5734 (aarch64_fma4_elt_to_64df): Likewise.
5735 (fnma<mode>4): Likewise.
5736 (aarch64_fnma4_elt<mode>): Likewise.
5737 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5738 (aarch64_fnma4_elt_to_128v2df): Likewise.
5739 (aarch64_fnma4_elt_to_64df): Likewise.
5740 * config/aarch64/iterators.md (VDQSF): New.
5741 * config/aarch64/arm_neon.h
5742 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5743 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5745 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5747 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5748 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5749 (aarch64_mul3_elt_to_128df): Likewise.
5750 (aarch64_mul3_elt_to_64v2df): Likewise.
5751 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5753 (VMUL_CHANGE_NLANES) Likewise.
5756 * config/aarch64/arm_neon.h
5757 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5759 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5761 * config/aarch64/arm_neon.h
5762 (vcvtx_high_f32_f64): Fix parameters.
5764 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5765 Uros Bizjak <ubizjak@gmail.com>
5767 * config/alpha.c: Include tree-ssanames.h.
5769 2013-09-16 Richard Biener <rguenther@suse.de>
5771 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5772 (dot_rdg_1): Handle control_dd.
5773 (create_edge_for_control_dependence): New function.
5774 (create_rdg_edges): Add control dependences if asked for.
5775 (build_rdg): Likewise.
5776 (generate_loops_for_partition): If there are not necessary
5777 control stmts remove all their dependencies.
5778 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5779 (distribute_loop): Pass on control dependences.
5780 (tree_loop_distribution): Compute control dependences and remove
5781 restriction on number of loop nodes.
5783 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5785 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5788 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5790 * cse.c (try_const_anchors): Punt on CC modes.
5792 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5794 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5796 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5799 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5802 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5805 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5807 (darwin_objc1_section): Likewise.
5808 (darwin_file_end): Emit Image Info section when required.
5810 2013-09-14 Jan Hubicka <jh@suse.cz>
5812 * tree-into-ssa.c (gate_into_ssa): New.
5813 (pass_data_build_ssa): Use it.
5814 * cgraph.h (expand_thunk): Update prototype.
5815 * cgraphunit.c (analyze_function): Expand thunks early.
5816 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5817 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5818 set lowered flag; do not add new function.
5819 (assemble_thunks_and_aliases): Update.
5820 * tree-ssa.c (gate_init_datastructures): New gate.
5821 (pass_data_init_datastructures): Use it.
5823 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5826 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5827 xmm register set described in the psABI.
5829 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5831 * dwarf2out.c (should_emit_struct_debug): Add check
5832 for type_decl variable is not NULL.
5834 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5836 * config.gcc: Use new winnt-c.c target hooks
5837 * config/t-winnt: New file
5838 * config/winnt-c.c: New file
5839 * doc/tm.texi.in: Document new hook
5840 * doc/tm.texi: Regenerated
5842 2013-09-13 Jan Hubicka <jh@suse.cz>
5845 * ipa-inline.c (check_callers): New function.
5846 (check_caller_edge): Remove.
5847 (want_inline_function_to_all_callers_p): Also permit alises that are
5849 (inline_to_all_callers): Terminate the walk when devirtualization
5850 introduce new calls.
5852 2013-09-13 Jan Hubicka <jh@suse.cz>
5854 * ipa-inline-analysis.c (struct growth_data): Add node.
5855 (do_estimate_growth_1): Fix detection of recursion.
5857 2013-09-13 Jakub Jelinek <jakub@redhat.com>
5859 PR tree-optimization/58392
5860 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
5861 to avoid shadowing of outer loop variable. If
5862 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
5863 replace_by_duplicate_decl simduid of loops that have it set and
5864 set dest_cfun->has_simduid_loops and/or
5865 dest_cfun->has_force_vect_loops.
5866 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
5867 instead of maybe_lookup_decl.
5868 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
5869 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
5870 call. Copy over force_vect and copy and remap simduid. Set
5871 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
5872 (copy_cfg_body): Remove blocks_to_copy argument. Use
5873 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
5874 caller. Don't set cfun->has_simduid_loops and/or
5875 cfun->has_force_vect_loops here.
5876 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
5878 (expand_call_inline, tree_function_versioning): Adjust copy_body
5881 2013-09-13 Martin Jambor <mjambor@suse.cz>
5884 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
5885 the assert if the edge was a speculative one.
5887 2013-09-13 Richard Biener <rguenther@suse.de>
5889 * tree-data-ref.h (known_dependences_p): Move here ...
5890 * tree-loop-distribution.c (known_dependences_p): ... from here.
5891 (dump_rdg_component, debug_rdg_component): Remove.
5893 (generate_loops_for_partition): Use gimple_uid instead of
5894 relying on matching stmt visit order.
5895 (rdg_build_partitions): Take starting stmt vector.
5896 (ldist_gen): Merge into ...
5897 (distribute_loop): ... this function. Do not compute starting
5899 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
5901 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5903 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
5904 Set type attribute correctly. Set predicable_short_it attribute.
5905 (cmpsi_shiftsi): Remove %? from output template.
5907 2013-09-13 Richard Biener <rguenther@suse.de>
5909 * tree-loop-distribution.c (struct rdg_component,
5910 rdg_defs_used_in_other_loops_p, free_rdg_components,
5911 rdg_build_components): Remove.
5912 (stmts_from_loop): Do not record virtual PHIs.
5913 (generate_loops_for_partition): Skip virtual PHIs.
5914 (build_rdg_partition_for_component): Rename to ...
5915 (build_rdg_partition_for_vertex): ... this and adjust.
5916 (rdg_build_partitions): Take a vector of starting vertices
5917 instead of components. Remove unnecessary leftover handling.
5918 (ldist_gen): Do not build components or record other stores.
5919 (distribute_loop): Do not distribute loops containing stmts
5922 2013-09-13 Christian Bruel <christian.bruel@st.com>
5925 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
5927 2013-09-13 Kai Tietz <ktietz@redhat.com>
5929 * config.gcc: Separate cases for mingw and cygwin targets,
5930 and add 64-bit cygwin target case.
5932 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
5933 dll-export inline-functions.
5934 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
5937 2013-09-13 Jeff Law <law@redhat.com>
5941 2013-09-06 Jeff Law <law@redhat.com>
5943 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5944 edge implied equivalences into successor phis.
5946 2013-09-12 DJ Delorie <dj@redhat.com>
5948 * config/rl78/rl78-virt.md: Change from | to \; for asm line
5951 2013-09-12 Brooks Moses <bmoses@google.com>
5954 * Makefile.in: Do not install driver binaries in $(target)/bin.
5956 2013-09-12 DJ Delorie <dj@redhat.com>
5958 * config/rl78/rl78.opt (mrelax): New.
5959 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
5960 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
5962 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
5964 (rl78_expand_epilogue): Likewise.
5966 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
5969 * lra-eliminations.c (remove_reg_equal_offset_note): New.
5970 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
5971 pointer elimination with using remove_reg_equal_offset_note.
5973 2013-09-12 DJ Delorie <dj@redhat.com>
5975 * config/msp430/: New port.
5976 * config.gcc (msp430): Added.
5977 * doc/invoke.texi: Document MSP430 options.
5978 * doc/install.texi: Document msp430-elf
5979 * doc/md.texi: Document msp430-elf
5980 * doc/contrib.texi: Document msp430-elf
5982 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
5985 2013-09-12 Martin Jambor <mjambor@suse.cz>
5988 * ipa-prop.c (remove_described_reference): Give up if the edge in the
5989 reference descriptor is NULL.
5990 (ipa_edge_removal_hook): If owning a reference descriptor, set its
5993 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
5995 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
5996 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
5997 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
5998 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
5999 make_temp_ssa_name): move to tree-ssanames.h
6000 * tree-ssa-alias.h: Move prototype.
6001 * tree-ssa.h: Include tree-ssanames.h.
6002 * tree-ssanames.c (FREE_SSANAMES): Move to here.
6003 * tree-ssanames.h: New. Move items from tree-flow*.h
6004 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
6006 2013-09-12 Richard Biener <rguenther@suse.de>
6008 PR tree-optimization/58404
6009 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
6010 propagate non-invariant addresses into dereferences wrapped
6011 in component references.
6013 2013-09-12 Richard Biener <rguenther@suse.de>
6015 PR tree-optimization/58402
6016 * passes.def: Move pass_late_warn_uninitialized later.
6018 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6020 * tree-ssa.h: New. Move content from tree-flow.h and
6022 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
6023 Move prototypes belonging to tree-ssa.c.
6024 * tree-flow-inline.h (redirect_edge_var_map_def,
6025 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
6027 * gimple.h: Adjust prototypes.
6028 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
6030 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
6031 * tree.h: Move prototype to tree-ssa.h.
6032 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
6033 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
6034 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
6035 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
6036 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
6037 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
6038 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
6039 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
6040 graphite-blocking.c, graphite-clast-to-gimple.c,
6041 graphite-dependences.c, graphite-interchange.c,
6042 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
6043 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
6044 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
6045 ipa-reference.c, ipa-split.c, ipa-utils.c,
6046 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
6047 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
6048 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
6049 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6050 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
6051 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
6052 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
6053 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
6054 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
6055 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
6056 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
6057 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6058 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
6059 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
6060 value-prof.c, var-tracking.c,
6061 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
6063 2013-09-12 Richard Biener <rguenther@suse.de>
6065 PR tree-optimization/58396
6066 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
6067 (build_rdg): Take a loop-nest parameter, fix memleaks.
6068 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
6070 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6072 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
6075 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
6077 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
6078 The FMA instruction names should have a 'v' prefix.
6080 2013-09-12 Richard Biener <rguenther@suse.de>
6082 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
6083 (dot_rdg): Use popen instead of system in optional code.
6084 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
6085 (already_processed_vertex_p): Adjust.
6086 (has_anti_or_output_dependence, predecessor_has_mem_write,
6087 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
6088 rdg_flag_uses): Remove.
6089 (rdg_flag_vertex): Simplify.
6090 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
6092 (build_rdg_partition_for_component): Process the first vertex
6093 of a component only.
6094 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
6096 2013-09-12 Alan Modra <amodra@gmail.com>
6098 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
6100 2013-09-11 DJ Delorie <dj@redhat.com>
6101 Nick Clifton <nickc@redhat.com>
6103 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
6104 (rl78_stack_based_mem): New.
6105 * config/rl78/constraints.md (Iv08): New.
6114 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
6116 (movsi): Change from expand to insn-and-split.
6117 (ashrsi3): Clobber AX.
6121 * config/rl78/rl78.md (CC_REG): Fix.
6122 (addsi3): Allow memory and immediate operands.
6123 (addsi3_internal): Split into...
6124 (addsi3_internal_virt): ...new, and ...
6125 (addsi3_internal_real): ...new.
6127 (subsi3_internal_virt): New.
6128 (subsi3_internal_real): New.
6129 (mulsi3): Add memory operand.
6130 (mulsi3_rl78): Likewise.
6131 (mulsi3_g13): Likewise.
6132 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
6133 (cbranchqi4_real): Add more constraint options.
6134 (cbranchhi4_real): Expand pattern.
6135 (cbranchhi4_real_signed): New.
6136 (cbranchhi4_real_inverted): New.
6137 (cbranchsi4_real_lt): New.
6138 (cbranchsi4_real_ge): New.
6139 (cbranchsi4_real_signed): New.
6140 (cbranchsi4_real): New.
6142 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
6144 (lshrsi3_virt): Likewise.
6145 (ashlsi3_virt): Likewise.
6146 (cbranchqi4_virt_signed): New.
6147 (cbranchhi4_virt_signed): New.
6148 (cbranchsi4_virt): New.
6149 * config/rl78/rl78.c: Whitespace fixes throughout.
6150 (move_elim_pass): New.
6151 (pass_data_rl78_move_elim): New.
6152 (pass_rl78_move_elim): New.
6153 (make_pass_rl78_move_elim): New.
6154 (rl78_devirt_info): Run devirt earlier.
6155 (rl78_move_elim_info): New.
6156 (rl78_asm_file_start): Register it.
6157 (rl78_split_movsi): New.
6158 (rl78_as_legitimate_address): Allow virtual base registers when
6160 (rl78_addr_space_convert): Remove spurious debug stuff.
6161 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
6162 (rl78_print_operand): More cases for not printing '#'.
6163 (rl78_expand_compare): Remove most of the logic.
6164 (content_memory): New.
6165 (clear_content_memory): New.
6166 (get_content_index): New.
6167 (get_content_name): New.
6168 (display_content_memory): New.
6169 (update_content): New.
6170 (record_content): New.
6171 (already_contains): New.
6172 (insn_ok_now): Re-recog insns with virtual registers.
6173 (add_postponed_content_update): New.
6174 (process_postponed_content_update): New.
6175 (gen_and_emit_move): New.
6176 (transcode_memory_rtx): Record new location content.
6177 Use gen_and_emit_move.
6178 (force_into_acc): New.
6179 (move_to_acc): Use gen_and_emit_move.
6180 (move_from_acc): Likewise.
6181 (move_acc_to_reg): Likewise.
6182 (move_to_x): Likewise.
6183 (move_to_hl): Likewise.
6184 (move_to_de): Likewise.
6185 (rl78_alloc_physical_registers_op1): Record location content.
6186 (has_constraint): New.
6187 (rl78_alloc_physical_registers_op2): Record location content.
6189 (rl78_alloc_physical_registers_ro1): Likewise.
6190 (rl78_alloc_physical_registers_cmp): Likewise.
6191 (rl78_alloc_physical_registers_umul): Likewise.
6192 (rl78_alloc_address_registers_macax): New.
6193 (rl78_alloc_physical_registers): Initialize and set location
6194 content memory as needed.
6195 (rl78_reorg): Make sure split2 is called.
6196 (rl78_rtx_costs): New.
6198 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6200 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6201 for (not (neg ...)) and (neg (not ...)) cases.
6203 2013-09-11 Richard Biener <rguenther@suse.de>
6206 * passes.def: Split critical edges before late uninit warning passes.
6207 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6209 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6211 PR tree-optimization/58385
6212 * fold-const.c (build_range_check): If both low and high are NULL,
6213 use omit_one_operand_loc to preserve exp side-effects.
6215 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6217 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6219 2013-09-11 Richard Biener <rguenther@suse.de>
6221 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6222 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6223 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6224 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6225 stmts_from_loop, known_dependences_p, build_empty_rdg,
6226 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6227 * tree-loop-distribution.c: ... here.
6228 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6229 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6230 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6231 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6232 * tree-loop-distribution.c: ... here.
6233 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6234 (struct partition_s): Add loops member.
6235 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6236 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6237 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6239 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6240 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6241 Sergey Lega <sergey.s.lega@intel.com>
6242 Anna Tikhonova <anna.tikhonova@intel.com>
6243 Ilya Tocar <ilya.tocar@intel.com>
6244 Andrey Turetskiy <andrey.turetskiy@intel.com>
6245 Ilya Verbin <ilya.verbin@intel.com>
6246 Kirill Yukhin <kirill.yukhin@intel.com>
6247 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6249 * config/i386/constraints.md (k): New.
6251 * config/i386/i386.c (const regclass_map): Add new mask registers.
6252 (dbx_register_map): Ditto.
6253 (dbx64_register_map): Ditto.
6254 (svr4_dbx_register_map): Ditto.
6255 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6257 (ix86_preferred_reload_class): Disable constants for mask registers.
6258 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6259 (ix86_hard_regno_mode_ok): Support new mask registers.
6260 (x86_order_regs_for_local_alloc): Ditto.
6261 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6262 (FIXED_REGISTERS): Add new mask registers.
6263 (CALL_USED_REGISTERS): Ditto.
6264 (REG_ALLOC_ORDER): Ditto.
6265 (VALID_MASK_REG_MODE): New.
6266 (FIRST_MASK_REG): Ditto.
6267 (LAST_MASK_REG): Ditto.
6268 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6269 (MAYBE_MASK_CLASS_P): New.
6270 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6271 (REG_CLASS_CONTENTS): Ditto.
6272 (MASK_REGNO_P): New.
6273 (ANY_MASK_REG_P): Ditto.
6274 (HI_REGISTER_NAMES): Add new mask registers.
6275 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6276 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6278 (attribute "type"): Add mskmov, msklog.
6279 (attribute "length_immediate"): Support them.
6280 (attribute "memory"): Ditto.
6281 (attribute "prefix_0f"): Ditto.
6282 (*movhi_internal): Support new mask registers.
6283 (*movqi_internal): Ditto.
6284 (define_split): Split out clobber pattern is a logic
6285 insn on mask registers.
6286 (*k<logic><mode>): New.
6287 (*andhi_1): Extend to support mask regs.
6288 (*andqi_1): Extend to support mask regs.
6290 (define_split): Split and-not to and and not if operands
6292 (*<code><mode>_1): Separate HI mode to new pattern...
6293 (*<code>hi_1): This.
6294 (*<code>qi_1): Extend to support mask regs.
6296 (kortestzhi): Ditto.
6297 (kortestchi): Ditto.
6299 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6300 (*one_cmplhi2_1): ...Here, now with mask registers support.
6301 (*one_cmplqi2_1): Support new mask registers.
6302 (HI/QImode arithmetics splitter): Don't split if mask registers
6304 (HI/QImode not splitter): Ditto.
6305 * config/i386/predicated.md (mask_reg_operand): New.
6306 (general_reg_operand): Ditto.
6308 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6310 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6311 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6313 2013-09-10 Jeff Law <law@redhat.com>
6315 PR tree-optimization/58380
6316 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6317 of threading through a buried loop header.
6319 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6320 return value for single successor case.
6322 2013-09-10 Jan Hubicka <jh@suse.cz>
6324 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6326 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6329 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6330 support conditional execution.
6331 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6333 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6335 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6337 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6338 Check destination too to check move insn.
6339 (undo_optional_reloads): Add check that the original peudo did not
6340 changed its allocation and the optional reload was inherited on last
6341 inheritance pass. Break loop after deciding to keep optional reload.
6342 (lra_undo_inheritance): Add check that inherited pseudo still in
6345 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6347 * config/aarch64/aarch64.md (generic_sched): New.
6348 * config/aarch64/aarch64-generic.md (load): Make conditional
6349 on generic_sched attribute.
6350 (nonload): Likewise.
6352 2013-09-10 Jan Hubicka <jh@suse.cz>
6354 * lto-cgraph.c: Include ipa-utils.h.
6355 (compute_ltrans_boundary): Also add possible targets into the boundary.
6357 2013-09-10 Jan Hubicka <jh@suse.cz>
6359 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6360 VAR_DECL of vtable rather than full expression.
6362 2013-09-10 Jan Hubicka <jh@suse.cz>
6363 Paolo Carlini <paolo.carlini@oracle.com>
6365 * cgraphunit.c (analyze_functions): Save input_location, set it
6366 to UNKNOWN_LOCATION and restore it at the end.
6368 2013-09-10 Martin Jambor <mjambor@suse.cz>
6370 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6371 represented by a thunk.
6373 2013-09-10 Jeff Law <law@redhat.com>
6375 PR tree-optimization/58343
6376 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6377 jump threading requests through loop headers buried in the middle
6378 of a jump threading path.
6380 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6381 in return value/type.
6383 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6385 PR rtl-optimization/58365
6386 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6387 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6390 2013-09-10 Richard Biener <rguenther@suse.de>
6392 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6393 * tree-data-ref.c (create_rdg_vertices): Collect all data
6394 references, signal failure to the caller, use data-ref API.
6395 (build_rdg): Compute data references only once. Maintain lifetime
6396 of data references and data dependences from within RDG.
6397 (free_rdg): Free dependence relations.
6398 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6399 inventing extra dependences.
6400 (distribute_loop): Update for RDG API changes.
6402 2013-09-10 Kai Tietz <ktietz@redhat.com>
6404 * doc/invoke.texi (fms-extensions): Document changed
6405 behavior for ms-abi targets.
6406 * config/i386/i386.c (ix86_option_override_internal):
6407 Set default value of option -fms-extension for ms-abi targets.
6409 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6411 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6413 2013-09-10 Alan Modra <amodra@gmail.com>
6416 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6418 2013-09-10 Alan Modra <amodra@gmail.com>
6420 * config/rs6000/predicates.md (add_cint_operand): New.
6421 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6422 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6423 using add_cint_operand.
6424 (largetoc_high_plus_aix): Likewise.
6426 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6428 PR tree-optimization/58364
6429 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6430 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6431 the current range can't be an unconditional true or false.
6433 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6435 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6437 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6439 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6441 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6444 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6446 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6448 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6449 NO_REGS for immediate that can't be moved directly into FP_REGS.
6451 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6453 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6454 comparison with negated operand.
6455 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6458 2013-09-09 Richard Biener <rguenther@suse.de>
6461 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6462 of a subloop record all its block as affecting loop-closed SSA form.
6464 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6466 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6467 of an rtx/bitpos pair.
6468 (store_fixed_bit_field): Update accordingly.
6470 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6472 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6474 * builtins.c (expand_errno_check): Likewise.
6475 * dwarf2cfi.c (init_return_column_size): Likewise.
6476 * except.c (sjlj_mark_call_sites): Likewise.
6477 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6478 * lra-constraints.c (emit_inc): Likewise.
6479 * ree.c (combine_set_extension): Likewise.
6480 * regmove.c (fixup_match_2): Likewise.
6481 * reload1.c (inc_for_reload): Likewise.
6483 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6485 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6486 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6487 Use gen_int_mode with the mode of the associated simplify_* call.
6488 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6489 * expmed.c (expand_shift_1): Likewise.
6490 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6491 * loop-iv.c (iv_number_of_iterations): Likewise.
6492 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6493 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6495 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6497 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6498 of the associated expand_* call.
6499 (asan_emit_stack_protection): Likewise.
6500 * builtins.c (round_trampoline_addr): Likewise.
6501 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6502 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6503 (emit_store_flag): Likewise.
6504 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6505 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6507 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6508 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6509 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6510 * modulo-sched.c (generate_prolog_epilog): Likewise.
6511 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6512 (expand_ctz, expand_ffs, expand_unop): Likewise.
6514 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6516 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6517 of the associated gen_rtx_* call.
6518 * caller-save.c (init_caller_save): Likewise.
6519 * combine.c (find_split_point, make_extraction): Likewise.
6520 (make_compound_operation): Likewise.
6521 * dwarf2out.c (mem_loc_descriptor): Likewise.
6522 * explow.c (plus_constant, probe_stack_range): Likewise.
6523 * expmed.c (expand_mult_const): Likewise.
6524 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6525 * reload1.c (init_reload): Likewise.
6526 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6527 * var-tracking.c (adjust_mems): Likewise.
6528 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6529 rather than gen_rtx_fmt_ee.
6531 2013-09-09 Jan Hubicka <jh@suse.cz>
6534 * value-prof.c (gimple_ic): Copy also abnormal edges.
6536 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6538 * asan.c (asan_shadow_cst): Use gen_int_mode.
6540 2013-09-08 Jan Hubicka <jh@suse.cz>
6542 * ipa-profile.c: Add toplevel comment.
6543 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6544 (contains_hot_call_p): New function.
6545 (ipa_propagate_frequency): Set frequencies based on counts when
6547 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6548 profile; do not tamper with profile after inlining if it is read.
6550 2013-09-08 Jan Hubicka <jh@suse.cz>
6552 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6555 2013-09-08 Jan Hubicka <jh@suse.cz>
6557 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6560 2013-09-08 Jeff Law <law@redhat.com>
6563 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6566 2013-09-08 Andi Kleen <ak@linux.intel.com>
6568 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6570 2013-09-08 Jan Hubicka <jh@suse.cz>
6572 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6574 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6576 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6577 for non-debug insns.
6578 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6579 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6580 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6583 2013-09-08 Jan Hubicka <jh@suse.cz>
6585 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6586 targets and devirtualize to BUILT_IN_UNREACHABLE.
6587 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6588 * ipa.c (walk_polymorphic_call_targets): New function.
6589 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6590 functions; use the new timevar.
6591 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6592 was removed from the program.
6593 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6594 not consider it in the walk when its vtable is dead.
6595 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6597 (devirt_variable_node_removal_hook): New function.
6598 (possible_polymorphic_call_targets): Also register
6599 devirt_variable_node_removal_hook.
6600 (ipa_devirt): Do not do non-speculative devirtualization.
6601 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6603 2013-09-08 Jan Hubicka <jh@suse.cz>
6605 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6606 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6607 varpool_remove_variable_insertion_hook): Declare.
6608 * varpool.c (varpool_node_hook_list): New structure.
6609 (first_varpool_node_removal_hook,
6610 first_varpool_variable_insertion_hook): New variables.
6611 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6612 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6613 varpool_remove_variable_insertion_hook,
6614 varpool_call_variable_insertion_hooks): New functions.
6615 (varpool_remove_node): Use it.
6617 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6620 * diagnostic.c (diagnostic_build_prefix): When s.file is
6621 "<built-in>" don't output line and column numbers.
6623 2013-09-06 Jan Hubicka <jh@suse.cz>
6625 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6626 * lto-streamer-out.c: Stream thunks, too.
6627 * lto-streamer-in.c (input_function): Pop cfun here
6628 (lto_read_body): Instead of here.
6630 2013-09-06 Caroline Tice <cmtice@google.com>
6632 * doc/install.texi: Add documentation for the --enable-vtable-verify
6633 and the --disable-libvtv configure options.
6635 2013-09-06 Jeff Law <law@redhat.com>
6637 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6638 edge implied equivalences into successor phis.
6640 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6642 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6644 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6646 * resource.c (mark_target_live_regs): Compute resources taking
6647 into account if a call is predicated or not.
6649 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6651 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6652 the compiler in the function names.
6654 2013-09-06 Jan Hubicka <jh@suse.cz>
6657 * ipa-inline.c (has_caller_p): New function.
6658 (want_inline_function_to_all_callers_p): Use it.
6659 (sum_callers, inline_to_all_callers): Break out from ...
6660 (ipa_inline): ... here.
6662 2013-09-06 Jan Hubicka <jh@suse.cz>
6664 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6665 only when AVX is enabled.
6667 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6669 * config/aarch64/aarch64.md
6670 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6671 is fpsimd_<load/store>2.
6672 (load_pair<mode>): Likewise.
6673 (store_pair<mode>): Likewise.
6675 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6677 * config/arm/types.md (type): Add "mrs" type.
6678 * config/aarch64/aarch64.md
6679 (aarch64_load_tp_hard): Make type "mrs".
6681 (load_tp_hard): Make type "mrs".
6682 * config/arm/cortex-a15.md: Update with new attributes.
6683 * config/arm/cortex-a5.md: Update with new attributes.
6684 * config/arm/cortex-a53.md: Update with new attributes.
6685 * config/arm/cortex-a7.md: Update with new attributes.
6686 * config/arm/cortex-a8.md: Update with new attributes.
6687 * config/arm/cortex-a9.md: Update with new attributes.
6688 * config/arm/cortex-m4.md: Update with new attributes.
6689 * config/arm/cortex-r4.md: Update with new attributes.
6690 * config/arm/fa526.md: Update with new attributes.
6691 * config/arm/fa606te.md: Update with new attributes.
6692 * config/arm/fa626te.md: Update with new attributes.
6693 * config/arm/fa726te.md: Update with new attributes.
6695 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6697 * config/aarch64/aarch64.md
6698 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6699 (*movtf_aarch64): Likewise.
6701 (thumb1_movdi_insn): Use "multiple" for type where more than one
6702 instruction is used for a move.
6703 (*arm32_movhf): Likewise.
6704 (*thumb_movdf_insn): Likewise.
6706 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6708 * config/arm/types.md (type): Rename fcpys to fmov.
6710 (*arm_movsi_vfp): Rename type fcpys as fmov.
6711 (*thumb2_movsi_vfp): Likewise
6712 (*movhf_vfp_neon): Likewise
6713 (*movhf_vfp): Likewise
6714 (*movsf_vfp): Likewise
6715 (*thumb2_movsf_vfp): Likewise
6716 (*movsfcc_vfp): Likewise
6717 (*thumb2_movsfcc_vfp): Likewise
6718 * config/aarch64/aarch64-simd.md
6719 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6720 * config/aarch64/aarch64.md
6721 (*movsi_aarch64): Replace type mov_reg with fmovs.
6722 (*movdi_aarch64): Likewise
6723 (*movsf_aarch64): Likewise
6724 (*movdf_aarch64): Likewise
6726 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6727 * config/arm/iwmmxt.md
6728 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6729 * config/arm/arm1020e.md: Update with new attributes.
6730 * config/arm/cortex-a15-neon.md: Update with new attributes.
6731 * config/arm/cortex-a5.md: Update with new attributes.
6732 * config/arm/cortex-a53.md: Update with new attributes.
6733 * config/arm/cortex-a7.md: Update with new attributes.
6734 * config/arm/cortex-a8-neon.md: Update with new attributes.
6735 * config/arm/cortex-a9.md: Update with new attributes.
6736 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6737 * config/arm/cortex-r4f.md: Update with new attributes.
6738 * config/arm/marvell-pj4.md: Update with new attributes.
6739 * config/arm/vfp11.md: Update with new attributes.
6741 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6743 * config/aarch64/aarch64.md
6744 (*madd<mode>): Fix type attribute.
6745 (*maddsi_uxtw): Likewise.
6746 (*msub<mode>): Likewise.
6747 (*msubsi_uxtw): Likewise.
6748 (<su_optab>maddsidi4): Likewise.
6749 (<su_optab>msubsidi4): Likewise.
6751 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6753 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6754 * config/arm/arm.md (core_cycles): Remove fdiv.
6755 * config/arm/vfp.md:
6756 (*sqrtsf2_vfp): Update for attribute changes.
6757 (*sqrtdf2_vfp): Likewise.
6758 * config/aarch64/aarch64.md:
6759 (sqrt<mode>2): Update for attribute changes.
6760 * config/arm/arm1020e.md: Update with new attributes.
6761 * config/arm/cortex-a15-neon.md: Update with new attributes.
6762 * config/arm/cortex-a5.md: Update with new attributes.
6763 * config/arm/cortex-a53.md: Update with new attributes.
6764 * config/arm/cortex-a7.md: Update with new attributes.
6765 * config/arm/cortex-a8-neon.md: Update with new attributes.
6766 * config/arm/cortex-a9.md: Update with new attributes.
6767 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6768 * config/arm/cortex-r4f.md: Update with new attributes.
6769 * config/arm/marvell-pj4.md: Update with new attributes.
6770 * config/arm/vfp11.md: Update with new attributes.
6772 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6774 * config/arm/types.md
6775 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6776 * config/aarch64/aarch64.md
6777 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6779 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6780 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6781 (float<GPI:mode><GPF:mode>2): Likewise.
6783 (*truncsisf2_vfp): Update with new attributes.
6784 (*truncsidf2_vfp): Likewise.
6785 (fixuns_truncsfsi2): Likewise.
6786 (fixuns_truncdfsi2): Likewise.
6787 (*floatsisf2_vfp): Likewise.
6788 (*floatsidf2_vfp): Likewise.
6789 (floatunssisf2): Likewise.
6790 (floatunssidf2): Likewise.
6791 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6792 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6793 * config/arm/arm1020e.md: Update with new attributes.
6794 * config/arm/cortex-a15-neon.md: Update with new attributes.
6795 * config/arm/cortex-a5.md: Update with new attributes.
6796 * config/arm/cortex-a53.md: Update with new attributes.
6797 * config/arm/cortex-a7.md: Update with new attributes.
6798 * config/arm/cortex-a8-neon.md: Update with new attributes.
6799 * config/arm/cortex-a9.md: Update with new attributes.
6800 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6801 * config/arm/cortex-r4f.md: Update with new attributes.
6802 * config/arm/marvell-pj4.md: Update with new attributes.
6803 * config/arm/vfp11.md: Update with new attributes.
6805 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6807 * config/aarch64/arm_neon.h
6808 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6809 (vqtbx<1,2,3,4><q>_s8): Likewise.
6811 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6813 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6814 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6815 (add<mode>3): Add type attribute.
6816 (add<mode>3): Likewise.
6817 (usadd<mode>3): Likewise.
6818 (ssadd<mode>3): Likewise.
6819 (sub<mode>3): Likewise.
6820 (sub<mode>3): Likewise.
6821 (ussub<mode>3): Likewise.
6822 (sssub<mode>3): Likewise.
6823 (ssmulsa3): Likewise.
6824 (usmulusa3): Likewise.
6825 (arm_usatsihi): Likewise.
6827 (*movdi_vfp): Add types for all instructions.
6828 (*movdi_vfp_cortexa8): Likewise.
6829 (*movhf_vfp_neon): Likewise.
6830 (*movhf_vfp): Likewise.
6831 (*movdf_vfp): Likewise.
6832 (*thumb2_movdf_vfp): Likewise.
6833 (*thumb2_movdfcc_vfp): Likewise.
6834 * config/arm/arm.md: Add type attribute to all insn patterns.
6835 (*thumb1_adddi3): Add type attribute.
6836 (*arm_adddi3): Likewise.
6837 (*adddi_sesidi_di): Likewise.
6838 (*adddi_zesidi_di): Likewise.
6839 (*thumb1_addsi3): Likewise.
6840 (addsi3_compare0): Likewise.
6841 (*addsi3_compare0_scratch): Likewise.
6842 (*compare_negsi_si): Likewise.
6843 (cmpsi2_addneg): Likewise.
6844 (*addsi3_carryin_<optab>): Likewise.
6845 (*addsi3_carryin_alt2_<optab>): Likewise.
6846 (*addsi3_carryin_clobercc_<optab>): Likewise.
6847 (*subsi3_carryin): Likewise.
6848 (*subsi3_carryin_const): Likewise.
6849 (*subsi3_carryin_compare): Likewise.
6850 (*subsi3_carryin_compare_const): Likewise.
6851 (*arm_subdi3): Likewise.
6852 (*thumb_subdi3): Likewise.
6853 (*subdi_di_zesidi): Likewise.
6854 (*subdi_di_sesidi): Likewise.
6855 (*subdi_zesidi_di): Likewise.
6856 (*subdi_sesidi_di): Likewise.
6857 (*subdi_zesidi_ze): Likewise.
6858 (thumb1_subsi3_insn): Likewise.
6859 (*arm_subsi3_insn): Likewise.
6860 (*anddi3_insn): Likewise.
6861 (*anddi_zesidi_di): Likewise.
6862 (*anddi_sesdi_di): Likewise.
6863 (*ne_zeroextracts): Likewise.
6864 (*ne_zeroextracts): Likewise.
6865 (*ite_ne_zeroextr): Likewise.
6866 (*ite_ne_zeroextr): Likewise.
6867 (*anddi_notdi_di): Likewise.
6868 (*anddi_notzesidi): Likewise.
6869 (*anddi_notsesidi): Likewise.
6870 (andsi_notsi_si): Likewise.
6871 (thumb1_bicsi3): Likewise.
6872 (*iordi3_insn): Likewise.
6873 (*iordi_zesidi_di): Likewise.
6874 (*iordi_sesidi_di): Likewise.
6875 (*thumb1_iorsi3_insn): Likewise.
6876 (*xordi3_insn): Likewise.
6877 (*xordi_zesidi_di): Likewise.
6878 (*xordi_sesidi_di): Likewise.
6879 (*arm_xorsi3): Likewise.
6880 (*andsi_iorsi3_no): Likewise.
6881 (*smax_0): Likewise.
6882 (*smax_m1): Likewise.
6883 (*arm_smax_insn): Likewise.
6884 (*smin_0): Likewise.
6885 (*arm_smin_insn): Likewise.
6886 (*arm_umaxsi3): Likewise.
6887 (*arm_uminsi3): Likewise.
6888 (*minmax_arithsi): Likewise.
6889 (*minmax_arithsi_): Likewise.
6890 (*satsi_<SAT:code>): Likewise.
6891 (arm_ashldi3_1bit): Likewise.
6892 (arm_ashrdi3_1bit): Likewise.
6893 (arm_lshrdi3_1bit): Likewise.
6894 (*arm_negdi2): Likewise.
6895 (*thumb1_negdi2): Likewise.
6896 (*arm_negsi2): Likewise.
6897 (*thumb1_negsi2): Likewise.
6898 (*negdi_extendsid): Likewise.
6899 (*negdi_zero_extend): Likewise.
6900 (*arm_abssi2): Likewise.
6901 (*thumb1_abssi2): Likewise.
6902 (*arm_neg_abssi2): Likewise.
6903 (*thumb1_neg_abss): Likewise.
6904 (one_cmpldi2): Likewise.
6905 (extend<mode>di2): Likewise.
6906 (*compareqi_eq0): Likewise.
6907 (*arm_extendhisi2addsi): Likewise.
6908 (*arm_movdi): Likewise.
6909 (*thumb1_movdi_insn): Likewise.
6910 (*arm_movt): Likewise.
6911 (*thumb1_movsi_insn): Likewise.
6912 (pic_add_dot_plus_four): Likewise.
6913 (pic_add_dot_plus_eight): Likewise.
6914 (tls_load_dot_plus_eight): Likewise.
6915 (*thumb1_movhi_insn): Likewise.
6916 (*thumb1_movsf_insn): Likewise.
6917 (*movdf_soft_insn): Likewise.
6918 (*thumb_movdf_insn): Likewise.
6919 (cbranchsi4_insn): Likewise.
6920 (cbranchsi4_scratch): Likewise.
6921 (*negated_cbranchsi4): Likewise.
6922 (*tbit_cbranch): Likewise.
6923 (*tlobits_cbranch): Likewise.
6924 (*tstsi3_cbranch): Likewise.
6925 (*cbranchne_decr1): Likewise.
6926 (*addsi3_cbranch): Likewise.
6927 (*addsi3_cbranch_scratch): Likewise.
6928 (*arm_cmpdi_insn): Likewise.
6929 (*arm_cmpdi_unsig): Likewise.
6930 (*arm_cmpdi_zero): Likewise.
6931 (*thumb_cmpdi_zero): Likewise.
6932 (*deleted_compare): Likewise.
6933 (*mov_scc): Likewise.
6934 (*mov_negscc): Likewise.
6935 (*mov_notscc): Likewise.
6936 (*cstoresi_eq0_thumb1_insn): Likewise.
6937 (cstoresi_nltu_thumb1): Likewise.
6938 (cstoresi_ltu_thu): Likewise.
6939 (thumb1_addsi3_addgeu): Likewise.
6940 (*arm_jump): Likewise.
6941 (*thumb_jump): Likewise.
6942 (*check_arch2): Likewise.
6943 (arm_casesi_internal): Likewise.
6944 (thumb1_casesi_dispatch): Likewise.
6945 (*arm_indirect_jump): Likewise.
6946 (*thumb1_indirect_jump): Likewise.
6948 (*and_scc): Likewise.
6949 (*ior_scc): Likewise.
6950 (*compare_scc): Likewise.
6951 (*cond_move): Likewise.
6952 (*cond_arith): Likewise.
6953 (*cond_sub): Likewise.
6954 (*cmp_ite0): Likewise.
6955 (*cmp_ite1): Likewise.
6956 (*cmp_and): Likewise.
6957 (*cmp_ior): Likewise.
6958 (*ior_scc_scc): Likewise.
6959 (*ior_scc_scc_cmp): Likewise.
6960 (*and_scc_scc): Likewise.
6961 (*and_scc_scc_cmp): Likewise.
6962 (*and_scc_scc_nod): Likewise.
6963 (*negscc): Likewise.
6964 (movcond_addsi): Likewise.
6965 (movcond): Likewise.
6966 (*ifcompare_plus_move): Likewise.
6967 (*if_plus_move): Likewise.
6968 (*ifcompare_move_plus): Likewise.
6969 (*if_move_plus): Likewise.
6970 (*ifcompare_arith_arith): Likewise.
6971 (*if_arith_arith): Likewise.
6972 (*ifcompare_arith_move): Likewise.
6973 (*if_arith_move): Likewise.
6974 (*ifcompare_move_arith): Likewise.
6975 (*if_move_arith): Likewise.
6976 (*ifcompare_move_not): Likewise.
6977 (*if_move_not): Likewise.
6978 (*ifcompare_not_move): Likewise.
6979 (*if_not_move): Likewise.
6980 (*ifcompare_shift_move): Likewise.
6981 (*if_shift_move): Likewise.
6982 (*ifcompare_move_shift): Likewise.
6983 (*if_move_shift): Likewise.
6984 (*ifcompare_shift_shift): Likewise.
6985 (*ifcompare_not_arith): Likewise.
6986 (*ifcompare_arith_not): Likewise.
6987 (*if_arith_not): Likewise.
6988 (*ifcompare_neg_move): Likewise.
6989 (*if_neg_move): Likewise.
6990 (*ifcompare_move_neg): Likewise.
6991 (*if_move_neg): Likewise.
6992 (prologue_thumb1_interwork): Likewise.
6993 (*cond_move_not): Likewise.
6994 (*sign_extract_onebit): Likewise.
6995 (*not_signextract_onebit): Likewise.
6996 (stack_tie): Likewise.
6997 (align_4): Likewise.
6998 (align_8): Likewise.
6999 (consttable_end): Likewise.
7000 (consttable_1): Likewise.
7001 (consttable_2): Likewise.
7002 (consttable_4): Likewise.
7003 (consttable_8): Likewise.
7004 (consttable_16): Likewise.
7005 (*thumb1_tablejump): Likewise.
7006 (prefetch): Likewise.
7007 (force_register_use): Likewise.
7008 (thumb_eh_return): Likewise.
7009 (load_tp_hard): Likewise.
7010 (load_tp_soft): Likewise.
7011 (tlscall): Likewise.
7012 (*arm_movtas_ze): Likewise.
7013 (*arm_rev): Likewise.
7014 (*arm_revsh): Likewise.
7015 (*arm_rev16): Likewise.
7016 * config/arm/thumb2.md
7017 (*thumb2_smaxsi3): Likewise.
7018 (*thumb2_sminsi3): Likewise.
7019 (*thumb32_umaxsi3): Likewise.
7020 (*thumb2_uminsi3): Likewise.
7021 (*thumb2_negdi2): Likewise.
7022 (*thumb2_abssi2): Likewise.
7023 (*thumb2_neg_abss): Likewise.
7024 (*thumb2_movsi_insn): Likewise.
7025 (tls_load_dot_plus_four): Likewise.
7026 (*thumb2_movhi_insn): Likewise.
7027 (*thumb2_mov_scc): Likewise.
7028 (*thumb2_mov_negs): Likewise.
7029 (*thumb2_mov_negs): Likewise.
7030 (*thumb2_mov_nots): Likewise.
7031 (*thumb2_mov_nots): Likewise.
7032 (*thumb2_movsicc_): Likewise.
7033 (*thumb2_movsfcc_soft_insn): Likewise.
7034 (*thumb2_indirect_jump): Likewise.
7035 (*thumb2_and_scc): Likewise.
7036 (*thumb2_ior_scc): Likewise.
7037 (*thumb2_ior_scc_strict_it): Likewise.
7038 (*thumb2_cond_move): Likewise.
7039 (*thumb2_cond_arith): Likewise.
7040 (*thumb2_cond_ari): Likewise.
7041 (*thumb2_cond_sub): Likewise.
7042 (*thumb2_negscc): Likewise.
7043 (*thumb2_movcond): Likewise.
7044 (thumb2_casesi_internal): Likewise.
7045 (thumb2_casesi_internal_pic): Likewise.
7046 (*thumb2_alusi3_short): Likewise.
7047 (*thumb2_mov<mode>_shortim): Likewise.
7048 (*thumb2_addsi_short): Likewise.
7049 (*thumb2_subsi_short): Likewise.
7050 (thumb2_addsi3_compare0): Likewise.
7051 (*thumb2_cbz): Likewise.
7052 (*thumb2_cbnz): Likewise.
7053 (*thumb2_one_cmplsi2_short): Likewise.
7054 (*thumb2_negsi2_short): Likewise.
7055 (*orsi_notsi_si): Likewise.
7056 * config/arm/arm1020e.md: Update with new attributes.
7057 * config/arm/arm1026ejs.md: Update with new attributes.
7058 * config/arm/arm1136jfs.md: Update with new attributes.
7059 * config/arm/arm926ejs.md: Update with new attributes.
7060 * config/arm/cortex-a15.md: Update with new attributes.
7061 * config/arm/cortex-a5.md: Update with new attributes.
7062 * config/arm/cortex-a53.md: Update with new attributes.
7063 * config/arm/cortex-a7.md: Update with new attributes.
7064 * config/arm/cortex-a8.md: Update with new attributes.
7065 * config/arm/cortex-a9.md: Update with new attributes.
7066 * config/arm/cortex-m4.md: Update with new attributes.
7067 * config/arm/cortex-r4.md: Update with new attributes.
7068 * config/arm/fa526.md: Update with new attributes.
7069 * config/arm/fa606te.md: Update with new attributes.
7070 * config/arm/fa626te.md: Update with new attributes.
7071 * config/arm/fa726te.md: Update with new attributes.
7073 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7075 * config/aarch64/aarch64-simd.md
7076 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
7077 <vwx> iterator to ensure correct register choice.
7078 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7079 (aarch64_sqdmull_n<mode>): Likewise.
7080 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7081 * config/aarch64/arm_neon.h
7082 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
7083 (vml<as><q>_n_<su>16): Likewise.
7084 (vml<as>l_high_lane<q>_<su>16): Likewise.
7085 (vml<as>l_high_n_<su>16): Likewise.
7086 (vml<as>l_lane<q>_<su>16): Likewise.
7087 (vml<as>l_n_<su>16): Likewise.
7088 (vmul<q>_lane<q>_<su>16): Likewise.
7089 (vmul<q>_n_<su>16): Likewise.
7090 (vmull_lane<q>_<su>16): Likewise.
7091 (vmull_n_<su>16): Likewise.
7092 (vmull_high_lane<q>_<su>16): Likewise.
7093 (vmull_high_n_<su>16): Likewise.
7094 (vqrdmulh<q>_n_s16): Likewise.
7096 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
7098 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
7099 have the correct lane parameter.
7101 2013-09-06 Richard Biener <rguenther@suse.de>
7103 * cfganal.c (control_dependences::~control_dependences):
7104 Properly free all of the vector.
7106 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
7109 * config/i386/i386.c (ix86_conditional_register_usage):
7110 Proper initialize extended SSE registers.
7112 2013-09-06 Jan Hubicka <jh@suse.cz>
7114 PR tree-optimization/58311
7115 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
7117 2013-09-06 Jan Hubicka <jh@suse.cz>
7119 * Makefile.in (tree-sra.o): Update dependencies.
7120 * tree-sra.c: Include ipa-utils.h
7121 (scan_function): Use recursive_call_p.
7122 (has_caller_p): New function.
7123 (cgraph_for_node_and_aliases): Count also callers of aliases.
7125 2013-09-06 Jan Hubicka <jh@suse.cz>
7128 * cgraph.h (symtab_semantically_equivalent_p): Declare.
7129 * tree-tailcall.c: Include ipa-utils.h.
7130 (find_tail_calls): Use it.
7131 * ipa-pure-const.c (check_call): Likewise.
7132 * ipa-utils.c (recursive_call_p): New function.
7133 * ipa-utils.h (recursive_call_p): Dclare.
7134 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
7135 (symtab_semantically_equivalent_p): New function.
7136 * Makefile.in (tree-tailcall.o): Update dependencies.
7138 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7140 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
7143 2013-09-06 Richard Biener <rguenther@suse.de>
7145 * lto-streamer.h (lto_global_var_decls): Remove.
7146 * Makefile.in (OBJS): Remove lto-symtab.o.
7147 (lto-symtab.o): Remove.
7148 (GTFILES): Remove lto-symtab.c
7149 * lto-symtab.c: Move to lto/
7151 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7153 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
7154 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
7155 (UNSPEC_FPINT_RINT): New constant definitions.
7156 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
7157 definition with 2 attributes.
7158 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
7159 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
7162 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7164 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
7165 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
7166 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
7167 attribute to "z196_alone".
7168 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
7171 2013-09-06 Richard Biener <rguenther@suse.de>
7173 * basic-block.h (class control_dependences): New.
7174 * tree-ssa-dce.c (control_dependence_map): Remove.
7176 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7177 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7178 find_pdom, find_control_dependence, find_all_control_dependences):
7180 (mark_control_dependent_edges_necessary,
7181 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7182 tree_dce_done, perform_tree_ssa_dce): Adjust.
7183 * cfganal.c (set_control_dependence_map_bit,
7184 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7185 find_all_control_dependences): Move from tree-ssa-dce.c and
7186 implement as methods of control_dependences class.
7187 (control_dependences::control_dependences): New.
7188 (control_dependences::~control_dependences): Likewise.
7189 (control_dependences::get_edges_dependent_on): Likewise.
7190 (control_dependences::get_edge): Likewise.
7192 2013-09-04 Jan Hubicka <jh@suse.cz>
7194 * tree.c (types_same_for_odr): Drop overactive check.
7195 * ipa-devirt.c (hash_type_name): Likewise.
7197 2013-09-04 Jan Hubicka <jh@suse.cz>
7199 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7200 (analyze_functions): ... here.
7202 2013-09-04 Jan Hubicka <jh@suse.cz>
7205 * cgraphunit.c (analyze_functions): Clear AUX fields
7206 after processing; initialize assembler name has.
7208 2013-09-05 Jeff Law <law@redhat.com>
7210 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7211 from thread_around_empty_block. Record threading path into PATH.
7212 Recurse if threading through the initial block is successful.
7213 (thread_across_edge): Corresponding changes to slightly simplify.
7215 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7217 * config/aarch64/aarch64.md
7218 (type): Remove frecpe, frecps, frecpx.
7219 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7220 fix to be a TARGET_SIMD instruction.
7221 (aarch64_frecps): Remove.
7222 * config/aarch64/aarch64-simd.md
7223 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7224 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7226 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7227 Sofiane Naci <sofiane.naci@arm.com>
7229 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7230 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7231 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7232 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7233 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7234 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7235 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7236 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7237 into "clz, "rbit". Rename "shift" to "shift_imm".
7238 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7239 changes. Update for attribute changes all occurrences of arlo_* and
7241 * config/arm/arm-fixed.md: Update for attribute changes
7242 all occurrences of arlo_* types.
7243 * config/arm/thumb2.md: Update for attribute changes all occurrences
7245 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7246 (cortexa7_older_only): Likewise.
7247 (cortexa7_younger): Likewise.
7248 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7249 (1020alu_shift_op): Likewise.
7250 (1020alu_shift_reg_op): Likewise.
7251 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7252 (alu_shift_op): Likewise.
7253 (alu_shift_reg_op): Likewise.
7254 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7255 (11_alu_shift_op): Likewise.
7256 (11_alu_shift_reg_op): Likewise.
7257 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7258 (9_alu_shift_reg_op): Likewise.
7259 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7261 (cortex_a15_alu_shift): Likewise.
7262 (cortex_a15_alu_shift_reg): Likewise.
7263 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7265 (cortex_a5_alu_shift): Likewise.
7266 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7268 (cortex_a53_alu_shift): Likewise.
7269 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7271 (cortex_a7_alu_reg): Likewise.
7272 (cortex_a7_alu_shift): Likewise.
7273 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7275 (cortex_a8_alu_shift): Likewise.
7276 (cortex_a8_alu_shift_reg): Likewise.
7277 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7278 (cortex_a9_dp_shift): Likewise.
7279 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7281 * config/arm/cortex-r4.md
7282 (cortex_r4_alu): Update for attribute changes.
7283 (cortex_r4_mov): Likewise.
7284 (cortex_r4_alu_shift_reg): Likewise.
7285 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7286 (526_alu_shift_op): Likewise.
7287 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7288 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7289 (626te_alu_shift_op): Likewise.
7290 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7291 (726te_alu_shift_op): Likewise.
7292 (726te_alu_shift_reg_op): Likewise.
7293 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7294 (mp626_alu_shift_op): Likewise.
7295 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7296 (pj4_alu_conds): Likewise.
7297 (pj4_shift): Likewise.
7298 (pj4_shift_conds): Likewise.
7299 (pj4_alu_shift): Likewise.
7300 (pj4_alu_shift_conds): Likewise.
7301 * config/aarch64/aarch64.md: Update for attribute change
7302 all occurrences of arlo_* and shift* types.
7304 2013-09-05 Mike Stump <mikestump@comcast.net>
7306 * tree.h: Move documentation for tree_function_decl to tree-core.h
7307 with the declaration.
7309 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7312 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7313 looking for widest mode.
7315 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7317 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7319 2013-09-05 Richard Biener <rguenther@suse.de>
7321 PR tree-optimization/58137
7322 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7323 Do not create vectors of pointers.
7324 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7325 types for the components of the vector initializer.
7326 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7327 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7329 2013-09-05 Martin Jambor <mjambor@suse.cz>
7331 * ipa-prop.c (remove_described_reference): Accept missing references,
7332 return false if that hppens, otherwise return true.
7333 (cgraph_node_for_jfunc): New function.
7334 (try_decrement_rdesc_refcount): Likewise.
7335 (try_make_edge_direct_simple_call): Use them.
7336 (ipa_edge_removal_hook): Remove references from rdescs.
7337 (ipa_edge_duplication_hook): Clone rdescs and their references
7338 when the new edge has the same caller as the old one.
7339 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7340 reference before removing any edges.
7342 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7344 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7346 * thumb2.md (thumb2_storewb_parisi): New pattern.
7348 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7350 * config/aarch64/aarch64-option-extensions.def: Add
7351 AARCH64_OPT_EXTENSION of 'crc'.
7352 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7353 (AARCH64_ISA_CRC): Ditto.
7354 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7355 description of the CRC extension.
7357 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7359 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7360 * config/rs6000/linux.h: Ditto.
7361 * alpha/linux.h: Ditto.
7362 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7363 no_c99_libc_has_function.
7364 * config/c6x/uclinux-elf.h: Ditto.
7365 * config/lm32/uclinux-elf.h: Ditto.
7366 * config/m68k/uclinux.h: Ditto.
7367 * config/moxie/uclinux.h: Ditto.
7368 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7369 (crisv32-*-linux*, cris-*-linux*): Ditto.
7370 * config/bfin/bfin.c: Include "tm_p.h".
7372 2013-09-05 Richard Biener <rguenther@suse.de>
7374 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7375 check for a definition without a basic-block.
7377 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7378 Sofiane Naci <sofiane.naci@arm.com>
7380 * config/aarch64/aarch64.md
7381 (*movti_aarch64): Rename r_2_f and f_2_r.
7382 (*movsf_aarch64): Likewise.
7383 (*movdf_aarch64): Likewise.
7384 (*movtf_aarch64): Likewise.
7385 (aarch64_movdi_<mode>low): Likewise.
7386 (aarch64_movdi_<mode>high): Likewise.
7387 (aarch64_mov<mode>high_di): Likewise.
7388 (aarch64_mov<mode>low_di): Likewise.
7389 (aarch64_movtilow_tilow): Likewise.
7390 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7391 values to config/arm/types.md
7392 (attribute "conds"): Update for attribute change.
7393 (anddi3_insn): Likewise.
7394 (iordi3_insn): Likewise.
7395 (xordi3_insn): Likewise.
7396 (one_cmpldi2): Likewise.
7397 * config/arm/types.md (type): Add Neon types.
7398 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7399 use "type" attribute.
7400 (movmisalign<mode>_neon_store): Likewise.
7401 (movmisalign<mode>_neon_load): Likewise.
7402 (vec_set<mode>_internal): Likewise.
7403 (vec_setv2di_internal): Likewise.
7404 (vec_extract<mode>): Likewise.
7405 (vec_extractv2di): Likewise.
7406 (add<mode>3_neon): Likewise.
7407 (adddi3_neon): Likewise.
7408 (sub<mode>3_neon): Likewise.
7409 (subdi3_neon): Likewise.
7410 (mul<mode>3_neon): Likewise.
7411 (mul<mode>3add<mode>_neon): Likewise.
7412 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7413 (fma<VCVTF:mode>4)): Likewise.
7414 (fma<VCVTF:mode>4_intrinsic): Likewise.
7415 (fmsub<VCVTF:mode>4)): Likewise.
7416 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7417 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7418 (ior<mode>3): Likewise.
7419 (and<mode>3): Likewise.
7420 (anddi3_neon): Likewise.
7421 (orn<mode>3_neon): Likewise.
7422 (orndi3_neon): Likewise.
7423 (bic<mode>3_neon): Likewise.
7424 (bicdi3_neon): Likewise.
7425 (xor<mode>3): Likewise.
7426 (one_cmpl<mode>2): Likewise.
7427 (abs<mode>2): Likewise.
7428 (neg<mode>2): Likewise.
7429 (umin<mode>3_neon): Likewise.
7430 (umax<mode>3_neon): Likewise.
7431 (smin<mode>3_neon): Likewise.
7432 (smax<mode>3_neon): Likewise.
7433 (vashl<mode>3): Likewise.
7434 (vashr<mode>3_imm): Likewise.
7435 (vlshr<mode>3_imm): Likewise.
7436 (ashl<mode>3_signed): Likewise.
7437 (ashl<mode>3_unsigned): Likewise.
7438 (neon_load_count): Likewise.
7439 (ashldi3_neon_noclobber): Likewise.
7440 (signed_shift_di3_neon): Likewise.
7441 (unsigned_shift_di3_neon): Likewise.
7442 (ashrdi3_neon_imm_noclobber): Likewise.
7443 (lshrdi3_neon_imm_noclobber): Likewise.
7444 (widen_ssum<mode>3): Likewise.
7445 (widen_usum<mode>3): Likewise.
7446 (quad_halves_<code>v4si): Likewise.
7447 (quad_halves_<code>v4sf): Likewise.
7448 (quad_halves_<code>v8hi): Likewise.
7449 (quad_halves_<code>v16qi): Likewise.
7450 (reduc_splus_v2di): Likewise.
7451 (neon_vpadd_internal<mode>): Likewise.
7452 (neon_vpsmin<mode>): Likewise.
7453 (neon_vpsmax<mode>): Likewise.
7454 (neon_vpumin<mode>): Likewise.
7455 (neon_vpumax<mode>): Likewise.
7456 (ss_add<mode>_neon): Likewise.
7457 (us_add<mode>_neon): Likewise.
7458 (ss_sub<mode>_neon): Likewise.
7459 (us_sub<mode>_neon): Likewise.
7460 (neon_vadd<mode>_unspec): Likewise.
7461 (neon_vaddl<mode>): Likewise.
7462 (neon_vaddw<mode>): Likewise.
7463 (neon_vhadd<mode>): Likewise.
7464 (neon_vqadd<mode>): Likewise.
7465 (neon_vaddhn<mode>): Likewise.
7466 (neon_vmul<mode>): Likewise.
7467 (neon_vmla<mode>): Likewise.
7468 (neon_vmlal<mode>): Likewise.
7469 (neon_vmls<mode>): Likewise.
7470 (neon_vmlsl<mode>): Likewise.
7471 (neon_vqdmulh<mode>): Likewise.
7472 (neon_vqdmlal<mode>): Likewise.
7473 (neon_vqdmlsl<mode>): Likewise.
7474 (neon_vmull<mode>): Likewise.
7475 (neon_vqdmull<mode>): Likewise.
7476 (neon_vsub<mode>_unspec): Likewise.
7477 (neon_vsubl<mode>): Likewise.
7478 (neon_vsubw<mode>): Likewise.
7479 (neon_vqsub<mode>): Likewise.
7480 (neon_vhsub<mode>): Likewise.
7481 (neon_vsubhn<mode>): Likewise.
7482 (neon_vceq<mode>): Likewise.
7483 (neon_vcge<mode>): Likewise.
7484 (neon_vcgeu<mode>): Likewise.
7485 (neon_vcgt<mode>): Likewise.
7486 (neon_vcgtu<mode>): Likewise.
7487 (neon_vcle<mode>): Likewise.
7488 (neon_vclt<mode>): Likewise.
7489 (neon_vcage<mode>): Likewise.
7490 (neon_vcagt<mode>): Likewise.
7491 (neon_vtst<mode>): Likewise.
7492 (neon_vabd<mode>): Likewise.
7493 (neon_vabdl<mode>): Likewise.
7494 (neon_vaba<mode>): Likewise.
7495 (neon_vabal<mode>): Likewise.
7496 (neon_vmax<mode>): Likewise.
7497 (neon_vmin<mode>): Likewise.
7498 (neon_vpaddl<mode>): Likewise.
7499 (neon_vpadal<mode>): Likewise.
7500 (neon_vpmax<mode>): Likewise.
7501 (neon_vpmin<mode>): Likewise.
7502 (neon_vrecps<mode>): Likewise.
7503 (neon_vrsqrts<mode>): Likewise.
7504 (neon_vqabs<mode>): Likewise.
7505 (neon_vqneg<mode>): Likewise.
7506 (neon_vcls<mode>): Likewise.
7507 (clz<mode>2): Likewise.
7508 (popcount<mode>2): Likewise.
7509 (neon_vrecpe): Likewise.
7510 (neon_vrsqrte): Likewise.
7511 (neon_vget_lane<mode>_sext_internal): Likewise.
7512 (neon_vget_lane<mode>_zext_internal): Likewise.
7513 (neon_vdup_n<mode>): Likewise.
7514 (neon_vdup_nv2di): Likewise.
7515 (neon_vdpu_lane<mode>_internal): Likewise.
7516 (neon_vswp<mode>): Likewise.
7517 (float<mode><V_cvtto>2): Likewise.
7518 (floatuns<mode><V_cvtto>2): Likewise.
7519 (fix_trunc<mode><V_cvtto>)2): Likewise
7520 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7521 (neon_vcvt<mode>): Likewise.
7522 (neon_vcvtv4sfv4hf): Likewise.
7523 (neon_vcvtv4hfv4sf): Likewise.
7524 (neon_vcvt_n<mode>): Likewise.
7525 (neon_vmovn<mode>): Likewise.
7526 (neon_vqmovn<mode>): Likewise.
7527 (neon_vqmovun<mode>): Likewise.
7528 (neon_vmovl<mode>): Likewise.
7529 (neon_vmul_lane<mode>): Likewise.
7530 (neon_vmull_lane<mode>): Likewise.
7531 (neon_vqdmull_lane<mode>): Likewise.
7532 (neon_vqdmulh_lane<mode>): Likewise.
7533 (neon_vmla_lane<mode>): Likewise.
7534 (neon_vmlal_lane<mode>): Likewise.
7535 (neon_vqdmlal_lane<mode>): Likewise.
7536 (neon_vmls_lane<mode>): Likewise.
7537 (neon_vmlsl_lane<mode>): Likewise.
7538 (neon_vqdmlsl_lane<mode>): Likewise.
7539 (neon_vext<mode>): Likewise.
7540 (neon_vrev64<mode>): Likewise.
7541 (neon_vrev32<mode>): Likewise.
7542 (neon_vrev16<mode>): Likewise.
7543 (neon_vbsl<mode>_internal): Likewise.
7544 (neon_vshl<mode>): Likewise.
7545 (neon_vqshl<mode>): Likewise.
7546 (neon_vshr_n<mode>): Likewise.
7547 (neon_vshrn_n<mode>): Likewise.
7548 (neon_vqshrn_n<mode>): Likewise.
7549 (neon_vqshrun_n<mode>): Likewise.
7550 (neon_vshl_n<mode>): Likewise.
7551 (neon_vqshl_n<mode>): Likewise.
7552 (neon_vqshlu_n<mode>): Likewise.
7553 (neon_vshll_n<mode>): Likewise.
7554 (neon_vsra_n<mode>): Likewise.
7555 (neon_vsri_n<mode>): Likewise.
7556 (neon_vsli_n<mode>): Likewise.
7557 (neon_vtbl1v8qi): Likewise.
7558 (neon_vtbl2v8qi): Likewise.
7559 (neon_vtbl3v8qi): Likewise.
7560 (neon_vtbl4v8qi): Likewise.
7561 (neon_vtbx1v8qi): Likewise.
7562 (neon_vtbx2v8qi): Likewise.
7563 (neon_vtbx3v8qi): Likewise.
7564 (neon_vtbx4v8qi): Likewise.
7565 (neon_vtrn<mode>_internal): Likewise.
7566 (neon_vzip<mode>_internal): Likewise.
7567 (neon_vuzp<mode>_internal): Likewise.
7568 (neon_vld1<mode>): Likewise.
7569 (neon_vld1_lane<mode>): Likewise.
7570 (neon_vld1_dup<mode>): Likewise.
7571 (neon_vld1_dupv2di): Likewise.
7572 (neon_vst1<mode>): Likewise.
7573 (neon_vst1_lane<mode>): Likewise.
7574 (neon_vld2<mode>): Likewise.
7575 (neon_vld2_lane<mode>): Likewise.
7576 (neon_vld2_dup<mode>): Likewise.
7577 (neon_vst2<mode>): Likewise.
7578 (neon_vst2_lane<mode>): Likewise.
7579 (neon_vld3<mode>): Likewise.
7580 (neon_vld3qa<mode>): Likewise.
7581 (neon_vld3qb<mode>): Likewise.
7582 (neon_vld3_lane<mode>): Likewise.
7583 (neon_vld3_dup<mode>): Likewise.
7584 (neon_vst3<mode>): Likewise.
7585 (neon_vst3qa<mode>): Likewise.
7586 (neon_vst3qb<mode>): Likewise.
7587 (neon_vst3_lane<mode>): Likewise.
7588 (neon_vld4<mode>): Likewise.
7589 (neon_vld4qa<mode>): Likewise.
7590 (neon_vld4qb<mode>): Likewise.
7591 (neon_vld4_lane<mode>): Likewise.
7592 (neon_vld4_dup<mode>): Likewise.
7593 (neon_vst4<mode>): Likewise.
7594 (neon_vst4qa<mode>): Likewise.
7595 (neon_vst4qb<mode>): Likewise.
7596 (neon_vst4_lane<mode>): Likewise.
7597 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7598 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7599 (neon_vec_<US>mult_lo_<mode>): Likewise.
7600 (neon_vec_<US>mult_hi_<mode>): Likewise.
7601 (neon_vec_<US>shiftl_<mode>): Likewise.
7602 (neon_unpack<US>_<mode>): Likewise.
7603 (neon_vec_<US>mult_<mode>): Likewise.
7604 (vec_pack_trunc_<mode>): Likewise.
7605 (neon_vec_pack_trunk_<mode>): Likewise.
7606 (neon_vabd<mode>_2): Likewise.
7607 (neon_vabd<mode>_3): Likewise.
7608 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7609 (thumb2_movsi_vfp): Likewise.
7610 (movdi_vfp): Likewise.
7611 (movdi_vfp_cortexa8): Likewise.
7612 (movhf_vfp_neon): Likewise.
7613 (movhf_vfp): Likewiwse.
7614 (movsf_vfp): Likewiwse.
7615 (thumb2_movsf_vfp): Likewiwse.
7616 (movdf_vfp): Likewise.
7617 (thumb2_movdf_vfp): Likewise.
7618 (movsfcc_vfp): Likewise.
7619 (thumb2_movsfcc_vfp): Likewise.
7620 (movdfcc_vfp): Likewise.
7621 (thumb2_movdfcc_vfp): Likewise.
7622 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7623 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7624 (v10_v2c): Likewise.
7625 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7627 (cortex_a15_neon_int_2): Likewise.
7628 (cortex_a15_neon_int_3): Likewise.
7629 (cortex_a15_neon_int_4): Likewise.
7630 (cortex_a15_neon_int_5): Likewise.
7631 (cortex_a15_neon_vqneg_vqabs): Likewise.
7632 (cortex_a15_neon_vmov): Likewise.
7633 (cortex_a15_neon_vaba): Likewise.
7634 (cortex_a15_neon_vaba_qqq): Likewise.
7635 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7636 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7637 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7638 scalar_64_32_long_scalar): Likewise.
7639 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7640 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7641 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7642 lotype_qdd_64_32_long): Likewise.
7643 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7644 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7645 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7646 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7647 (cortex_a15_neon_shift_1): Likewise.
7648 (cortex_a15_neon_shift_2): Likewise.
7649 (cortex_a15_neon_shift_3): Likewise.
7650 (cortex_a15_neon_vshl_ddd): Likewise.
7651 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7652 (cortex_a15_neon_vsra_vrsra): Likewise.
7653 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7654 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7655 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7656 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7657 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7658 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7659 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7660 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7661 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7662 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7663 (cortex_a15_neon_bp_simple): Likewise.
7664 (cortex_a15_neon_bp_2cycle): Likewise.
7665 (cortex_a15_neon_bp_3cycle): Likewise.
7666 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7667 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7668 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7669 (cortex_a15_neon_vld2_4_regs): Likewise.
7670 (cortex_a15_neon_vld3_vld4): Likewise.
7671 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7672 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7673 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7674 (cortex_a15_neon_vst3_vst4): Likewise.
7675 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7676 (cortex_a15_neon_vld3_vld4_lane" 10
7677 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7678 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7679 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7680 (cortex_a15_neon_ldm_2): Likewise.0
7681 (cortex_a15_neon_stm_2): Likewise.
7682 (cortex_a15_neon_mcr): Likewise.
7683 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7684 (cortex_a15_neon_mrc): Likewise.
7685 (cortex_a15_neon_mrrc): Likewise.
7686 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7688 (cortex_a15_alu_shift): Likewise.
7689 (cortex_a15_alu_shift_reg): Likewise.
7690 (cortex_a15_mult32): Likewise.
7691 (cortex_a15_mult64): Likewise.
7692 (cortex_a15_block): Likewise.
7693 (cortex_a15_branch): Likewise.
7694 (cortex_a15_load1): Likewise.
7695 (cortex_a15_load3): Likewise.
7696 (cortex_a15_store1): Likewise.
7697 (cortex_a15_store3): Likewise.
7698 (cortex_a15_call): Likewise.
7699 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7700 (cortex_a5_f2r): Likewise.
7701 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7703 (cortex_a53_f2r): Likewise.
7704 * config/arm/cortex-a7.md
7705 (cortex_a7_branch): Update for attribute change.
7706 (cortex_a7_call): Likewise.
7707 (cortex_a7_alu_imm): Likewise.
7708 (cortex_a7_alu_reg): Likewise.
7709 (cortex_a7_alu_shift): Likewise.
7710 (cortex_a7_mul): Likewise.
7711 (cortex_a7_load1): Likewise.
7712 (cortex_a7_store1): Likewise.
7713 (cortex_a7_load2): Likewise.
7714 (cortex_a7_store2): Likewise.
7715 (cortex_a7_load3): Likewise.
7716 (cortex_a7_store3): Likewise.
7717 (cortex_a7_load4): Likewise.
7718 (cortex_a7_store4): Likewise.
7719 (cortex_a7_fpalu): Likewise.
7720 (cortex_a7_fconst): Likewise.
7721 (cortex_a7_fpmuls): Likewise.
7722 (cortex_a7_neon_mul): Likewise.
7723 (cortex_a7_fpmacs): Likewise.
7724 (cortex_a7_neon_mla: Likewise.
7725 (cortex_a7_fpmuld: Likewise.
7726 (cortex_a7_fpmacd: Likewise.
7727 (cortex_a7_fpfmad: Likewise.
7728 (cortex_a7_fdivs: Likewise.
7729 (cortex_a7_fdivd: Likewise.
7730 (cortex_a7_r2f: Likewise.
7731 (cortex_a7_f2r: Likewise.
7732 (cortex_a7_f_flags: Likewise.
7733 (cortex_a7_f_loads: Likewise.
7734 (cortex_a7_f_loadd: Likewise.
7735 (cortex_a7_f_stores: Likewise.
7736 (cortex_a7_f_stored: Likewise.
7737 (cortex_a7_neon): Likewise.
7738 * config/arm/cortex-a8-neon.md
7739 (cortex_a8_neon_mrc): Update for attribute change.
7740 (cortex_a8_neon_mrrc): Likewise.
7741 (cortex_a8_neon_int_1): Likewise.
7742 (cortex_a8_neon_int_2): Likewise.
7743 (cortex_a8_neon_int_3): Likewise.
7744 (cortex_a8_neon_int_4): Likewise.
7745 (cortex_a8_neon_int_5): Likewise.
7746 (cortex_a8_neon_vqneg_vqabs): Likewise.
7747 (cortex_a8_neon_vmov): Likewise.
7748 (cortex_a8_neon_vaba): Likewise.
7749 (cortex_a8_neon_vaba_qqq): Likewise.
7750 (cortex_a8_neon_vsma): Likewise.
7751 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7752 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7753 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7754 long_scalar): Likewise.
7755 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7756 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7757 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7758 long_scalar_qdd_64_32_long): Likewise.
7759 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7760 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7761 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7762 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7763 (cortex_a8_neon_shift_1): Likewise.
7764 (cortex_a8_neon_shift_2): Likewise.
7765 (cortex_a8_neon_shift_3): Likewise.
7766 (cortex_a8_neon_vshl_ddd): Likewise.
7767 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7768 (cortex_a8_neon_vsra_vrsra): Likewise.
7769 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7770 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7771 (cortex_a8_neon_fp_vsum): Likewise.
7772 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7773 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7774 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7775 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7776 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7777 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7778 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7779 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7780 (cortex_a8_neon_bp_simple): Likewise.
7781 (cortex_a8_neon_bp_2cycle): Likewise.
7782 (cortex_a8_neon_bp_3cycle): Likewise.
7783 (cortex_a8_neon_ldr): Likewise.
7784 (cortex_a8_neon_str): Likewise.
7785 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7786 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7787 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7788 (cortex_a8_neon_vld2_4_regs): Likewise.
7789 (cortex_a8_neon_vld3_vld4): Likewise.
7790 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7791 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7792 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7793 (cortex_a8_neon_vst3_vst4): Likewise.
7794 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7795 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7796 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7797 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7798 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7799 (cortex_a8_neon_mcr): Likewise.
7800 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7801 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7802 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7804 (ca9_neon_mrrc): Likewise.
7805 (cortex_a9_neon_int_1): Likewise.
7806 (cortex_a9_neon_int_2): Likewise.
7807 (cortex_a9_neon_int_3): Likewise.
7808 (cortex_a9_neon_int_4): Likewise.
7809 (cortex_a9_neon_int_5): Likewise.
7810 (cortex_a9_neon_vqneg_vqabs): Likewise.
7811 (cortex_a9_neon_vmov): Likewise.
7812 (cortex_a9_neon_vaba): Likewise.
7813 (cortex_a9_neon_vaba_qqq): Likewise.
7814 (cortex_a9_neon_vsma): Likewise.
7815 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7816 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7817 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7818 long_scalar): Likewise.
7819 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7820 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7821 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7822 long_scalar_qdd_64_32_long): Likewise.
7823 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7824 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7825 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7826 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7827 (cortex_a9_neon_shift_1): Likewise.
7828 (cortex_a9_neon_shift_2): Likewise.
7829 (cortex_a9_neon_shift_3): Likewise.
7830 (cortex_a9_neon_vshl_ddd): Likewise.
7831 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7832 (cortex_a9_neon_vsra_vrsra): Likewise.
7833 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7834 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7835 (cortex_a9_neon_fp_vsum): Likewise.
7836 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7837 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7838 (cortex_a9_neon_fp_vmla_ddd): Likewise.
7839 (cortex_a9_neon_fp_vmla_qqq): Likewise.
7840 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
7841 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
7842 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7843 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7844 (cortex_a9_neon_bp_simple): Likewise.
7845 (cortex_a9_neon_bp_2cycle): Likewise.
7846 (cortex_a9_neon_bp_3cycle): Likewise.
7847 (cortex_a9_neon_ldr): Likewise.
7848 (cortex_a9_neon_str): Likewise.
7849 (cortex_a9_neon_vld1_1_2_regs): Likewise.
7850 (cortex_a9_neon_vld1_3_4_regs): Likewise.
7851 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7852 (cortex_a9_neon_vld2_4_regs): Likewise.
7853 (cortex_a9_neon_vld3_vld4): Likewise.
7854 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7855 (cortex_a9_neon_vst1_3_4_regs): Likewise.
7856 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
7857 (cortex_a9_neon_vst3_vst4): Likewise.
7858 (cortex_a9_neon_vld1_vld2_lane): Likewise.
7859 (cortex_a9_neon_vld3_vld4_lane): Likewise.
7860 (cortex_a9_neon_vst1_vst2_lane): Likewise.
7861 (cortex_a9_neon_vst3_vst4_lane): Likewise.
7862 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
7863 (cortex_a9_neon_mcr): Likewise.
7864 (cortex_a9_neon_mcr_2_mcrr): Likewise.
7865 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
7866 (cortex_a9_fps): Likewise.
7867 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
7869 (cortex_m4_fmuls): Likewise.
7870 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
7872 (cortex_r4_mrc): Likewise.
7873 * config/arm/iterators.md: Update comment referring to neon_type.
7874 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
7875 (iwmmxt_movsi_insn): Likewise.
7876 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
7878 (pj4_core_to_vfp): Likewise.
7879 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
7881 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
7882 (vfp_fstore): Likewise.
7883 * doc/md.texi: Change references to neon_type to refer to type.
7885 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7887 * tree.h (DECL_BUILT_IN): Fix typo in comment.
7889 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
7891 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
7894 2013-09-04 Easwaran Raman <eraman@google.com>
7897 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
7898 (build_and_add_sum): Use it.
7899 (appears_later_in_bb): Simplify code.
7901 2013-09-04 Teresa Johnson <tejohnson@google.com>
7903 * dumpfile.c (dump_finish): Don't close stderr/stdout.
7905 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
7907 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
7909 2013-09-04 Jan Hubicka <jh@suse.cz>
7911 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
7912 * ipa-devirt.c: Include diganostic.h
7913 (odr_type_d): Add types and types_set.
7914 (hash_type_name): Work for types with vtables during LTO.
7915 (odr_hasher::remove): Fix comment; destroy types_set.
7916 (add_type_duplicate): New function,
7917 (get_odr_type): Use it.
7918 (dump_type_inheritance_graph): Dump type duplicates.
7919 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
7921 * tree.c (types_same_for_odr): Give exact answers on types with
7924 2013-09-04 Dodji Seketeli <dodji@redhat.com>
7926 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7927 explaining their differences.
7929 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
7931 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
7933 2013-09-03 Jeff Law <law@redhat.com>
7935 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
7936 when not threading through a joiner block. Pass joiner/no joiner
7937 state to register_jump_thread.
7938 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
7939 state from argument rather than implying on path length.
7940 Dump the entire jump thread path into debugging dump.
7941 * tree-flow.h (register_jump_thread): Update prototype.
7943 2013-08-29 Xinliang David Li <davidxl@google.com>
7945 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7946 Remove a trivial gcc_assert.
7948 2013-08-29 Xinliang David Li <davidxl@google.com>
7950 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
7951 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
7952 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
7953 Delay base decl alignment adjustment.
7954 * tree-vectorizer.c (vect_destroy_datarefs): New function.
7955 * tree-vectorizer.h: New data structure.
7956 (set_dr_misalignment): New function.
7957 (dr_misalignment): Ditto.
7958 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
7959 (vectorizable_load): Ditto.
7960 (ensure_base_align): New function.
7961 (vectorize_loops): Add dbg_cnt support.
7962 (execute_vect_slp): Ditto.
7963 * dbgcnt.def: New debug counter.
7964 * Makefile: New dependency.
7966 2013-09-03 Meador Inge <meadori@codesourcery.com>
7970 2013-08-30 Meador Inge <meadori@codesourcery.com>
7972 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7974 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
7976 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
7977 function descriptor.
7979 2013-09-03 Richard Biener <rguenther@suse.de>
7981 * tree-affine.c (add_elt_to_tree): Fix association issue,
7982 avoid useless converts and make sure to always return a
7983 properly typed result.
7985 2013-09-03 Richard Biener <rguenther@suse.de>
7988 * fold-const.c (negate_expr_p): Fix division case.
7989 (negate_expr): Likewise.
7991 2013-09-03 Richard Biener <rguenther@suse.de>
7994 * tree-streamer-out.c: Include tm.h.
7995 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
7997 2013-09-03 Jan Hubicka <jh@suse.cz>
7999 * tree-profile.c (tree_profiling): Cleanup CFG when done.
8001 2013-09-03 Alan Modra <amodra@gmail.com>
8003 * config.gcc (powerpc*-*-linux*): Add support for little-endian
8004 multilibs to big-endian target and vice versa.
8005 * config/rs6000/t-linux64: Use := assignment on all vars.
8006 (MULTILIB_EXTRA_OPTS): Remove fPIC.
8007 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
8008 * config/rs6000/t-linux64le: New file.
8009 * config/rs6000/t-linux64bele: New file.
8010 * config/rs6000/t-linux64lebe: New file.
8012 2013-09-02 Jan Hubicka <jh@suse.cz>
8014 * ipa-inline-transform.c (inline_transform): Do not
8015 optimize_inline_calls when not optimizing.
8017 2013-09-02 Jan Hubicka <jh@suse.cz>
8019 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
8020 duplicated nodes for assembler names.
8021 * symtab.c (symtab_unregister_node): Do not attempt to unlink
8022 hard registers from assembler name hash.
8024 2013-09-02 Jan Hubicka <jh@suse.cz>
8026 * ipa-split.c (execute_split_functions): Split externally visible
8027 functions called once.
8029 2013-09-02 Martin Jambor <mjambor@suse.cz>
8032 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
8033 linked list. When finding the correct duplicate, also consider also
8034 the caller in additon to its inlined_to node.
8036 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
8038 * config/aarch64/aarch64-simd-builtins.def
8039 (dup_lane_scalar): Remove.
8040 * config/aarch64/aarch64-simd.md
8041 (aarch64_simd_dup): Add 'w->w' alternative.
8042 (aarch64_dup_lane<mode>): Allow for VALL.
8043 (aarch64_dup_lane_scalar<mode>): Remove.
8044 (aarch64_dup_lane_<vswap_width_name><mode>): New.
8045 (aarch64_get_lane_signed<mode>): Add w->w altenative.
8046 (aarch64_get_lane_unsigned<mode>): Likewise.
8047 (aarch64_get_lane<mode>): Likewise.
8048 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
8049 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
8050 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
8051 (VCON): Change container of V2SF.
8052 (vswap_width_name): Likewise.
8053 * config/aarch64/arm_neon.h
8054 (__aarch64_vdup_lane_any): New.
8055 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8056 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
8057 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8059 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
8062 * expr.c (emit_move_complex): Do not move complex FP values as parts if
8063 the source or the destination is a single hard register.
8065 2013-09-02 Richard Biener <rguenther@suse.de>
8068 * tree-scalar-evolution.c (instantiate_scev_name): Allow
8071 2013-09-02 Richard Biener <rguenther@suse.de>
8073 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
8074 arithmetic to sizetype.
8076 2013-09-02 Bin Cheng <bin.cheng@arm.com>
8078 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
8079 Find auto-increment use both before and after candidate.
8081 2013-09-02 Marek Polacek <polacek@redhat.com>
8083 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
8085 2013-09-01 Jan Hubicka <jh@suse.cz>
8087 * Makefile.in: Add ipa-profile.o
8088 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
8089 * cgraph.c (struct cgraph_propagate_frequency_data,
8090 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
8091 ipa-profile.c; replace cgraph_ by ipa_ prefix.
8092 * cgraph.h (cgraph_propagate_frequency): Remove.
8093 * ipa-inline-analysis.c: Include ipa-utils.h;
8094 drop duplicated cfgloop.h.
8095 (inline_update_callee_summaries): Update.
8096 * ipa-profile.c: New file.
8097 * ipa-utils.h (ipa_propagate_frequency): Declare.
8098 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
8099 data-streamer.h, value-prof.h.
8100 (symtab_remove_unreachable_nodes): Update profile.
8101 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
8102 account_time_size, cmp_counts, dump_histogram,
8103 ipa_profile_generate_summary, ipa_profile_write_summary,
8104 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
8105 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
8106 Move to ipa-profile.c.
8108 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
8110 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
8112 2013-09-01 Jan Hubicka <jh@suse.cz>
8114 * common.opt (fdevirtualize-speculatively): New function.
8115 * invoke.texi (fdevirtualize-speculatively): Document.
8116 * ipa-devirt.c: Include ipa-inline.h
8117 (likely_target_p): New function.
8118 (ipa_devirt): New function.
8119 (gate_ipa_devirt): New function.
8120 (pass_data_ipa_devirt): New static var.
8121 (pass_ipa_devirt): Likewise.
8122 (make_pass_ipa_devirt): New function.
8123 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
8124 (common_handle_option): Disable devirtualization when
8125 value range profiling is available.
8126 * passes.def (pass_ipa_devirt): Add.
8127 * timever.def (TV_IPA_DEVIRT): New timevar.
8128 * tree-pass.h (make_pass_ipa_devirt):
8130 2013-09-01 Iain Sandoe <iain@codesourcery.com>
8132 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
8133 include sanitize(undefined).
8135 2013-08-31 Diego Novillo <dnovillo@google.com>
8137 * Makefile.in (TREE_CORE_H): Define.
8139 (GTFILES): Add tree-core.h.
8140 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
8142 * tree-core.h: New file.
8143 Move all data structures, enum, typedefs, global
8144 declarations and constants from ...
8147 2013-08-31 Jan Hubicka <jh@suse.cz>
8149 * bulitins.c (expand_builtin): Do not early exit for gcov
8150 instrumented functions.
8152 2013-08-31 Marek Polacek <polacek@redhat.com>
8154 * ubsan.c: Include tm_p.h.
8156 2013-08-31 Jan Hubicka <jh@suse.cz>
8158 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
8161 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
8162 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
8163 * tree-cfg.c (verify_gimple_label): ... here.
8164 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
8165 (ipa_merge_profiles): New function.
8166 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
8167 (lto_input_function_body): Likewise.
8168 * ipa-utils.h (ipa_merge_profiles): Declare.
8169 * lto-streamer.h (lto_input_function_body): Update prototype.
8170 (emit_label_in_global_context_p): Remove.
8171 * lto-symtab.c: Include ipa-utils.h
8172 (lto_cgraph_replace_node): Use ipa_merge_profiles.
8174 2013-08-31 Jan Hubicka <jh@suse.cz>
8176 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8178 2013-08-31 Jan Hubicka <jh@suse.cz>
8180 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8182 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8184 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8185 "cmp" RTX before signed_comparison_operator check to account
8188 2013-08-30 Jan Hubicka <jh@suse.cz>
8190 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8191 (detect_type_change_1): Rename to ...
8192 (detect_type_change): ... this one; early return on non-polymorphic
8194 (detect_type_change_ssa): Add comp_type parameter; update
8195 use of detect_type_change.
8196 (compute_complex_assign_jump_func): Add param_type parameter;
8197 update use of detect_type_change_ssa.
8198 (compute_complex_ancestor_jump_func): Likewise.
8199 (ipa_get_callee_param_type): New function.
8200 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8201 update calls to the jump function computation functions.
8203 2013-08-30 Teresa Johnson <tejohnson@google.com>
8204 Steven Bosscher <steven@gcc.gnu.org>
8206 * cfgrtl.c (fixup_new_cold_bb): New routine.
8207 (commit_edge_insertions): Invoke fixup_partitions.
8208 (find_partition_fixes): New routine.
8209 (fixup_partitions): Ditto.
8210 (verify_hot_cold_block_grouping): Update comments.
8211 (rtl_verify_edges): Invoke find_partition_fixes.
8212 (rtl_verify_bb_pointers): Update comments.
8213 (rtl_verify_bb_layout): Ditto.
8214 * basic-block.h (probably_never_executed_edge_p): Declare.
8215 (fixup_partitions): Ditto.
8216 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8217 * bb-reorder.c (sanitize_hot_paths): New function.
8218 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8220 * predict.c (probably_never_executed_edge_p): New routine.
8221 * cfg.c (check_bb_profile): Add partition insanity warnings.
8223 2013-08-30 Meador Inge <meadori@codesourcery.com>
8225 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8227 2013-08-30 Marek Polacek <polacek@redhat.com>
8229 * Makefile.in (ubsan.o): Add.
8230 (c-family/c-ubsan.o): Add.
8231 (builtins.o): Add ubsan.h dependency.
8232 * ubsan.h: New file.
8233 * ubsan.c: New file.
8234 * common.opt: Add -fsanitize=undefined option.
8235 (flag_sanitize): Add variable.
8236 (fsanitize=): Add option. Add Driver.
8237 (fsanitize=thread): Remove option.
8238 (fsanitize=address): Likewise.
8239 (static-libubsan): New option.
8240 * doc/invoke.texi: Document the new flag and -static-libubsan.
8241 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8242 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8243 * builtin-attrs.def (ATTR_COLD): Define.
8244 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8245 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8246 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8247 * flag-types.h (sanitize_code): New enum.
8248 * opts.c (common_handle_option): Parse command line arguments
8249 of -fsanitize=. Add -fsanitize=unreachable option.
8250 * varasm.c (get_variable_section): Adjust.
8251 (assemble_noswitch_variable): Likewise.
8252 (assemble_variable): Likewise.
8253 (output_constant_def_contents): Likewise.
8254 (categorize_decl_for_section): Likewise.
8255 (place_block_symbol): Likewise.
8256 (output_object_block): Likewise.
8257 * builtins.def: Likewise.
8258 * toplev.c (compile_file): Likewise.
8259 (process_options): Likewise.
8260 * cppbuiltin.c: Likewise.
8261 * tsan.c (tsan_pass): Likewise.
8262 (tsan_gate): Likewise.
8263 (tsan_gate_O0): Likewise.
8264 * cfgexpand.c (partition_stack_vars): Likewise.
8265 (expand_stack_vars): Likewise.
8266 (defer_stack_allocation): Likewise.
8267 (expand_used_vars): Likewise.
8268 * cfgcleanup.c (old_insns_match_p): Likewise.
8269 * asan.c (asan_finish_file): Likewise.
8270 (asan_instrument): Likewise.
8271 (gate_asan): Likewise.
8272 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8273 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8274 (asan_global_struct): Use pointer_sized_int_node instead
8275 calling build_nonstandard_integer_type.
8276 (initialize_sanitizer_builtins): Likewise.
8277 (asan_finish_file): Likewise.
8278 * gcc.c: Document %{%:function(args):X}.
8279 (static_spec_functions): Add sanitize.
8280 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8281 store funcval != NULL there.
8282 (do_spec_1): Adjust handle_spec_function caller.
8283 (handle_braces): Allow %:function(args) as condition.
8284 (sanitize_spec_function): New function.
8285 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8286 (LIBUBSAN_SPEC): Likewise.
8287 (LIBUBSAN_EARLY_SPEC): Likewise.
8288 (SANITIZER_SPEC): Handle libubsan.
8289 (SANITIZER_EARLY_SPEC): Likewise.
8290 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8291 instead of fsanitize=address.
8292 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8293 instead of fsanitize=address*.
8294 * builtins.c: Include ubsan.h.
8295 (fold_builtin_0): Instrument __builtin_unreachable.
8296 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8297 instead of flag_asan.
8298 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8299 (pointer_sized_int_node): Define.
8300 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8302 2013-08-30 Mike Stump <mikestump@comcast.net>
8304 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8307 2013-08-29 Jan Hubicka <jh@suse.cz>
8309 * cgraph.c (cgraph_function_body_availability): Handle weakref
8311 * passes.def: Remove pass_fixup_cfg.
8312 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8313 track when we need to remove functions.
8314 (gate_ipa_inline): Execute inlining always; add comment why.
8315 (pass_data_ipa_inline): Remove TODO_remove_functions.
8316 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8317 do not produce summaries.
8318 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8319 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8320 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8321 constant pool and vtable.
8323 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8325 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8326 New arm_neon.h's internal macros to specify 64-bit constants.
8327 Avoid using stdint.h's macros.
8329 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8331 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8332 before applying REG_P.
8334 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8336 PR tree-optimization/58277
8337 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8338 after seeing too many stmts with vdef in between dombb and current
8339 bb, invalidate everything.
8341 2013-08-30 Richard Biener <rguenther@suse.de>
8343 * fold-const.c (fold_single_bit_test): Fix overflow test.
8345 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8347 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8348 and which can live in a register, always retrieve the value on entry.
8349 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8350 passed by invisible reference specially.
8351 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8352 (vt_add_function_parameter): Correctly deal with a parameter passed by
8353 invisible reference.
8355 2013-08-30 Jan Hubicka <jh@suse.cz>
8357 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8359 2013-08-30 Richard Biener <rguenther@suse.de>
8361 PR tree-optimization/58228
8362 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8363 allow invariant loads in nested loop vectorization.
8365 2013-08-30 Richard Biener <rguenther@suse.de>
8367 PR tree-optimization/58223
8368 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8369 (has_anti_or_output_dependence): ... this and adjust to also
8370 look for output dependences.
8371 (mark_nodes_having_upstream_mem_writes): Adjust.
8372 (rdg_flag_uses): Likewise.
8374 2013-08-30 Richard Biener <rguenther@suse.de>
8376 PR tree-optimization/58010
8377 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8378 assert that we have a loop-closed PHI.
8380 2013-08-29 Jan Hubicka <jh@suse.cz>
8382 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8383 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8384 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8385 (lto_free_function_in_decl_state_for_node): New function.
8387 2013-08-29 Xinliang David Li <davidxl@google.com>
8389 * loop-unroll.c (report_unroll_peel): Minor message change.
8390 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8391 Emit alignment peeling message with default -fopt-info.
8392 (vect_loop_versioning): Emit loop version info message.
8393 * tree-vectorizer.c (vectorize_loops): Minor message change.
8394 (execute_vect_slp): Ditto.
8396 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8398 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8399 of the clone from the DECL_NAME of the original function.
8401 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8403 * passes.c (register_pass): Add overload.
8404 * tree-pass.h (register_pass): Forward declare it. Add comment.
8406 2013-08-29 Jan Hubicka <jh@suse.cz>
8408 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8409 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8410 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8411 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8412 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8413 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8414 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8415 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8417 2013-08-29 Teresa Johnson <tejohnson@google.com>
8419 * dumpfile.c (dump_loc): Output column number.
8420 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8421 * doc/invoke.texi: Document optall -fopt-info flag.
8422 * profile.c (read_profile_edge_counts): Use new dump framework.
8423 (compute_branch_probabilities): Ditto.
8424 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8425 when pass not in any opt group.
8426 * pass_manager.h (pass_manager::get_pass_profile): New method.
8427 * value-prof.c (check_counter): Use new dump framework.
8428 (check_ic_target): Ditto.
8429 * coverage.c (get_coverage_counts): Ditto.
8430 (coverage_init): Setup new dump framework.
8432 2013-08-29 Richard Biener <rguenther@suse.de>
8434 PR tree-optimization/58246
8435 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8436 handle the dominance check inside a basic-block.
8438 2013-08-29 Richard Biener <rguenther@suse.de>
8441 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8442 of default defs that appear in abnormal PHI nodes.
8444 2013-08-29 Richard Biener <rguenther@suse.de>
8446 PR tree-optimization/57685
8447 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8448 single-use operands to avoid exponential complexity.
8450 2013-08-28 Dehao Chen <dehao@google.com>
8452 * ipa-inline.c (edge_badness): Fix integer underflow.
8454 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8456 * gtm-builtins.def (_ITM_free): Declare leaf.
8458 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8461 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8462 (*tls_local_dynamic_base_64_largepic): Likewise.
8463 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8464 Remove predicate from call operand.
8465 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8466 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8468 2013-08-28 Jeff Law <law@redhat.com>
8470 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8471 checks for the number of predecessors and successors allowed.
8472 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8473 which require copying a joiner block if there is a request which
8474 is a subpath that requires no joiner block copying.
8476 2013-08-28 Jan Hubicka <jh@suse.cz>
8478 * lto-streamer-out.c (DFS_write_tree_body): Drop
8479 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8480 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8481 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8482 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8483 Do not read DECL_ERROR_ISSUED.
8484 (unpack_ts_decl_with_vis_value_fields): Do not read
8486 (lto_input_ts_binfo_tree_pointers): Do not read
8487 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8488 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8489 write DECL_ERROR_ISSUED..
8490 (pack_ts_decl_with_vis_value_fields): Do not write
8492 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8493 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8494 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8495 * tree.h (tree_decl_common): Update comment.
8496 (DECL_ERROR_ISSUED): Remove.
8498 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8501 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8503 2013-08-28 Jan Hubicka <jh@suse.cz>
8505 * builtins.def (free): Declare leaf.
8507 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8510 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8511 * configure: Regenerate.
8513 2013-08-27 Martin Jambor <mjambor@suse.cz>
8515 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8516 (ipa_ancestor_jf_data): Likewise.
8517 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8518 (ipa_get_jf_pass_through_type_preserved): New function.
8519 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8520 (ipa_get_jf_ancestor_type_preserved): New function.
8521 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8522 (ipa_get_jf_ancestor_result): Likewise.
8523 (propagate_vals_accross_pass_through): Use
8524 ipa_get_jf_pass_through_result to do all the value mappings.
8525 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8526 type_preserved flag.
8527 (ipa_set_jf_cst_copy): New function.
8528 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8529 (ipa_set_jf_arith_pass_through): Likewise.
8530 (ipa_set_ancestor_jf): Likewise.
8531 (compute_complex_assign_jump_func): Set type_preserved instead of
8533 (ipa_compute_jump_functions_for_edge): Likewise.
8534 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8535 (update_jump_functions_after_inlining): Update type_preserved.
8536 Explicitely create jump functions when combining one with pass_through.
8537 (ipa_write_jump_function): Stream the type_preserved flags.
8538 (ipa_read_jump_function): Likewise.
8540 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8541 Aldy Hernandez <aldyh@redhat.com>
8543 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8544 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8545 * function.h (struct function): Add has_force_vect_loops and
8547 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8548 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8550 * gimple.def: Update CLAUSES comments.
8551 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8552 (gimple_build_omp_for): Add argument to prototype.
8553 (gimple_omp_for_kind): New.
8554 (gimple_omp_for_set_kind): New.
8555 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8556 GOVD_DATA_SHARE_CLASS.
8557 (enum omp_region_type): Add ORT_SIMD.
8558 (gimple_add_tmp_var): Handle ORT_SIMD.
8559 (gimplify_var_or_parm_decl): Same.
8560 (is_gimple_stmt): Same.
8561 (omp_firstprivatize_variable): Same.
8562 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8563 (omp_notice_variable): Handle ORT_SIMD.
8564 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8565 (omp_check_private): Handle ORT_SIMD.
8566 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8568 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8569 Handle OMP_CLAUSE_LASTPRIVATE.
8570 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8572 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8573 (gimplify_expr): Handle OMP_SIMD.
8574 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8575 (expand_GOMP_SIMD_VF): New.
8576 (expand_GOMP_SIMD_LAST_LANE): New.
8577 * internal-fn.def (GOMP_SIMD_LANE): New.
8578 (GOMP_SIMD_VF): New.
8579 (GOMP_SIMD_LAST_LANE): New.
8580 * omp-low.c: Include target.h.
8581 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8583 (check_omp_nesting_restrictions): Same.
8585 (lower_rec_simd_input_clauses): New.
8586 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8588 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8589 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8590 (expand_omp_build_assign): New.
8591 (expand_omp_for_init_counts): New.
8592 (expand_omp_for_init_vars): New.
8593 (extract_omp_for_update_vars): New.
8594 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8595 and rewrite accordingly.
8596 (expand_omp_simd): New.
8597 (expand_omp_for): Use expand_omp_simd.
8598 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8599 (lower_omp_for): Do not lower the body.
8600 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8602 * tree-flow.h (find_omp_clause): Remove prototype.
8603 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8604 forcing vectorization of the loop, or if flag_tree_vectorize.
8605 (gate_tree_if_conversion): Similarly.
8606 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8607 gimple_build_omp_for.
8608 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8609 * tree-parloops (create_parallel_loop): Pass kind argument to
8610 gimple_build_omp_for.
8611 * tree-pretty-print.c (dump_omp_clause): Add cases for
8612 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8613 OMP_CLAUSE__SIMDUID_.
8614 (dump_generic_node): Handle OMP_SIMD.
8615 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8616 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8617 unroll OMP_SIMD loops here.
8618 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8619 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8621 (vect_analyze_data_refs): Handle simd loops.
8622 * tree-vect-loop.c (vectorizable_live_operation): Handle
8624 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8625 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8626 (vectorizable_load): Same.
8627 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8628 (struct simduid_to_vf): New.
8629 (simduid_to_vf::hash): New.
8630 (simduid_to-vf::equal): New.
8631 (struct simd_array_to_simduid): New.
8632 (simd_array_to_simduid::hash): New.
8633 (simd_array_to_simduid::equal): New.
8634 (adjust_simduid_builtins): New.
8635 (struct note_simd_array_uses_struct): New.
8636 (note_simd_array_uses_cb): New.
8637 (note_simd_array_uses): New.
8638 (vectorize_loops): Handle simd hints and adjust simd builtins
8640 * tree-vectorizer.h (struct _stmt_vec_info): Add
8641 simd_lane_access_p field.
8642 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8643 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8644 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8645 (omp_clause_code_name): Same.
8646 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8647 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8648 * tree.def (OMP_SIMD): New entry.
8649 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8650 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8651 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8652 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8653 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8654 (OMP_CLAUSE_LINEAR_STEP): New.
8655 (OMP_CLAUSE_SAFELEN_EXPR): New.
8656 (OMP_CLAUSE__SIMDUID__DECL): New.
8657 (find_omp_clause): New prototype.
8659 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8661 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8662 Haswell processor detection.
8664 2013-08-27 Christian Widmer <shadow@umbrox.de>
8667 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8668 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8669 AVX2 capable processors.
8671 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8673 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8674 of vget_low_* with implementations in terms of other intrinsics.
8676 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8679 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8680 values to -1, 0 and 1.
8682 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8684 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8685 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8686 (<optab><mode>3_insn): Remove.
8687 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8688 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8689 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8690 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8691 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8692 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8693 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8694 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8695 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8696 (ror<mode>3_insn): Likewise.
8697 * config/aarch64/predicates.md (aarch64_simd_register): New.
8699 2013-08-27 Richard Biener <rguenther@suse.de>
8701 PR tree-optimization/57521
8702 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8703 one edge is non-critical.
8704 (find_phi_replacement_condition): Make sure to use a non-critical
8705 edge. Cleanup and remove old bug workarounds.
8706 (bb_postdominates_preds): Remove.
8707 (if_convertible_loop_p_1): Do not compute post-dominators.
8708 (combine_blocks): Do not free post-dominators.
8709 (main_tree_if_conversion): Likewise.
8710 (pass_data_if_conversion): Add TODO_verify_ssa.
8712 2013-08-27 DJ Delorie <dj@redhat.com>
8714 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8716 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8718 * function.c (assign_parm_find_data_types): Set passed_mode and
8719 nominal_mode to the TYPE_MODE of nominal_type for the built
8720 pointer type in case of the struct-pass-by-reference.
8722 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8724 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8725 (UINT16_TYPE): Change default to "unsigned int".
8727 * config/avr/avr.opt (mfract-convert-truncate): New option.
8728 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8729 is set, round negative fractional integers according to n1169
8730 when converting to integer types.
8732 2013-08-26 Jan Hubicka <jh@suse.cz>
8734 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8735 methods can not be called indirectly when their address is not taken.
8737 2013-08-26 Jan Hubicka <jh@suse.cz>
8739 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8742 2013-08-26 Jan Hubicka <jh@suse.cz>
8744 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8747 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8749 * reload.c (find_valid_class): Allow classes that do not include
8750 FIRST_PSEUDO_REGISTER - 1.
8752 2013-08-26 Jan Hubicka <jh@suse.cz>
8754 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8755 fix edge count/frequency when speculation failed; fix type check
8756 for the direct call.
8758 2013-08-26 Jan Hubicka <jh@suse.cz>
8760 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8762 2013-08-26 Jan Hubicka <jh@suse.cz>
8764 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8765 to be changed by edge redirection.
8767 2013-08-26 Jan Hubicka <jh@suse.cz>
8769 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8770 formating; add sanity check.
8771 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8772 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8773 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8774 (resolve_noninline_speculation): Update callee keys, too.
8776 2013-08-26 Jan Hubicka <jh@suse.cz>
8778 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8779 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8781 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8783 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8786 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8788 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8789 (x86_64_elf_select_section): Ditto.
8790 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8792 (ix86_pass_by_reference): Ditto.
8793 (output_set_got): Ditto.
8794 (ix86_unary_operator_ok): Ditto.
8795 (ix86_expand_builtin): Ditto.
8797 2013-08-23 Jan Hubicka <jh@suse.cz>
8799 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8801 2013-08-23 Jan Hubicka <jh@suse.cz>
8803 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8804 (tree_decl_with_vis): Add FINAL field.
8806 2013-08-23 Jeff Law <law@redhat.com>
8808 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8809 the debugging dump when the expression is fully redundant.
8811 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8813 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8814 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8815 * pretty-print.c (pp_formatted_text_data): Likewise.
8816 (pp_write_text_to_stream): Likewise.
8817 (pp_write_text_as_dot_label_to_stream): Likewise.
8818 (pp_append_r): Likewise.
8819 (pp_format): Likewise.
8820 (pp_flush): Likewise.
8821 (pp_clear_output_area): Likewise.
8822 (pp_append_text): Likewise.
8823 (pp_formatted_text): Likewise.
8824 (pp_remaining_character_count_for_line): Likewise.
8825 (pp_newline): Likewise.
8826 (pp_character): Likewise.
8827 (output_buffer::~output_buffer): Define.
8828 (pretty_printer::~pretty_printer): Destruct output buffer.
8829 * pretty-print.h (output_buffer::~output_buffer): Declare.
8830 (pretty_printer::~pretty_printer): Declare virtual.
8832 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8835 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8836 HOST_WIDE_INT_M1U): New macros.
8837 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8838 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
8839 unsigned -1 for lshift.
8840 * cse.c (cse_insn): Likewise.
8841 * double-int.c (rshift_double, lshift_double): Likewise.
8842 * builtins.c (fold_builtin_bitop): Likewise.
8843 * combine.c (force_to_mode): Likewise.
8844 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
8845 * simplify-rtx.c (simplify_const_unary_operation,
8846 simplify_const_binary_operation): Likewise.
8847 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
8848 check_va_list_escapes): Likewise.
8849 * rtlanal.c (nonzero_bits1): Likewise.
8850 * expmed.c (expand_smod_pow2): Likewise.
8851 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
8853 2013-08-23 Jan Hubicka <jh@suse.cz>
8855 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
8856 as having address taken.
8858 2013-08-23 Jan Hubicka <jh@suse.cz>
8860 * ipa-utils.h (method_class_type): Declare.
8861 * ipa-devirt.c (method_class_type): Export.
8863 * cgraphunit.c (analyze_functions): Do basic devirtualization;
8864 do not walk base classes of anonymous types.
8866 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
8868 PR rtl-optimization/58220
8870 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
8871 handle SEQUENCE insns properly.
8873 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8875 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
8877 (pp_newline_and_indent): Likewise.
8878 (pp_separate_with): Likewise.
8879 * pretty-print.c (pp_newline_and_flush): Define.
8880 (pp_newline_and_indent): Likewise.
8881 (pp_separate_with): Likewise.
8883 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8886 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
8887 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
8889 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
8891 * config/i386/predicates.md (ext_sse_reg_operand): New.
8892 * config/i386/i386.md (*movti_internal): Use
8893 predicate to determine if EVEX is needed.
8894 (*movsi_internal): Ditto.
8895 (*movdf_internal): Ditto.
8896 (*movsf_internal): Ditto.
8897 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
8899 2013-08-23 Jakub Jelinek <jakub@redhat.com>
8901 PR tree-optimization/58209
8902 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
8903 (find_tail_calls): Give up for pointer result types if m is non-NULL.
8904 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
8905 emit POINTER_PLUS_EXPR.
8906 (create_tailcall_accumulator): For pointer result type accumulate in
8909 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
8911 * configure.ac: Add backslashes missing from the last change.
8912 * configure: Regenerate.
8914 2013-08-22 Jan Hubicka <jh@susue.cz>
8916 * ipa.c (function_and_variable_visibility): First remember function
8917 was global and then make it local.
8919 2013-08-22 Julian Brown <julian@codesourcery.com>
8921 * configure.ac: Add aarch64 to list of arches which use "nop" in
8923 * configure: Regenerate.
8925 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
8927 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
8928 gnu_libc_has_function.
8929 * config/s390/tpf.h: Likewise.
8931 2013-08-22 Jan Hubicka <jh@susue.cz>
8933 * timevar.c (validate_phases): Add cast.
8935 2013-08-22 Jan Hubicka <jh@susue.cz>
8937 * timevar.c (validate_phases): Use size_t for memory.
8938 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
8940 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
8942 * pretty-print.h (output_buffer::output_buffer): Declare.
8943 (pretty_printer::pretty_printer): Likewise.
8944 (pp_construct): Remove.
8945 * pretty-print.c (output_buffer::output_buffer): Define.
8946 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
8947 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
8948 (print_gimple_expr): Likewise.
8949 (print_gimple_seq): Likewise.
8950 (gimple_dump_bb): Likewise.
8951 * sched-vis.c (dump_value_slim): Likewise.
8952 (dump_insn_slim): Likewise.
8953 (dump_rtl_slim): Likewise.
8954 (str_pattern_slim): Likewise.
8955 * tree-mudflap.c (mf_varname_tree): Likewise.
8956 * graph.c (print_graph_cfg): Likewise.
8957 (start_graph_dump): Likewise.
8958 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
8960 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
8961 pretty printer initialization.
8962 * coretypes.h (diagnostic_context): Remove superflous type alias
8964 (pretty_printer): Likewise. Declare directly as a class.
8965 (pretty_print_info): Remove declaration as class.
8966 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
8967 and pp_clear_output_area.
8968 (asan_add_global): Likewise.
8970 2013-08-22 Jan Hubicka <jh@suse.cz>
8972 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
8973 * ipa-utils.h (update_type_inheritance_graph): Declare.
8974 (possible_polymorphic_call_target_p): Declare.
8975 (possible_polymorphic_call_target_p): New.
8976 * ipa-devirt.c: Update toplevel comments.
8977 (cached_polymorphic_call_targets): Move up.
8978 (odr_type_d): Move ID down.
8979 (polymorphic_type_binfo_p): Update comment.
8980 (odr_hasher::remove): Likewise;
8981 (get_odr_type): Set anonymous_namespace.
8982 (dump_odr_type): Dump it.
8983 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
8984 (maybe_record_node): Record node in cached_polymorphic_call_targets.
8985 (record_binfo): Add comment.
8986 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
8988 (devirt_node_removal_hook): Do not iCE when cache is freed.
8989 (possible_polymorphic_call_target_p): New predicate.
8990 (update_type_inheritance_graph): New function.
8992 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
8993 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8994 Sergey Lega <sergey.s.lega@intel.com>
8995 Anna Tikhonova <anna.tikhonova@intel.com>
8996 Ilya Tocar <ilya.tocar@intel.com>
8997 Andrey Turetskiy <andrey.turetskiy@intel.com>
8998 Ilya Verbin <ilya.verbin@intel.com>
8999 Kirill Yukhin <kirill.yukhin@intel.com>
9000 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9002 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
9003 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
9004 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
9005 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
9006 (OPTION_MASK_ISA_AVX2_UNSET): Update.
9007 (OPTION_MASK_ISA_AVX512F_UNSET): New.
9008 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
9009 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
9010 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
9011 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
9012 OPT_mavx512pf, OPT_mavx512er cases.
9013 * config/i386/constraints.md (v): New constraint.
9014 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
9015 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
9016 (bit_AVX512CD): New.
9017 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9018 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
9019 * config/i386/i386-c.c (ix86_target_macros_internal):
9020 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
9022 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
9023 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
9024 * config/i386/i386.c (regclass_map, dbx_register_map)
9025 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
9026 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
9027 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
9029 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
9030 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
9031 -mavx512pf options. Fix formatting.
9032 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
9033 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
9034 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
9035 -mavx512cd, -mavx512pf options.
9036 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
9037 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
9038 (ix86_preferred_output_reload_class): Replace SSE_REGS with
9040 (ix86_hard_regno_mode_ok): Support 512-bit registers.
9041 (ix86_set_reg_reg_cost): Ditto.
9042 (x86_order_regs_for_local_alloc): Ditto.
9043 (MAX_VECT_LEN): Extend to 64-byte.
9044 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
9045 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
9046 (TARGET_AVX512ER, TARGET_AVX512CD): New.
9047 (BIGGEST_ALIGNMENT): Extend to 512-bits.
9048 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
9049 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
9050 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
9051 (SSE_REG_MODE_P): Support new modes.
9052 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
9053 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
9054 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
9055 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
9056 (REG_CLASS_CONTENTS): Add new registers.
9057 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
9058 (EXT_REX_SSE_REGNO_P): New.
9059 (HI_REGISTER_NAMES): Add new registers.
9060 * config/i386/i386.md: Define constants for new registers.
9061 (mode): Add new 512-bit modes.
9062 (prefix): Support evex prefix.
9063 (isa): Support avx512f, noavx512f, fma_avx512f.
9064 (ssemodesuffix): Add new 512-bit modes.
9066 (*movxi_internal_avx512f): Ditto.
9067 (*movdi_internal): Replace constraint "x" with the new constraint "v".
9069 (*movsi_internal): Likewise.
9070 (*movdf_internal): Likewise.
9071 (*movsf_internal): Likewise.
9072 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
9073 (<code><mode>3): Likewise.
9074 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
9076 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
9077 with the new constraint "v".
9078 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
9080 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
9081 with the new constraint "v".
9082 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
9083 (<sse2>_storedqu<avxsizesuffix>): Likewise.
9084 (*<plusminus_insn><mode>3): Likewise.
9085 (<sse>_vm<plusminus_insn><mode>3): Likewise.
9086 (*mul<mode>3): Likewise.
9087 (<sse>_vmmul<mode>3): Likewise.
9088 (<sse>_div<mode>3): Likewise.
9089 (<sse>_vmdiv<mode>3): Likewise.
9090 (<sse>_sqrt<mode>2): Likewise.
9091 (<sse>_vmsqrt<mode>2): Likewise.
9092 (*<code><mode>3_finite): Likewise.
9093 (*<code><mode>3) <smaxmin>: Likewise.
9094 (<sse>_vm<code><mode>3): Likewise.
9095 (*<code><mode>3) <any_logic>: Likewise.
9096 (*fma_fmadd_<mode>): Likewise.
9097 (*fma_fmsub_<mode>): Likewise.
9098 (*fma_fnmadd_<mode>): Likewise.
9099 (*fma_fnmsub_<mode>): Likewise.
9100 (*fma_fmaddsub_<mode>): Likewise.
9101 (*fma_fmsubadd_<mode>): Likewise.
9102 (*fmai_fmadd_<mode>): Likewise.
9103 (*fmai_fmsub_<mode>): Likewise.
9104 (*fmai_fnmadd_<mode>): Likewise.
9105 (*fmai_fnmsub_<mode>): Likewise.
9106 (sse_cvtsi2ss): Likewise.
9107 (sse_cvtsi2ssq): Likewise.
9108 (sse_cvtss2si): Likewise.
9109 (sse_cvtss2si_2): Likewise.
9110 (sse_cvtss2siq): Likewise.
9111 (sse_cvtss2siq_2): Likewise.
9112 (sse_cvttss2si): Likewise.
9113 (sse_cvtss2siq_2): Likewise.
9114 (float<sseintvecmodelower><mode>2): Likewise.
9115 (sse2_cvtsd2si_2): Likewise.
9116 (sse2_cvtsd2siq_2): Likewise.
9117 (*<plusminus_insn><mode>3): Likewise.
9118 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
9119 (*<sse4_1_avx2>_mul<mode>3): Likewise.
9120 (ashr<mode>3): Likewise.
9121 (<shift_insn><mode>3): Likewise.
9122 (avx2_<code><mode>3): Likewise.
9123 (*avx2_<code><mode>3): Likewise.
9124 (*andnot<mode>3): Likewise.
9125 (*<code><mode>3) <any_logic>: Likewise.
9126 (abs<mode>2): Likewise.
9127 (avx2_permvar<mode>): Likewise.
9128 (avx2_perm<mode>_1): Likewise.
9129 (*avx_vpermilp<mode>): Likewise.
9130 (avx_vpermilvar<mode>3): Likewise.
9131 (avx2_ashrv<mode>): Likewise.
9132 (avx2_<shift_insn>v<mode>): Likewise.
9133 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
9135 * doc/rtl.texi: Document XImode.
9137 2013-08-21 Jeff Law <law@redhat.com>
9139 * tree-flow.h (register_jump_thread): Pass vector of edges
9140 instead of each important edge.
9141 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
9142 thread path into a vector and pass that to register_jump_thread.
9143 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
9144 passed in edge vector to the current 3-edge form.
9147 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9149 * dce.c (fini_dce): Call df_analyze again just in case
9150 delete_unmarked_insns removed anything.
9152 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
9154 * reload.h (struct reg_equivs): Rename to ..
9155 (struct reg_equivs_s): .. this.
9157 2013-08-20 Martin Liska <marxin.liska@gmail.com>
9159 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
9161 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9163 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
9165 2013-08-21 Jeff Law <law@redhat.com>
9167 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
9168 simplify assignments too. If the RHS collapses to a singleton
9169 range, then return the value for the range.
9171 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
9173 * config/i386/sse.md (V16): Rename to...
9175 (mov<mode>): Update iterator name.
9176 (*mov<mode>_internal): Ditto.
9177 (push<mode>1): Ditto.
9178 (movmisalign<mode>): Ditto.
9180 2013-08-20 Jan Hubicka <jh@suse.cz>
9183 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9184 entry for direct edges.
9185 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9187 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9189 Revert my last two changes, r201865 and r201864:
9192 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9194 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9195 instances can own GC refs.
9197 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9198 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9199 (gcc::context::gt_pch_nx): Likewise.
9200 (gcc::context::gt_pch_nx): Likewise.
9201 * ggc.h (gt_ggc_mx <T>): New.
9202 (gt_pch_nx_with_op <T>): New.
9203 (gt_pch_nx <T>): New.
9204 * passes.c (opt_pass::gt_ggc_mx): New.
9205 (opt_pass::gt_pch_nx): New.
9206 (opt_pass::gt_pch_nx_with_op): New.
9207 (pass_manager::gt_ggc_mx): New.
9208 (pass_manager::gt_pch_nx): New.
9209 (pass_manager::gt_pch_nx_with_op): New.
9210 (pass_manager::operator new): Use
9211 ggc_internal_cleared_alloc_stat rather than xcalloc.
9212 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9213 (pass_manager::gt_ggc_mx): New.
9214 (pass_manager::gt_pch_nx): New.
9215 (pass_manager::gt_pch_nx_with_op): New.
9216 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9217 (opt_pass::operator new): New.
9218 (opt_pass::gt_ggc_mx): New.
9219 (opt_pass::gt_pch_nx): New.
9220 (opt_pass::gt_pch_nx_with_op): New.
9223 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9225 * Makefile.in (GTFILES): Add context.h.
9226 * context.c (gcc::context::operator new): New.
9227 (gcc::context::gt_ggc_mx): New.
9228 (gcc::context::gt_pch_nx): New.
9229 (gcc::context::gt_pch_nx): New.
9230 * context.h (gcc::context): Add GTY((user)) marking.
9231 (gcc::context::operator new): New.
9232 (gcc::context::gt_ggc_mx): New.
9233 (gcc::context::gt_pch_nx): New.
9234 (gcc::context::gt_pch_nx): New.
9235 (g): Add GTY marking.
9236 (gt_ggc_mx (gcc::context *)): New.
9237 (gt_pch_nx (gcc::context *)): New.
9238 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9239 void *cookie)): New.
9240 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9242 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9244 * dce.c (fini_dce): Call df_analyze again just in case
9245 delete_unmarked_insns removed anything.
9247 2013-08-20 Teresa Johnson <tejohnson@google.com>
9249 PR rtl-optimizations/57451
9250 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9251 from crossing split section boundaries.
9253 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9255 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9256 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9257 with MULTLIB_DEFAULTS.
9259 2013-08-20 Nick Clifton <nickc@redhat.com>
9261 * target.def (narrow_volatile_bitfield): Note that the default
9262 value is false, not !TARGET_STRICT_ALIGN.
9263 * doc/tm.texi: Regenerate.
9265 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9267 Fix LIB_SPEC for systems without libpthread.
9269 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9270 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9272 * config/i386/linux-common.h: Likewise.
9273 * config/mips/linux-common.h: Likewise.
9275 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9277 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9280 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9282 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9283 instances can own GC refs.
9285 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9286 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9287 (gcc::context::gt_pch_nx): Likewise.
9288 (gcc::context::gt_pch_nx): Likewise.
9289 * ggc.h (gt_ggc_mx <T>): New.
9290 (gt_pch_nx_with_op <T>): New.
9291 (gt_pch_nx <T>): New.
9292 * passes.c (opt_pass::gt_ggc_mx): New.
9293 (opt_pass::gt_pch_nx): New.
9294 (opt_pass::gt_pch_nx_with_op): New.
9295 (pass_manager::gt_ggc_mx): New.
9296 (pass_manager::gt_pch_nx): New.
9297 (pass_manager::gt_pch_nx_with_op): New.
9298 (pass_manager::operator new): Use
9299 ggc_internal_cleared_alloc_stat rather than xcalloc.
9300 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9301 (pass_manager::gt_ggc_mx): New.
9302 (pass_manager::gt_pch_nx): New.
9303 (pass_manager::gt_pch_nx_with_op): New.
9304 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9305 (opt_pass::operator new): New.
9306 (opt_pass::gt_ggc_mx): New.
9307 (opt_pass::gt_pch_nx): New.
9308 (opt_pass::gt_pch_nx_with_op): New.
9310 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9312 * Makefile.in (GTFILES): Add context.h.
9313 * context.c (gcc::context::operator new): New.
9314 (gcc::context::gt_ggc_mx): New.
9315 (gcc::context::gt_pch_nx): New.
9316 (gcc::context::gt_pch_nx): New.
9317 * context.h (gcc::context): Add GTY((user)) marking.
9318 (gcc::context::operator new): New.
9319 (gcc::context::gt_ggc_mx): New.
9320 (gcc::context::gt_pch_nx): New.
9321 (gcc::context::gt_pch_nx): New.
9322 (g): Add GTY marking.
9323 (gt_ggc_mx (gcc::context *)): New.
9324 (gt_pch_nx (gcc::context *)): New.
9325 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9326 void *cookie)): New.
9327 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9329 2013-08-20 Alan Modra <amodra@gmail.com>
9332 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9333 (rs6000_emit_epilogue): Likewise.
9335 2013-08-19 Dehao Chen <dehao@google.com>
9337 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9339 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9340 Jakub Jelinek <jakub@redhat.com>
9342 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9343 (BUILT_IN_FABSD64): Likewise.
9344 (BUILT_IN_FABSD128): Likewise.
9345 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9346 (fold_builtin_1): Likewise.
9347 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9348 destination and source operands.
9349 (*abstd2_fpr): Likewise.
9350 (*nabstd2_fpr): Likewise.
9352 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9354 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9357 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9359 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9362 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9364 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9365 * builtins.c (default_libc_has_function): New.
9366 (gnu_libc_has_function): Ditto.
9367 (no_c99_libc_has_function): Ditto.
9368 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9369 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9370 (fold_builtin_sincos): Likewise.
9371 (fold_builtin_cexp): Likewise.
9372 * builtins.def (DEF_C94_BUILTIN): Likewise.
9373 (DEF_C99_BUILTIN): Likewise.
9374 (DEF_C99_C90RES_BUILTIN): Likewise.
9375 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9376 definitions to using this define.
9377 * config/darwin-protos.h (darwin_libc_has_function): New.
9378 * config/darwin.c (darwin_libc_has_function): Ditto.
9379 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9380 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9381 * config/darwin.h: Ditto.
9382 * config/elfos.h: Ditto.
9383 * config/freebsd.h: Ditto.
9384 * config/i386/cygming.h: Ditto.
9385 * config/i386/djgpp.h: Ditto.
9386 * config/i386/i386-interix.h: Ditto.
9387 * config/microblaze/microblaze.h: Ditto.
9388 * config/mmix/mmix.h: Ditto.
9389 * config/gnu-user.h: Ditto.
9390 * config/ia64/hpux.h: Ditto.
9391 * config/pa/pa-hpux.h: Ditto.
9392 * config/pdp11/pdp11.h: Ditto.
9393 * config/picochip/picochip.h: Ditto.
9394 * config/linux.h: Ditto.
9395 * config/netbsd.h: Ditto.
9396 * config/openbsd.h: Ditto.
9397 * config/rs6000/aix43.h: Ditto.
9398 * config/rs6000/aix51.h: Ditto.
9399 * config/rs6000/aix52.h: Ditto.
9400 * config/rs6000/aix53.h: Ditto.
9401 * config/rs6000/aix61.h: Ditto.
9402 * config/rs6000/darwin.h: Ditto.
9403 * config/rs6000/linux.h: Ditto.
9404 * config/rs6000/linux64.h: Ditto.
9405 * config/s390/tpf.h: Ditto.
9406 * config/sol2-10.h: Ditto.
9407 * config/sol2.h: Ditto.
9408 * config/vms/vms.h: Ditto.
9409 * config/vxworks.h: Ditto.
9410 * config/linux-android.c (linux_android_libc_has_function):
9411 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9412 * config/linux-protos.h (linux_android_libc_has_function):
9414 * config/i386/i386.c (ix86_libc_has_function): New.
9415 * config/i386/i386-protos.h
9416 (ix86_libc_has_function): New declaration.
9417 * config/i386/i386.md
9418 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9419 ("isinf<mode>2): Likewise.
9420 * convert.c (convert_to_integer): Using new target hook
9421 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9422 TARGET_C99_FUNCTIONS.
9423 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9424 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9425 * coretypes.h (function_class): New enum for different
9426 classes of functions.
9427 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9428 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9429 (TARGET_HAS_SINCOS): Likewise.
9430 (TARGET_LIBC_HAS_FUNCTION): New.
9431 * doc/tm.texi: Regenerated.
9432 * targhooks.h (default_libc_has_function): New declaration.
9433 (no_c99_libc_has_function): Ditto.
9434 (gnu_libc_has_function): Ditto.
9435 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9436 and TARGET_HAS_SINCOS.
9438 2013-08-18 Jan Hubicka <jh@suse.cz>
9440 * Makeifle-in (ipa-devirt.o): New.
9441 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9442 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9443 (analyze_functions): Look into possible targets of polymorphic call.
9444 * dumpfile.c (dump_files): Add type-inheritance dump.
9445 * dumpfile.h (TDI_inheritance): New.
9446 * ipa-devirt.c: New file.
9447 * ipa-utils.h (odr_type_d): Forward declare.
9448 (odr_type): New type.
9449 (build_type_inheritance_graph): Declare.
9450 (possible_polymorphic_call_targets): Declare and introduce inline
9451 variant when only edge is pased.
9452 (dump_possible_polymorphic_call_targets): Likewise.
9453 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9454 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9455 (types_same_for_odr): ... here.
9456 * tree.h (type_in_anonymous_namespace_p): Declare.
9458 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9460 PR tree-optimization/58006
9461 * tree-parloops.c (take_address_of): Don't ICE if get_name
9463 (eliminate_local_variables_stmt): Remove clobber stmts.
9465 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9467 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9468 error message is issued for an alias to undefined symbol.
9470 2013-08-18 Jan Hubicka <jh@suse.cz>
9472 * cgraph.c (cgraph_create_indirect_edge): Discover
9473 polymorphic calls and record basic info into indirect_info.
9474 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9475 devirtualization, ignore objc function calls.
9476 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9477 call with no parm index info.
9478 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9479 * tree.c (virtual_method_call_p): New function.
9480 * tree.h (virtual_method_call_p): Declare.
9482 2013-08-16 Jan Hubicka <jh@suse.cz>
9485 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9487 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9489 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9492 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9494 * gengtype.c (type_for_name): Add special-case support for
9495 locating types within the "gcc::" namespace.
9496 (open_base_files): Emit a "using namespace gcc" directive.
9498 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9501 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9502 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9504 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9505 array instead of each individual operand as a separate argument.
9506 (emit_fusion_gpr_load): Likewise.
9507 (expand_fusion_gpr_load): Add new function declaration.
9509 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9510 signature to have the operands passed as an array, instead of as
9511 separate arguments. Allow ZERO_EXTEND to be in the memory
9512 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9513 depend on the register live/dead flags when peepholes are run.
9514 (expand_fusion_gpr_load): New function to be called from the
9515 peephole2 pass, to change the register that addis sets to be the
9517 (emit_fusion_gpr_load): Change the calling signature to have the
9518 operands passed as an array, instead of as separate arguments.
9519 Allow ZERO_EXTEND to be in the memory address, and also
9520 SIGN_EXTEND if -mpower8-fusion-sign.
9522 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9524 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9525 adjust the register addis loads up in the peephole2 pass. Do not
9526 depend on the register live/dead state when the peephole pass is done.
9528 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9530 * gengtype.c (create_user_defined_type): Ensure that the kind
9531 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9532 declaration is seen before the GTY((user)) marking.
9534 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9537 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9539 2013-08-16 Jan Hubicka <jh@suse.cz>
9541 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9542 arugment expected_type.
9543 (gimple_fold_call): Use it.
9544 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9545 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9546 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9547 (try_make_edge_direct_virtual_call): Likewise.
9548 * tree.c (obj_type_ref_class): New.
9549 * tree.h (obj_type_ref_class): Use it.
9551 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9553 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9554 (rtl_slim_pp): Likewise.
9555 (init_rtl_slim_pretty_print): Likewise.
9556 (dump_value_slim): Don't call it. Use local pretty printer.
9557 (dump_insn_slim): Likewise.
9558 (dump_rtl_slim): Likewise.
9559 (str_pattern_slim): Likewise.
9560 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9563 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9565 PR tree-optimization/58164
9566 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9567 walk gimple_goto_dest of GIMPLE_GOTO.
9569 PR tree-optimization/58165
9570 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9571 bi_call must be the last stmt in a bb, don't split_block, instead
9572 use fallthru edge from it and give up if there is none.
9573 Release conds vector when returning early.
9575 2013-08-14 Xinliang David Li <davidxl@google.com>
9577 * config/i386/i386.c (ix86_option_override_internal):
9578 Remove unused variable and field.
9580 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9583 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9584 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9585 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9586 and Linux, correct BLKmode alignment when 128-bit alignment is
9587 required and compatibility flag is not set.
9588 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9589 for zero-size arguments when compatibility flag is not set.
9591 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9593 PR tree-optimization/58145
9594 * tree-sra.c (build_ref_for_offset): If prev_base has
9595 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9597 2013-08-14 Xinliang David Li <davidxl@google.com>
9599 * config/i386/i386.c (ix86_option_override_internal):
9600 Fix uninitialized variable error.
9602 2013-08-14 Xinliang David Li <davidxl@google.com>
9604 * config/i386/i386.opt: Define two new options.
9605 * config/i386/x86-tune.def: Add arch selector field in macros.
9606 * config/i386/i386.h: Adjust macro definition.
9607 * config/i386/i386.c (ix86_option_override_internal):
9609 (parse_mtune_ctrl_str): New function.
9610 (set_ix86_tune_features): New function.
9611 (ix86_function_specific_restore): Call the new helper function.
9613 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9615 PR rtl-optimization/57662
9616 * sel-sched.c (code_motion_process_successors): When the current insn
9617 is removed after the recursive traversal, break from the loop.
9618 Add comments and debug printouts.
9620 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9621 Alexandre Oliva <aoliva@redhat.com>
9624 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9625 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9626 there also UNSPEC_PLTOFF.
9628 2013-08-14 Marek Polacek <polacek@redhat.com>
9630 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9631 NUM_CONDITIONS bit positions.
9633 2013-08-13 Cary Coutant <ccoutant@google.com>
9635 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9636 (attr_checksum): Hash vector contents instead of pointer.
9637 (attr_checksum_ordered): Likewise.
9639 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9641 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9642 when Pmode != word_mode. Add length_address attribute.
9643 (sse3_monitor_<mode>): Merge from sse3_monitor and
9644 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9645 Pmode != word_mode. Update insn length attribute.
9646 * config/i386/i386.c (ix86_option_override_internal): Update
9647 ix86_gen_monitor selection for merged sse3_monitor insn.
9649 2013-08-13 Julian Brown <julian@codesourcery.com>
9651 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9652 perform invalid legitimization on greater-than-word-size modes for
9655 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9657 * ira.c (setup_class_translate_array): Use aclass instead of cl
9658 for classes not fully covered by allocno classes.
9660 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9662 PR tree-optimization/57661
9663 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9664 * tree-inline.c (tree_function_versioning): Initialize it.
9665 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9666 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9667 that is not being copied.
9670 * asan.c (instrument_strlen_call): Fix typo in comment.
9671 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9673 2013-08-13 Steve Ellcey <sellcey@mips.com>
9675 * config/mips/mips.md (prefetch): Use lw instead of ld on
9676 loongson in 32bit mode.
9678 2013-08-13 Nick Clifton <nickc@redhat.com>
9680 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9682 2013-08-13 Jan Hubicka <jh@suse.cz>
9684 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9685 introduced edge; fix typo in sanity check.
9686 (cgraph_resolve_speculation): Export; improve diagnostic.
9687 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9688 speculation at type mismatch.
9689 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9690 (cgraph_resolve_speculation): Declare.
9691 (symtab_can_be_discarded): New function.
9692 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9693 * ipa-inline-transform.c (speculation_removed): New global var.
9694 (clone_inlined_nodes): See if speculation can be removed.
9695 (inline_call): If speculations was removed, we growths may not match.
9696 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9697 (speculation_useful_p): New function.
9698 (resolve_noninline_speculation): New function.
9699 (inline_small_functions): Resolve useless speculations.
9700 * ipa-inline.h (speculation_useful_p): Declare
9701 * ipa.c (can_replace_by_local_alias): Simplify.
9702 (ipa_profile): Produce speculative calls in non-lto, too;
9703 add simple cost model; produce local aliases.
9705 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9707 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9710 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9712 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9714 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9716 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9717 iterator instead of X87MODEF.
9719 2013-08-12 Perez Read <netfirewall@gmail.com>
9722 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9723 operand 0 for intel asm alternative.
9724 (*movabs<mode>_2): Ditto for operand 1.
9726 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9728 * config/aarch64/arm_none.h
9729 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9731 2013-08-12 Nick Clifton <nickc@redhat.com>
9733 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9735 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9737 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9738 expand for QI/HImode operand to produce more effictive code for
9739 unsigned char(short) --> float(double) conversion.
9741 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9743 * doc/invoke.texi: Mention that -ftls-model does not force the final
9746 2013-08-12 Marek Polacek <polacek@redhat.com>
9747 Marc Glisse <marc.glisse@inria.fr>
9749 PR tree-optimization/57980
9750 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9751 when creating -1 constant.
9753 2013-08-10 Jan Hubicka <jh@suse.cz>
9755 Workaround binutils PR14342.
9756 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9757 (gimple_init_edge_profiler): Likewise.
9758 (gimple_gen_ic_func_profiler): Likewise.
9760 2013-08-09 Jan Hubicka <jh@suse.cz>
9762 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9764 2013-08-09 Xinliang David Li <davidxl@google.com>
9766 * config/i386/stringop.def: New file.
9767 * config/i386/stringop.opt: New file.
9768 * config/i386/i386-opts.h: Include stringopt.def.
9769 * config/i386/i386.opt: Include stringopt.opt.
9770 * config/i386/i386.c (ix86_option_override_internal):
9771 Override default size based stringop inline strategies with options.
9772 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9775 2013-08-09 Jan Hubicka <jh@suse.cz>
9777 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9779 2013-08-09 Jan Hubicka <jh@suse.cz>
9781 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9783 (dump_cgraph_node): Dump profile-id.
9784 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9785 and common_target_probability.
9786 * lto-cgraph.c (lto_output_edge): Stream common targets.
9787 (lto_output_node): Stream profile ids.
9788 (input_node): Stream profile ids.
9789 (input_edge): Stream common targets.
9790 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9791 * ipa.c: Include value-prof.h
9792 (ipa_profile_generate_summary): Turn indirect call statement histograms
9793 into common targets.
9794 (ipa_profile): Turn common targets into speculative edges.
9796 2013-08-09 Jan Hubicka <jh@suse.cz>
9798 * cgraph.h (cgraph_node): Add profile_id.
9799 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9800 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9801 (del_node_map): Update.
9802 (find_func_by_funcdef_no): Replace by ...
9803 (find_func_by_profile_id): ... this one.
9804 (gimple_ic_transform): Do not remove useful histograms when
9805 speculation is not done; dump info when indirect call removal
9807 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9808 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9809 (__gcov_indirect_call_profiler_v2): .. this one.
9810 * profile.h (init_node_map): Update.
9811 * coverage.c (coverage_compute_profile_id): New function.
9812 * coverage.h (coverage_compute_profile_id): Declare.
9813 * tree-profile.c (init_ic_make_global_vars): Make
9814 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9815 (gimple_init_edge_profiler): Update prototype of
9816 __gcov_indirect_call_profiler.
9817 (gimple_gen_ic_func_profiler): Simplify.
9818 (tree_profiling): Use init_node_map
9820 2013-08-09 Jan Hubicka <jh@suse.cz>
9822 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9823 non-speculative refs.
9824 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9825 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9826 (cgraph_set_call_stmt): Likewise.
9827 (cgraph_create_edge_1): Fix release checking compilatoin;
9829 (cgraph_free_edge): Free indirect info.
9830 (cgraph_turn_edge_to_speculative): New function.
9831 (cgraph_speculative_call_info): New function.
9832 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9833 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9834 (dump_cgraph_node): Dump speculation.
9835 (verify_edge_count_and_frequency): Accept speculative edges.
9836 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9837 (verify_cgraph_node): Handle speculation.
9838 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
9839 (cgraph_set_call_stmt): Update prototype.
9840 (cgraph_make_edge_direct): Update prototype.
9841 (cgraph_speculative_call_info): Declare.
9842 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
9843 to change; update call of ipa_find_references.
9844 * ipa-ref.c (ipa_record_reference): Fix return value; clear
9845 lto_stmt_uid and speculative flags.
9846 (ipa_dump_references): Dump speculation.
9847 (ipa_clone_references): Clone speculative flag.
9848 (ipa_clone_referring): Likewise.
9849 (ipa_clone_ref): New function.
9850 (ipa_find_reference): Look into lto_stmt_uids
9851 (ipa_clear_stmts_in_references): Do not clear speculative calls.
9852 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
9853 (ipa_find_reference): Update declaration.
9854 (ipa_clone_ref): Declare.
9855 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
9856 stream speculative flag.
9857 (lto_output_ref): Stream statements uids and speculation.
9858 (input_ref): Likewise.
9859 (input_edge): Stream speuclation.
9860 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
9861 (cgraph_set_call_stmt_including_clones): Handle speculation.
9862 * ipa-inline.c (heap_edge_removal_hook): New function.
9863 (inline_small_functions): Register it.
9864 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
9865 also initialize refs.
9866 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
9868 (try_make_edge_direct_simple_call): Likewise.
9869 (try_make_edge_direct_simple_call): Likewise.
9870 (update_indirect_edges_after_inlining): Likewise.
9871 (remove_described_reference): Look proper lto_stmt_uid.
9872 (propagate_controlled_uses): Likewise.
9873 (propagate_controlled_uses): Liekwise.
9874 * tree-inline.c (copy_bb): Copy speculative edges.
9875 (redirect_all_calls): New function.
9876 (copy_cfg_body): Do redirection after loop info is updated.
9877 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
9879 2013-08-09 Jan Hubicka <jh@suse.cz>
9881 * lto-streamer-out.c (output_function): Renumber PHIs.
9882 * lto-streamer-in.c (input_function): Likewise.
9884 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
9886 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
9887 (get_lane_unsigned): Likewise.
9888 (dup_lane_scalar): Likewise.
9889 (get_lane): enable for VALL.
9890 * config/aarch64/aarch64-simd.md
9891 (aarch64_dup_lane_scalar<mode>): Remove.
9892 (aarch64_get_lane_signed<mode>): Likewise.
9893 (aarch64_get_lane_unsigned<mode>): Likewise.
9894 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
9895 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
9896 (aarch64_get_lane<mode>): Enable for all vector modes.
9897 (aarch64_get_lanedi): Remove misleading constraints.
9898 * config/aarch64/arm_neon.h
9899 (__aarch64_vget_lane_any): Define.
9900 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
9901 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
9902 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
9903 * config/aarch64/iterators.md (VDQQH): New.
9907 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
9909 * configure.ac: Add GAS check for LEON instructions on SPARC.
9910 * configure: Regenerate.
9911 * config.in: Likewise.
9912 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
9914 * config/sparc/sparc.opt (LEON, LEON3): New masks.
9915 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
9917 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
9918 (AS_LEON_FLAG): New macro.
9919 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
9920 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
9921 Deal with LEON and LEON3 for the memory model.
9922 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
9923 (atomic_compare_and_swap<mode>_1): Likewise.
9924 (*atomic_compare_and_swap<mode>_1): Likewise.
9926 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
9928 * config/arm/neon.md (vcond): Fix floating-point vector
9929 comparisons against 0.
9931 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
9933 * lra-constraints.c (emit_spill_move): Remove assert.
9934 (process_alt_operands): Add more debugging
9935 output. Increase reject for spilling into memory. Decrease
9936 reject for reloading scratch.
9937 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
9939 2013-08-08 Steve Ellcey <sellcey@mips.com>
9941 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
9942 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
9943 micromips incompatible. Add nan2008.
9944 (MULTILIB_DIRNAMES): Add nan2008.
9945 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9946 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
9947 and micromips incompatible. Add nan2008.
9948 (MULTILIB_DIRNAMES): Add nan2008.
9949 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
9951 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
9953 PR rtl-optimization/58079
9954 * combine.c (combine_simplify_rtx): Avoid using SUBST if
9955 simplify_comparison has widened a comparison with an integer.
9957 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9959 * config/arm/neon.md (movmisalign<mode>): Disable when we
9960 don't allow unaligned accesses.
9961 (*movmisalign<mode>_neon_store): Likewise.
9962 (*movmisalign<mode>_neon_load): Likewise.
9963 (*movmisalign<mode>_neon_store): Likewise.
9964 (*movmisalign<mode>_neon_load): Likewise.
9966 2013-08-08 Jan Hubicka <jh@suse.cz>
9968 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
9969 (make_pass_rebuild_cgraph_edges): Also clear references.
9970 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
9971 * ipa-inline-transform.c (inline_transform): Remove all references
9973 * cgraphunit.c (expand_function): Remove all references after
9975 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
9976 (ipa_find_reference): Rewrite to iterator.
9977 (remove_stmt_references): Likewise.
9978 (ipa_clear_stmts_in_references): New function.
9979 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
9980 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
9982 * ipa-split.c (split_function): Remove references in split function.
9984 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9987 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
9988 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
9990 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
9993 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
9994 suggested mode for the assignment isn't compatible with the
9997 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10000 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
10002 2013-08-07 Xinliang David Li <davidxl@google.com>
10004 * config/i386/i386.opt: New option -mtune-ctrl=.
10005 * config/i386/x86-tune.def: New file.
10006 * config/i386/i386.h: include x86-tune.def.
10007 * config/i386/i386.c (ix86_option_override_internal):
10008 Parsing -mtune-ctrl= option and set tune features.
10010 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
10013 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
10014 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
10017 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10019 * rtl.h (update_alignments): Declare.
10020 * final.c (grow_label_align): New function extracted from...
10021 (shorten_branches): ...here. Call it.
10022 (update_alignments): New function.
10023 * reorg.c (sibling_labels): New variable.
10024 (get_label_before): Add SIBLING parameter. If it is non-zero, push
10025 the new label along with it onto the sibling_labels vector.
10026 (fill_simple_delay_slots): Adjust call to get_label_before.
10027 (fill_slots_from_thread): Likewise.
10028 (relax_delay_slots): Likewise.
10029 (make_return_insns): Likewise.
10030 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
10032 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10034 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
10035 document its semantics.
10036 (diagnostic_report_diagnostic): Adjust accordingly.
10038 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10040 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
10041 (sparc_option_override): ...and port to new C++ pass API.
10042 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
10044 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
10046 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
10048 2013-08-06 Caroline Tice <cmtice@google.com>
10050 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
10051 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
10052 * tree-pass.h: Add pass_vtable_verify.
10053 * varasm.c (assemble_variable): Add code to properly set the comdat
10054 section and name for the .vtable_map_vars section.
10055 (assemble_vtyv_preinit_initializer): New function.
10056 (default_sectin_type_flags): Make sure .vtable_map_vars section has
10058 * output.h: Add function decl for assemble_vtv_preinit_initializer.
10059 * vtable-verify.c: New file.
10060 * vtable-verify.h: New file.
10061 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
10062 initialiation levels.
10063 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
10064 * passes.def: Insert pass_vtable_verify.
10065 * aclocal.m4: Reorder includes.
10066 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
10067 -fvtv-counts options.
10068 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
10069 as appropriate, if -fvtable-verify=... is used.
10070 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
10071 -fvtable-verify=... is used.
10072 * Makefile.in (OBJS): Add vtable-verify.o to list.
10073 (vtable-verify.o): Add new build rule.
10074 (GTFILES): Add vtable-verify.c to list.
10075 * common.opt (fvtable-verify=): New flag.
10076 (vtv_priority): Values for fvtable-verify= flag.
10077 (fvtv-counts): New flag.
10078 (fvtv-debug): New flag.
10079 * tree.h (save_vtable_map_decl): New extern function decl.
10081 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10083 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
10084 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
10085 (pass_data_rl78_devirt): ...new pass_data instance and...
10086 (make_pass_rl78_devirt): ...new function.
10087 (rl78_asm_file_start): Port pass registration to new C++ API.
10089 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10091 * coretypes.h (rtl_opt_pass): Add.
10092 (gcc::context): Add.
10093 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
10094 (epiphany_init): Port to new C++ pass API.
10095 (epiphany_optimize_mode_switching): Likewise.
10096 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
10097 (pass_manager::get_pass_mode_switching): New.
10098 (pass_manager::get_pass_peephole2): New.
10099 * mode-switching.c (pass_mode_switching): Add clone method.
10100 * recog.c (pass_peephole2): Add clone method.
10101 (pass_split_all_insns): Add clone method.
10103 2013-08-06 David Malcolm <dmalcolm@redhat.com>
10105 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
10106 (mips_option_override): ...here, porting to new C++ API for passes.
10108 2013-08-06 Jan Hubicka <jh@suse.cz>
10110 * cgraph.c (cgraph_get_body): New function based on lto.c
10112 * cgraph.h (cgraph_get_body): Declare.
10113 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
10115 * cgraphunit.c (expand_function): Get body prior expanding.
10116 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
10117 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
10119 * passes.c (do_per_function_toporder): Get body.
10120 * tree-inline.c (expand_call_inline): Get body prior inlining it.
10121 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
10123 2013-08-06 Martin Jambor <mjambor@suse.cz>
10126 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
10127 is not re-finalized. Rename second parameter to no_collect.
10129 2013-08-06 Martin Jambor <mjambor@suse.cz>
10131 PR middle-end/58041
10132 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
10133 MEM_REF has proper alignment information.
10135 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
10138 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
10140 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
10141 rtx (*) (rtx, ...) with insn_gen_fn.
10142 * genoutput.c (output_insn_data): Cast gen_? function pointers to
10143 insn_gen_fn::stored_funcptr. Add initializer braces.
10145 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10147 Rewrite how instances of passes are cloned to remove assumptions
10148 about their sizes (thus allowing pass subclasses to have
10149 additional data fields, albeit non-GC-managed ones at this point).
10151 * passes.c (make_pass_instance): Now that passes have clone
10152 methods, rewrite this function to eliminate XNEW and memcpy
10153 calls that used hardcoded sizes. Since this function no longer
10154 creates pass instances, rename it to...
10155 (add_pass_instance): ...this. Document the old way that passes were
10156 numbered and flagged, and rework this function to continue using it.
10157 (next_pass_1): Add an initial_pass argument for use by
10159 (position_pass): When adding multiple instances of a pass, use
10160 the pass's clone method, rather than relying on the XNEW/memcpy
10161 within the former make_pass_instance (now add_pass_instance).
10162 (pass_manager::pass_manager): When invoking next_pass_1, also supply
10163 the initial instance of the current pass within the pass manager.
10165 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10167 This is the automated part of the conversion of passes from C
10168 structs to C++ classes.
10170 Patch autogenerated by refactor_passes.py from
10171 https://github.com/davidmalcolm/gcc-refactoring-scripts
10172 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
10174 * asan.c (pass_asan): Convert from a global struct to a subclass of
10175 gimple_opt_pass along with...
10176 (pass_data_asan): ...new pass_data instance and...
10177 (make_pass_asan): ...new function.
10178 (pass_asan_O0): Convert from a global struct to a subclass of
10179 gimple_opt_pass along with...
10180 (pass_data_asan_O0): ...new pass_data instance and...
10181 (make_pass_asan_O0): ...new function.
10182 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10183 subclass of rtl_opt_pass along with...
10184 (pass_data_inc_dec): ...new pass_data instance and...
10185 (make_pass_inc_dec): ...new function.
10186 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10187 a subclass of rtl_opt_pass along with...
10188 (pass_data_reorder_blocks): ...new pass_data instance and...
10189 (make_pass_reorder_blocks): ...new function.
10190 (pass_duplicate_computed_gotos): Convert from a global struct to a
10191 subclass of rtl_opt_pass along with...
10192 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10193 (make_pass_duplicate_computed_gotos): ...new function.
10194 (pass_partition_blocks): Convert from a global struct to a subclass of
10195 rtl_opt_pass along with...
10196 (pass_data_partition_blocks): ...new pass_data instance and...
10197 (make_pass_partition_blocks): ...new function.
10198 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10199 struct to a subclass of rtl_opt_pass along with...
10200 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10202 (make_pass_branch_target_load_optimize1): ...new function.
10203 (pass_branch_target_load_optimize2): Convert from a global struct to a
10204 subclass of rtl_opt_pass along with...
10205 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10207 (make_pass_branch_target_load_optimize2): ...new function.
10208 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10209 of rtl_opt_pass along with...
10210 (pass_data_jump): ...new pass_data instance and...
10211 (make_pass_jump): ...new function.
10212 (pass_jump2): Convert from a global struct to a subclass of
10213 rtl_opt_pass along with...
10214 (pass_data_jump2): ...new pass_data instance and...
10215 (make_pass_jump2): ...new function.
10216 * cfgexpand.c (pass_expand): Convert from a global struct to a
10217 subclass of rtl_opt_pass along with...
10218 (pass_data_expand): ...new pass_data instance and...
10219 (make_pass_expand): ...new function.
10220 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10221 of rtl_opt_pass along with...
10222 (pass_data_free_cfg): ...new pass_data instance and...
10223 (make_pass_free_cfg): ...new function.
10224 (pass_into_cfg_layout_mode): Convert from a global struct to a
10225 subclass of rtl_opt_pass along with...
10226 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10227 (make_pass_into_cfg_layout_mode): ...new function.
10228 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10229 subclass of rtl_opt_pass along with...
10230 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10231 (make_pass_outof_cfg_layout_mode): ...new function.
10232 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10233 struct to a subclass of gimple_opt_pass along with...
10234 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10235 (make_pass_build_cgraph_edges): ...new function.
10236 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10237 subclass of gimple_opt_pass along with...
10238 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10239 (make_pass_rebuild_cgraph_edges): ...new function.
10240 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10241 subclass of gimple_opt_pass along with...
10242 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10244 (make_pass_remove_cgraph_callee_edges): ...new function.
10245 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10246 struct to a subclass of rtl_opt_pass along with...
10247 (pass_data_stack_adjustments): ...new pass_data instance and...
10248 (make_pass_stack_adjustments): ...new function.
10249 * combine.c (pass_combine): Convert from a global struct to a subclass
10250 of rtl_opt_pass along with...
10251 (pass_data_combine): ...new pass_data instance and...
10252 (make_pass_combine): ...new function.
10253 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10254 global struct to a subclass of rtl_opt_pass along with...
10255 (pass_data_compare_elim_after_reload): ...new pass_data instance
10257 (make_pass_compare_elim_after_reload): ...new function.
10258 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10259 of rtl_opt_pass along with...
10260 (pass_data_rtl_cprop): ...new pass_data instance and...
10261 (make_pass_rtl_cprop): ...new function.
10262 * cse.c (pass_cse): Convert from a global struct to a subclass of
10263 rtl_opt_pass along with...
10264 (pass_data_cse): ...new pass_data instance and...
10265 (make_pass_cse): ...new function.
10266 (pass_cse2): Convert from a global struct to a subclass of
10267 rtl_opt_pass along with...
10268 (pass_data_cse2): ...new pass_data instance and...
10269 (make_pass_cse2): ...new function.
10270 (pass_cse_after_global_opts): Convert from a global struct to a
10271 subclass of rtl_opt_pass along with...
10272 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10273 (make_pass_cse_after_global_opts): ...new function.
10274 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10275 of rtl_opt_pass along with...
10276 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10277 (make_pass_ud_rtl_dce): ...new function.
10278 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10279 rtl_opt_pass along with...
10280 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10281 (make_pass_fast_rtl_dce): ...new function.
10282 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10283 a subclass of rtl_opt_pass along with...
10284 (pass_data_df_initialize_opt): ...new pass_data instance and...
10285 (make_pass_df_initialize_opt): ...new function.
10286 (pass_df_initialize_no_opt): Convert from a global struct to a
10287 subclass of rtl_opt_pass along with...
10288 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10289 (make_pass_df_initialize_no_opt): ...new function.
10290 (pass_df_finish): Convert from a global struct to a subclass of
10291 rtl_opt_pass along with...
10292 (pass_data_df_finish): ...new pass_data instance and...
10293 (make_pass_df_finish): ...new function.
10294 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10295 rtl_opt_pass along with...
10296 (pass_data_rtl_dse1): ...new pass_data instance and...
10297 (make_pass_rtl_dse1): ...new function.
10298 (pass_rtl_dse2): Convert from a global struct to a subclass of
10299 rtl_opt_pass along with...
10300 (pass_data_rtl_dse2): ...new pass_data instance and...
10301 (make_pass_rtl_dse2): ...new function.
10302 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10303 subclass of rtl_opt_pass along with...
10304 (pass_data_dwarf2_frame): ...new pass_data instance and...
10305 (make_pass_dwarf2_frame): ...new function.
10306 * except.c (pass_set_nothrow_function_flags): Convert from a global
10307 struct to a subclass of rtl_opt_pass along with...
10308 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10310 (make_pass_set_nothrow_function_flags): ...new function.
10311 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10312 subclass of rtl_opt_pass along with...
10313 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10315 (make_pass_convert_to_eh_region_ranges): ...new function.
10316 * final.c (pass_compute_alignments): Convert from a global struct to a
10317 subclass of rtl_opt_pass along with...
10318 (pass_data_compute_alignments): ...new pass_data instance and...
10319 (make_pass_compute_alignments): ...new function.
10320 (pass_final): Convert from a global struct to a subclass of
10321 rtl_opt_pass along with...
10322 (pass_data_final): ...new pass_data instance and...
10323 (make_pass_final): ...new function.
10324 (pass_shorten_branches): Convert from a global struct to a subclass of
10325 rtl_opt_pass along with...
10326 (pass_data_shorten_branches): ...new pass_data instance and...
10327 (make_pass_shorten_branches): ...new function.
10328 (pass_clean_state): Convert from a global struct to a subclass of
10329 rtl_opt_pass along with...
10330 (pass_data_clean_state): ...new pass_data instance and...
10331 (make_pass_clean_state): ...new function.
10332 * function.c (pass_instantiate_virtual_regs): Convert from a global
10333 struct to a subclass of rtl_opt_pass along with...
10334 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10335 (make_pass_instantiate_virtual_regs): ...new function.
10336 (pass_leaf_regs): Convert from a global struct to a subclass of
10337 rtl_opt_pass along with...
10338 (pass_data_leaf_regs): ...new pass_data instance and...
10339 (make_pass_leaf_regs): ...new function.
10340 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10341 subclass of rtl_opt_pass along with...
10342 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10344 (make_pass_thread_prologue_and_epilogue): ...new function.
10345 (pass_match_asm_constraints): Convert from a global struct to a
10346 subclass of rtl_opt_pass along with...
10347 (pass_data_match_asm_constraints): ...new pass_data instance and...
10348 (make_pass_match_asm_constraints): ...new function.
10349 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10350 subclass of rtl_opt_pass along with...
10351 (pass_data_rtl_fwprop): ...new pass_data instance and...
10352 (make_pass_rtl_fwprop): ...new function.
10353 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10354 rtl_opt_pass along with...
10355 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10356 (make_pass_rtl_fwprop_addr): ...new function.
10357 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10358 rtl_opt_pass along with...
10359 (pass_data_rtl_pre): ...new pass_data instance and...
10360 (make_pass_rtl_pre): ...new function.
10361 (pass_rtl_hoist): Convert from a global struct to a subclass of
10362 rtl_opt_pass along with...
10363 (pass_data_rtl_hoist): ...new pass_data instance and...
10364 (make_pass_rtl_hoist): ...new function.
10365 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10366 subclass of gimple_opt_pass along with...
10367 (pass_data_lower_cf): ...new pass_data instance and...
10368 (make_pass_lower_cf): ...new function.
10369 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10370 from a global struct to a subclass of gimple_opt_pass along with...
10371 (pass_data_strength_reduction): ...new pass_data instance and...
10372 (make_pass_strength_reduction): ...new function.
10373 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10374 of rtl_opt_pass along with...
10375 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10376 (make_pass_rtl_ifcvt): ...new function.
10377 (pass_if_after_combine): Convert from a global struct to a subclass of
10378 rtl_opt_pass along with...
10379 (pass_data_if_after_combine): ...new pass_data instance and...
10380 (make_pass_if_after_combine): ...new function.
10381 (pass_if_after_reload): Convert from a global struct to a subclass of
10382 rtl_opt_pass along with...
10383 (pass_data_if_after_reload): ...new pass_data instance and...
10384 (make_pass_if_after_reload): ...new function.
10385 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10386 a subclass of rtl_opt_pass along with...
10387 (pass_data_initialize_regs): ...new pass_data instance and...
10388 (make_pass_initialize_regs): ...new function.
10389 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10390 of ipa_opt_pass_d along with...
10391 (pass_data_ipa_cp): ...new pass_data instance and...
10392 (make_pass_ipa_cp): ...new function.
10393 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10394 global struct to a subclass of gimple_opt_pass along with...
10395 (pass_data_inline_parameters): ...new pass_data instance and...
10396 (make_pass_inline_parameters): ...new function.
10397 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10398 subclass of gimple_opt_pass along with...
10399 (pass_data_early_inline): ...new pass_data instance and...
10400 (make_pass_early_inline): ...new function.
10401 (pass_ipa_inline): Convert from a global struct to a subclass of
10402 ipa_opt_pass_d along with...
10403 (pass_data_ipa_inline): ...new pass_data instance and...
10404 (make_pass_ipa_inline): ...new function.
10405 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10406 struct to a subclass of gimple_opt_pass along with...
10407 (pass_data_local_pure_const): ...new pass_data instance and...
10408 (make_pass_local_pure_const): ...new function.
10409 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10410 ipa_opt_pass_d along with...
10411 (pass_data_ipa_pure_const): ...new pass_data instance and...
10412 (make_pass_ipa_pure_const): ...new function.
10413 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10414 to a subclass of ipa_opt_pass_d along with...
10415 (pass_data_ipa_reference): ...new pass_data instance and...
10416 (make_pass_ipa_reference): ...new function.
10417 * ipa-split.c (pass_split_functions): Convert from a global struct to
10418 a subclass of gimple_opt_pass along with...
10419 (pass_data_split_functions): ...new pass_data instance and...
10420 (make_pass_split_functions): ...new function.
10421 (pass_feedback_split_functions): Convert from a global struct to a
10422 subclass of gimple_opt_pass along with...
10423 (pass_data_feedback_split_functions): ...new pass_data instance and...
10424 (make_pass_feedback_split_functions): ...new function.
10425 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10426 global struct to a subclass of simple_ipa_opt_pass along with...
10427 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10429 (make_pass_ipa_function_and_variable_visibility): ...new function.
10430 (pass_ipa_free_inline_summary): Convert from a global struct to a
10431 subclass of simple_ipa_opt_pass along with...
10432 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10433 (make_pass_ipa_free_inline_summary): ...new function.
10434 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10435 subclass of ipa_opt_pass_d along with...
10436 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10438 (make_pass_ipa_whole_program_visibility): ...new function.
10439 (pass_ipa_profile): Convert from a global struct to a subclass of
10440 ipa_opt_pass_d along with...
10441 (pass_data_ipa_profile): ...new pass_data instance and...
10442 (make_pass_ipa_profile): ...new function.
10443 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10444 ipa_opt_pass_d along with...
10445 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10446 (make_pass_ipa_cdtor_merge): ...new function.
10447 * ira.c (pass_ira): Convert from a global struct to a subclass of
10448 rtl_opt_pass along with...
10449 (pass_data_ira): ...new pass_data instance and...
10450 (make_pass_ira): ...new function.
10451 (pass_reload): Convert from a global struct to a subclass of
10452 rtl_opt_pass along with...
10453 (pass_data_reload): ...new pass_data instance and...
10454 (make_pass_reload): ...new function.
10455 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10456 subclass of rtl_opt_pass along with...
10457 (pass_data_cleanup_barriers): ...new pass_data instance and...
10458 (make_pass_cleanup_barriers): ...new function.
10459 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10460 of rtl_opt_pass along with...
10461 (pass_data_loop2): ...new pass_data instance and...
10462 (make_pass_loop2): ...new function.
10463 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10464 rtl_opt_pass along with...
10465 (pass_data_rtl_loop_init): ...new pass_data instance and...
10466 (make_pass_rtl_loop_init): ...new function.
10467 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10468 rtl_opt_pass along with...
10469 (pass_data_rtl_loop_done): ...new pass_data instance and...
10470 (make_pass_rtl_loop_done): ...new function.
10471 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10472 subclass of rtl_opt_pass along with...
10473 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10474 (make_pass_rtl_move_loop_invariants): ...new function.
10475 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10476 rtl_opt_pass along with...
10477 (pass_data_rtl_unswitch): ...new pass_data instance and...
10478 (make_pass_rtl_unswitch): ...new function.
10479 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10480 subclass of rtl_opt_pass along with...
10481 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10483 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10484 (pass_rtl_doloop): Convert from a global struct to a subclass of
10485 rtl_opt_pass along with...
10486 (pass_data_rtl_doloop): ...new pass_data instance and...
10487 (make_pass_rtl_doloop): ...new function.
10488 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10489 a subclass of rtl_opt_pass along with...
10490 (pass_data_lower_subreg): ...new pass_data instance and...
10491 (make_pass_lower_subreg): ...new function.
10492 (pass_lower_subreg2): Convert from a global struct to a subclass of
10493 rtl_opt_pass along with...
10494 (pass_data_lower_subreg2): ...new pass_data instance and...
10495 (make_pass_lower_subreg2): ...new function.
10496 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10497 struct to a subclass of ipa_opt_pass_d along with...
10498 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10499 (make_pass_ipa_lto_gimple_out): ...new function.
10500 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10501 of ipa_opt_pass_d along with...
10502 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10503 (make_pass_ipa_lto_finish_out): ...new function.
10504 * mode-switching.c (pass_mode_switching): Convert from a global struct
10505 to a subclass of rtl_opt_pass along with...
10506 (pass_data_mode_switching): ...new pass_data instance and...
10507 (make_pass_mode_switching): ...new function.
10508 * modulo-sched.c (pass_sms): Convert from a global struct to a
10509 subclass of rtl_opt_pass along with...
10510 (pass_data_sms): ...new pass_data instance and...
10511 (make_pass_sms): ...new function.
10512 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10513 subclass of gimple_opt_pass along with...
10514 (pass_data_expand_omp): ...new pass_data instance and...
10515 (make_pass_expand_omp): ...new function.
10516 (pass_lower_omp): Convert from a global struct to a subclass of
10517 gimple_opt_pass along with...
10518 (pass_data_lower_omp): ...new pass_data instance and...
10519 (make_pass_lower_omp): ...new function.
10520 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10521 of gimple_opt_pass along with...
10522 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10523 (make_pass_diagnose_omp_blocks): ...new function.
10524 * passes.c (pass_early_local_passes): Convert from a global struct to
10525 a subclass of simple_ipa_opt_pass along with...
10526 (pass_data_early_local_passes): ...new pass_data instance and...
10527 (make_pass_early_local_passes): ...new function.
10528 (pass_all_early_optimizations): Convert from a global struct to a
10529 subclass of gimple_opt_pass along with...
10530 (pass_data_all_early_optimizations): ...new pass_data instance and...
10531 (make_pass_all_early_optimizations): ...new function.
10532 (pass_all_optimizations): Convert from a global struct to a subclass
10533 of gimple_opt_pass along with...
10534 (pass_data_all_optimizations): ...new pass_data instance and...
10535 (make_pass_all_optimizations): ...new function.
10536 (pass_all_optimizations_g): Convert from a global struct to a subclass
10537 of gimple_opt_pass along with...
10538 (pass_data_all_optimizations_g): ...new pass_data instance and...
10539 (make_pass_all_optimizations_g): ...new function.
10540 (pass_rest_of_compilation): Convert from a global struct to a subclass
10541 of rtl_opt_pass along with...
10542 (pass_data_rest_of_compilation): ...new pass_data instance and...
10543 (make_pass_rest_of_compilation): ...new function.
10544 (pass_postreload): Convert from a global struct to a subclass of
10545 rtl_opt_pass along with...
10546 (pass_data_postreload): ...new pass_data instance and...
10547 (make_pass_postreload): ...new function.
10548 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10549 subclass of rtl_opt_pass along with...
10550 (pass_data_gcse2): ...new pass_data instance and...
10551 (make_pass_gcse2): ...new function.
10552 * postreload.c (pass_postreload_cse): Convert from a global struct to
10553 a subclass of rtl_opt_pass along with...
10554 (pass_data_postreload_cse): ...new pass_data instance and...
10555 (make_pass_postreload_cse): ...new function.
10556 * predict.c (pass_profile): Convert from a global struct to a subclass
10557 of gimple_opt_pass along with...
10558 (pass_data_profile): ...new pass_data instance and...
10559 (make_pass_profile): ...new function.
10560 (pass_strip_predict_hints): Convert from a global struct to a subclass
10561 of gimple_opt_pass along with...
10562 (pass_data_strip_predict_hints): ...new pass_data instance and...
10563 (make_pass_strip_predict_hints): ...new function.
10564 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10565 of rtl_opt_pass along with...
10566 (pass_data_peephole2): ...new pass_data instance and...
10567 (make_pass_peephole2): ...new function.
10568 (pass_split_all_insns): Convert from a global struct to a subclass of
10569 rtl_opt_pass along with...
10570 (pass_data_split_all_insns): ...new pass_data instance and...
10571 (make_pass_split_all_insns): ...new function.
10572 (pass_split_after_reload): Convert from a global struct to a subclass
10573 of rtl_opt_pass along with...
10574 (pass_data_split_after_reload): ...new pass_data instance and...
10575 (make_pass_split_after_reload): ...new function.
10576 (pass_split_before_regstack): Convert from a global struct to a
10577 subclass of rtl_opt_pass along with...
10578 (pass_data_split_before_regstack): ...new pass_data instance and...
10579 (make_pass_split_before_regstack): ...new function.
10580 (pass_split_before_sched2): Convert from a global struct to a subclass
10581 of rtl_opt_pass along with...
10582 (pass_data_split_before_sched2): ...new pass_data instance and...
10583 (make_pass_split_before_sched2): ...new function.
10584 (pass_split_for_shorten_branches): Convert from a global struct to a
10585 subclass of rtl_opt_pass along with...
10586 (pass_data_split_for_shorten_branches): ...new pass_data instance
10588 (make_pass_split_for_shorten_branches): ...new function.
10589 * ree.c (pass_ree): Convert from a global struct to a subclass of
10590 rtl_opt_pass along with...
10591 (pass_data_ree): ...new pass_data instance and...
10592 (make_pass_ree): ...new function.
10593 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10594 subclass of rtl_opt_pass along with...
10595 (pass_data_stack_regs): ...new pass_data instance and...
10596 (make_pass_stack_regs): ...new function.
10597 (pass_stack_regs_run): Convert from a global struct to a subclass of
10598 rtl_opt_pass along with...
10599 (pass_data_stack_regs_run): ...new pass_data instance and...
10600 (make_pass_stack_regs_run): ...new function.
10601 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10602 subclass of rtl_opt_pass along with...
10603 (pass_data_cprop_hardreg): ...new pass_data instance and...
10604 (make_pass_cprop_hardreg): ...new function.
10605 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10606 subclass of rtl_opt_pass along with...
10607 (pass_data_reginfo_init): ...new pass_data instance and...
10608 (make_pass_reginfo_init): ...new function.
10609 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10610 of rtl_opt_pass along with...
10611 (pass_data_regmove): ...new pass_data instance and...
10612 (make_pass_regmove): ...new function.
10613 * regrename.c (pass_regrename): Convert from a global struct to a
10614 subclass of rtl_opt_pass along with...
10615 (pass_data_regrename): ...new pass_data instance and...
10616 (make_pass_regrename): ...new function.
10617 * reorg.c (pass_delay_slots): Convert from a global struct to a
10618 subclass of rtl_opt_pass along with...
10619 (pass_data_delay_slots): ...new pass_data instance and...
10620 (make_pass_delay_slots): ...new function.
10621 (pass_machine_reorg): Convert from a global struct to a subclass of
10622 rtl_opt_pass along with...
10623 (pass_data_machine_reorg): ...new pass_data instance and...
10624 (make_pass_machine_reorg): ...new function.
10625 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10626 of rtl_opt_pass along with...
10627 (pass_data_sched): ...new pass_data instance and...
10628 (make_pass_sched): ...new function.
10629 (pass_sched2): Convert from a global struct to a subclass of
10630 rtl_opt_pass along with...
10631 (pass_data_sched2): ...new pass_data instance and...
10632 (make_pass_sched2): ...new function.
10633 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10634 to a subclass of rtl_opt_pass along with...
10635 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10636 (make_pass_stack_ptr_mod): ...new function.
10637 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10638 to a subclass of rtl_opt_pass along with...
10639 (pass_data_rtl_store_motion): ...new pass_data instance and...
10640 (make_pass_rtl_store_motion): ...new function.
10641 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10642 of gimple_opt_pass along with...
10643 (pass_data_tracer): ...new pass_data instance and...
10644 (make_pass_tracer): ...new function.
10645 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10646 to a subclass of gimple_opt_pass along with...
10647 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10648 (make_pass_diagnose_tm_blocks): ...new function.
10649 (pass_lower_tm): Convert from a global struct to a subclass of
10650 gimple_opt_pass along with...
10651 (pass_data_lower_tm): ...new pass_data instance and...
10652 (make_pass_lower_tm): ...new function.
10653 (pass_tm_init): Convert from a global struct to a subclass of
10654 gimple_opt_pass along with...
10655 (pass_data_tm_init): ...new pass_data instance and...
10656 (make_pass_tm_init): ...new function.
10657 (pass_tm_mark): Convert from a global struct to a subclass of
10658 gimple_opt_pass along with...
10659 (pass_data_tm_mark): ...new pass_data instance and...
10660 (make_pass_tm_mark): ...new function.
10661 (pass_tm_edges): Convert from a global struct to a subclass of
10662 gimple_opt_pass along with...
10663 (pass_data_tm_edges): ...new pass_data instance and...
10664 (make_pass_tm_edges): ...new function.
10665 (pass_tm_memopt): Convert from a global struct to a subclass of
10666 gimple_opt_pass along with...
10667 (pass_data_tm_memopt): ...new pass_data instance and...
10668 (make_pass_tm_memopt): ...new function.
10669 (pass_ipa_tm): Convert from a global struct to a subclass of
10670 simple_ipa_opt_pass along with...
10671 (pass_data_ipa_tm): ...new pass_data instance and...
10672 (make_pass_ipa_tm): ...new function.
10673 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10674 subclass of gimple_opt_pass along with...
10675 (pass_data_call_cdce): ...new pass_data instance and...
10676 (make_pass_call_cdce): ...new function.
10677 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10678 subclass of gimple_opt_pass along with...
10679 (pass_data_build_cfg): ...new pass_data instance and...
10680 (make_pass_build_cfg): ...new function.
10681 (pass_split_crit_edges): Convert from a global struct to a subclass of
10682 gimple_opt_pass along with...
10683 (pass_data_split_crit_edges): ...new pass_data instance and...
10684 (make_pass_split_crit_edges): ...new function.
10685 (pass_warn_function_return): Convert from a global struct to a
10686 subclass of gimple_opt_pass along with...
10687 (pass_data_warn_function_return): ...new pass_data instance and...
10688 (make_pass_warn_function_return): ...new function.
10689 (pass_warn_function_noreturn): Convert from a global struct to a
10690 subclass of gimple_opt_pass along with...
10691 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10692 (make_pass_warn_function_noreturn): ...new function.
10693 (pass_warn_unused_result): Convert from a global struct to a subclass
10694 of gimple_opt_pass along with...
10695 (pass_data_warn_unused_result): ...new pass_data instance and...
10696 (make_pass_warn_unused_result): ...new function.
10697 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10698 a subclass of gimple_opt_pass along with...
10699 (pass_data_merge_phi): ...new pass_data instance and...
10700 (make_pass_merge_phi): ...new function.
10701 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10702 a subclass of gimple_opt_pass along with...
10703 (pass_data_lower_complex): ...new pass_data instance and...
10704 (make_pass_lower_complex): ...new function.
10705 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10706 gimple_opt_pass along with...
10707 (pass_data_lower_complex_O0): ...new pass_data instance and...
10708 (make_pass_lower_complex_O0): ...new function.
10709 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10710 subclass of gimple_opt_pass along with...
10711 (pass_data_lower_eh): ...new pass_data instance and...
10712 (make_pass_lower_eh): ...new function.
10713 (pass_refactor_eh): Convert from a global struct to a subclass of
10714 gimple_opt_pass along with...
10715 (pass_data_refactor_eh): ...new pass_data instance and...
10716 (make_pass_refactor_eh): ...new function.
10717 (pass_lower_resx): Convert from a global struct to a subclass of
10718 gimple_opt_pass along with...
10719 (pass_data_lower_resx): ...new pass_data instance and...
10720 (make_pass_lower_resx): ...new function.
10721 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10722 of gimple_opt_pass along with...
10723 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10724 (make_pass_lower_eh_dispatch): ...new function.
10725 (pass_cleanup_eh): Convert from a global struct to a subclass of
10726 gimple_opt_pass along with...
10727 (pass_data_cleanup_eh): ...new pass_data instance and...
10728 (make_pass_cleanup_eh): ...new function.
10729 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10730 to a subclass of simple_ipa_opt_pass along with...
10731 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10732 (make_pass_ipa_lower_emutls): ...new function.
10733 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10734 a subclass of gimple_opt_pass along with...
10735 (pass_data_if_conversion): ...new pass_data instance and...
10736 (make_pass_if_conversion): ...new function.
10737 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10738 subclass of gimple_opt_pass along with...
10739 (pass_data_build_ssa): ...new pass_data instance and...
10740 (make_pass_build_ssa): ...new function.
10741 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10742 global struct to a subclass of gimple_opt_pass along with...
10743 (pass_data_loop_distribution): ...new pass_data instance and...
10744 (make_pass_loop_distribution): ...new function.
10745 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10746 subclass of gimple_opt_pass along with...
10747 (pass_data_mudflap_1): ...new pass_data instance and...
10748 (make_pass_mudflap_1): ...new function.
10749 (pass_mudflap_2): Convert from a global struct to a subclass of
10750 gimple_opt_pass along with...
10751 (pass_data_mudflap_2): ...new pass_data instance and...
10752 (make_pass_mudflap_2): ...new function.
10753 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10754 subclass of gimple_opt_pass along with...
10755 (pass_data_mudflap_1): ...new pass_data instance and...
10756 (make_pass_mudflap_1): ...new function.
10757 (pass_mudflap_2): Convert from a global struct to a subclass of
10758 gimple_opt_pass along with...
10759 (pass_data_mudflap_2): ...new pass_data instance and...
10760 (make_pass_mudflap_2): ...new function.
10761 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10762 gimple_opt_pass along with...
10763 (pass_data_nrv): ...new pass_data instance and...
10764 (make_pass_nrv): ...new function.
10765 (pass_return_slot): Convert from a global struct to a subclass of
10766 gimple_opt_pass along with...
10767 (pass_data_return_slot): ...new pass_data instance and...
10768 (make_pass_return_slot): ...new function.
10769 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10770 to a subclass of gimple_opt_pass along with...
10771 (pass_data_object_sizes): ...new pass_data instance and...
10772 (make_pass_object_sizes): ...new function.
10773 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10774 global struct to a subclass of gimple_opt_pass along with...
10775 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10777 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10778 (pass_fixup_cfg): Convert from a global struct to a subclass of
10779 gimple_opt_pass along with...
10780 (pass_data_fixup_cfg): ...new pass_data instance and...
10781 (make_pass_fixup_cfg): ...new function.
10782 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10783 (make_pass_mudflap_1): ...new function.
10784 (pass_mudflap_2): Replace declaration with that of...
10785 (make_pass_mudflap_2): ...new function.
10786 (pass_asan): Replace declaration with that of...
10787 (make_pass_asan): ...new function.
10788 (pass_asan_O0): Replace declaration with that of...
10789 (make_pass_asan_O0): ...new function.
10790 (pass_tsan): Replace declaration with that of...
10791 (make_pass_tsan): ...new function.
10792 (pass_tsan_O0): Replace declaration with that of...
10793 (make_pass_tsan_O0): ...new function.
10794 (pass_lower_cf): Replace declaration with that of...
10795 (make_pass_lower_cf): ...new function.
10796 (pass_refactor_eh): Replace declaration with that of...
10797 (make_pass_refactor_eh): ...new function.
10798 (pass_lower_eh): Replace declaration with that of...
10799 (make_pass_lower_eh): ...new function.
10800 (pass_lower_eh_dispatch): Replace declaration with that of...
10801 (make_pass_lower_eh_dispatch): ...new function.
10802 (pass_lower_resx): Replace declaration with that of...
10803 (make_pass_lower_resx): ...new function.
10804 (pass_build_cfg): Replace declaration with that of...
10805 (make_pass_build_cfg): ...new function.
10806 (pass_early_tree_profile): Replace declaration with that of...
10807 (make_pass_early_tree_profile): ...new function.
10808 (pass_cleanup_eh): Replace declaration with that of...
10809 (make_pass_cleanup_eh): ...new function.
10810 (pass_sra): Replace declaration with that of...
10811 (make_pass_sra): ...new function.
10812 (pass_sra_early): Replace declaration with that of...
10813 (make_pass_sra_early): ...new function.
10814 (pass_early_ipa_sra): Replace declaration with that of...
10815 (make_pass_early_ipa_sra): ...new function.
10816 (pass_tail_recursion): Replace declaration with that of...
10817 (make_pass_tail_recursion): ...new function.
10818 (pass_tail_calls): Replace declaration with that of...
10819 (make_pass_tail_calls): ...new function.
10820 (pass_tree_loop): Replace declaration with that of...
10821 (make_pass_tree_loop): ...new function.
10822 (pass_tree_loop_init): Replace declaration with that of...
10823 (make_pass_tree_loop_init): ...new function.
10824 (pass_lim): Replace declaration with that of...
10825 (make_pass_lim): ...new function.
10826 (pass_tree_unswitch): Replace declaration with that of...
10827 (make_pass_tree_unswitch): ...new function.
10828 (pass_predcom): Replace declaration with that of...
10829 (make_pass_predcom): ...new function.
10830 (pass_iv_canon): Replace declaration with that of...
10831 (make_pass_iv_canon): ...new function.
10832 (pass_scev_cprop): Replace declaration with that of...
10833 (make_pass_scev_cprop): ...new function.
10834 (pass_empty_loop): Replace declaration with that of...
10835 (make_pass_empty_loop): ...new function.
10836 (pass_record_bounds): Replace declaration with that of...
10837 (make_pass_record_bounds): ...new function.
10838 (pass_graphite): Replace declaration with that of...
10839 (make_pass_graphite): ...new function.
10840 (pass_graphite_transforms): Replace declaration with that of...
10841 (make_pass_graphite_transforms): ...new function.
10842 (pass_if_conversion): Replace declaration with that of...
10843 (make_pass_if_conversion): ...new function.
10844 (pass_loop_distribution): Replace declaration with that of...
10845 (make_pass_loop_distribution): ...new function.
10846 (pass_vectorize): Replace declaration with that of...
10847 (make_pass_vectorize): ...new function.
10848 (pass_slp_vectorize): Replace declaration with that of...
10849 (make_pass_slp_vectorize): ...new function.
10850 (pass_complete_unroll): Replace declaration with that of...
10851 (make_pass_complete_unroll): ...new function.
10852 (pass_complete_unrolli): Replace declaration with that of...
10853 (make_pass_complete_unrolli): ...new function.
10854 (pass_parallelize_loops): Replace declaration with that of...
10855 (make_pass_parallelize_loops): ...new function.
10856 (pass_loop_prefetch): Replace declaration with that of...
10857 (make_pass_loop_prefetch): ...new function.
10858 (pass_iv_optimize): Replace declaration with that of...
10859 (make_pass_iv_optimize): ...new function.
10860 (pass_tree_loop_done): Replace declaration with that of...
10861 (make_pass_tree_loop_done): ...new function.
10862 (pass_ch): Replace declaration with that of...
10863 (make_pass_ch): ...new function.
10864 (pass_ccp): Replace declaration with that of...
10865 (make_pass_ccp): ...new function.
10866 (pass_phi_only_cprop): Replace declaration with that of...
10867 (make_pass_phi_only_cprop): ...new function.
10868 (pass_build_ssa): Replace declaration with that of...
10869 (make_pass_build_ssa): ...new function.
10870 (pass_build_alias): Replace declaration with that of...
10871 (make_pass_build_alias): ...new function.
10872 (pass_build_ealias): Replace declaration with that of...
10873 (make_pass_build_ealias): ...new function.
10874 (pass_dominator): Replace declaration with that of...
10875 (make_pass_dominator): ...new function.
10876 (pass_dce): Replace declaration with that of...
10877 (make_pass_dce): ...new function.
10878 (pass_dce_loop): Replace declaration with that of...
10879 (make_pass_dce_loop): ...new function.
10880 (pass_cd_dce): Replace declaration with that of...
10881 (make_pass_cd_dce): ...new function.
10882 (pass_call_cdce): Replace declaration with that of...
10883 (make_pass_call_cdce): ...new function.
10884 (pass_merge_phi): Replace declaration with that of...
10885 (make_pass_merge_phi): ...new function.
10886 (pass_split_crit_edges): Replace declaration with that of...
10887 (make_pass_split_crit_edges): ...new function.
10888 (pass_pre): Replace declaration with that of...
10889 (make_pass_pre): ...new function.
10890 (pass_profile): Replace declaration with that of...
10891 (make_pass_profile): ...new function.
10892 (pass_strip_predict_hints): Replace declaration with that of...
10893 (make_pass_strip_predict_hints): ...new function.
10894 (pass_lower_complex_O0): Replace declaration with that of...
10895 (make_pass_lower_complex_O0): ...new function.
10896 (pass_lower_complex): Replace declaration with that of...
10897 (make_pass_lower_complex): ...new function.
10898 (pass_lower_vector): Replace declaration with that of...
10899 (make_pass_lower_vector): ...new function.
10900 (pass_lower_vector_ssa): Replace declaration with that of...
10901 (make_pass_lower_vector_ssa): ...new function.
10902 (pass_lower_omp): Replace declaration with that of...
10903 (make_pass_lower_omp): ...new function.
10904 (pass_diagnose_omp_blocks): Replace declaration with that of...
10905 (make_pass_diagnose_omp_blocks): ...new function.
10906 (pass_expand_omp): Replace declaration with that of...
10907 (make_pass_expand_omp): ...new function.
10908 (pass_expand_omp_ssa): Replace declaration with that of...
10909 (make_pass_expand_omp_ssa): ...new function.
10910 (pass_object_sizes): Replace declaration with that of...
10911 (make_pass_object_sizes): ...new function.
10912 (pass_strlen): Replace declaration with that of...
10913 (make_pass_strlen): ...new function.
10914 (pass_fold_builtins): Replace declaration with that of...
10915 (make_pass_fold_builtins): ...new function.
10916 (pass_stdarg): Replace declaration with that of...
10917 (make_pass_stdarg): ...new function.
10918 (pass_early_warn_uninitialized): Replace declaration with that of...
10919 (make_pass_early_warn_uninitialized): ...new function.
10920 (pass_late_warn_uninitialized): Replace declaration with that of...
10921 (make_pass_late_warn_uninitialized): ...new function.
10922 (pass_cse_reciprocals): Replace declaration with that of...
10923 (make_pass_cse_reciprocals): ...new function.
10924 (pass_cse_sincos): Replace declaration with that of...
10925 (make_pass_cse_sincos): ...new function.
10926 (pass_optimize_bswap): Replace declaration with that of...
10927 (make_pass_optimize_bswap): ...new function.
10928 (pass_optimize_widening_mul): Replace declaration with that of...
10929 (make_pass_optimize_widening_mul): ...new function.
10930 (pass_warn_function_return): Replace declaration with that of...
10931 (make_pass_warn_function_return): ...new function.
10932 (pass_warn_function_noreturn): Replace declaration with that of...
10933 (make_pass_warn_function_noreturn): ...new function.
10934 (pass_cselim): Replace declaration with that of...
10935 (make_pass_cselim): ...new function.
10936 (pass_phiopt): Replace declaration with that of...
10937 (make_pass_phiopt): ...new function.
10938 (pass_forwprop): Replace declaration with that of...
10939 (make_pass_forwprop): ...new function.
10940 (pass_phiprop): Replace declaration with that of...
10941 (make_pass_phiprop): ...new function.
10942 (pass_tree_ifcombine): Replace declaration with that of...
10943 (make_pass_tree_ifcombine): ...new function.
10944 (pass_dse): Replace declaration with that of...
10945 (make_pass_dse): ...new function.
10946 (pass_nrv): Replace declaration with that of...
10947 (make_pass_nrv): ...new function.
10948 (pass_rename_ssa_copies): Replace declaration with that of...
10949 (make_pass_rename_ssa_copies): ...new function.
10950 (pass_sink_code): Replace declaration with that of...
10951 (make_pass_sink_code): ...new function.
10952 (pass_fre): Replace declaration with that of...
10953 (make_pass_fre): ...new function.
10954 (pass_check_data_deps): Replace declaration with that of...
10955 (make_pass_check_data_deps): ...new function.
10956 (pass_copy_prop): Replace declaration with that of...
10957 (make_pass_copy_prop): ...new function.
10958 (pass_vrp): Replace declaration with that of...
10959 (make_pass_vrp): ...new function.
10960 (pass_uncprop): Replace declaration with that of...
10961 (make_pass_uncprop): ...new function.
10962 (pass_return_slot): Replace declaration with that of...
10963 (make_pass_return_slot): ...new function.
10964 (pass_reassoc): Replace declaration with that of...
10965 (make_pass_reassoc): ...new function.
10966 (pass_rebuild_cgraph_edges): Replace declaration with that of...
10967 (make_pass_rebuild_cgraph_edges): ...new function.
10968 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
10969 (make_pass_remove_cgraph_callee_edges): ...new function.
10970 (pass_build_cgraph_edges): Replace declaration with that of...
10971 (make_pass_build_cgraph_edges): ...new function.
10972 (pass_local_pure_const): Replace declaration with that of...
10973 (make_pass_local_pure_const): ...new function.
10974 (pass_tracer): Replace declaration with that of...
10975 (make_pass_tracer): ...new function.
10976 (pass_warn_unused_result): Replace declaration with that of...
10977 (make_pass_warn_unused_result): ...new function.
10978 (pass_diagnose_tm_blocks): Replace declaration with that of...
10979 (make_pass_diagnose_tm_blocks): ...new function.
10980 (pass_lower_tm): Replace declaration with that of...
10981 (make_pass_lower_tm): ...new function.
10982 (pass_tm_init): Replace declaration with that of...
10983 (make_pass_tm_init): ...new function.
10984 (pass_tm_mark): Replace declaration with that of...
10985 (make_pass_tm_mark): ...new function.
10986 (pass_tm_memopt): Replace declaration with that of...
10987 (make_pass_tm_memopt): ...new function.
10988 (pass_tm_edges): Replace declaration with that of...
10989 (make_pass_tm_edges): ...new function.
10990 (pass_split_functions): Replace declaration with that of...
10991 (make_pass_split_functions): ...new function.
10992 (pass_feedback_split_functions): Replace declaration with that of...
10993 (make_pass_feedback_split_functions): ...new function.
10994 (pass_strength_reduction): Replace declaration with that of...
10995 (make_pass_strength_reduction): ...new function.
10996 (pass_ipa_lower_emutls): Replace declaration with that of...
10997 (make_pass_ipa_lower_emutls): ...new function.
10998 (pass_ipa_function_and_variable_visibility): Replace declaration with
11000 (make_pass_ipa_function_and_variable_visibility): ...new function.
11001 (pass_ipa_tree_profile): Replace declaration with that of...
11002 (make_pass_ipa_tree_profile): ...new function.
11003 (pass_early_local_passes): Replace declaration with that of...
11004 (make_pass_early_local_passes): ...new function.
11005 (pass_ipa_whole_program_visibility): Replace declaration with that
11007 (make_pass_ipa_whole_program_visibility): ...new function.
11008 (pass_ipa_lto_gimple_out): Replace declaration with that of...
11009 (make_pass_ipa_lto_gimple_out): ...new function.
11010 (pass_ipa_increase_alignment): Replace declaration with that of...
11011 (make_pass_ipa_increase_alignment): ...new function.
11012 (pass_ipa_inline): Replace declaration with that of...
11013 (make_pass_ipa_inline): ...new function.
11014 (pass_ipa_free_lang_data): Replace declaration with that of...
11015 (make_pass_ipa_free_lang_data): ...new function.
11016 (pass_ipa_free_inline_summary): Replace declaration with that of...
11017 (make_pass_ipa_free_inline_summary): ...new function.
11018 (pass_ipa_cp): Replace declaration with that of...
11019 (make_pass_ipa_cp): ...new function.
11020 (pass_ipa_reference): Replace declaration with that of...
11021 (make_pass_ipa_reference): ...new function.
11022 (pass_ipa_pure_const): Replace declaration with that of...
11023 (make_pass_ipa_pure_const): ...new function.
11024 (pass_ipa_pta): Replace declaration with that of...
11025 (make_pass_ipa_pta): ...new function.
11026 (pass_ipa_lto_finish_out): Replace declaration with that of...
11027 (make_pass_ipa_lto_finish_out): ...new function.
11028 (pass_ipa_tm): Replace declaration with that of...
11029 (make_pass_ipa_tm): ...new function.
11030 (pass_ipa_profile): Replace declaration with that of...
11031 (make_pass_ipa_profile): ...new function.
11032 (pass_ipa_cdtor_merge): Replace declaration with that of...
11033 (make_pass_ipa_cdtor_merge): ...new function.
11034 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
11036 (make_pass_cleanup_cfg_post_optimizing): ...new function.
11037 (pass_init_datastructures): Replace declaration with that of...
11038 (make_pass_init_datastructures): ...new function.
11039 (pass_fixup_cfg): Replace declaration with that of...
11040 (make_pass_fixup_cfg): ...new function.
11041 (pass_expand): Replace declaration with that of...
11042 (make_pass_expand): ...new function.
11043 (pass_instantiate_virtual_regs): Replace declaration with that of...
11044 (make_pass_instantiate_virtual_regs): ...new function.
11045 (pass_rtl_fwprop): Replace declaration with that of...
11046 (make_pass_rtl_fwprop): ...new function.
11047 (pass_rtl_fwprop_addr): Replace declaration with that of...
11048 (make_pass_rtl_fwprop_addr): ...new function.
11049 (pass_jump): Replace declaration with that of...
11050 (make_pass_jump): ...new function.
11051 (pass_jump2): Replace declaration with that of...
11052 (make_pass_jump2): ...new function.
11053 (pass_lower_subreg): Replace declaration with that of...
11054 (make_pass_lower_subreg): ...new function.
11055 (pass_cse): Replace declaration with that of...
11056 (make_pass_cse): ...new function.
11057 (pass_fast_rtl_dce): Replace declaration with that of...
11058 (make_pass_fast_rtl_dce): ...new function.
11059 (pass_ud_rtl_dce): Replace declaration with that of...
11060 (make_pass_ud_rtl_dce): ...new function.
11061 (pass_rtl_dce): Replace declaration with that of...
11062 (make_pass_rtl_dce): ...new function.
11063 (pass_rtl_dse1): Replace declaration with that of...
11064 (make_pass_rtl_dse1): ...new function.
11065 (pass_rtl_dse2): Replace declaration with that of...
11066 (make_pass_rtl_dse2): ...new function.
11067 (pass_rtl_dse3): Replace declaration with that of...
11068 (make_pass_rtl_dse3): ...new function.
11069 (pass_rtl_cprop): Replace declaration with that of...
11070 (make_pass_rtl_cprop): ...new function.
11071 (pass_rtl_pre): Replace declaration with that of...
11072 (make_pass_rtl_pre): ...new function.
11073 (pass_rtl_hoist): Replace declaration with that of...
11074 (make_pass_rtl_hoist): ...new function.
11075 (pass_rtl_store_motion): Replace declaration with that of...
11076 (make_pass_rtl_store_motion): ...new function.
11077 (pass_cse_after_global_opts): Replace declaration with that of...
11078 (make_pass_cse_after_global_opts): ...new function.
11079 (pass_rtl_ifcvt): Replace declaration with that of...
11080 (make_pass_rtl_ifcvt): ...new function.
11081 (pass_into_cfg_layout_mode): Replace declaration with that of...
11082 (make_pass_into_cfg_layout_mode): ...new function.
11083 (pass_outof_cfg_layout_mode): Replace declaration with that of...
11084 (make_pass_outof_cfg_layout_mode): ...new function.
11085 (pass_loop2): Replace declaration with that of...
11086 (make_pass_loop2): ...new function.
11087 (pass_rtl_loop_init): Replace declaration with that of...
11088 (make_pass_rtl_loop_init): ...new function.
11089 (pass_rtl_move_loop_invariants): Replace declaration with that of...
11090 (make_pass_rtl_move_loop_invariants): ...new function.
11091 (pass_rtl_unswitch): Replace declaration with that of...
11092 (make_pass_rtl_unswitch): ...new function.
11093 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
11094 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11095 (pass_rtl_doloop): Replace declaration with that of...
11096 (make_pass_rtl_doloop): ...new function.
11097 (pass_rtl_loop_done): Replace declaration with that of...
11098 (make_pass_rtl_loop_done): ...new function.
11099 (pass_web): Replace declaration with that of...
11100 (make_pass_web): ...new function.
11101 (pass_cse2): Replace declaration with that of...
11102 (make_pass_cse2): ...new function.
11103 (pass_df_initialize_opt): Replace declaration with that of...
11104 (make_pass_df_initialize_opt): ...new function.
11105 (pass_df_initialize_no_opt): Replace declaration with that of...
11106 (make_pass_df_initialize_no_opt): ...new function.
11107 (pass_reginfo_init): Replace declaration with that of...
11108 (make_pass_reginfo_init): ...new function.
11109 (pass_inc_dec): Replace declaration with that of...
11110 (make_pass_inc_dec): ...new function.
11111 (pass_stack_ptr_mod): Replace declaration with that of...
11112 (make_pass_stack_ptr_mod): ...new function.
11113 (pass_initialize_regs): Replace declaration with that of...
11114 (make_pass_initialize_regs): ...new function.
11115 (pass_combine): Replace declaration with that of...
11116 (make_pass_combine): ...new function.
11117 (pass_if_after_combine): Replace declaration with that of...
11118 (make_pass_if_after_combine): ...new function.
11119 (pass_ree): Replace declaration with that of...
11120 (make_pass_ree): ...new function.
11121 (pass_partition_blocks): Replace declaration with that of...
11122 (make_pass_partition_blocks): ...new function.
11123 (pass_match_asm_constraints): Replace declaration with that of...
11124 (make_pass_match_asm_constraints): ...new function.
11125 (pass_regmove): Replace declaration with that of...
11126 (make_pass_regmove): ...new function.
11127 (pass_split_all_insns): Replace declaration with that of...
11128 (make_pass_split_all_insns): ...new function.
11129 (pass_fast_rtl_byte_dce): Replace declaration with that of...
11130 (make_pass_fast_rtl_byte_dce): ...new function.
11131 (pass_lower_subreg2): Replace declaration with that of...
11132 (make_pass_lower_subreg2): ...new function.
11133 (pass_mode_switching): Replace declaration with that of...
11134 (make_pass_mode_switching): ...new function.
11135 (pass_sms): Replace declaration with that of...
11136 (make_pass_sms): ...new function.
11137 (pass_sched): Replace declaration with that of...
11138 (make_pass_sched): ...new function.
11139 (pass_ira): Replace declaration with that of...
11140 (make_pass_ira): ...new function.
11141 (pass_reload): Replace declaration with that of...
11142 (make_pass_reload): ...new function.
11143 (pass_clean_state): Replace declaration with that of...
11144 (make_pass_clean_state): ...new function.
11145 (pass_branch_prob): Replace declaration with that of...
11146 (make_pass_branch_prob): ...new function.
11147 (pass_value_profile_transformations): Replace declaration with that
11149 (make_pass_value_profile_transformations): ...new function.
11150 (pass_postreload_cse): Replace declaration with that of...
11151 (make_pass_postreload_cse): ...new function.
11152 (pass_gcse2): Replace declaration with that of...
11153 (make_pass_gcse2): ...new function.
11154 (pass_split_after_reload): Replace declaration with that of...
11155 (make_pass_split_after_reload): ...new function.
11156 (pass_branch_target_load_optimize1): Replace declaration with that
11158 (make_pass_branch_target_load_optimize1): ...new function.
11159 (pass_thread_prologue_and_epilogue): Replace declaration with that
11161 (make_pass_thread_prologue_and_epilogue): ...new function.
11162 (pass_stack_adjustments): Replace declaration with that of...
11163 (make_pass_stack_adjustments): ...new function.
11164 (pass_peephole2): Replace declaration with that of...
11165 (make_pass_peephole2): ...new function.
11166 (pass_if_after_reload): Replace declaration with that of...
11167 (make_pass_if_after_reload): ...new function.
11168 (pass_regrename): Replace declaration with that of...
11169 (make_pass_regrename): ...new function.
11170 (pass_cprop_hardreg): Replace declaration with that of...
11171 (make_pass_cprop_hardreg): ...new function.
11172 (pass_reorder_blocks): Replace declaration with that of...
11173 (make_pass_reorder_blocks): ...new function.
11174 (pass_branch_target_load_optimize2): Replace declaration with that
11176 (make_pass_branch_target_load_optimize2): ...new function.
11177 (pass_leaf_regs): Replace declaration with that of...
11178 (make_pass_leaf_regs): ...new function.
11179 (pass_split_before_sched2): Replace declaration with that of...
11180 (make_pass_split_before_sched2): ...new function.
11181 (pass_compare_elim_after_reload): Replace declaration with that of...
11182 (make_pass_compare_elim_after_reload): ...new function.
11183 (pass_sched2): Replace declaration with that of...
11184 (make_pass_sched2): ...new function.
11185 (pass_stack_regs): Replace declaration with that of...
11186 (make_pass_stack_regs): ...new function.
11187 (pass_stack_regs_run): Replace declaration with that of...
11188 (make_pass_stack_regs_run): ...new function.
11189 (pass_df_finish): Replace declaration with that of...
11190 (make_pass_df_finish): ...new function.
11191 (pass_compute_alignments): Replace declaration with that of...
11192 (make_pass_compute_alignments): ...new function.
11193 (pass_duplicate_computed_gotos): Replace declaration with that of...
11194 (make_pass_duplicate_computed_gotos): ...new function.
11195 (pass_variable_tracking): Replace declaration with that of...
11196 (make_pass_variable_tracking): ...new function.
11197 (pass_free_cfg): Replace declaration with that of...
11198 (make_pass_free_cfg): ...new function.
11199 (pass_machine_reorg): Replace declaration with that of...
11200 (make_pass_machine_reorg): ...new function.
11201 (pass_cleanup_barriers): Replace declaration with that of...
11202 (make_pass_cleanup_barriers): ...new function.
11203 (pass_delay_slots): Replace declaration with that of...
11204 (make_pass_delay_slots): ...new function.
11205 (pass_split_for_shorten_branches): Replace declaration with that of...
11206 (make_pass_split_for_shorten_branches): ...new function.
11207 (pass_split_before_regstack): Replace declaration with that of...
11208 (make_pass_split_before_regstack): ...new function.
11209 (pass_convert_to_eh_region_ranges): Replace declaration with that
11211 (make_pass_convert_to_eh_region_ranges): ...new function.
11212 (pass_shorten_branches): Replace declaration with that of...
11213 (make_pass_shorten_branches): ...new function.
11214 (pass_set_nothrow_function_flags): Replace declaration with that of...
11215 (make_pass_set_nothrow_function_flags): ...new function.
11216 (pass_dwarf2_frame): Replace declaration with that of...
11217 (make_pass_dwarf2_frame): ...new function.
11218 (pass_final): Replace declaration with that of...
11219 (make_pass_final): ...new function.
11220 (pass_rtl_seqabstr): Replace declaration with that of...
11221 (make_pass_rtl_seqabstr): ...new function.
11222 (pass_release_ssa_names): Replace declaration with that of...
11223 (make_pass_release_ssa_names): ...new function.
11224 (pass_early_inline): Replace declaration with that of...
11225 (make_pass_early_inline): ...new function.
11226 (pass_inline_parameters): Replace declaration with that of...
11227 (make_pass_inline_parameters): ...new function.
11228 (pass_update_address_taken): Replace declaration with that of...
11229 (make_pass_update_address_taken): ...new function.
11230 (pass_convert_switch): Replace declaration with that of...
11231 (make_pass_convert_switch): ...new function.
11232 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11233 to a subclass of simple_ipa_opt_pass along with...
11234 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11235 (make_pass_ipa_tree_profile): ...new function.
11236 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11237 subclass of gimple_opt_pass along with...
11238 (pass_data_sra_early): ...new pass_data instance and...
11239 (make_pass_sra_early): ...new function.
11240 (pass_sra): Convert from a global struct to a subclass of
11241 gimple_opt_pass along with...
11242 (pass_data_sra): ...new pass_data instance and...
11243 (make_pass_sra): ...new function.
11244 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11245 gimple_opt_pass along with...
11246 (pass_data_early_ipa_sra): ...new pass_data instance and...
11247 (make_pass_early_ipa_sra): ...new function.
11248 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11249 subclass of gimple_opt_pass along with...
11250 (pass_data_ccp): ...new pass_data instance and...
11251 (make_pass_ccp): ...new function.
11252 (pass_fold_builtins): Convert from a global struct to a subclass of
11253 gimple_opt_pass along with...
11254 (pass_data_fold_builtins): ...new pass_data instance and...
11255 (make_pass_fold_builtins): ...new function.
11256 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11257 subclass of gimple_opt_pass along with...
11258 (pass_data_copy_prop): ...new pass_data instance and...
11259 (make_pass_copy_prop): ...new function.
11260 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11261 global struct to a subclass of gimple_opt_pass along with...
11262 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11263 (make_pass_rename_ssa_copies): ...new function.
11264 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11265 subclass of gimple_opt_pass along with...
11266 (pass_data_dce): ...new pass_data instance and...
11267 (make_pass_dce): ...new function.
11268 (pass_dce_loop): Convert from a global struct to a subclass of
11269 gimple_opt_pass along with...
11270 (pass_data_dce_loop): ...new pass_data instance and...
11271 (make_pass_dce_loop): ...new function.
11272 (pass_cd_dce): Convert from a global struct to a subclass of
11273 gimple_opt_pass along with...
11274 (pass_data_cd_dce): ...new pass_data instance and...
11275 (make_pass_cd_dce): ...new function.
11276 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11277 subclass of gimple_opt_pass along with...
11278 (pass_data_dominator): ...new pass_data instance and...
11279 (make_pass_dominator): ...new function.
11280 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11281 gimple_opt_pass along with...
11282 (pass_data_phi_only_cprop): ...new pass_data instance and...
11283 (make_pass_phi_only_cprop): ...new function.
11284 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11285 subclass of gimple_opt_pass along with...
11286 (pass_data_dse): ...new pass_data instance and...
11287 (make_pass_dse): ...new function.
11288 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11289 a subclass of gimple_opt_pass along with...
11290 (pass_data_forwprop): ...new pass_data instance and...
11291 (make_pass_forwprop): ...new function.
11292 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11293 struct to a subclass of gimple_opt_pass along with...
11294 (pass_data_tree_ifcombine): ...new pass_data instance and...
11295 (make_pass_tree_ifcombine): ...new function.
11296 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11297 subclass of gimple_opt_pass along with...
11298 (pass_data_ch): ...new pass_data instance and...
11299 (make_pass_ch): ...new function.
11300 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11301 subclass of gimple_opt_pass along with...
11302 (pass_data_tree_loop): ...new pass_data instance and...
11303 (make_pass_tree_loop): ...new function.
11304 (pass_tree_loop_init): Convert from a global struct to a subclass of
11305 gimple_opt_pass along with...
11306 (pass_data_tree_loop_init): ...new pass_data instance and...
11307 (make_pass_tree_loop_init): ...new function.
11308 (pass_lim): Convert from a global struct to a subclass of
11309 gimple_opt_pass along with...
11310 (pass_data_lim): ...new pass_data instance and...
11311 (make_pass_lim): ...new function.
11312 (pass_tree_unswitch): Convert from a global struct to a subclass of
11313 gimple_opt_pass along with...
11314 (pass_data_tree_unswitch): ...new pass_data instance and...
11315 (make_pass_tree_unswitch): ...new function.
11316 (pass_predcom): Convert from a global struct to a subclass of
11317 gimple_opt_pass along with...
11318 (pass_data_predcom): ...new pass_data instance and...
11319 (make_pass_predcom): ...new function.
11320 (pass_vectorize): Convert from a global struct to a subclass of
11321 gimple_opt_pass along with...
11322 (pass_data_vectorize): ...new pass_data instance and...
11323 (make_pass_vectorize): ...new function.
11324 (pass_graphite): Convert from a global struct to a subclass of
11325 gimple_opt_pass along with...
11326 (pass_data_graphite): ...new pass_data instance and...
11327 (make_pass_graphite): ...new function.
11328 (pass_graphite_transforms): Convert from a global struct to a subclass
11329 of gimple_opt_pass along with...
11330 (pass_data_graphite_transforms): ...new pass_data instance and...
11331 (make_pass_graphite_transforms): ...new function.
11332 (pass_check_data_deps): Convert from a global struct to a subclass of
11333 gimple_opt_pass along with...
11334 (pass_data_check_data_deps): ...new pass_data instance and...
11335 (make_pass_check_data_deps): ...new function.
11336 (pass_iv_canon): Convert from a global struct to a subclass of
11337 gimple_opt_pass along with...
11338 (pass_data_iv_canon): ...new pass_data instance and...
11339 (make_pass_iv_canon): ...new function.
11340 (pass_scev_cprop): Convert from a global struct to a subclass of
11341 gimple_opt_pass along with...
11342 (pass_data_scev_cprop): ...new pass_data instance and...
11343 (make_pass_scev_cprop): ...new function.
11344 (pass_record_bounds): Convert from a global struct to a subclass of
11345 gimple_opt_pass along with...
11346 (pass_data_record_bounds): ...new pass_data instance and...
11347 (make_pass_record_bounds): ...new function.
11348 (pass_complete_unroll): Convert from a global struct to a subclass of
11349 gimple_opt_pass along with...
11350 (pass_data_complete_unroll): ...new pass_data instance and...
11351 (make_pass_complete_unroll): ...new function.
11352 (pass_complete_unrolli): Convert from a global struct to a subclass of
11353 gimple_opt_pass along with...
11354 (pass_data_complete_unrolli): ...new pass_data instance and...
11355 (make_pass_complete_unrolli): ...new function.
11356 (pass_parallelize_loops): Convert from a global struct to a subclass
11357 of gimple_opt_pass along with...
11358 (pass_data_parallelize_loops): ...new pass_data instance and...
11359 (make_pass_parallelize_loops): ...new function.
11360 (pass_loop_prefetch): Convert from a global struct to a subclass of
11361 gimple_opt_pass along with...
11362 (pass_data_loop_prefetch): ...new pass_data instance and...
11363 (make_pass_loop_prefetch): ...new function.
11364 (pass_iv_optimize): Convert from a global struct to a subclass of
11365 gimple_opt_pass along with...
11366 (pass_data_iv_optimize): ...new pass_data instance and...
11367 (make_pass_iv_optimize): ...new function.
11368 (pass_tree_loop_done): Convert from a global struct to a subclass of
11369 gimple_opt_pass along with...
11370 (pass_data_tree_loop_done): ...new pass_data instance and...
11371 (make_pass_tree_loop_done): ...new function.
11372 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11373 struct to a subclass of gimple_opt_pass along with...
11374 (pass_data_cse_reciprocals): ...new pass_data instance and...
11375 (make_pass_cse_reciprocals): ...new function.
11376 (pass_cse_sincos): Convert from a global struct to a subclass of
11377 gimple_opt_pass along with...
11378 (pass_data_cse_sincos): ...new pass_data instance and...
11379 (make_pass_cse_sincos): ...new function.
11380 (pass_optimize_bswap): Convert from a global struct to a subclass of
11381 gimple_opt_pass along with...
11382 (pass_data_optimize_bswap): ...new pass_data instance and...
11383 (make_pass_optimize_bswap): ...new function.
11384 (pass_optimize_widening_mul): Convert from a global struct to a
11385 subclass of gimple_opt_pass along with...
11386 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11387 (make_pass_optimize_widening_mul): ...new function.
11388 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11389 subclass of gimple_opt_pass along with...
11390 (pass_data_phiopt): ...new pass_data instance and...
11391 (make_pass_phiopt): ...new function.
11392 (pass_cselim): Convert from a global struct to a subclass of
11393 gimple_opt_pass along with...
11394 (pass_data_cselim): ...new pass_data instance and...
11395 (make_pass_cselim): ...new function.
11396 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11397 subclass of gimple_opt_pass along with...
11398 (pass_data_phiprop): ...new pass_data instance and...
11399 (make_pass_phiprop): ...new function.
11400 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11401 subclass of gimple_opt_pass along with...
11402 (pass_data_pre): ...new pass_data instance and...
11403 (make_pass_pre): ...new function.
11404 (pass_fre): Convert from a global struct to a subclass of
11405 gimple_opt_pass along with...
11406 (pass_data_fre): ...new pass_data instance and...
11407 (make_pass_fre): ...new function.
11408 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11409 subclass of gimple_opt_pass along with...
11410 (pass_data_reassoc): ...new pass_data instance and...
11411 (make_pass_reassoc): ...new function.
11412 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11413 subclass of gimple_opt_pass along with...
11414 (pass_data_sink_code): ...new pass_data instance and...
11415 (make_pass_sink_code): ...new function.
11416 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11417 subclass of gimple_opt_pass along with...
11418 (pass_data_strlen): ...new pass_data instance and...
11419 (make_pass_strlen): ...new function.
11420 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11421 struct to a subclass of gimple_opt_pass along with...
11422 (pass_data_build_alias): ...new pass_data instance and...
11423 (make_pass_build_alias): ...new function.
11424 (pass_build_ealias): Convert from a global struct to a subclass of
11425 gimple_opt_pass along with...
11426 (pass_data_build_ealias): ...new pass_data instance and...
11427 (make_pass_build_ealias): ...new function.
11428 (pass_ipa_pta): Convert from a global struct to a subclass of
11429 simple_ipa_opt_pass along with...
11430 (pass_data_ipa_pta): ...new pass_data instance and...
11431 (make_pass_ipa_pta): ...new function.
11432 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11433 subclass of gimple_opt_pass along with...
11434 (pass_data_uncprop): ...new pass_data instance and...
11435 (make_pass_uncprop): ...new function.
11436 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11437 global struct to a subclass of gimple_opt_pass along with...
11438 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11439 (make_pass_late_warn_uninitialized): ...new function.
11440 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11441 to a subclass of gimple_opt_pass along with...
11442 (pass_data_init_datastructures): ...new pass_data instance and...
11443 (make_pass_init_datastructures): ...new function.
11444 (pass_early_warn_uninitialized): Convert from a global struct to a
11445 subclass of gimple_opt_pass along with...
11446 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11447 (make_pass_early_warn_uninitialized): ...new function.
11448 (pass_update_address_taken): Convert from a global struct to a
11449 subclass of gimple_opt_pass along with...
11450 (pass_data_update_address_taken): ...new pass_data instance and...
11451 (make_pass_update_address_taken): ...new function.
11452 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11453 struct to a subclass of gimple_opt_pass along with...
11454 (pass_data_release_ssa_names): ...new pass_data instance and...
11455 (make_pass_release_ssa_names): ...new function.
11456 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11457 subclass of gimple_opt_pass along with...
11458 (pass_data_stdarg): ...new pass_data instance and...
11459 (make_pass_stdarg): ...new function.
11460 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11461 global struct to a subclass of gimple_opt_pass along with...
11462 (pass_data_convert_switch): ...new pass_data instance and...
11463 (make_pass_convert_switch): ...new function.
11464 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11465 to a subclass of gimple_opt_pass along with...
11466 (pass_data_tail_recursion): ...new pass_data instance and...
11467 (make_pass_tail_recursion): ...new function.
11468 (pass_tail_calls): Convert from a global struct to a subclass of
11469 gimple_opt_pass along with...
11470 (pass_data_tail_calls): ...new pass_data instance and...
11471 (make_pass_tail_calls): ...new function.
11472 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11473 struct to a subclass of gimple_opt_pass along with...
11474 (pass_data_lower_vector): ...new pass_data instance and...
11475 (make_pass_lower_vector): ...new function.
11476 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11477 gimple_opt_pass along with...
11478 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11479 (make_pass_lower_vector_ssa): ...new function.
11480 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11481 to a subclass of gimple_opt_pass along with...
11482 (pass_data_slp_vectorize): ...new pass_data instance and...
11483 (make_pass_slp_vectorize): ...new function.
11484 (pass_ipa_increase_alignment): Convert from a global struct to a
11485 subclass of simple_ipa_opt_pass along with...
11486 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11487 (make_pass_ipa_increase_alignment): ...new function.
11488 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11489 gimple_opt_pass along with...
11490 (pass_data_vrp): ...new pass_data instance and...
11491 (make_pass_vrp): ...new function.
11492 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11493 subclass of simple_ipa_opt_pass along with...
11494 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11495 (make_pass_ipa_free_lang_data): ...new function.
11496 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11497 gimple_opt_pass along with...
11498 (pass_data_tsan): ...new pass_data instance and...
11499 (make_pass_tsan): ...new function.
11500 (pass_tsan_O0): Convert from a global struct to a subclass of
11501 gimple_opt_pass along with...
11502 (pass_data_tsan_O0): ...new pass_data instance and...
11503 (make_pass_tsan_O0): ...new function.
11504 * var-tracking.c (pass_variable_tracking): Convert from a global
11505 struct to a subclass of rtl_opt_pass along with...
11506 (pass_data_variable_tracking): ...new pass_data instance and...
11507 (make_pass_variable_tracking): ...new function.
11508 * web.c (pass_web): Convert from a global struct to a subclass of
11509 rtl_opt_pass along with...
11510 (pass_data_web): ...new pass_data instance and...
11511 (make_pass_web): ...new function.
11512 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11513 declaration with that of...
11514 (make_pass_mode_switch_use): ...new function.
11515 (pass_resolve_sw_modes): Replace declaration with that of...
11516 (make_pass_resolve_sw_modes): ...new function.
11517 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11518 from a global struct to a subclass of rtl_opt_pass along with...
11519 (pass_data_mode_switch_use): ...new pass_data instance and...
11520 (make_pass_mode_switch_use): ...new function.
11521 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11522 from a global struct to a subclass of rtl_opt_pass along with...
11523 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11524 (make_pass_resolve_sw_modes): ...new function.
11525 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11526 struct to a subclass of rtl_opt_pass along with...
11527 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11528 (make_pass_insert_vzeroupper): ...new function.
11529 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11530 global struct to a subclass of rtl_opt_pass along with...
11531 (pass_data_work_around_errata): ...new pass_data instance and...
11532 (make_pass_work_around_errata): ...new function.
11533 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11534 struct to a subclass of rtl_opt_pass along with...
11535 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11536 (make_pass_mips_machine_reorg2): ...new function.
11538 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11540 * passes.c (pass_manager::operator new): New.
11542 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11544 Handwritten part of conversion of passes to C++ classes.
11546 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11547 (toplev.o): Add dep on PASS_MANAGER_H.
11548 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11549 of early local pases to reflect this moving from a global to a
11550 member of gcc::pass_manager.
11551 (cgraph_add_new_function): Likewise.
11552 * lto-cgraph.c (lto_output_node): Update for conversion of
11553 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11554 * passes.c (opt_pass::clone): New.
11555 (opt_pass::gate): New.
11556 (opt_pass::execute): New.
11557 (opt_pass::opt_pass): New.
11558 (pass_manager::execute_early_local_passes): New.
11559 (pass_manager::execute_pass_mode_switching): new.
11560 (finish_optimization_passes): Convert to...
11561 (pass_manager::finish_optimization_passes): ...this.
11562 (finish_optimization_passes): Update for conversion of passes to
11564 (register_dump_files_1): Use has_gate since we cannot portably
11565 check a vtable entry against NULL.
11566 (dump_one_pass): Likewise.
11567 (ipa_write_summaries_2): Likewise.
11568 (ipa_write_optimization_summaries_1): Likewise.
11569 (ipa_read_summaries_1): Likewise.
11570 (ipa_read_optimization_summaries_1): Likewise.
11571 (execute_ipa_stmt_fixups): Likewise.
11572 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11573 pass-creation functions rather than wiring up globals, and
11574 storing the results in fields of pass_manager generated using
11575 pass-instances.def.
11576 (pass_manager::dump_profile_report): Update for conversion of
11577 passes to C++ classes.
11578 (pass_manager::execute_ipa_summary_passes): Likewise.
11579 (execute_one_ipa_transform_pass): Likewise.
11580 (execute_one_pass): Use has_gate and has_execute since we cannot
11581 portably check a vtable entry against NULL.
11582 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11583 (pass_manager): Use pass-instances.def to add fields for the
11584 various pass instances.
11585 * toplev.c (finalize): Update for move of
11586 finish_optimization_passes to a method of gcc::pass_manager.
11587 * toplev.h (finish_optimization_passes): Move to method of class
11589 * tree-pass.h (struct pass_data): New.
11590 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11591 (opt_pass::gate): Convert to virtual function.
11592 (opt_pass::~opt_pass): New.
11593 (opt_pass::clone): New.
11594 (opt_pass::execute): Convert to virtual function.
11595 (opt_pass::opt_pass): New.
11596 (opt_pass::ctxt_): new.
11597 (gimple_opt_pass): Convert to subclass of opt_pass.
11598 (gimple_opt_pass::gimple_opt_pass): New.
11599 (rtl_opt_pass): Convert to subclass of opt_pass.
11600 (rtl_opt_pass::rtl_opt_pass): New.
11601 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11602 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11603 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11604 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11605 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11606 invocation of pass_mode_switching to reflect this moving from a
11607 global to a member of gcc::pass_manager.
11608 (ix86_option_override): Rework how pass_insert_vzeroupper is
11609 added to the pass_manager to reflect autogenerated changes.
11610 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11612 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11614 PR rtl-optimization/57708
11615 * recog.c (peep2_find_free_register): Validate all regs in a
11618 2013-08-05 Jan Hubicka <jh@suse.cz>
11621 * cgraph.c (verify_cgraph_node): Accept local flags from other
11623 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11624 (function_and_variable_visibility): Likewise.
11625 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11627 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11629 * graph.c (init_graph_slim_pretty_print): Remove.
11630 (print_graph_cfg): Do not call it. Use local pretty printer.
11631 (start_graph_dump): Likewise.
11633 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11635 * gimple-pretty-print.c (buffer): Remove.
11636 (initialized): Likewise.
11637 (maybe_init_pretty_print): Likewise.
11638 (print_gimple_stmt): Do not call it. Use non-static local
11639 pretty_printer variable.
11640 (print_gimple_expr): Likewise.
11641 (print_gimple_seq): Likewise.
11642 (gimple_dump_bb): Likewise.
11644 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11646 * asan.c (asan_pp): Remove.
11647 (asan_pp_initialized): Likewise.
11648 (asan_pp_initialize): Likewise.
11649 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11650 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11651 (asan_add_global): Likewise.
11653 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11655 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11656 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11657 * pretty-print.c (pp_formatted_text_data): Likewise.
11658 (pp_indent): Rename from pp_base_indent.
11659 (pp_format): Rename from pp_base_format.
11660 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11661 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11662 (pp_flush): Rename from pp_base_flush.
11663 (pp_set_line_maximum_length): Rename from
11664 pp_base_set_line_maximum_length.
11665 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11666 (pp_set_prefix): Rename from pp_base_set_prefix.
11667 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11668 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11669 (pp_append_text): Rename from pp_base_append_text.
11670 (pp_formatted_text): Rename from pp_base_formatted_text.
11671 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11672 (pp_remaining_character_count_for_line): Rename from
11673 pp_base_remaining_character_count_for_line.
11674 (pp_newline): Rename from pp_base_newline.
11675 (pp_character): Rename from pp_base_character.
11676 (pp_string): Rename from pp_base_string.
11677 (pp_maybe_space): Rename from pp_base_maybe_space.
11678 * asan.c (asan_pp_string): Adjust.
11679 (asan_emit_stack_protection): Likewise.
11680 (asan_add_global): Likewise.
11681 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11682 * tree-mudflap.c (mf_varname_tree): Likewise.
11683 * tree-pretty-print.c (pp_tree_identifier): Rename from
11684 pp_base_tree_identifier.
11685 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11686 Declare as function.
11688 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11690 * pretty-print.h (pp_bar_bar): New.
11691 (pp_ampersand_ampersand): Likewise.
11692 (pp_less_equal): Likewise.
11693 (pp_greater_equal): Likewise.
11694 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11695 printer functions instead of pp_string or operators and punctuators.
11696 (dump_gimple_call): Likewise.
11697 (dump_gimple_omp_for): Likewise.
11698 (dump_gimple_transaction): Likewise.
11699 (dump_gimple_phi): Likewise.
11700 (pp_gimple_stmt_1): Likewise.
11701 * sched-vis.c (print_insn): Likewise.
11702 * tree-mudflap.c (mf_varname_tree): Likewise.
11703 * tree-pretty-print.c (dump_block_node): Likewise.
11704 (dump_generic_node): Likewise.
11706 2013-08-02 Jan Hubicka <jh@suse.cz>
11708 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11710 * lto-streamer-out.c (tree_is_indexable): Results decls and
11711 parm decls are not indexable.
11712 (DFS_write_tree_body): Do not follow args and results.
11713 (hash_tree): Likewise.
11714 (output_functions): Rearrange so struct function is needed
11715 only when real body is output; be able to also ouptut abstract
11716 functions; output DECL_ARGUMENTS and DECL_RESULT.
11717 (lto_output): When not in WPA, ale store abstract functions.
11718 (write_symbol): Do not care about RESULT_DECL.
11719 (output_symbol_p): Handle correctly sbtract decls.
11720 * lto-streamer-in.c (input_function): Rearrange so struct
11721 function can be NULL at entry; allow streaming of
11722 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11723 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11724 sanity check during LTO.
11725 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11726 RESULT_DECl and DECL_ARGUMENTS.
11727 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11730 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11732 * pretty-print.h (pp_underscore): New.
11734 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11735 printer functions instead of pp_character.
11736 (dump_binary_rhs): Likewise.
11737 (dump_ternary_rhs): Likewise.
11738 (dump_gimple_call_args): Likewise.
11739 (pp_points_to_solution): Likewise.
11740 (dump_gimple_call): Likewise.
11741 (dump_gimple_switch): Likewise.
11742 (dump_gimple_cond): Likewise.
11743 (dump_gimple_bind): Likewise.
11744 (dump_gimple_try): Likewise.
11745 (dump_gimple_omp_for): Likewise.
11746 (dump_gimple_omp_continue): Likewise.
11747 (dump_gimple_omp_single): Likewise.
11748 (dump_gimple_omp_sections): Likewise.
11749 (dump_gimple_omp_block): Likewise.
11750 (dump_gimple_omp_critical): Likewise.
11751 (dump_gimple_transaction): Likewise.
11752 (dump_gimple_asm): Likewise.
11753 (dump_gimple_phi): Likewise.
11754 (dump_gimple_omp_parallel): Likewise.
11755 (dump_gimple_omp_task): Likewise.
11756 (dump_gimple_omp_atomic_load): Likewise.
11757 (dump_gimple_omp_atomic_store): Likewise.
11758 (dump_gimple_mem_ops): Likewise.
11759 (pp_gimple_stmt_1): Likewise.
11760 (pp_cfg_jump): Likewise.
11761 (dump_implicit_edges): Likewise.
11762 (gimple_dump_bb_for_graph): Likewise.
11763 * graph.c (draw_cfg_node): Likewise.
11764 * langhooks.c (lhd_print_error_function): Likewise.
11765 * sched-vis.c (print_exp): Likewise.
11766 (print_value): Likewise.
11767 (print_pattern): Likewise.
11768 (print_insn): Likewise.
11769 (rtl_dump_bb_for_graph): Likewise.
11770 * tree-pretty-print.c (dump_function_declaration): Likewise.
11771 (dump_array_domain): Likewise.
11772 (dump_omp_clause): Likewise.
11773 (dump_location): Likewise.
11774 (dump_generic_node): Likewise.
11775 (print_struct_decl): Likewise.
11776 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11778 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11780 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11781 candidate table when replacing a candidate statement.
11782 (replace_rhs_if_not_dup): Likewise.
11783 (replace_one_candidate): Likewise.
11785 2013-08-02 Jan Hubicka <jh@suse.cz>
11786 Martin Liska <marxin.liska@gmail.com>
11788 * cgraphunit.c (add_new_function): Fix logic when adding from
11790 (assemble_thunk): Rename to ...
11791 (expand_thunk); .. this one; export; get it working with
11792 general functions; make produced gimple valid.
11793 * cgraph.h (expand_thunk): Declare.
11795 2013-08-02 Jan Hubicka <jh@suse.cz>
11797 * ipa-cp.c (gather_context_independent_values): Use
11798 ipa_get_param_move_cost.
11799 (get_replacement_map): Remove PARAM; move parameter folding
11801 (create_specialized_node): Update.
11802 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11803 assert that we have gimple body; update move_cost.
11804 (count_formal_params): Assert that we have gimple body.
11805 (ipa_dump_param): New function.
11806 (ipa_alloc_node_params): Break out from ...
11807 (ipa_initialize_node_params): ... here.
11808 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11809 (ipa_write_node_info): Stream move costs.
11810 (ipa_read_node_info): Read move costs.
11811 (ipa_update_after_lto_read): Do not recompute node params.
11812 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11813 (ipa_get_param): Check we are not in WPA.
11814 (ipa_get_param_move_cost): New.
11815 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11816 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11817 parm numbers to be present.
11819 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11821 PR rtl-optimization/58048
11822 * lra-constraints.c (process_alt_operands): Don't check asm
11823 operand on register.
11825 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11827 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11828 the implied StoreLoad barrier for atomic operations if before.
11830 2013-08-02 Jan Hubicka <jh@suse.cz>
11831 Martin Liska <marxin.liska@gmail.com>
11833 * cgraph.c (cgraph_function_body_availability): Do not check
11835 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11836 symtab_node_availability): Declare.
11837 * ipa.c (can_replace_by_local_alias): New.
11838 (function_and_variable_visibility): Use it.
11839 * symtab.c (symtab_for_node_and_aliases,
11840 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
11842 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11844 PR rtl-optimization/57963
11845 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
11846 (lra_constraints): Use them.
11848 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
11850 * config/arm/types.md (define_attr "type"): Add "load_acq"
11852 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
11854 (cortex_a53_store1): Likewise.
11856 2013-08-01 Jan Hubicka <jh@suse.cz>
11858 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
11859 partitions are not needed.
11861 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
11863 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
11864 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
11865 MAYBE_NON_Q_CLASS_P where appropriate.
11867 2013-08-01 Jan Hubicka <jh@suse.cz>
11869 * cgraph.h (release_function_body): Declare.
11870 * tree.c (free_lang_data_in_decl): Free, parameters and return values
11871 of unused delcarations.
11873 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11875 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
11876 RTL form when subtracting a constant.
11878 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11880 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
11881 Generate canonical plus rtx with negated immediate instead of minus
11883 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
11885 2013-08-01 Jan Hubicka <jh@suse.cz>
11887 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
11888 (cgraph_release_function_body): Likewise.
11889 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
11890 * cgraph.h (cgrpah_node): Rename abstract_and_needed
11891 to used_as_abstract_origin.
11892 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
11893 symbols used as abstract origins.
11894 * cgraphunit.c (analyze_functions): Update.
11895 * ipa.c (symtab_remove_unreachable_nodes): Recompute
11896 used_as_abstract_origin.
11897 * tree-inline.c (tree_function_versioning): Update
11898 used_as_abstract_origin; be ready for DECL_RESULT and
11899 DECL_ARGUMENTS to be NULL.
11901 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
11902 for abstract functions.
11903 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
11906 2013-08-01 Jan Hubicka <jh@suse.cz>
11908 * profile.c (compute_value_histograms): Fix thinko.
11910 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11912 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
11913 aarch-common-protos.h to extra_headers.
11914 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11915 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
11916 * config/aarch64/t-aarch64 (aarch-common.o): Define.
11918 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
11920 * config/aarch64/aarch64.md (define_attr "type"): Delete.
11921 Include "../arm/types.md". Define "type" attribute for all patterns.
11922 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
11925 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
11927 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
11928 to support power8 load fusion.
11929 (fusion_gpr_mem_load): Likewise.
11931 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
11933 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
11934 declarations for power8 load fusion.
11935 (emit_fusion_gpr_load): Likewise.
11937 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
11938 tuning for power8, turn on fusion mode by default. Turn on sign
11939 extending fusion mode if normal fusion mode is on, and we are at
11941 (fusion_gpr_load_p): New function, return true if we can fuse an
11942 addis instruction with a dependent load to a GPR.
11943 (emit_fusion_gpr_load): Emit the instructions for power8 load
11946 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
11947 (VSX load fusion peepholes): New peepholes to fuse together an
11948 addi instruction with a VSX load instruction.
11950 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
11951 peepholes to fuse an addis instruction with a load to a GPR base
11952 register. If we are supporting sign extending fusions, convert
11953 sign extending loads to zero extending loads and add an explicit
11956 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11958 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
11959 aarch-common-protos.h to extra_headers.
11960 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
11961 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
11962 (arm_early_store_addr_dep): Likewise.
11963 (arm_no_early_alu_shift_dep): Likewise.
11964 (arm_no_early_alu_shift_value_dep): Likewise.
11965 (arm_no_early_mul_dep): Likewise.
11966 (arm_no_early_store_addr_dep): Likewise.
11967 (arm_mac_accumulator_is_mul_result): Likewise.
11968 (arm_mac_accumulator_is_result): Likewise.
11969 * config/arm/aarch-common.c: ... here. New file.
11970 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
11972 (arm_early_store_addr_dep): Likewise.
11973 (arm_no_early_alu_shift_dep): Likewise.
11974 (arm_no_early_alu_shift_value_dep): Likewise.
11975 (arm_no_early_mul_dep): Likewise.
11976 (arm_no_early_store_addr_dep): Likewise.
11977 (arm_mac_accumulator_is_mul_result): Likewise.
11978 (arm_mac_accumulator_is_result): Likewise.
11979 * config/arm/aarch-common-protos.h: ... here. New file.
11980 * config/arm/t-arm (aarch-common.o): Define.
11982 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
11984 * config/arm/arm.md: Include new file "types.md".
11985 (define_attr "type"): Move from here to ...
11986 (define_attr "mul32"): Likewise.
11987 (define_attr "mul64"): Likewise.
11988 * config/arm/types.md: ... here. New file.
11990 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
11992 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
11993 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
11995 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
11997 * gen-pass-instances.awk: Fix offset of substr().
11999 2013-07-31 David Malcolm <dmalcolm@redhat.com>
12001 * Makefile.in (pass-instances.def): New.
12002 (passes.o): Replace dependency on passes.def with one on
12005 * gen-pass-instances.awk: New.
12007 * passes.c (pass_manager::pass_manager): Use pass-instances.def
12008 rather than passes.def, updating local definition of NEXT_PASS
12009 macro to add an extra NUM parameter (currently unused).
12011 2013-07-30 David Malcolm <dmalcolm@redhat.com>
12013 * Makefile.in (PASS_MANAGER_H): New.
12014 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
12015 (passes.o): Likewise.
12016 (statistics.o): Likewise.
12017 (cgraphunit.o): Likewise.
12018 (context.o): Depend on PASS_MANAGER_H.
12020 * pass_manager.h: New.
12022 * cgraphunit.c (cgraph_add_new_function): Update for moves
12023 of globals to fields of pass_manager.
12024 (analyze_function): Likewise.
12025 (expand_function): Likewise.
12026 (ipa_passes): Likewise.
12027 (compile): Likewise.
12029 * context.c (context::context): New.
12030 * context.h (context::context): New.
12031 (context::get_passes): New.
12032 (context::passes_): New.
12034 * lto-cgraph.c (input_node): Update for moves of globals to
12035 fields of pass_manager.
12037 * passes.c (all_passes): Remove, in favor of a field of the
12038 same name within the new class pass_manager.
12039 (all_small_ipa_passes): Likewise.
12040 (all_lowering_passes): Likewise.
12041 (all_regular_ipa_passes): Likewise.
12042 (all_late_ipa_passes): Likewise.
12043 (all_lto_gen_passes): Likewise.
12044 (passes_by_id): Likewise.
12045 (passes_by_id_size): Likewise.
12046 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12047 the new class pass_manager.
12048 (set_pass_for_id): Convert to...
12049 (pass_manager::set_pass_for_id): ...method.
12050 (get_pass_for_id): Convert to...
12051 (pass_manager::get_pass_for_id): ...method.
12052 (register_one_dump_file): Move body of implementation into...
12053 (pass_manager::register_one_dump_file): ...here.
12054 (register_dump_files_1): Convert to...
12055 (pass_manager::register_dump_files_1): ...method.
12056 (register_dump_files): Convert to...
12057 (pass_manager::register_dump_files): ...method.
12058 (create_pass_tab): Update for moves of globals to fields of
12060 (dump_passes): Move body of implementation into...
12061 (pass_manager::dump_passes): ...here.
12062 (register_pass): Move body of implementation into...
12063 (pass_manager::register_pass): ...here.
12064 (init_optimization_passes): Convert into...
12065 (pass_manager::pass_manager): ...constructor for new
12066 pass_manager class, and initialize the pass_lists array.
12067 (check_profile_consistency): Update for moves of globals to
12068 fields of pass_manager.
12069 (dump_profile_report): Move body of implementation into...
12070 (pass_manager::dump_profile_report): ...here.
12071 (ipa_write_summaries_1): Update for moves of pass lists from
12072 being globals to fields of pass_manager.
12073 (ipa_write_optimization_summaries): Likewise.
12074 (ipa_read_summaries): Likewise.
12075 (ipa_read_optimization_summaries): Likewise.
12076 (execute_all_ipa_stmt_fixups): Likewise.
12078 * statistics.c (statistics_fini): Update for moves of globals to
12079 fields of pass_manager.
12081 * toplev.c (general_init): Replace call to
12082 init_optimization_passes with construction of the pass_manager
12085 * tree-pass.h (all_passes): Remove, in favor of a field of the
12086 same name within the new class pass_manager.
12087 (all_small_ipa_passes): Likewise.
12088 (all_lowering_passes): Likewise.
12089 (all_regular_ipa_passes): Likewise.
12090 (all_lto_gen_passes): Likewise.
12091 (all_late_ipa_passes): Likewise.
12092 (passes_by_id): Likewise.
12093 (passes_by_id_size): Likewise.
12094 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12095 the new class pass_manager.
12096 (get_pass_for_id): Remove.
12098 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12100 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
12103 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12105 * arm.md (mulhi3): New expand pattern.
12107 2013-07-30 Jan Hubicka <jh@suse.cz>
12108 Martin Liska <marxin.liska@gmail.com>
12110 * profile.c (compute_value_histograms): Do not ICE when
12111 there is mismatch only on some counters.
12113 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12115 PR rtl-optimization/57637
12116 * function.c (move_insn_for_shrink_wrap): Also check the
12117 GEN set of the LIVE problem for the liveness analysis
12118 if it exists, otherwise give up.
12120 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12122 PR tree-optimization/57993
12123 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
12124 replaced statement in the candidate table.
12125 (phi_add_costs): Return infinite cost when the hidden basis does
12126 not dominate all phis on which the candidate is dependent.
12127 (replace_one_candidate): Record replaced statement in the
12130 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
12132 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
12133 (ashlv2si3): New expander.
12134 (*ashlv2si3_i): New define_insn_and_split.
12135 * predicates.md (float_operation): Allow patterns with three
12136 basic sub-patterns.
12138 PR rtl-optimization/58021
12139 * mode-switching.c (create_pre_exit): Always split off preceding
12140 insns if we are not at the basic block head.
12142 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
12144 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
12145 (UCLIBC_DYNAMIC_LINKER): New macro.
12146 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
12148 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
12149 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
12151 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
12152 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
12153 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
12154 for SF and DF modes. Use ieee_quad_format for TF mode.
12155 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
12156 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
12157 (mips_option_override): Handle `-mnan=legacy'.
12158 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
12159 `-mabs=2008' and `-mnan=2008'.
12160 (OPTION_DEFAULT_SPECS): Add "nan" default.
12161 (ASM_SPEC): Handle `-mnan='.
12162 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
12163 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
12164 comment accordingly.
12165 (neg<mode>2): Likewise.
12166 * config/mips/mips.opt (mabs, mnan): New options.
12167 * doc/install.texi (Configuration): Document `--with-nan=' option.
12168 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
12170 (MIPS Options): Document them.
12171 * config.gcc <mips*-*-*>: Handle `--with-nan='.
12172 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
12173 * configure: Regenerate.
12174 * config.in: Regenerate.
12176 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12178 * config/i386/i386.md (float post-reload splitters): Do not check
12179 for subregs of SSE registers.
12181 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12182 H.J. Lu <hongjiu.lu@intel.com>
12186 * config/i386/i386.md (post-reload splitter
12187 to avoid partial SSE reg dependency stalls): New pattern.
12189 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12191 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12193 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12195 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12196 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12198 2013-07-26 Cary Coutant <ccoutant@google.com>
12200 * dwarf2out.c (die_checksum_ordered): Don't include template
12201 instantiations in signature.
12202 (is_template_parameter): New function.
12203 (is_template_instantiation): New function.
12204 (generate_skeleton_bottom_up): Don't include template instantiations
12206 (generate_skeleton): Likewise.
12207 (break_out_comdat_types): Move recursive call to break out nested
12209 (prune_unused_types_mark_generic_parms_dies): Call
12210 is_template_parameter.
12212 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12214 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12215 uses vector registers.
12216 * config/aarch64/iterators.md: Add attributes rtn and vas.
12218 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12219 Richard Earnshaw <richard.earnshaw@arm.com>
12221 * combine.c (simplify_comparison): Re-canonicalize operands
12223 * config/arm/arm.md (movcond_addsi): New splitter.
12225 2013-07-25 Sterling Augustine <saugustine@google.com>
12227 * dwarf2out.c (size_of_pubnames): Move code to...
12228 (include_pubname_in_output): ...here. New.
12229 (want_pubnames): Rearrange.
12230 (output_pubnames): Call include_pubname_in_output. Move assertion.
12232 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12234 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12236 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12239 * doc/extend.texi: Remove obsolete builtins. Fix
12240 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12242 2013-07-25 Jan Hubicka <jh@suse.cz>
12244 * cgraph.c (release_function_body): Break out from ...
12245 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12246 and DECL_ARGUMENTS.
12247 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12248 old_tree in the map.
12249 (create_specialized_node): Update.
12250 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12252 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12253 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12254 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12255 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12258 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12260 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12261 addsi3_carryin_alt2_<optab>): Correct output template.
12263 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12265 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12266 Adjust for arm_restrict_it.
12267 Remove trailing whitespace.
12269 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12271 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12272 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12274 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12276 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12278 PR rtl-optimization/57960
12279 * lra-constraints.c (process_alt_operands): Use the right mode
12280 when checking strict_low.
12282 2013-07-25 Jan Hubicka <jh@suse.cz>
12284 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12285 * cgraph.c (cgraph_remove_node): Do not release function body
12286 when in cgraph streaming.
12287 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12288 in other partitions are not considered reachable; fix handling of
12291 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12293 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12295 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12300 * config/arm/arm.md ("*sibcall_insn): Replace use of
12301 Ss with US. Adjust output for v5 and v4t.
12302 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12304 * config/arm/constraints.md ("Ss"): Rename to US.
12306 2013-07-25 Terry Guo <terry.guo@arm.com>
12308 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12309 shift_add/shift_sub0/shift_sub1 RTXs.
12311 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12312 Anton Blanchard <anton@au1.ibm.com>
12314 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12315 (altivec_vpks<VI_char>ss): Likewise.
12316 (altivec_vpks<VI_char>us): Likewise.
12317 (altivec_vpku<VI_char>us): Likewise.
12318 (altivec_vpku<VI_char>um): Likewise.
12320 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12322 Introduce context class.
12324 * Makefile.in (CONTEXT_H): New.
12325 (OBJS): Add context.o.
12326 (toplev.o): Add CONTEXT_H to dependencies.
12329 * toplev.c (general_init): Create the singleton gcc::context instance.
12335 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12337 PR rtl-optimization/57968
12338 * mode-switching.c (create_pre_exit): Allow instructions that
12339 don't set a return register to need a non-exit mode.
12341 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12342 Anton Blanchard <anton@au1.ibm.com>
12344 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12345 operands to vperm for little endian.
12346 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12347 of lvsl to create the control mask for a vperm for little endian.
12349 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12350 Anton Blanchard <anton@au1.ibm.com>
12352 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12353 two operands for little-endian.
12355 2013-07-23 Steve Ellcey <sellcey@mips.com>
12357 * config/mips/mips.c (mips_case_values_threshold): New.
12358 (TARGET_CASE_VALUES_THRESHOLD): Define.
12360 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12361 Anton Blanchard <anton@au1.ibm.com>
12363 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12364 selection of field for vector splat in little endian mode.
12366 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12368 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12369 expanders to rs6000.md.
12370 (ior<mode>3): Likewise.
12371 (and<mode>3): Likewise.
12372 (one_cmpl<mode>2): Likewise.
12373 (nor<mode>3): Likewise.
12374 (andc<mode>3): Likewise.
12375 (eqv<mode>3): Likewise.
12376 (nand<mode>3): Likewise.
12377 (orc<mode>3): Likewise.
12379 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12382 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12383 to split multi-word logical operations.
12384 (rs6000_split_logical_di): Likewise.
12385 (rs6000_split_logical): Likewise.
12387 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12388 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12389 and allow TImode operations in 32-bit.
12390 (vsx_and<mode>3_64bit): Likewise.
12391 (vsx_ior<mode>3_32bit): Likewise.
12392 (vsx_ior<mode>3_64bit): Likewise.
12393 (vsx_xor<mode>3_32bit): Likewise.
12394 (vsx_xor<mode>3_64bit): Likewise.
12395 (vsx_one_cmpl<mode>2_32bit): Likewise.
12396 (vsx_one_cmpl<mode>2_64bit): Likewise.
12397 (vsx_nor<mode>3_32bit): Likewise.
12398 (vsx_nor<mode>3_64bit): Likewise.
12399 (vsx_andc<mode>3_32bit): Likewise.
12400 (vsx_andc<mode>3_64bit): Likewise.
12401 (vsx_eqv<mode>3_32bit): Likewise.
12402 (vsx_eqv<mode>3_64bit): Likewise.
12403 (vsx_nand<mode>3_32bit): Likewise.
12404 (vsx_nand<mode>3_64bit): Likewise.
12405 (vsx_orc<mode>3_32bit): Likewise.
12406 (vsx_orc<mode>3_64bit): Likewise.
12408 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12409 logical types in GPRs.
12411 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12412 logical insns to rs6000.md, and allow TImode operations in
12414 (altivec_ior<mode>3): Likewise.
12415 (altivec_xor<mode>3): Likewise.
12416 (altivec_one_cmpl<mode>2): Likewise.
12417 (altivec_nor<mode>3): Likewise.
12418 (altivec_andc<mode>3): Likewise.
12420 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12421 attributes for moving the 128-bit logical operations into
12423 (BOOL_REGS_OUTPUT): Likewise.
12424 (BOOL_REGS_OP1): Likewise.
12425 (BOOL_REGS_OP2): Likewise.
12426 (BOOL_REGS_UNARY): Likewise.
12427 (BOOL_REGS_AND_CR0): Likewise.
12428 (one_cmpl<mode>2): Add support for DI logical operations on
12429 32-bit, splitting the operations to 32-bit.
12430 (anddi3): Likewise.
12431 (iordi3): Likewise.
12432 (xordi3): Likewise.
12433 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12434 changes to combine the 32/64-bit code, allow logical operations on
12435 TI mode in 32-bit, and to use similar match_operator patterns like
12436 scalar mode uses. Combine the Altivec and VSX code for logical
12437 operations, and move it here.
12438 (ior<mode>3, 128-bit types): Likewise.
12439 (xor<mode>3, 128-bit types): Likewise.
12440 (one_cmpl<mode>3, 128-bit types): Likewise.
12441 (nor<mode>3, 128-bit types): Likewise.
12442 (andc<mode>3, 128-bit types): Likewise.
12443 (eqv<mode>3, 128-bit types): Likewise.
12444 (nand<mode>3, 128-bit types): Likewise.
12445 (orc<mode>3, 128-bit types): Likewise.
12446 (and<mode>3_internal): Likewise.
12447 (bool<mode>3_internal): Likewise.
12448 (boolc<mode>3_internal1): Likewise.
12449 (boolc<mode>3_internal2): Likewise.
12450 (boolcc<mode>3_internal1): Likewise.
12451 (boolcc<mode>3_internal2): Likewise.
12452 (eqv<mode>3_internal1): Likewise.
12453 (eqv<mode>3_internal2): Likewise.
12454 (one_cmpl1<mode>3_internal): Likewise.
12456 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12458 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12459 Rename flag_stack_usage to flag_stack_usage_info.
12461 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12463 * config/microblaze/sync.md: New file.
12464 * config/microblaze/microblaze.md: Include sync.md
12465 * config/microblaze/microblaze.c: Add print_operand 'y'.
12466 * config/microblaze/constraints.md: Add memory_contraint
12467 'Q' which is a single register.
12469 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12471 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12473 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12475 * reload.c (find_reloads): Exit loop once we find this operand
12476 cannot be reloaded somehow for this alternative.
12478 * reload.c (find_reloads): Exit loop once we find a hard register.
12480 * rtlanal.c (computed_jump_p): Exit loop once we find label
12483 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12485 * cfgloopmanip.c (remove_path): Exit loop after setting
12488 * gensupport.c (subst_dup): Avoid loop if code is not
12489 MATCH_DUP nor MATCH_OP_DUP.
12491 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12493 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12495 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12497 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12498 true for SP_REGNUM if mode == ptr_mode.
12499 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12500 with value R0_REGNUM + 31.
12502 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12504 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12505 pad pointer-typed argument downward.
12507 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12509 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12510 and __ILP32__ when the ILP32 model is in use.
12512 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12514 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12515 (aarch64_load_symref_appropriately): In the case of
12516 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12517 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12518 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12519 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12520 if 'mode' doesn't equal to 'ptr_mode'.
12521 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12522 'vcall_offset'; change to call aarch64_emit_move differently depending
12523 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12524 to calculate the upper bound of 'vcall_offset'.
12525 (aarch64_cannot_force_const_mem): Change to also return true if
12527 (aarch64_legitimize_reload_address): In the case of large
12528 displacements, add new local variable 'xmode' and an assertion
12529 based on it; change to use 'xmode' to generate the new rtx and
12531 (aarch64_asm_trampoline_template): Change to generate the template
12532 differently depending on TARGET_ILP32 or not; change to use
12533 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12534 (aarch64_trampoline_size): Removed.
12535 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12536 and replace immediate literals with it. Change to use 'ptr_mode'
12537 instead of 'DImode' and call convert_memory_address if the mode
12538 of 'fnaddr' doesn't equal to 'ptr_mode'.
12539 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12541 (aarch64_elf_asm_destructor): Likewise.
12542 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12543 on TARGET_ILP32 instead of aarch64_trampoline_size.
12544 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12545 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12546 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12547 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12548 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12549 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12550 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12551 depending on the value of 'mode'.
12552 (add_losym_<mode>): New.
12553 (ldr_got_small_<mode>): New, based on ldr_got_small.
12554 (ldr_got_small): Remove.
12555 (ldr_got_small_sidi): New.
12556 * config/aarch64/iterators.md (P): New.
12557 (PTR): Change to 'ptr_mode' in the condition.
12559 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12561 * config.gcc (aarch64*-*-*): Support --with-abi.
12562 (aarch64*-*-elf): Support --with-multilib-list.
12563 (aarch64*-*-linux*): Likewise.
12564 (supported_defaults): Add abi to aarch64*-*-*.
12565 * configure.ac: Mention AArch64 for --with-multilib-list.
12566 * configure: Re-generated.
12567 * config/aarch64/biarchilp32.h: New file.
12568 * config/aarch64/biarchlp64.h: New file.
12569 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12571 (MULTILIB_DEFAULTS): Ditto.
12572 (DRIVER_SELF_SPECS): Ditto.
12573 (ASM_SPEC): Update to also substitute -mabi.
12574 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12575 file whose name depends on -mabi= and -mbig-endian.
12576 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12578 (POINTER_SIZE): New define.
12579 (POINTERS_EXTEND_UNSIGNED): Ditto.
12580 (enum aarch64_abi_type): New enumeration tag.
12581 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12582 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12583 (TARGET_ILP32): New define.
12584 * config/aarch64/aarch64.opt (mabi): New.
12585 (aarch64_abi): New.
12586 (ilp32, lp64): New values for -mabi.
12587 * config/aarch64/t-aarch64 (comma): New define.
12588 (MULTILIB_OPTIONS): Ditto.
12589 (MULTILIB_DIRNAMES): Ditto.
12590 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12591 * doc/invoke.texi: Document -mabi for AArch64.
12593 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12595 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12597 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12598 Anton Blanchard <anton@au1.ibm.com>
12600 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12601 endianness when selecting field to splat.
12603 2013-07-22 Eric Christopher <echristo@gmail.com>
12605 * dwarf2out.c (die_odr_checksum): New function to use
12606 CHECKSUM_ macros and ULEB128 for DIE tag.
12607 (generate_type_signature): Use.
12609 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12611 * config.gcc (sparc*-*-*): Accept leon3 processor.
12612 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12613 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12614 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12615 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12616 (mfix-ut699): Adjust comment.
12617 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12618 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12619 (CPP_CPU_SPEC): Likewise.
12620 (ASM_CPU_SPEC): Likewise.
12621 * config/sparc/sparc.c (leon3_cost): New constant.
12622 (sparc_option_override): Add leon3 support.
12623 (mem_ref): New function.
12624 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12625 (sparc_do_work_around_errata): Look into the instruction in the delay
12626 slot and adjust accordingly. Add fix for the data cache nullify issues
12627 of the UT699. Change insertion position for the NOP.
12628 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12629 (leon3_load): New reservation.
12630 (leon_store): Bump latency to 2.
12631 (grfpu): New automaton.
12632 (grfpu_alu): New unit.
12633 (grfpu_ds): Likewise.
12634 (leon_fp_alu): Adjust.
12635 (leon_fp_mult): Delete.
12636 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12637 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12638 * config/sparc/sparc.md (cpu): Add leon3.
12639 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12640 (swapsi): Likewise.
12641 (atomic_test_and_set): Likewise.
12642 (ldstub): Likewise.
12644 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12646 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12647 default for R5900 targets.
12648 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12649 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12650 * config/mips/mips.c (mips_option_override): Report an error for
12651 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12652 for -march=r5900 -mhard-float.
12654 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12656 * df-problems.c (can_move_insns_across): Exit loop once we
12657 find a non-fixed, non-global register.
12659 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12662 * omega.c (omega_eliminate_red): Break after setting red_found.
12663 (omega_problem_has_red_equations): Similarly after setting found.
12664 (omega_query_variable): Similarly after setting coupled.
12666 2013-07-22 Marek Polacek <polacek@redhat.com>
12668 * gimplify.c: Don't include gimple.h twice.
12670 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12672 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12673 instead of TARGET_THUMB1.
12674 (Pz): New constraint.
12675 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12677 (compare_negsi_si): Likewise.
12678 (compare_addsi2_op0): Likewise.
12679 (compare_addsi2_op1): Likewise.
12680 (addsi3_carryin_<optab>): Likewise.
12681 (addsi3_carryin_alt2_<optab>): Likewise.
12682 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12683 for arm_restrict_it.
12684 (subsi3_carryin): Likewise.
12685 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12686 (minmax_arithsi): Disable for arm_restrict_it.
12687 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12688 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12689 (satsi_<SAT:code>_shift): Likewise.
12690 (arm_shiftsi3): Add alternative for 16-bit encoding.
12691 (arm32_movhf): Disable for arm_restrict_it.
12692 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12693 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12695 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12697 * config/arm/arm.md (attribute "insn"): Delete.
12698 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12699 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12700 (not_shiftsi): Update for attribute change.
12701 (not_shiftsi_compare0): Likewise.
12702 (not_shiftsi_compare0_scratch): Likewise.
12703 (arm_one_cmplsi2): Likewise.
12704 (thumb1_one_cmplsi2): Likewise.
12705 (notsi_compare0): Likewise.
12706 (notsi_compare0_scratch): Likewise.
12707 (thumb1_movdi_insn): Likewise.
12708 (arm_movsi_insn): Likewise.
12709 (movhi_insn_arch4): Likewise.
12710 (movhi_bytes): Likewise.
12711 (arm_movqi_insn): Likewise.
12712 (thumb1_movqi_insn): Likewise.
12713 (arm32_movhf): Likewise.
12714 (thumb1_movhf): Likewise.
12715 (arm_movsf_soft_insn): Likewise.
12716 (thumb1_movsf_insn): Likewise.
12717 (thumb_movdf_insn): Likewise.
12718 (movsicc_insn): Likewise.
12719 (movsfcc_soft_insn): Likewise.
12720 (and_scc): Likewise.
12721 (cond_move): Likewise.
12722 (if_move_not): Likewise.
12723 (if_not_move): Likewise.
12724 (if_shift_move): Likewise.
12725 (if_move_shift): Likewise.
12726 (if_shift_shift): Likewise.
12727 (if_not_arith): Likewise.
12728 (if_arith_not): Likewise.
12729 (cond_move_not): Likewise.
12730 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12731 (neon_mov<mode>): Likewise.
12732 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12733 (thumb2_movsi_vfp): Likewise.
12734 (movsf_vfp): Likewise.
12735 (thumb2_movsf_vfp): Likewise.
12736 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12738 (cortexa7_older_only): Likewise.
12739 (cortexa7_younger): Likewise.
12740 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12741 (1020alu_shift_op): Likewise.
12742 (1020alu_shift_reg_op): Likewise.
12743 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12744 (alu_shift_op): Likewise.
12745 (alu_shift_reg_op): Likewise.
12746 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12747 (11_alu_shift_op): Likewise.
12748 (11_alu_shift_reg_op): Likewise.
12749 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12750 (9_alu_shift_reg_op): Likewise.
12751 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12753 (cortex_a15_alu_shift): Likewise.
12754 (cortex_a15_alu_shift_reg): Likewise.
12755 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12756 (cortex_a5_alu_shift): Likewise.
12757 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12759 (cortex_a53_alu_shift): Likewise.
12760 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12762 (cortex_a7_alu_reg): Likewise.
12763 (cortex_a7_alu_shift): Likewise.
12764 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12765 (cortex_a8_alu_shift): Likewise.
12766 (cortex_a8_alu_shift_reg): Likewise.
12767 (cortex_a8_mov): Likewise.
12768 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12769 (cortex_a9_dp_shift): Likewise.
12770 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12771 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12772 (cortex_r4_mov): Likewise.
12773 (cortex_r4_alu_shift): Likewise.
12774 (cortex_r4_alu_shift_reg): Likewise.
12775 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12776 (526_alu_shift_op): Likewise.
12777 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12778 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12779 (626te_alu_shift_op): Likewise.
12780 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12781 (726te_alu_op): Likewise.
12782 (726te_alu_shift_op): Likewise.
12783 (726te_alu_shift_reg_op): Likewise.
12784 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12785 (mp626_alu_shift_op): Likewise.
12786 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12787 (pj4_alu_e1_conds): Likewise.
12788 (pj4_alu): Likewise.
12789 (pj4_alu_conds): Likewise.
12790 (pj4_shift): Likewise.
12791 (pj4_shift_conds): Likewise.
12792 (pj4_alu_shift): Likewise.
12793 (pj4_alu_shift_conds): Likewise.
12795 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12797 * config/arm/predicates.md (shiftable_operator_strict_it):
12799 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12800 Disable cond_exec version for arm_restrict_it.
12801 (thumb2_smaxsi3): Convert to generate cond_exec.
12802 (thumb2_sminsi3): Likewise.
12803 (thumb32_umaxsi3): Likewise.
12804 (thumb2_uminsi3): Likewise.
12805 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12806 (thumb2_neg_abssi2): Likewise.
12807 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12808 (thumb2_movsicc_insn): Adjust alternatives.
12809 (thumb2_mov_negscc): Disable for arm_restrict_it.
12810 (thumb2_mov_negscc_strict_it): New pattern.
12811 (thumb2_mov_notscc_strict_it): New pattern.
12812 (thumb2_mov_notscc): Disable for arm_restrict_it.
12813 (thumb2_ior_scc): Likewise.
12814 (thumb2_ior_scc_strict_it): New pattern.
12815 (thumb2_cond_move): Adjust for arm_restrict_it.
12816 (thumb2_cond_arith): Disable for arm_restrict_it.
12817 (thumb2_cond_arith_strict_it): New pattern.
12818 (thumb2_cond_sub): Adjust for arm_restrict_it.
12819 (thumb2_movcond): Likewise.
12820 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12821 (thumb2_zero_extendhisi2_v6): Likewise.
12822 (thumb2_zero_extendqisi2_v6): Likewise.
12823 (orsi_notsi_si): Likewise.
12824 (orsi_not_shiftsi_si): Likewise.
12826 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12828 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12829 instruction sequence is 1 byte shorter.
12831 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12833 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12834 it is not needed after split.
12836 2013-07-20 Iain Sandoe <iain@codesourcery.com>
12839 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
12840 second label for nonlocal goto receivers. Don't output pic base labels
12841 unless we're producing PIC; mark that action unreachable().
12842 (ix86_save_reg): If the function contains a nonlocal label, save the
12844 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
12845 * config/darwin.c (emitted_pic_label_num): New GTY.
12846 (update_pic_label_number_if_needed): New.
12847 (machopic_output_function_base_name): Adjust for nonlocal receiver
12849 (machopic_should_output_picbase_label): New.
12850 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
12851 (nonlocal_goto_receiver): New insn and split.
12853 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
12855 * config/aarch64/aarch64-builtins.c
12856 (aarch64_fold_builtin): Fold abs in all modes.
12857 * config/aarch64/aarch64-simd-builtins.def
12858 (abs): Enable for all modes.
12859 * config/aarch64/arm_neon.h
12860 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
12861 (vabs_f64): Add missing intrinsic.
12863 2013-07-19 Ian Bolton <ian.bolton@arm.com>
12865 * config/aarch64/arm_neon.h (vabs_s64): New function
12867 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
12870 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
12871 * config/avr/avr.md (adjust_len): Add `round'.
12872 * config/avr/avr-protos.h (avr_out_round): New prototype.
12873 (avr_out_plus): Add `out_label' argument.
12874 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
12875 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
12876 Handle the case where `insn' is just a pattern.
12877 (avr_out_bitop): Handle the case where `insn' is just a pattern.
12878 (avr_out_round): New function.
12879 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
12881 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
12883 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12884 Add check for flag_stack_usage to handle -fstack-usage support
12886 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
12888 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
12889 interaction for new Power8 flags and VSX.
12891 2013-07-18 Sriraman Tallam <tmsriram@google.com>
12893 PR middle-end/57698
12894 * tree-inline.c (expand_call_inline): Emit errors during
12895 early_inlining only if optimization is not turned on.
12897 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12901 * passes.c (init_optimization_passes): Move the construction of
12902 the pass hierarchy into a new passes.def file.
12904 * Makefile.in (passes.o): Add dependency on passes.def.
12906 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12908 * passes.c (init_optimization_passes): Introduce macros for
12909 constructing the tree of passes (INSERT_PASSES_AFTER,
12910 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
12911 TERMINATE_PASS_LIST).
12913 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
12914 Wei Mi <wmi@google.com>
12916 PR rtl-optimization/57878
12917 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
12919 (reload_pseudo_compare_func): Check nregs first for reload
12922 2013-07-18 David Malcolm <dmalcolm@redhat.com>
12924 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
12926 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
12928 * read-rtl.c (validate_const_int): Once an invalid character is
12929 seen, quit the loop.
12931 * gengtype.c (write_roots): Similarly once we find the "deletable"
12932 or "if_marked" option.
12934 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12936 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
12937 "xtab" and "sat". Move value "clz" from here to ...
12938 (attriubte "type"): ... here.
12939 (satsi_<SAT:code>): Delete "insn" attribute.
12940 (satsi_<SAT:code>_shift): Likewise.
12941 (arm_zero_extendqisi2addsi): Likewise.
12942 (arm_extendqisi2addsi): Likewise.
12943 (clzsi2): Update for attribute changes.
12944 (rbitsi2): Likewise.
12945 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
12947 (arm_usatsihi): Likewise.
12948 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12950 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
12952 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
12953 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
12954 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
12955 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
12956 in alphabetical order.
12957 (attribute "core_cycles"): Update for attribute changes.
12958 (arm_addsi3): Likewise.
12959 (addsi3_compare0): Likewise.
12960 (addsi3_compare0_scratch): Likewise.
12961 (addsi3_compare_op1): Likewise.
12962 (addsi3_compare_op2): Likewise.
12963 (compare_addsi2_op0): Likewise.
12964 (compare_addsi2_op1): Likewise.
12965 (addsi3_carryin_shift_<optab>): Likewise.
12966 (subsi3_carryin_shift): Likewise.
12967 (rsbsi3_carryin_shift): Likewise.
12968 (arm_subsi3_insn): Likewise.
12969 (subsi3_compare0): Likewise.
12970 (subsi3_compare): Likewise.
12971 (arm_andsi3_insn): Likewise.
12972 (thumb1_andsi3_insn): Likewise.
12973 (andsi3_compare0): Likewise.
12974 (andsi3_compare0_scratch): Likewise.
12975 (zeroextractsi_compare0_scratch
12976 (andsi_not_shiftsi_si): Likewise.
12977 (iorsi3_insn): Likewise.
12978 (iorsi3_compare0): Likewise.
12979 (iorsi3_compare0_scratch): Likewise.
12980 (arm_xorsi3): Likewise.
12981 (thumb1_xorsi3_insn): Likewise.
12982 (xorsi3_compare0): Likewise.
12983 (xorsi3_compare0_scratch): Likewise.
12984 (satsi_<SAT:code>_shift): Likewise.
12986 (arm_shiftsi3): Likewise.
12987 (shiftsi3_compare0): Likewise.
12988 (not_shiftsi): Likewise.
12989 (not_shiftsi_compare0): Likewise.
12990 (not_shiftsi_compare0_scratch): Likewise.
12991 (arm_one_cmplsi2): Likewise.
12992 (thumb_one_complsi2): Likewise.
12993 (notsi_compare0): Likewise.
12994 (notsi_compare0_scratch): Likewise.
12995 (thumb1_zero_extendhisi2): Likewise.
12996 (arm_zero_extendhisi2): Likewise.
12997 (arm_zero_extendhisi2_v6): Likewise.
12998 (arm_zero_extendhisi2addsi): Likewise.
12999 (thumb1_zero_extendqisi2): Likewise.
13000 (thumb1_zero_extendqisi2_v6): Likewise.
13001 (arm_zero_extendqisi2): Likewise.
13002 (arm_zero_extendqisi2_v6): Likewise.
13003 (arm_zero_extendqisi2addsi): Likewise.
13004 (thumb1_extendhisi2): Likewise.
13005 (arm_extendhisi2): Likewise.
13006 (arm_extendhisi2_v6): Likewise.
13007 (arm_extendqisi): Likewise.
13008 (arm_extendqisi_v6): Likewise.
13009 (arm_extendqisi2addsi): Likewise.
13010 (thumb1_extendqisi2): Likewise.
13011 (thumb1_movdi_insn): Likewise.
13012 (arm_movsi_insn): Likewise.
13013 (movsi_compare0): Likewise.
13014 (movhi_insn_arch4): Likewise.
13015 (movhi_bytes): Likewise.
13016 (arm_movqi_insn): Likewise.
13017 (thumb1_movqi_insn): Likewise.
13018 (arm32_movhf): Likewise.
13019 (thumb1_movhf): Likewise.
13020 (arm_movsf_soft_insn): Likewise.
13021 (thumb1_movsf_insn): Likewise.
13022 (movdf_soft_insn): Likewise.
13023 (thumb_movdf_insn): Likewise.
13024 (arm_cmpsi_insn): Likewise.
13025 (cmpsi_shiftsi): Likewise.
13026 (cmpsi_shiftsi_swp): Likewise.
13027 (arm_cmpsi_negshiftsi_si): Likewise.
13028 (movsicc_insn): Likewise.
13029 (movsfcc_soft_insn): Likewise.
13030 (arith_shiftsi): Likewise.
13031 (arith_shiftsi_compare0
13032 (arith_shiftsi_compare0_scratch
13033 (sub_shiftsi): Likewise.
13034 (sub_shiftsi_compare0
13035 (sub_shiftsi_compare0_scratch
13036 (and_scc): Likewise.
13037 (cond_move): Likewise.
13038 (if_plus_move): Likewise.
13039 (if_move_plus): Likewise.
13040 (if_move_not): Likewise.
13041 (if_not_move): Likewise.
13042 (if_shift_move): Likewise.
13043 (if_move_shift): Likewise.
13044 (if_shift_shift): Likewise.
13045 (if_not_arith): Likewise.
13046 (if_arith_not): Likewise.
13047 (cond_move_not): Likewise.
13048 (thumb1_ashlsi3): Set type attribute.
13049 (thumb1_ashrsi3): Likewise.
13050 (thumb1_lshrsi3): Likewise.
13051 (thumb1_rotrsi3): Likewise.
13052 (shiftsi3_compare0_scratch): Likewise.
13053 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
13054 (neon_mov<mode>): Likewise.
13055 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
13057 (thumb2_movsi_insn): Likewise.
13058 (thumb2_cmpsi_neg_shiftsi): Likewise.
13059 (thumb2_extendqisi_v6): Likewise.
13060 (thumb2_zero_extendhisi2_v6): Likewise.
13061 (thumb2_zero_extendqisi2_v6): Likewise.
13062 (thumb2_shiftsi3_short): Likewise.
13063 (thumb2_addsi3_compare0_scratch): Likewise.
13064 (orsi_not_shiftsi_si): Likewise.
13065 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
13066 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
13068 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
13069 (1020alu_shift_op): Likewise.
13070 (1020alu_shift_reg_op): Likewise.
13071 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13072 (alu_shift_op): Likewise.
13073 (alu_shift_reg_op): Likewise.
13074 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13075 (11_alu_shift_op): Likewise.
13076 (11_alu_shift_reg_op): Likewise.
13077 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13078 (9_alu_shift_reg_op): Likewise.
13079 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13081 (cortex_a15_alu_shift): Likewise.
13082 (cortex_a15_alu_shift_reg): Likewise.
13083 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
13085 (cortex_a5_alu_shift): Likewise.
13086 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
13088 (cortex_a53_alu_shift): Likewise.
13089 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13091 (cortex_a7_alu_reg): Likewise.
13092 (cortex_a7_alu_shift): Likewise.
13093 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
13095 (cortex_a8_alu_shift): Likewise.
13096 (cortex_a8_alu_shift_reg): Likewise.
13097 (cortex_a8_mov): Likewise.
13098 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13099 (cortex_a9_dp_shift): Likewise.
13100 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
13102 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
13104 (cortex_r4_mov): Likewise.
13105 (cortex_r4_alu_shift): Likewise.
13106 (cortex_r4_alu_shift_reg): Likewise.
13107 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13108 (526_alu_shift_op): Likewise.
13109 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13110 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13111 (626te_alu_shift_op): Likewise.
13112 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
13113 (726te_alu_op): Likewise.
13114 (726te_alu_shift_op): Likewise.
13115 (726te_alu_shift_reg_op): Likewise.
13116 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13117 (mp626_alu_shift_op): Likewise.
13118 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
13119 (pj4_alu_e1_conds): Likewise.
13120 (pj4_alu): Likewise.
13121 (pj4_alu_conds): Likewise.
13122 (pj4_shift): Likewise.
13123 (pj4_shift_conds): Likewise.
13124 (pj4_alu_shift): Likewise.
13125 (pj4_alu_shift_conds): Likewise.
13126 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13128 (cortexa7_older_only): Likewise.
13129 (cortexa7_younger): Likewise.
13131 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13133 * ipa-pure-const.c (generate_summary): Rename to...
13134 (pure_const_generate_summary): ... this.
13136 2013-07-17 Iain Sandoe <iain@codesourcery.com>
13138 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
13140 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
13143 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
13146 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13148 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
13149 enabled without -march=zEC12.
13150 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
13153 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13155 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
13156 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
13157 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
13158 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
13159 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
13160 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
13161 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
13163 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
13164 and ISA_HAS_NMADD3_NMSUB3.
13165 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
13166 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
13167 (nmsub4<mode>, nmsub3<mode>): Likewise.
13168 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
13170 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13172 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
13173 TARGET_MIPS5400 checking.
13175 2013-07-16 Jakub Jelinek <jakub@redhat.com>
13176 Peter Bergner <bergner@vnet.ibm.com>
13178 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13179 registers in the comment.
13180 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13181 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13182 rather than FIRST_PSEUDO_REGISTERS.
13184 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13186 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13187 enable extra ISA flags with TARGET_HTM.
13189 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13191 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13194 2013-07-15 Cong Hou <congh@google.com>
13196 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13197 in compare function for sorting.
13199 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13201 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13202 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13203 * config/rs6000/rs6000.opt: Add -mhtm option.
13204 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13205 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13206 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13207 __HTM__ if the HTM instructions are available.
13208 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13209 htm_spr_reg_operand): New define_predicates.
13210 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13211 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13213 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13214 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13215 HTM builtin functions.
13216 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13217 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13218 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13219 (rs6000_builtin_mask_calculate): Likewise.
13220 (rs6000_option_override_internal): Likewise.
13221 (bdesc_htm): Add new HTM builtin support.
13222 (htm_spr_num): New function.
13223 (htm_spr_regno): Likewise.
13224 (rs6000_htm_spr_icode): Likewise.
13225 (htm_expand_builtin): Likewise.
13226 (htm_init_builtins): Likewise.
13227 (rs6000_expand_builtin): Add support for HTM builtin functions.
13228 (rs6000_init_builtins): Likewise.
13229 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13231 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13232 (TARGET_HTM, MASK_HTM): Define macros.
13233 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13234 (FIXED_REGISTERS): Likewise.
13235 (CALL_USED_REGISTERS): Likewise.
13236 (CALL_REALLY_USED_REGISTERS): Likewise.
13237 (REG_ALLOC_ORDER): Likewise.
13238 (enum reg_class): Likewise.
13239 (REG_CLASS_NAMES): Likewise.
13240 (REG_CLASS_CONTENTS): Likewise.
13241 (REGISTER_NAMES): Likewise.
13242 (ADDITIONAL_REGISTER_NAMES): Likewise.
13243 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13244 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13245 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13246 * config/rs6000/htm.md: New file.
13247 * config/rs6000/htmintrin.h: New file.
13248 * config/rs6000/htmxlintrin.h: New file.
13250 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13252 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13253 Define SYMBOL_TINY_GOT, update comment.
13254 * config/aarch64/aarch64.c
13255 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13256 (aarch64_expand_mov_immediate): Likewise.
13257 (aarch64_print_operand): Likewise.
13258 (aarch64_classify_symbol): Likewise.
13259 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13260 (ldr_got_tiny): Define.
13262 2013-07-13 Tobias Grosser <tobias@grosser.es>
13264 PR tree-optimization/54094
13265 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13266 scheduling dimension for the parallelism check from the polyhedral
13267 information in the AST.
13268 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13271 2013-07-13 Jason Merrill <jason@redhat.com>
13273 * print-tree.c (debug_vec_tree): Use debug_raw.
13274 (debug_raw (vec<tree, va_gc> &)): New.
13275 (debug_raw (vec<tree, va_gc> *)): New.
13276 * tree.h: Declare them.
13278 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13280 * ifcvt.c (ifcvt_after_combine): New static variable.
13281 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13283 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13284 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13285 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13287 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13289 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13291 2013-07-12 Michael Matz <matz@suse.de>
13293 PR middle-end/55771
13294 * convert.c (convert_to_real): Reject non-float inner types.
13296 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13298 * config/aarch64/aarch64-protos.h
13299 (aarch64_simd_immediate_valid_for_move): Remove.
13300 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13301 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13303 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13305 2013-07-11 Steve Ellcey <sellcey@mips.com>
13307 * config/mips/mips.c (mips_conditional_register_usage): Do not
13308 use t[0-7] registers in MIPS16 mode when optimizing for size.
13310 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13312 * config/i386/i386.c (dispatch_function_versions): Fix array
13313 indexing of function_version_info to match actual_versions.
13315 2013-07-11 Teresa Johnson <tejohnson@google.com>
13317 * vec.h (struct va_gc): Move release out-of-line.
13318 (va_gc::release): Call ggc_free on released vec.
13320 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13322 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13323 Require GOT register as additional operand in UNSPEC.
13324 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13325 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13326 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13327 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13328 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13329 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13330 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13332 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13335 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13336 name seen by assembler/linker rather if available.
13338 2013-07-11 Andreas Schwab <schwab@suse.de>
13340 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13342 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13344 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13346 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13348 * doc/tm.texi.in: Move hook documentation to ....
13349 * target.def: ... here.
13351 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13352 text on @hook line.
13353 * doc/tm.texi: Regenerate.
13355 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13358 * doc/invoke.texi: Document -Wconditionally-supported.
13360 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13363 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13366 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13369 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13370 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13371 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13372 Remove duplicate devices.
13373 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13374 * config/avr/t-multilib: Regenerate.
13375 * config/avr/avr-tables.opt: Regenerate.
13376 * doc/avr-mmcu.texi: Regenerate.
13378 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13381 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13383 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13385 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13386 the cost of MULT when optimizing for size.
13388 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13390 * config/cr16/cr16-protos.h: Don't include target.h.
13392 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13394 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13395 adjust register size for TDmode and TFmode for VSX registers.
13397 2013-07-08 Kai Tietz <ktietz@redhat.com>
13400 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13401 hook_bool_const_tree_true.
13403 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13405 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13406 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13407 * config/s390/s390.md: Define FPR*_REGNUM constants.
13408 Fix FPR2_REGNUM constant (18 -> 17).
13409 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13410 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13411 ("trunc<BFP:mode><DFP_ALL:mode>2")
13412 ("trunc<DFP_ALL:mode><BFP:mode>2")
13413 ("*extend<BFP:mode><DFP_ALL:mode>2")
13414 ("*extend<DFP_ALL:mode><BFP:mode>2")
13415 ("extend<BFP:mode><DFP_ALL:mode>2")
13416 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13419 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13421 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13423 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13425 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13426 and cfun_fpr_bit_p to cfun_fpr_save_p.
13427 (s390_frame_area, s390_register_info, s390_frame_info)
13428 (s390_emit_prologue, s390_emit_epilogue)
13429 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13431 * config/s390/s390.h: Define *_REGNUM macros for floating point
13434 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13436 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13438 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13440 PR rtl-optimization/57786
13441 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13442 and break out of the loop when it is set to false.
13444 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13447 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13448 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13449 (const_int 63)) 0)).
13450 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13451 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13452 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13454 PR rtl-optimization/57829
13455 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13456 mask bits outside of mode are just sign-extension from mode to HWI.
13458 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13460 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13461 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13462 adjust_address instead of change_address to keep info about alignment.
13463 (emit_strmov): Remove.
13464 (emit_memmov): New function.
13465 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13466 (expand_movmem_epilogue): Likewise and return updated rtx for
13468 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13469 destination and source.
13470 (decide_alignment): Refactor, handle vector_loop.
13471 (ix86_expand_movmem): Likewise.
13472 (ix86_expand_setmem): Likewise.
13473 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13475 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13477 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13478 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13480 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13482 * config/i386/sse.md (sse_movlhps): Change alternative 3
13483 of operand 2 to "m".
13485 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13488 * config/i386/sse.md (iptr): New mode attribute.
13489 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13490 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13491 (<sse>_vmmul<mode>3): Ditto.
13492 (<sse>_vmdiv<mode>3): Ditto.
13493 (sse_vmrcpv4sf2): Ditto.
13494 (<sse>_vmsqrt<mode>2): Ditto.
13495 (sse_vmrsqrtv4sf2): Ditto.
13496 (<sse>_vm<code><mode>3): Ditto.
13497 (avx_vmcmp<mode>3): Ditto.
13498 (<sse>_vmmaskcmp<mode>3): Ditto.
13499 (<sse>_comi): Ditto.
13500 (<sse>_ucomi): Ditto.
13501 (*xop_vmfrcz_<mode>): Ditto.
13502 (*fmai_fmadd_<mode>): Ditto.
13503 (*fmai_fmsub_<mode>): Ditto.
13504 (*fmai_fnmadd_<mode>): Ditto.
13505 (*fmai_fnmsub_<mode>): Ditto.
13506 (*fma4i_vmfmadd_<mode>): Ditto.
13507 (*fma4i_vmfmsub_<mode>): Ditto.
13508 (*fma4i_vmfnmadd_<mode>): Ditto.
13509 (*fma4i_vmfnmsub_<mode>): Ditto.
13510 (*xop_vmfrcz_<mode>): Ditto.
13511 (sse_cvtps2pi): Ditto.
13512 (sse_cvttps2pi): Ditto.
13513 (sse_cvtss2si): Ditto.
13514 (sse_cvtss2si_2): Ditto.
13515 (sse_cvtss2siq_2): Ditto.
13516 (sse_cvttss2si): Ditto.
13517 (sse_cvttss2siq): Ditto.
13518 (sse_cvtsd2si): Ditto.
13519 (sse_cvtsd2si_2): Ditto.
13520 (sse_cvtsd2siq_2): Ditto.
13521 (sse_cvttsd2si): Ditto.
13522 (sse_cvttsd2siq): Ditto.
13523 (sse_cvtsd2ss): Ditto.
13524 (sse_cvtss2sd): Ditto.
13525 (avx2_pbroadcast<mode>): Ditto.
13526 (avx2_pbroadcast<mode>_1): Ditto.
13527 (*avx_vperm_broadcast_v4sf): Ditto.
13529 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13530 (sse_movlhps): Ditto.
13531 (sse_storehps): Ditto.
13532 (sse_loadhps): Ditto.
13533 (sse_storelps): Ditto.
13534 (sse_loadlps): Ditto.
13535 (*vec_concatv4sf): Ditto.
13536 (*vec_interleave_highv2df): Ditto.
13537 (*vec_interleave_lowv2df): Ditto.
13538 (*vec_extractv2df_1_sse): Ditto.
13539 (*vec_extractv2df_0_sse): Ditto.
13540 (sse2_storelpd): Ditto.
13541 (sse2_loadlpd): Ditto.
13542 (sse2_movsd): Ditto.
13543 (*vec_concatv4si): Ditto.
13544 (vec_concatv2di): Ditto.
13546 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13547 for Intel asm dialect.
13548 (mmx_punpcklwd): Ditto.
13549 (mmx_punpckldq): Ditto.
13551 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13552 for intel assembler dialect.
13554 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13557 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13558 for BUILT_IN_C{LZ,LRSB}*.
13559 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13560 * tree-vrp.c (extract_range_basic): Handle
13561 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13562 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13563 fall thru to code calling set_value*.
13564 * builtins.c (expand_builtin): Remove *IMAX cases.
13565 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13566 if width is bigger than 2*HWI.
13568 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13570 PR rtl-optimization/55342
13571 * lra-int.h (lra_subreg_reload_pseudos): New.
13572 * lra.c: Add undoing optional reloads to the block diagram.
13573 (lra_subreg_reload_pseudos): New.
13574 (lra_optional_reload_pseudos): Change comments.
13575 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13576 lra_optional_reload_pseudos after undo transformations.
13577 * lra-assigns.c (pseudo_prefix_title): New.
13578 (lra_setup_reg_renumber): Use it.
13579 (spill_for): Ditto. Check subreg reload pseudos too.
13580 (assign_by_spills): Consider subreg reload pseudos too.
13581 * lra-constraints.c (simplify_operand_subreg): Use
13582 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13583 (curr_insn_transform): Recognize and do optional reloads.
13584 (undo_optional_reloads): New.
13585 (lra_undo_inheritance): Call undo_optional_reloads.
13587 2013-07-05 Thomas Quinot <quinot@adacore.com>
13589 * tree-complex.c (expand_complex_operations_1): Fix typo.
13591 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13593 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13594 (tune_params): New member 'const vec_costs'.
13595 * config/aarch64/aarch64.c (generic_vector_cost): New.
13596 (generic_tunings): New member 'generic_vector_cost'.
13597 (aarch64_builtin_vectorization_cost): New.
13598 (aarch64_add_stmt_cost): New.
13599 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13600 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13602 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13605 * config/i386/predicates.md (vsib_address_operand): Disallow
13606 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13608 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13610 PR middle-end/55030
13611 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13612 expand_builtin_setjmp_receiver.
13613 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13614 with NULL for the label parameter.
13615 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13616 the frame-pointer. Adjust comments.
13617 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13618 only if LABEL is non-NULL.
13620 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13622 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13623 (ARM_ABI_AAPCS64): Ditto.
13625 (ARM_DEFAULT_ABI): Ditto.
13627 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13629 * config/aarch64/aarch64-builtins.c
13630 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13631 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13633 * config/aarch64/aarch64-simd.md
13634 (aarch64_ld1<VALL:mode>): New.
13635 (aarch64_st1<VALL:mode>): Likewise.
13636 * config/aarch64/arm_neon.h
13637 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13639 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13641 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13643 (ix86_option_override_internal):Turn on all -mavx target flags by
13644 default as they are dependent on AVX anyway.
13646 2013-07-02 Cary Coutant <ccoutant@google.com>
13648 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13649 deterministic hash.
13650 (loc_checksum_ordered): Likewise.
13651 (hash_loc_operands): Remove inline keyword.
13653 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13655 PR tree-optimization/57741
13656 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13657 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13658 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13659 Allow REAL_CST step_exprs if flag_associative_math.
13660 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13662 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13664 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13666 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13668 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13670 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13672 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13674 (iorsi3_insn): Likewise.
13675 (arm_xorsi3): Likewise.
13677 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13679 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13681 (attribute "type"): ... here, and prefix with "wmmx_".
13682 (attribute "core_cycles"): Update for attribute changes.
13683 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13684 (tbcstv4hi): Likewise.
13685 (tbcstv2si): Likewise.
13686 (iwmmxt_iordi3): Likewise.
13687 (iwmmxt_xordi3): Likewise.
13688 (iwmmxt_anddi3): Likewise.
13689 (iwmmxt_nanddi3): Likewise.
13690 (iwmmxt_arm_movdi): Likewise.
13691 (iwmmxt_movsi_insn): Likewise.
13692 (mov<mode>_internal): Likewise.
13693 (and<mode>3_iwmmxt): Likewise.
13694 (ior<mode>3_iwmmxt): Likewise.
13695 (xor<mode>3_iwmmxt): Likewise.
13696 (add<mode>3_iwmmxt): Likewise.
13697 (ssaddv8qi3): Likewise.
13698 (ssaddv4hi3): Likewise.
13699 (ssaddv2si3): Likewise.
13700 (usaddv8qi3): Likewise.
13701 (usaddv4hi3): Likewise.
13702 (usaddv2si3): Likewise.
13703 (sub<mode>3_iwmmxt): Likewise.
13704 (sssubv8qi3): Likewise.
13705 (sssubv4hi3): Likewise.
13706 (sssubv2si3): Likewise.
13707 (ussubv8qi3): Likewise.
13708 (ussubv4hi3): Likewise.
13709 (ussubv2si3): Likewise.
13710 (mulv4hi3_iwmmxt): Likewise.
13711 (smulv4hi3_highpart): Likewise.
13712 (umulv4hi3_highpart): Likewise.
13713 (iwmmxt_wmacs): Likewise.
13714 (iwmmxt_wmacsz): Likewise.
13715 (iwmmxt_wmacu): Likewise.
13716 (iwmmxt_wmacuz): Likewise.
13717 (iwmmxt_clrdi): Likewise.
13718 (iwmmxt_clrv8qi): Likewise.
13719 (iwmmxt_clr4hi): Likewise.
13720 (iwmmxt_clr2si): Likewise.
13721 (iwmmxt_uavgrndv8qi3): Likewise.
13722 (iwmmxt_uavgrndv4hi3): Likewise.
13723 (iwmmxt_uavgv8qi3): Likewise.
13724 (iwmmxt_uavgv4hi3): Likewise.
13725 (iwmmxt_tinsrb): Likewise.
13726 (iwmmxt_tinsrh): Likewise.
13727 (iwmmxt_tinsrw): Likewise.
13728 (iwmmxt_textrmub): Likewise.
13729 (iwmmxt_textrmsb): Likewise.
13730 (iwmmxt_textrmuh): Likewise.
13731 (iwmmxt_textrmsh): Likewise.
13732 (iwmmxt_textrmw): Likewise.
13733 (iwmxxt_wshufh): Likewise.
13734 (eqv8qi3): Likewise.
13735 (eqv4hi3): Likewise.
13736 (eqv2si3): Likewise.
13737 (gtuv8qi3): Likewise.
13738 (gtuv4hi3): Likewise.
13739 (gtuv2si3): Likewise.
13740 (gtv8qi3): Likewise.
13741 (gtv4hi3): Likewise.
13742 (gtv2si3): Likewise.
13743 (smax<mode>3_iwmmxt): Likewise.
13744 (umax<mode>3_iwmmxt): Likewise.
13745 (smin<mode>3_iwmmxt): Likewise.
13746 (umin<mode>3_iwmmxt): Likewise.
13747 (iwmmxt_wpackhss): Likewise.
13748 (iwmmxt_wpackwss): Likewise.
13749 (iwmmxt_wpackdss): Likewise.
13750 (iwmmxt_wpackhus): Likewise.
13751 (iwmmxt_wpackwus): Likewise.
13752 (iwmmxt_wpackdus): Likewise.
13753 (iwmmxt_wunpckihb): Likewise.
13754 (iwmmxt_wunpckihh): Likewise.
13755 (iwmmxt_wunpckihw): Likewise.
13756 (iwmmxt_wunpckilb): Likewise.
13757 (iwmmxt_wunpckilh): Likewise.
13758 (iwmmxt_wunpckilw): Likewise.
13759 (iwmmxt_wunpckehub): Likewise.
13760 (iwmmxt_wunpckehuh): Likewise.
13761 (iwmmxt_wunpckehuw): Likewise.
13762 (iwmmxt_wunpckehsb): Likewise.
13763 (iwmmxt_wunpckehsh): Likewise.
13764 (iwmmxt_wunpckehsw): Likewise.
13765 (iwmmxt_wunpckelub): Likewise.
13766 (iwmmxt_wunpckeluh): Likewise.
13767 (iwmmxt_wunpckeluw): Likewise.
13768 (iwmmxt_wunpckelsb): Likewise.
13769 (iwmmxt_wunpckelsh): Likewise.
13770 (iwmmxt_wunpckelsw): Likewise.
13771 (ror<mode>3): Likewise.
13772 (ashr<mode>3_iwmmxt): Likewise.
13773 (lshr<mode>3_iwmmxt): Likewise.
13774 (ashl<mode>3_iwmmxt): Likewise.
13775 (ror<mode>3_di): Likewise.
13776 (ashr<mode>3_di): Likewise.
13777 (lshr<mode>3_di): Likewise.
13778 (ashl<mode>3_di): Likewise.
13779 (iwmmxt_wmadds): Likewise.
13780 (iwmmxt_wmaddu): Likewise.
13781 (iwmmxt_tmia): Likewise.
13782 (iwmmxt_tmiaph): Likewise.
13783 (iwmmxt_tmiabb): Likewise.
13784 (iwmmxt_tmiatb): Likewise.
13785 (iwmmxt_tmiabt): Likewise.
13786 (iwmmxt_tmiatt): Likewise.
13787 (iwmmxt_tmovmskb): Likewise.
13788 (iwmmxt_tmovmskh): Likewise.
13789 (iwmmxt_tmovmskw): Likewise.
13790 (iwmmxt_waccb): Likewise.
13791 (iwmmxt_wacch): Likewise.
13792 (iwmmxt_waccw): Likewise.
13793 (iwmmxt_waligni): Likewise.
13794 (iwmmxt_walignr): Likewise.
13795 (iwmmxt_walignr0): Likewise.
13796 (iwmmxt_walignr1): Likewise.
13797 (iwmmxt_walignr2): Likewise.
13798 (iwmmxt_walignr3): Likewise.
13799 (iwmmxt_wsadb): Likewise.
13800 (iwmmxt_wsadh): Likewise.
13801 (iwmmxt_wsadbz): Likewise.
13802 (iwmmxt_wsadhz): Likewise.
13803 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13804 (iwmmxt_wabsdiffb): Likewise.
13805 (iwmmxt_wabsdiffh): Likewise.
13806 (iwmmxt_wabsdiffw): Likewise.
13807 (iwmmxt_waddsubhx): Likewise
13808 (iwmmxt_wsubaddhx): Likewise.
13809 (addc<mode>3): Likewise.
13810 (iwmmxt_avg4): Likewise.
13811 (iwmmxt_avg4r): Likewise.
13812 (iwmmxt_wmaddsx): Likewise.
13813 (iwmmxt_wmaddux): Likewise.
13814 (iwmmxt_wmaddsn): Likewise.
13815 (iwmmxt_wmaddun): Likewise.
13816 (iwmmxt_wmulwsm): Likewise.
13817 (iwmmxt_wmulwum): Likewise.
13818 (iwmmxt_wmulsmr): Likewise.
13819 (iwmmxt_wmulumr): Likewise.
13820 (iwmmxt_wmulwsmr): Likewise.
13821 (iwmmxt_wmulwumr): Likewise.
13822 (iwmmxt_wmulwl): Likewise.
13823 (iwmmxt_wqmulm): Likewise.
13824 (iwmmxt_wqmulwm): Likewise.
13825 (iwmmxt_wqmulmr): Likewise.
13826 (iwmmxt_wqmulwmr): Likewise.
13827 (iwmmxt_waddbhusm): Likewise.
13828 (iwmmxt_waddbhusl): Likewise.
13829 (iwmmxt_wqmiabb): Likewise.
13830 (iwmmxt_wqmiabt): Likewise.
13831 (iwmmxt_wqmiatb): Likewise.
13832 (iwmmxt_wqmiatt): Likewise.
13833 (iwmmxt_wqmiabbn): Likewise.
13834 (iwmmxt_wqmiabtn): Likewise.
13835 (iwmmxt_wqmiatbn): Likewise.
13836 (iwmmxt_wqmiattn): Likewise.
13837 (iwmmxt_wmiabb): Likewise.
13838 (iwmmxt_wmiabt): Likewise.
13839 (iwmmxt_wmiatb): Likewise.
13840 (iwmmxt_wmiatt): Likewise.
13841 (iwmmxt_wmiabbn): Likewise.
13842 (iwmmxt_wmiabtn): Likewise.
13843 (iwmmxt_wmiatbn): Likewise.
13844 (iwmmxt_wmiattn): Likewise.
13845 (iwmmxt_wmiawbb): Likewise.
13846 (iwmmxt_wmiawbt): Likewise.
13847 (iwmmxt_wmiawtb): Likewise.
13848 (iwmmxt_wmiawtt): Likewise.
13849 (iwmmxt_wmiawbbn): Likewise.
13850 (iwmmxt_wmiawbtn): Likewise.
13851 (iwmmxt_wmiawtbn): Likewise.
13852 (iwmmxt_wmiawttn): Likewise.
13853 (iwmmxt_wmerge): Likewise.
13854 (iwmmxt_tandc<mode>3): Likewise.
13855 (iwmmxt_torc<mode>3): Likewise.
13856 (iwmmxt_torvsc<mode>3): Likewise.
13857 (iwmmxt_textrc<mode>3): Likewise.
13858 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
13859 (wmmxt_pack): Likewise.
13860 (wmmxt_mult_c1): Likewise.
13861 (wmmxt_mult_c2): Likewise.
13862 (wmmxt_alu_c1): Likewise.
13863 (wmmxt_alu_c2): Likewise.
13864 (wmmxt_alu_c3): Likewise.
13865 (wmmxt_transfer_c1): Likewise.
13866 (wmmxt_transfer_c2): Likewise.
13867 (wmmxt_transfer_c3): Likewise.
13868 (marvell_f_iwmmxt_wstr): Likewise.
13869 (marvell_f_iwmmxt_wldr): Likewise.
13871 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
13873 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
13875 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13878 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13879 * lra-constraints.c (need_for_split_p): Check call used hard regs
13880 living through calls.
13882 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
13883 call used regs for call insn.
13885 2013-06-28 Jakub Jelinek <jakub@redhat.com>
13888 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
13889 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
13891 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
13893 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
13895 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
13897 * lra-constraints.c (need_for_split_p): Check call used hard regs
13898 living through calls.
13900 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
13903 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
13904 to tie with any other modes. Eliminate Altivec vector mode tests,
13905 since these are a subset of ALTIVEC or VSX vector modes. Simplify
13906 code, to return 0 if testing MODE2 for a condition, if we've
13907 already tested MODE1 for the same condition.
13909 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13911 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
13914 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13916 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13917 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
13919 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
13921 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
13923 (aarch64_symbolic_constant_p): Remove.
13924 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
13925 static. Fix line length and white space.
13926 (aarch64_symbolic_constant_p): Remove.
13927 * config/aarch64/predicates.md (aarch64_valid_symref):
13928 Use aarch64_classify_symbol_expression.
13930 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13932 * config/arm/constraints.md (Ts): New constraint.
13933 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
13935 (compare_scc): Use "Ts" constraint for operand 0.
13936 (ior_scc_scc): Likewise.
13937 (and_scc_scc): Likewise.
13938 (and_scc_scc_nodom): Likewise.
13939 (ior_scc_scc_cmp): Likewise for operand 7.
13940 (and_scc_scc_cmp): Likewise.
13941 * config/arm/thumb2.md (thumb2_movsi_insn):
13942 Add alternatives for 16-bit encodings.
13943 (thumb2_movhi_insn): Likewise.
13944 (thumb2_movsicc_insn): Likewise.
13945 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
13946 (thumb2_negscc): Use "Ts" constraint.
13947 Move mvn instruction outside cond_exec block.
13948 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
13949 for 16-bit encodings.
13951 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13953 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
13955 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
13956 (mulsi3subsi): Likewise.
13957 (mulsidi3adddi): Likewise.
13958 (mulsidi3_v6): Likewise.
13959 (umulsidi3_v6): Likewise.
13960 (umulsidi3adddi_v6): Likewise.
13961 (smulsi3_highpart_v6): Likewise.
13962 (umulsi3_highpart_v6): Likewise.
13963 (mulhisi3tb): Likewise.
13964 (mulhisi3bt): Likewise.
13965 (mulhisi3tt): Likewise.
13966 (maddhisi4): Likewise.
13967 (maddhisi4tb): Likewise.
13968 (maddhisi4tt): Likewise.
13969 (maddhidi4): Likewise.
13970 (maddhidi4tb): Likewise.
13971 (maddhidi4tt): Likewise.
13972 (zeroextractsi_compare0_scratch): Likewise.
13973 (insv_zero): Likewise.
13974 (insv_t2): Likewise.
13975 (anddi_notzesidi_di): Likewise.
13976 (anddi_notsesidi_di): Likewise.
13977 (andsi_notsi_si): Likewise.
13978 (iordi_zesidi_di): Likewise.
13979 (xordi_zesidi_di): Likewise.
13980 (andsi_iorsi3_notsi): Likewise.
13981 (smax_0): Likewise.
13982 (smax_m1): Likewise.
13983 (smin_0): Likewise.
13984 (not_shiftsi): Likewise.
13985 (unaligned_loadsi): Likewise.
13986 (unaligned_loadhis): Likewise.
13987 (unaligned_loadhiu): Likewise.
13988 (unaligned_storesi): Likewise.
13989 (unaligned_storehi): Likewise.
13990 (extv_reg): Likewise.
13991 (extzv_t2): Likewise.
13992 (divsi3): Likewise.
13993 (udivsi3): Likewise.
13994 (arm_zero_extendhisi2addsi): Likewise.
13995 (arm_zero_extendqisi2addsi): Likewise.
13996 (compareqi_eq0): Likewise.
13997 (arm_extendhisi2_v6): Likewise.
13998 (arm_extendqisi2addsi): Likewise.
13999 (arm_movt): Likewise.
14000 (thumb2_ldrd): Likewise.
14001 (thumb2_ldrd_base): Likewise.
14002 (thumb2_ldrd_base_neg): Likewise.
14003 (thumb2_strd): Likewise.
14004 (thumb2_strd_base): Likewise.
14005 (thumb2_strd_base_neg): Likewise.
14006 (arm_negsi2): Add alternative for 16-bit encoding.
14007 (arm_one_cmplsi2): Likewise.
14009 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14011 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
14012 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
14013 (movdfcc): Likewise.
14014 * config/arm/vfp.md (*thumb2_movsf_vfp):
14015 Disable predication for arm_restrict_it.
14016 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
14017 (*thumb2_movdfcc_vfp): Likewise.
14018 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
14019 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
14020 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
14021 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
14022 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
14023 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
14024 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
14025 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
14026 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
14027 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
14028 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
14029 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
14030 Disable predication for arm_restrict_it.
14032 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
14034 * config/i386/bmiintrin.h (_bextr_u32): New.
14035 (_bextr_u64): Ditto.
14037 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
14039 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
14040 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
14041 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
14042 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
14043 * config/mips/n32-elf.h: ...this new file.
14045 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
14048 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
14049 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
14051 2013-06-27 Catherine Moore <clm@codesourcery.com>
14053 * config/mips/mips-tables.opt: Regenerate.
14054 * config/mips/mips-cpus.def: Add m14ke and m14kec.
14055 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
14056 * doc/invoke.texi: Add -m14kc.
14058 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14061 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
14062 constraints of operand 1 and 2.
14065 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
14066 to match RTL canonicalization. Swap predicates and
14067 constraints of operand 1 and 2.
14069 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
14071 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
14072 Process OP_INOUT regs for splitting too.
14074 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14076 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
14077 decl before the loop, initialize to NULL.
14078 (vectorizable_load): Initialize ptr_incr to NULL.
14080 2013-06-27 Martin Jambor <mjambor@suse.cz>
14083 * ipa-ref.h (ipa_maybe_record_reference): Declare.
14084 * ipa-ref.c (ipa_maybe_record_reference): New function.
14085 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
14086 * ipa-cp.c (create_specialized_node): Record potential references from
14088 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
14090 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14092 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
14093 parameter 'mode' of type 'enum machine_mode mode'; change to pass
14094 'mode' to force_reg.
14095 (aarch64_add_offset): Update calls to aarch64_force_temporary.
14096 (aarch64_expand_mov_immediate): Likewise.
14098 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14100 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
14101 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
14103 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14105 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
14106 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14107 (struct machine_function): Add tbegin_p.
14108 (s390_canonicalize_comparison): Fold CC mode compares to
14109 conditional jump if possible.
14110 (s390_emit_jump): Return the emitted jump.
14111 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
14112 Handle CCRAWmode compares.
14113 (s390_option_override): Default to -mhtm if available.
14114 (s390_reg_clobbered_rtx): Handle floating point regs as well.
14115 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
14116 FPRs instead of df_regs_ever_live_p.
14117 (s390_optimize_nonescaping_tx): New function.
14118 (s390_init_frame_layout): Extend clobbered_regs array to cover
14120 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
14121 (s390_expand_tbegin): New function.
14122 (enum s390_builtin): New enum definition.
14123 (code_for_builtin): New array definition.
14124 (s390_init_builtins): New function.
14125 (s390_expand_builtin): New function.
14126 (TARGET_INIT_BUILTINS): Define.
14127 (TARGET_EXPAND_BUILTIN): Define.
14128 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
14129 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
14130 (s390_alc_comparison): Likewise.
14131 * config/s390/s390-modes.def: Add CCRAWmode.
14132 * config/s390/s390.h (processor_flags): Add PF_TX.
14133 (TARGET_CPU_HTM): Define macro.
14134 (TARGET_HTM): Define macro.
14135 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
14136 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
14137 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14138 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
14139 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
14140 (TBEGIN_MASK, TBEGINC_MASK): New constants.
14141 ("*cc_to_int"): Move up.
14142 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
14143 constants other than 0.
14144 ("*ccraw_to_int"): New insn and splitter definition.
14145 ("tbegin", "tbegin_nofloat", "tbegin_retry")
14146 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
14147 ("tx_assist"): New expander.
14148 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
14149 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
14150 * config/s390/s390.opt: Add -mhtm option.
14151 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
14152 * config/s390/htmxlintrin.h: New file.
14153 * config/s390/htmintrin.h: New file.
14154 * config/s390/s390intrin.h: New file.
14155 * doc/extend.texi: Document htm builtins.
14156 * config.gcc: Add the new header files to extra_headers.
14158 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
14160 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
14161 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
14163 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
14164 Pat Haugen <pthaugen@us.ibm.com>
14165 Peter Bergner <bergner@vnet.ibm.com>
14167 * config/rs6000/power8.md: New.
14168 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
14169 setting for power8 entry.
14170 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
14171 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
14172 test for Power4/Power5 only.
14173 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
14175 (force_new_group): Adjust comment.
14176 * config/rs6000/rs6000.md: Include power8.md.
14178 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14180 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14181 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14183 (tune_params): Update comment.
14184 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14185 (arm_max_conditional_execute): New function.
14186 (thumb2_final_prescan_insn): Use max_insn_skipped and
14187 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14189 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14191 PR tree-optimization/57705
14192 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14193 SSA_NAME step, provided that it is not defined inside the loop.
14194 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14195 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14197 2013-06-25 Martin Jambor <mjambor@suse.cz>
14199 PR middle-end/57670
14200 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14201 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14203 (ipa_note_param_call): Initialize member_ptr flag.
14204 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14205 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14206 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14207 (ipa_read_indirect_edge_info): Likewise.
14209 2013-06-25 Richard Biener <rguenther@suse.de>
14211 PR middle-end/56977
14212 * passes.c (init_optimization_passes): Move pass_fold_builtins
14213 and pass_dce earlier with -Og.
14215 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14217 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14218 <BIT_FIELD_REF>: Remove trailing TAB.
14219 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14222 2013-06-24 Martin Jambor <mjambor@suse.cz>
14224 PR tree-optimization/57358
14225 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14226 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14227 (ipa_analyze_params_uses): Generate pessimistic info when true.
14229 2013-06-24 Martin Jambor <mjambor@suse.cz>
14231 PR tree-optimization/57539
14232 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14233 global.inlined_to of the new node to it. All callers changed.
14234 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14235 inlining_into, pass it to cgraph_clone_node.
14236 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14237 ipa_free_edge_args_substructures.
14238 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14239 rdesc linked list. Do not assert rdesc edges have inlined caller.
14240 Assert we have found an rdesc in the rdesc list.
14242 2013-06-24 Richard Biener <rguenther@suse.de>
14244 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14245 (pointer_set_lookup): Declare.
14246 (class pointer_map): New template class implementing a
14247 generic pointer to T map.
14248 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14249 pointer_map<T>::contains, pointer_map<T>::insert,
14250 pointer_map<T>::traverse): New functions.
14251 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14252 (pointer_set_lookup): New function.
14253 (pointer_set_contains): Use pointer_set_lookup.
14254 (pointer_set_insert): Likewise.
14255 (insert_aux): Remove.
14256 (struct pointer_map_t): Embed a pointer_set_t.
14257 (pointer_map_create): Adjust.
14258 (pointer_map_destroy): Likewise.
14259 (pointer_map_contains): Likewise.
14260 (pointer_map_insert): Likewise.
14261 (pointer_map_traverse): Likewise.
14262 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14263 pointer_map<unsigned> instead of a pointer_map_t.
14264 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14265 (streamer_tree_cache_lookup): Likewise.
14266 (streamer_tree_cache_create): Likewise.
14267 (streamer_tree_cache_delete): Likewise.
14268 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14269 pointer_map<unsigned> instead of a pointer_map_t.
14270 (lto_init_tree_ref_encoder): Adjust.
14271 (lto_destroy_tree_ref_encoder): Likewise.
14272 * lto-section-out.c (lto_output_decl_index): Likewise.
14273 (lto_record_function_out_decl_state): Likewise.
14274 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14276 2013-06-24 Richard Biener <rguenther@suse.de>
14278 PR tree-optimization/57488
14279 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14281 2013-06-24 Alan Modra <amodra@gmail.com>
14283 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14284 (gen_easy_altivec_constant): Likewise.
14285 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14286 easy_vector_constant_msb): Likewise.
14288 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14291 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14292 add missing return true.
14294 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14297 * config/sh/predicates.md (general_extend_operand): Invoke
14298 general_movsrc_operand for memory operands.
14299 (general_movsrc_operand): Allow reg+reg addressing, do not use
14300 general_operand for memory operands.
14302 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14304 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14305 when current target options does not apply.
14306 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14307 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14308 * config/i386/bmiintrin.h: Pass appropriate target
14309 attributes to header.
14310 * config/i386/mmintrin.h: Ditto.
14311 * config/i386/nmmintrin.h: Ditto.
14312 * config/i386/avx2intrin.h: Ditto.
14313 * config/i386/fxsrintrin.h: Ditto.
14314 * config/i386/tbmintrin.h: Ditto.
14315 * config/i386/xsaveintrin.h: Ditto.
14316 * config/i386/f16cintrin.h: Ditto.
14317 * config/i386/xtestintrin.h: Ditto.
14318 * config/i386/xsaveoptintrin.h: Ditto.
14319 * config/i386/bmi2intrin.h: Ditto.
14320 * config/i386/lzcntintrin.h: Ditto.
14321 * config/i386/smmintrin.h: Ditto.
14322 * config/i386/wmmintrin.h: Ditto.
14323 * config/i386/x86intrin.h: Remove all header include guards.
14324 * config/i386/prfchwintrin.h: Ditto.
14325 * config/i386/pmmintrin.h: Ditto.
14326 * config/i386/tmmintrin.h: Ditto.
14327 * config/i386/xmmintrin.h: Ditto.
14328 * config/i386/popcntintrin.h: Ditto.
14329 * config/i386/rdseedintrin.h: Ditto.
14330 * config/i386/ammintrin.h: Ditto.
14331 * config/i386/emmintrin.h: Ditto.
14332 * config/i386/immintrin.h: Remove all header include guards.
14333 * config/i386/fma4intrin.h: Ditto.
14334 * config/i386/lwpintrin.h: Ditto.
14335 * config/i386/xopintrin.h: Ditto.
14336 * config/i386/ia32intrin.h: Ditto.
14337 * config/i386/avxintrin.h: Ditto.
14338 * config/i386/rtmintrin.h: Ditto.
14339 * config/i386/fmaintrin.h: Ditto.
14340 * config/i386/mm3dnow.h: Ditto.
14342 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14344 * common/config/i386/i386-common.c: Handle LZCNT.
14346 2013-06-22 Andi Kleen <ak@linux.intel.com>
14348 * doc/extend.texi: Use __atomic_store_n instead of
14349 __atomic_store in HLE example.
14351 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14353 * config/sh/sh.c: Remove <cstdlib> workaround.
14355 2013-06-21 Andi Kleen <ak@linux.intel.com>
14357 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14359 2013-06-21 Andi Kleen <ak@linux.intel.com>
14361 * doc/extend.texi: Document that __atomic_clear and
14362 __atomic_test_and_set should only be used with bool.
14364 2013-06-20 Jan Hubicka <jh@suse.cz>
14366 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14367 types_same_for_odr.
14368 * tree.c (decls_same_for_odr): New function.
14369 (same_for_edr): New function.
14370 (types_same_for_odr): New function.
14371 (get_binfo_at_offset): Use it.
14372 * tree.h (types_same_for_odr): Declare.
14374 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14375 Jason Merrill <jason@redhat.com>
14377 * system.h: Include <cstdlib> as well as <stdlib.h>.
14379 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14382 * config/i386/i386.c (construct_container): Report error if
14383 long double is used with disabled x87 float returns.
14385 2013-06-20 Jan Hubicka <jh@suse.cz>
14387 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14389 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14391 PR rtl-optimization/57425
14392 PR rtl-optimization/57569
14393 * alias.c (write_dependence_p): Remove parameters mem_mode and
14394 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14395 Changed all callers.
14396 (canon_anti_dependence): Get comments and semantics in sync.
14397 Add parameter mem_canonicalized. Changed all callers.
14398 * rtl.h (canon_anti_dependence): Update prototype.
14400 2013-06-20 Richard Biener <rguenther@suse.de>
14402 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14403 byte case, inline streamer_read_uchar and defer section
14406 2013-06-20 Richard Biener <rguenther@suse.de>
14408 PR tree-optimization/57584
14409 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14410 SSA names into the expanded expression that take part in
14411 abnormal coalescing.
14413 2013-06-19 Sharad Singhai <singhai@google.com>
14415 * gcov.c (print_usage): Handle new option.
14416 (process_args): Ditto.
14417 (get_gcov_intermediate_filename): New function.
14418 (output_intermediate_file): New function.
14419 (output_gcov_file): New function
14420 (generate_results): Handle new option.
14421 (release_function): Relase demangled name.
14422 (read_graph_file): Handle demangled name.
14423 (output_lines): Ditto.
14424 * doc/gcov.texi: Document gcov intermediate format.
14426 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14429 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14430 (lra_emit_add): Use the functions. Add comment about Y as an
14433 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14436 * collect2.c (collect_atexit): New.
14437 (collect_exit): Delete.
14438 (main): Register collect_atexit with atexit.
14439 (collect_wait): Change collect_exit to exit.
14441 * collect2.h (collect_exit): Delete.
14442 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14444 2013-06-19 Wei Mi <wmi@google.com>
14446 PR rtl-optimization/57518
14447 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14448 if regno is used in paradoxical subreg.
14449 (update_equiv_regs): Check pdx_subregs[regno] before
14450 set a reg to be equivalent with a mem.
14452 2013-06-19 Matthias Klose <doko@ubuntu.com>
14455 * file-find.h (find_a_file): Add a mode parameter.
14456 * file-find.c (find_a_file): Likewise.
14457 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14458 with X_OK for the executables.
14459 * collect2.c (main): Call find_a_file with X_OK.
14461 2013-06-19 Steve Ellcey <sellcey@mips.com>
14464 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14465 Use NEXT_INSN instead of next_real_insn.
14467 2013-06-19 Jan Hubicka <jh@suse.cz>
14469 * cgraph.h (const_value_known_p): Replace by ...
14470 (ctor_for_folding): .. this one.
14471 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14472 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14473 * expr.c (expand_expr_real_1): Likewise.
14474 (string_constant): Likewise.
14475 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14476 * ipa.c (process_references): Likewise.
14477 (symtab_remove_unreachable_nodes): Likewise.
14478 * ipa-inline-analysis.c (param_change_prob): Likewise.
14479 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14480 (get_base_constructor): Likwise.
14481 * varpool.c (varpool_remove_node): Likewise.
14482 (varpool_remove_initializer): LIkewise.
14483 (dump_varpool_node): LIkwise.
14484 (const_value_known_p): Rewrite to ...
14485 (ctor_for_folding): ... this one.
14487 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14490 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14491 PERSONALITY in $PATH derived prefixes.
14493 2013-06-19 Jeff Law <law@redhat.com>
14495 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14498 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14499 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14502 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14504 * config/arm/vfp.md: Move VFP instruction classification documentation
14506 * config/arm/arm.md: ... here. Update instruction classification
14509 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14511 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14512 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14513 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14515 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14518 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14519 with NEXT_INSN. Use tablejump_p to check for jump table data
14522 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14525 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14526 that now in C++ the value is correct per the C++ standards.
14528 2013-06-19 Richard Biener <rguenther@suse.de>
14530 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14531 for global context.
14533 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14536 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14539 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14540 with next_active_insn.
14542 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14544 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14545 functions are inlined during failures to flag an error.
14546 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14547 in early inline pass.
14549 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14551 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14554 2013-06-18 Julian Brown <julian@codesourcery.com>
14556 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14557 Permit virtual register pre-reload if !strict.
14558 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14560 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14562 * config/arm/neon.md (movmisalign<mode>): Use
14563 neon_perm_struct_or_reg_operand instead of
14564 neon_struct_or_register_operand.
14565 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14566 neon_permissive_struct_operand instead of neon_struct_operand.
14567 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14568 neon_vector_mem_operand.
14569 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14570 neon_vector_mem_operand.
14571 (neon_permissive_struct_operand): New.
14572 (neon_struct_or_register_operand): Rename to...
14573 (neon_perm_struct_or_reg_operand): This. Adjust call to
14574 neon_vector_mem_operand.
14576 2013-06-18 Richard Biener <rguenther@suse.de>
14578 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14579 * lto-streamer.h: Include pointer-set.h.
14580 (struct lto_decl_slot): Remove.
14581 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14582 Remove next_index entry.
14583 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14584 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14585 (lto_init_tree_ref_encoder): Adjust.
14586 (lto_destroy_tree_ref_encoder): Likewise.
14587 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14588 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14589 (lto_output_decl_index): Adjust.
14590 (lto_new_out_decl_state): Likewise.
14591 (lto_record_function_out_decl_state): Likewise.
14592 * lto-streamer-out.c (copy_function): Likewise.
14594 2013-06-18 Richard Biener <rguenther@suse.de>
14596 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14597 * cgraphunit.c: Include cfgloop.h.
14598 (init_lowered_empty_function): Initialize the loop tree.
14599 (assemble_thunk): Insert new BBs into loops.
14601 2013-06-18 Richard Biener <rguenther@suse.de>
14603 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14604 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14605 the map from cache entry to cache index optional.
14606 (streamer_tree_cache_replace_tree): Adjust accordingly.
14607 (streamer_tree_cache_append): Likewise.
14608 (streamer_tree_cache_delete): Likewise.
14609 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14610 streamer cache map from cache entry to cache index.
14611 * lto-streamer-out.c (create_output_block): Adjust.
14613 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14615 * config/arm/arm.md (attribute "insn"): Move multiplication and
14616 division attributes to...
14617 (attribute "type"): ... here. Remove mult.
14618 (attribute "mul32"): New attribute.
14619 (attribute "mul64"): Add umaal.
14620 (*arm_mulsi3): Update attributes.
14621 (*arm_mulsi3_v6): Likewise.
14622 (*thumb_mulsi3): Likewise.
14623 (*thumb_mulsi3_v6): Likewise.
14624 (*mulsi3_compare0): Likewise.
14625 (*mulsi3_compare0_v6): Likewise.
14626 (*mulsi_compare0_scratch): Likewise.
14627 (*mulsi_compare0_scratch_v6): Likewise.
14628 (*mulsi3addsi): Likewise.
14629 (*mulsi3addsi_v6): Likewise.
14630 (*mulsi3addsi_compare0): Likewise.
14631 (*mulsi3addsi_compare0_v6): Likewise.
14632 (*mulsi3addsi_compare0_scratch): Likewise.
14633 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14634 (*mulsi3subsi): Likewise.
14635 (*mulsidi3adddi): Likewise.
14636 (*mulsi3addsi_v6): Likewise.
14637 (*mulsidi3adddi_v6): Likewise.
14638 (*mulsidi3_nov6): Likewise.
14639 (*mulsidi3_v6): Likewise.
14640 (*umulsidi3_nov6): Likewise.
14641 (*umulsidi3_v6): Likewise.
14642 (*umulsidi3adddi): Likewise.
14643 (*umulsidi3adddi_v6): Likewise.
14644 (*smulsi3_highpart_nov6): Likewise.
14645 (*smulsi3_highpart_v6): Likewise.
14646 (*umulsi3_highpart_nov6): Likewise.
14647 (*umulsi3_highpart_v6): Likewise.
14648 (mulhisi3): Likewise.
14649 (*mulhisi3tb): Likewise.
14650 (*mulhisi3bt): Likewise.
14651 (*mulhisi3tt): Likewise.
14652 (maddhisi4): Likewise.
14653 (*maddhisi4tb): Likewise.
14654 (*maddhisi4tt): Likewise.
14655 (maddhidi4): Likewise.
14656 (*maddhidi4tb): Likewise.
14657 (*maddhidi4tt): Likewise.
14658 (divsi3): Likewise.
14659 (udivsi3): Likewise.
14660 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14661 (thumb2_mulsi_short_compare0): Likewise.
14662 (thumb2_mulsi_short_compare0_scratch): Likewise.
14663 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14664 (1020mult2): Likewise.
14665 (1020mult3): Likewise.
14666 (1020mult4): Likewise.
14667 (1020mult5): Likewise.
14668 (1020mult6): Likewise.
14669 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14671 (cortex_a15_mult64): Likewise.
14672 (cortex_a15_sdiv): Likewise.
14673 (cortex_a15_udiv): Likewise.
14674 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14680 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14681 (pj4_ir_div): Likewise.
14682 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14683 (11_mult2): Likewise.
14684 (11_mult3): Likewise.
14685 (11_mult4): Likewise.
14686 (11_mult5): Likewise.
14687 (11_mult6): Likewise.
14688 (11_mult7): Likewise.
14689 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14690 (cortex_a8_mla): Likewise.
14691 (cortex_a8_mull): Likewise.
14692 (cortex_a8_smulwy): Likewise.
14693 (cortex_a8_smlald): Likewise.
14694 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14695 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14696 (cortex_r4_mul_3): Likewise.
14697 (cortex_r4_mla_4): Likewise.
14698 (cortex_r4_mla_3): Likewise.
14699 (cortex_r4_smlald): Likewise.
14700 (cortex_r4_mull): Likewise.
14701 (cortex_r4_sdiv): Likewise.
14702 (cortex_r4_udiv): Likewise.
14703 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14704 (cortex_a7_idiv): Likewise.
14705 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14706 (9_mult2): Likewise.
14707 (9_mult3): Likewise.
14708 (9_mult4): Likewise.
14709 (9_mult5): Likewise.
14710 (9_mult6): Likewise.
14711 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14712 (cortex_a53_sdiv): Likewise.
14713 (cortex_a53_udiv): Likewise.
14714 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14715 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14716 (mp626_mult2): Likewise.
14717 (mp626_mult3): Likewise.
14718 (mp626_mult4): Likewise.
14719 * config/arm/fa526.md (526_mult1): Update attribute change.
14720 (526_mult2): Likewise.
14721 * config/arm/arm-generic.md (mult): Update attribute change.
14722 (mult_ldsched_strongarm): Likewise.
14723 (mult_ldsched): Likewise.
14724 (multi_cycle): Likewise.
14725 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14726 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14727 (606te_mult2): Likewise.
14728 (606te_mult3): Likewise.
14729 (606te_mult4): Likewise.
14730 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14731 (cortex_a9_mac16): Likewise.
14732 (cortex_a9_multiply): Likewise.
14733 (cortex_a9_mac): Likewise.
14734 (cortex_a9_multiply_long): Likewise.
14735 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14736 (626te_mult2): Likewise.
14737 (626te_mult3): Likewise.
14738 (626te_mult4): Likewise.
14740 2013-06-18 Richard Biener <rguenther@suse.de>
14743 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14745 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14748 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14749 with next_active_insn.
14751 2013-06-18 Alan Modra <amodra@gmail.com>
14753 * config/rs6000/rs6000.h (enum data_align): New.
14754 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14755 (DATA_ABI_ALIGNMENT): Define.
14756 (CONSTANT_ALIGNMENT): Correct comment.
14757 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14758 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14760 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14762 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14763 ATTRIBUTE_UNUSED marking.
14765 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14767 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14768 alternative and update.
14769 (aarch64_dup_lanedi): Delete.
14770 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14771 * config/aarch64/aarch64-simd-builtins.def: Update.
14773 2013-06-17 Richard Biener <rguenther@suse.de>
14775 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14776 (lto_input_scc): Declare.
14777 (lto_input_tree_1): Likewise.
14778 (struct lto_stats_d): Add num_tree_bodies_output and
14779 num_pickle_refs_output.
14780 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14781 (lto_read_tree_1): Split out from ...
14782 (lto_read_tree): ... this.
14783 (lto_input_scc): New function.
14784 (lto_input_tree_1): Split out from ...
14785 (lto_input_tree): ... this. Handle LTO_tree_scc.
14786 (lto_data_in_create): Create the streamer cache without hashes.
14787 * lto-streamer-out.c (create_output_block): Create the streamer
14788 cache with hashes when not doing WPA.
14789 (lto_write_tree_1): Split out from ...
14790 (lto_write_tree): ... this.
14791 (get_symbol_initial_value): New function.
14792 (lto_output_tree_1): Split out from ...
14793 (lto_output_tree): ... this. Write trees as series of SCCs
14794 using a DFS walk via DFS_write_tree.
14795 (struct sccs, struct scc_entry): New types.
14796 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14797 (DFS_write_tree_body): New function.
14798 (DFS_write_tree): Likewise.
14799 (hash_tree): Likewise.
14800 (scc_entry_compare): Likewise.
14801 (hash_scc): Likewise.
14802 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14803 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14804 TREE_CHAIN as regular reference.
14805 (streamer_read_integer_cst): Remove.
14806 (streamer_get_pickled_tree): Adjust.
14807 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14808 of DECL_EXTERNALs in BLOCK_VARS for now.
14809 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14811 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14812 Add hash value argument and record that if hashes are recorded
14814 (streamer_tree_cache_insert_1): Adjust.
14815 (streamer_tree_cache_insert): Likewise.
14816 (streamer_tree_cache_insert_at): Rename to ...
14817 (streamer_tree_cache_replace_tree): ... this and adjust.
14818 (streamer_tree_cache_append): Adjust.
14819 (record_common_node): Likewise.
14820 (streamer_tree_cache_create): Add argument whether to
14821 record hash values together with trees.
14822 (streamer_tree_cache_delete): Adjust.
14823 * tree-streamer.h (struct streamer_tree_cache_d): Add
14825 (streamer_read_integer_cst): Remove.
14826 (streamer_tree_cache_insert): Adjust.
14827 (streamer_tree_cache_append): Likewise.
14828 (streamer_tree_cache_insert_at): Rename to ...
14829 (streamer_tree_cache_replace_tree): ... this and adjust.
14830 (streamer_tree_cache_create): Add argument whether to record hashes.
14831 (streamer_tree_cache_get): Rename to ...
14832 (streamer_tree_cache_get_tree): ... this.
14833 (streamer_tree_cache_get_hash): New function.
14834 * tree.c (cache_integer_cst): New function.
14835 * tree.h (cache_integer_cst): Declare.
14836 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14837 * lto-symtab.c (lto_varpool_replace_node): Only release
14838 DECL_INITIAL of non-prevailing decls.
14839 * varpool.c (varpool_remove_initializer): Do not release
14840 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
14842 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14844 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
14845 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
14846 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
14847 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
14848 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
14849 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
14850 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
14851 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
14852 instead of TARGET_64BIT.
14853 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
14854 Require ISA_HAS_<D>DIV.
14856 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
14858 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
14859 (mips*-*-linux*): Move default with_llsc setting to where other
14861 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
14863 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
14864 Likewise. Remove default with_tune setting. Move default float
14865 setting to its own block. Handle with_llsc in the same block as above.
14867 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
14869 PR rtl-optimization/57425
14870 PR rtl-optimization/57569
14871 * alias.c (write_dependence_p): Add new parameters mem_mode,
14872 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
14873 Changed all callers.
14874 (canon_anti_dependence): New function.
14875 * cse.c (check_dependence): Use canon_anti_dependence.
14876 * cselib.c (cselib_invalidate_mem): Likewise.
14877 * rtl.h (canon_anti_dependence): Declare.
14879 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
14881 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
14882 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
14883 ".set mips3" for 64-bit targets.
14885 2013-06-15 Dehao Chen <dehao@google.com>
14887 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
14888 * gimple-low.c (gimple_check_call_matching_types): Likewise.
14889 (gimple_check_call_args): Likewise.
14890 * value-prof.c (check_ic_target): Likewise.
14891 * ipa-inline.c (early_inliner): Likewise.
14892 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
14893 * cgraph.c (cgraph_create_edge_1): Likewise.
14894 (cgraph_make_edge_direct): Likewise.
14896 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
14899 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
14900 rs6000_output_move_128bit to handle emitting quad memory
14901 operations. Set attribute length to 8 bytes.
14903 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
14905 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
14907 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
14908 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
14909 (aarch64_<su>mlsl<mode>): Likewise.
14911 2013-06-14 Mike Stump <mikestump@comcast.net>
14913 * Makefile.in (TARGET_H): Add insn-codes.h.
14915 2013-06-14 Alan Modra <amodra@gmail.com>
14917 PR middle-end/57134
14918 PR middle-end/57586
14919 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
14920 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
14921 bitfield expansion when EXPAND_MEMORY.
14922 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
14924 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
14926 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
14927 test for clearing quad memory on 32-bit later.
14929 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14931 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
14932 (fold_negate_expr): Likewise.
14933 (fold_real_zero_addition_p): Handle vectors.
14934 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
14936 2013-06-14 Alan Modra <amodra@gmail.com>
14938 * varasm.c (force_const_mem): Revert 2013-06-07 change.
14940 2013-06-13 Jan Hubicka <jh@suse.cz>
14942 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
14943 Local comdats are not externally visible.
14944 * symtab.c (dump_symtab_base): Dump externally visible.
14945 (verify_symtab_base): Verify back links in the symtab hash.
14947 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14949 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
14950 CONVERT_EXPR as equal nodes.
14952 2013-06-13 Bin Cheng <bin.cheng@arm.com>
14954 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
14956 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
14958 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
14959 Generalize to complex and vector.
14960 * tree.c (build_all_ones_cst): New function.
14961 * tree.h (build_all_ones_cst): Declare it.
14963 2013-06-13 Alan Modra <amodra@gmail.com>
14965 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
14966 * config/rs6000/rs6000.md (signbittf2): New insn.
14967 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
14968 (abstf2_internal, cmptf_internal2): Likewise.
14969 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
14971 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
14972 Pat Haugen <pthaugen@us.ibm.com>
14973 Peter Bergner <bergner@vnet.ibm.com>
14975 * config/rs6000/rs6000.c (emit_load_locked): Add support for
14976 power8 byte, half-word, and quad-word atomic instructions.
14977 (emit_store_conditional): Likewise.
14978 (rs6000_expand_atomic_compare_and_swap): Likewise.
14979 (rs6000_expand_atomic_op): Likewise.
14981 * config/rs6000/sync.md (larx): Add new modes for power8.
14983 (AINT): New mode iterator to include TImode as well as normal
14984 integer modes on power8.
14985 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
14986 that VSX registers are not considered. Use AINT mode iterator
14987 instead of INT1 to allow inclusion of quad word atomic operations
14989 (load_locked<mode>): Likewise.
14990 (store_conditional<mode>): Likewise.
14991 (atomic_compare_and_swap<mode>): Likewise.
14992 (atomic_exchange<mode>): Likewise.
14993 (atomic_nand<mode>): Likewise.
14994 (atomic_fetch_<fetchop_name><mode>): Likewise.
14995 (atomic_nand_fetch<mode>): Likewise.
14996 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
14998 (ATOMIC): On power8, add QImode, HImode modes.
14999 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
15000 modes that promote to SImode.
15001 (load_lockedti): Convert TImode arguments to PTImode, so that we
15002 get a guaranteed even/odd register pair.
15003 (load_lockedpti): Likewise.
15004 (store_conditionalti): Likewise.
15005 (store_conditionalpti): Likewise.
15007 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
15008 atomic load/store instructions.
15011 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
15013 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
15015 (insn_count): New attribute, with most cases extracted from...
15016 (length): ...here. Redefine most cases in terms of insn_count.
15017 (single_insn): Delete.
15018 (can_delay): Use insn_count to check for single instructions.
15019 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
15020 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
15021 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
15022 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
15023 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
15024 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
15025 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
15026 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
15027 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
15028 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
15029 rather than "length".
15030 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
15031 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
15032 Use "insn_count" rather than "length".
15033 * config/mips/mips-dsp.md
15034 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
15035 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
15038 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
15040 PR tree-optimization/57361
15041 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
15043 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
15045 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
15047 (aarch64_simd_combine<mode>): New instruction expansion.
15048 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
15049 function prototype.
15050 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
15051 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
15053 2013-06-12 Jan Hubicka <jh@suse.cz>
15055 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
15056 decl has when in streaming stage.
15057 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
15058 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
15060 2013-06-12 Roland Stigge <stigge@antcom.de>
15063 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
15065 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15067 PR tree-optimization/57537
15068 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
15069 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
15071 2013-06-12 Richard Biener <rguenther@suse.de>
15073 * data-streamer.h (streamer_write_char_stream): CSE
15074 obs->current_pointer.
15075 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
15076 streamer_write_char_stream manually and optimize the resulting loop.
15077 (streamer_write_hwi_stream): Likewise.
15079 2013-06-12 Jan Hubicka <jh@suse.cz>
15081 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
15082 * cgraph.h (varpool_create_empty_node): Declare.
15083 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
15085 * symtab.c (symtab_unregister_node): Be lax about missin entries
15087 (symtab_get_node): Update comment.
15088 * varpool.c (varpool_create_empty_node): Break out from ...
15089 (varpool_node_for_decl): ... here.
15090 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
15092 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
15094 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
15095 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
15096 part. Use straight-line flow at the end.
15097 <COMPONENT_REF>: Remove superfluous else.
15098 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
15100 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15103 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
15104 target hook even for !TREE_PUBLIC decls. If no resolution info
15105 is available, return false for common and external decls.
15107 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15109 * config/rl78/constraints.md (U): New constraint.
15110 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
15113 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15116 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
15117 to allow returning address to AT_PLATFORM name.
15119 2013-06-11 Jan Hubicka <jh@suse.cz>
15121 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
15122 * cgraph.h (symtab_node_base): Add weakref flag.
15123 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
15124 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
15125 (output_weakrefs): Use weakref flag.
15126 * fold-const.c (simple_operand_p): Handle WEAK.
15127 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
15128 * ipa.c (varpool_externally_visible_p): Drop weakref.
15129 (function_and_variable_visibility): Update comment; fix weakref
15130 sanity checks; do not clear DECL_WEAK on them.
15131 * lto-cgraph.c (lto_output_node): update.
15132 (lto_output_varpool_node): Update.
15133 (input_overwrite_node): Update.
15134 (input_node): Update.
15135 (input_varpool_node): Update.
15136 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
15137 (lto_symtab_merge_symbols): Add sanity check.
15138 (lto_symtab_prevailing_decl): Do not special case weakrefs.
15139 * passes.c (rest_of_decl_compilation): Set static flag, too.
15140 * symtab.c (dump_symtab_base): Dump weakref.
15141 (verify_symtab_base): Sanity check weakrefs.
15142 (symtab_make_decl_local): Remove duplicated code.
15143 (symtab_alias_ultimate_target): Simplify.
15144 * varpool.c (varpool_create_variable_alias): Set weakref flag.
15146 2013-06-11 Tom de Vries <tom@codesourcery.com>
15148 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
15149 sequence_vect == NULL.
15151 2013-06-11 DJ Delorie <dj@redhat.com>
15153 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
15154 (rl78_unwind_word_mode): New.
15156 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15158 * final.c (debug_prefix_maps): Make static.
15160 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15162 * function.c (initial_trampoline): Remove stray copy.
15164 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
15166 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
15168 2013-06-11 Martin Jambor <mjambor@suse.cz>
15170 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
15171 within bounds at the beginning of the function.
15173 2013-06-11 Alan Modra <amodra@gmail.com>
15175 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15177 (get_named_section): Don't NULL !DECL_P decl.
15179 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15181 * doc/invoke.texi (core-avx2): Document.
15183 (atom): Updated with MOVBE.
15185 2013-06-11 Richard Biener <rguenther@suse.de>
15187 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15189 2013-06-11 Anton Blanchard <anton@samba.org>
15191 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15192 correct shift value in little-endian mode.
15194 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15197 * varasm.c (get_variable_align): Move #endif to the right place.
15199 2013-06-10 Cary Coutant <ccoutant@google.com>
15201 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15202 for hash so that hash table traversal order is deterministic.
15204 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15205 Pat Haugen <pthaugen@us.ibm.com>
15206 Peter Bergner <bergner@vnet.ibm.com>
15208 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15209 of vectors in GPRS if they are direct moves or quad word load or
15212 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15214 (direct_move_p): Likewise.
15215 (quad_load_store_p): Likewise.
15217 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15218 classes into bins based on the physical register type.
15219 (reg_class_to_reg_type): Likewise.
15220 (IS_STD_REG_TYPE): Likewise.
15221 (IS_FP_VECT_REG_TYPE): Likewise.
15222 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15223 use direct move instructions.
15224 (reload_gpr_vsx): Likewise.
15225 (reload_vsx_gpr): Likewise.
15226 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15227 information that is a simplification of register classes. Also
15228 precalculate direct move reload helpers.
15229 (direct_move_p): New function to return true if the operation can
15230 be done as a direct move instruciton.
15231 (quad_load_store_p): New function to return true if the operation
15232 is a quad memory operation.
15233 (rs6000_legitimize_address): If quad memory, only allow register
15234 indirect for TImode addresses.
15235 (rs6000_legitimate_address_p): Likewise.
15236 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15237 (rs6000_reload_register_type): Likewise.
15238 (register_to_reg_type): Return register type.
15239 (rs6000_secondary_reload_simple_move): New helper function for
15240 secondary reload and secondary memory needed to identify anything
15241 that is a simple move, and does not need reloading.
15242 (rs6000_secondary_reload_direct_move): New helper function for
15243 secondary reload to identify cases that can be done with several
15244 instructions via the direct move instructions.
15245 (rs6000_secondary_reload_move): New helper function for secondary
15246 reload to identify moves between register types that can be done.
15247 (rs6000_secondary_reload): Add support for quad memory operations
15248 and for direct move.
15249 (rs6000_secondary_memory_needed): Likewise.
15250 (rs6000_debug_secondary_memory_needed): Change argument names.
15251 (rs6000_output_move_128bit): New function to return the move to
15252 use for 128-bit moves, including knowing about the various
15253 limitations of quad memory operations.
15255 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15256 memory operations. call rs6000_output_move_128bit for the actual
15257 instruciton(s) to generate.
15258 (vsx_movti_64bit): Likewise.
15260 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15261 (UNSPEC_P8V_MTVSRWZ): Likewise.
15262 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15263 (UNSPEC_P8V_MTVSRD): Likewise.
15264 (UNSPEC_P8V_XXPERMDI): Likewise.
15265 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15266 (UNSPEC_FUSION_GPR): Likewise.
15267 (FMOVE128_GPR): New iterator for direct move.
15268 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15269 (f32_sv): Likewise.
15270 (f32_dm): Likewise.
15271 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15272 loads and direct move instructions.
15273 (zero_extendsidi2_lfiwzx): Likewise.
15274 (extendsidi2_lfiwax): Likewise.
15275 (extendsidi2_nocell): Likewise.
15276 (floatsi<mode>2_lfiwax): Likewise.
15277 (lfiwax): Likewise.
15278 (floatunssi<mode>2_lfiwzx): Likewise.
15279 (lfiwzx): Likewise.
15280 (fix_trunc<mode>_stfiwx): Likewise.
15281 (fixuns_trunc<mode>_stfiwx): Likewise.
15282 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15283 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15284 (parity<mode>2_cmpb): Set length/type attr.
15285 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15286 for 'mr.' to fast_compare.
15287 (bpermd_<mode>): Change type attr to popcnt.
15288 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15289 (p8_mtvsrwz_1): Likewise.
15290 (p8_mtvsrwz_2): Likewise.
15291 (reload_fpr_from_gpr<mode>): Likewise.
15292 (p8_mtvsrd_1): Likewise.
15293 (p8_mtvsrd_2): Likewise.
15294 (p8_xxpermdi_<mode>): Likewise.
15295 (reload_vsx_from_gpr<mode>): Likewise.
15296 (reload_vsx_from_gprsf): Likewise.
15297 (p8_mfvsrd_3_<mode>): LIkewise.
15298 (reload_gpr_from_vsx<mode>): Likewise.
15299 (reload_gpr_from_vsxsf): Likewise.
15300 (p8_mfvsrd_4_disf): Likewise.
15301 (multi-word GPR splits): Do not split direct moves or quad memory
15304 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15306 * tree-into-ssa.c (interesting_blocks): Make static.
15308 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15311 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15312 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15313 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15314 (get_variable_align): New function.
15315 (get_variable_section, emit_bss, emit_common,
15316 assemble_variable_contents, place_block_symbol): Use
15317 get_variable_align instead of DECL_ALIGN.
15318 (assemble_noswitch_variable): Add align argument, use it
15319 instead of DECL_ALIGN.
15320 (assemble_variable): Adjust caller. Use get_variable_align
15321 instead of DECL_ALIGN.
15322 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15324 (DATA_ABI_ALIGNMENT): Define.
15325 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15326 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15327 opt is false, only return the psABI mandated alignment increase.
15328 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15329 (DATA_ABI_ALIGNMENT): ... this.
15330 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15331 (DATA_ABI_ALIGNMENT): ... this.
15332 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15333 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15334 (DATA_ABI_ALIGNMENT): ... this.
15335 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15336 * doc/tm.texi: Regenerated.
15338 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15340 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15341 cmp_code to construct REG_EQUAL note.
15343 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15346 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15347 that operands[2] doesn't overlap with operands[0].
15349 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15350 Jan Hubicka <jh@suse.cz>
15352 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15353 hack to mark symbols as used.
15355 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15357 PR rtl-optimization/57559
15358 * lra-constraints.c (process_alt_operands): Don't discourage
15359 memory with known offset for offsetable memory constraint.
15360 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15362 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15364 * varasm.c (struct oc_local_state): Reorder fields.
15365 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15366 and adjust accordingly.
15367 (output_constructor): Reorder initialization code and adjust call to
15368 output_constructor_bitfield.
15370 2013-06-07 Jan Hubicka <jh@suse.cz>
15372 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15374 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15376 * tree-object-size.c (unknown): Make const.
15378 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15380 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15381 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15382 for last alternative in the cpu_facility attribute.
15384 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15387 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15388 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15389 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15390 * config/arm/constraints.md (Dg): New constraint.
15391 * config/arm/neon.md (xordi3_neon): Remove.
15392 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15393 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15395 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15397 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15398 Clean up alternatives.
15400 2013-06-07 Alan Modra <amodra@gmail.com>
15402 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15405 2013-06-07 Alan Modra <amodra@gmail.com>
15407 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15409 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15411 * config/arm/constraints.md (Df): New constraint.
15412 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15413 Correct length attribute for last two alternatives.
15415 2013-06-07 Alan Modra <amodra@gmail.com>
15417 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15418 override user -mfp-in-toc.
15419 (offsettable_ok_by_alignment): Consider just the current access
15420 rather than the whole object, unless BLKmode. Handle
15421 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15422 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15423 for -mcmodel=medium.
15424 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15425 override user -mfp-in-toc or -msum-in-toc. Default to
15426 -mno-fp-in-toc for -mcmodel=medium.
15428 2013-06-06 DJ Delorie <dj@redhat.com>
15430 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15431 TARGET_VALID_POINTER_MODE.
15433 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15434 Pat Haugen <pthaugen@us.ibm.com>
15435 Peter Bergner <bergner@vnet.ibm.com>
15437 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15438 Document new power8 builtins.
15440 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15441 condition code register, to allow 128-bit logical operations to be
15442 done in the VSX or GPR registers.
15443 (nor<mode>3): Use the canonical form for nor.
15444 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15445 vclz*, and vpopcnt* vector instructions.
15446 (nand<mode>3): Likewise.
15447 (orc<mode>3): Likewise.
15448 (clz<mode>2): LIkewise.
15449 (popcount<mode>2): Likewise.
15451 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15452 that only the GPRs are recognized.
15454 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15455 support for new power8 builtins.
15457 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15459 (xscvdpspn): Likewise.
15465 (vpopcnt): Likewise.
15466 (vpopcntb): Likewise.
15467 (vpopcnth): Likewise.
15468 (vpopcntw): Likewise.
15469 (vpopcntd): Likewise.
15471 (vmrgew): Likewise.
15472 (vmrgow): Likewise.
15474 (eqv_v16qi3): Likewise.
15475 (eqv_v8hi3): Likewise.
15476 (eqv_v4si3): Likewise.
15477 (eqv_v2di3): Likewise.
15478 (eqv_v4sf3): Likewise.
15479 (eqv_v2df3): Likewise.
15481 (nand_v16qi3): Likewise.
15482 (nand_v8hi3): Likewise.
15483 (nand_v4si3): Likewise.
15484 (nand_v2di3): Likewise.
15485 (nand_v4sf3): Likewise.
15486 (nand_v2df3): Likewise.
15488 (orc_v16qi3): Likewise.
15489 (orc_v8hi3): Likewise.
15490 (orc_v4si3): Likewise.
15491 (orc_v2di3): Likewise.
15492 (orc_v4sf3): Likewise.
15493 (orc_v2df3): Likewise.
15495 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15496 allow power8 quad mode in 64-bit.
15497 (rs6000_builtin_vectorized_function): Add support to vectorize
15498 ISA 2.07 count leading zeros, population count builtins.
15499 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15500 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15501 (builtin_function_type): Add vgbbd builtin function which takes an
15503 (altivec_expand_vec_perm_const): Add support for new power8 merge
15506 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15507 that does not include TImdoe for use with 32-bit.
15508 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15510 (UNSPEC_VSX_CVDPSPN): Likewise.
15511 (vsx_xscvdpspn): Likewise.
15512 (vsx_xscvspdpn): Likewise.
15513 (vsx_xscvdpspn_scalar): Likewise.
15514 (vsx_xscvspdpn_directmove): Likewise.
15515 (vsx_and<mode>3): Split logical operations into 32-bit and
15516 64-bit. Add support to do logical operations on TImode as well as
15517 VSX vector types. Allow logical operations to be done in either
15518 VSX registers or in general purpose registers in 64-bit mode. Add
15519 splitters if GPRs were used. For AND, add clobber of CCmode to
15520 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15522 (vsx_and<mode>3_32bit): Likewise.
15523 (vsx_and<mode>3_64bit): Likewise.
15524 (vsx_ior<mode>3): Likewise.
15525 (vsx_ior<mode>3_32bit): Likewise.
15526 (vsx_ior<mode>3_64bit): Likewise.
15527 (vsx_xor<mode>3): Likewise.
15528 (vsx_xor<mode>3_32bit): Likewise.
15529 (vsx_xor<mode>3_64bit): Likewise.
15530 (vsx_one_cmpl<mode>2): Likewise.
15531 (vsx_one_cmpl<mode>2_32bit): Likewise.
15532 (vsx_one_cmpl<mode>2_64bit): Likewise.
15533 (vsx_nor<mode>3): Likewise.
15534 (vsx_nor<mode>3_32bit): Likewise.
15535 (vsx_nor<mode>3_64bit): Likewise.
15536 (vsx_andc<mode>3): Likewise.
15537 (vsx_andc<mode>3_32bit): Likewise.
15538 (vsx_andc<mode>3_64bit): Likewise.
15539 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15540 and xxlorc instructions.
15541 (vsx_eqv<mode>3_64bit): Likewise.
15542 (vsx_nand<mode>3_32bit): Likewise.
15543 (vsx_nand<mode>3_64bit): Likewise.
15544 (vsx_orc<mode>3_32bit): Likewise.
15545 (vsx_orc<mode>3_64bit): Likewise.
15547 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15549 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15551 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15552 (p8_vmrgow): Likewise.
15553 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15554 GPRs to be split under VSX.
15555 (p8v_clz<mode>2): Add power8 count leading zero support.
15556 (p8v_popcount<mode>2): Add power8 population count support.
15557 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15560 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15563 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15565 (vec_nand): Likewise.
15566 (vec_vclz): Likewise.
15567 (vec_vclzb): Likewise.
15568 (vec_vclzd): Likewise.
15569 (vec_vclzh): Likewise.
15570 (vec_vclzw): Likewise.
15571 (vec_vgbbd): Likewise.
15572 (vec_vmrgew): Likewise.
15573 (vec_vmrgow): Likewise.
15574 (vec_vpopcnt): Likewise.
15575 (vec_vpopcntb): Likewise.
15576 (vec_vpopcntd): Likewise.
15577 (vec_vpopcnth): Likewise.
15578 (vec_vpopcntw): Likewise.
15580 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15582 PR rtl-optimization/57468
15583 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15586 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15588 PR rtl-optimization/57459
15589 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15590 type when setting live regs.
15592 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15594 * config/s390/s390.opt (mlra): New option.
15595 * config/s390/s390.c (s390_decompose_address): Check displacement
15596 for all registers for LRA.
15597 (s390_secondary_reload): Don't used secondary reloads for LRA.
15598 (s390_lra_p): New function.
15599 (TARGET_LRA_P): Define.
15600 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15601 of attribute cpu_facility to zarch for the last alternative.
15602 (*cmpmem_short): Ditto.
15604 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15606 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15607 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15608 (arm_expand_prologue): Likewise.
15610 2013-06-06 Teresa Johnson <tejohnson@google.com>
15613 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15614 as this is now done by redirect_edge_and_branch_force.
15615 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15616 barriers, and fix interaction with splitting.
15617 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15618 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15619 reflect changes made in the routine.
15620 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15621 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15622 since this is called in cfglayout mode, and replace partition fixup
15623 with assert as that is now done by force_nonfallthru_and_redirect.
15624 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15625 already be marked with region crossing note.
15626 (insert_section_boundary_note): Make non-static, gate on flag
15627 has_bb_partition, rewrite to also check for multiple partitions.
15628 (rest_of_handle_reorder_blocks): Remove call to
15629 insert_section_boundary_note, now done later during free_cfg.
15630 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15631 * bb-reorder.h (insert_section_boundary_note): Declare.
15632 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15633 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15634 invoke insert_section_boundary_note.
15635 (try_redirect_by_replacing_jump): Remove unnecessary
15636 check for region crossing note.
15637 (fixup_partition_crossing): New function.
15638 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15639 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15640 in non-cfglayout mode.
15641 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15642 remove old code that tried to do this. Emit barrier correctly
15643 when we are in cfglayout mode.
15644 (last_bb_in_partition): New function.
15645 (rtl_split_edge): Correctly fixup partition boundaries.
15646 (commit_one_edge_insertion): Remove old code that tried to
15647 fixup region crossing edge since this is now handled in
15648 split_block, and set up insertion point correctly since
15649 block may now end in a jump.
15650 (verify_hot_cold_block_grouping): Guard against checking when not in
15651 linearized RTL mode.
15652 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15654 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15655 rtl_verify_flow_info, so not called in cfglayout mode.
15656 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15657 (fixup_reorder_chain): Remove old code that attempted to fixup region
15658 crossing note as this is now handled in force_nonfallthru_and_redirect.
15659 (duplicate_insn_chain): Don't duplicate switch section notes.
15660 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15662 * basic-block.h (emit_barrier_after_bb): Declare.
15664 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15666 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15667 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15668 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15670 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15672 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15674 * config/arm/ldmstm.md: Regenerate.
15676 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15678 * config/arm/sync.md (atomic_loaddi_1):
15679 Disable predication for arm_restrict_it.
15680 (arm_load_exclusive<mode>): Likewise.
15681 (arm_load_exclusivesi): Likewise.
15682 (arm_load_exclusivedi): Likewise.
15683 (arm_load_acquire_exclusive<mode>): Likewise.
15684 (arm_load_acquire_exclusivesi): Likewise.
15685 (arm_load_acquire_exclusivedi): Likewise.
15686 (arm_store_exclusive<mode>): Likewise.
15687 (arm_store_exclusive<mode>): Likewise.
15688 (arm_store_release_exclusivedi): Likewise.
15689 (arm_store_release_exclusive<mode>): Likewise.
15691 2013-06-06 Richard Biener <rguenther@suse.de>
15693 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15695 (lto_tag_is_tree_code_p): Adjust.
15696 (lto_tag_is_gimple_code_p): Likewise.
15697 (lto_gimple_code_to_tag): Likewise.
15698 (lto_tag_to_gimple_code): Likewise.
15699 (lto_tree_code_to_tag): Likewise.
15700 (lto_tag_to_tree_code): Likewise.
15701 * data-streamer.h (streamer_write_hwi_in_range): Use
15702 uhwi streaming to stream the normalized range.
15703 (streamer_read_hwi_in_range): Likewise.
15705 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15707 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15708 (predicable_short_it): Likewise.
15709 (predicated): Likewise.
15710 (enabled): Handle above.
15711 (define_cond_exec): Set predicated attribute to yes.
15713 2013-06-05 Mike Stump <mikestump@comcast.net>
15715 * gdbinit.in (__FUNCTION__): Add.
15717 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15719 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15720 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15722 2013-06-05 Jan Hubicka <jh@suse.cz>
15724 * varasm.c (mark_decl_referenced): Revert the removal until targets
15727 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15729 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15730 instead of mark_decl_referenced.
15732 2013-06-05 Jan Hubicka <jh@suse.cz>
15734 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15735 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15736 and symtab_used_from_object_file_p.
15737 (cgraph_make_node_local_1): Clear forced_by_abi.
15738 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15739 * cgraph.h (symtab_node_base): Add forced_by_abi.
15740 (decide_is_variable_needed): Remove.
15741 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15742 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15743 (decide_is_symbol_needed): ... this one; handle symbols in general;
15744 always analyze virtuals; honnor forced_by_abi.
15745 (cgraph_finalize_function): Update.
15746 (varpool_finalize_decl): Update.
15747 (symbol_defined_and_needed): Remove.
15748 (analyze_functions): Update.
15749 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15750 output_refs, input_overwrite_node): Handle forced_by_abi.
15751 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15752 (address_taken_from_non_vtable_p): ... this one.
15753 (comdat_can_be_unshared_p_1): New function.
15754 (cgraph_comdat_can_be_unshared_p): Rename to ...
15755 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15756 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15757 (function_and_variable_visibility): Clear forced_by_abi as needed.
15758 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15759 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15760 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15761 * varpool.c (decide_is_variable_needed): Remove.
15763 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15765 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15766 (arm_option_override): Override arm_restrict_it where appropriate.
15767 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15768 * config/arm/arm.opt (mrestrict-it): New command-line option.
15769 * doc/invoke.texi: Document -mrestrict-it.
15771 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15773 * tsan.c (tsan_atomic_table): Make const.
15775 2013-06-05 Richard Biener <rguenther@suse.de>
15777 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15778 index associated with the tree we are supposed to replace.
15779 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15780 TREE_ASM_WRITTEN as zero for everything but SSA names.
15782 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15784 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15786 2013-06-04 Jan Hubicka <jh@suse.cz>
15788 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15789 (input_node, input_varpool_node): Handle correctly external same
15791 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15792 nodes at ltrans stage.
15794 2013-06-04 Jan Hubicka <jh@suse.cz>
15796 * ipa-inline.c (update_caller_keys): Fix availability test.
15797 (update_callee_keys): Likewise.
15798 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15799 to follow ELF standard.
15801 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15803 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15804 (mips64r5900el-*-elf*): New configurations.
15805 * config/mips/mips-cpus.def (r5900): New processor.
15806 * config/mips/mips-tables.opt: Regenerate.
15807 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15808 (mips_issue_rate): Handle PROCESSOR_R5900.
15809 (mips_reorg_process_insns): Force reorder mode for the R5900.
15810 * config/mips/mips.h (TARGET_MIPS5900): Define.
15811 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15813 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15815 * config/mips/mips.md (processor): Add r5900.
15816 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15818 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15820 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15821 into function to generate MOVI instruction.
15822 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15823 (aarch64_preferred_simd_mode): Turn into wrapper.
15824 (aarch64_output_scalar_simd_mov_immediate): New function.
15825 * config/aarch64/aarch64-protos.h: Add prototype for above.
15827 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15829 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15830 element_char member.
15831 (sizetochar): Return signed char.
15832 (aarch64_simd_valid_immediate): Remove elchar and other
15833 unnecessary variables.
15834 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15835 Calculate element_char as required.
15836 * config/aarch64/aarch64-protos.h: Update and move prototype
15837 for aarch64_output_simd_mov_immediate.
15838 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
15841 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15843 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
15844 information completed by aarch64_simd_valid_immediate.
15845 (aarch64_legitimate_constant_p): Update arguments.
15846 (aarch64_simd_valid_immediate): Work with struct rather than many
15848 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
15849 (aarch64_simd_make_constant): Update arguments.
15850 (aarch64_output_simd_mov_immediate): Work with struct rather than
15851 many pointers. Output immediate directly rather than as operand.
15852 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15854 * config/aarch64/constraints.md (Dn): Update arguments.
15856 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15858 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
15860 (aarch64_simd_immediate_valid_for_move): Remove.
15861 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
15862 (aarch64_simd_make_constant): Update call.
15863 (aarch64_output_simd_mov_immediate): Update call.
15864 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
15866 * config/aarch64/constraints.md (Dn): Update call.
15868 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15870 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
15871 return type to bool for prototype.
15872 (aarch64_legitimate_constant_p): Check for true instead of not -1.
15873 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
15874 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
15876 2013-06-04 Catherine Moore <clm@codesourcery.com>
15878 * config/mips/mips.opt (meva): New.
15879 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
15880 (ASM_SPEC): Handle -meva.
15881 * doc/invoke.texi (meva): Document.
15883 2013-06-04 Alan Modra <amodra@gmail.com>
15885 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
15888 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15890 * rtl.def: Add extra fourth optional field to define_cond_exec.
15891 * gensupport.c (process_one_cond_exec): Process attributes from
15893 * doc/md.texi: Document fourth field in define_cond_exec.
15895 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
15897 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
15898 out the processing order as in store_bit_field_1.
15900 2013-06-04 Jan Hubicka <jh@suse.cz>
15902 PR middle-end/57500
15903 * cgraphunit.c (cgraph_process_same_body_aliases): Create
15904 non-VAR_DECL node if it does not exist yet.
15906 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
15908 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
15909 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
15910 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
15911 target_cpu_default setting.
15913 2013-06-03 Teresa Johnson <tejohnson@google.com>
15915 * dumpfile.c (opt_info_switch_p): Change -fopt-info
15916 default to -fopt-info=optimized instead of all.
15917 * doc/invoke.texi: Ditto.
15918 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
15919 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
15920 (execute_vect_slp): Emit BB vectorization success under
15921 MSG_OPTIMIZED_LOCATIONS.
15922 * tree-vect-slp.c (vect_slp_transform_bb): Change
15923 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
15924 * tree-vect-loop.c (vect_transform_loop): Ditto.
15926 2013-06-03 Jason Merrill <jason@redhat.com>
15929 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
15930 Use TARGET_EXPR for C++.
15932 2013-06-03 Jakub Jelinek <jakub@redhat.com>
15934 PR rtl-optimization/57268
15935 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
15936 if DEBUG_INSN_P (insn).
15939 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
15941 PR rtl-optimization/57268
15942 * sched-deps.c (sched_analyze_2): Flush dependence lists if
15943 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
15945 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15947 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
15948 (ix86_avoid_lea_for_addr): Likewise.
15949 (exact_dependency_1): Likewise.
15950 (ix86_adjust_cost): Likewise.
15951 (swap_top_of_ready_list): Fix formatting and !reload_completed check
15953 (do_reorder_for_imul): Fix typo, formatting and
15954 !reload_completed check removed.
15955 (ix86_sched_reorder): Fix typo and formatting.
15956 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
15959 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
15961 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
15963 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
15965 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
15966 <STRING_CST>: Likewise.
15967 <VECTOR_CST>: Likewise.
15969 2013-06-01 Janus Weil <janus@gcc.gnu.org>
15970 Mikael Morin <mikael@gcc.gnu.org>
15972 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
15973 * config.in: Regenerated.
15974 * configure: Regenerated.
15976 2013-06-01 Jan Hubicka <jh@suse.cz>
15978 PR middle-end/57366
15979 * cgraphunit.c (compile): When weakref is not supported,
15980 set up transparent aliases before final output pass.
15981 * varasm.c (assemble_alias): Do not try to do it here.
15983 2013-06-01 Jan Hubicka <jh@suse.cz>
15985 PR middle-end/57467
15986 * passes.c (for_per_function): Skip unanalyzed functions.
15988 2013-06-01 Jan Hubicka <jh@suse.cz>
15990 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
15991 (lto_symtab_merge_symbols_1): ... this one.
15992 (lto_symtab_merge_cgraph_nodes): Rename to ...
15993 (lto_symtab_merge_symbols): ... this one; simplify.
15994 * cgraph.c (same_body_aliases_done): Rename to ...
15995 (cpp_implicit_aliases_done): ... this one.
15996 (cgraph_create_function_alias): Update.
15997 (cgraph_same_body_alias): Update.
15998 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
15999 (verify_edge_corresponds_to_fndecl): Simplify.
16000 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
16001 (cgraph_node): Remove same_body_alias.
16002 (varpool_node): Remove alias_of and extra_name_alias.
16003 (same_body_aliases_done): Rename to ..
16004 (cpp_implicit_aliases_done): ... this one.
16005 (symtab_alias_ultimate_target): Add default parameter.
16006 (symtab_resolve_alias): New function.
16007 (fixup_same_cpp_alias_visibility): Declare.
16008 (cgraph_function_node): Add default parameter.
16009 (cgraph_node_asm_name): Likewise.
16010 (cgraph_function_or_thunk_node): Add default parameter; do
16011 not ICE when it is NULL.
16012 (varpool_variable_node): Likewise.
16013 * tree-emutls.c (create_emultls_var): Update.
16014 (ipa_lower_emutls): Update.
16015 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
16016 (cgraph_reset_node): Reset alias info.
16017 (cgraph_finalize_function): Update.
16018 (fixup_same_cpp_alias_visibility): Move to symtab.c.
16019 (analyze_function): Simplify.
16020 (cgraph_process_same_body_aliases): Simplify.
16021 (analyze_functions): Fixup same body aliases.
16022 (handle_alias_pairs): Simplify.
16023 (assemble_thunk): Update.
16024 (assemble_thunks_and_aliases): Update.
16025 (output_weakrefs): Rewrite.
16026 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
16027 (lto_output_varpool_node): Likewise.
16028 (compute_ltrans_boundary): Remve assert.
16029 (get_alias_symbol): New functoin.
16030 (input_node): Rewrite alias handling.
16031 (input_varpool_node): Likewise.
16032 * ipa-pure-const.c (propagate_pure_const): Fix formating.
16033 * ipa.c (process_references): Handle weakrefs correctly.
16034 (symtab_remove_unreachable_nodes): Likewise.
16035 * trans-mem.c (get_cg_data): Update.
16036 (ipa_tm_create_version_alias): Update.
16037 (ipa_tm_execute): Update.
16038 * symtab.c (dump_symtab_base): Dump aliases.
16039 (verify_symtab_base): Verify aliases.
16040 (symtab_node_availability): New function.
16041 (symtab_alias_ultimate_target): Simplify.
16042 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
16043 handle all the fixup cases.
16044 (symtab_resolve_alias): New function.
16045 * passes.c (ipa_write_summaries): Handle weakrefs.
16046 * varpool.c (varpool_analyze_node): Simplify.
16047 (assemble_aliases): Update.
16048 (varpool_create_variable_alias): Simplify.
16049 (varpool_extra_name_alias): Simplify.
16050 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
16051 (lto_symtab_merge_symbols): ... this one.
16053 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
16056 PR rtl-optimization/57268
16057 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16058 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16060 2013-06-01 Tobias Burnus <burnus@net-b.de>
16062 Partially reverted:
16063 2013-05-31 Tobias Burnus <burnus@net-b.de>
16065 PR middle-end/57073
16066 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
16069 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16071 PR rtl-optimization/57268
16072 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16073 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16075 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16077 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
16078 unordered comparison operators when -fno-trapping-math is in effect
16080 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
16081 and implement unordered comparison operators properly on the e500.
16083 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16085 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
16086 for constant scalar integers.
16087 (simplify_relational_operation_1): Likewise.
16089 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
16091 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
16092 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
16095 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16096 Igor Zamyatin <igor.zamyatin@intel.com>
16098 Silvermont (SLM) architecture performance tuning.
16099 * config/i386/i386.h (enum ix86_tune_indices): Add
16100 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
16101 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
16103 * config/i386/i386.c (initial_ix86_tune_features)
16104 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
16105 (ix86_lea_outperforms): Handle Silvermont tuning.
16106 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
16108 (ix86_use_lea_for_mov): Likewise.
16109 (ix86_avoid_lea_for_addr): Likewise.
16110 (ix86_lea_for_add_ok): Likewise.
16111 (exact_dependency_1): New function.
16112 (exact_store_load_dependency): Likewise.
16113 (ix86_adjust_cost): Handle Silvermont tuning.
16114 (do_reoder_for_imul): Likewise.
16115 (swap_top_of_ready_list): New function.
16116 (ix86_sched_reorder): Changed to handle Silvermont tuning.
16118 * config/i386/i386.md (peepholes that split memory operand in fp
16121 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16123 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
16124 Remove un-necessary braces.
16126 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16128 * config/aarch64/aarch64.c (aarch64_classify_symbol):
16129 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
16131 2013-05-31 Tobias Burnus <burnus@net-b.de>
16133 PR middle-end/57073
16134 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
16136 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16139 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
16140 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
16141 * config/arm/neon.md (iordi3_neon): Remove.
16142 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
16143 * config/arm/predicates.md (imm_for_neon_logic_operand):
16144 Move to earlier in the file.
16145 (neon_logic_op2): Likewise.
16146 (arm_iordi_operand_neon): New predicate.
16148 2013-05-31 Richard Biener <rguenther@suse.de>
16150 PR tree-optimization/57478
16151 PR tree-optimization/57453
16152 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
16155 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16157 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
16158 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
16160 2013-05-30 Tobias Burnus <burnus@net-b.de>
16161 Thomas Koenig <tkoenig@gcc.gnu.org>
16163 PR middle-end/57073
16164 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
16165 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
16167 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
16169 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
16171 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
16173 * target.def (register_usage_leveling_p): New hook.
16174 * targhooks.c (default_register_usage_leveling_p): New.
16175 * targhooks.h (default_register_usage_leveling_p): New prototype.
16176 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16177 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16178 * doc/tm.texi: Update.
16179 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16181 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16183 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16184 (*insv_reg<mode>): New define_insn.
16186 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16188 PR rtl-optimization/57439
16189 * postreload.c (move2add_valid_value_p): Check that we have
16190 a zero subreg_regno_offset when accessing the register in
16191 the requested mode.
16193 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16194 Igor Zamyatin <igor.zamyatin@intel.com>
16196 Silvermont (SLM) architecture pipeline model, tuning and
16198 * config.gcc: Add slm config options and target.
16200 * config/i386/slm.md: New.
16202 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16204 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16206 (ix86_target_macros_internal): Likewise.
16208 * config/i386/i386.c (slm_cost): New cost.
16209 (m_SLM): New macro flag.
16210 (initial_ix86_tune_features): Set m_SLM.
16211 (x86_accumulate_outgoing_args): Likewise.
16212 (x86_arch_always_fancy_math_387): Likewise.
16213 (processor_target_table): Add slm cost.
16214 (cpu_names): Add slm cpu name.
16215 (x86_option_override_internal): Set SLM ISA.
16216 (ix86_issue_rate): New case PROCESSOR_SLM.
16217 (ia32_multipass_dfa_lookahead): Likewise.
16218 (fold_builtin_cpu): Add slm.
16220 * config/i386/i386.h (TARGET_SLM): New target macro.
16221 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16222 (processor_type): Add PROCESSOR_SLM.
16224 * config/i386/i386.md (cpu): Add new value "slm".
16225 (slm.md): Include slm.md.
16227 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16228 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16230 * config/arm/arm-protos.h: Add and update function protos.
16231 * config/arm/arm.c (use_simple_return_p): New added.
16232 (thumb2_expand_return): Check simple_return flag.
16233 * config/arm/arm.md: Add simple_return and conditional simple_return.
16234 * config/arm/iterators.md: Add iterator for return and simple_return.
16236 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16238 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16239 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16240 (arm_emit_vfp_multi_reg_pop): Likewise.
16241 (thumb2_emit_ldrd_pop): Likewise.
16242 (arm_expand_epilogue): Add misc REG_CFA notes.
16243 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16245 2013-05-29 Lawrence Crowl <crowl@google.com>
16247 * config/arm/t-arm: Update for below.
16249 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16250 Change type to hash_table. Update dependent calls and types.
16252 * config/i386/t-cygming: Update for below.
16254 * config/i386/t-interix: Update for below.
16256 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16257 Change type to hash_table. Update dependent calls and types.
16258 (i386_find_on_wrapper_list::wrappers): Likewise.
16260 * config/ia64/t-ia64: Update for below.
16262 * config/ia64/ia64.c (bundle_state_table):
16263 Change type to hash_table. Update dependent calls and types.
16265 * config/mips/mips.c (mips_reorg_process_insns::htab):
16266 Change type to hash_table. Update dependent calls and types.
16268 * config/sol2.c (solaris_comdat_htab):
16269 Change type to hash_table. Update dependent calls and types.
16271 * config/t-sol2: Update for above.
16273 2013-05-29 Teresa Johnson <tejohnson@google.com>
16275 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16276 functions are not yet marked as defined.
16278 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16279 Pat Haugen <pthaugen@us.ibm.com>
16280 Peter Bergner <bergner@vnet.ibm.com>
16282 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16286 (vrotl<mode>3): Likewise.
16287 (vashl<mode>3): Likewise.
16288 (vlshr<mode>3): Likewise.
16289 (vashr<mode>3): Likewise.
16291 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16292 support for power8 V2DI builtins.
16294 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16295 power8 V2DI builtins.
16296 (vupkhsw): Likewise.
16297 (vupklsw): Likewise.
16298 (vaddudm): Likewise.
16299 (vminsd): Likewise.
16300 (vmaxsd): Likewise.
16301 (vminud): Likewise.
16302 (vmaxud): Likewise.
16303 (vpkudum): Likewise.
16304 (vpksdss): Likewise.
16305 (vpkudus): Likewise.
16306 (vpksdus): Likewise.
16311 (vsubudm): Likewise.
16312 (vcmpequd): Likewise.
16313 (vcmpgtsd): Likewise.
16314 (vcmpgtud): Likewise.
16315 (vcmpequd_p): Likewise.
16316 (vcmpgtsd_p): Likewise.
16317 (vcmpgtud_p): Likewise.
16318 (vupkhsw): Likewise.
16319 (vupklsw): Likewise.
16320 (vaddudm): Likewise.
16321 (vmaxsd): Likewise.
16322 (vmaxud): Likewise.
16323 (vminsd): Likewise.
16324 (vminud): Likewise.
16325 (vpksdss): Likewise.
16326 (vpksdus): Likewise.
16327 (vpkudum): Likewise.
16328 (vpkudus): Likewise.
16333 (vsubudm): Likewise.
16335 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16336 support for power8 V2DI instructions.
16338 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16339 power8 V2DI instructions. Combine pack and unpack insns to use an
16340 iterator for each mode. Check whether a particular mode supports
16341 Altivec instructions instead of just checking TARGET_ALTIVEC.
16342 (UNSPEC_VPKUWUM): Likewise.
16343 (UNSPEC_VPKSHSS): Likewise.
16344 (UNSPEC_VPKSWSS): Likewise.
16345 (UNSPEC_VPKUHUS): Likewise.
16346 (UNSPEC_VPKSHUS): Likewise.
16347 (UNSPEC_VPKUWUS): Likewise.
16348 (UNSPEC_VPKSWUS): Likewise.
16349 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16350 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16351 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16352 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16353 (UNSPEC_VUPKHSB): Likewise.
16354 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16355 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16356 (UNSPEC_VUPKHSH): Likewise.
16357 (UNSPEC_VUPKLSB): Likewise.
16358 (UNSPEC_VUPKLSH): Likewise.
16360 (VI_char): Likewise.
16361 (VI_scalar): Likewise.
16362 (VI_unit): Likewise.
16364 (VP_small): Likewise.
16365 (VP_small_lc): Likewise.
16366 (VU_char): Likewise.
16367 (add<mode>3): Likewise.
16368 (altivec_vaddcuw): Likewise.
16369 (altivec_vaddu<VI_char>s): Likewise.
16370 (altivec_vadds<VI_char>s): Likewise.
16371 (sub<mode>3): Likewise.
16372 (altivec_vsubcuw): Likewise.
16373 (altivec_vsubu<VI_char>s): Likewise.
16374 (altivec_vsubs<VI_char>s): Likewise.
16375 (altivec_vavgs<VI_char>): Likewise.
16376 (altivec_vcmpbfp): Likewise.
16377 (altivec_eq<mode>): Likewise.
16378 (altivec_gt<mode>): Likewise.
16379 (altivec_gtu<mode>): Likewise.
16380 (umax<mode>3): Likewise.
16381 (smax<mode>3): Likewise.
16382 (umin<mode>3): Likewise.
16383 (smin<mode>3): Likewise.
16384 (altivec_vpkuhum): Likewise.
16385 (altivec_vpkuwum): Likewise.
16386 (altivec_vpkshss): Likewise.
16387 (altivec_vpkswss): Likewise.
16388 (altivec_vpkuhus): Likewise.
16389 (altivec_vpkshus): Likewise.
16390 (altivec_vpkuwus): Likewise.
16391 (altivec_vpkswus): Likewise.
16392 (altivec_vpks<VI_char>ss): Likewise.
16393 (altivec_vpks<VI_char>us): Likewise.
16394 (altivec_vpku<VI_char>us): Likewise.
16395 (altivec_vpku<VI_char>um): Likewise.
16396 (altivec_vrl<VI_char>): Likewise.
16397 (altivec_vsl<VI_char>): Likewise.
16398 (altivec_vsr<VI_char>): Likewise.
16399 (altivec_vsra<VI_char>): Likewise.
16400 (altivec_vsldoi_<mode>): Likewise.
16401 (altivec_vupkhsb): Likewise.
16402 (altivec_vupkhs<VU_char>): Likewise.
16403 (altivec_vupkls<VU_char>): Likewise.
16404 (altivec_vupkhsh): Likewise.
16405 (altivec_vupklsb): Likewise.
16406 (altivec_vupklsh): Likewise.
16407 (altivec_vcmpequ<VI_char>_p): Likewise.
16408 (altivec_vcmpgts<VI_char>_p): Likewise.
16409 (altivec_vcmpgtu<VI_char>_p): Likewise.
16410 (abs<mode>2): Likewise.
16411 (vec_unpacks_hi_v16qi): Likewise.
16412 (vec_unpacks_hi_v8hi): Likewise.
16413 (vec_unpacks_lo_v16qi): Likewise.
16414 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16415 (vec_unpacks_lo_v8hi): Likewise.
16416 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16417 (vec_pack_trunc_v8h): Likewise.
16418 (vec_pack_trunc_v4si): Likewise.
16419 (vec_pack_trunc_<mode>): Likewise.
16421 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16423 (vec_vmaxsd): Likewise.
16424 (vec_vmaxud): Likewise.
16425 (vec_vminsd): Likewise.
16426 (vec_vminud): Likewise.
16427 (vec_vpksdss): Likewise.
16428 (vec_vpksdus): Likewise.
16429 (vec_vpkudum): Likewise.
16430 (vec_vpkudus): Likewise.
16431 (vec_vrld): Likewise.
16432 (vec_vsld): Likewise.
16433 (vec_vsrad): Likewise.
16434 (vec_vsrd): Likewise.
16435 (vec_vsubudm): Likewise.
16436 (vec_vupkhsw): Likewise.
16437 (vec_vupklsw): Likewise.
16439 2013-05-29 Jan Hubicka <jh@suse.cz>
16441 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16442 flags; reorder rest of fields in more consistent way.
16443 (varpool_node): Remove analyzed, finalized and alias.
16444 (cgraph_ndoe): Likewise.
16445 (symtab_alias_ultimate_target): New function.
16446 (cgraph_function_node): Move offline.
16447 (cgraph_reset_node): Declare.
16448 (cgraph_comdat_can_be_unshared_p): Remove.
16449 (varpool_remove_initializer): Declare.
16450 (varpool_first_defined_variable, varpool_next_defined_variable
16451 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16452 (cgraph_function_with_gimple_body_p): Update.
16453 (varpool_all_refs_explicit_p): Update.
16454 (symtab_alias_target): New function.
16455 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16456 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16457 (cgraph_function_or_thunk_node): Simplify using
16458 symtab_alias_ultimate_target.
16459 (varpool_variable_node): Likewise.
16460 * cgraph.c (cgraph_create_function_alias): Update.
16461 (cgraph_add_thunk): Update.
16462 (cgraph_remove_node): Update.
16463 (dump_cgraph_node): Do not dump removed flags.
16464 (cgraph_function_body_availability): Update.
16465 (cgraph_propagate_frequency): Update.
16466 (verify_cgraph_node): Check sanity of local flag.
16467 (cgraph_function_node): Move here from cgraph.h; revamp for
16468 cgraph_function_or_thunk_node.
16469 * lto-symtab.c (lto_varpool_replace_node): Update.
16470 (lto_symtab_resolve_can_prevail_p): Update.
16471 (lto_symtab_merge_cgraph_nodes): Update.
16472 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16473 propagate_constants_accross_call, devirtualization_time_bonus,
16474 ipcp_propagate_stage): Update.
16475 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16476 * ipa-inline-transform.c (clone_inlined_nodes,
16477 preserve_function_body_p): Update.
16478 * ipa-reference.c (propagate): Update.
16479 (write_node_summary_p): Update.
16480 * toplev.c (wrapup_global_declaration_2): Update.
16481 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16482 (analyze_function) ... this one.
16483 (cgraph_process_new_functions): Update.
16484 (cgraph_reset_node): Export.
16485 (cgraph_finalize_function): Update.
16486 (cgraph_add_new_function): Update.
16487 (process_function_and_variable_attributes): Update.
16488 (varpool_finalize_decl): Update.
16489 (symbol_finalized): Remove.
16490 (symbol_finalized_and_needed): Rename to ...
16491 (symbol_defined_and_needed): ... update.
16492 (cgraph_analyze_functions): Update.
16493 (handle_alias_pairs): Update.
16494 (mark_functions_to_output): Update.
16495 (assemble_thunk): Update.
16496 (output_in_order): Update.
16497 (output_weakrefs): Update.
16498 (finalize_compilation_unit): Update.
16499 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16500 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16501 input_node, input_varpool_node): Update.
16502 * dbxout.c (dbxout_expand_expr): Update.
16503 * cgraphclones.c (cgraph_clone_node): Update.
16504 (cgraph_copy_node_for_versioning): Update.
16505 (cgraph_materialize_clone): Update.
16506 (cgraph_materialize_all_clones): Update.
16507 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16508 propagate_pure_const, propagate_nothrow): Update.
16509 * lto-streamer-out.c (lto_output, write_symbol): Update.
16510 * ipa-utils.c (ipa_reverse_postorder): Update.
16511 * ipa-inline.c (can_inline_edge_p): Update.
16512 (update_caller_keys, ipa_inline): Update.
16513 * dwarf2out.c (reference_to_unused,
16514 premark_types_used_by_global_vars_helper): Update.
16515 * tree-eh.c (tree_could_trap_p): Update.
16516 * ipa-split.c (consider_split, execute_split_functions): Update.
16517 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16518 has_addr_references_p): Update; move ahead in file for better
16520 (process_references): Simplify.
16521 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16522 bodies are removed.
16523 (cgraph_comdat_can_be_unshared_p): Make static.
16524 (cgraph_externally_visible_p): Update.
16525 (varpool_externally_visible_p): Update.
16526 (function_and_variable_visibility): Update.
16527 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16528 ipa_tm_mark_force_output_node): Update.
16529 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16530 estimate_edge_devirt_benefit, inline_generate_summary,
16531 inline_write_summary): Update.
16532 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16533 * ipa-prop.c (ipa_compute_jump_functions): Update.
16534 (ipa_print_node_params, ipa_prop_read_section,
16535 ipa_update_after_lto_read, read_replacements_section): Update.
16536 * varasm.c (mark_decl_referenced): Update.
16537 (assemble_alias, dump_tm_clone_pairs): Update.
16538 * tree-inline.c (copy_bb): Update.
16539 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16541 * symtab.c (dump_symtab_base): Print new flags.
16542 (verify_symtab_base): Verify new flags.
16543 (symtab_alias_ultimate_target): New function.
16544 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16545 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16547 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16549 * i386.c (ix86_get_function_versions_dispatcher,
16550 ix86_generate_version_dispatcher_body): Update.
16551 (fold_builtin_cpu): Use varpool_add_new_variable.
16552 * varpool.c (varpool_remove_initializer): Break out from ...
16553 (varpool_remove_node): ... this one.
16554 (dump_varpool_node, varpool_node_for_asm,
16555 cgraph_variable_initializer_availability, varpool_analyze_node,
16556 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16557 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16560 2013-05-29 Jan Hubicka <jh@suse.cz>
16562 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16564 2013-05-29 Easwaran Raman <eraman@google.com>
16566 PR tree-optimization/57442
16567 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16568 when control exits the main loop.
16570 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16572 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16574 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16575 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16576 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16579 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16581 PR tree-optimization/57441
16582 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16583 Don't limit size of incr_vec to number of candidates.
16585 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16587 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16588 and mips16 directories.
16589 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16590 (MULTILIB_DIRNAMES): Ditto.
16591 (MULTILIB_EXCEPTIONS): Add new exceptions.
16592 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16593 (MULTILIB_DIRNAMES): Ditto.
16594 (MULTILIB_EXCEPTIONS): Add new exceptions.
16596 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16597 Marcus Shawcroft <marcus.shawcroft@arm.com>
16599 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16600 SYMBOL_TINY_ABSOLUTE.
16601 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16602 SYMBOL_TINY_ABSOLUTE.
16603 (aarch64_expand_mov_immediate): Likewise.
16604 (aarch64_classify_symbol): Likewise.
16605 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16606 Permit SYMBOL_TINY_ABSOLUTE.
16607 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16609 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16610 Marcus Shawcroft <marcus.shawcroft@arm.com>
16612 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16613 Refactor if/switch. Replace gcc_assert with if.
16615 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16617 * config/i386/i386.c (initial_ix86_tune_features): Enable
16618 FP Reassociation for AMD bdver1 and bdver2.
16620 2013-05-29 Martin Jambor <mjambor@suse.cz>
16622 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16623 and IMAGPART_EXPR do not occur within other handled_components.
16625 2013-05-29 Richard Biener <rguenther@suse.de>
16627 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16628 access on whether the use is in the BB we currently try to
16630 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16631 vectorize to vect_bb_slp_scalar_cost.
16633 2013-05-29 Richard Biener <rguenther@suse.de>
16635 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16636 computing scalar cost offsetted by stmts that are kept live
16638 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16639 for computation of scalar cost.
16641 2013-05-28 Steve Ellcey <sellcey@mips.com>
16643 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16645 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16647 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16648 array notation built-in reduction functions.
16649 * doc/passes.texi (Passes): Added documentation about changes done
16651 * doc/invoke.texi (C Dialect Options): Added documentation about
16652 the -fcilkplus flag.
16653 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16654 (BUILTINS_DEF): Depend on cilkplus.def.
16655 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16656 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16657 * cilkplus.def: New file.
16659 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16661 PR rtl-optimization/57439
16662 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16664 2013-05-28 Easwaran Raman <eraman@google.com>
16666 PR tree-optimization/57337
16667 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16668 (find_insert_point): Correctly identify the insertion point
16669 when two statements with the same UID is compared.
16671 2013-05-28 Richard Biener <rguenther@suse.de>
16673 PR tree-optimization/56787
16674 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16675 from the list of data references.
16676 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16678 (vect_analyze_loop_operations): Likewise.
16679 (vect_transform_loop): Remove clobbers.
16681 2013-05-28 Martin Jambor <mjambor@suse.cz>
16683 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16684 and REALPART_EXPRs have scalar type.
16686 2013-05-28 Richard Biener <rguenther@suse.de>
16688 PR tree-optimization/57411
16689 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16691 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16692 virtual operand propagation.
16694 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16696 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16697 destination register for bmasksi_vis.
16698 (vector_init_bshuffle): Likewise.
16699 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16701 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16703 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16704 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16705 mode if the instruction isn't available in the original mode.
16706 * config/sparc/sparc.opt (mfix-ut699): New option.
16707 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16708 (divdf3): Turn into expander.
16709 (divdf3_nofix): New insn.
16710 (divdf3_fix): Likewise.
16711 (divsf3): Disable if -mfix-ut699.
16712 (sqrtdf2): Turn into expander.
16713 (sqrtdf2_nofix): New insn.
16714 (sqrtdf2_fix): Likewise.
16715 (sqrtsf2): Disable if -mfix-ut699.
16717 2013-05-27 Richard Biener <rguenther@suse.de>
16719 PR middle-end/57412
16720 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16721 block for the new loop.
16723 2013-05-27 Richard Biener <rguenther@suse.de>
16725 PR tree-optimization/57343
16726 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16727 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16728 (number_of_iterations_cond): Do not build the folded tree.
16730 2013-05-27 Richard Biener <rguenther@suse.de>
16733 PR middle-end/57381
16734 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16735 OEP_CONSTANT_ADDRESS_OF retained.
16737 PR tree-optimization/57417
16738 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16739 for unchanged base.
16740 (set_ssa_val_to): Compare addresses using
16741 get_addr_base_and_unit_offset.
16743 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16745 PR rtl-optimization/56833
16746 * postreload.c (move2add_record_mode): New function.
16747 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16748 (move2add_use_add2_insn): Use move2add_record_sym_value.
16749 (move2add_use_add3_insn): Likewise.
16750 (reload_cse_move2add): Use move2add_valid_value_p and
16751 move2add_record_mode. Invalidate call-clobbered and REG_INC
16752 affected regs by setting reg_mode to VOIDmode.
16753 (move2add_note_store): Don't pretend the inside of a SUBREG is
16754 the actual destination. Invalidate single/leading registers by
16755 setting reg_mode to VOIDmode.
16756 Use move2add_record_sym_value, move2add_valid_value_p and
16757 move2add_record_mode.
16759 2013-05-27 Richard Biener <rguenther@suse.de>
16761 PR tree-optimization/57396
16762 * tree-affine.c (double_int_constant_multiple_p): Properly
16763 return false for val == 0 and div != 0.
16765 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16767 * config/mips/mips.h: Use #elif in preprocessor conditions.
16769 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16772 * config/mips/constraints.md (kl): New constraint.
16773 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16774 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16775 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16776 from LO for MIPS16.
16777 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16779 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16782 * config/mips/mips.c (mips_can_inline_p): New function.
16783 (TARGET_CAN_INLINE_P): Define.
16785 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16787 * sched-int.h (ds_t, dw_t): Make unsigned int.
16788 Fix documentation that describes how all the ds_t bits are used.
16789 Reserve the last bit for delayed-branch scheduling.
16790 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16791 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16792 (gen_dep_weak_1): Remove prototype.
16793 * sched-deps.c (get_dep_weak_1): Make static.
16794 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16795 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16796 * doc/tm.texi: Regenerate.
16797 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16799 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16802 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16804 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16805 Sandra Loosemore <sandra@codesourcery.com>
16807 * config.gcc (powerpc-*): Allow native for with-cpu.
16809 2013-05-24 Jeff Law <law@redhat.com>
16811 PR tree-optimization/57124
16812 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16813 conversion feeding a condition if the range has an overflow
16814 if -fstrict-overflow. Add warnings for when we do make the
16817 2013-05-24 Dehao Chen <dehao@google.com>
16819 * tree-cfg.c (locus_discrim_map): Fix the typo.
16820 (locus_discrim_hasher): Likewise.
16821 (locus_discrim_hasher::hash): Likewise.
16822 (locus_discrim_hasher::equal): Likewise.
16824 2013-05-24 Martin Jambor <mjambor@suse.cz>
16826 PR tree-optimization/57294
16827 * cgraph.h (ipa_record_stmt_references): Declare.
16828 * cgraphbuild.c (ipa_record_stmt_references): New function.
16829 (build_cgraph_edges): Use ipa_record_stmt_references.
16830 (rebuild_cgraph_edges): Likewise.
16831 (cgraph_rebuild_references): Likewise.
16832 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16833 associated with the old statement and build references from the
16834 newly built statements.
16835 * ipa-ref.c (ipa_remove_stmt_references): New function.
16836 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16838 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
16840 * lra-constraints.c (emit_spill_move): Use smaller mode for
16842 (check_and_process_move): Consider mem-reg moves for secondary
16844 (curr_insn_transform): Don't lose insns emitted before for
16845 secondary memory moves.
16846 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
16847 reg set up in the current insn.
16849 2013-05-24 Dehao Chen <dehao@google.com>
16851 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
16853 (locus_descrim_hasher::equal): Likewise.
16854 (build_gimple_cfg): New discriminator assignment algorithm.
16855 (make_edges): Likewise.
16856 (next_discriminator_for_locus): Likewise.
16857 (same_line_p): Likewise.
16858 (assign_discriminators): Likewise.
16859 (make_cond_expr_edges): Likewise.
16860 (make_gimple_switch_edges): Likewise.
16861 (make_goto_expr_edges): Likewise.
16862 (make_gimple_asm_edges): Likewise.
16864 2013-05-24 Ian Bolton <ian.bolton@arm.com>
16866 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
16867 X format specifier to only display bottom 16 bits.
16868 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
16869 immediate to match for operand 2, since it will be masked.
16871 2013-05-24 Richard Biener <rguenther@suse.de>
16873 PR tree-optimization/57287
16874 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
16875 all SSA names that occur in abnormal PHIs.
16877 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
16880 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
16881 that index is not negative.
16883 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
16885 PR rtl-optimization/55177
16886 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
16887 (simplify_byte_swapping_operation): New.
16888 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
16889 (simplify_relational_operation_1): Deal with BSWAP.
16891 2013-05-23 Richard Henderson <rth@redhat.com>
16894 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
16895 (ix86_reorg): Call it.
16897 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
16900 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
16901 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
16902 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
16904 2013-05-23 Christian Bruel <christian.bruel@st.com>
16907 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
16909 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16910 Marcus Shawcroft <marcus.shawcroft@arm.com>
16912 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
16913 * config/aarch64/constraints.md (Usa): Remove.
16914 * doc/md.texi (AArch64 Usa): Remove.
16916 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16917 Marcus Shawcroft <marcus.shawcroft@arm.com>
16919 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
16920 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
16921 * config/aarch64/predicates.md (aarch64_const_address): Remove.
16922 (aarch64_mov_operand): Use aarch64_mov_operand_p.
16924 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
16926 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
16927 instruction (AdvSIMD).
16928 * config/aarch64/aarch64-builtins.c
16929 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
16930 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
16932 2013-05-23 Martin Jambor <mjambor@suse.cz>
16934 PR middle-end/57347
16935 * tree.h (contains_bitfld_component_ref_p): Declare.
16936 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
16937 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
16939 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
16940 not access a bit-field. Assert all final offsets are byte-aligned.
16942 2013-05-23 Richard Biener <rguenther@suse.de>
16944 PR tree-optimization/57380
16945 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
16946 least one invariant or re-used load.
16947 * passes.c (init_optimization_passes): Move pass_phiprop before
16950 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
16952 * config/aarch64/aarch64-simd.md
16953 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
16955 2013-05-23 Richard Biener <rguenther@suse.de>
16957 PR middle-end/57381
16958 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16959 OEP_CONSTANT_ADDRESS_OF retained.
16961 2013-05-23 Jakub Jelinek <jakub@redhat.com>
16963 PR middle-end/57344
16964 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
16965 don't lower unit. Handle unit not being always BITS_PER_WORD.
16967 2013-05-23 Richard Biener <rguenther@suse.de>
16969 PR rtl-optimization/57341
16970 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
16971 instead of true_dependence.
16973 2013-05-22 David Malcolm <dmalcolm@redhat.com>
16975 * bb-reorder.c (branch_threshold): Make const.
16976 (exec_threshold): Ditto.
16978 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
16979 Pat Haugen <pthaugen@us.ibm.com>
16980 Peter Bergner <bergner@vnet.ibm.com>
16982 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
16983 documentation for the power8 crypto builtins.
16985 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
16987 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
16988 macros for defining power8 builtin functions.
16989 (BU_P8V_AV_2): Likewise.
16990 (BU_P8V_AV_P): Likewise.
16991 (BU_P8V_VSX_1): Likewise.
16992 (BU_P8V_OVERLOAD_1): Likewise.
16993 (BU_P8V_OVERLOAD_2): Likewise.
16994 (BU_CRYPTO_1): Likewise.
16995 (BU_CRYPTO_2): Likewise.
16996 (BU_CRYPTO_3): Likewise.
16997 (BU_CRYPTO_OVERLOAD_1): Likewise.
16998 (BU_CRYPTO_OVERLOAD_2): Likewise.
16999 (XSCVSPDP): Fix typo, point to the correct instruction.
17000 (VCIPHER): Add power8 crypto builtins.
17001 (VCIPHERLAST): Likewise.
17002 (VNCIPHER): Likewise.
17003 (VNCIPHERLAST): Likewise.
17004 (VPMSUMB): Likewise.
17005 (VPMSUMH): Likewise.
17006 (VPMSUMW): Likewise.
17007 (VPERMXOR_V2DI): Likewise.
17008 (VPERMXOR_V4SI: Likewise.
17009 (VPERMXOR_V8HI: Likewise.
17010 (VPERMXOR_V16QI: Likewise.
17011 (VSHASIGMAW): Likewise.
17012 (VSHASIGMAD): Likewise.
17013 (VPMSUM): Likewise.
17014 (VPERMXOR): Likewise.
17015 (VSHASIGMA): Likewise.
17017 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17018 __CRYPTO__ if the crypto instructions are available.
17019 (altivec_overloaded_builtins): Add support for overloaded power8
17022 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
17023 support for power8 crypto builtins.
17024 (builtin_function_type): Likewise.
17025 (altivec_init_builtins): Add support for builtins that take vector
17026 long long (V2DI) arguments.
17028 * config/rs6000/crypto.md: New file, define power8 crypto
17031 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17032 Pat Haugen <pthaugen@us.ibm.com>
17033 Peter Bergner <bergner@vnet.ibm.com>
17035 * doc/invoke.texi (Option Summary): Add power8 options.
17036 (RS/6000 and PowerPC Options): Likewise.
17038 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
17039 constraints.md instead of rs6000.h. Reorder w* constraints. Add
17040 wm, wn, wr documentation.
17042 * config/rs6000/constraints.md (wm): New constraint for VSX
17043 registers if direct move instructions are enabled.
17044 (wn): New constraint for no registers.
17045 (wq): New constraint for quad word even GPR registers.
17046 (wr): New constraint if 64-bit instructions are enabled.
17047 (wv): New constraint if power8 vector instructions are enabled.
17048 (wQ): New constraint for quad word memory locations.
17050 * config/rs6000/predicates.md (const_0_to_15_operand): New
17051 constraint for 0..15 for crypto instructions.
17052 (gpc_reg_operand): If VSX allow registers in VSX registers as well
17053 as GPR and floating point registers.
17054 (int_reg_operand): New predicate to match only GPR registers.
17055 (base_reg_operand): New predicate to match base registers.
17056 (quad_int_reg_operand): New predicate to match even GPR registers
17057 for quad memory operations.
17058 (vsx_reg_or_cint_operand): New predicate to allow vector logical
17059 operations in both GPR and VSX registers.
17060 (quad_memory_operand): New predicate for quad memory operations.
17061 (reg_or_indexed_operand): New predicate for direct move support.
17063 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
17064 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
17065 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
17066 (POWERPC_MASKS): Add power8 options.
17067 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
17070 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
17071 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
17073 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
17074 (-mpower8-fusion): New power8 options.
17075 (-mpower8-fusion-sign): Likewise.
17076 (-mpower8-vector): Likewise.
17077 (-mcrypto): Likewise.
17078 (-mdirect-move): Likewise.
17079 (-mquad-memory): Likewise.
17081 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
17082 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
17084 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
17085 (rs6000_debug_vector_unit): Add p8_vector.
17086 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
17087 definitions. Also print fusion state.
17088 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
17089 (rs6000_builtin_mask_calculate): Add power8 builtin support.
17090 (rs6000_option_override_internal): Add support for power8.
17091 (rs6000_common_init_builtins): Add debugging for skipped builtins
17092 if -mdebug=builtin.
17093 (rs6000_adjust_cost): Add power8 support.
17094 (rs6000_issue_rate): Likewise.
17095 (insn_must_be_first_in_group): Likewise.
17096 (insn_must_be_last_in_group): Likewise.
17097 (force_new_group): Likewise.
17098 (rs6000_register_move_cost): Likewise.
17099 (rs6000_opt_masks): Likewise.
17101 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
17102 power8 capable assembler, default to power7 options.
17103 (TARGET_DIRECT_MOVE): Likewise.
17104 (TARGET_CRYPTO): Likewise.
17105 (TARGET_P8_VECTOR): Likewise.
17106 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
17107 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
17108 (VECTOR_MEM_P8_VECTOR_P): Likewise.
17109 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
17110 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
17111 (TARGET_XSCVDPSPN): Likewise.
17112 (TARGET_XSCVSPDPN): Likewsie.
17113 (TARGET_SYNC_HI_QI): Likewise.
17114 (TARGET_SYNC_TI): Likewise.
17115 (MASK_CRYPTO): Likewise.
17116 (MASK_DIRECT_MOVE): Likewise.
17117 (MASK_P8_FUSION): Likewise.
17118 (MASK_P8_VECTOR): Likewise.
17119 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
17120 temporary used by some of the direct move instructions to get two FP
17121 temporary registers does not force creation of a stack frame.
17122 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
17123 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
17124 that any VSX registers are tieable, even if they are also an
17125 Altivec vector mode.
17126 (r6000_reg_class_enum): Add wm, wr, wv constraints.
17127 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
17128 (RS6000_BTM_CRYPTO): Likewise.
17129 (RS6000_BTM_COMMON): Likewise.
17131 * config/rs6000/rs6000.md (cpu attribute): Add power8.
17132 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
17133 (enum rs6000_vector): Add power8 vector support.
17135 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17139 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
17140 (any_sibcall_could_use_r3): this and handle indirect calls.
17141 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
17143 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17145 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
17147 2013-05-22 Richard Biener <rguenther@suse.de>
17149 PR middle-end/57349
17150 * profile.c (branch_prob): Do not split blocks that are
17151 abnormally receiving from ECF_RETURNS_TWICE functions.
17153 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17155 * recog.c (offsettable_address_addr_space_p): Fix calculation of
17156 address mode. Move pointer mode initialization to the same place.
17158 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
17160 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
17161 while it has any effect.
17163 2013-05-21 Easwaran Raman <eraman@google.com>
17165 PR tree-optimization/57322
17166 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
17167 UID of the statement added to the BB to be 1.
17169 2013-05-21 Jakub Jelinek <jakub@redhat.com>
17171 PR tree-optimization/57331
17172 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
17173 of conversion from pointer type to integral type with integer.
17175 2013-05-21 Martin Jambor <mjambor@suse.cz>
17178 * ipa-prop.c (ipa_read_node_info): Process param_used and
17179 controlled_uses in the same order as when writing.
17181 2013-05-21 Magnus Granberg <baldrick@free.fr>
17184 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17186 2013-05-21 Richard Biener <rguenther@suse.de>
17188 PR tree-optimization/57318
17189 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17190 estimate stmts with side-effects as likely eliminated.
17192 2013-05-21 Richard Biener <rguenther@suse.de>
17194 PR tree-optimization/57330
17195 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17196 preserve the call stmts fntype.
17198 2013-05-21 Richard Biener <rguenther@suse.de>
17200 PR tree-optimization/57303
17201 * tree-ssa-sink.c (statement_sink_location): Improve killing
17202 stmt detection and properly handle self-assignments.
17204 2013-05-21 Christian Bruel <christian.bruel@st.com>
17206 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17207 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17208 registers. Set register size out of the PARALLEL loop.
17210 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17213 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17214 (*fmasf4, *fmasf4_media): New insns.
17216 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17218 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17219 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17220 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17221 (mips_idiv_insns): Update the comments to say that the returned
17222 instruction counts are in units of BASE_INSN_LENGTH.
17223 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17224 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17225 using 2 rather than 4 as the length of indirect MIPS16 and
17226 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17227 length of a NOP. Don't divide MIPS16 lengths by 2.
17228 (mips16_split_long_branches): Assume a branch is long if the
17229 length is greater than 4 rather than 8.
17230 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17231 rather than multiplying them by 2. Multiply instruction counts
17232 by BASE_INSN_LENGTH rather than 4.
17233 (*jump_mips16, tls_get_tp_mips16_<mode>)
17234 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17236 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17238 * config/mips/mips.md (extended_mips16): Remove branch case.
17239 (length): Remove duplicated extended_mips16 test.
17241 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17243 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17245 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17247 * recog.h (Recog_data): Rename to...
17248 (recog_data_d): ...this.
17249 (recog_data): Update accordingly.
17250 * recog.c (recog_data): Likewise.
17251 * reload.c (save_recog_data): Likewise.
17252 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17253 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17255 2013-05-17 Julian Brown <julian@codesourcery.com>
17257 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17258 found in a REG_EQUAL note, invalidate it.
17260 2013-05-17 Easwaran Raman <eraman@google.com>
17262 * tree-ssa-reassoc.c (find_insert_point): New function.
17263 (insert_stmt_after): Likewise.
17264 (get_def_stmt): Likewise.
17265 (ensure_ops_are_available): Likewise.
17266 (not_dominated_by): Likewise.
17267 (rewrite_expr_tree): Do not move statements beyond what is
17268 necessary. Remove call to swap_ops_for_binary_stmt...
17269 (reassociate_bb): ... and move it here.
17270 (build_and_add_sum): Assign UIDs for new statements.
17271 (linearize_expr): Likewise.
17272 (do_reassoc): Renumber gimple statement UIDs.
17274 2013-05-17 Jan Hubicka <jh@suse.cz>
17276 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17278 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17279 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17281 (output_weakrefs): Update.
17283 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17284 Martin Jambor <mjambor@suse.cz>
17286 PR middle-end/57276
17287 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17288 value that corresponds to the given aggval is found in values vector.
17290 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17292 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17293 sse, sse2, sse3, ssse3 and sse4a flags to options.
17295 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17297 * gengtype-state.c: (s_expr_writer): New class, to handle
17298 prettifying of output layout of s-expressions.
17299 (state_writer): New class, to write out gtype.state.
17300 (state_written_type_count): Move this variable into member data of
17302 (s_expr_writer::s_expr_writer): New code: constructor for new class
17303 (state_writer::state_writer(): ditto
17304 (s_expr_writer::write_new_line): New function
17305 (s_expr_writer::write_any_indent): ditto
17306 (s_expr_writer::begin_s_expr): ditto
17307 (s_expr_writer::end_s_expr): ditto
17308 (write_state_fileloc): convert to method of state_writer...
17309 (state_writer:: write_state_fileloc): ...and use methods of
17310 s_expr_writer to write indentation into the gtype.state output file
17311 to visually represent the hierarchical structure of the list
17313 (write_state_fields): ditto, renaming to...
17314 (state_writer::write_state_fields)
17315 (write_state_a_string): ditto, renaming to...
17316 (state_writer::write_state_a_string)
17317 (write_state_string_option): ditto, renaming to...
17318 (state_writer::write_state_string_option)
17319 (write_state_type_option): ditto, renaming to...
17320 (state_writer::write_state_type_option)
17321 (write_state_nested_option): ditto, renaming to...
17322 (state_writer::write_state_nested_option)
17323 (write_state_option): ditto, renaming to...
17324 (state_writer::write_state_option)
17325 (write_state_options): ditto, renaming to...
17326 (state_writer::write_state_options)
17327 (write_state_lang_bitmap): ditto, renaming to...
17328 (state_writer::write_state_lang_bitmap)
17329 (write_state_version): ditto, renaming to...
17330 (state_writer::write_state_version)
17331 (write_state_scalar_type): ditto, renaming to...
17332 (state_writer::write_state_scalar_type)
17333 (write_state_string_type): ditto, renaming to...
17334 (state_writer::write_state_string_type)
17335 (write_state_undefined_type): ditto, renaming to...
17336 (state_writer::write_state_undefined_type)
17337 (write_state_struct_union_type): ditto, renaming to...
17338 (state_writer::write_state_struct_union_type)
17339 (write_state_struct_type): ditto, renaming to...
17340 (state_writer::write_state_struct_type)
17341 (write_state_user_struct_type): ditto, renaming to...
17342 (state_writer::write_state_user_struct_type)
17343 (write_state_lang_struct_type): ditto, renaming to...
17344 (state_writer::write_state_lang_struct_type)
17345 (write_state_param_struct_type): ditto, renaming to...
17346 (state_writer::write_state_param_struct_type)
17347 (write_state_pointer_type): ditto, renaming to...
17348 (state_writer::write_state_pointer_type)
17349 (write_state_array_type): ditto, renaming to...
17350 (state_writer::write_state_array_type)
17351 (write_state_gc_used): ditto, renaming to...
17352 (state_writer::write_state_gc_used)
17353 (write_state_common_type_content): ditto, renaming to...
17354 (state_writer::write_state_common_type_content)
17355 (write_state_type): ditto, renaming to...
17356 (state_writer::write_state_type)
17357 (write_state_pair_list): ditto, renaming to...
17358 (state_writer::write_state_pair_list)
17359 (write_state_pair): ditto, renaming to...
17360 (state_writer::write_state_pair)
17361 (write_state_typedefs): ditto, renaming to...
17362 (state_writer::write_state_typedefs)
17363 (write_state_structures): ditto, renaming to...
17364 (state_writer::write_state_structures)
17365 (write_state_param_structs): ditto, renaming to...
17366 (state_writer::write_state_param_structs)
17367 (write_state_variables): ditto, renaming to...
17368 (state_writer::write_state_variables)
17369 (write_state_srcdir): ditto, renaming to...
17370 (state_writer::write_state_srcdir)
17371 (write_state_files_list): ditto, renaming to...
17372 (state_writer::write_state_files_list)
17373 (write_state_languages): ditto, renaming to...
17374 (state_writer::write_state_languages)
17375 (write_state): create a state_writer instance and use it when
17376 writing out the state file
17378 2013-05-17 Mike Stump <mikestump@comcast.net>
17380 PR rtl-optimization/57304
17381 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17382 accessing DF_REF_REAL_LOC.
17384 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17386 PR rtl-optimization/57281
17387 PR rtl-optimization/57300
17388 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17389 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17390 what the other splitter did if the registers are dead.
17392 2013-05-17 Richard Biener <rguenther@suse.de>
17394 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17397 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17399 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17402 2013-05-17 Marek Polacek <polacek@redhat.com>
17404 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17405 length when doing non-zero store of storing '\0' to '\0'.
17407 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17409 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17410 vect_external_def oprnd1 with loop_vinfo, try to emit
17411 optional cast, negation and and stmts on the loop preheader
17412 edge instead of into the pattern def seq.
17414 PR tree-optimization/57051
17415 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17416 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17418 2013-05-16 Nick Clifton <nickc@redhat.com>
17420 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17421 (rl78_is_naked_func): New function.
17422 (rl78_expand_prologue): Skip prologue generation for naked functions.
17423 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17424 * doc/extend.texi (naked): Add RL78 to the list of processors
17425 that supports this attribute.
17427 2013-05-16 Jeff Law <law@redhat.com>
17429 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17431 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17433 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17434 cache parameters using detect_caches_amd also for CYRIX,
17435 NSC and TM2 signatures.
17437 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17438 Dzianis Kahanovich <mahatma@eu.by>
17442 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17443 VIA/Centaur processors and determine their cache parameters
17444 using detect_caches_amd.
17446 2013-05-16 Teresa Johnson <tejohnson@google.com>
17448 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17449 (rtl_verify_edges): New function.
17450 (rtl_verify_bb_insns): Ditto.
17451 (rtl_verify_bb_pointers): Ditto.
17452 (rtl_verify_bb_insn_chain): Ditto.
17453 (rtl_verify_fallthru): Ditto.
17454 (rtl_verify_bb_layout): Ditto.
17455 (rtl_verify_flow_info_1): Outline checks into new functions.
17456 (rtl_verify_flow_info): Ditto.
17458 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17460 * cfghooks.c (copy_bbs): Add update_dominance argument.
17461 * cfghooks.h (copy_bbs): Update prototype.
17462 * tree-cfg.c (gimple_duplicate_sese_region):
17463 Add update_dominance argument.
17464 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17465 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17466 gimple_duplicate_sese_region call.
17467 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17468 Update copy_bbs call.
17469 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17470 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17472 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17474 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17475 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17476 vect_recog_rotate_pattern.
17477 (vect_recog_rotate_pattern): New function.
17479 2013-05-16 Jason Merrill <jason@redhat.com>
17481 * Makefile.in (LLINKER): New variable.
17482 (mostlyclean): Remove link mutex.
17483 * configure.ac: Handle --enable-link-mutex.
17484 * lock-and-run.sh: New script.
17486 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17489 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17492 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17494 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17496 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17498 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17499 * config/arm/arm.c (next_consecutive_mem): New function.
17500 (gen_movmem_ldrd_strd): Likewise.
17501 * config/arm/arm.md (movmemqi): Update condition and code.
17502 (unaligned_loaddi, unaligned_storedi): New patterns.
17504 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17506 * config.gcc: Obsolete *-*-solaris2.9*.
17507 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17509 2013-05-16 Richard Biener <rguenther@suse.de>
17511 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17512 earlier, after GRAPHITE transforms and IV canonicalization.
17514 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17516 * omp-low.c (extract_omp_for_data): For collapsed loops,
17517 if at least one of the loops is known at compile time to
17518 iterate zero times, set count to 0.
17519 (expand_omp_regimplify_p): New function.
17520 (expand_omp_for_generic): For collapsed loops, if at least
17521 one of the loops isn't known to iterate at least once,
17522 add runtime check with setting count to 0.
17523 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17524 For unsigned types if it isn't known at compile time that
17525 the loop will iterate at least once, add runtime check to bypass
17526 the whole loop if initial condition isn't true.
17528 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17530 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17532 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17534 PR middle-end/57286
17535 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17536 transformations to avoid an infinite loop.
17538 2013-05-16 Marek Polacek <polacek@redhat.com>
17540 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17542 2013-05-15 Leif Ekblad <leif@rdos.net>
17544 * config/i386/i386.c (ix86_decompose_address): Use
17545 DEFAULT_TLS_SEG_REG to access TLS segment register.
17546 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17547 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17548 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17550 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17553 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17554 sibling calls to functions that would normally be lazily bound,
17555 unless $gp is call-clobbered.
17557 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17559 * config/i386/i386.c (ix86_option_override_internal): Update
17560 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17561 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17562 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17563 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17565 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17567 2013-05-15 Andreas Schwab <schwab@suse.de>
17569 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17570 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17573 2013-05-15 Teresa Johnson <tejohnson@google.com>
17575 * loop-unroll.c (report_unroll_peel): Check decision before
17576 emitting unroll/peel message.
17578 2013-05-15 Teresa Johnson <tejohnson@google.com>
17580 * function.h (has_bb_partition): New rtl_data flag.
17581 (bb_reorder_complete): Ditto.
17582 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17583 instead of flag_reorder_blocks_and_partition.
17584 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17585 with some enhancements.
17586 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17587 * bb-reorder.c (connect_traces): Check for has_bb_partition
17588 instead of flag_reorder_blocks_and_partition.
17589 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17590 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17591 verify_hot_cold_block_grouping.
17592 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17594 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17597 * config/arm/predicates.md (call_insn_operand): New predicate.
17598 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17599 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17600 if insn is not a tail call.
17601 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17603 * config/arm/arm.h (enum reg_class): New caller save register class.
17604 (REG_CLASS_NAMES): Likewise.
17605 (REG_CLASS_CONTENTS): Likewise.
17606 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17609 2013-05-15 Richard Biener <rguenther@suse.de>
17611 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17612 of MSG_OPTIMIZED_LOCATIONS.
17613 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17614 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17616 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17617 of MSG_OPTIMIZED_LOCATIONS.
17618 (execute_vect_slp): Likewise.
17619 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17620 (vect_create_cond_for_alias_checks): Likewise.
17621 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17622 (vect_recog_widen_mult_pattern): Likewise.
17623 (vect_recog_widen_sum_pattern): Likewise.
17624 (vect_recog_over_widening_pattern): Likewise.
17625 (vect_recog_widen_shift_pattern): Likewise.
17626 (vect_recog_vector_vector_shift_pattern): Likewise.
17627 (vect_recog_divmod_pattern): Likewise.
17628 (vect_recog_mixed_size_cond_pattern): Likewise.
17629 (vect_recog_bool_pattern): Likewise.
17630 (vect_pattern_recog_1): Likewise.
17632 2013-05-15 Martin Jambor <mjambor@suse.cz>
17634 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17635 non-functions to builtin_unreachable.
17636 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17637 correct when new direct edges were discovered.
17639 2013-05-15 Martin Jambor <mjambor@suse.cz>
17641 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17642 header, print symbol order instead of node uid, print more information
17643 about indirect edge targets.
17644 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17646 (ipa_make_edge_direct_to_target): Likewise.
17647 (remove_described_reference): Likewise.
17648 (propagate_controlled_uses): Likewise.
17649 (ipa_print_node_params): Also print symbol order.
17650 (ipcp_transform_function): Print symbol order instead of node uids.
17651 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17652 (cgraph_get_create_real_symbol_node): Likewise.
17653 * ipa-cp.c (print_lattice): Likewise.
17654 (print_all_lattices): Likewise.
17655 (determine_versionability): Likewise.
17656 (initialize_node_lattices): Likewise.
17657 (estimate_local_effects): Likewise.
17658 (update_profiling_info): Likewise.
17659 (create_specialized_node): Likewise.
17660 (perhaps_add_new_callers): Likewise.
17661 (decide_about_value): Likewise.
17662 (decide_whether_version_node): Likewise.
17663 (identify_dead_nodes): Likewise.
17664 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17665 (dump_inline_summary): Likewise.
17666 (estimate_node_size_and_time): Likewise.
17667 (inline_analyze_function): Likewise.
17668 * ipa-inline.c (report_inline_failed_reason): Likewise.
17669 (want_early_inline_function_p): Likewise.
17670 (edge_badness): Likewise.
17671 (update_edge_key): Likewise.
17672 (inline_small_functions): Likewise. Add dumping of order to two other
17674 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17675 instead of node uids.
17676 (propagate_pure_const): Likewise.
17677 (propagate_pure_const): Likewise.
17678 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17679 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17680 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17682 * tree-pretty-print.c (dump_function_header): Likewise.
17683 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17684 Print symbol order instead of node uids.
17686 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17688 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17689 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17691 2013-05-15 Richard Biener <rguenther@suse.de>
17693 PR tree-optimization/57275
17694 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17695 return value for fail to do runtime alias checks for gather loads.
17697 2013-05-15 Jan Hubicka <jh@suse.cz>
17701 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17702 weakrefs are not external.
17703 (lto_symtab_merge_decls): Fix thinko when dealing with
17704 non-lto_symtab decls.
17705 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17706 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17707 * varpool.c (dump_varpool_node): Dump more flags.
17709 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17711 * config/i386/i386.c (processor_alias_table): Add instruction
17712 FSGSBASE for AMD bdver3 architecture.
17714 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17716 * tree.c (warn_deprecated_use): Print file:line using locus color.
17717 * diagnostic.c (diagnostic_report_current_module): Print file:line
17718 and file:line:column using locus color.
17720 2013-05-14 Mike Stump <mikestump@comcast.net>
17722 * gdbinit.in: Add __null.
17724 2013-05-14 Mike Stump <mikestump@comcast.net>
17726 * recog.h: Rename struct recog_data to Recog_data.
17727 * recog.c: Likewise.
17728 * reload.c (can_reload_into): Likewise.
17729 * config/picochip/picochip.c: Likewise.
17731 2013-05-14 Mike Stump <mikestump@comcast.net>
17733 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17735 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17737 * resource.h (struct resources): Remove unch_memory member.
17738 (CLEAR_RESOURCE): Don't clear unch_memory.
17739 * resource.c (mark_referenced_resources): Don't set it.
17740 (mark_set_resources): Likewise.
17741 (mark_target_live_regs): Don't clear it.
17742 (init_resource_info): Likewise.
17743 * reorg.c (resource_conflicts_p): Don't compare it.
17744 (redundant_insn): Don't set it.
17746 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17748 * emit-rtl.c (next_label): Remove unused function.
17749 (skip_consecutive_labels, link_cc0_insns): Move to ...
17750 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17751 only place where these functions are used, and make them static.
17753 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17755 * fold-const.c (fold_negate_expr): Handle vectors.
17756 (fold_truth_not_expr): Make it static.
17757 (fold_invert_truthvalue): New static function.
17758 (invert_truthvalue_loc): Handle vectors. Do not call
17759 fold_truth_not_expr directly.
17760 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17761 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17762 (fold_comparison): Handle vector constants.
17763 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17764 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17765 * tree.h (fold_truth_not_expr): Remove declaration.
17767 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17769 * config/aarch64/aarch64-simd.md
17770 (aarch64_vcond_internal<mode>): Rename to...
17771 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17772 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17773 float modes. Clarify all iterator modes.
17774 (vcond<mode><mode>): Use new name for vcond expanders.
17775 (vcond<v_cmp_result><mode>): Likewise.
17776 (vcondu<mode><mode>: Likewise.
17777 * config/aarch64/iterators.md (VDQF_COND): New.
17779 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17782 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17783 variable for the shift amount. Check that we shift by non-negative
17786 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17789 * config/arm/arm.h (EPILOGUE_USES): Only return true
17790 for LR_REGNUM after epilogue_completed.
17792 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17794 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17795 is error_mark_node.
17797 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17800 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17801 and Solaris 11+/x86 with gld.
17802 * configure: Regenerate.
17804 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17806 * expmed.c (expand_shift_1): Canonicalize rotates by
17807 constant bitsize / 2 to bitsize - 1.
17808 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17809 case ROTATERT>: Likewise.
17812 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17814 * config/i386/i386.md (rotateinv): New code attr.
17815 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17816 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17817 roll $31, %eax, etc.
17819 2013-05-14 Richard Biener <rguenther@suse.de>
17821 PR middle-end/57235
17822 * tree-eh.c (sink_clobbers): Give up for successors with
17823 multiple predecessors and no virtual uses.
17825 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17827 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17828 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17830 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17832 PR middle-end/57251
17833 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17834 the case when both op0 and op1 have VOIDmode.
17836 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17838 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
17839 in multiply-accumulate mode.
17841 2013-05-13 Guozhi Wei <carrot@google.com>
17843 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
17845 2013-05-13 Kai Tietz <ktietz@redhat.com>
17848 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
17849 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
17850 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
17851 * config/i386/i386.c (ix86_option_override_internal): Likewise.
17852 (ix86_expand_prologue): Likewise.
17853 (ix86_expand_split_stack_prologue): Likewise.
17854 (legitimate_pic_address_disp_p): Likewise.
17855 (legitimize_pic_address): Likewise.
17856 (legitimize_tls_address): Likewise.
17857 (legitimize_pe_coff_symbol): Likewise.
17858 (output_pic_addr_const): Likewise.
17859 (construct_plt_address): Likewise.
17860 (ix86_expand_call): Likewise.
17861 (x86_output_mi_thunk): Likewise.
17862 (x86_function_profiler): Likewise.
17864 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
17866 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
17867 similar switch cases.
17868 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
17869 (aarch64_simd_mov_to_<mode>low): Delete.
17870 (aarch64_simd_mov_to_<mode>high): Delete.
17871 (move_lo_quad_<mode>): Add w<-r alternative.
17872 (aarch64_simd_move_hi_quad_<mode>): Likewise.
17873 (aarch64_simd_mov_from_*): Update type attribute.
17874 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
17877 2013-05-13 Jan Hubicka <jh@suse.cz>
17879 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
17880 * config/i386/i386.c (ix86_compute_frame_layout,
17881 ix86_expand_epilogue, emit_i387_cw_initialization,
17882 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
17883 ix86_local_alignment): Fix use of size/speed predicates.
17885 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17887 PR tree-optimization/45216
17888 PR tree-optimization/57157
17889 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
17890 the (-Y) & (B - 1) variant if OP is |.
17891 * expmed.c (expand_shift_1): For rotations by const0_rtx just
17892 return shifted. Use (-op1) & (prec - 1) as other_amount
17893 instead of prec - op1.
17895 2013-05-13 Martin Jambor <mjambor@suse.cz>
17897 PR middle-end/42371
17898 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
17899 (ipa_constant_data): New type.
17900 (ipa_jump_func): Use ipa_constant_data to hold information about
17901 constant jump functions.
17902 (ipa_get_jf_constant): Adjust to jump function type changes.
17903 (ipa_get_jf_constant_rdesc): New function.
17904 (ipa_param_descriptor): New field controlled_uses.
17905 (ipa_get_controlled_uses): New function.
17906 (ipa_set_controlled_uses): Likewise.
17907 * ipa-ref.h (ipa_find_reference): Declare.
17908 * ipa-prop.c (ipa_cst_ref_desc): New type.
17909 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
17911 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
17912 New parameter cs. Adjust all callers.
17913 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
17914 (remove_described_reference): New function.
17915 (jfunc_rdesc_usable): Likewise.
17916 (try_make_edge_direct_simple_call): Decrement controlled use count,
17917 attempt to remove reference if it hits zero.
17918 (combine_controlled_uses_counters): New function.
17919 (propagate_controlled_uses): Likewise.
17920 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
17921 (ipa_edge_duplication_hook): Duplicate reference descriptions.
17922 (ipa_print_node_params): Print described use counter.
17923 (ipa_write_jump_function): Adjust to jump function type changes.
17924 (ipa_read_jump_function): New parameter CS, pass it to
17925 ipa_set_jf_constant. Adjust caller.
17926 (ipa_write_node_info): Stream controlled use count
17927 (ipa_read_node_info): Likewise.
17928 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
17930 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
17931 count. Remove cloning-added reference if it reaches zero.
17932 * ipa-ref.c (ipa_find_reference): New function.
17934 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17936 * config/i386/i386.c (processor_target_table): Modified default
17937 alignment values for AMD BD and BT architectures.
17939 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
17941 * tree-vect-generic.c (uniform_vector_p): Move ...
17942 * tree.c (uniform_vector_p): ... here.
17943 * tree.h (uniform_vector_p): Declare it.
17944 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
17947 2013-05-13 Jakub Jelinek <jakub@redhat.com>
17949 PR tree-optimization/57230
17950 * tree-ssa-strlen.c (handle_char_store): Record length for
17951 array store from STRING_CST.
17953 PR tree-optimization/57230
17954 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
17957 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
17959 * config/epiphany/epiphany.c (epiphany_init): Check size of
17960 NUM_MODES_FOR_MODE_SWITCHING.
17961 (epiphany_expand_prologue):
17962 Remove CONFIG_REGNUM initial value handling code.
17963 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
17964 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
17965 (emit_set_fp_mode, epiphany_mode_after): Likewise.
17966 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
17967 Don't return 1 for FP_MODE_NONE.
17968 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
17969 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
17970 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
17971 * config/epiphany/epiphany.md (save_config): New pattern.
17973 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
17975 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
17977 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
17979 * config/i386/i386.md (memory): Handle sseishft1.
17980 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
17981 (*vec_extractv2di_1): Ditto.
17983 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
17985 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
17988 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
17990 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
17991 Add mthumb/march=armv7-a multilib.
17992 Add mthumb/march=armv7-r multilib.
17993 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
17995 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
17997 * config/v850/t-rtems: Add more multilibs.
17999 2013-05-10 Richard Biener <rguenther@suse.de>
18001 PR tree-optimization/57214
18002 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
18003 not propagate from SSA names that occur in abnormal PHI nodes.
18005 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
18007 * stor-layout.c (element_precision): New function.
18008 * machmode.h (element_precision): Declare it.
18009 * tree.c (build_minus_one_cst): New function.
18010 (element_precision): Likewise.
18011 * tree.h (build_minus_one_cst): Declare new function.
18012 (element_precision): Likewise.
18013 * fold-const.c (operand_equal_p): Use element_precision.
18014 (fold_binary_loc): Handle vector types.
18015 * convert.c (convert_to_integer): Use element_precision.
18016 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
18019 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
18021 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
18022 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
18023 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
18024 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18025 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
18026 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
18027 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
18028 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18029 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
18030 (Uuw8): New constraints.
18031 (Usb4): Move into alphabetical order.
18032 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
18033 (sd8_operand, ub8_operand, uw8_operand): New predicates.
18034 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
18035 previously unnamed patterns.
18036 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
18037 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
18038 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
18039 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
18040 of set_attr_alternative/if_then_else. Use extended_mips16 instead
18041 of specific lengths.
18043 2013-05-10 Jakub Jelinek <jakub@redhat.com>
18045 * config/i386/i386.md (rotateinv): New code attr.
18046 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
18047 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
18048 roll $31, %eax, etc.
18050 PR tree-optimization/45216
18051 PR tree-optimization/57157
18052 * tree-ssa-forwprop.c (simplify_rotate): New function.
18053 (ssa_forward_propagate_and_combine): Call it.
18055 2013-05-10 Richard Biener <rguenther@suse.de>
18057 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
18058 disable peeling when we version for aliasing.
18059 (vector_alignment_reachable_p): Honor explicit user alignment.
18060 (vect_supportable_dr_alignment): Likewise.
18061 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
18062 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
18063 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
18064 then peeling to arrange for the cost-model check to come first.
18066 2013-05-10 Alan Modra <amodra@gmail.com>
18068 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
18069 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
18070 * configure: Regenerate.
18072 2013-05-10 Alan Modra <amodra@gmail.com>
18075 * varasm.c (default_elf_select_section): Move !DECL_P check..
18076 (get_named_section): ..to here before calling get_section_name.
18078 (default_section_type_flags): Add DECL_P check.
18079 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
18080 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
18082 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
18084 * config/epiphany/epiphany.c (epiphany_expand_prologue):
18085 When using gen_stack_adjust_str with a register offset, add a
18086 REG_FRAME_RELATED_EXPR note.
18088 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
18090 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
18091 (*vec_extractv4si_zext_mem): Ditto.
18092 (*vec_extractv2di): Add 0->x and x->x alternatives.
18093 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
18094 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
18096 2013-05-09 Jason Merrill <jason@redhat.com>
18098 N3639 C++1y VLA support
18099 * gimplify.c (gimplify_vla_decl): Don't touch an existing
18102 * tree.c (build_constructor_va): New.
18103 * tree.h: Declare it.
18105 2013-05-09 Martin Jambor <mjambor@suse.cz>
18108 * gimple-fold.c (canonicalize_constructor_val): Call
18109 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
18111 2013-05-09 Jan Hubicka <jh@suse.cz>
18112 Richard Biener <rguenther@suse.de>
18115 * symtab.c (symtab_make_decl_local): Do not add private names.
18117 2013-05-09 Jan Hubicka <jh@suse.cz>
18120 * symtab.c (insert_to_assembler_name_hash): Handle clones.
18121 (unlink_from_assembler_name_hash): Likewise.
18122 (symtab_prevail_in_asm_name_hash, symtab_register_node,
18123 symtab_unregister_node, symtab_initialize_asm_name_hash,
18124 change_decl_assembler_name): Update.
18126 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
18128 * config/aarch64/aarch64.md: New movtf split.
18129 (*movtf_aarch64): Update.
18130 (aarch64_movdi_tilow): Handle TF modes and rename to
18131 aarch64_movdi_<mode>low.
18132 (aarch64_movdi_tihigh): Handle TF modes and rename to
18133 aarch64_movdi_<mode>high
18134 (aarch64_movtihigh_di): Handle TF modes and rename to
18135 aarch64_mov<mode>high_di
18136 (aarch64_movtilow_di): Handle TF modes and rename to
18137 aarch64_mov<mode>low_di
18138 (aarch64_movtilow_tilow): Remove spurious whitespace.
18139 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
18141 (aarch64_print_operand): Update.
18143 2013-05-09 Alan Modra <amodra@gmail.com>
18145 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
18147 * configure: Regenerate.
18149 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18151 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
18152 splitter preparation statements.
18153 * config/i386/sse.md (*vec_extract* splitters): Ditto.
18154 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
18157 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18159 * gimple-ssa-strength-reduction.c (count_candidates): Change
18160 return value to int.
18161 (analyze_candidates_and_replace): Change type of length to int.
18163 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18165 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
18166 (*vec_extract<mode>): Use VI12_128 mode iterator.
18167 (*vec_extract<mode>_mem): Ditto.
18168 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
18171 2013-05-08 Diego Novillo <dnovillo@google.com>
18176 2012-08-17 Diego Novillo <dnovillo@google.com>
18179 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18180 * config.in: Regenerate.
18181 * configure: Regenerate.
18182 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18184 2013-05-08 Jan Hubicka <jh@suse.cz>
18187 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18188 * cgraph.h (symtab_node_base): Add unique_name.
18189 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18190 input_overwrite_node, input_varpool_node): Stream unique_name.
18191 * cgraphclones.c (cgraph_create_virtual_clone,
18192 cgraph_function_versioning): Set unique_name.
18193 * ipa.c (function_and_variable_visibility): Set unique_name.
18195 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18197 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18198 (alloc_cand_and_find_basis): Restrict conditional candidate
18199 processing to CAND_MULTs.
18201 2013-05-08 Jan Hubicka <jh@suse.cz>
18204 lto-symtab.c (lto_symtab_symbol_p): New function.
18205 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18206 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18207 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18208 Skip static symbols.
18210 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18212 PR tree-optimization/57200
18213 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18214 Only call inform if the preceding warning_at returns true.
18216 2013-05-07 Han Shen <shenhan@google.com>
18218 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18219 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18220 * common.opt (fstack-protector-strong): New option.
18221 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18222 * doc/invoke.texi (Optimization Options): Document
18223 "-fstack-protector-strong".
18224 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18226 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18228 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18230 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18232 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18233 it is not lhs of the insn.
18234 (setup_reg_equiv): Remove insn having equiv memory which it is not
18236 * lra-constraints.c (process_address): Try to improve generation
18237 code for address base + disp.
18238 (lra_constraints): Make correct the code for checking insn setting
18239 up backward equivalence. Remove insn only if it is in the init
18241 * lra-eliminations.c (update_reg_eliminate): Change return value.
18242 (lra_eliminate): Use the result.
18244 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18246 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18247 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18248 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18249 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18250 register target operands.
18251 (*vec_extractv8hi_sse2): New pattern.
18252 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18253 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18254 (*vec_extract<mode>_mem): New insn and split pattern.
18256 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18258 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18259 (TARGET_ASAN_SHADOW_OFFSET): Define.
18260 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18261 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18263 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18265 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18266 (incr_vec_index): Return -1 if increment not found.
18267 (create_add_on_incoming_edge): Assert if increment not found.
18268 (record_increment): Limit number of increments recorded.
18269 (all_phi_incrs_profitable): Return false if an increment not found.
18270 (replace_profitable_candidates): Don't process increments that were
18272 (analyze_candidates_and_replace): Limit size of incr_vec.
18274 2013-05-07 Richard Biener <rguenther@suse.de>
18276 * calls.c (special_function_p): setjmp-like functions are leaf.
18277 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18278 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18280 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18282 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18283 (aarch64_simd_mov<mode>): New expander.
18284 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18285 (aarch64_simd_mov_to_<mode>high): Likewise.
18286 (aarch64_simd_mov_from_<mode>low): Likewise.
18287 (aarch64_simd_mov_from_<mode>high): Likewise.
18288 (aarch64_dup_lane<mode>): Update.
18289 (aarch64_dup_lanedi): New instruction pattern.
18290 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18291 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18293 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18295 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18296 (replace_mult_candidate): Remove unnecessary argument; remove
18297 unnecessary parameter from call to introduce_cast_before_cand.
18298 (replace_unconditional_candidate): Remove unnecessary parameter
18299 from call to replace_mult_candidate.
18300 (replace_conditional_candidate): Likewise.
18301 (insert_initializers): Use make_temp_ssa_name.
18302 (introduce_cast_before_cand): Remove unnecessary argument; use
18303 make_temp_ssa_name.
18304 (replace_one_candidate): Remove unnecessary argument; remove
18305 unnecessary parameter from calls to introduce_cast_before_cand.
18306 (replace_profitable_candidates): Remove unnecessary parameters
18307 from calls to replace_one_candidate.
18309 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18311 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18312 phi def as possibly hiding a basis for a CAND_ADD whose operands
18313 have been commuted in the analysis.
18314 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18316 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18318 * config/aarch64/aarch64.md
18319 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18320 shift value between 0-4.
18322 2013-05-07 Richard Biener <rguenther@suse.de>
18324 * double-int.h (rshift): New overload.
18325 * double-int.c (rshift): New function.
18326 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18327 (create_reference_ops_from_ref): Remove.
18328 (vn_reference_insert): Use shared ops for constructing the
18329 reference and copy it.
18331 2013-05-07 Richard Biener <rguenther@suse.de>
18333 PR middle-end/57190
18334 * tree-eh.c (sink_clobbers): Properly propagate
18335 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18337 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18339 PR tree-optimization/57149
18340 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18341 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18342 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18343 uninit_undefined_value_p instead of ssa_undefined_value_p.
18346 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18347 for modifier == EXPAND_INITIALIZER.
18349 2013-05-07 Anton Blanchard <anton@samba.org>
18351 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18352 for powerpc64 little endian.
18353 * configure: Regenerate.
18355 2013-05-06 Graham Stott <grahams@btinternet.com>
18357 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18358 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18359 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18360 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18362 2013-05-06 Graham Stott <grahams@btinternet.com>
18364 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18365 codes which allow non-lvalues.
18367 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18369 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18370 components are all 1s.
18371 (integer_minus_onep): New function.
18372 * tree.h (integer_minus_onep): Declare it.
18373 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18374 integer_minus_onep instead of integer_all_onesp.
18376 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18379 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18380 variations of these patterns.
18382 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18384 * config/i386/i386.md (isa): Add x64_sse4 member.
18385 (enabled): Handle x64_sse4.
18386 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18387 instruction for 64bit SSE4_1 targets. Update insn attributes.
18388 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18389 instruction for SSE4_1 targets. Update insn attributes.
18390 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18391 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18392 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18394 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18395 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18396 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18398 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18401 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18403 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18405 * final.c (do_assembler_dialects): Don't handle curly braces and
18406 vertical bar escaped by % as dialect delimiters.
18407 (output_asm_insn): Print curly braces and vertical bar if escaped
18408 by % and ASSEMBLER_DIALECT defined.
18409 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18410 * doc/tm.texi: Regenerated.
18412 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18414 * config/mips/mips.c: Include tree-pass.h.
18415 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18416 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18417 (pass_mips_machine_reorg2): New machine specific pass.
18418 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18419 (mips_option_override): Register the new pass.
18420 * rtl.h (cleanup_barriers): Remove prototype.
18421 (dbr_schedule): Likewise.
18422 * jump.c (cleanup_barriers): Make static.
18423 * reorg.c (dbr_schedule): Likewise.
18425 2013-05-06 Richard Biener <rguenther@suse.de>
18427 PR tree-optimization/57185
18428 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18429 SSA names properly.
18431 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18434 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18435 between "first_arg" and "insn", not "last" and "insn".
18437 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18439 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18440 (find_candidates_in_block): Re-enable slsr_process_phi.
18441 (create_phi_basis): Fix double counting of candidate adjustment.
18443 2013-05-06 Richard Biener <rguenther@suse.de>
18445 PR middle-end/57147
18446 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18447 the edge is also fallthru, preserve it and just clear the
18449 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18450 also complex, preserve that and just clear the fallthru flag.
18451 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18452 update virtual operands.
18454 2013-05-06 Alan Modra <amodra@gmail.com>
18456 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18457 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18458 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18459 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18460 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18461 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18463 2013-05-06 Alan Modra <amodra@gmail.com>
18465 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18466 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18467 (DEFAULT_ASM_ENDIAN): Define.
18468 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18469 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18470 Update -K PIC clause from sysv4.h.
18471 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18472 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18474 2013-05-06 Alan Modra <amodra@gmail.com>
18476 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18477 twice for little-endian.
18478 (ashrdi3_no_power, ashrdi3): Support little-endian.
18480 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18483 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18484 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18486 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18487 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18488 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18491 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18492 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18494 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18495 * config/pa/pa.opt: Make mbig-switch a no-op.
18496 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18497 (CASE_VECTOR_MODE): Always return SImode.
18498 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18499 for the !TARGET_BIG_SWITCH case.
18500 * config/pa/pa-linux.h: Likewise.
18501 * config/pa/pa-openbsd.h: Likewise.
18502 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18503 * config/pa/pa.md (short_jump): Remove define_insn.
18504 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18505 (casesi0): Remove define_insn.
18506 (type): Remove btable_branch.
18507 (pa_combine_type): Likewise.
18508 (in_nullified_branch_delay): Likewise.
18509 (in_call_delay): Likewise.
18510 (define_delay): Likewise.
18511 (define_insn_reservation "Z3"): Likewise.
18512 (define_insn_reservation "Z4"): Likewise.
18513 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18514 (pa_adjust_insn_length): Remove adjustment for btable branches.
18515 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18518 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18520 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18521 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18522 Add m->r,x alternatives.
18523 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18524 splitters using SWI48x mode iterator.
18525 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18526 TARGET_64BIT. Add m->x alternative.
18527 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18528 Add o->x alternative. Enable for TARGET_SSE.
18529 (sse_storeq): Remove expander.
18530 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18531 with memory input operand.
18532 (*vec_extractv2di_1 splitter): New.
18533 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18534 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18536 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18538 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18539 (INT_LOWPART): Delete.
18540 (extract_MB): Adjust.
18541 (extract_ME): Adjust.
18542 (print_operand): Adjust.
18544 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18546 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18547 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18548 (reg_or_logical_cint_operand, easy_fp_constant,
18549 logical_const_operand): Delete "CONST_DOUBLE" case.
18550 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18551 "HOST_BITS_PER_WIDE_INT == 64" test.
18552 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18553 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18554 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18555 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18557 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18558 CONST_DOUBLE DImode/VOIDmode case.
18559 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18560 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18561 CONST_DOUBLE VOIDmode case.
18562 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18563 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18564 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18565 Delete CONST_DOUBLE case.
18566 (splitters for mov FMOVE64 const_double): Delete
18567 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18568 "HOST_BITS_PER_WIDE_INT >= 64" test.
18569 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18571 (mov DI const_double): Delete.
18573 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18575 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18576 on op shows all bits zero in mode of a lowpart subreg, return zero.
18578 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18581 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18582 to save TFmode registers and DImode to save TImode registers for
18583 caller save operations.
18584 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18585 mark being partially clobbered since they only use the first
18588 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18589 and TDmode only use the upper 64-bits of each VSX register.
18591 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18593 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18594 (find_candidates_in_block): Disable slsr_process_phi.
18596 2013-05-03 Guozhi Wei <carrot@google.com>
18598 * coverage.c (coverage_obj_init): Move the construction of gcov
18600 (build_init_ctor): ... here.
18602 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18604 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18605 (slsr_cand_d): Redefine def_phi.
18606 (stride_status, phi_adjust_status, count_phis_status): New enums.
18607 (find_phi_def): New.
18608 (find_basis_for_base_expr): New.
18609 (find_basis_for_candidate): Handle hidden bases.
18610 (alloc_cand_and_find_basis): Handle phi candidates.
18611 (slsr_process_phi): New.
18612 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18613 (create_mul_imm_cand): Likewise.
18614 (create_add_ssa_cand): Exclude phi base candidates.
18615 (create_add_imm_cand): Likewise.
18616 (slsr_process_cast): Likewise.
18617 (slsr_process_copy): Likewise.
18618 (find_candidates_in_block): Handle phi candidates.
18619 (dump_candidate): Likewise.
18620 (unconditional_cands): Delete.
18621 (unconditional_cands_with_known_stride_p): Delete.
18622 (phi_dependent_cand_p): New.
18623 (cand_increment): Handle phi-dependent candidates.
18624 (replace_dependent): Delete.
18625 (replace_mult_candidate): New.
18626 (replace_unconditional_candidate): New.
18627 (incr_vec_index): Move to avoid forward reference.
18628 (create_add_on_incoming_edge): New.
18629 (create_phi_basis): New.
18630 (replace_dependents): Delete.
18631 (replace_conditional_candidate): New.
18632 (phi_add_costs): New.
18633 (replace_uncond_cands_and_profitable_phis): New.
18634 (record_increment): Handle phi adjustments.
18635 (record_phi_increments): New.
18636 (record_increments): Handle phi adjustments.
18637 (phi_incr_cost): New.
18638 (lowest_cost_path): Handle phis.
18639 (total_savings): Likewise.
18640 (analyze_increments): Likewise.
18641 (ncd_with_phi): New.
18642 (ncd_of_cand_and_phis): New.
18643 (nearest_common_dominator_for_cands): Handle phi increments.
18644 (all_phi_incrs_profitable): New.
18645 (replace_profitable_candidates): Handle phi-dependent candidates.
18646 (analyze_candidates_and_replace): Likewise.
18648 2013-05-03 Teresa Johnson <tejohnson@google.com>
18651 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18652 do not exceed REG_BR_PROB_BASE.
18654 2013-05-03 Jeff Law <law@redhat.com>
18656 PR tree-optimization/57144
18657 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18658 operand of the condition will bit into the new type when eliminating
18659 a cast feeding a condition.
18661 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18663 PR rtl-optimization/57130
18664 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18665 of COMPARE as in_code to the recursive call if needed.
18667 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18669 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18670 (enabled): Handle new members.
18671 * config/i386/sse.md (*vec_concatv2si): Merge from
18672 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18673 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18675 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18677 PR tree-optimization/57027
18678 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18679 for fnms opportunity, check we got the prerequisite kind
18680 of tree / gimple before using accessor functions.
18682 2013-05-03 Richard Biener <rguenther@suse.de>
18684 * double-int.h (lshift): New overload without precision
18685 and arith argument.
18686 (operator *=, operator +=, operator -=): Move ...
18687 * double-int.c (operator *=, operator +=, operator -=): ... here
18688 and implement more efficiently.
18689 (mul_double_with_sign): Remove.
18690 (lshift_double): Adjust to take unsinged shift argument, push
18691 dispatching code to callers.
18692 (mul_double_wide_with_sign): Add early out for callers that
18693 are not interested in high parts or overflow.
18694 (lshift): New function.
18695 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18696 dispatch code here.
18697 (lrotate, rrotate): Use logical shifts.
18698 * expr.c (get_inner_reference): Use lshift.
18699 * fixed-value.c (do_fixed_divide): Likewise.
18700 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18701 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18702 (indirect_refs_may_alias_p): Likewise.
18703 (stmt_kills_ref_p_1): Likewise.
18705 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18707 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18709 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18711 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18712 scalar form of FABD instruction.
18714 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18716 * lra-constraints.c (process_alt_operands): Add checking alt
18717 number to choose the best alternative.
18719 2013-05-02 Richard Biener <rguenther@suse.de>
18721 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18722 bitmap and its handling.
18723 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18725 2013-05-02 Richard Biener <rguenther@suse.de>
18727 PR middle-end/57140
18728 * tree-inline.c (copy_loops): Properly handle removed loops.
18729 (copy_cfg_body): Mark destination loops for fixup if source
18730 loops needed fixup.
18732 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18735 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18736 generating simple_return for naked functions.
18738 2013-05-02 Martin Jambor <mjambor@suse.cz>
18740 PR middle-end/56988
18741 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18742 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18744 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18745 ipa_agg_replacement_value structures.
18746 (known_aggs_to_agg_replacement_list): Likewise.
18747 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18748 (read_agg_replacement_chain): Likewise.
18749 (ipcp_transform_function): Also check that by_ref flags match.
18751 2013-05-02 Richard Biener <rguenther@suse.de>
18753 * graphds.h (struct graph): Add obstack member.
18754 * graphds.c (new_graph): Initialize obstack and allocate
18756 (add_edge): Allocate edge from the obstack.
18757 (free_graph): Free the obstack instead of all edges and vertices.
18759 2013-05-02 Teresa Johnson <tejohnson@google.com>
18761 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18763 * cfg.c (update_bb_profile_for_threading): Ditto.
18764 * tree-inline.c (copy_bb): Ditto.
18765 (copy_edges_for_bb): Ditto.
18766 (initialize_cfun): Ditto.
18767 (copy_cfg_body): Ditto.
18768 (expand_call_inline): Ditto.
18769 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18770 (estimate_node_size_and_time): Ditto.
18771 (inline_merge_summary): Ditto.
18772 * cgraphclones.c (cgraph_clone_edge): Ditto.
18773 (cgraph_clone_node): Ditto.
18774 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18775 (compute_trg_info): Ditto.
18777 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18779 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18780 S reg when fp attribute set.
18781 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18783 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18785 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18787 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18788 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18789 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18791 2013-05-02 Richard Biener <rguenther@suse.de>
18793 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18794 (struct instantiate_cache_entry): New type.
18795 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18796 (struct instantiate_cache_type): New type.
18797 (set_instantiated_value, get_instantiated_value): Remove.
18798 (get_instantiated_value_entry): New function.
18799 (instantiate_scev_name): Use the new cache and adjust.
18800 (instantiate_scev_poly): Adjust.
18801 (instantiate_scev_binary): Likewise.
18802 (instantiate_array_ref): Likewise.
18803 (instantiate_scev_convert): Likewise.
18804 (instantiate_scev_not): Likewise.
18805 (instantiate_scev_3): Likewise.
18806 (instantiate_scev_2): Likewise.
18807 (instantiate_scev_r): Likewise.
18808 (instantiate_scev): Likewise.
18809 (resolve_mixers): Likewise.
18811 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18814 * lra-constraints.c (best_small_class_operands_num): Remove.
18815 (process_alt_operands): Remove small_class_operands_num. Take
18816 small classes operands into losers and only if the operand is not
18817 matched. Modify debugging output.
18818 (curr_insn_transform): Remove best_small_class_operands_num.
18821 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18823 * config/aarch64/aarch64-builtins.c
18824 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18825 * config/aarch64/aarch64-simd-builtins.def
18826 (reduc_splus_): Add new modes.
18827 (reduc_uplus_): New.
18828 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18829 (reduc_uplus_v4sf): Likewise.
18830 (reduc_splus_v4sf): Likewise.
18831 (aarch64_addv<mode>): Likewise.
18832 (reduc_uplus_<mode>): Likewise.
18833 (reduc_splus_<mode>): Likewise.
18834 (aarch64_addvv2di): Likewise.
18835 (reduc_uplus_v2di): Likewise.
18836 (reduc_splus_v2di): Likewise.
18837 (aarch64_addvv2si): Likewise.
18838 (reduc_uplus_v2si): Likewise.
18839 (reduc_splus_v2si): Likewise.
18840 (reduc_<sur>plus_<mode>): New.
18841 (reduc_<sur>plus_v2di): Likewise.
18842 (reduc_<sur>plus_v2si): Likewise.
18843 (reduc_<sur>plus_v4sf): Likewise.
18844 (aarch64_addpv4sf): Likewise.
18845 * config/aarch64/arm_neon.h
18846 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
18847 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
18848 add UNSPEC_SADDV, UNSPEC_UADDV.
18850 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
18852 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18854 * config/aarch64/arm_neon.h
18855 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
18857 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18859 * config/aarch64/aarch64-builtins
18860 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
18862 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18864 * config/aarch64/aarch64-simd-builtins.def
18865 (reduc_smax_): New.
18866 (reduc_smin_): Likewise.
18867 (reduc_umax_): Likewise.
18868 (reduc_umin_): Likewise.
18869 (reduc_smax_nan_): Likewise.
18870 (reduc_smin_nan_): Likewise.
18873 (smax): Update for V2SF, V4SF and V2DF modes.
18876 (smin_nan): Likewise.
18877 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
18878 (<su><maxmin><mode>3): ...This, refactor.
18879 (s<maxmin><mode>3): New.
18880 (<maxmin_uns><mode>3): Likewise.
18881 (reduc_<maxmin_uns>_<mode>): Refactor.
18882 (reduc_<maxmin_uns>_v4sf): Likewise.
18883 (reduc_<maxmin_uns>_v2si): Likewise.
18884 (aarch64_<fmaxmin><mode>: Remove.
18885 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
18887 (vmin<q>_f<32,64>): Likewise.
18888 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18890 (su): Add mappings for smax, smin, umax, umin.
18892 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
18893 (FMAXMIN): Rename as...
18894 (FMAXMIN_UNS): ...This.
18896 (fmaxminv): Likewise.
18897 (fmaxmin): Likewise.
18899 (maxmin_uns_op): Likewise.
18901 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18903 * config/aarch64/arm_neon.h
18904 (vac<ge, gt><sd>_f<32, 64>): Rename to...
18905 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
18906 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
18908 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18910 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
18911 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
18913 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18915 * config/aarch64/aarch64-simd.md
18916 (vcond<mode>_internal): Handle special cases for constant masks.
18917 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
18918 (vcondu<mode><mode>): Likewise.
18919 (vcond<v_cmp_result><mode>): New.
18921 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18923 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
18924 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
18925 * config/aarch64/aarch64-simd-builtins.def
18926 (cmeq): Update to BUILTIN_VALLDI.
18931 * config/aarch64/arm_neon.h
18932 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
18933 to builtins or C as appropriate.
18935 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18937 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
18939 (cmhi): Rename to...
18941 * config/aarch64/aarch64-simd.md
18942 (simd_mode): Add SF.
18943 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
18944 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
18945 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
18946 (cstore<mode>_neg): ...This.
18947 * config/aarch64/iterators.md
18949 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
18950 (COMPARISONS): New.
18951 (UCOMPARISONS): Likewise.
18952 (optab): Add missing comparisons.
18958 (VCMP_S): Likewise.
18959 (VCMP_U): Likewise.
18960 (V_cmp_result): Add DF, SF modes.
18961 (v_cmp_result): Likewise.
18963 (vmtype): Likewise.
18964 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
18966 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
18968 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
18969 define_insn to define_insn_and_split.
18970 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
18971 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
18972 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
18973 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
18974 (thumb2_negscc): Likewise.
18976 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18978 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
18980 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
18982 * config/arm/thumb2.md: Remove trailing whitespaces.
18984 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
18986 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
18987 Use gen_int_mode rather than GEN_INT.
18989 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
18991 * value-prof.c (stream_in_histogram_value): Remove the strayed
18994 2013-04-30 Richard Biener <rguenther@suse.de>
18996 PR middle-end/57122
18997 * cfghooks.c (split_edge): Properly check for the loop latch edge.
18999 2013-04-30 Richard Biener <rguenther@suse.de>
19001 PR middle-end/57107
19002 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
19004 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
19006 PR rtl-optimization/56957
19007 PR rtl-optimization/57105
19008 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
19009 variable. Use just INSN_UID for determining whether an insn
19010 should be only disconnected from the insn stream.
19011 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
19013 2013-04-30 Jakub Jelinek <jakub@redhat.com>
19015 PR tree-optimization/57104
19016 * tsan.c (instrument_expr): Don't instrument accesses to
19017 DECL_HARD_REGISTER VAR_DECLs.
19019 2013-04-30 Richard Biener <rguenther@suse.de>
19021 * function.h (loops_for_fn): New inline function.
19022 (set_loops_for_fn): Likewise.
19023 * cfgloop.h (place_new_loop): Add struct function parameter.
19024 (get_loop): Likewise.
19025 (get_loops): Likewise.
19026 (number_of_loops): Likewise.
19027 (fel_next): Adjust.
19028 (fel_init): Likewise.
19029 * cfg.c (get_loop_copy): Adjust.
19030 * cfgloop.c (flow_loops_dump): Likewise.
19031 (record_loop_exits): Likewise.
19032 (verify_loop_structure): Likewise.
19033 * cfgloopanal.c (mark_irreducible_loops): Likewise.
19034 (estimate_reg_pressure_cost): Likewise.
19035 (mark_loop_exit_edges): Likewise.
19036 * cfgloopmanip.c (place_new_loop): Likewise.
19037 (add_loop): Likewise.
19038 (duplicate_loop): Likewise.
19039 * graph.c (draw_cfg_nodes): Likewise.
19040 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
19041 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
19042 (extract_affine_chrec): Likewise.
19043 (build_scop_iteration_domain): Likewise.
19044 * graphite.c (graphite_initialize): Likewise.
19045 * ira-build.c (create_loop_tree_nodes): Likewise.
19046 (more_one_region_p): Likewise.
19047 (rebuild_regno_allocno_maps): Likewise.
19048 (mark_loops_for_removal): Likewise.
19049 (mark_all_loops_for_removal): Likewise.
19050 (remove_unnecessary_regions): Likewise.
19051 (ira_build): Likewise.
19052 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19053 * loop-init.c (fix_loop_structure): Likewise.
19054 (gate_rtl_move_loop_invariants): Likewise.
19055 (gate_rtl_unswitch): Likewise.
19056 (gate_rtl_unroll_and_peel_loops): Likewise.
19057 (rtl_doloop): Likewise.
19058 * lto-streamer-in.c (input_cfg): Likewise.
19059 * lto-streamer-out.c (output_cfg): Likewise.
19060 * modulo-sched.c (sms_schedule): Likewise.
19061 * predict.c (tree_estimate_probability): Likewise.
19062 (tree_estimate_probability_driver): Likewise.
19063 (estimate_loops): Likewise.
19064 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
19065 (move_sese_region_to_fn): Likewise.
19066 (debug_loop_num): Likewise.
19067 * tree-chrec.c (chrec_evaluate): Likewise.
19068 (hide_evolution_in_other_loops_than_loop): Likewise.
19069 (chrec_component_in_loop_num): Likewise.
19070 (reset_evolution_in_loop): Likewise.
19071 (evolution_function_is_invariant_rec_p): Likewise.
19072 * tree-if-conv.c (main_tree_if_conversion): Likewise.
19073 * tree-inline.c (copy_loops): Likewise.
19074 (copy_cfg_body): Likewise.
19075 (tree_function_versioning): Likewise.
19076 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
19077 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
19079 (add_to_evolution_1): Likewise.
19080 (scev_const_prop): Likewise.
19081 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
19082 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
19083 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
19084 (tree_ssa_lim_initialize): Likewise.
19085 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
19086 (verify_loop_closed_ssa): Likewise.
19087 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
19088 (tree_ssa_loop_im): Likewise.
19089 (tree_ssa_loop_unswitch): Likewise.
19090 (tree_vectorize): Likewise.
19091 (check_data_deps): Likewise.
19092 (tree_ssa_loop_ivcanon): Likewise.
19093 (tree_ssa_loop_bounds): Likewise.
19094 (tree_complete_unroll): Likewise.
19095 (tree_complete_unroll_inner): Likewise.
19096 (tree_parallelize_loops): Likewise.
19097 (tree_ssa_loop_prefetch): Likewise.
19098 (tree_ssa_loop_ivopts): Likewise.
19099 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
19100 * tree-vectorizer.c (vectorize_loops): Likewise.
19102 2013-04-29 Mike Frysinger <vapier@gentoo.org>
19104 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
19105 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
19106 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
19107 with EABI_LINK_SPEC.
19109 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19112 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
19115 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
19118 * lra-constraints.c (process_alt_operands): Discourage a bit more
19119 using memory for pseudos. Print cost dump for alternatives.
19120 Modify cost values for conflicts with early clobbers.
19121 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
19123 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19126 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
19128 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19130 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
19131 from/to S register.
19132 (movdi_aarch64): Support LDR/STR from/to D register.
19134 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19136 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
19137 or higher by default.
19139 2013-04-29 Richard Biener <rguenther@suse.de>
19141 PR middle-end/57075
19142 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
19143 even if not adding abnormal edges for calls that can make
19146 2013-04-29 Richard Biener <rguenther@suse.de>
19148 PR middle-end/57103
19149 * tree-cfg.c (move_stmt_op): Fix condition under which to update
19151 (move_stmt_r): Remove redundant checking.
19153 2013-04-29 Teresa Johnson <tejohnson@google.com>
19156 * basic-block.h (apply_scale): New function.
19157 (apply_probability): Use apply_scale.
19158 * gimple-streamer-in.c (input_bb): Ditto.
19159 * lto-streamer-in.c (input_cfg): Ditto.
19160 * lto-cgraph.c (merge_profile_summaries): Ditto.
19161 * tree-optimize.c (execute_fixup_cfg): Ditto.
19162 * tree-inline.c (copy_bb): Update comment to use apply_scale.
19163 (copy_edges_for_bb): Ditto.
19164 (copy_cfg_body): Ditto.
19166 2013-04-29 Tom de Vries <tom@codesourcery.com>
19168 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
19169 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
19170 (tail_merge_optimize): Handle current_loops == NULL.
19172 2013-04-26 Jeff Law <law@redhat.com>
19174 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
19175 (simplify_cond_using_ranges): Generalize code to simplify
19176 COND_EXPRs where one argument is a constant and the other
19177 is an SSA_NAME created by an integral type conversion.
19179 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19181 * config/arm/arm.md (store_minmaxsi): Use only when
19182 optimize_insn_for_size_p.
19184 2013-04-29 Christian Bruel <christian.bruel@st.com>
19187 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19189 2013-04-29 Richard Biener <rguenther@suse.de>
19191 PR middle-end/57089
19192 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19193 loop tree make sure to schedule a fixup for the child as well.
19194 (expand_omp_for_generic): Properly add loops.
19195 (expand_omp_for_static_nochunk): Likewise.
19196 (expand_omp_for_static_chunk): Likewise.
19197 (expand_omp_for): For the degenerate case fixup loops.
19198 (expand_omp_sections): Fix default bb placement in loops.
19199 (expand_omp_atomic_pipeline): Properly add loops.
19201 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19203 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19205 2013-04-29 Tom de Vries <tom@codesourcery.com>
19207 * tree-ssa-tail-merge.c: Update header comment.
19209 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19211 * config/aarch64/arm_neon.h
19212 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19213 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19214 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19215 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19216 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19217 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19218 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19219 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19221 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19223 * config/aarch64/aarch64-simd.md
19224 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19225 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19226 fix_trunc, fixuns_trunc.
19227 (ftrunc<VDQF:mode>2): New.
19228 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19229 (fix_trunc_optab): New.
19231 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19233 * config/aarch64/aarch64-builtins.c
19234 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19235 iceilf, lround, iroundf.
19237 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19240 * config/i386/i386.h (enum ix86_tune_indices)
19241 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19242 New, split from X86_TUNE_INTER_UNIT_MOVES.
19243 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19244 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19245 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19246 (TARGET_INTER_UNIT_MOVES): Remove.
19247 * config/i386/i386.c (initial_ix86_tune_features): Update.
19248 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19249 (ix86_expand_convert_uns_didf_sse): Use
19250 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19251 (ix86_expand_vector_init_one_nonzero): Ditto.
19252 (ix86_expand_vector_init_interleave): Ditto.
19253 (inline_secondary_memory_needed): Return true for moves from SSE class
19254 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19255 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19256 * config/i386/constraints.md (Yi, Ym): Depend on
19257 TARGET_INTER_UNIT_MOVES_TO_VEC.
19258 (Yj, Yn): New constraints.
19259 * config/i386/i386.md (*movdi_internal): Change constraints of
19260 operand 1 from Yi to Yj and from Ym to Yn.
19261 (*movsi_internal): Ditto.
19262 (*movdf_internal): Ditto.
19263 (*movsf_internal): Ditto.
19264 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19265 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19266 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19267 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19268 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19269 * config/i386/sse.md (movdi_to_sse): Ditto.
19270 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19271 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19272 TARGET_INTER_UNIT_MOVES.
19273 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19274 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19275 instead of TARGET_INTER_UNIT_MOVES.
19276 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19277 operand 1 from Yi to Yj and from Ym to Yn.
19279 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19281 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19282 (float_truncate_hi_): Likewise.
19283 (float_extend_lo_): Likewise.
19284 (float_truncate_lo_): Likewise.
19285 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19286 (aarch64_float_extend_lo_v2df): Likewise.
19287 (vec_unpacks_hi_v4sf): Likewise.
19288 (aarch64_float_truncate_lo_v2sf): Likewise.
19289 (aarch64_float_truncate_hi_v4sf): Likewise.
19290 (vec_pack_trunc_v2df): Likewise.
19291 (vec_pack_trunc_df): Likewise.
19293 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19295 * config/aarch64/aarch64-builtins.c
19296 (aarch64_fold_builtin): Fold float conversions.
19297 * config/aarch64/aarch64-simd-builtins.def
19298 (floatv2si, floatv4si, floatv2di): New.
19299 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19300 * config/aarch64/aarch64-simd.md
19301 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19302 * config/aarch64/iterators.md (FLOATUORS): New.
19303 (optab): Add float, floatuns.
19304 (su_optab): Likewise.
19306 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19308 * config/aarch64/aarch64-builtins.c
19309 (aarch64_builtin_vectorized_function): Use new names for
19311 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19312 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19313 (fcvtzu): Split as...
19314 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19315 (fcvtas): Split as...
19316 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19317 (fcvtau): Split as...
19318 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19319 (fcvtps): Split as...
19320 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19321 (fcvtpu): Split as...
19322 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19323 (fcvtms): Split as...
19324 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19325 (fcvtmu): Split as...
19326 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19327 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19328 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19329 (lfrintnusf, lfrintnudf): Likewise.
19330 * config/aarch64/aarch64-simd.md
19331 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19333 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19334 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19335 (fcvt_pattern): Likewise.
19337 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19339 * config/aarch64/aarch64-simd.md
19340 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19341 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19343 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19345 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19346 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19347 (vrnd<a,m,n,p>_f32): Implement using builtins.
19348 (vrnd<i,x><q>_f<32, 64>): New.
19350 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19352 * config/aarch64/aarch64-builtins.c
19353 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19354 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19355 (frintz): Rename to...
19357 (frintp): Rename to...
19359 (frintm): Rename to...
19361 (frinti): Rename to...
19362 (nearbyint): ...this.
19363 (frintx): Rename to...
19365 (frinta): Rename to...
19367 * config/aarch64/aarch64-simd.md
19368 (aarch64_frint<frint_suffix><mode>): Delete.
19369 (<frint_pattern><mode>2): Convert to insn.
19370 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19371 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19372 (frint_pattern): Likewise.
19373 (frint_suffix): Likewise.
19375 2013-04-29 Richard Biener <rguenther@suse.de>
19377 PR tree-optimization/57081
19378 * loop-init.c: Include tree-flow.h.
19379 (loop_optimizer_finalize): Free number of iteration estimates.
19380 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19382 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19384 PR tree-optimization/57083
19385 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19386 non-singleton shift count range, zero extend low_bound for uns case.
19388 * config/i386/predicates.md (general_vector_operand): New predicate.
19389 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19390 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19391 if they aren't nonimmediate operands. If their original values
19392 satisfy const_vector_equal_evenodd_p, don't shift them.
19393 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19394 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19395 if not nonimmediate_operand.
19396 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19397 instead of register_operand.
19398 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19400 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19402 * stor-layout.c (finalize_size_functions): Allocate a structure and
19403 reset cfun before dumping the functions.
19405 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19407 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19410 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19411 use xop_pmacsdqh if uns_p.
19412 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19413 the immediate rotate count.
19415 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19417 * rtl.h (struct rtx_def): Add comment for field jump.
19418 (LRA_SUBREG_P): New macro.
19419 * recog.c (register_operand): Check LRA_SUBREG_P.
19420 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19422 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19423 (lra_final_code_change): Skip subreg change for operators.
19424 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19425 if there are no operand changes.
19426 * lra-constraints.c (curr_insn_set): New.
19427 (match_reload): Set LRA_SUBREG_P.
19428 (emit_spill_move): Ditto.
19429 (check_and_process_move): Use curr_insn_set. Process only single
19430 set insns. Don't initialize sec_mem_p and change_p.
19431 (simplify_operand_subreg): Use LRA_SUBREG_P.
19432 (reg_in_class_p): New function.
19433 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19434 of #ifdef. Add code to remove cycling.
19435 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19436 non-null disp. Reload inner instead of disp when base and index
19437 are null. Try to put lo_sum into register.
19438 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19439 (check_and_process_move): Move code for move cost check to
19440 simple_move_p. Remove equiv_substitution.
19441 (simple_move_p): New function.
19442 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19443 curr_insn_set. Call check_and_process_move only for single set
19444 insns. Use the new function. Move call of check_and_process_move
19445 after operand equiv substitution and address process.
19447 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19450 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19451 with nonlocal goto receivers or returns twice calls, ignore
19452 unininitialized values from abnormal edges to nl goto receiver
19453 or returns twice call.
19455 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19457 PR tree-optimization/57051
19458 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19459 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19462 2013-04-26 Richard Biener <rguenther@suse.de>
19464 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19465 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19466 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19467 (fixup_loop_arrays_after_move): New function.
19468 (move_sese_region_to_fn): Properly outline the loop tree parts
19469 of the SESE region.
19471 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19473 * config/i386/i386.md (type, unit): Fix long lines.
19475 2013-04-26 Richard Biener <rguenther@suse.de>
19477 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19478 (lto-streamer-out.o): Likewise.
19479 * cfgloop.c (init_loops_structure): Export, add struct function
19480 argument and adjust.
19481 (flow_loops_find): Adjust.
19482 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19483 (init_loops_structure): Declare.
19484 * lto-streamer-in.c: Include cfgloop.h.
19485 (input_cfg): Input the loop tree.
19486 * lto-streamer-out.c: Include cfgloop.h.
19487 (output_cfg): Output the loop tree.
19488 (output_struct_function_base): Do not drop PROP_loops.
19490 2013-03-26 Richard Biener <rguenther@suse.de>
19492 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19493 (pass_build_cfg): Provide PROP_loops.
19494 (move_sese_region_to_fn): Remove loops that are outlined into fn
19496 * tree-inline.c: Include cfgloop.h.
19497 (initialize_cfun): Do not drop PROP_loops.
19498 (copy_loops): New function.
19499 (copy_cfg_body): Copy loop structure.
19500 (tree_function_versioning): Initialize destination loop tree.
19501 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19502 (pass_parallelize_loops): Do IL verification.
19503 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19504 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19505 the CFG make sure we fixup loops as well.
19506 * tree-ssa-tail-merge.c: Include cfgloop.h.
19507 (replace_block_by): When merging loop latches mark loops for fixup.
19508 * lto-streamer-out.c (output_struct_function_base): Drop
19509 PROP_loops for now.
19510 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19511 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19512 * ipa-split.c: Include cfgloop.h.
19513 (split_function): Add the new return block to the loop tree root.
19514 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19515 whether we have removed the forwarder block.
19516 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19517 * cfgloop.h (place_new_loop): Declare.
19518 * cfgloopmanip.c (place_new_loop): Export.
19519 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19520 (tree-switch-conversion.o): Likewise.
19521 (tree-complex.o): Likewise.
19522 (tree-inline.o): Likewise.
19523 (tree-ssa-tailmerge.o): Likewise.
19524 (ipa-split.o): Likewise.
19525 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19526 (tree-ssa-copy.o): Likewise.
19527 * tree-switch-conversion.c: Include cfgloop.h
19528 (process_switch): If we emit a bit-test cascade, schedule loops
19530 * tree-complex.c: Include cfgloop.h.
19531 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19532 * asan.c: Include cfgloop.h.
19533 (create_cond_insert_point): Properly add new basic-blocks to
19534 loops, schedule loop fixup.
19535 * cfgloop.c (verify_loop_structure): Check that looks are not
19537 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19539 (expand_omp_for_generic): Likewise.
19540 (expand_omp_sections): Likewise.
19541 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19542 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19543 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19544 is initialized, not when loops are present.
19545 * tree-parloops.c (parallelize_loops): Remove checking here.
19546 * passes.c (init_optimization_passes): Schedule a copy-propagation
19547 pass before complete unrolling of inner loops.
19549 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19551 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19552 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19553 (colorize_init): Add argument to _WIN32 version.
19554 * toplev.c: Include diagnostic-color.h.
19555 (process_options): Default to -fdiagnostics-color=auto if
19556 GCC_COLORS env var is in the environment.
19557 * common.opt (fdiagnostics-color=): Add Var and Init.
19558 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19559 env var is in the environment, the default is auto rather than never.
19561 * diagnostic.h (file_name_as_prefix): Add context argument.
19562 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19563 the string as locus.
19564 * langhooks.c (lhd_print_error_function): Adjust caller.
19566 2013-04-25 Lawrence Crowl <crowl@google.com>
19568 * var-tracking.c (shared_hash_def::htab):
19569 Change type to hash_table. Update dependent calls and types.
19571 2013-04-25 Lawrence Crowl <crowl@google.com>
19573 * Makefile.in: Update as needed below.
19575 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19576 Move declaration to after the type's method definitons.
19578 * attribs.c (htab_t scoped_attributes::attribute_hash):
19579 Change type to hash_table. Update dependent calls and types.
19581 * bitmap.c (htab_t bitmap_desc_hash):
19582 Change type to hash_table. Update dependent calls and types.
19584 * cselib.c (htab_t cselib_hash_table):
19585 Change type to hash_table. Update dependent calls and types.
19587 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19588 (hash_string_slot_node): Move implementation into lto-streamer.h
19589 struct string_slot_hasher.
19590 (eq_string_slot_node): Likewise.
19592 * data-streamer-out.c: Update output_block::string_hash_table
19593 dependent calls and types.
19595 * dwarf2cfi.c (htab_t trace_index):
19596 Change type to hash_table. Update dependent calls and types.
19598 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19599 Change type to hash_table. Update dependent calls and types.
19600 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19601 (htab_t optimize_external_refs::map): Likewise.
19602 (htab_t output_comp_unit::extern_map): Likewise.
19603 (htab_t output_comdat_type_unit::extern_map): Likewise.
19604 (htab_t output_macinfo::macinfo_htab): Likewise.
19605 (htab_t optimize_location_lists::htab): Likewise.
19606 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19608 * except.c (htab_t ehspec_hash_type):
19609 Change type to hash_table. Update dependent calls and types.
19610 (assign_filter_values::ttypes): Likewise.
19611 (assign_filter_values::ehspec): Likewise.
19612 (sjlj_assign_call_site_values::ar_hash): Likewise.
19613 (convert_to_eh_region_ranges::ar_hash): Likewise.
19615 * gcse.c (htab_t pre_ldst_table):
19616 Change type to hash_table. Update dependent calls and types.
19618 * ggc-common.c (htab_t saving_htab):
19619 Change type to hash_table. Update dependent calls and types.
19620 (htab_t loc_hash): Likewise.
19621 (htab_t ptr_hash): Likewise.
19622 (call_count): Rename ggc_call_count.
19623 (call_alloc): Rename ggc_call_alloc.
19624 (loc_descriptor): Rename make_loc_descriptor.
19625 (add_statistics): Rename ggc_add_statistics.
19627 * ggc-common.c (saving_htab):
19628 Change type to hash_table. Update dependent calls and types.
19630 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19631 (push_gimplify_context): Likewise.
19632 (pop_gimplify_context): Likewise.
19633 (struct gimple_temp_hash_elt): Added.
19634 (struct gimplify_hasher): Likewise.
19635 (struct gimplify_ctx.temp_htab):
19636 Change type to hash_table. Update dependent calls and types.
19638 * gimple-fold.c: Include gimplify-ctx.h.
19640 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19641 Change type to hash_table. Update dependent calls and types.
19642 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19643 avoid potential global name collision.
19645 * gimplify.c: Include gimplify-ctx.h.
19646 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19647 (htab_t gimplify_ctx::temp_htab):
19648 Update dependent calls and types for new type hash_table.
19649 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19650 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19652 * gimplify-ctx.h: New.
19653 (struct gimple_temp_hash_elt): Move from gimplify.c.
19654 (class gimplify_hasher): New.
19655 (struct gimplify_ctx): Move from gimple.h.
19656 (htab_t gimplify_ctx::temp_htab):
19657 Change type to hash_table. Update dependent calls and types.
19659 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19660 (htab_t ivs_params::newivs_index):
19661 Change type to hash_table. Update dependent calls and types.
19662 (htab_t ivs_params::params_index): Likewise.
19663 (htab_t print_generated_program::params_index): Likewise.
19664 (htab_t gloog::newivs_index): Likewise.
19665 (htab_t gloog::params_index): Likewise.
19667 * graphite.c: Include graphite-htab.h.
19668 4htab_t graphite_transform_loops::bb_pbb_mapping):
19669 Change type to hash_table. Update dependent calls and types.
19671 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19672 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19673 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19675 * graphite-dependences.c: Include graphite-htab.h.
19676 (loop_is_parallel_p): Change hash table type of parameter.
19678 * graphite-htab.h: New.
19679 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19680 (extern find_pbb_via_hash): Move from graphite-poly.h.
19681 (extern loop_is_parallel_p): Move from graphite-poly.h.
19682 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19684 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19685 (extern loop_is_parallel_p): Move to graphite-htab.h.
19686 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19688 * haifa-sched.c (htab_t delay_htab):
19689 Change type to hash_table. Update dependent calls and types.
19690 (htab_t delay_htab_i2): Likewise.
19692 * ira-color.c (htab_t allocno_hard_regs_htab):
19693 Change type to hash_table. Update dependent calls and types.
19695 * ira-costs.c (htab_t cost_classes_htab):
19696 Change type to hash_table. Update dependent calls and types.
19698 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19699 Change type to hash_table. Update dependent calls and types.
19701 * loop-iv.c (htab_t bivs):
19702 Change type to hash_table. Update dependent calls and types.
19704 * loop-unroll.c (htab_t opt_info::insns_to_split):
19705 Change type to hash_table. Update dependent calls and types.
19706 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19708 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19709 (struct string_slot_hasher): New.
19710 (htab_t output_block::string_hash_table):
19711 Change type to hash_table. Update dependent calls and types.
19713 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19714 (htab_t file_name_hash_table):
19715 Change type to hash_table. Update dependent calls and types.
19717 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19720 * lto-streamer.c (htab_t tree_htab):
19721 Change type to hash_table. Update dependent calls and types.
19723 * omp-low.c: Include gimplify-ctx.h.
19725 * passes.c (htab_t name_to_pass_map):
19726 Change type to hash_table. Update dependent calls and types.
19727 (pass_traverse): Rename to passes_pass_traverse.
19729 * plugin.c (htab_t event_tab):
19730 Change type to hash_table. Update dependent calls and types.
19732 * postreload-gcse.c (htab_t expr_table):
19733 Change type to hash_table. Update dependent calls and types.
19734 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19736 * sese.c (debug_rename_map_1): Make extern.
19737 (htab_t copy_bb_and_scalar_dependences::rename_map):
19738 Change type to hash_table. Update dependent calls and types.
19740 * sese.h (extern debug_rename_map): Move to .c file.
19742 * store-motion.c (htab_t store_motion_mems_table):
19743 Change type to hash_table. Update dependent calls and types.
19745 * trans-mem.c (htab_t tm_new_mem_hash):
19746 Change type to hash_table. Update dependent calls and types.
19748 * tree-browser.c (htab_t TB_up_ht):
19749 Change type to hash_table. Update dependent calls and types.
19751 * tree-cfg.c (htab_t discriminator_per_locus):
19752 Change type to hash_table. Update dependent calls and types.
19754 * tree-complex.c: Include tree-hasher.h
19755 (htab_t complex_variable_components):
19756 Change type to hash_table. Update dependent calls and types.
19758 * tree-eh.c (htab_t finally_tree):
19759 Change type to hash_table. Update dependent calls and types.
19761 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19762 struct int_tree_hasher.
19763 (extern int_tree_map_eq): Likewise.
19764 (uid_decl_map_hash): Removed.
19765 (extern decl_tree_map_eq): Likewise.
19767 * tree-hasher.h: New.
19768 (struct int_tree_hasher): New.
19769 (typedef int_tree_htab_type): New.
19771 * tree-inline.c: Include gimplify-ctx.h.
19773 * tree-mudflap.c: Include gimplify-ctx.h.
19775 * tree-parloops.c: Include tree-hasher.h.
19776 (htab_t eliminate_local_variables_stmt::decl_address):
19777 Change type to hash_table. Update dependent calls and types.
19778 (htab_t separate_decls_in_region::decl_copies): Likewise.
19780 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19781 Change type to hash_table. Update dependent calls and types.
19783 * tree-sra.c (candidates):
19784 Change type to hash_table. Update dependent calls and types.
19786 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19788 (int_tree_map_hash): Likewise.
19790 * tree-ssa-dom.c (htab_t avail_exprs):
19791 Change type to hash_table. Update dependent calls and types.
19793 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19794 Change type to hash_table. Update dependent calls and types.
19796 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19797 Change type to hash_table. Update dependent calls and types.
19799 * tree-ssa-phiopt.c (seen_ssa_names):
19800 Change type to hash_table. Update dependent calls and types.
19802 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19803 Change type to hash_table. Update dependent calls and types.
19805 * tree-ssa-uncprop.c (equiv):
19806 Change type to hash_table. Update dependent calls and types.
19808 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19810 PR rtl-optimization/57003
19811 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19812 call note_stores with kill_clobbered_value callback again after
19813 killing regs_invalidated_by_call.
19815 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19817 * config/aarch64/aarch64-simd.md
19818 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19819 (aarch64_simd_bsl<mode>): Likewise.
19820 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19822 2013-04-25 Marek Polacek <polacek@redhat.com>
19824 PR tree-optimization/57066
19825 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19827 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19829 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19831 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19833 * config/aarch64/aarch64-builtins.c
19834 (aarch64_fold_builtin): New.
19835 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19836 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19837 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19838 * config/aarch64/arm_neon.h
19839 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
19841 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19842 Tejas Belagod <tejas.belagod@arm.com>
19844 * config/aarch64/aarch64-builtins.c
19845 (aarch64_gimple_fold_builtin): New.
19846 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
19847 * config/aarch64/aarch64-simd-builtins.def (addv): New.
19848 * config/aarch64/aarch64-simd.md (addpv4sf): New.
19849 (addvv4sf): Update.
19850 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
19852 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19854 * config/aarch64/aarch64.md
19855 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
19857 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19859 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
19860 (*ngcsi_uxtw): New pattern.
19862 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19863 Julian Brown <julian@codesourcery.com>
19865 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
19866 (TB_DREG): Add T_V4HF.
19867 (v4hf_UP): New macro.
19868 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19869 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19870 Handle initialisation of V4HF. Adjust initialisation of reinterpret
19872 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
19873 (arm_vector_mode_supported_p): Handle V4HF.
19874 (arm_mangle_map): Handle V4HFmode.
19875 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
19876 * config/arm/arm_neon_builtins.def: Add entries for
19877 vcvtv4hfv4sf, vcvtv4sfv4hf.
19878 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
19879 (neon_vcvtv4hfv4sf): Likewise.
19880 * config/arm/neon-gen.ml: Handle half-precision floating point
19882 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
19883 * config/arm/arm_neon.h: Regenerate.
19884 * config/arm/neon.ml (type elts): Add F16.
19885 (type vectype): Add T_float16x4, T_floatHF.
19886 (type vecmode): Add V4HF.
19887 (type features): Add Requires_FP_bit feature.
19888 (elt_width): Handle F16.
19889 (elt_class): Likewise.
19890 (elt_of_class_width): Likewise.
19891 (mode_of_elt): Refactor.
19892 (type_for_elt): Handle F16, fix error messages.
19893 (vectype_size): Handle T_float16x4.
19894 (vcvt_sh): New function.
19895 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
19896 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
19897 (string_of_mode): Handle V4HF.
19898 * doc/arm-neon-intrinsics.texi: Regenerate.
19900 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19902 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
19903 format specifier in 'X' case.
19905 2013-04-25 Alan Modra <amodra@gmail.com>
19908 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
19909 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
19910 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
19911 Repeat for many other rotate/shift and mask patterns using subregs.
19912 Name lshiftrt insns.
19913 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
19914 on WORDS_BIG_ENDIAN.
19916 2013-04-25 Alan Modra <amodra@gmail.com>
19918 * config.gcc: Support little-endian powerpc-linux targets.
19919 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
19920 (LINK_OS_LINUX_SPEC): Define.
19921 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
19922 Preserve MASK_LITTLE_ENDIAN.
19923 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
19924 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
19925 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
19926 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
19927 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
19928 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
19929 Correct fp word order for little-endian. Don't shift toc entries
19930 smaller than a word for little-endian.
19931 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
19932 (bswapdi2 splits): Correct low-part subreg for little-endian.
19933 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
19934 low/high where such is correct only for be.
19935 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
19936 little-endian for -mcall-aixdesc.
19938 2013-04-25 Alan Modra <amodra@gmail.com>
19940 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
19941 replace_equiv_address_nv.
19943 2013-04-25 Alan Modra <amodra@gmail.com>
19945 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
19947 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19950 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19951 * rtl.h (struct rtx_def): ...
19953 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19955 PR rtl-optimizations/57046
19956 * lra-constraints (split_reg): Set up lra_risky_transformations_p
19957 for multi-reg splits.
19959 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
19961 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
19963 2013-04-24 Sterling Augustine <saugustine@google.com>
19965 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
19966 (comp_dir_string, debug_str_dwo_section): New.
19967 (DEBUG_STR_DWO_SECTION): Rename to ...
19968 (DEBUG_DWO_STR_SECTION): ... this.
19969 (DEBUG_NORM_STR_SECTION): Delete.
19970 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
19971 (DEBUG_STR_DWO_SECTION_FLAGS): New.
19972 (find_AT_string): Move most logic to ...
19973 (find_AT_string_in_table): ... here. New.
19974 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
19975 add_skeleton_AT_string. Delete logic.
19976 (output_skeleton_debug_sections): Remove call to
19977 add_top_level_skeleton_die_attrs.
19978 (add_comp_dir_attribute): Move logic to comp_dir_string.
19979 (dwarf2out_init): Initialize debug_str_dwo_section.
19980 (output_indirect_string): Call find_string_form.
19981 (output_indirect_strings): Rewrite.
19982 (prune_unused_types): Empty skeleton_debug_str_hash.
19983 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
19984 (dwarf2out_finish): Call output_indirect_strings.
19986 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
19988 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
19990 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
19992 * rtl.h (struct rtx_def): Add comment for field jump.
19993 (LRA_SUBREG_P): New macro.
19994 * recog.c (register_operand): Check LRA_SUBREG_P.
19995 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19997 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19998 (lra_final_code_change): Skip subreg change for operators.
19999 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
20000 if there are no operand changes.
20001 * lra-constraints.c (curr_insn_set): New.
20002 (match_reload): Set LRA_SUBREG_P.
20003 (emit_spill_move): Ditto.
20004 (check_and_process_move): Use curr_insn_set. Process only single
20005 set insns. Don't initialize sec_mem_p and change_p.
20006 (simplify_operand_subreg): Use LRA_SUBREG_P.
20007 (reg_in_class_p): New function.
20008 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
20009 of #ifdef. Add code to remove cycling.
20010 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
20011 non-null disp. Reload inner instead of disp when base and index
20012 are null. Try to put lo_sum into register.
20013 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20014 (check_and_process_move): Move code for move cost check to
20015 simple_move_p. Remove equiv_substitution.
20016 (simple_move_p): New function.
20017 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20018 curr_insn_set. Call check_and_process_move only for single set
20019 insns. Use the new function. Move call of check_and_process_move
20020 after operand equiv substitution and address process.
20022 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
20024 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
20025 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
20026 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
20028 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20030 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
20032 2013-04-24 Marek Polacek <polacek@redhat.com>
20034 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
20035 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
20036 (select_loops_exit_conditions): Likewise.
20037 (number_of_iterations_for_all_loops): Likewise.
20038 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
20039 (scev_analysis): Likewise.
20041 2013-04-02 Catherine Moore <clm@codesourcery.com>
20042 Chao-ying Fu <fu@mips.com>
20044 * config/mips/micromips.md (jraddiusp): New pattern.
20045 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
20046 instruction if possible.
20048 2013-04-24 Alan Modra <amodra@gmail.com>
20050 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
20052 2013-04-24 Julian Brown <julian@codesourcery.com>
20053 Chung-Lin Tang <cltang@codesourcery.com>
20055 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
20056 dependency behavior in enumeration type DIE generation. Add TODO note
20057 to comments about future DW_FORM_sdata/udata re-work of related code.
20059 2013-04-23 Lawrence Crowl <crowl@google.com>
20061 * Makefile.in: Update as needed below.
20063 * hash-table.h (class hash_table):
20064 Correct many methods with parameter types compare_type to the correct
20065 value_type. (Correct code was unlikely to notice the change.)
20066 (hash_table::elements_with_deleted) New.
20067 (class hashtable::iterator): New.
20068 (hashtable::begin()): New.
20069 (hashtable::end()): New.
20070 (FOR_EACH_HASH_TABLE_ELEMENT): New.
20072 * statistics.c (statistics_hashes):
20073 Change type to hash_table. Update dependent calls and types.
20075 * tree-into-ssa.c (var_infos):
20076 Change type to hash_table. Update dependent calls and types.
20078 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
20079 Change type to hash_table. Update dependent calls and types.
20081 * tree-ssa-loop-im.c (struct mem_ref.refs):
20082 Change type to hash_table. Update dependent calls and types.
20084 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
20085 Change type to hash_table. Update dependent calls and types.
20087 * tree-ssa-sccvn.c (vn_tables_s::nary):
20088 Change type to hash_table. Update dependent calls and types.
20089 (vn_tables_s::phis): Likewise.
20090 (vn_tables_s::references): Likewise.
20092 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
20093 (vn_reference_eq): Update parameter and return types.
20095 * tree-ssa-structalias.c (pointer_equiv_class_table):
20096 Change type to hash_table. Update dependent calls and types.
20097 (location_equiv_class_table): Likewise.
20099 * tree-vect-data-refs.c: Consequential changes for making
20100 peeling a hash_table.
20102 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
20103 (destroy_loop_vec_info): Dependent hash_table update.
20105 * tree-vectorizer.h (peeling_htab):
20106 Change type to hash_table. Update dependent calls and types.
20108 2013-04-23 Shiva Chen <shiva0217@gmail.com>
20110 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
20111 to check the register content is equal or not.
20112 * lra-constraints.c (match_reload): Use lra_assign_reg_val
20113 to assign register content record.
20114 * lra-eliminations.c (update_reg_eliminate): Use
20115 lra_update_reg_val_offset to update register content offset.
20116 * lra-int.h (struct lra_reg): Add offset member.
20117 (lra_reg_val_equal_p): New static inline function.
20118 (lra_update_reg_val_offset): New static inline function.
20119 (lra_assign_reg_val): New static inline function.
20120 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
20121 to assign register content record.
20122 (initialize_lra_reg_info_element): Initial offset to zero.
20124 2013-04-23 Catherine Moore <clm@codesourcery.com>
20126 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
20127 operands. Record compression.
20129 2013-04-23 Xinliang David Li <davidxl@google.com>
20131 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
20133 2013-04-23 Richard Biener <rguenther@suse.de>
20135 PR middle-end/57036
20136 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
20137 parameter, only add abnormal goto edges from the copied body
20138 if the call could perform abnormal gotos.
20139 (copy_cfg_body): Adjust.
20141 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
20143 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
20145 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
20147 * coretypes.h (gimple_stmt_iterator): Add struct to make
20150 2013-04-23 Richard Biener <rguenther@suse.de>
20152 PR tree-optimization/57026
20153 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
20154 from SSA names occuring in abnormal PHI nodes.
20156 2013-04-22 Andi Kleen <ak@linux.intel.com>
20158 * lto/lto.c (print_lto_report_1): Fix LTO report names.
20160 2013-04-22 Andi Kleen <ak@linux.intel.com>
20162 * lto/lto.c (print_lto_report_1): Declare early.
20163 (read_cgraph_and_symbols): Call print_lto_report_1 early.
20165 2013-04-22 Andi Kleen <ak@linux.intel.com>
20167 * common.opt (-flto-report-wpa): Add.
20168 * doc/invoke.texi (-flto-report-wpa): Add.
20169 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
20172 2013-04-22 Xinliang David Li <davidxl@google.com>
20174 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
20175 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
20176 * Makefile.in: New dependency
20178 David Daney <ddaney.cavm@gmail.com>
20180 * configure.ac (gcc_cv_as_micromips_support): Use the
20181 --fatal-warnings option.
20182 * configure: Regenerate.
20184 2013-04-22 Marek Polacek <polacek@redhat.com>
20187 * tsan.c (instrument_expr): Don't instrument expression
20188 in case its size is zero.
20190 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20194 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20196 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20198 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20200 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20201 (gimple_stmt_iterator): New typedef.
20202 * gimple.h (gimple_stmt_iterator): Rename to...
20203 (gimple_stmt_iterator_d): ... This.
20204 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20205 trees be valid for GIMPLE and GENERIC.
20206 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20207 * gimple-fold.c (gimple_fold_call): Call target hook
20208 gimple_fold_builtin.
20209 * hooks.c (hook_bool_gsiptr_false): New.
20210 * hooks.h (hook_bool_gsiptr_false): New.
20211 * target.def (fold_stmt): New.
20212 * doc/tm.texi: Regenerate.
20214 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20217 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20218 a set sp if no stack realignment.
20220 2013-04-22 Nick Clifton <nickc@redhat.com>
20222 * config.gcc (tilegx-linux): Extend extra_objs rather than
20224 (tilepro-linux): Likewise.
20226 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20228 * config/aarch64/aarch64-builtins.c
20230 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20231 (VAR<1-12>): Add MAP parameter.
20232 (BUILTIN_*): Likewise.
20233 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20234 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20235 (aarch64_ushl_n<mode>): Likewise.
20236 (aarch64_sshr_n<mode>): Likewise.
20237 (aarch64_ushr_n<mode>): Likewise.
20238 (aarch64_<maxmin><mode>): Likewise.
20239 (aarch64_sqrt<mode>): Likewise.
20240 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20241 (vshr<q>_n_*): Likewise.
20243 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20245 * config/aarch64/aarch64-builtins.c
20246 (aarch64_simd_builtin_type_mode): Handle SF types.
20248 (BUILTIN_GPF): Define.
20249 (aarch64_init_simd_builtins): Handle SF types.
20250 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20251 (frecps): Likewise.
20252 (frecpx): Likewise.
20253 * config/aarch64/aarch64-simd.md
20254 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20255 (aarch64_frecpe<mode>): New.
20256 (aarch64_frecps<mode>): Likewise.
20257 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20258 (v8type): Add frecp<esx>.
20259 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20260 (aarch64_frecps<mode>): Likewise.
20261 * config/aarch64/iterators.md (FRECP): New.
20262 (frecp_suffix): Likewise.
20263 * config/aarch64/arm_neon.h
20264 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20266 2013-04-22 Christian Bruel <christian.bruel@st.com>
20269 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20270 (REG_CLASS_NAMES): Idem.
20271 (REG_CLASS_CONTENTS): Idem.
20272 (REGCLASS_HAS_FP_REG): Idem.
20273 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20274 (sh_conditional_register_usage): Idem.
20276 2013-04-21 Jeff Law <law@redhat.com>
20278 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20279 (ssa_forward_propagate_and_combine): Use it.
20281 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20283 * lra.c: Update the flow chart diagram.
20285 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20287 PR rtl-optimization/56847
20288 * lra-constraints.c (process_alt_operands): Discourage alternative
20289 with non-matche doffsettable memory constraint fro memory with
20292 2013-04-19 Richard Biener <rguenther@suse.de>
20294 PR tree-optimization/56982
20295 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20297 * gimplify.c (gimplify_call_expr): Notice special calls.
20298 (gimplify_modify_expr): Likewise.
20299 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20300 abnormal control flow receivers.
20301 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20302 in the same way as cfun->has_nonlocal_labels.
20303 (gimple_purge_dead_abnormal_call_edges): Likewise.
20304 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20305 receivers start a basic-block.
20307 2013-04-19 Richard Biener <rguenther@suse.de>
20309 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20311 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20312 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20313 (SLP_TREE_LOAD_PERMUTATION): Add.
20314 (vect_transform_slp_perm_load): Adjust prototype.
20315 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20316 (vect_free_slp_instance): Likewise.
20317 (vect_create_new_slp_node): Likewise.
20318 (vect_supported_slp_permutation_p): Remove.
20319 (vect_slp_rearrange_stmts): Adjust.
20320 (vect_supported_load_permutation_p): Likewise. Inline
20321 vect_supported_slp_permutation_p here.
20322 (vect_analyze_slp_instance): Compute load permutations per
20323 slp node instead of per instance.
20324 (vect_get_slp_defs): Adjust.
20325 (vect_transform_slp_perm_load): Likewise.
20326 (vect_schedule_slp_instance): Remove redundant code.
20327 (vect_schedule_slp): Remove hack for PR56270, add it ...
20328 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20329 CSE loads for SLP. Adjust.
20331 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20333 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20334 spelling in two comments.
20336 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20339 * config/arm/arm.c (load_multiple_sequence): Require SP
20340 as base register for loads if SP is in the register list.
20342 2013-04-19 Martin Jambor <mjambor@suse.cz>
20344 PR tree-optimization/56718
20345 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20346 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20347 and made public. Adjusted all callers.
20348 (ipa_intraprocedural_devirtualization): New function.
20349 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20350 (ipa_intraprocedural_devirtualization): Likewise.
20351 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20353 2013-04-19 Richard Biener <rguenther@suse.de>
20355 PR tree-optimization/57000
20356 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20358 2013-04-19 Terry Guo <terry.guo@arm.com>
20360 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20362 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20363 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20364 (cortex_m4_fmacs): Use new reservations.
20365 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20367 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20369 PR rtl-optimization/56999
20370 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20372 (lra_coalesce): Remove split_origin_bitmap and related code.
20373 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20374 ranges if necessary.
20376 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20378 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20380 (ix86_expand_call): Remove clobbered_registers array and use
20381 x86_64_ms_sysv_extra_clobbered_registers instead.
20382 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20384 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20386 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20387 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20388 (*call_value_rex64_ms_sysv): Ditto.
20390 2013-04-18 Cary Coutant <ccoutant@google.com>
20392 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20393 parent instead of die_mark.
20395 2013-04-18 Diego Novillo <dnovillo@google.com>
20397 * gimple.c (create_gimple_tmp): New.
20398 (get_expr_type): New.
20399 (build_assign): New.
20400 (build_type_cast): New.
20401 * gimple.h (enum ssa_mode): Define.
20402 (gimple_seq_set_location): New.
20403 * asan.c (build_check_stmt): Change some gimple_build_* calls
20404 to use build_assign and build_type_cast.
20406 2013-04-18 Richard Biener <rguenther@suse.de>
20408 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20409 handle negative step. Remove redundant checks.
20410 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20411 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20412 for negative step and grouped loads fail to vectorize.
20414 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20416 * emit-rtl.c (reset_insn_used_flags): New function.
20417 (reset_all_used_flags): Use it.
20418 (verify_insn_sharing): New function.
20419 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20421 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20423 PR tree-optimization/56984
20424 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20425 and (x >> M) >= N don't register any assertion if N << M is the
20428 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20430 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20431 do not delete_insn insns that have not yet been emitted, only
20432 unlink them with remove_insn.
20433 * df-scan.c (df_insn_delete): Revert r197492.
20435 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20437 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20438 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20440 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20442 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20443 define_insn_and_split.
20444 (and_scc,ior_scc,negscc): Likewise.
20445 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20447 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20449 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20450 can benefit from using a sequence of LDRD instructions in epilogue
20451 instead of a single LDM instruction.
20453 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20456 * doc/extend.texi: Fix typo.
20458 2013-04-17 Richard Biener <rguenther@suse.de>
20460 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20461 (vect_build_slp_tree): ... here.
20462 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20463 match. Remove special-casing of mismatched complex loads.
20464 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20465 re-try the match with swapped commutative operands.
20466 (vect_supported_load_permutation_p): Remove special-casing of
20467 mismatched complex loads.
20468 (vect_analyze_slp_instance): Adjust.
20470 2013-04-17 Richard Biener <rguenther@suse.de>
20472 PR rtl-optimization/56921
20473 * cfgloop.h (struct loop): Add simple_loop_desc member.
20474 (struct niter_desc): Mark with GTY(()).
20475 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20476 * loop-iv.c (get_simple_loop_desc): Likewise.
20477 (free_simple_loop_desc): Likewise.
20480 2013-04-16 Richard Biener <rguenther@suse.de>
20482 PR rtl-optimization/56921
20483 * loop-init.c (pass_rtl_move_loop_invariants): Add
20484 TODO_do_not_ggc_collect to todo_flags_finish.
20485 (pass_rtl_unswitch): Same.
20486 (pass_rtl_unroll_and_peel_loops): Same.
20487 (pass_rtl_doloop): Same.
20489 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20491 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20492 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20493 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20495 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20496 * tree-streamer.c (record_common_node): Adjust reference in comment.
20498 2013-04-17 Terry Guo <terry.guo@arm.com>
20500 * config/arm/cortex-m4.md: Add a new bypass.
20502 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20504 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20506 (*subs_<optab><mode>_multp2): New pattern.
20507 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20508 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20510 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20512 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20513 (*subs_mul_imm_<mode>): New pattern.
20515 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20518 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20519 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20520 (vsx_movti_32bit): Change j->wa to O->wa.
20522 2013-04-16 Richard Biener <rguenther@suse.de>
20524 PR rtl-optimization/56921
20525 * loop-init.c (pass_rtl_move_loop_invariants): Add
20526 TODO_do_not_ggc_collect to todo_flags_finish.
20527 (pass_rtl_unswitch): Same.
20528 (pass_rtl_unroll_and_peel_loops): Same.
20529 (pass_rtl_doloop): Same.
20531 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20533 * config/arm/arm.c (emit_multi_reg_push): New declaration
20534 for an existing function.
20535 (arm_emit_strd_push): New function.
20536 (arm_expand_prologue): Used here.
20537 (arm_emit_ldrd_pop): New function.
20538 (arm_expand_epilogue): Used here.
20539 (arm_get_frame_offsets): Update condition.
20540 (arm_emit_multi_reg_pop): Add a special case for load of a single
20541 register with writeback.
20543 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20545 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20548 2013-04-16 Richard Biener <rguenther@suse.de>
20550 PR tree-optimization/56756
20551 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20552 (first_mem_ref_loc): New.
20553 (execute_sm): Place the load temporarily before a previous
20554 access instead of in the latch edge to ensure its SSA dependencies
20555 are defined at points dominating the load.
20557 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20559 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20560 correct fix by moving header and footer insn to the footer of
20561 the merged basic block. Clear BB_END of the merged-away block.
20563 PR middle-end/43631
20564 * emit-rtl.c (make_note_raw): New function.
20565 (link_insn_into_chain): New static inline function.
20566 (add_insn): Use it.
20567 (add_insn_before, add_insn_after): Factor insn chain linking code...
20568 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20569 using link_insn_into_chain.
20570 (note_outside_basic_block_p): New helper function for emit_note_after
20571 and emit_note_before.
20572 (emit_note_after): Use nobb variant of add_insn_after if the note
20573 should not be contained in a basic block.
20574 (emit_note_before): Use nobb variant of add_insn_before if the note
20575 should not be contained in a basic block.
20576 (emit_note_copy): Use make_note_raw.
20577 (emit_note): Likewise.
20578 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20579 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20580 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20581 the moved barrier the tail of the basic block it follows.
20582 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20584 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20586 PR tree-optimization/56962
20587 * gimple-ssa-strength-reduction.c (record_increment): Only set
20588 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20589 either rhs1 or rhs2 is equal to c->base_expr.
20591 2013-04-15 Richard Biener <rguenther@suse.de>
20593 PR tree-optimization/56933
20594 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20596 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20597 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20598 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20599 dependence check ...
20600 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20602 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20603 GROUP_READ_WRITE_DEPENDENCE.
20605 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20607 * emit-rtl.c (reset_all_used_flags): New function.
20608 (verify_rtl_sharing): Call reset_all_used_flags before and after
20609 performing the checks.
20611 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20613 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20614 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20615 * config/arm/constraints.md (De): New constraint.
20616 * config/arm/neon.md (anddi3_neon): Delete.
20617 (neon_vand<mode>): Expand to standard anddi3 pattern.
20618 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20619 Move earlier in the file.
20620 (neon_inv_logic_op2): Likewise.
20621 (arm_anddi_operand_neon): New predicate.
20623 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20625 * configure.ac (gcc_cv_ld_as_needed): Set
20626 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20627 Use -z ignore, -z record on *-*-solaris2*.
20628 (HAVE_LD_AS_NEEDED): Update comment.
20629 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20630 * configure: Regenerate.
20631 * config.in: Regenerate.
20632 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20633 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20634 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20635 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20636 equivalents. Fix markup.
20637 * doc/tm.texi: Regenerate.
20639 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20641 * config/i386/i386.opt: New option mstack-protector-guard=.
20642 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20643 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20644 TARGET_SSP_TLS_GUARD.
20645 * config/i386/i386.c (ix86_option_override_internal): Set
20646 ix86_stack_protector_guard.
20647 * config/i386/i386.md (stack_protect_set): Enable for
20648 TARGET_SSP_TLS_GUARD only.
20649 (stack_protect_set_<mode>): Ditto.
20650 (stack_protect_test): Ditto.
20651 (stack_protect_test_<mode>): Ditto.
20652 * doc/invoke.texi (i386 Option): Document.
20654 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20657 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20658 (S_MODES): Set H_MODE bit.
20659 (SF_MODES): Set only S_MODE and SF_MODE bits.
20660 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20661 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20662 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20663 <MODE_FLOAT>: Likewise.
20665 2013-04-15 Joey Ye <joey.ye@arm.com>
20667 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20669 2013-04-15 Joey Ye <joey.ye@arm.com>
20671 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20673 (thumb_far_jump_used_p): Count instruction size and set
20676 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20678 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20679 * resource.c (mark_target_live_regs): Reformat conditional block.
20681 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20683 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20684 notes, they are emitted only just before final.
20685 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20687 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20689 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20690 * cfgrtl.c (delete_insn): Call it here instead.
20691 * lra-spills.c (lra_final_code_change): Use delete_insn.
20692 * haifa-sched.c (sched_remove_insn): Likewise.
20693 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20694 returning to the nop pool.
20695 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20696 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20698 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20700 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20701 * doc/tm.texi: Regenerated.
20703 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20705 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20708 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20710 * df-core.c (df_find_def): Compare register numbers.
20711 (df_find_use): Likewise.
20713 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20716 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20717 lra_in_progress for return.
20719 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20721 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20722 define_insn into define_insn_and_split and emit movsicc patterns.
20724 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20726 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20728 2013-04-12 Richard Biener <rguenther@suse.de>
20730 * tree-pass.h (TODO_do_not_ggc_collect): New.
20731 * passes.c (execute_one_ipa_transform_pass): Honor
20732 TODO_do_not_ggc_collect.
20733 (execute_one_pass): Likewise.
20736 2013-04-10 Richard Biener <rguenther@suse.de>
20738 * passes.c (init_optimization_passes): Remove reload pass.
20739 * ira.c (do_reload): Merge into ...
20741 (rest_of_handle_reload): Remove.
20742 (pass_reload): Likewise.
20743 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20744 of reload for vzeroupper pass placement.
20746 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20748 PR tree-optimization/56918
20749 PR tree-optimization/56920
20750 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20751 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20752 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20753 use wide_mul_with_sign method.
20755 2013-04-12 Richard Biener <rguenther@suse.de>
20757 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20758 not be considered a gimple constant.
20760 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20762 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20763 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20766 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20767 Jakub Jelinek <jakub@redhat.com>
20769 * opts.c: Include diagnostic-color.h.
20770 (common_handle_option): Handle OPT_fdiagnostics_color_.
20771 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20772 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20773 (diagnostic-color.o): New.
20774 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20775 (diagnostic_color_rule): New enum.
20776 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20777 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20778 the location string.
20779 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20780 either NULL, or color kind.
20781 * diagnostic-color.c: New file.
20782 * diagnostic-color.h: New file.
20783 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20785 * doc/invoke.texi (-fdiagnostics-color): Document.
20786 * pretty-print.h (pp_show_color): Define.
20787 (struct pretty_print_info): Add show_color field.
20788 * diagnostic.c: Include diagnostic-color.h.
20789 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20790 macros. Colorize error:, warning: etc. strings and also the location
20792 (diagnostic_show_locus): Colorize the caret line.
20793 * pretty-print.c: Include diagnostic-color.h.
20794 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20795 inside of %< %> quotes or quoted through q format modifier.
20797 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20799 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20801 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20803 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20804 code in CC_NZ mode.
20805 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20808 2013-04-11 Marek Polacek <polacek@redhat.com>
20810 PR tree-optimization/48184
20811 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20813 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20815 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20816 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20817 (skip_simple_arithmetic): Tidy up.
20818 * tree.h (skip_simple_constant_arithmetic): Declare.
20820 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20822 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20824 2013-04-11 Richard Biener <rguenther@suse.de>
20826 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20827 generate vector constants.
20829 2013-04-11 Richard Biener <rguenther@suse.de>
20831 PR tree-optimization/56878
20832 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20833 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20835 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20836 Prefer to align the DR with the most invariant base address.
20838 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20840 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
20843 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
20845 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
20846 floating-point vector comparisons against 0.
20848 2013-04-11 Jakub Jelinek <jakub@redhat.com>
20850 PR tree-optimization/56899
20851 * fold-const.c (extract_muldiv_1): Apply distributive law
20852 only if TYPE_OVERFLOW_WRAPS (ctype).
20854 2013-04-11 Bin Cheng <bin.cheng@arm.com>
20857 * ira-costs.c (scan_one_insn): Check whether the source rtx of
20858 loading has side effect.
20860 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
20862 * config/sparc/sparc.c: Include tree-pass.h.
20863 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
20864 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
20865 head of file. Change return type. Split off gate function.
20866 (sparc_gate_work_around_errata): New function.
20867 (pass_work_around_errata): New pass definition.
20868 (insert_pass_work_around_errata) New pass insert definition to
20869 insert pass_work_around_errata just after delayed-branch scheduling.
20870 (sparc_option_override): Insert the pass.
20871 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
20873 2013-04-10 David S. Miller <davem@davemloft.net>
20875 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
20876 or -mcpu=hypersparc.
20878 * target.def (cstore_mode): New hook.
20879 * target.h: Include insn-codes.h
20880 * targhooks.c: Likewise.
20881 (default_cstore_mode): New function.
20882 * targhooks.h: Declare it.
20883 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
20884 * doc/tm.texi: Rebuild.
20885 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
20886 target hook, rather than inspecting the insn_data.
20887 * config/sparc/sparc.c (sparc_cstore_mode): New function.
20888 (TARGET_CSTORE_MODE): Redefine.
20889 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
20891 * config/sparc/predicates.md (cstore_result_operand): New special
20893 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
20894 Use it for operand 0.
20895 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
20896 (*snesi_special): Likewise.
20897 (*snesi_zero): Likewise.
20898 (*seqsi_zero): Likewise.
20899 (*sltu_insn): Likewise.
20900 (*sgeu_insn): Likewise.
20901 (*seqdi_special): Make operand 0 and comparison operation be of
20903 (*snedi_special): Likewise.
20904 (*snedi_special_vis3): Likewise.
20905 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
20906 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
20907 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
20908 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
20909 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
20910 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
20911 (*sltu_extend_sp64): Likewise.
20912 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
20913 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
20914 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
20915 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
20916 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
20918 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
20920 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
20921 (aarch64_start_file): Use the new function.
20923 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
20924 Jason Merrill <jason@redhat.com>
20926 * common.opt: Add -gdwarf.
20927 * opts.c (common_handle_option): Handle it.
20928 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
20930 2013-04-10 Richard Biener <rguenther@suse.de>
20932 * passes.c (execute_todo): Do not call ggc_collect conditional here.
20933 (execute_one_ipa_transform_pass): But unconditionally here.
20934 (execute_one_pass): And here.
20935 (init_optimization_passes): Remove reload pass.
20936 * tree-pass.h (TODO_ggc_collect): Remove.
20937 (pass_reload): Likewise.
20938 * ira.c (do_reload): Merge into ...
20940 (rest_of_handle_reload): Remove.
20941 (pass_reload): Likewise.
20942 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20943 of reload for vzeroupper pass placement.
20944 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
20945 and todo_flags_finish of all passes.
20947 2013-04-10 Richard Biener <rguenther@suse.de>
20949 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
20950 first_const_oprnd field, rename first_def_type to first_op_type.
20951 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
20952 (vect_get_and_check_slp_defs): Always use the type of the
20953 operand. Allow mixed vect_external_def, vect_constant_def types.
20954 (vect_get_constant_vectors): Handle mixed vect_external_def,
20955 vect_constant_def types.
20957 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
20959 PR tree-optimization/55524
20960 * tree-ssa-math-opts.c
20961 (convert_mult_to_fma): Don't use an fms construct
20962 when we don't have an fms operation, but fnma, and it looks
20963 likely that we'll be able to use the latter.
20965 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
20967 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
20969 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
20970 inline fail caused by overwritable functions.
20972 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
20974 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
20975 unnecessary bits in the constant power of two case.
20977 2013-04-10 Richard Biener <rguenther@suse.de>
20979 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
20980 broken code swapping operands.
20981 (vect_build_slp_tree): Do not compute load permutations here.
20982 (vect_analyze_slp_instance): Compute load permutations here,
20983 after building the SLP tree.
20985 2013-04-09 Christian Bruel <christian.bruel@st.com>
20987 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
20988 of next/prev_real_insn.
20990 2013-04-09 Jan Hubicka <jh@suse.cz>
20992 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
20993 Drop aliased parameter.
20994 (function_and_variable_visibility): Do not handle alias pairs.
20995 * cgraph.c (varpool_externally_visible_p): Update prototype.
20996 * varpool.c (varpool_add_new_variable): Update.
20998 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21000 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
21002 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
21004 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
21006 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
21008 2013-04-09 Marek Polacek <polacek@redhat.com>
21010 PR tree-optimization/48762
21011 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
21013 2013-04-09 Richard Biener <rguenther@suse.de>
21015 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
21017 (vect_build_slp_tree): Likewise.
21018 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
21019 calculating the cost of a SLP instance.
21020 (vect_analyze_slp_instance): Use it from here, after building
21023 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21025 PR middle-end/56883
21026 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
21027 expand_omp_for_static_chunk): Use simple_p = true in
21028 force_gimple_operand_gsi calls when assigning to addressable decls.
21030 2013-04-09 Jeff Law <law@redhat.com>
21032 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
21033 when the boolean was created by converting a wider object which
21034 had a boolean range.
21036 2013-04-09 Richard Biener <rguenther@suse.de>
21038 * tree-vectorizer.h (slp_void_p): Remove.
21039 (slp_tree): Typedef before _slp_tree declaration.
21040 (struct _slp_tree): Use a vector of slp_tree as children.
21041 (vect_get_place_in_interleaving_chain): Remove.
21042 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
21044 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
21046 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
21047 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
21048 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
21049 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
21050 Use slp_node instead of slp_void_p and adjust.
21052 2013-04-09 Richard Biener <rguenther@suse.de>
21054 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
21055 work that is not necessary.
21057 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21059 PR tree-optimization/56854
21060 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
21061 forward into clobber stmts if it would change MEM_REF lhs into
21064 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
21066 * tree.c (type_hash_lookup, type_hash_add): Make static.
21067 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
21069 2013-04-09 Richard Biener <rguenther@suse.de>
21071 * tree.h (unsave_expr_now): Remove.
21072 * tree-inline.c (mark_local_for_remap_r): Remove.
21073 (unsave_expr_1): Likewise.
21074 (unsave_r): Likewise.
21075 (unsave_expr_now): Likewise.
21076 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
21077 (propagate_tree_value): Likewise.
21079 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
21081 * doc/rtl.texi (sequence): Rewrite documentation to match the
21082 current use of SEQUENCE rtl objects.
21083 * rtl.def (SEQUENCE): Likewise.
21085 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
21086 Update documentation.
21087 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
21088 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
21090 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
21092 2013-04-08 Teresa Johnson <tejohnson@google.com>
21094 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
21095 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
21097 (estimate_edge_size_and_time): Add comment to suggest using rounding
21099 (estimate_node_size_and_time): Ditto.
21100 (remap_edge_change_prob): Use helper rounding divide methods.
21101 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
21102 (gimple_mod_pow2_value_transform): Ditto.
21103 (gimple_mod_subtract_transform): Ditto.
21104 (gimple_ic_transform): Ditto.
21105 (gimple_stringops_transform): Ditto.
21106 * stmt.c (conditional_probability): Ditto.
21107 (emit_case_dispatch_table): Ditto.
21108 * lto-cgraph.c (merge_profile_summaries): Ditto.
21109 * tree-optimize.c (execute_fixup_cfg): Ditto.
21110 * cfgcleanup.c (try_forward_edges): Ditto.
21111 * cfgloopmanip.c (scale_loop_profile): Ditto.
21113 (duplicate_loop_to_header_edge): Ditto.
21114 (lv_adjust_loop_entry_edge): Ditto.
21115 * tree-vect-loop.c (vect_transform_loop): Ditto.
21116 * profile.c (compute_branch_probabilities): Ditto.
21117 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
21118 * lto-streamer-in.c (input_cfg): Ditto.
21119 * gimple-streamer-in.c (input_bb): Ditto.
21120 * ipa-cp.c (update_profiling_info): Ditto.
21121 (update_specialized_profile): Ditto.
21122 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
21123 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
21125 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21126 (compute_trg_info): Ditto.
21127 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
21128 (purge_dead_edges): Ditto.
21129 * loop-unswitch.c (unswitch_loop): Ditto.
21130 * cgraphclones.c (cgraph_clone_edge): Ditto.
21131 (cgraph_clone_node): Ditto.
21132 * tree-inline.c (copy_bb): Ditto.
21133 (copy_edges_for_bb): Ditto.
21134 (initialize_cfun): Ditto.
21135 (copy_cfg_body): Ditto.
21136 (expand_call_inline): Ditto.
21138 2013-04-08 Kai Tietz <ktietz@redhat.com>
21140 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
21141 TARGET_CYGWIN64 by TARGET_64BIT.
21143 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21145 * config/epiphany/epiphany.md (GPR_1): New constant.
21146 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
21147 * config/epiphany/epiphany.c (gen_compare_reg):
21148 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
21149 is already in place.
21150 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
21151 Don't require being called during rtl expansion; If y operlaps r0,
21153 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
21154 (epiphany_expand_epilogue): Likewise.
21156 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
21157 Don't use CC_FPmode for ORDERED / UNORDERED.
21158 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
21160 * config/epiphany/constraints.md (CnL): New constraint.
21161 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
21162 * config/epiphany/predicates.md (add_operand): Allow 1024.
21164 * config/epiphany/epiphany.md (logical_op): New code iterator.
21165 (op_mnc): New code attribute.
21166 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
21167 (mov_f+1, mov_f+2): New peephole2 patterns.
21169 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
21170 (cstoresi4): Also allow re-use of zero result when doing a NE
21171 comparison to a non-zero operand.
21172 Use (clobber (scratch)) for first insn if the gpr output is not needed.
21174 * config/epiphany/epiphany.md (<insn_opname>v2si3):
21175 Use gen_addsi3_i / gen_subsi3_i.
21177 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21181 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21182 contain anything but clobbers, at most one __builtin_stack_restore,
21183 optionally debug stmts and final resx, and if it has at least one
21184 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21185 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21186 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21187 which isn't defaut definition, remove them.
21188 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21189 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21190 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21191 with MEM_REF LHS with SSA_NAME address.
21193 2013-04-08 Jeff Law <law@redhat.com>
21195 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21197 2013-04-08 Richard Biener <rguenther@suse.de>
21199 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21201 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21202 determined vector type.
21203 (vect_analyze_data_refs): Likewise.
21204 (vect_get_new_vect_var): Adjust.
21205 (vect_create_destination_var): Preserve SSA name versions.
21206 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21207 not dump anything here.
21209 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21211 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21212 Add member lr_slot_known.
21213 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21215 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21216 Remove code that sets lr_slot_offset according to what a previous
21217 version of epiphany_emit_save_restore used to do.
21218 (epiphany_emit_save_restore): When doing an lr save or restore,
21219 set/verify lr_slot_known and lr_slot_offset.
21221 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21224 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21227 2013-04-08 Richard Biener <rguenther@suse.de>
21229 * alias.c (find_base_term): Fix thinko in previous change.
21231 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21233 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21234 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21235 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21236 if possible to compute val.
21237 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21238 For QImode integers don't require anything about precision. Use
21239 const_with_all_bytes_same to find out if the constant doesn't have
21240 repeated bytes in it.
21242 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21244 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21247 2013-04-08 Marek Polacek <polacek@redhat.com>
21249 PR rtl-optimization/48182
21250 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21253 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21256 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21257 nuses, make sure we have a label.
21259 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21262 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21263 (rs6000_emit_swdiv_low_precision): Remove.
21264 (rs6000_emit_swdiv): Rewrite to handle between one and four
21265 iterations of Newton-Raphson generally; modify required number of
21266 iterations for some cases.
21267 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21269 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21271 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21272 set-but-unused variable.
21274 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21275 basic blocks of released function bodies garbage-collectable.
21277 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21278 (struct rtl_opt_pass): Add TODO_df_finish.
21280 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21282 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21284 * config/arm/constraints.md (q): New constraint.
21285 * config/arm/ldrdstrd.md: New file.
21286 * config/arm/arm.md (ldrdstrd.md) New include.
21287 (arm_movdi): Use "q" instead of "r" constraint
21288 for double-word memory access.
21289 (movdf_soft_insn): Likewise.
21290 * config/arm/vfp.md (movdi_vfp): Likewise.
21291 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21292 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21293 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21294 (mem_ok_for_ldrd_strd): Likewise.
21295 (output_move_double): Update assertion.
21297 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21299 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21301 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21303 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21304 define_insn_and_split.
21305 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21307 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21309 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21310 define_insn_and_split.
21311 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21312 (shiftsi3_compare): New pattern.
21313 (rrx): New pattern.
21314 * config/arm/unspecs.md (UNSPEC_RRX): New.
21316 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21318 * config/arm/arm.md (negdi_extendsidi): New pattern.
21319 (negdi_zero_extendsidi): Likewise.
21321 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21323 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21324 define_insn_and_split.
21325 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21326 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21328 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21330 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21331 define_insn_and_split.
21332 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21333 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21335 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21337 * config/arm/arm.md (subsi3_carryin): New pattern.
21338 (subsi3_carryin_const): Likewise.
21339 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21340 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21342 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21344 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21346 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21348 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21349 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21351 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21353 * config/arm/arm.c (arm_expand_builtin): Change fcode
21354 type to unsigned int.
21356 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21358 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21360 2013-04-04 Ian Lance Taylor <iant@google.com>
21362 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21365 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21367 PR middle-end/56729
21368 * df-scan.c (df_insn_delete): Disable failing assert.
21370 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21372 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21373 New function prototype.
21374 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21375 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21376 (arm_builtin_vectorized_function): New function.
21378 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21380 * config/arm/arm_neon_builtins.def: New file.
21381 * config/arm/arm.c (neon_builtin_data): Move contents to
21382 arm_neon_builtins.def.
21383 (enum arm_builtins): Include neon builtin definitions.
21384 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21385 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21387 2013-04-04 Marek Polacek <polacek@redhat.com>
21389 PR tree-optimization/48186
21390 * predict.c (maybe_hot_frequency_p): Return false if
21391 HOT_BB_FREQUENCY_FRACTION is 0.
21392 (cgraph_maybe_hot_edge_p): Likewise.
21394 2013-04-04 Richard Biener <rguenther@suse.de>
21396 PR tree-optimization/56826
21397 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21400 2013-04-04 Richard Biener <rguenther@suse.de>
21402 PR tree-optimization/56213
21403 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21404 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21406 2013-04-04 Richard Biener <rguenther@suse.de>
21408 PR tree-optimization/56837
21409 * tree-loop-distribution.c (classify_partition): For non-zero
21410 values require that the value has the same precision as its
21411 mode to be useful as memset value.
21413 2013-04-03 Nick Clifton <nickc@redhat.com>
21415 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21416 (fmssf4): Use fmsf.s on E3V5 architectures.
21417 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21418 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21420 2013-04-03 Jeff Law <law@redhat.com>
21422 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21423 (lra-eliminations.o): Likewise.
21425 2013-04-03 Teresa Johnson <tejohnson@google.com>
21427 * gcov-io.c (compute_working_sets): Moved most of body of old
21428 compute_working_sets here from profile.c.
21429 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21430 (gcov_working_set_t): Moved typedef here from basic-block.h
21431 (compute_working_set): Declare.
21432 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21433 (get_working_sets): Renamed from compute_working_set,
21434 replace most of body with call to new compute_working_sets.
21435 (get_exec_counts): Replace call to compute_working_sets
21436 to get_working_sets.
21437 * profile.h (get_working_sets): Renamed from compute_working_set.
21438 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21439 to get_working_sets.
21440 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21441 * gcov-dump.c (dump_working_sets): New function.
21443 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21445 * hwint.c (sext_hwi, zext_hwi): New functions.
21446 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21447 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21448 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21449 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21450 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21451 (sext_hwi, zext_hwi): New functions.
21453 2013-04-03 Jeff Law <law@redhat.com>
21455 PR tree-optimization/56799
21456 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21457 back test for widening conversion erroneously dropped in prior change.
21459 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21462 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21463 instead of next_real_insn.
21465 2013-04-03 Marek Polacek <polacek@redhat.com>
21468 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21470 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21473 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21475 (thumb1_output_casesi): Likewise.
21476 (thumb2_output_casesi): Likewise.
21478 2013-04-03 Richard Biener <rguenther@suse.de>
21480 PR tree-optimization/56817
21481 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21483 (tree_unroll_loops_completely_1): ... new function to manually
21484 walk the loop tree, properly defering outer loops of unrolled
21485 loops to later iterations.
21487 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21489 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21490 (vectorizable_load): Likewise.
21491 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21492 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21494 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21496 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21499 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21501 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21503 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21505 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21507 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21509 PR tree-optimization/56790
21510 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21513 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21515 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21517 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21518 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21521 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21524 * tree.h (force_folding_builtin_constant_p): New decl.
21525 * builtins.c (force_folding_builtin_constant_p): New variable.
21526 (fold_builtin_constant_p): Fold immediately also if
21527 force_folding_builtin_constant_p.
21529 2013-04-03 Richard Biener <rguenther@suse.de>
21531 PR tree-optimization/56812
21532 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21533 DRs of the same interleaving chain are independent.
21535 2013-04-02 Jason Merrill <jason@redhat.com>
21537 * gdbinit.in (pbb): Use debug fn.
21539 2013-04-02 Lawrence Crowl <crowl@google.com>
21541 * sese.h (struct ivtype_map_elt_s): Remove unused.
21542 (extern debug_ivtype_map): Remove unused.
21543 (extern eq_ivtype_map_elts): Remove unused.
21544 * sese.c (debug_ivtype_map): Removed unused.
21545 (debug_ivtype_map_1): Removed unused.
21546 (debug_ivtype_elt): Remove unused.
21547 (eq_ivtype_map_elts): Remove unused.
21550 2013-04-02 Kai Tietz <ktietz@redhat.com>
21553 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21554 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21555 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21557 (legitimize_pe_coff_symbol): Likewise.
21558 (is_imported_p): New helper-function.
21559 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21560 for Windows x64 targets.
21561 (ix86_expand_prologue): Optimize for pe-coff targets.
21562 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21563 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21564 medium/large code-model.
21565 (legitimize_pic_address): Likewise.
21566 (legitimize_tls_address): Likewise.
21567 (ix86_expand_call): Likewise.
21568 (x86_output_mi_thunk): Likewise.
21569 (get_dllimport_decl): Add new beimport argument.
21570 (construct_plt_address): Don't assert for x64 pe-coff targets.
21571 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21573 (SYMBOL_FLAG_STUBVAR): New macro.
21574 (SYMBOL_REF_STUBVAR_P): Likewise.
21575 * config/i386/winnt.c (stub_list): New structure.
21576 (stub_head): New local variable.
21577 (i386_pe_record_stub): New function.
21578 (i386_pe_file_end): Emit refptr-stubs.
21580 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21582 PR rtl-optimization/56745
21583 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21584 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21587 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21588 and both of them are MEM_REFs, just compare first argument for
21589 equality and attempt to deal even with differing offsets.
21592 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21593 of gimple_clobber_p to be MEM_REF.
21594 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21595 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21596 after gimplification.
21597 * asan.c (get_mem_ref_of_assignment): Don't instrument
21598 gimple_clobber_p stmts.
21599 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21600 gimple_clobber_p stmt if they have MEM_REF lhs and
21601 are dead because of another gimple_clobber_p stmt.
21602 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21603 gimple_clobber_p stmts like debug stmts.
21604 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21605 that refer to unused VAR_DECLs or uninitialized values.
21606 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21607 gimple_clobber_p stmts if they refer to removed parameters.
21608 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21611 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21613 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21614 using SWI48 mode attribute.
21616 2013-04-02 Wei Mi <wmi@google.com>
21618 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21619 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21620 *<rotate_insn><mode>3_mask in i386.md.
21622 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21624 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21626 2013-04-02 Richard Biener <rguenther@suse.de>
21628 PR tree-optimization/56778
21629 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21630 Runtime alias tests are not supported for gather loads.
21631 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21632 stmts referenced from SSA operands before updating SSA form.
21634 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21635 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21637 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21638 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21639 * config/arm/cortex-a53.md: New file.
21640 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21641 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21642 * config/arm/arm.c (arm_issue_rate): Likewise.
21643 * config/arm/arm-tune.md: Regenerate
21644 * config/arm/arm-tables.opt: Regenerate.
21645 * config/arm/arm-cores.def: Add cortex-a53.
21647 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21649 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21652 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21654 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21655 scalar load/store operations using B/H registers.
21656 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21658 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21660 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21662 * config/aarch64/aarch64.c
21663 (aarch64_simd_scalar_immediate_valid_for_move): New.
21664 * config/aarch64/aarch64-protos.h
21665 (aarch64_simd_scalar_immediate_valid_for_move): New.
21666 * config/aarch64/constraints.md (Dh, Dq): New.
21667 * config/aarch64/iterators.md (hq): New.
21669 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21671 * reorg.c (get_branch_condition): Deal with conditional returns.
21672 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21674 2013-04-01 Wei Mi <wmi@google.com>
21676 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21677 Truncate operand 2 using %b asm operand modifier.
21678 (*<shift_insn><mode>3_mask): Ditto.
21679 (*<rotate_insn><mode>3_mask): Ditto.
21681 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21683 PR middle-end/56798
21684 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21686 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21688 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21690 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21692 2013-03-30 Lawrence Crowl <crowl@google.com>
21694 * dse.c (clear_alias_sets): Remove never set.
21695 (disqualified_clear_alias_sets): Remove never set.
21696 (clear_alias_mode_pool): Remove never set.
21697 (dse_step0): Remove condition that is never true.
21698 (canon_address): Remove condition that is never true.
21699 (dse_step7): Remove condition that is never true.
21700 (rest_of_handle_dse): Remove condition that is never true.
21701 (rest_of_handle_dse::did_global): Remove never read from above.
21702 (dse_step2_spill): Remove never called from above.
21703 (dse_step5_spill): Remove never called from above.
21705 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21707 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21708 JUMP_TABLE_DATA changes.
21709 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21710 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21711 (Insns) <jump_table_data>: New entry.
21712 * doc/tm.texi: Regenerate.
21714 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21716 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21717 for table jump at the end of a basic block using tablejump_p.
21718 * targhooks.c (default_invalid_within_doloop): Likewise.
21719 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21720 target hook implementation that is identical to the default hook.
21721 (rs6000_invalid_within_doloop): Remove.
21723 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21724 unused variable from tablejump_p call.
21726 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21727 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21728 (INSN_DELETED_P): Likewise.
21729 (emit_jump_table_data): New prototype.
21730 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21731 after 4th as unused.
21732 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21733 * sched-vis.c (print_insn): Likewise.
21734 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21735 insn for compatibility with back ends that use next_active_insn to
21736 identify jump table data.
21737 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21738 (remove_insn): Likewise.
21739 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21741 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21742 (emit_jump_table_data): New function.
21744 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21745 basic block, a JUMP_TABLE_DATA never is.
21746 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21747 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21748 off from code handling real insns.
21749 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21750 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21751 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21752 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21753 is not a NONDEBUG_INSN_P.
21754 * ira-costs.c (scan_one_insn): Likewise.
21755 * jump.c (mark_all_labels): Likewise.
21756 (mark_jump_label_1): Likewise.
21757 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21758 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21760 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21761 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21762 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21763 JUMP_TABLE_DATA_P insns.
21764 (calculate_elim_costs_all_insns): Likewise.
21765 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21766 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21767 (delete_output_reload): Code style fixups.
21768 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21769 insn flags on this non-insn.
21770 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21771 as scheduling barriers, for pre-change compatibility.
21772 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21773 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21775 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21776 redundant JUMP_TABLE_DATA_P test.
21777 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21778 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21779 (frv_for_each_packet): Likewise.
21780 * config/i386/i386.c (min_insn_size): Likewise.
21781 (ix86_avoid_jump_mispredicts): Likewise.
21782 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21783 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21784 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21785 (mips16_insn_length): Robustify.
21786 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21787 (mips16_split_long_branches): Likewise.
21788 * config/pa/pa.c (pa_combine_instructions): Likewise.
21789 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21790 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21791 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21792 as contributing to pool range lengths.
21793 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21794 Remove redundant JUMP_TABLE_DATA_P test.
21795 (sh_loop_align): Likewise.
21796 (split_branches): Likewise.
21797 (sh_insn_length_adjustment): Likewise.
21798 * config/spu/spu.c (get_branch_target): Likewise.
21800 2013-03-29 Jan Hubicka <jh@suse.cz>
21802 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21803 gcov streaming; stream hot bb threshold to ltrans.
21804 * predict.c (get_hot_bb_threshold): Break out from ....
21805 (maybe_hot_count_p): ... here.
21806 (set_hot_bb_threshold): New function.
21807 * lto-section-in.c (lto_section_name): Add profile.
21808 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21809 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21810 and data-streamer.h
21811 (histogram_entry): New structure.
21812 (histogram, histogram_pool): New global vars.
21813 (histogram_hash): New structure.
21814 (histogram_hash::hash): New method.
21815 (histogram_hash::equal): Likewise.
21816 (account_time_size): New function.
21817 (cmp_counts): New function.
21818 (dump_histogram): New function.
21819 (ipa_profile_generate_summary): New function.
21820 (ipa_profile_write_summary): New function.
21821 (ipa_profile_read_summary): New function.
21822 (ipa_profile): Decide on threshold.
21823 (pass_ipa_profile): Add ipa_profile_write_summary and
21824 ipa_profile_read_summary.
21825 * Makefile.in (ipa.o): Update dependencies.
21826 * lto-streamer.h (LTO_section_ipa_profile): New section.
21828 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21830 * tree.h (VAR_P): New.
21832 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21835 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21837 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
21839 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
21840 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21841 (control_flow_insn_p): Likewise.
21842 * cfgrtl.c (duplicate_insn_chain): Likewise.
21843 * final.c (get_attr_length_1): Likewise.
21844 (shorten_branches): Likewise.
21845 (final_scan_insn): Likewise.
21846 * function.c (instantiate_virtual_regs): Likewise.
21847 * gcse.c (insert_insn_end_basic_block): Likewise.
21848 * ira-costs.c (scan_one_insn): Likewise.
21849 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21850 * lra.c (check_rtl): Likewise.
21851 * reload1.c (elimination_costs_in_insn): Likewise.
21852 * reorg.c (follow_jumps): Likewise.
21854 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
21855 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
21856 (thumb_far_jump_used_p): Likewise.
21857 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
21858 (workaround_speculation): Likewise.
21859 (add_sched_insns_for_speculation): Likewise.
21860 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
21861 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21862 (frv_for_each_packet): Likewise.
21863 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
21864 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
21865 (final_emit_insn_group_barriers): Likewise.
21866 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21867 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21868 (mips16_insn_length): Likewise.
21869 * config/pa/pa.c (pa_reorg): Likewise.
21870 (pa_combine_instructions): Likewise.
21871 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
21872 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
21873 (sh_reorg): Likewise.
21874 (split_branches): Likewise.
21875 * config/spu/spu.c (get_branch_target): Likewise.
21877 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
21880 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
21882 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
21883 Fix declaration name.
21885 2013-03-28 Lawrence Crowl <crowl@google.com>
21887 * graphds.h (struct graph.indicies): Remove unused.
21888 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
21889 (SCOP_ORIGINAL_PDDRS): Remove unused.
21890 * sese.h (extern insert_loop_close_phis): Removed unused.
21891 (extern insert_guard_phis): Removed unused.
21892 (extern ivtype_map_elt_info): Removed unused.
21893 (new_ivtype_map_elt): Removed unused.
21894 * sese.c (ivtype_map_elt_info): Removed unused.
21896 2013-03-28 Lawrence Crowl <crowl@google.com>
21898 * Makefile.in: Add several missing include dependences.
21900 (test-dump.o): New. This object is not added to any executable,
21901 but is present for ad-hoc testing.
21903 (debug (const bitmap_head_def &)): New.
21904 (debug (const bitmap_head_def *)): New.
21906 (extern debug (const bitmap_head_def &)): New.
21907 (extern debug (const bitmap_head_def *)): New.
21909 (debug (edge_def &)): New.
21910 (debug (edge_def *)): New.
21912 (debug (basic_block_def &)): New.
21913 (debug (basic_block_def *)): New.
21915 (dump_node (const_tree, int, FILE *)): Correct source file.
21917 (debug (die_struct &)): New.
21918 (debug (die_struct *)): New.
21920 (extern debug (die_struct &)): New.
21921 (extern debug (die_struct *)): New.
21922 * gimple-pretty-print.c
21923 (debug (gimple_statement_d &)): New.
21924 (debug (gimple_statement_d *)): New.
21925 * gimple-pretty-print.h
21926 (extern debug (gimple_statement_d &)): New.
21927 (extern debug (gimple_statement_d *)): New.
21929 (debug (ira_allocno_copy &)): New.
21930 (debug (ira_allocno_copy *)): New.
21931 (debug (ira_allocno &)): New.
21932 (debug (ira_allocno *)): New.
21934 (extern debug (ira_allocno_copy &)): New.
21935 (extern debug (ira_allocno_copy *)): New.
21936 (extern debug (ira_allocno &)): New.
21937 (extern debug (ira_allocno *)): New.
21939 (debug (live_range &)): New.
21940 (debug (live_range *)): New.
21942 (debug (lra_live_range &)): New.
21943 (debug (lra_live_range *)): New.
21945 (debug (lra_live_range &)): New.
21946 (debug (lra_live_range *)): New.
21948 (debug (omega_pb_d &)): New.
21949 (debug (omega_pb_d *)): New.
21951 (extern debug (omega_pb_d &)): New.
21952 (extern debug (omega_pb_d *)): New.
21954 (debug (const rtx_def &)): New.
21955 (debug (const rtx_def *)): New.
21957 (debug_tree (tree): Move within file.
21958 (debug_raw (const tree_node &)): New.
21959 (debug_raw (const tree_node *)): New.
21960 (dump_tree_via_hooks (const tree_node *, int)): New.
21961 (debug (const tree_node &)): New.
21962 (debug (const tree_node *)): New.
21963 (debug_verbose (const tree_node &)): New.
21964 (debug_verbose (const tree_node *)): New.
21965 (debug_head (const tree_node &)): New.
21966 (debug_head (const tree_node *)): New.
21967 (debug_body (const tree_node &)): New.
21968 (debug_body (const tree_node *)): New.
21969 (debug_vec_tree (tree): Move and reimplement in terms of dump.
21970 (debug (vec<tree, va_gc> &)): New.
21971 (debug (vec<tree, va_gc> *)): New.
21973 (extern debug (const rtx_def &)): New.
21974 (extern debug (const rtx_def *)): New.
21976 (debug_raw (simple_bitmap_def &)): New.
21977 (debug_raw (simple_bitmap_def *)): New.
21978 (debug (simple_bitmap_def &)): New.
21979 (debug (simple_bitmap_def *)): New.
21981 (extern debug (simple_bitmap_def &)): New.
21982 (extern debug (simple_bitmap_def *)): New.
21983 (extern debug_raw (simple_bitmap_def &)): New.
21984 (extern debug_raw (simple_bitmap_def *)): New.
21986 (debug (vinsn_def &)): New.
21987 (debug (vinsn_def *)): New.
21988 (debug_verbose (vinsn_def &)): New.
21989 (debug_verbose (vinsn_def *)): New.
21990 (debug (expr_def &)): New.
21991 (debug (expr_def *)): New.
21992 (debug_verbose (expr_def &)): New.
21993 (debug_verbose (expr_def *)): New.
21994 (debug (vec<rtx> &)): New.
21995 (debug (vec<rtx> *)): New.
21997 (extern debug (vinsn_def &)): New.
21998 (extern debug (vinsn_def *)): New.
21999 (extern debug_verbose (vinsn_def &)): New.
22000 (extern debug_verbose (vinsn_def *)): New.
22001 (extern debug (expr_def &)): New.
22002 (extern debug (expr_def *)): New.
22003 (extern debug_verbose (expr_def &)): New.
22004 (extern debug_verbose (expr_def *)): New.
22005 (extern debug (vec<rtx> &)): New.
22006 (extern debug (vec<rtx> *)): New.
22008 (_list_iter_cond_expr): Make inline instead of static.
22010 (debug (sreal &)): New.
22011 (debug (sreal *)): New.
22013 (extern debug (sreal &)): New.
22014 (extern debug (sreal *)): New.
22016 (extern debug_raw (const tree_node &)): New.
22017 (extern debug_raw (const tree_node *)): New.
22018 (extern debug (const tree_node &)): New.
22019 (extern debug (const tree_node *)): New.
22020 (extern debug_verbose (const tree_node &)): New.
22021 (extern debug_verbose (const tree_node *)): New.
22022 (extern debug_head (const tree_node &)): New.
22023 (extern debug_head (const tree_node *)): New.
22024 (extern debug_body (const tree_node &)): New.
22025 (extern debug_body (const tree_node *)): New.
22026 (extern debug (vec<tree, va_gc> &)): New.
22027 (extern debug (vec<tree, va_gc> *)): New.
22029 (debug (struct loop &)): New.
22030 (debug (struct loop *)): New.
22031 (debug_verbose (struct loop &)): New.
22032 (debug_verbose (struct loop *)): New.
22033 * tree-dump.c: Add header dependence.
22035 (extern debug (struct loop &)): New.
22036 (extern debug (struct loop *)): New.
22037 (extern debug_verbose (struct loop &)): New.
22038 (extern debug_verbose (struct loop *)): New.
22040 (debug (data_reference &)): New.
22041 (debug (data_reference *)): New.
22042 (debug (vec<data_reference_p> &)): New.
22043 (debug (vec<data_reference_p> *)): New.
22044 (debug (vec<ddr_p> &)): New.
22045 (debug (vec<ddr_p> *)): New.
22047 (extern debug (data_reference &)): New.
22048 (extern debug (data_reference *)): New.
22049 (extern debug (vec<data_reference_p> &)): New.
22050 (extern debug (vec<data_reference_p> *)): New.
22051 (extern debug (vec<ddr_p> &)): New.
22052 (extern debug (vec<ddr_p> *)): New.
22054 (debug (pt_solution &)): New.
22055 (debug (pt_solution *)): New.
22057 (extern debug (pt_solution &)): New.
22058 (extern debug (pt_solution *)): New.
22060 (debug (_var_map &)): New.
22061 (debug (_var_map *)): New.
22062 (debug (tree_live_info_d &)): New.
22063 (debug (tree_live_info_d *)): New.
22065 (extern debug (_var_map &)): New.
22066 (extern debug (_var_map *)): New.
22067 (extern debug (tree_live_info_d &)): New.
22068 (extern debug (tree_live_info_d *)): New.
22070 2013-03-28 Jan Hubicka <jh@suse.cz>
22072 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
22074 2013-03-28 Ian Bolton <ian.bolton@arm.com>
22076 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
22077 record only when desired or required.
22079 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
22081 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
22082 *vec_extractv2di_1_rex64. Use x64 isa attribute.
22084 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22086 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
22087 (*andsi3_compare0_uxtw): New pattern.
22088 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
22089 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
22091 2013-03-28 Jan Hubicka <jh@suse.cz>
22093 * data-streamer-in.c (streamer_read_gcov_count): New function.
22094 * gimple-streamer-out.c: Include value-prof.h.
22095 (output_gimple_stmt): Output histogram.
22096 (output_bb): Use streamer_write_gcov_count.
22097 * value-prof.c: Include data-streamer.h
22098 (dump_histogram_value): Add HIST_TYPE_MAX.
22099 (stream_out_histogram_value): New function.
22100 (stream_in_histogram_value): New function.
22101 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
22102 (stream_out_histogram_value, stream_in_histogram_value): Declare.
22103 * data-streamer-out.c (streamer_write_gcov_count): New function.
22104 (streamer_write_gcov_count_stream): New function.
22105 * lto-cgraph.c (lto_output_edge): Update counter streaming.
22106 (lto_output_node): Likewise.
22107 (input_node, input_edge): Likewise.
22108 * lto-streamer-out.c (output_cfg): Update streaming.
22109 * lto-streamer-in.c (input_cfg): Likewise.
22110 * data-streamer.h (streamer_write_gcov_count,
22111 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
22112 * gimple-streamer-in.c: Include value-prof.h
22113 (input_gimple_stmt): Input histograms.
22114 (input_bb): Update profile streaming.
22116 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22118 * genmodes.c (emit_max_int): New function.
22119 (emit_insn_modes_h): Added call to emit_max_function.
22120 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
22122 * machmode.def: Fixed comment.
22124 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22126 * combine.c (try_combine): Removed useless assert.
22127 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
22129 2013-03-28 Marek Polacek <polacek@redhat.com>
22130 Richard Biener <rguenther@suse.de>
22132 PR tree-optimization/56695
22133 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
22134 build signed result of a vector comparison.
22135 * tree-cfg.c (verify_gimple_comparison): Check that a result
22136 of a vector comparison has signed type.
22138 2013-03-28 Richard Biener <rguenther@suse.de>
22140 PR tree-optimization/37021
22141 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
22142 do not restrict gaps between groups.
22143 * tree-vect-stmts.c (vectorizable_load): Properly account for
22144 a gap between groups.
22146 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
22148 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
22149 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
22152 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
22154 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
22155 * doc/extend.texi (Named Address Spaces): Ditto.
22156 (Variable Attributes): Ditto.
22158 2013-03-27 Kai Tietz <ktietz@redhat.com>
22160 * config.build: Add support for cygwin x64 target.
22161 * config.gcc: Likewise.
22162 * config.host: Likewise.
22163 * configure.ac: Likewise
22164 * configure: Regenerated.
22166 2013-03-27 Kai Tietz <ktietz@redhat.com>
22168 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
22169 * config/i386/t-cygwin-w64: New file.
22170 * config/i386/cygwin-w64.h: New file.
22171 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
22172 and add support for x64-cygwin target.
22173 (CPP_SPEC): Likewise.
22174 (CXX_WRAP_SPEC_LIST): Undefine before define.
22175 (LIBGCJ_SONAME): Use 15 as version.
22177 2013-03-27 Richard Biener <rguenther@suse.de>
22179 PR tree-optimization/56716
22180 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22181 dumping for ref nodes.
22183 2013-03-27 Martin Jambor <mjambor@suse.cz>
22185 PR tree-optimization/55334
22186 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22187 restricted pointers to arrays.
22189 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22191 * Makefile.in (.SUFFIXES): Add .cc.
22192 (.c.o): Apply same recipe for implicit rule .cc.o.
22194 2013-03-27 Richard Biener <rguenther@suse.de>
22196 PR tree-optimization/37021
22197 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22198 REALPART/IMAGPART_EXPRs around the supported refs.
22199 * tree-ssa-structalias.c (find_func_aliases): Assume that
22200 floating-point values are not used to transfer pointers.
22202 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22204 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22205 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22206 * doc/tm.texi: Regenerate.
22207 * targhooks.h (default_has_ifunc_p): New.
22208 * targhooks.c (default_has_ifunc_p): Ditto.
22209 * config/linux-protos.h: New file.
22210 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22211 hook for linux which disables support of indirect functions in android.
22212 * config/linux-android.c: New file.
22213 * config/t-linux-android.c: Ditto.
22214 * config.gcc: Added new object file linux-android.o.
22215 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22216 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22217 * varasm.c (do_assemble_alias): Likewise.
22218 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22219 doesn't support indirect functions.
22220 * configure: Regenerate.
22222 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22225 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22226 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22229 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22231 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22233 2013-03-27 Terry Guo <terry.guo@arm.com>
22235 * config/arm/arm-cores.def: Added core cortex-r7.
22236 * config/arm/arm-tune.md: Regenerated.
22237 * config/arm/arm-tables.opt: Regenerated.
22238 * doc/invoke.texi: Added entry for core cortex-r7.
22240 2013-03-27 Walter Lee <walt@tilera.com>
22242 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22243 double-decrement of next_scratch_regno.
22245 2013-03-27 Walter Lee <walt@tilera.com>
22247 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22249 (insn_v1mulus): Ditto.
22250 (insn_v2muls): Ditto.
22252 2013-03-27 Walter Lee <walt@tilera.com>
22254 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22255 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22257 2013-03-27 Walter Lee <walt@tilera.com>
22259 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22260 (*sibcall_value): Ditto.
22262 2013-03-27 Walter Lee <walt@tilera.com>
22264 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22265 (insn_mnz_v8qi): ... this ...
22266 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22268 (insn_v<n>mnz): Replaced by ...
22269 (insn_v1mnz): ... this ...
22270 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22272 (insn_mz_<mode>): Replaced by ...
22273 (insn_mz_v8qi): ... this ...
22274 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22276 (insn_v<n>mz): Replaced by ...
22277 (insn_v1mz): ... this ...
22278 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22281 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22283 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22285 2013-03-26 Roland McGrath <mcgrathr@google.com>
22287 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22288 than fprintf with a non-constant, non-format string.
22290 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22292 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22293 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22294 operand 0 predicate.
22295 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22296 attribute. Use general_x64nomem_operand as operand 1 predicate.
22297 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22298 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22299 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22300 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22301 (mov<mode>_insv_1): Remove expander. Merge insn with
22302 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22303 Use general_x64nomem_operand as operand 1 predicate.
22304 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22305 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22306 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22307 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22308 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22309 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22310 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22311 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22312 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22313 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22314 (general_x64nomem_operand): Ditto.
22316 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22318 * config/rtems.opt: Add -pthread option.
22320 2013-03-26 Richard Biener <rguenther@suse.de>
22322 * alias.c (find_base_term): Avoid redundant and not used recursion.
22323 (base_alias_check): Get the initial base term from the caller.
22324 (true_dependence_1): Compute and pass base terms to base_alias_check.
22325 (write_dependence_p): Likewise.
22326 (may_alias_p): Likewise.
22328 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22330 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22331 PC-relative load in SI modes and above only.
22333 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22335 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22336 * config/arm/iwmmxt.md (WCGR0): Update.
22337 (WCGR1, WCGR2, WCGR3): Likewise.
22339 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22341 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22342 Use x64 and nox64 isa attributes.
22344 2013-03-26 Richard Biener <rguenther@suse.de>
22346 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22347 alignment computations and rely on get_object_alignment_1
22348 for the !TYPE_P case.
22349 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22351 2013-03-26 Walter Lee <walt@tilera.com>
22353 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22354 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22356 2013-03-25 Jeff Law <law@redhat.com>
22358 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22359 check for INTEGRAL_TYPE_P that was missing due to checking in
22360 wrong version of prior patch.
22362 2013-03-25 Walter Lee <walt@tilera.com>
22364 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22365 TILEGX_INSN_SHUFFLEBYTES1.
22366 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22368 (tilegx_builtins): Ditto.
22369 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22371 2013-03-25 Walter Lee <walt@tilera.com>
22373 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22374 (floatunssisf2): New pattern.
22375 (floatsidf2): New pattern.
22376 (floatunssidf2): New pattern.
22378 2013-03-25 Walter Lee <walt@tilera.com>
22380 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22381 tests for constraint J, K, N, P.
22383 2013-03-25 Walter Lee <walt@tilera.com>
22385 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22386 Use indirect/pcrel encoding.
22387 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22390 2013-03-25 Steve Ellcey <sellcey@mips.com>
22392 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22393 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22394 * config/mips/mips.c (mips_option_override): Set IMADD default.
22395 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22396 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22397 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22398 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22399 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22401 2013-03-25 Jeff Law <law@redhat.com>
22403 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22404 slightly to avoid creating and folding useless trees. Simplify
22405 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22407 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22409 * config/i386/i386.md (*zero_extendsidi2): Merge with
22410 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22411 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22412 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22413 targets. Clarify comment.
22415 2013-03-25 Martin Jambor <mjambor@suse.cz>
22417 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22418 pass-through jump functions differently.
22419 (ipa_read_jump_function): Likewise. Also use setter functions to set
22422 2013-03-25 Martin Jambor <mjambor@suse.cz>
22424 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22425 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22427 (ipa_get_indirect_edge_target): New function.
22428 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22429 ipa_get_indirect_edge_target. Update all callers.
22430 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22431 ipa_get_indirect_edge_target_1 instead of calling
22432 ipa_get_indirect_edge_target.
22433 (create_specialized_node): Pass aggvlas to
22434 ipcp_discover_new_direct_edges.
22436 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22438 * config/arm/arm.md (f_sels, f_seld): New types.
22439 (*cmov<mode>): New pattern.
22440 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22443 2013-03-25 Kai Tietz <ktietz@redhat.com>
22445 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22446 POSIX-printf for mingw-hosted builds.
22448 2013-03-25 Richard Biener <rguenther@suse.de>
22450 PR middle-end/56694
22451 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22452 must-not-throw stmt location.
22454 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22456 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22457 Emit load-acquire versions when acq is true.
22458 (arm_emit_store_exclusive): Add rel parameter.
22459 Emit store-release versions when rel is true.
22460 (arm_split_compare_and_swap): Use acquire-release instructions
22462 of barriers when appropriate.
22463 (arm_split_atomic_op): Likewise.
22464 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22465 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22466 (VUNSPEC_SLX): Likewise.
22467 (VUNSPEC_LDA): Likewise.
22468 (VUNSPEC_STL): Likewise.
22469 * config/arm/sync.md (atomic_load<mode>): New pattern.
22470 (atomic_store<mode>): Likewise.
22471 (arm_load_acquire_exclusive<mode>): Likewise.
22472 (arm_load_acquire_exclusivesi): Likewise.
22473 (arm_load_acquire_exclusivedi): Likewise.
22474 (arm_store_release_exclusive<mode>): Likewise.
22476 2013-03-25 Catherine Moore <clm@codesourcery.com>
22478 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22479 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22480 * config/mip/predicates.md (lwsp_swsp_operand,
22481 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22482 sb16_operand, db4_operand, db7_operand, ib3_operand,
22483 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22484 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22485 andi16_operand): New predicates.
22486 * config/mips/mips.md (compression): New attribute.
22487 (enabled): New attribute.
22488 (length): Consider compression in computing length.
22489 (shift_compression): New code attribute.
22490 (*add<mode>3): New operands. Record compression.
22491 (sub<mode>3): Likewise.
22492 (one_cmpl<mode>2): Likewise.
22493 (*and<mode>3): Likewise.
22494 (*ior<mode>3): Likewise.
22495 (unnamed pattern for xor): Likewise.
22496 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22497 (*<optab><mode>3): Likewise.
22498 (*mov<mode>_internal: Likewise.
22499 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22500 (mips_unsigned_immediate_p): New.
22501 (umips_lwsp_swsp_address_p): New.
22502 (m16_based_address_p): New.
22503 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22504 (mips_unsigned_immediate_p): New prototype.
22505 (lwsp_swsp_address_p): New prototype.
22506 (m16_based_address_p): New prototype.
22507 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22508 (mips_signed_immediate_p): New function.
22509 (m16_based_address_p): New function.
22510 (lwsp_swsp_address_p): New function.
22511 (mips_print_operand_punctuation): Recognize short delay slot insns
22512 for microMIPS.add<mode>3"
22514 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22517 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22518 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22520 2013-03-25 Richard Biener <rguenther@suse.de>
22522 PR tree-optimization/56689
22523 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22526 2013-03-25 Richard Biener <rguenther@suse.de>
22528 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22530 (memory_references): Likewise.
22531 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22532 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22533 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22534 (gather_mem_refs_in_loops): Fold into ...
22535 (analyze_memory_references): ... this. Move initialization
22536 to tree_ssa_lim_initialize.
22537 (fill_always_executed_in): Rename to ...
22538 (fill_always_executed_in_1): ... this.
22539 (fill_always_executed_in): Move contains_call computation to
22540 this new function from ...
22541 (tree_ssa_lim_initialize): ... here.
22542 (tree_ssa_lim): Call fill_always_executed_in.
22544 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22546 * postreload.c (reload_combine): Fix code detecting returns.
22548 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22550 * function.c (emit_use_return_register_into_block): On cc0 targets,
22551 do not emit the sequence between cc0 setter and user.
22553 2013-03-25 Kai Tietz <ktietz@redhat.com>
22555 * config/i386/predicates.md (local_symbolic_operand): Interpret
22556 dll-imported symbols as none-local.
22558 2013-03-25 Richard Biener <rguenther@suse.de>
22560 * tree-ssa-loop-im.c (struct depend): Remove.
22561 (struct lim_aux_data): Make depends a vec of gimples.
22562 (free_lim_aux_data): Adjust.
22563 (add_dependency): Likewise.
22564 (set_level): Likewise.
22566 2013-03-25 Richard Biener <rguenther@suse.de>
22568 PR middle-end/56434
22569 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22570 the pointer returned by calls with ECF_MALLOC set.
22572 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22574 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22576 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22578 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22579 using MMXMODE mode iterator.
22580 (*move<mode>_internal): Merge with *movv2sf_internal and
22581 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22583 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22585 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22586 (record_last_mem_set_info): Likewise.
22588 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22589 of XNEWVEC followed by memset.
22590 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22592 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22594 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22595 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22596 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22597 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22598 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22599 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22600 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22601 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22602 BARRIER_P instead of GET_CODE.
22604 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22606 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22607 inaccuracy in the probing code.
22609 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22610 (ctrapdi4): Likewise.
22612 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22614 * calls.c (expand_call): Add missing guard to code handling return
22615 of non-BLKmode structures in MSB.
22616 * function.c (expand_function_end): Likewise.
22618 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22620 * combine.c (try_combine): Adjust comment. Do not add the set of
22621 insn #0 if the destination indirectly is set or dies in insn #2.
22622 Tidy up code to distribute a new note.
22624 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22626 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22627 also for alternatives 16 and 17.
22629 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22631 * config/i386/sse.md (*mov<mode>_internal): Merge with
22632 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22633 Emit insn template depending on type attribute. Use
22634 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22635 movd instead of movq mnemonic for interunit moves. Rewrite mode
22636 attribute calculation. Remove unit attribute calculation.
22637 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22638 Set prefix_data16 attribute for DImode ssemov types.
22639 Use Ym instead of y for SSE-MMX conversion alternatives.
22640 Reorder operand constraints.
22642 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22644 * df.h (df_insn_delete): Adjust prototype.
22645 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22646 and let it decide whether mark the basic block dirty.
22647 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22648 * df-scan.c (df_insn_info_delete): New helper function, split
22649 off from df_insn_delete.
22650 (df_scan_free_bb_info): Use it.
22651 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22653 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22654 that the insn is actually an insn and it has a non-NULL basic block.
22655 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22657 2013-03-22 Richard Biener <rguenther@suse.de>
22659 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22661 (mem_ref_alloc): Do not allocate them.
22662 (refs_independent_p): Do not query or maintain a cache.
22664 2013-03-22 Richard Biener <rguenther@suse.de>
22666 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22667 (gather_mem_refs_in_loops): Do not compute it.
22668 (analyze_memory_references): Do not allocate it.
22669 (tree_ssa_lim_finalize): Do not free it.
22670 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22672 2013-03-22 Richard Biener <rguenther@suse.de>
22674 * is-a.h (as_a): Use gcc_checking_assert.
22676 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22678 * config/aarch64/aarch64.c (aarch64_print_operand): New
22679 format specifier for printing a constant in hex.
22680 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22681 format specifier for printing second operand.
22683 2013-03-22 Richard Biener <rguenther@suse.de>
22685 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22687 (gather_mem_refs_in_loops): Perform store accumulation here.
22688 (create_vop_ref_mapping_loop): Remove.
22689 (create_vop_ref_mapping): Likewise.
22690 (analyze_memory_references): Initialize refs_stored_in_loop.
22691 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22692 (record_indep_loop): Remove.
22693 (record_dep_loop): New function.
22694 (ref_indep_loop_p_1): Adjust to only walk over references
22695 in the loop, not its subloops.
22696 (ref_indep_loop_p): Rename to ...
22697 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22698 maintaining a more fine-grained cache.
22699 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22700 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22702 2013-03-22 Richard Biener <rguenther@suse.de>
22704 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22705 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22706 aggregate mem_ref_loc.
22707 (free_mem_ref_locs): Inline into ...
22708 (memref_free): ... this and adjust.
22709 (mem_ref_alloc): Adjust.
22710 (mem_ref_locs_alloc): Remove.
22711 (record_mem_ref_loc): Adjust.
22712 (get_all_locs_in_loop): Rewrite into ...
22713 (for_all_locs_in_loop): ... this iterator.
22714 (rewrite_mem_ref_loc): New functor.
22715 (rewrite_mem_refs): Use for_all_locs_in_loop.
22716 (sm_set_flag_if_changed): New functor.
22717 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22718 (ref_always_accessed): New functor.
22719 (ref_always_accessed_p): Use for_all_locs_in_loop.
22721 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22723 * tree-pass.h (PROP_gimple_lvec): New.
22724 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22725 (init_optimization_passes): Move pass_lower_vector.
22726 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22728 (pass_lower_vector): Provide PROP_gimple_lvec.
22729 (pass_lower_vector_ssa): Likewise.
22730 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22732 2013-03-21 Mark Wielaard <mjw@redhat.com>
22734 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22736 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22738 * config/i386/i386.md (*movdi_internal): Disparage slightly
22739 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22740 conversion alternatives.
22742 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22744 PR middle-end/48087
22745 * diagnostic.def (DK_WERROR): New kind.
22746 * diagnostic.h (werrorcount): Define.
22747 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22748 promoted to DK_ERROR, increment DK_WERROR counter instead of
22750 * toplev.c (toplev_main): Call print_ignored_options even if
22751 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22752 even if just werrorcount is non-zero.
22755 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22757 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22758 (string_cst_pool_decl): New function.
22759 (optimize_one_addr_into_implicit_ptr): New function.
22760 (resolve_addr_in_expr): Optimize DWARF location expression
22761 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22762 which doesn't live in memory, but has DW_AT_location or
22763 DW_AT_const_value, or refers to a string literal, into
22764 DW_OP_GNU_implicit_pointer.
22765 (optimize_location_into_implicit_ptr): New function.
22766 (resolve_addr): If removing DW_AT_location of a variable because
22767 it was DW_OP_addr of address of the variable, but the variable doesn't
22768 live in memory, try to emit const value attribute for the initializer.
22770 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22772 * tree.h (VECTOR_TYPE_P): New macro.
22773 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22774 TYPE_MODE): Use it.
22775 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22776 VEC_COND_EXPR cannot be lvalues.
22777 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22779 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22781 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22782 Restrict the transformation to equal modes.
22784 2013-03-21 Richard Biener <rguenther@suse.de>
22786 PR tree-optimization/39326
22787 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22788 (MEM_ANALYZABLE): Adjust.
22789 (record_mem_ref_loc): Move bitmap ops ...
22790 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22791 unanalyzable refs, do not record locations for it.
22792 (analyze_memory_references): Allocate ref zero as shared
22794 (refs_independent_p): Do not test for unanalyzed mems here.
22795 (ref_indep_loop_p_1): Special-case disambiguation against
22796 the unanalyzed ref.
22797 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22799 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22801 * config/arm/arm-protos.h (tune_params): Add
22802 prefer_neon_for_64bits field.
22803 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22804 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22805 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22806 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22807 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22808 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22809 (arm_option_override): Handle -mneon-for-64bits new option.
22810 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22811 (prefer_neon_for_64bits): Declare new variable.
22812 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22813 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22814 (arch_enabled): Handle new arch types. Remove support for onlya8
22816 (one_cmpldi2): Use new arch names.
22817 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22818 * config/arm/arm.opt (mneon-for-64bits): Add option.
22819 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22820 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22821 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22823 * doc/invoke.texi (-mneon-for-64bits): Document.
22825 2013-03-21 Richard Biener <rguenther@suse.de>
22827 PR tree-optimization/39326
22828 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22829 (sort_bbs_in_loop_postorder_cmp): New function.
22830 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22832 2013-03-21 Richard Biener <rguenther@suse.de>
22834 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22835 (vect_insert_into_interleaving_chain): Likewise.
22836 (vect_drs_dependent_in_basic_block): Inline ...
22837 (vect_slp_analyze_data_ref_dependence): ... here. New function,
22839 (vect_analyze_data_ref_dependence): ... here. Simplify.
22840 (vect_check_interleaving): Simplify.
22841 (vect_analyze_data_ref_dependences): Likewise. Split out ...
22842 (vect_slp_analyze_data_ref_dependences): ... this new function.
22843 (dr_group_sort_cmp): New function.
22844 (vect_analyze_data_ref_accesses): Compute data-reference groups
22845 here instead of in vect_analyze_data_ref_dependence. Use
22846 a more efficient algorithm.
22847 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
22848 vect_slp_analyze_data_ref_dependences. Call
22849 vect_analyze_data_ref_accesses earlier.
22850 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
22851 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
22852 (vect_slp_analyze_data_ref_dependences): New prototype.
22854 2013-03-21 Richard Biener <rguenther@suse.de>
22856 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
22857 ref is stored in the loop.
22858 (find_refs_for_sm): Walk only over all stores.
22859 (store_motion_loop): Allocate from lim_bitmap_obstack.
22860 (store_motion): Likewise.
22862 2013-03-21 Richard Biener <rguenther@suse.de>
22864 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
22865 Update virtual SSA form.
22867 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22869 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
22870 * configure: Regenerate.
22871 * config.in: Regenerate.
22872 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
22873 if !HAVE_LD_EH_FRAME_CIEV3.
22875 2013-03-21 Richard Biener <rguenther@suse.de>
22877 * tree-cfg.c (verify_expr_no_block): New function.
22878 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
22879 nor DECL_VALUE_EXPR have locations with associated blocks.
22880 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
22881 (clear_unused_block_pointer): Remove code dealing with
22882 blocks in DECL_DEBUG_EXPR locations.
22884 2013-03-21 Richard Biener <rguenther@suse.de>
22886 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
22887 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
22888 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
22889 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
22890 instead of DECL_DEBUG_EXPR_IS_FROM.
22891 * gimplify.c (gimplify_modify_expr): Likewise.
22892 * tree-cfg.c (verify_expr_location_1): Likewise.
22893 * tree-complex.c (create_one_component_var): Likewise.
22894 * tree-sra.c (create_access_replacement): Likewise.
22895 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
22896 (clear_unused_block_pointer): Likewise.
22897 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
22898 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
22899 * var-tracking.c (var_debug_decl): Likewise.
22900 (track_expr_p): Likewise.
22901 * tree-inline.c (add_local_variables): Likewise. Set
22902 DECL_HAS_DEBUG_EXPR_P after copying it.
22903 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
22904 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
22906 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22909 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
22910 * configure: Regenerate.
22911 * config.in: Regenerate.
22912 * config/i386/i386.md (*movdf_internal): Use
22913 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22914 movd instead of movq mnemonic for interunit moves.
22915 (*movdi_internal): Ditto.
22917 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22919 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
22920 (abd<mode>_3): New pattern.
22921 (aba<mode>_3): New pattern.
22922 (fabd<mode>_3): New pattern.
22924 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22926 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
22927 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
22928 occurrence of REGISTER_PREFIX as its empty string.
22930 2013-03-20 Jeff Law <law@redhat.com>
22932 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
22933 addititional equivalences for equality comparisons between an SSA_NAME
22934 and a constant where the SSA_NAME was set from a widening conversion.
22936 2013-03-20 Walter Lee <walt@tilera.com>
22938 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
22940 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22942 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
22943 depending on type attribute.
22944 (*movti_internal): Ditto.
22945 (*movtf_internal): Ditto.
22946 (*movxf_internal): Ditto.
22947 (*movdf_internal): Ditto.
22948 (*movsf_internal): Ditto.
22950 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22952 * config/i386/i386.md (*movti_internal): Set prefix attribute to
22953 maybe_vex for sselog1 and ssemov types.
22954 (*movdi_internal): Reorder operand constraints.
22955 (*movsi_internal): Ditto. Set prefix attribute to
22956 maybe_vex for sselog1 and ssemov types.
22957 (*movtf_internal): Set prefix attribute to maybe_vex
22958 for sselog1 and ssemov types.
22959 (*movdf_internal): Ditto. Set prefix_data16 attribute for
22960 DImode ssemov types. Reorder operand constraints.
22961 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
22962 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
22963 attribute for SImode ssemov types. Reorder operand constraints.
22965 2013-03-20 Martin Jambor <mjambor@suse.cz>
22967 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
22968 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
22970 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
22972 * config/rs6000/predicates.md (indexed_address, update_address_mem
22973 update_indexed_address_mem): New predicates.
22974 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
22975 attribute for load/store instructions.
22976 * config/rs6000/dfp.md (movsd_store): Likewise.
22977 (movsd_load): Likewise.
22978 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
22979 (unnamed HI->DI extend define_insn): Likewise.
22980 (unnamed SI->DI extend define_insn): Likewise.
22981 (unnamed QI->SI extend define_insn): Likewise.
22982 (unnamed QI->HI extend define_insn): Likewise.
22983 (unnamed HI->SI extend define_insn): Likewise.
22984 (unnamed HI->SI extend define_insn): Likewise.
22985 (extendsfdf2_fpr): Likewise.
22986 (movsi_internal1): Likewise.
22987 (movsi_internal1_single): Likewise.
22988 (movhi_internal): Likewise.
22989 (movqi_internal): Likewise.
22990 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
22991 attribute for load/store instructions.
22992 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
22994 (mov<mode>_softfloat): Likewise.
22995 (mov<mode>_hardfloat32): Likewise.
22996 (mov<mode>_hardfloat64): Likewise.
22997 (mov<mode>_softfloat64): Likewise.
22998 (movdi_internal32): Likewise.
22999 (movdi_internal64): Likewise.
23000 (probe_stack_<mode>): Likewise.
23002 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
23004 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
23005 floating point, and decimal floating point to reload iterator.
23007 * config/rs6000/constraints.md (wl constraint): New constraints to
23008 return FLOAT_REGS if certain options are used to reduce the number
23009 of separate patterns that exist in the file.
23010 (wx constraint): Likewise.
23011 (wz constraint): Likewise.
23013 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23014 -mdebug=reg, print wg, wl, wx, and wz constraints.
23015 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23016 Initialize the reload functions for 64-bit binary/decimal floating
23018 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23019 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23020 create the buffer on the stack to overcome not having a 32-bit
23022 (rs6000_emit_move): Likewise.
23023 (rs6000_secondary_memory_needed_rtx): Likewise.
23024 (rs6000_alloc_sdmode_stack_slot): Likewise.
23025 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23026 via xxlxor, just like DFmode 0.0.
23028 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
23029 define as 1 if we are running on a power7 or newer.
23030 (enum r6000_reg_class_enum): Add new constraints.
23032 * config/rs6000/dfp.md (movsd): Delete, combine with binary
23033 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23034 with other moves by using conditional constraits (wg). Use LFIWZX
23035 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
23036 (movsd splitter): Likewise.
23037 (movsd_hardfloat): Likewise.
23038 (movsd_softfloat): Likewise.
23040 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
23041 binary and decimal floating point moves.
23042 (fmove_ok): New attributes to combine binary and decimal floating
23043 point moves, and to combine power6x (mfpgpr) moves along normal
23045 (real_value_to_target): Likewise.
23046 (f32_lr): Likewise.
23047 (f32_lm): Likewise.
23048 (f32_li): Likewise.
23049 (f32_sr): Likewise.
23050 (f32_sm): Likewise.
23051 (f32_si): Likewise.
23052 (movsf): Combine binary and decimal floating point moves. Combine
23053 power6x (mfpgpr) moves with other moves by using conditional
23054 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
23055 (mov<mode> for SFmode/SDmode); Likewise.
23056 (SFmode/SDmode splitters): Likewise.
23057 (movsf_hardfloat): Likewise.
23058 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
23059 (movsf_softfloat): Likewise.
23060 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
23062 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
23063 wx and wz constraints.
23065 * config/rs6000/constraints.md (wg constraint): New constraint to
23066 return FLOAT_REGS if -mmfpgpr (power6x) was used.
23068 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
23071 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23072 -mdebug=reg, print wg, wl, wx, and wz constraints.
23073 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23074 Initialize the reload functions for 64-bit binary/decimal floating
23076 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23077 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23078 create the buffer on the stack to overcome not having a 32-bit
23080 (rs6000_emit_move): Likewise.
23081 (rs6000_secondary_memory_needed_rtx): Likewise.
23082 (rs6000_alloc_sdmode_stack_slot): Likewise.
23083 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23084 via xxlxor, just like DFmode 0.0.
23086 * config/rs6000/dfp.md (movdd): Delete, combine with binary
23087 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23088 with other moves by using conditional constraits (wg). Use LFIWZX
23089 and STFIWX for loading SDmode on power7.
23090 (movdd splitters): Likewise.
23091 (movdd_hardfloat32): Likewise.
23092 (movdd_softfloat32): Likewise.
23093 (movdd_hardfloat64_mfpgpr): Likewise.
23094 (movdd_hardfloat64): Likewise.
23095 (movdd_softfloat64): Likewise.
23097 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
23098 64-bit binary and decimal floating point moves.
23099 (FMOVE64X): Likewise.
23100 (movdf): Combine 64-bit binary and decimal floating point moves.
23101 Combine power6x (mfpgpr) moves with other moves by using
23102 conditional constraits (wg).
23103 (mov<mode> for DFmode/DDmode): Likewise.
23104 (DFmode/DDmode splitters): Likewise.
23105 (movdf_hardfloat32): Likewise.
23106 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
23107 (movdf_softfloat32): Likewise.
23108 (movdf_hardfloat64_mfpgpr): Likewise.
23109 (movdf_hardfloat64): Likewise.
23110 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
23111 (movdf_softfloat64): Likewise.
23112 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
23113 (reload_<mode>_load): Move to later in the file so they aren't in
23114 the middle of the floating point move insns.
23115 (reload_<mode>_store): Likewise.
23117 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
23120 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
23121 constraint if -mdebug=reg.
23122 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
23123 Enable using dd reload support if needed.
23125 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
23126 binary and decimal floating point moves in rs6000.md.
23127 (movtd_internal): Likewise.
23129 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
23130 decimal floating point moves.
23132 (movtf_internal): Likewise.
23133 (mov<mode>_internal, TDmode/TFmode): Likewise.
23134 (movtf_softfloat): Likewise.
23135 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
23137 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
23138 movdi_internal64, using wg constraint for move direct operations.
23139 (movdi_internal64): Likewise.
23141 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
23142 MODES_TIEABLE_P for selected modes. Print the numerical value of
23143 the various virtual registers. Use GPR/FPR first/last values,
23144 instead of hard coding the register numbers. Print which modes
23145 have reload functions registered.
23146 (rs6000_option_override_internal): If -mdebug=reg, trace the options
23147 settings before/after setting cpu, target and subtarget settings.
23148 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
23149 and for secondary reload failures in rs6000_secondary_reload_inner.
23150 (rs6000_secondary_reload_fail): Likewise.
23151 (rs6000_secondary_reload_inner): Likewise.
23153 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
23154 macros for first/last GPR and FPR registers.
23155 (LAST_GPR_REGNO): Likewise.
23156 (FIRST_FPR_REGNO): Likewise.
23157 (LAST_FPR_REGNO): Likewise.
23159 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
23160 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
23161 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
23162 (vcond<mode><mode>): Likewise.
23163 (vcondu<mode><mode>): Likewise.
23164 (vector_gtu<mode>): Likewise.
23165 (vector_gte<mode>): Likewise.
23166 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
23167 to prevent the compiler from converting DImode operations to TImode.
23168 (ior<mode>3): Likewise.
23169 (and<mode>3): Likewise.
23170 (one_cmpl<mode>2): Likewise.
23171 (nor<mode>3): Likewise.
23172 (andc<mode>3): Likewise.
23174 * config/rs6000/constraints.md (wt constraint): New constraint
23175 that returns VSX_REGS if TImode is allowed in VSX registers.
23177 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23178 constant under VSX.
23180 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23181 similar to TImode, but it is restricted to being in the GPRs.
23183 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23184 TImode to occupy a single VSX register.
23186 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23187 -mvsx-timode for power7/power8.
23188 (power7 cpu): Likewise.
23189 (power8 cpu): Likewise.
23191 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23192 sure that TFmode/TDmode take up two registers if they are ever
23193 allowed in the upper VSX registers.
23194 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23196 (rs6000_init_hard_regno_mode_ok): Likewise.
23197 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23198 constraint. Print if LRA is turned on.
23199 (rs6000_option_override_internal): Give an error if -mvsx-timode
23200 and VSX is not enabled.
23201 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23202 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23203 to reg+offset addressing. Use PTImode when checking offset
23204 addresses for validity.
23205 (reg_offset_addressing_ok_p): Likewise.
23206 (rs6000_legitimate_offset_address_p): Likewise.
23207 (rs6000_legitimize_address): Likewise.
23208 (rs6000_legitimize_reload_address): Likewise.
23209 (rs6000_legitimate_address_p): Likewise.
23210 (rs6000_eliminate_indexed_memrefs): Likewise.
23211 (rs6000_emit_move): Likewise.
23212 (rs6000_secondary_reload): Likewise.
23213 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23214 reloads to fpr registers to continue to use reg+offset addressing,
23215 but 64-bit reloads to altivec registers need reg+reg addressing.
23216 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23217 it. Treat LO_SUM like a PLUS operation.
23218 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23219 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23220 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23221 registers to share a register with a smaller sized type, since VSX
23222 puts scalars in the upper 64-bits.
23223 (print_operand): Add support for PTImode.
23224 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23225 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23226 registers, but don't have arithmetic support.
23227 (rs6000_memory_move_cost): Add test for VSX.
23228 (rs6000_opt_masks): Add -mvsx-timode.
23230 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23233 (VSr): Use wt constraint for TImode.
23234 (VSv): Drop TImode support.
23235 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23236 (vsx_movti_64bit): Likewise.
23237 (vsx_movti_32bit): Likewise.
23238 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23239 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23240 one '?' on the appropriate output constraint. Do not allow TImode
23241 logical operations on 32-bit systems.
23242 (vsx_ior<mode>3): Likewise.
23243 (vsx_xor<mode>3): Likewise.
23244 (vsx_one_cmpl<mode>2): Likewise.
23245 (vsx_nor<mode>3): Likewise.
23246 (vsx_andc<mode>3): Likewise.
23247 (vsx_concat_<mode>): Likewise.
23248 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23250 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23251 OPTION_MASK_VSX_TIMODE.
23252 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23253 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23255 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23256 (TI2 iterator): New iterator for TImode, PTImode.
23257 (wd mode attribute): Add values for vector types.
23258 (movti_string): Replace TI move operations with operations for TImode
23259 and PTImode. Add support for TImode being allowed in VSX registers.
23260 (mov<mode>_string, TImode/PTImode): Likewise.
23261 (movti_ppc64): Likewise.
23262 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23263 (TI mode splitters): Likewise.
23265 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23268 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23270 PR tree-optimization/56355
23271 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23272 Also handle integers with undefined overflow.
23274 2013-03-20 Catherine Moore <clm@codesourcery.com>
23275 Maciej W. Rozycki <macro@codesourcery.com>
23276 Tom de Vries <tom@codesourcery.com>
23277 Nathan Sidwell <nathan@codesourcery.com>
23278 Iain Sandoe <iain@codesourcery.com>
23279 Nathan Froyd <froydnj@codesourcery.com>
23280 Chao-ying Fu <fu@mips.com>
23282 * doc/extend.texi: (micromips, nomicromips, nocompression):
23283 Document new function attributes.
23284 * doc/invoke.texi (minterlink-compressed, mmicromips,
23285 m14k, m14ke, m14kec): Document new options.
23286 (minterlink-mips16): Update documentation.
23287 * doc/md.texi (ZC, ZD): Document new constraints.
23288 * configure.ac (gcc_cv_as_micromips): Check if linker
23289 supports the .set micromips directive.
23290 * configure: Regenerate.
23291 * config.in: Regenerate.
23292 * config/mips/mips-tables.opt: Regenerate.
23293 * config/mips/micromips.md: New file.
23294 * constraints.md (ZC, ZD): New constraints.
23295 * config/mips/predicates.md (movep_src_register): New predicate.
23296 (movep_src_operand): New predicate.
23297 (non_volatile_mem_operand): New predicate.
23298 * config/mips/mips.md (multimem): New type.
23299 (length): Differentiate between 17-bit and 18-bit branch offsets.
23300 (MOVEP1, MOVEP2): New mode iterator.
23301 (mov_<load>l): Use ZC constraint.
23302 (mov_<load>r): Likewise.
23303 (mov_<store>l): Likewise.
23304 (mov_<store>r): Likewise.
23305 (*branch_equality<mode>_inverted): Add microMIPS support.
23306 (*branch_equality<mode>): Likewise.
23307 (*jump_absolute): Likewise.
23308 (indirect_jump_<mode>): Likewise.
23309 (tablejump_<mode>): Likewise.
23310 (<optab>_internal): Likewise.
23311 (sibcall_internal): Likewise.
23312 (sibcall_value_internal): Likewise.
23313 (prefetch): Use constraint ZD.
23314 * config/mips/mips.opt (minterlink-compressed): New option.
23315 (minterlink-mips16): Now an alias for minterlink-compressed.
23316 (mmicromips): New option.
23317 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23318 (compare_and_swap_12): Likewise.
23319 (sync_add<mode>): Likewise.
23320 (sync_<optab>_12): Likewise.
23321 (sync_old_<optab>_12): Likewise.
23322 (sync_new_<optab>_12): Likewise.
23323 (sync_nand_12): Likewise.
23324 (sync_old_nand_12): Likewise.
23325 (sync_new_nand_12): Likewise.
23326 (sync_sub<mode>): Likewise.
23327 (sync_old_add<mode>): Likewise.
23328 (sync_old_sub<mode>): Likewise.
23329 (sync_new_add<mode>): Likewise.
23330 (sync_new_sub<mode>): Likewise.
23331 (sync_<optab><mode>): Likewise.
23332 (sync_old_<optab><mode>): Likewise.
23333 (sync_new_<optab><mode>): Likewise.
23334 (sync_nand<mode>): Likewise.
23335 (sync_old_nand<mode>): Likewise.
23336 (sync_new_nand<mode>): Likewise.
23337 (sync_lock_test_and_set<mode>): Likewise.
23338 (test_and_set_12): Likewise.
23339 (atomic_compare_and_swap<mode>): Likewise.
23340 (atomic_exchange<mode>_llsc): Likewise.
23341 (atomic_fetch_add<mode>_llsc): Likewise.
23342 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23343 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23344 (umips_save_restore_pattern_p): Likewise.
23345 (umips_load_store_pair_p): Likewise.
23346 (umips_output_load_store_pair): Likewise.
23347 (umips_movep_target_p): Likewise.
23348 (umips_12bit_offset_address_p): Likewise.
23349 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23350 (mips_base_mips16): Rename this...
23351 (mips_base_compression_flags): ...to this. Update all uses.
23352 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23353 (mips_mips16_decl_p): Delete.
23354 (mips_nomips16_decl_p): Delete.
23355 (mips_get_compress_on_flags): New function.
23356 (mips_get_compress_off_flags): New function.
23357 (mips_get_compress_mode): New function.
23358 (mips_get_compress_on_name): New function.
23359 (mips_get_compress_off_name): New function.
23360 (mips_insert_attributes): Support multiple compression types.
23361 (mips_merge_decl_attributes): Likewise.
23362 (umips_12bit_offset_address_p): New function.
23363 (mips_start_function_definition): Emit .set micromips directive.
23364 (mips_call_may_need_jalx_p): New function.
23365 (mips_function_ok_for_sibcall): Add microMIPS support.
23366 (mips_print_operand_punctuation): Support short delay slots and
23368 (umips_swm_mask, umips_swm_encoding): New.
23369 (umips_build_save_restore): New function.
23370 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23371 (was_mips16_p): Remove.
23372 (old_compression_mode): New.
23373 (mips_set_compression_mode): New function.
23374 (mips_set_current_function): Add microMIPS support.
23375 (mips_option_override): Likewise.
23376 (umips_save_restore_pattern_p): New function.
23377 (umips_output_save_restore): New function.
23378 (umips_load_store_pair_p_1): New function.
23379 (umips_load_store_pair_p): New function.
23380 (umips_output_load_store_pair_1): New function.
23381 (umips_output_load_store_pair): New function.
23382 (umips_movep_target_p) New function.
23383 (mips_prepare_pch_save): Add microMIPS support.
23384 * config/mips/mips.h (TARGET_COMPRESSION): New.
23385 (TARGET_CPU_CPP_BUILTINS): Update macro
23386 to use new compression flags and to support microMIPS.
23387 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23388 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23389 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23390 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23391 (ASM_SPEC): Support mmicromips and mno-micromips.
23392 (M16STORE_REG_P): New macro.
23393 (MIPS_CALL): Support TARGET_MICROMIPS.
23394 (MICROMIPS_J): New macro.
23395 (mips_base_mips16): Rename this...
23396 (mips_base_compression_flags): ...to this.
23397 (UMIPS_12BIT_OFFSET_P): New macro.
23398 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23399 (MULTILIB_DIRNAMES): Likewise.
23400 2013-03-20 Richard Biener <rguenther@suse.de>
23402 PR tree-optimization/56661
23403 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23404 the result does not have to be distinct.
23406 2013-03-20 Richard Biener <rguenther@suse.de>
23408 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23411 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23412 Steven Bosscher <steven@gcc.gnu.org>
23414 PR rtl-optimization/56605
23415 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23417 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23420 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23421 that require movd instead of movq.
23423 2013-03-20 Richard Biener <rguenther@suse.de>
23425 * tree-ssa-structalias.c (struct variable_info): Add pointer
23426 to the first field of an aggregate with sub-vars. Make
23427 this and the pointer to the next subfield its ID.
23428 (vi_next): New function.
23429 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23430 storedanything_id, integer_id): Increment by one.
23431 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23432 get_call_clobber_vi): Adjust.
23433 (solution_set_expand): Simplify and speedup.
23434 (solution_set_add): Inline into ...
23435 (set_union_with_increment): ... this. Adjust accordingly.
23436 (do_sd_constraint): Likewise.
23437 (do_ds_constraint): Likewise.
23438 (do_complex_constraint): Simplify.
23439 (build_pred_graph): Adjust.
23440 (solve_graph): Likewise. Simplify and speedup.
23441 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23442 get_constraint_for_component_ref, get_constraint_for_1,
23443 first_vi_for_offset, first_or_preceding_vi_for_offset,
23444 create_function_info_for, create_variable_info_for_1,
23445 create_variable_info_for, intra_create_variable_infos): Adjust.
23446 (init_base_vars): Push NULL for ID zero.
23447 (compute_points_to_sets): Adjust.
23449 2013-03-20 Richard Biener <rguenther@suse.de>
23451 * cfgloop.c (verify_loop_structure): Streamline and avoid
23452 ICEing on corrupt loop tree.
23453 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23456 2013-03-20 Richard Biener <rguenther@suse.de>
23458 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23459 check whether an SSA update is needed.
23461 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23463 * config/mips/constraints.md (T): Rename to...
23467 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23468 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23470 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23472 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23473 (*subsi3_carryin_uxtw): Likewise.
23475 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23477 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23478 (*rorsi3_insn_uxtw): Likewise.
23480 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23482 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23483 (*extrsi5_insn_uxtw): Likewise.
23485 2013-03-19 Richard Biener <rguenther@suse.de>
23487 PR tree-optimization/56273
23488 * passes.c (init_optimization_passes): Move second VRP after DOM.
23490 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23492 * config/i386/i386.md (*movti_internal): Merge from
23493 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23494 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23495 nox64 isa attributes.
23497 2013-03-18 Richard Biener <rguenther@suse.de>
23499 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23501 (merge_node_constraints): Likewise.
23502 (build_succ_graph): Likewise.
23503 (valid_graph_edge): Inline into single caller.
23504 (unify_nodes): Likewise. Use bitmap_set_bit return value
23506 (scc_visit): Fix formatting and variable use.
23507 (do_sd_constraint): Use gcc_checking_assert.
23508 (do_ds_constraint): Likewise.
23509 (do_complex_constraint): Likewise.
23510 (condense_visit): Likewise. Cleanup.
23511 (dump_pred_graph): New function.
23512 (perform_var_substitution): Dump the pred-graph before
23513 variable substitution.
23514 (find_equivalent_node): Use gcc_checking_assert.
23515 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23517 2013-03-18 Richard Biener <rguenther@suse.de>
23519 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23520 Remove cond_expr_stmt_list argument and do not gimplify the
23522 (vect_loop_versioning): Adjust.
23523 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23524 Cleanup to use less temporaries.
23525 (vect_create_data_ref_ptr): Cleanup.
23527 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23529 PR tree-optimization/56635
23530 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23531 require types_compatible_p types.
23533 2013-03-18 Nick Clifton <nickc@redhat.com>
23535 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23536 spurious backslash.
23538 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23539 Add missing line to comment describing function.
23541 2013-03-18 Richard Biener <rguenther@suse.de>
23543 PR tree-optimization/56210
23544 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23545 Handle string / character search functions.
23546 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23548 2013-03-18 Richard Biener <rguenther@suse.de>
23550 PR middle-end/56483
23551 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23552 and implement properly.
23553 * gimple.h (gimple_cond_single_var_p): Remove.
23555 2013-03-18 Richard Biener <rguenther@suse.de>
23557 * tree-data-ref.h (find_data_references_in_loop): Declare.
23558 * tree-data-ref.c (get_references_in_stmt): Use a stack
23559 vector pre-allocated in the callers.
23560 (find_data_references_in_stmt): Adjust.
23561 (graphite_find_data_references_in_stmt): Likewise.
23562 (create_rdg_vertices): Likewise.
23563 (find_data_references_in_loop): Export.
23564 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23565 Compute dependences here...
23566 (vect_analyze_data_refs): ...not here. When we encounter
23567 a non-vectorizable data reference in basic-block vectorization
23568 truncate the data reference vector. Do not bother to
23569 fixup data-dependence information for gather loads.
23570 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23571 of data references, as reported.
23573 2013-03-18 Richard Biener <rguenther@suse.de>
23575 PR tree-optimization/3713
23576 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23577 has_constants and expr.
23578 (stmt_has_constants): Properly valueize SSA names when deciding
23579 whether the stmt has constants.
23581 2013-03-18 Richard Biener <rguenther@suse.de>
23583 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23584 whole function when there is nothing to do.
23585 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23586 * tree-vectorizer.c (vectorize_loops): Update virtual and
23587 loop-closed SSA once.
23588 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23590 2013-03-18 Richard Biener <rguenther@suse.de>
23592 PR middle-end/56113
23593 * domwalk.c (bb_postorder): New global static.
23594 (cmp_bb_postorder): New function.
23595 (walk_dominator_tree): Replace scheme imposing an order for
23596 visiting dominator sons by one sorting them at the time they
23597 are pushed on the stack.
23599 2013-03-18 Richard Biener <rguenther@suse.de>
23601 PR tree-optimization/39326
23602 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23603 (struct mem_ref): Replace mem member with ao_ref typed member.
23604 (MEM_ANALYZABLE): Adjust.
23605 (memref_eq): Likewise.
23606 (mem_ref_alloc): Likewise.
23607 (gather_mem_refs_stmt): Likewise.
23608 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23609 (execute_sm_if_changed_flag_set): Adjust.
23610 (execute_sm): Likewise.
23611 (ref_always_accessed_p): Likewise.
23612 (refs_independent_p): Likewise.
23613 (can_sm_ref_p): Likewise.
23615 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23618 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23619 return 1 even for !unsignedp.
23621 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23623 * config/i386/i386.md (isa): Add x64 and nox64.
23624 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23625 (*pushtf): Enable *roF alternative for x64 isa only.
23626 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23627 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23628 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23629 (*movtf_internal): Merge from *movtf_internal_rex64 and
23630 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23631 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23632 nox64 isa attributes.
23633 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23634 nox64 isa attributes.
23635 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23637 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23639 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23641 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23644 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23645 class names. Remove trailing comma after "ALL_REGS".
23647 2013-03-16 Jan Hubicka <jh@suse.cz>
23649 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23650 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23651 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23652 of cgraph_get_create_node.
23653 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23655 2013-03-16 Jason Merrill <jason@redhat.com>
23658 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23659 with DW_AT_default_value.
23661 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23663 * BASE-VER: Set to 4.9.0.
23665 2013-03-14 Andi Kleen <ak@linux.intel.com>
23668 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23669 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23670 Document _x* TSX intrinsics.
23672 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23673 David Holsgrove <david.holsgrove@xilinx.com>
23675 * configure.ac: Add MicroBlaze TLS support detection.
23676 * configure: Regenerate.
23677 * config/microblaze/microblaze-protos.h
23678 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23679 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23680 * config/microblaze/microblaze.c (microblaze_address_type): Add
23681 ADDRESS_TLS and tls_reloc address types.
23682 (microblaze_address_info): Add tls_reloc.
23683 (TARGET_HAVE_TLS): Define.
23684 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23685 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23686 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23687 load_tls_operand, microblaze_call_tls_get_addr,
23688 microblaze_legitimize_tls_address): New functions.
23689 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23690 (get_base_reg): Use microblaze_tls_symbol_p.
23691 (microblaze_classify_address): Handle TLS.
23692 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23693 label_mentioned_p and microblaze_tls_referenced_p.
23694 (microblaze_legitimize_address): Handle TLS.
23695 (microblaze_address_insns): Handle ADDRESS_TLS.
23696 (pic_address_needs_scratch): Handle TLS.
23697 (print_operand_address): Handle TLS.
23698 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23699 (microblaze_expand_move): Handle TLS.
23700 (microblaze_legitimate_constant_p): Check
23701 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23702 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23703 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23704 (PIC_OFFSET_TABLE_REGNUM): Set.
23705 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23706 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23707 (addsi3, movsi_internal2, movdf_internal): Update constraints
23708 * config/microblaze/predicates.md (arith_plus_operand): Define
23709 (move_operand): Redefine as move_src_operand,
23710 check microblaze_tls_referenced_p.
23712 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23714 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23715 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23717 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23719 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23722 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23724 PR tree-optimization/53265
23725 * common.opt (Waggressive-loop-optimizations): New option.
23726 * tree-ssa-loop-niter.c: Include tree-pass.h.
23727 (do_warn_aggressive_loop_optimizations): New function.
23728 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23729 if number_of_latch_executions returned constant.
23730 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23731 early. If number_of_latch_executions returned constant, set
23732 nb_iterations_upper_bound back to it.
23733 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23735 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23736 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23738 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23739 (MULTILIB_OSDIRNAMES): Set.
23740 * genmultilib: If defaultosdirname doesn't start with :: , set
23741 defaultosdirname2 instead, clear it and emit two . multilib_raw
23742 entries instead of just one.
23744 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23746 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23747 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23748 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23749 (SUBTARGET_OVERRIDE_OPTIONS): New.
23751 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23754 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23755 (musermode): Convert to Var(TARGET_USERMODE).
23756 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23757 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23758 * config/sh/sh.c (sh_option_override): Use
23759 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23760 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23762 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23764 (udivsi3_i4_single, divsi3_i4_single): Use
23765 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23767 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
23769 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23772 2013-03-13 Richard Biener <rguenther@suse.de>
23774 PR tree-optimization/56608
23775 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23776 calls when vectorizing basic-blocks.
23778 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23781 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23784 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23786 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23788 2013-03-11 Jan Hubicka <jh@suse.cz>
23791 * lto-streamer-out.c (output_symbol_p): Skip references from
23792 constructors of external variables.
23794 2013-03-11 Jan Hubicka <jh@suse.cz>
23796 PR middle-end/56571
23797 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23799 * emit-rtl.c (verify_rtx_sharing): Likewise.
23800 (copy_insn_1): Likewise.
23801 * rtl.c (copy_rtx): Likewise.
23803 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23806 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23807 output_operand_lossage message.
23809 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23812 * arm.c (shift_op): Validate RTL pattern on the fly.
23813 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23816 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23819 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23820 new call patterns that clobber %fr12.
23821 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23822 split and postreload patterns.
23823 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23824 registers %fr12 and %fr12R as call used.
23826 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23828 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23829 (canon_address, record_store, replace_read, check_mem_read_rtx,
23830 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23831 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23832 rest_of_handle_dse): Likewise.
23834 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23836 PR middle-end/56524
23837 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
23839 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
23840 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
23841 (save_optabs_if_changed): Replace with...
23842 (init_tree_optimization_optabs): ...this.
23843 * optabs.c (save_optabs_if_changed): Rename to...
23844 (init_tree_optimization_optabs): ...this. Take the optimization node
23845 as argument. Do nothing if the base optabs are already correct.
23846 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
23847 to recompute optabs.
23848 * function.h (function): Remove optabs field.
23849 * function.c (invoke_set_current_function_hook): Call
23850 init_tree_optimization_optabs. Use the result to initialize
23853 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
23855 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
23856 if GTMA_HAS_NO_INSTRUMENTATION.
23857 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
23858 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
23859 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
23860 * gimple-pretty-print.c (dump_gimple_transaction): Handle
23861 GTMA_HAS_NO_INSTRUMENTATION.
23863 2013-03-08 Jakub Jelinek <jakub@redhat.com>
23865 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
23868 2013-03-08 Marek Polacek <polacek@redhat.com>
23869 Jakub Jelinek <jakub@redhat.com>
23871 PR tree-optimization/56478
23872 * predict.c (is_comparison_with_loop_invariant_p): Change the
23873 type of loop_step to tree.
23874 (predict_loops): Adjust.
23875 (predict_iv_comparison): Perform the computations on double_ints.
23877 2013-03-08 Richard Biener <rguenther@suse.de>
23879 PR tree-optimization/56570
23880 * tree-cfg.c (verify_expr_location_1): Verify locations for
23882 * tree-sra.c (create_access_replacement): Strip locations
23883 from DECL_DEBUG_EXPRs.
23885 2013-03-08 Richard Biener <rguenther@suse.de>
23887 * tree-inline.c (expand_call_inline): Do not associate
23888 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
23889 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
23891 2013-03-08 Richard Biener <rguenther@suse.de>
23893 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
23894 or block changes with -Og. Fix for location / block encoding
23895 changes and PHI arguments with locations.
23897 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
23899 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
23901 (struct output_info): Likewise.
23902 (register_overhead): Remove bad gcc_assert.
23903 (bitmap_find_bit): If there is only a single bitmap element, do not
23904 count a miss as a search.
23905 (print_statistics): Update for counter type changes.
23906 (dump_bitmap_statistics): Likewise. Print headers such that they
23907 are properly lined up with the printed counters.
23909 2013-03-07 Jakub Jelinek <jakub@redhat.com>
23911 PR tree-optimization/56559
23912 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
23913 check that it has only a single use.
23915 2013-03-07 Richard Biener <rguenther@suse.de>
23917 * doc/invoke.texi (fwhole-program): Discourage use in combination
23920 2013-03-06 Jakub Jelinek <jakub@redhat.com>
23922 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
23924 PR tree-optimization/56539
23925 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
23926 instead of GSI_CONTINUE_LINKING as last argument to
23927 force_gimple_operand_gsi. Adjust function comment.
23929 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
23932 PR middle-end/56548
23933 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
23934 promoted mode, convert the result back to the original mode.
23936 2013-03-06 Richard Biener <rguenther@suse.de>
23938 PR middle-end/56294
23939 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
23940 (insert_updated_phi_nodes_compare_uids): New function.
23941 (update_ssa): Sort symbols_to_rename after UID before
23942 traversing it to insert PHI nodes.
23944 2013-03-06 Richard Biener <rguenther@suse.de>
23946 PR middle-end/50494
23947 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
23948 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
23951 2013-02-13 Richard Biener <rguenther@suse.de>
23954 * varasm.c (output_constant_def_1): Get the decl representing
23955 the constant as argument.
23956 (output_constant_def): Wrap output_constant_def_1.
23957 (make_decl_rtl): Use output_constant_def_1 with the decl
23958 representing the constant.
23959 (build_constant_desc): Optionally re-use a decl already
23960 representing the constant.
23961 (tree_output_constant_def): Adjust.
23963 2013-03-06 Joey Ye <joey.ye@arm.com>
23966 * gcc.c (convert_white_space): New function.
23967 (main): Handles white space in function name.
23969 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
23972 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
23973 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
23974 to SH_DIV_CALL_TABLE for TARGET_SH2.
23975 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
23977 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
23978 call-table options.
23980 2013-03-05 Sterling Augustine <saugustine@google.com>
23981 Cary Coutant <ccoutant@google.com>
23984 * dwarf2out.c (resolve_addr): Don't call
23985 remove_loc_list_addr_table_entries a second time for the same
23988 2013-03-05 Jakub Jelinek <jakub@redhat.com>
23991 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
23992 (avoid_complex_debug_insns): New function.
23993 (expand_debug_locations): Call it.
23995 PR rtl-optimization/56484
23996 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
23997 lifetimes of hard registers on small register class machines.
23999 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24001 * config/microblaze/microblaze-protos.h: Rename
24002 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
24003 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
24005 (microblaze_fast_interrupt_function_p): New function.
24006 (microblaze_is_interrupt_handler): Rename to
24007 microblaze_is_interrupt_variant and add fast_interrupt check.
24008 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
24009 (save_restore_insns): Likewise.
24010 (compute_frame_size): Likewise.
24011 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
24012 (microblaze_globalize_label): Likewise.
24013 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
24014 * config/microblaze/microblaze.md: Use wrapper
24015 microblaze_is_interrupt_variant.
24017 2013-03-05 Kai Tietz <ktietz@redhat.com>
24019 * sdbout.c (sdbout_one_type): Switch to current function's section
24020 supporting cold/hot.
24022 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24024 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
24027 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24029 PR middle-end/56461
24030 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
24031 if VALGRIND_GET_VBITS is defined, temporarily make object
24032 memory all defined, and restore previous valgrind addressability
24033 and definability afterwards. Free this_object at the end.
24035 PR middle-end/56461
24036 * lra.c (lra): Call lra_clear_live_ranges if live_p,
24037 right before calling lra_create_live_ranges, also call it
24038 when clearing live_p. Only call lra_clear_live_ranges
24039 at the end if live_p.
24041 PR middle-end/56461
24042 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
24044 2013-03-05 Richard Biener <rguenther@suse.de>
24046 PR tree-optimization/56521
24047 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
24050 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
24053 * except.h (remove_unreachable_eh_regions): New prototype.
24054 * except.c (remove_eh_handler_splicer): New function, split out
24055 of remove_eh_handler.
24056 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
24057 warning about running it on many EH regions one at a time.
24058 (remove_unreachable_eh_regions_worker): New function, walk the
24059 EH tree in depth-first order and remove non-marked regions.
24060 (remove_unreachable_eh_regions): New function.
24061 * tree-eh.c (mark_reachable_handlers): New function, split out
24062 from remove_unreachable_handlers.
24063 (remove_unreachable_handlers): Use mark_reachable_handlers and
24064 remove_unreachable_eh_regions.
24065 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
24066 and remove_unreachable_eh_regions.
24068 2013-03-05 Richard Biener <rguenther@suse.de>
24070 PR middle-end/56525
24071 * loop-init.c (fix_loop_structure): Remove loops in two stages,
24072 not freeing them until the end.
24074 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24076 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
24078 2013-03-05 Richard Biener <rguenther@suse.de>
24080 PR tree-optimization/56270
24081 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
24082 of loads after scheduling an SLP instance.
24084 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24086 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
24088 (check_gcc_parallelize): Run guality.exp as a separate job from
24089 vect.exp with unsorted.exp and $(dg_target_exps) separately from
24090 struct-layout-1.exp with stackalign.exp.
24092 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
24094 PR middle-end/56461
24095 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
24096 load_index sbitmap even if some bit in it isn't set.
24098 PR middle-end/56461
24099 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
24100 (discover_iteration_bound_by_body_walk): Change queues to
24101 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
24102 spelling in comment. Call safe_push on queues[bound_index] directly.
24103 Release queues[queue_index] in every iteration unconditionally.
24104 Release bounds vector.
24106 PR middle-end/56461
24107 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
24108 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
24109 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
24112 2013-03-05 Richard Biener <rguenther@suse.de>
24115 * tree-inline.c (remap_blocks_to_null): New function.
24116 (expand_call_inline): When expanding a call stmt without
24117 an associated block inline remap all callee blocks to NULL.
24119 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24121 PR rtl-optimization/56494
24122 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
24123 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
24124 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
24126 PR middle-end/56461
24127 * sel-sched-ir.c (free_sched_pools): Release
24128 succs_info_pool.stack[succs_info_pool.max_top] vectors too
24129 if succs_info_pool.max_top isn't -1.
24132 * opts.c (opts_obstack, opts_concat): Moved to...
24133 * opts-common.c (opts_obstack, opts_concat): ... here.
24135 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24137 PR middle-end/56461
24138 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
24140 2013-03-04 Martin Jambor <mjambor@suse.cz>
24142 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
24143 all appropriate places.
24145 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
24147 PR tree-optimization/56424
24148 * ipa-split.c (split_function): Do not set the RSO flag if result is
24149 not by reference and its type is a register type.
24151 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
24153 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
24154 (microblaze_legitimate_pic_operand): Likewise
24155 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
24156 new function microblaze_legitimate_pic_operand
24157 * config/microblaze/microblaze-protos.h
24158 (microblaze_legitimate_pic_operand): Declare.
24160 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24162 * config/microblaze/predicates.md (call_insn_simple_operand):
24163 New predicate for supported rtx code types.
24164 * config/microblaze/microblaze.md (call_internal1): Use
24165 call_insn_simple_operand predicate.
24167 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24169 PR middle-end/56461
24170 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
24171 partitions.ordered_remove.
24173 PR middle-end/56461
24174 * tree-vect-stmts.c (vectorizable_conversion): Don't call
24175 vec_oprnds0.create (1) for modifier == NONE.
24177 PR middle-end/56461
24178 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24179 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24180 vec_oprnds1 right before pushing anything to it for
24183 PR middle-end/56461
24184 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24185 set nbbs to 0 instead of having separate code path.
24186 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24187 instead of false as last argument if returning NULL.
24189 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24191 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24192 the attribute is now called "target" instead of "option".
24193 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24194 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24195 attribute/pragma name for TARGET_OPTION_VALID_P and
24196 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24197 * doc/tm.texi: Regenerated.
24199 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24201 * config/microblaze/microblaze.c:
24202 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24203 * config/microblaze/microblaze.h: Add -mxl-reorder to
24205 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24206 instructions emitted if TARGET_REORDER.
24207 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24208 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24211 2013-03-01 Xinliang David Li <davidxl@google.com>
24213 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24216 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24218 PR middle-end/56461
24219 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24220 vector even when returning true. Fix up function comment formatting.
24222 PR middle-end/56461
24223 * ira-build.c (ira_loop_nodes_count): New variable.
24224 (create_loop_tree_nodes): Initialize it.
24225 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24227 PR middle-end/56461
24228 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24229 method on dr_chain and result_chain.
24230 * tree-vect-stmts.c (vectorizable_store): Only call
24231 result_chain.create if j == 0.
24233 PR middle-end/56461
24234 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24235 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24236 before overwriting it.
24238 2013-03-01 Tobias Burnus <burnus@net-b.de>
24240 * doc/extended.texi (C Extensions): Change order in @menu
24242 (Other MIPS Built-in Functions): Move last MIPS entry before
24243 "picoChip Built-in Functions".
24244 (SH Built-in Functions): Move after RX Built-in Functions.
24245 * doc/gcc.texi (Introduction): Change order in @menu
24247 * doc/md.texi (Constraints): Ditto.
24248 * gty.texi (Type Information): Ditto.
24249 (User-provided marking routines for template types): Make
24251 * doc/invoke.texi (AArch64 Options): Move before
24252 "Adapteva Epiphany Options".
24254 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24255 Jakub Jelinek <jakub@redhat.com>
24258 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24259 no_address_safety_analysis attribute.
24260 * doc/extend.texi (no_address_safety_attribute): Rename to
24261 no_sanitize_address attribute, mention no_address_safety_analysis
24262 attribute as deprecated alias.
24264 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24266 PR middle-end/56461
24267 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24268 type to vec<vec<tree> > *.
24269 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24270 to be vec<tree> instead of vec<tree> *, set vec_defs
24271 to vNULL and call vec_defs.create (number_of_vects), adjust other
24273 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24274 vectorizable_condition): Adjust vect_get_slp_defs callers.
24276 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24278 * config/aarch64/aarch64.c
24279 (aarch64_float_const_representable): Remove unused variable.
24281 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24283 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24285 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24287 * config/aarch64/aarch64-builtins.c
24288 (aarch64_init_simd_builtins): Make static.
24290 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24292 * config/aarch64/aarch64.c
24293 (aarch64_simd_make_constant): Make static.
24295 2013-02-28 Martin Jambor <mjambor@suse.cz>
24297 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24298 with no initialization to the RHS of debug statements.
24300 2013-02-28 Martin Jambor <mjambor@suse.cz>
24302 PR tree-optimization/56294
24303 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24305 (get_access_replacement): Do not call create_access_replacement.
24306 Assert a replacement exists.
24307 (get_repl_default_def_ssa_name): Create the replacement declaration
24310 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24312 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24313 final_end_function.
24315 2013-02-28 Marek Polacek <polacek@redhat.com>
24317 PR rtl-optimization/56466
24318 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24319 if we're changing a loop.
24320 (peel_loops_completely): Likewise.
24322 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24325 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24327 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24330 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24331 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24332 INTX_FTYPE_FX, FX_FTYPE_INTX.
24333 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24335 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24337 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24338 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24339 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24340 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24341 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24342 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24343 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24344 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24345 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24346 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24347 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24348 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24349 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24350 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24351 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24352 (avrxmega6): Increase max flash segments from 5 to 6.
24353 * config/avr/t-multilib: Regenerate.
24354 * config/avr/avr-tables.opt: Regenerate.
24355 * doc/avr-mmcu.texi: Regenerate.
24357 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24359 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24360 (avr_device_to_arch): Rename to avr_device_to_ld.
24361 (avr_device_to_as): New prototype.
24362 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24363 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24364 * config/avr/driver-avr.c (avr_device_to_as): New.
24365 (avr_device_to_arch): Rename to avr_device_to_ld.
24367 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24369 PR middle-end/56461
24370 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24371 method on dr_chain and result_chain.
24373 PR middle-end/56461
24374 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24375 pointer_set_destroy on not_executed_last_iteration.
24377 PR middle-end/56461
24378 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24380 PR middle-end/56461
24381 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24382 FOR_EACH_DEFINED_FUNCTION when freeing state.
24384 PR middle-end/56461
24385 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24387 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24390 PR middle-end/56461
24391 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24392 known_aggs[i].items and release known_aggs vector.
24394 PR middle-end/56461
24395 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24397 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24399 * config/microblaze/microblaze.c (microblaze_emit_compare):
24400 Use xor for EQ/NE comparisions.
24401 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24402 (cbranchsf4): Adjust operator to comparison_operator.
24404 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24406 PR middle-end/56461
24407 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24409 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24410 vec_safe_push, always update *slot.
24411 (redirect_edge_var_map_clear): Use vec_free.
24412 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24413 (free_var_map_entry): Use vec_free.
24414 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24415 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24417 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24419 PR middle-end/45472
24420 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24421 when the may_trap_p bit of the exprs being merged differs.
24422 Reorder tests for speculativeness in the logical and operator.
24424 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24426 * incpath.c (add_standard_paths): Use reconcat instead of concat
24427 where appropriate and avoid leaking memory.
24429 * opts.h: Include obstack.h.
24430 (opts_concat): New prototype.
24431 (opts_obstack): New declaration.
24432 * opts.c (opts_concat): New function.
24433 (opts_obstack): New variable.
24434 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24435 (finish_options): Use opts_concat instead of concat
24436 and XOBNEWVEC instead of XNEWVEC.
24437 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24438 generate_option): Likewise.
24439 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24440 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24443 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24444 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24446 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24448 PR middle-end/56461
24449 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24451 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24453 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24454 (arm_block_move_unaligned_straight): Likewise.
24455 (arm_adjust_block_mem): Likewise.
24457 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24460 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24461 temp, cond and label.
24462 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24465 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24466 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24469 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24470 of prologue/epilogue functions.
24473 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24474 Remove unused variable cfa_offset.
24475 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24478 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24479 type promotion to unsigned.
24482 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24483 for HOST_WIDE_INT of 32 bit / same size as int.
24484 (arm_block_move_unaligned_straight): Likewise.
24485 (arm_adjust_block_mem): Likewise.
24488 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24491 2013-02-26 Marek Polacek <polacek@redhat.com>
24493 PR tree-optimization/56426
24494 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24496 2013-02-26 Richard Biener <rguenther@suse.de>
24499 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24500 unused variable loops.
24502 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24504 PR tree-optimization/56448
24505 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24506 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24507 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24508 later operands of the references, or even first operand for
24509 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24511 PR tree-optimization/56443
24512 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24513 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24514 to type_for_mode langhook.
24516 2013-02-25 Matt Turner <mattst88@gmail.com>
24518 * doc/invoke.texi: Document r4700.
24520 2013-02-25 Richard Biener <rguenther@suse.de>
24522 PR tree-optimization/56175
24523 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24525 (simplify_bitwise_binary): ... here. Also guard the conversion
24526 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24528 2013-02-25 Catherine Moore <clm@codesourcery.com>
24531 2013-02-24 Catherine Moore <clm@codesourcery.com>
24532 Maciej W. Rozycki <macro@codesourcery.com>
24533 Tom de Vries <tom@codesourcery.com>
24534 Nathan Sidwell <nathan@codesourcery.com>
24535 Iain Sandoe <iain@codesourcery.com>
24536 Nathan Froyd <froydnj@codesourcery.com>
24537 Chao-ying Fu <fu@mips.com>
24539 * doc/extend.texi: (micromips, nomicromips, nocompression):
24540 Document new function attributes.
24541 * doc/invoke.texi (minterlink-compressed, mmicromips,
24542 m14k, m14ke, m14kec): Document new options.
24543 (minterlink-mips16): Update documentation.
24544 * doc/md.texi (ZC, ZD): Document new constraints.
24545 * configure.ac (gcc_cv_as_micromips): Check if linker
24546 supports the .set micromips directive.
24547 * configure: Regenerate.
24548 * config.in: Regenerate.
24549 * config/mips/mips-tables.opt: Regenerate.
24550 * config/mips/micromips.md: New file.
24551 * constraints.md (ZC, AD): New constraints.
24552 * config/mips/predicates.md (movep_src_register): New predicate.
24553 (movep_src_operand): New predicate.
24554 (non_volatile_mem_operand): New predicate.
24555 * config/mips/mips.md (multimem): New type.
24556 (length): Differentiate between 17-bit and 18-bit branch offsets.
24557 (MOVEP1, MOVEP2): New mode iterator.
24558 (mov_<load>l): Use ZC constraint.
24559 (mov_<load>r): Likewise.
24560 (mov_<store>l): Likewise.
24561 (mov_<store>r): Likewise.
24562 (*branch_equality<mode>_inverted): Add microMIPS support.
24563 (*branch_equality<mode>): Likewise.
24564 (*jump_absolute): Likewise.
24565 (indirect_jump_<mode>): Likewise.
24566 (tablejump_<mode>): Likewise.
24567 (<optab>_internal): Likewise.
24568 (sibcall_internal): Likewise.
24569 (sibcall_value_internal): Likewise.
24570 (prefetch): Use constraint ZD.
24571 * config/mips/mips.opt (minterlink-compressed): New option.
24572 (minterlink-mips16): Now an alias for minterlink-compressed.
24573 (mmicromips): New option.
24574 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24575 (compare_and_swap_12): Likewise.
24576 (sync_add<mode>): Likewise.
24577 (sync_<optab>_12): Likewise.
24578 (sync_old_<optab>_12): Likewise.
24579 (sync_new_<optab>_12): Likewise.
24580 (sync_nand_12): Likewise.
24581 (sync_old_nand_12): Likewise.
24582 (sync_new_nand_12): Likewise.
24583 (sync_sub<mode>): Likewise.
24584 (sync_old_add<mode>): Likewise.
24585 (sync_old_sub<mode>): Likewise.
24586 (sync_new_add<mode>): Likewise.
24587 (sync_new_sub<mode>): Likewise.
24588 (sync_<optab><mode>): Likewise.
24589 (sync_old_<optab><mode>): Likewise.
24590 (sync_new_<optab><mode>): Likewise.
24591 (sync_nand<mode>): Likewise.
24592 (sync_old_nand<mode>): Likewise.
24593 (sync_new_nand<mode>): Likewise.
24594 (sync_lock_test_and_set<mode>): Likewise.
24595 (test_and_set_12): Likewise.
24596 (atomic_compare_and_swap<mode>): Likewise.
24597 (atomic_exchange<mode>_llsc): Likewise.
24598 (atomic_fetch_add<mode>_llsc): Likewise.
24599 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24600 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24601 (umips_save_restore_pattern_p): Likewise.
24602 (umips_load_store_pair_p): Likewise.
24603 (umips_output_load_store_pair): Likewise.
24604 (umips_movep_target_p): Likewise.
24605 (umips_12bit_offset_address_p): Likewise.
24606 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24607 (mips_base_mips16): Rename this...
24608 (mips_base_compression_flags): ...to this. Update all uses.
24609 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24610 (mips_mips16_decl_p): Delete.
24611 (mips_nomips16_decl_p): Delete.
24612 (mips_get_compress_on_flags): New function.
24613 (mips_get_compress_off_flags): New function.
24614 (mips_get_compress_mode): New function.
24615 (mips_get_compress_on_name): New function.
24616 (mips_get_compress_off_name): New function.
24617 (mips_insert_attributes): Support multiple compression types.
24618 (mips_merge_decl_attributes): Likewise.
24619 (umips_12bit_offset_address_p): New function.
24620 (mips_start_function_definition): Emit .set micromips directive.
24621 (mips_call_may_need_jalx_p): New function.
24622 (mips_function_ok_for_sibcall): Add microMIPS support.
24623 (mips_print_operand_punctuation): Support short delay slots and
24625 (umips_swm_mask, umips_swm_encoding): New.
24626 (umips_build_save_restore): New function.
24627 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24628 (was_mips16_p): Remove.
24629 (old_compression_mode): New.
24630 (mips_set_compression_mode): New function.
24631 (mips_set_current_function): Add microMIPS support.
24632 (mips_option_override): Likewise.
24633 (umips_save_restore_pattern_p): New function.
24634 (umips_output_save_restore): New function.
24635 (umips_load_store_pair_p_1): New function.
24636 (umips_load_store_pair_p): New function.
24637 (umips_output_load_store_pair_1): New function.
24638 (umips_output_load_store_pair): New function.
24639 (umips_movep_target_p) New function.
24640 (mips_prepare_pch_save): Add microMIPS support.
24641 * config/mips/mips.h (TARGET_COMPRESSION): New.
24642 (TARGET_CPU_CPP_BUILTINS): Update macro
24643 to use new compression flags and to support microMIPS.
24644 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24645 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24646 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24647 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24648 (ASM_SPEC): Support mmicromips and mno-micromips.
24649 (M16STORE_REG_P): New macro.
24650 (MIPS_CALL): Support TARGET_MICROMIPS.
24651 (MICROMIPS_J): New macro.
24652 (mips_base_mips16): Rename this...
24653 (mips_base_compression_flags): ...to this.
24654 (UMIPS_12BIT_OFFSET_P): New macro.
24655 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24656 (MULTILIB_DIRNAMES): Likewise.
24658 2013-02-25 Tom de Vries <tom@codesourcery.com>
24660 PR rtl-optimization/56131
24661 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24662 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24663 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24665 2013-02-25 Tobias Burnus <burnus@net-b.de>
24667 * doc/invoke.texi (-fsanitize=): Move from optimization
24668 to debugging options.
24670 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24672 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24674 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24675 Alexander Monakov <amonakov@ispras.ru>
24677 PR middle-end/56077
24678 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24679 flush pending lists also on non-jumps. Adjust comment.
24681 2013-02-24 Catherine Moore <clm@codesourcery.com>
24682 Maciej W. Rozycki <macro@codesourcery.com>
24683 Tom de Vries <tom@codesourcery.com>
24684 Nathan Sidwell <nathan@codesourcery.com>
24685 Iain Sandoe <iain@codesourcery.com>
24686 Nathan Froyd <froydnj@codesourcery.com>
24687 Chao-ying Fu <fu@mips.com>
24689 * doc/extend.texi: (micromips, nomicromips, nocompression):
24690 Document new function attributes.
24691 * doc/invoke.texi (minterlink-compressed, mmicromips,
24692 m14k, m14ke, m14kec): Document new options.
24693 (minterlink-mips16): Update documentation.
24694 * doc/md.texi (ZC, ZD): Document new constraints.
24695 * configure.ac (gcc_cv_as_micromips): Check if linker
24696 supports the .set micromips directive.
24697 * configure: Regenerate.
24698 * config.in: Regenerate.
24699 * config/mips/mips-tables.opt: Regenerate.
24700 * config/mips/micromips.md: New file.
24701 * constraints.md (ZC, AD): New constraints.
24702 * config/mips/predicates.md (movep_src_register): New predicate.
24703 (movep_src_operand): New predicate.
24704 (non_volatile_mem_operand): New predicate.
24705 * config/mips/mips.md (multimem): New type.
24706 (length): Differentiate between 17-bit and 18-bit branch offsets.
24707 (MOVEP1, MOVEP2): New mode iterator.
24708 (mov_<load>l): Use ZC constraint.
24709 (mov_<load>r): Likewise.
24710 (mov_<store>l): Likewise.
24711 (mov_<store>r): Likewise.
24712 (*branch_equality<mode>_inverted): Add microMIPS support.
24713 (*branch_equality<mode>): Likewise.
24714 (*jump_absolute): Likewise.
24715 (indirect_jump_<mode>): Likewise.
24716 (tablejump_<mode>): Likewise.
24717 (<optab>_internal): Likewise.
24718 (sibcall_internal): Likewise.
24719 (sibcall_value_internal): Likewise.
24720 (prefetch): Use constraint ZD.
24721 * config/mips/mips.opt (minterlink-compressed): New option.
24722 (minterlink-mips16): Now an alias for minterlink-compressed.
24723 (mmicromips): New option.
24724 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24725 (compare_and_swap_12): Likewise.
24726 (sync_add<mode>): Likewise.
24727 (sync_<optab>_12): Likewise.
24728 (sync_old_<optab>_12): Likewise.
24729 (sync_new_<optab>_12): Likewise.
24730 (sync_nand_12): Likewise.
24731 (sync_old_nand_12): Likewise.
24732 (sync_new_nand_12): Likewise.
24733 (sync_sub<mode>): Likewise.
24734 (sync_old_add<mode>): Likewise.
24735 (sync_old_sub<mode>): Likewise.
24736 (sync_new_add<mode>): Likewise.
24737 (sync_new_sub<mode>): Likewise.
24738 (sync_<optab><mode>): Likewise.
24739 (sync_old_<optab><mode>): Likewise.
24740 (sync_new_<optab><mode>): Likewise.
24741 (sync_nand<mode>): Likewise.
24742 (sync_old_nand<mode>): Likewise.
24743 (sync_new_nand<mode>): Likewise.
24744 (sync_lock_test_and_set<mode>): Likewise.
24745 (test_and_set_12): Likewise.
24746 (atomic_compare_and_swap<mode>): Likewise.
24747 (atomic_exchange<mode>_llsc): Likewise.
24748 (atomic_fetch_add<mode>_llsc): Likewise.
24749 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24750 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24751 (umips_save_restore_pattern_p): Likewise.
24752 (umips_load_store_pair_p): Likewise.
24753 (umips_output_load_store_pair): Likewise.
24754 (umips_movep_target_p): Likewise.
24755 (umips_12bit_offset_address_p): Likewise.
24756 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24757 (mips_base_mips16): Rename this...
24758 (mips_base_compression_flags): ...to this. Update all uses.
24759 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24760 (mips_mips16_decl_p): Delete.
24761 (mips_nomips16_decl_p): Delete.
24762 (mips_get_compress_on_flags): New function.
24763 (mips_get_compress_off_flags): New function.
24764 (mips_get_compress_mode): New function.
24765 (mips_get_compress_on_name): New function.
24766 (mips_get_compress_off_name): New function.
24767 (mips_insert_attributes): Support multiple compression types.
24768 (mips_merge_decl_attributes): Likewise.
24769 (umips_12bit_offset_address_p): New function.
24770 (mips_start_function_definition): Emit .set micromips directive.
24771 (mips_call_may_need_jalx_p): New function.
24772 (mips_function_ok_for_sibcall): Add microMIPS support.
24773 (mips_print_operand_punctuation): Support short delay slots and
24775 (umips_swm_mask, umips_swm_encoding): New.
24776 (umips_build_save_restore): New function.
24777 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24778 (was_mips16_p): Remove.
24779 (old_compression_mode): New.
24780 (mips_set_compression_mode): New function.
24781 (mips_set_current_function): Add microMIPS support.
24782 (mips_option_override): Likewise.
24783 (umips_save_restore_pattern_p): New function.
24784 (umips_output_save_restore): New function.
24785 (umips_load_store_pair_p_1): New function.
24786 (umips_load_store_pair_p): New function.
24787 (umips_output_load_store_pair_1): New function.
24788 (umips_output_load_store_pair): New function.
24789 (umips_movep_target_p) New function.
24790 (mips_prepare_pch_save): Add microMIPS support.
24791 * config/mips/mips.h (TARGET_COMPRESSION): New.
24792 (TARGET_CPU_CPP_BUILTINS): Update macro
24793 to use new compression flags and to support microMIPS.
24794 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24795 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24796 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24797 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24798 (ASM_SPEC): Support mmicromips and mno-micromips.
24799 (M16STORE_REG_P): New macro.
24800 (MIPS_CALL): Support TARGET_MICROMIPS.
24801 (MICROMIPS_J): New macro.
24802 (mips_base_mips16): Rename this...
24803 (mips_base_compression_flags): ...to this.
24804 (UMIPS_12BIT_OFFSET_P): New macro.
24805 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24806 (MULTILIB_DIRNAMES): Likewise.
24808 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24811 * target-globals.c (save_target_globals): For init_reg_sets and
24812 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24814 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
24816 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24817 * config/aarch64/t-aarch64
24818 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24820 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24822 PR inline-asm/56148
24823 * lra-constraints.c (process_alt_operands): Reload operand
24824 conflicting with earlier clobber only if no more other conflicting
24827 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24830 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24831 if not linking a shared library.
24833 2013-02-22 Seth LaForge <sethml@google.com>
24835 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24837 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
24839 * config/arm/arm.md (split for extendsidi): Update condition.
24840 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
24841 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
24842 (qhs_zextenddi_cstr): Likewise.
24844 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24846 PR middle-end/56420
24847 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
24848 avoid signed wrapping.
24849 (expand_mult): Handle properly multiplication by
24850 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
24851 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
24852 in the compiler if coeff is HOST_WIDE_INT_MIN.
24853 (expand_divmod): Don't make ext_op1 static, change it's type to
24854 uhwi. Avoid undefined behavior in -INTVAL (op1).
24856 PR rtl-optimization/50339
24857 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
24859 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
24860 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
24861 into splitting_ashiftrt field.
24862 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
24864 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
24867 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
24869 PR middle-end/56108
24870 * trans-mem.c (execute_tm_mark): Do not expand transactions that
24871 are sure to go irrevocable.
24873 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
24875 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
24876 scalars are valid operands.
24878 2013-02-21 Martin Jambor <mjambor@suse.cz>
24880 PR tree-optimization/56310
24881 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
24882 only matching indices and non-negative final offsets.
24883 (intersect_aggregates_with_edge): Pass src_idx to
24884 agg_replacements_to_vector. Pass src_idx insstead of index to
24885 intersect_with_agg_replacements.
24887 2013-02-21 Martin Jambor <mjambor@suse.cz>
24889 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
24890 instead of hard-wired defaults.
24892 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
24894 * doc/invoke.texi (MIPS Options): Update documentation of the
24895 floating-point multiply-accumulate instruction restrictions.
24897 2013-02-21 Kostya Serebryany <kcc@google.com>
24899 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
24900 asan_shadow_offset on x86_64 linux.
24902 2013-02-21 Richard Biener <rguenther@suse.de>
24904 PR tree-optimization/56415
24906 2013-02-11 Richard Biener <rguenther@suse.de>
24908 PR tree-optimization/56273
24909 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24912 2013-02-21 Jakub Jelinek <jakub@redhat.com>
24915 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
24918 PR inline-asm/56405
24919 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
24920 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
24922 2013-02-20 Jan Hubicka <jh@suse.cz>
24924 PR tree-optimization/56265
24925 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
24926 when target is referenced for first time.
24928 2013-02-20 Richard Biener <rguenther@suse.de>
24930 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
24931 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
24932 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
24933 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
24934 not return anything.
24935 (rename_ssa_copies): Do not remove unused locals.
24936 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
24937 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
24938 * passes.c (execute_function_todo): Do not schedule unused locals
24939 removal if cleanup_tree_cfg did something.
24940 * tree-ssa-live.c (remove_unused_locals): Dump statistics
24941 about the number of removed locals.
24943 2013-02-20 Richard Biener <rguenther@suse.de>
24945 PR tree-optimization/56398
24946 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
24948 2013-02-20 Martin Jambor <mjambor@suse.cz>
24950 PR tree-optimization/55334
24951 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
24952 restricted pointers to arrays.
24954 2013-02-20 Richard Biener <rguenther@suse.de>
24955 Jakub Jelinek <jakub@redhat.com>
24957 PR tree-optimization/56396
24958 * tree-ssa-ccp.c (n_const_val): New static variable.
24959 (get_value): Return NULL for SSA names we don't have a lattice
24961 (ccp_initialize): Initialize n_const_val.
24962 * tree-ssa-copy.c (n_copy_of): New static variable.
24963 (init_copy_prop): Initialize n_copy_of.
24964 (get_value): Return NULL_TREE for SSA names we don't have a
24967 2013-02-20 Martin Jambor <mjambor@suse.cz>
24969 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
24971 2013-02-20 Richard Biener <rguenther@suse.de>
24973 * genpreds.c (write_lookup_constraint): Do not compare first
24974 letter of the constraint again.
24976 2013-02-20 Richard Biener <rguenther@suse.de>
24978 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
24980 (get_use_iv_cost): Terminate hashtable walk when coming across
24983 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
24985 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
24986 reassociation for avx2 targets.
24988 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24990 * config/microblaze/microblaze.c: microblaze_has_clz = 0
24991 Add version check for v8.10.a to enable microblaze_has_clz
24992 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
24993 version and TARGET_PATTERN_COMPARE check
24994 * config/microblaze/microblaze.md: New clzsi2 instruction
24996 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24998 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
24999 function before branching.
25001 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25003 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
25005 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
25007 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25009 PR middle-end/55889
25010 * sel-sched.c: Include ira.h.
25011 (implicit_clobber_conflict_p): New function.
25012 (moveup_expr): Use it.
25013 * Makefile.in (sel-sched.o): Depend on ira.h.
25015 2013-02-19 Richard Biener <rguenther@suse.de>
25017 PR tree-optimization/56384
25018 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
25019 (vn_hash_type): Split out from ...
25020 (vn_hash_constant_with_type): ... here.
25021 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
25022 (vn_phi_eq): Compare types from vn_phi_s structure.
25023 (vn_phi_lookup): Populate vn_phi_s type.
25024 (vn_phi_insert): Likewise.
25026 2013-02-19 Jakub Jelinek <jakub@redhat.com>
25028 PR tree-optimization/56350
25029 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
25030 if haven't found reduction or nested cycle operand, rather than
25031 asserting we must find it.
25033 PR tree-optimization/56381
25034 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
25037 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
25038 Jakub Jelinek <jakub@redhat.com>
25041 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
25042 (swap_optab_enable): Same.
25043 (init_all_optabs): Use argument instead of global.
25044 * tree.h (struct tree_optimization_option): New field target_optabs.
25045 * expr.h (init_all_optabs): Add argument to prototype.
25046 (TREE_OPTIMIZATION_OPTABS): New.
25047 (save_optabs_if_changed): Protoize.
25048 * optabs.h: Declare this_fn_optabs.
25049 * optabs.c (save_optabs_if_changed): New.
25050 Declare this_fn_optabs.
25051 (init_optabs): Add argument to init_all_optabs() call.
25052 * function.c (invoke_set_current_function_hook): Handle per
25054 * function.h (struct function): New field optabs.
25055 * config/mips/mips.c (mips_set_mips16_mode): Handle when
25056 optimization_current_node has changed.
25057 * target-globals.h (save_target_globals_default_opts): Protoize.
25058 * target-globals.c (save_target_globals_default_opts): New.
25060 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25063 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
25064 registers %fr12 and %fr12R as call used.
25067 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
25068 and HImode, require all displacements to be an integer multiple of
25070 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
25071 only allow QImode and HImode when reload is in progress and strict is
25072 true. Likewise for symbolic addresses. Use base14_operand to check
25073 displacements in REG+BASE addresses.
25075 2013-02-18 Richard Biener <rguenther@suse.de>
25077 PR tree-optimization/56366
25078 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25079 handle sign-conversion of outer-loop initial induction value.
25081 2013-02-18 Richard Biener <rguenther@suse.de>
25083 PR middle-end/56349
25084 * cfghooks.c (merge_blocks): If we merge a latch into another
25085 block adjust references to it.
25086 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
25087 (verify_loop_structure): Verify that a recorded latch is in fact
25090 2013-02-18 Richard Biener <rguenther@suse.de>
25092 PR tree-optimization/56321
25093 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
25094 order SSA name release and virtual operand unlinking.
25096 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25098 * config/microblaze/microblaze.md (save_stack_block): Define.
25099 (restore_stack_block): Likewise.
25101 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25103 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
25104 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
25105 * config/microblaze/microblaze.c (microblaze_option_override):
25106 Bail out early for PIC modes when target does not support PIC.
25108 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25110 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
25111 Replace with a microblaze version.
25112 (microblaze_trampoline_init): Adapt for microblaze.
25113 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
25116 2013-02-16 Jakub Jelinek <jakub@redhat.com>
25117 Dodji Seketeli <dodji@redhat.com>
25120 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
25121 (instrument_mem_region_access): Do not forget to always put
25122 instrumentation of the of 'base' and 'base + len' in a "if (len !=
25123 0) statement, even for cases where either 'base' or 'base + len'
25124 are not instrumented -- because they have been previously
25125 instrumented. Simplify the logic by putting all the statements
25126 instrument 'base + len' inside a sequence, and then insert that
25127 sequence right before the current insertion point. Then, to
25128 instrument 'base + len', just get an iterator on that statement.
25129 And do not forget to update the pointer to iterator the function
25130 received as argument.
25132 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
25134 PR rtl-optimization/56348
25135 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
25137 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
25139 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
25140 (clean_graph_dump_file): Pass base to start_graph_dump.
25142 2013-02-14 Richard Henderson <rth@redhat.com>
25145 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
25147 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
25149 * collect2-aix.h: Define F_LOADONLY.
25151 2013-02-14 Richard Biener <rguenther@suse.de>
25154 * varasm.c (output_constant_def_1): Get the decl representing
25155 the constant as argument.
25156 (output_constant_def): Wrap output_constant_def_1.
25157 (make_decl_rtl): Use output_constant_def_1 with the decl
25158 representing the constant.
25159 (build_constant_desc): Optionally re-use a decl already
25160 representing the constant.
25161 (tree_output_constant_def): Adjust.
25163 2013-02-14 Dodji Seketeli <dodji@redhat.com>
25166 * asan.c (instrument_builtin_call): Really put the length of the
25167 second source argument into src1_len.
25169 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25171 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
25172 argument. If it is false, don't create edge from then_bb to
25174 (insert_if_then_before_iter): Pass true to it.
25175 (build_check_stmt): Pass false to it.
25176 (transform_statements): Flush hash table only on extended basic
25177 block boundaries, rather than at the beginning of every bb.
25178 Don't flush hash table on nonfreeing_call_p calls.
25179 * tree-flow.h (nonfreeing_call_p): New prototype.
25180 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25182 2013-02-13 David S. Miller <davem@davemloft.net>
25184 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25186 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25189 * ira.c (max_regno_before_ira): Move from ...
25191 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25194 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25196 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25198 2013-02-13 Richard Biener <rguenther@suse.de>
25201 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25202 globals in MEM_REFs.
25204 2013-02-13 Richard Biener <rguenther@suse.de>
25206 * loop-init.c (loop_optimizer_init): Clear loop state when
25207 re-initializing preserved loops.
25208 * loop-unswitch.c (unswitch_single_loop): Return whether
25209 we unswitched the loop. Do not verify loop state here.
25210 (unswitch_loops): When we unswitched a loop discover new loops.
25212 2013-02-13 Kostya Serebryany <kcc@google.com>
25214 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25216 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25218 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25220 Avoid instrumenting duplicated memory access in the same basic block
25221 * Makefile.in (asan.o): Add new dependency on hash-table.h
25222 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25223 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25224 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25225 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25226 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25227 (get_mem_ref_of_assignment): New functions.
25228 (get_mem_refs_of_builtin_call): Extract from
25229 instrument_builtin_call and tweak a little bit to make it fit with
25231 (instrument_builtin_call): Use the new
25232 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25233 of is_gimple_builtin_call.
25234 (instrument_derefs, instrument_mem_region_access): Insert the
25235 instrumented memory reference into the hash table.
25236 (maybe_instrument_assignment): Renamed instrument_assignment into
25237 this, and change it to advance the iterator when instrumentation
25238 actually happened and return true in that case. This makes it
25239 homogeneous with maybe_instrument_assignment, and thus give a
25240 chance to callers to be more 'regular'.
25241 (transform_statements): Clear the memory reference hash table
25242 whenever we enter a new BB, when we cross a function call, or when
25243 we are done transforming statements. Use
25244 maybe_instrument_assignment instead of instrumentation. No more
25245 need to special case maybe_instrument_assignment and advance the
25246 iterator after calling it; it's now handled just like
25247 maybe_instrument_call. Update comment.
25249 2013-02-13 Richard Biener <rguenther@suse.de>
25251 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25252 Fix loop discovery code.
25254 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25256 PR inline-asm/56148
25257 * lra-constraints.c (process_alt_operands): Match early clobber
25258 operand with itself. Check conflicts with earlyclobber only if
25259 the operand is not reloaded. Prefer to reload conflicting operand
25260 if earlyclobber and matching operands are the same.
25262 2013-02-12 Richard Biener <rguenther@suse.de>
25265 * lto-streamer-out.c (write_symbol): Do not output symbols
25266 for hard register variables.
25268 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25271 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25272 (umulsidi3_insn, mulsidi3_insn): New insns.
25274 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25276 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25277 (struct tune_params): Add vec_costs field.
25278 * config/arm/arm.c (arm_builtin_vectorization_cost)
25279 (arm_add_stmt_cost): New functions.
25280 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25281 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25282 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25283 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25284 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25285 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25286 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25288 2013-02-12 Richard Biener <rguenther@suse.de>
25291 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25292 decls again if possible.
25294 2013-02-12 Richard Biener <rguenther@suse.de>
25296 PR middle-end/56288
25297 * tree-ssa.c (verify_ssa_name): Fix check, move
25298 SSA_NAME_IN_FREE_LIST check up.
25300 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25301 Steven Bosscher <steven@gcc.gnu.org>
25303 PR rtl-optimization/56151
25304 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25305 equal to op0 or op1, and last_insn pattern is CODE operation
25306 with MEM dest and one of the operands matches that MEM.
25308 2013-02-11 Sriraman Tallam <tmsriram@google.com>
25310 * doc/extend.texi: Document Function Multiversioning and "default"
25311 parameter string to target attribute.
25312 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25313 target attribute parameter is "default".
25314 (ix86_compare_version_priority): Remove checks for target attribute.
25315 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25316 Remove check for target attribute equal to NULL. Add assert.
25317 (ix86_generate_version_dispatcher_body): Change error to sorry.
25319 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25320 Jack Howarth <howarth@bromo.med.uc.edu>
25321 Patrick Marlier <patrick.marlier@gmail.com>
25324 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25325 define ENDFILE_SPEC as TM_DESTRUCTOR.
25326 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25328 2013-02-11 Alexander Potapenko <glider@google.com>
25329 Jack Howarth <howarth@bromo.med.uc.edu>
25330 Jakub Jelinek <jakub@redhat.com>
25333 * config/darwin.c (cdtor_record): Rename ctor_record.
25334 (sort_cdtor_records): Rename sort_ctor_records.
25335 (finalize_dtors): New routine to sort destructors by
25336 priority before use in assemble_integer.
25337 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25339 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25341 PR rtl-optimization/56275
25342 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25343 offset is non-negative and less than cmode size before
25344 calling simplify_subreg.
25346 2013-02-11 Richard Biener <rguenther@suse.de>
25348 PR tree-optimization/56264
25349 * cfgloop.h (fix_loop_structure): Adjust prototype.
25350 * loop-init.c (fix_loop_structure): Return the number of
25351 newly discovered loops.
25352 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25353 are discovered, do a full loop-closed SSA rewrite.
25355 2013-02-11 Richard Biener <rguenther@suse.de>
25357 PR tree-optimization/56273
25358 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25360 (check_array_ref): Fix missing newline in dumps.
25361 (search_for_addr_array): Likewise.
25363 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25365 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25367 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25370 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25372 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25374 PR rtl-optimization/56246
25375 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25377 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25378 constraints are satisfied.
25380 2013-02-08 Jeff Law <law@redhat.com>
25383 * emit-rtl.c (reg_is_parm_p): New function.
25384 * regs.h (reg_is_parm_p): New prototype.
25385 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25386 callee-clobbered registers.
25388 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25391 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25392 If there is no implicit builtin declaration, just return NULL.
25394 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25396 * config/i386/sse.md (FMAMODEM): New mode iterator.
25397 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25398 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25400 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25402 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25403 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25404 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25406 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25408 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25409 (microblaze*-*-elf): Likewise.
25410 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25412 * config/microblaze/microblaze-c.c: Add builtin defines for
25413 _LITTLE_ENDIAN and _BIG_ENDIAN.
25414 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25415 add to TARGET_DEFAULT flags.
25416 Expand ASM_SPEC and LINK_SPEC.
25417 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25418 * config/microblaze/microblaze.md: Update extendsidi2 and
25419 movdi_internal instructions to use low-order / high-order reg
25421 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25422 options and inversemask / mask of LITTLE_ENDIAN.
25423 * config/microblaze/t-microblaze: Expand multilib options to
25424 include mlittle-endian (le) and update exceptions patterns.
25426 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25428 PR rtl-optimization/56195
25429 * lra-constraints.c (get_reload_reg): Don't reuse regs
25430 if they have smaller mode than requested, if they have
25431 wider mode than requested, try to return a SUBREG.
25433 PR tree-optimization/56250
25434 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25435 if type is unsigned and code isn't MULT_EXPR.
25437 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25439 PR tree-optimization/56064
25440 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25441 bits according to mode.
25442 * fixed-value.h (fixed_from_double_int)
25443 (const_fixed_from_double_int): Adjust comments.
25445 2013-02-08 Richard Biener <rguenther@suse.de>
25448 * lto-streamer.h (struct data_in): Remove current_file, current_line
25449 and current_col members.
25450 * lto-streamer-out.c (lto_output_location): Stream changed bits
25451 en-block for efficiency.
25452 * lto-streamer-in.c (clear_line_info): Remove.
25453 (lto_input_location): Cache current file, line and column
25454 globally via local statics. Read changed bits en-block.
25455 (input_function): Do not call clear_line_info.
25456 (lto_read_body): Likewise.
25457 (lto_input_toplevel_asms): Likewise.
25459 2013-02-08 Michael Matz <matz@suse.de>
25461 PR tree-optimization/52448
25462 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25463 (nt_call_phase): New static.
25464 (add_or_mark_expr): Only mark accesses with newer phase than any
25466 (nonfreeing_call_p): New.
25467 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25468 (nt_fini_block): Keep blocks marked as visited.
25469 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25471 2013-02-08 Richard Biener <rguenther@suse.de>
25473 * ira.c (ira): Free broken dominator information.
25475 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25477 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25479 2013-02-08 Marek Polacek <polacek@redhat.com>
25481 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25483 2013-02-08 Richard Biener <rguenther@suse.de>
25485 PR middle-end/56181
25486 * cfgloop.h (flow_loops_find): Adjust.
25487 (bb_loop_header_p): Declare.
25488 * cfgloop.c (bb_loop_header_p): New function split out from ...
25489 (flow_loops_find): ... here. Adjust function signature,
25490 support incremental loop structure update.
25491 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25492 * cfgloopmanip.c (fix_loop_structure): Move ...
25493 * loop-init.c (fix_loop_structure): ... here.
25494 (apply_loop_flags): Split out from ...
25495 (loop_optimizer_init): ... here.
25496 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25497 in incremental mode, only remove dead loops here.
25499 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25502 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25503 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25504 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25505 (*round<mode>3.libgcc): New insns for fixed-modes.
25506 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25507 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25508 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25509 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25510 implementations. Define to __builtin_avr_absFX,
25511 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25512 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25513 __builtin_avr_countlsFX, respectively.
25514 * config/avr/avr-c.c (target.h): Include it.
25515 (enum avr_builtin_id): New enum.
25516 (avr_resolve_overloaded_builtin): New static function.
25517 (avr_register_target_pragmas): Use it to set
25518 targetm.resolve_overloaded_builtin.
25519 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25520 tree nodes used by DEF_BUILTIN.
25521 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25522 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25523 <AVR_BUILTIN_xxBITS>: Same.
25525 2013-02-08 Richard Biener <rguenther@suse.de>
25527 * cfgloop.c (verify_loop_structure): Properly handle
25528 a loop exiting to another loop header.
25529 * ira-int.h (ira_loops): Remove.
25530 * ira.c (ira_loops): Remove.
25531 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25532 (do_reload): Use loop_optimizer_finalize.
25533 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25534 number_of_loops to access the loop tree.
25535 (more_one_region_p): Likewise.
25536 (finish_loop_tree_nodes): Likewise.
25537 (rebuild_regno_allocno_maps): Likewise.
25538 (mark_loops_for_removal): Likewise.
25539 (mark_all_loops_for_removal): Likewise.
25540 (remove_unnecessary_regions): Likewise.
25541 (ira_build): Likewise.
25542 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25544 2013-02-08 Richard Biener <rguenther@suse.de>
25546 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25547 * ipa-pure-const.c (analyze_function): Avoid calling
25548 mark_irreducible_loops twice.
25549 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25551 2013-02-07 David S. Miller <davem@davemloft.net>
25553 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25555 * var-tracking.c (vt_add_function_parameter): Test the presence of
25556 HAVE_window_save properly and do not remap argument registers when
25557 we have a leaf function.
25559 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25562 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25564 * config/i386/i386.c (ix86_print_operand): Ditto.
25566 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25568 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25570 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25572 PR rtl-optimization/56225
25573 * lra-constraints.c (process_alt_operands): Check that reload hard
25574 reg can hold value for strict_low_part.
25576 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25579 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25580 dwarf2out_end_function.
25581 (in_first_function_p, maybe_at_text_label_p,
25582 first_loclabel_num_not_at_text_label): New variables.
25583 (dwarf2out_var_location): In the first function find out
25584 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25585 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25588 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25590 PR rtl-optimization/56178
25591 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25592 SUBREG of a register. Tidy up related block of code.
25593 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25594 note if the source is a register or a SUBREG of a register.
25596 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25599 * config/rs6000/rs6000.md (ptrm): New mode attr.
25600 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25601 call_value_indirect_aix<pttrsize>,
25602 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25605 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25607 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25608 if -bnortl. Convert to strcmp and strncmp.
25610 2013-02-07 Alan Modra <amodra@gmail.com>
25613 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25614 addresses won't wrap when offsetting.
25615 (rs6000_secondary_reload): Provide secondary reloads needed for
25616 wrapping LO_SUM addresses.
25618 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25620 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25621 MACH, just __MACH__.
25623 2013-02-06 Richard Biener <rguenther@suse.de>
25625 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25626 instead of calling fix_loop_structure.
25628 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25630 PR middle-end/56217
25631 * omp-low.c (use_pointer_for_field): Return false if
25632 lower_send_shared_vars doesn't generate any copy-out code.
25634 2013-02-06 Tom de Vries <tom@codesourcery.com>
25636 PR rtl-optimization/56131
25637 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25638 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25639 of the label is NULL. Add comment.
25641 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25643 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25646 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25647 (STATIC_LIBTSAN_LIBS): Likewise.
25648 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25649 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25650 is defined, don't add anything else beyond that.
25651 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25652 (LINK_COMMAND_SPEC): Use them.
25654 PR tree-optimization/56205
25655 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25656 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25657 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25659 2013-02-05 Richard Biener <rguenther@suse.de>
25661 PR tree-optimization/53342
25662 PR tree-optimization/53185
25663 * tree-vectorizer.h (vect_check_strided_load): Remove.
25664 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25665 not disallow peeling for vectorized strided loads.
25666 (vect_check_strided_load): Make static and simplify.
25667 (vect_analyze_data_refs): Adjust.
25668 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25669 correctly when vectorizing strided loads.
25671 2013-02-05 Richard Biener <rguenther@suse.de>
25673 * doc/install.texi: Refer to ISL, not PPL.
25675 2013-02-05 Jan Hubicka <jh@suse.cz>
25677 PR tree-optimization/55789
25678 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25680 2013-02-05 Jan Hubicka <jh@suse.cz>
25682 PR tree-optimization/55789
25683 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25684 the dead call anyway.
25686 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25689 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25691 2013-02-04 Alexander Potapenko <glider@google.com>
25692 Jack Howarth <howarth@bromo.med.uc.edu>
25693 Jakub Jelinek <jakub@redhat.com>
25696 * config/darwin.c (sort_ctor_records): Stabilized qsort
25697 on constructor priority by using original position.
25698 (finalize_ctors): New routine to sort constructors by
25699 priority before use in assemble_integer.
25700 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25702 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25705 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25706 about visibility on artificial decls.
25707 * config/sol2.c (solaris_assemble_visibility): Likewise.
25709 2013-02-04 Kai Tietz <ktietz@redhat.com>
25712 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25713 argument and improve checking of return-argument types for 16-byte
25715 (ix86_function_value_1): Add additional valtype argument on call
25716 of function_value_64.
25717 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25718 handling infunction_value_64 function.
25720 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25722 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25724 2013-02-04 Richard Biener <rguenther@suse.de>
25726 PR tree-optimization/56188
25727 * tree-ssa-structalias.c (label_visit): Consider case with
25728 initially non-empty points-to set.
25729 (perform_var_substitution): Dump node mapping and clean up.
25731 2013-02-04 Richard Guenther <rguenther@suse.de>
25734 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25735 node prevail as last resort.
25736 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25737 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25739 2013-02-04 Richard Biener <rguenther@suse.de>
25741 PR tree-optimization/56113
25742 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25744 (equiv_class_lookup_or_add): ... this.
25745 (label_visit): Adjust and fix error in previous patch.
25746 (perform_var_substitution): Adjust.
25748 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25750 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25751 * config/sh/sh4-300.md: Likewise.
25752 * config/sh/sh4a.md: Likewise.
25753 * config/sh/constraints.md: Likewise.
25754 * config/sh/sh.md: Likewise.
25755 * config/sh/netbsd-elf.h: Likewise.
25756 * config/sh/predicates.md: Likewise.
25757 * config/sh/sh-protos.h: Likewise.
25758 * config/sh/ushmedia.h: Likewise.
25759 * config/sh/linux.h: Likewise.
25760 * config/sh/sh.c: Likewise.
25761 * config/sh/superh.h: Likewise.
25762 * config/sh/elf.h: Likewise.
25763 * config/sh/sh4.md: Likewise.
25764 * config/sh/sh.h: Likewise.
25766 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25768 * config/pa/constraints.md: Adjust unused letters. Change "T"
25769 constraint to match_test floating_point_store_memory_operand().
25770 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25771 (base14_operand): New.
25772 (floating_point_store_memory_operand): New.
25773 (integer_store_memory_operand): Revise to use base14_operand and
25774 reg_plus_base_memory_operand.
25775 (move_dest_operand): Allow symbolic_memory_operands.
25776 (symbolic_memory_operand): Check for LO_SOM.
25777 (symbolic_operand): Change default case to break.
25778 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25779 CONST_DOUBLE values to be reloaded by putting them into memory when
25780 the destination is a floating point register.
25781 (movdf): Remove code to handle CONST_DOUBLE.
25783 (reload_indf_r1): New.
25784 (reload_insf_r1): New.
25785 Consistently use "Q" and "T" constraints with integer and floating
25786 point move instructions, respectively.
25787 (movdi): Remove FAIL.
25788 Change predicate for source operand unamed DImode move from
25789 general_operand to move_src_operand.
25790 (umulsidi3): Change predicate for destination operand to
25792 Likewise for similar unamed patterns.
25793 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25794 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25795 (hppa_legitimize_address): Simplify mask calculation.
25796 (pa_emit_move_sequence): Revised handling of secondary reloads from
25797 REG+D addresses for floating point loads and stores. Directly handle
25798 loading CONST0_RTX (mode) to a floating point register.
25799 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25800 to floating point registers. Don't restrict secondary reloads to
25801 floating point registers to integer modes. Revise some comments and
25803 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25804 (pa_legitimate_address_p): New.
25805 (pa_legitimize_reload_address): New.
25806 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25807 (STRICT_REG_OK_FOR_BASE_P): New.
25808 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25809 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25811 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25812 Andrew Dixie <andrewd@gentrack.com>
25814 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25817 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25819 * expmed.c (extract_bit_field_1): Pass the full width of the
25820 structure to get_best_reg_extraction_insn.
25822 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25825 * configure.ac (use_cxa_atexit): Add AIX.
25826 * configure: Regenerate.
25828 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25830 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25833 * final.c (need_profile_function): New variable.
25834 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25835 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25836 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25837 notes, targetm.asm_out.function_prologue doesn't emit anything,
25838 HAVE_prologue and profiler should be emitted before prologue,
25839 set need_profile_function instead of emitting it.
25840 (final_scan_insn): If need_profile_function, emit
25841 profile_function on the first NOTE_INSN_BASIC_BLOCK or
25842 NOTE_INSN_FUNCTION_BEG note.
25844 2013-02-01 Richard Henderson <rth@redhat.com>
25846 * config/rs6000/rs6000.md (smulditi3): New.
25849 * config/alpha/alpha.md (umulditi3): New.
25851 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25853 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
25854 (ASM_OUTPUT_ALIGNED_LOCAL): New.
25856 2013-02-01 Richard Biener <rguenther@suse.de>
25858 PR tree-optimization/56113
25859 * tree-ssa-structalias.c (label_visit): Reduce work for
25860 single-predecessor nodes.
25862 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
25864 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
25865 range isn't testing for zero.
25867 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
25869 PR middle-end/56113
25870 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
25872 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
25873 Nick Clifton <nickc@redhat.com>
25875 * config/v850/constraints.md (Q): Define as a memory constraint.
25876 * config/v850/predicates.md (label_ref_operand): New predicate.
25877 (e3v5_shift_operand): New predicate.
25878 (ior_operator): New predicate.
25879 * config/v850/t-v850: Add e3v5 multilib.
25880 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
25881 (v850_gen_movdi): Prototype.
25882 * config/v850/v850.c: Add support for e3v5 architecture.
25883 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
25885 (construct_save_jarl): Add e3v5 long JARL support.
25886 (v850_adjust_insn_length): New function. Adjust length of call
25887 insns when using e3v5 instructions.
25888 (v850_gen_movdi): New function: Generate instructions to move a
25890 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
25891 (CPP_SPEC): Define __v850e3v5__ as appropriate.
25892 (TARGET_USE_FPU): Enable for e3v5.
25893 (CONST_OK_FOR_W): New macro.
25894 (ADJUST_INSN_LENGTH): Define.
25895 * config/v850/v850.md (UNSPEC_LOOP): Define.
25896 (attr cpu): Add v850e3v5.
25897 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
25898 (movdi): New pattern.
25899 (movdi_internal): New pattern.
25900 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
25901 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
25902 (cstoresf4): Likewise.
25903 (cstoredf4): Likewise.
25904 (insv): New pattern.
25905 (rotlso3_a): New pattern.
25906 (rotlsi3_b): New pattern
25907 (rotlsi3_v850e3v5): New pattern.
25908 (doloop_begin): New pattern.
25909 (fix_loop_counter): New pattern.
25910 (doloop_end): New pattern.
25911 (branch_normal): Add e3v5 long branch support.
25912 (branch_invert): Likewise.
25913 (branch_z_normal): Likewise.
25914 (branch_z_invert): Likewise.
25915 (branch_nz_normal): Likewise.
25916 (branch_nz_invert): Likewise.
25917 (call_internal_short): Add e3v5 register-indirect JARL support.
25918 (call_internal_long): Likewise.
25919 (call_value_internal_short): Likewise.
25920 (call_value_internal_long): Likewise.
25921 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
25922 (mloop): New option.
25923 * config.gcc: Add support for configuring v840e3v5 target.
25924 * doc/invoke.texi: Document new v850 specific command line options.
25926 2013-01-31 Paul Koning <ni1d@arrl.net>
25930 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
25931 children if parent is a class.
25932 (prune_unused_types_prune): Don't add DW_AT_declaration.
25934 2013-01-31 Richard Biener <rguenther@suse.de>
25936 PR tree-optimization/56157
25937 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
25938 match up operand with SLP child.
25940 2013-01-31 Jason Merrill <jason@redhat.com>
25943 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
25944 parameters the first time.
25945 (gen_scheduled_generic_parms_dies): Check completeness here.
25947 2013-01-31 Richard Biener <rguenther@suse.de>
25949 PR middle-end/53073
25950 * common.opt (faggressive-loop-optimizations): New flag,
25951 enabled by default.
25952 * doc/invoke.texi (faggressive-loop-optimizations): Document.
25953 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
25954 infer_loop_bounds_from_undefined by it.
25956 2013-01-31 Richard Biener <rguenther@suse.de>
25958 PR tree-optimization/56150
25959 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
25960 visit virtual operands.
25961 (find_uses_to_rename_bb): Likewise.
25963 2013-01-31 Richard Biener <rguenther@suse.de>
25965 PR tree-optimization/56150
25966 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
25967 mixed store non-store stmts.
25969 2013-01-30 Jakub Jelinek <jakub@redhat.com>
25972 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
25973 LIBASAN_EARLY_SPEC is defined.
25974 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
25975 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
25977 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
25980 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
25981 invalid args instead of ICEing on it.
25982 (ix86_valid_target_attribute_tree): Return error_mark_node if
25983 ix86_valid_target_attribute_inner_p failed.
25984 (ix86_valid_target_attribute_p): Return false only if
25985 ix86_valid_target_attribute_tree returned error_mark_node. Allow
25986 target("default") attribute.
25987 (sorted_attr_string): Change argument from const char * to tree,
25988 merge in all target attribute arguments rather than just one.
25989 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
25990 instead of free. Avoid using strcat.
25991 (ix86_mangle_function_version_assembler_name): Mangle
25992 target("default") as if no target attribute is present. Adjust
25993 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
25994 instead of xmalloc and XDELETEVEC instead of free.
25995 (ix86_function_versions): Don't return true if one of the decls
25996 doesn't have target attribute. If they don't and one of the decls
25997 is DECL_FUNCTION_VERSIONED, report an error. Adjust
25998 sorted_attr_string caller. Use XDELETEVEC instead of free.
25999 (ix86_supports_function_versions): Remove.
26000 (make_name): Fix up formatting.
26001 (make_dispatcher_decl): Remove resolver_name and its initialization.
26002 Avoid leaking memory.
26003 (is_function_default_version): Return true if there is
26004 target("default") attribute rather than no target attribute at all.
26005 (make_resolver_func): Avoid leaking memory.
26006 (ix86_generate_version_dispatcher_body): Likewise.
26007 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
26008 * target.def (supports_function_versions): Remove.
26009 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
26010 * doc/tm.texi: Regenerated.
26012 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
26014 PR rtl-optimization/56144
26015 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
26016 for values with side effects.
26018 2013-01-30 Richard Biener <rguenther@suse.de>
26020 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
26021 (sparseset_pop): Likewise.
26022 * cfganal.c (compute_idf): Likewise. Increase work-stack size
26023 to be able to use quick_push in the worker loop.
26025 2013-01-30 Marek Polacek <polacek@redhat.com>
26027 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
26029 2013-01-30 Richard Biener <rguenther@suse.de>
26032 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
26034 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
26036 PR tree-optimization/56064
26037 * fixed-value.c (fixed_from_double_int): New function.
26038 * fixed-value.h (fixed_from_double_int): New prototype.
26039 (const_fixed_from_double_int): New static inline function.
26040 * fold-const.c (native_interpret_fixed): New static function.
26041 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
26042 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
26043 (native_encode_fixed): New static function.
26044 (native_encode_expr) <FIXED_CST>: Use it.
26045 (native_interpret_int): Move double_int worker code to...
26046 * double-int.c (double_int::from_buffer): ...this new static method.
26047 * double-int.h (double_int::from_buffer): Prototype it.
26049 2013-01-30 Richard Biener <rguenther@suse.de>
26051 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
26052 New pointer-map and obstack.
26053 (init_alias_vars): Allocate pointer-map and obstack.
26054 (delete_points_to_sets): Free them.
26055 (find_what_var_points_to): Cache result.
26056 (find_what_p_points_to): Adjust for changed interface of
26057 find_what_var_points_to.
26058 (compute_points_to_sets): Likewise.
26059 (ipa_pta_execute): Likewise.
26061 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26063 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
26064 * configure: Regenerate.
26065 * config.in: Regenerate.
26066 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
26067 #nobits/#progbits if supported.
26069 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
26072 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
26073 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
26074 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
26076 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26078 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
26079 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
26081 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26083 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
26084 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
26086 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26088 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
26090 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
26091 * config/arm/cortex-a7.md: New bypasses using
26092 arm_mac_accumulator_is_result.
26094 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26096 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
26097 (cortex_a7_neon_mla): Likewise.
26098 (cortex_a7_fpfmad): New reservation.
26099 (cortex_a7_fpmacs): Use ffmas and update required units.
26100 (cortex_a7_fpmuld): Update required units and latency.
26101 (cortex_a7_fpmacd): Likewise.
26102 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
26103 (cortex_a7_neon). Likewise.
26104 (bypass) Update participating units.
26106 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26108 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
26109 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
26111 * config/arm/vfp11.md (vfp_farith): Use ffmas.
26112 (vfp_fmul): Use ffmad.
26113 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
26114 (cortex_r4_fmacd): Use ffmad.
26115 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
26116 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
26117 (cortex_a9_fmacd): Use ffmad.
26118 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
26119 (cortex_a8_vfp_macd): Use ffmad.
26120 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
26121 (cortex_a5_fpmacd): Use ffmad.
26122 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
26123 (cortex_a15_vfp_macd): Use ffmad.
26124 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
26126 2013-01-29 Jason Merrill <jason@redhat.com>
26129 * varasm.c (default_assemble_visibility): Don't warn about
26130 visibility on artificial decls.
26132 2013-01-29 Richard Biener <rguenther@suse.de>
26134 PR tree-optimization/56113
26135 * tree-ssa-structalias.c (equiv_class_lookup): Also return
26137 (label_visit): Free duplicate bitmaps and record the leader instead.
26138 (perform_var_substitution): Adjust.
26140 2013-01-29 Richard Biener <rguenther@suse.de>
26142 PR tree-optimization/55270
26143 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
26144 the CFG, schedule loops for fixup.
26146 2013-01-29 Nick Clifton <nickc@redhat.com>
26148 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
26151 2013-01-28 Leif Ekblad <leif@rdos.net>
26153 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
26154 * config/i386/i386.h (TARGET_RDOS): New macro.
26155 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
26156 * config/i386/i386.c (ix86_option_override_internal): For 64bit
26157 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
26158 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
26159 DEFAULT_LARGE_SECTION_THRESHOLD.
26160 * config/i386/i386.md (R14_REG, R15_REG): New constants.
26161 * config/i386/rdos.h: New file.
26162 * config/i386/rdos64.h: New file.
26164 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
26167 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
26170 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26172 PR rtl-optimization/56117
26173 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
26174 call cselib_lookup_from_insn on the MEM before calling
26175 add_insn_mem_dependence.
26177 2013-01-28 Richard Biener <rguenther@suse.de>
26179 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26180 to a stmt that didn't have one.
26181 (copy_phis_for_bb): Likewise for PHI arguments.
26182 (copy_debug_stmt): Likewise for debug stmts.
26184 2013-01-28 Richard Biener <rguenther@suse.de>
26186 PR tree-optimization/56034
26187 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26188 (partition_builtin_p): Adjust.
26189 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26190 it is the last partition.
26191 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26192 up the vertex for the definition.
26193 (classify_partition): Classify whether a partition is a
26194 PKIND_REDUCTION, thus has uses outside of the loop.
26195 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26196 Merge all PKIND_REDUCTION partitions into the last partition.
26197 (tree_loop_distribution): Seed partitions from reductions as well.
26199 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26201 PR tree-optimization/56125
26202 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26203 pow(x,c) into sqrt(x) * powi(x, n/2) or
26204 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26205 optimizing for size.
26206 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26207 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26210 PR tree-optimization/56094
26211 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26212 to UNKNOWN_LOCATION while gimplifying expr.
26214 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26217 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26218 operand 0 in movabs insn template for -masm=intel asm alternative.
26219 (*movabs<mode>_2): Ditto for operand 1.
26221 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26224 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26227 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26229 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26232 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26234 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26235 Undef to avoid warning.
26237 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26239 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26240 * configure: Regenerate.
26242 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26244 PR tree-optimization/56098
26245 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26246 for stmts with volatile ops.
26247 (cond_store_replacement): Don't optimize if assign has volatile ops.
26248 (cond_if_else_store_replacement_1): Don't optimize if either
26249 then_assign or else_assign have volatile ops.
26250 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26253 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26255 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26257 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26259 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26260 missing ':' in asm example.
26262 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26264 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26265 entries into lane and laneq entries.
26266 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26267 Remove AdvSIMD scalar modes.
26268 (aarch64_sq<r>dmulh_laneq<mode>): New.
26269 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26271 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26272 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26273 * config/aarch64/iterators.md (VCOND): New.
26276 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26279 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26280 Add NULL LIBNAME argument to existing definitions.
26281 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26282 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26283 * config/avr/avr.c (DEF_BUILTIN): Same.
26284 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26285 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26286 implementation is available (DECL_ASSEMBLER_NAME is set).
26287 (avr_fold_absfx): New static function.
26288 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26289 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26290 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26291 AVR_BUILTIN_ABSLLK.
26292 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26293 (abshk, absk, abslk, absllk): Provide as static inline functions.
26295 2013-01-25 Marek Polacek <polacek@redhat.com>
26297 PR tree-optimization/56035
26298 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26300 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26302 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26303 (*movtf_internal_rex64): Add (!o,C) alternative
26304 (*movxf_internal_rex64): Ditto.
26305 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26307 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26309 * doc/invoke.texi: fix typo.
26310 * doc/objc.texi: fix typo.
26312 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26314 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26315 for the first two alternatives.
26317 2013-01-24 Diego Novillo <dnovillo@google.com>
26319 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26320 (ggc-zone.o): Remove.
26321 * configure.ac: Remove option --with-gc.
26322 * configure: Re-generate.
26323 * doc/install.texi: Remove documentation for --with-gc.
26324 * gengtype.c (write_enum_defn): Remove. Update all users.
26325 (write_Types_process_field): Remove generation of gt_e_* argument.
26326 (output_type_enum): Remove. Update all users.
26327 (write_enum_defn): Remove. Update all users.
26328 (enum alloc_zone): Remove. Update all users.
26329 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26330 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26331 Update all callers.
26332 (struct ptr_data): Remove field TYPE. Update all users.
26333 (gt_pch_note_object): Remove argument TYPE. Update all users.
26334 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26336 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26337 (struct alloc_zone): Remove.
26338 (ggc_internal_alloc_zone_stat): Remove.
26339 (ggc_internal_cleared_alloc_zone_stat): Remove.
26340 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26341 (ggc_pch_count_object): Remove last argument. Update all users.
26342 (ggc_pch_alloc_object): Remove last argument. Update all users.
26343 (struct alloc_zone): Remove.
26344 * ggc-zone.c: Remove.
26345 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26346 (struct alloc_zone): Remove.
26347 (ggc_alloc_typed_stat): Remove.
26348 (ggc_alloc_typed): Remove.
26349 (ggc_splay_alloc): Remove first argument.
26350 (rtl_zone): Remove. Update all users.
26351 (tree_zone): Remove. Update all users.
26352 (tree_id_zone): Remove. Update all users.
26353 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26354 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26355 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26356 * tree-ssanames.c: Remove references to zone allocator in comments.
26358 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26360 * config/avr/avr.c (avr_out_fract): Make register numbers that
26361 might be outside of source operand signed.
26363 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26365 * config/i386/constraints.md (Yf): New constraint.
26366 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26367 of f constraint to conditionaly disable x87 register preferences.
26368 (*movdf_internal): Ditto.
26369 (*movsf_internal): Ditto.
26371 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26373 PR inline-asm/55934
26374 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26375 that have operands with impossible constraints.
26376 Add a FIXME for a speed-up opportunity.
26377 * lra-constraints.c (process_alt_operands): Verify that a class
26378 selected from constraints on asms is valid for the operand mode.
26379 (curr_insn_transform): Remove incorrect comment.
26381 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26383 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26384 TOC operand is a valid symbol ref in the constant pool.
26386 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26388 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26390 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26393 * config/avr/stdfix.h: New file.
26394 * t-avr (stdfix-gcc.h): New rule to build it.
26395 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26397 2013-01-23 Kostya Serebryany <kcc@google.com>
26399 * config/darwin.h: remove dependency on
26400 CoreFoundation (asan on Mac OS).
26402 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26405 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26406 instead of cmpdi_operand for first comparison operand.
26407 Don't assert that comparison operands aren't both constants.
26409 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26411 * doc/install.texi (Downloading the Source): Update references to
26412 downloading separate components.
26414 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26416 * doc/extend.texi (__int128): Improve grammar.
26418 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26421 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26422 alternative to (o,r).
26423 (*movdi_internal_rex64): Remove (!o,n) alternative.
26424 (DImode immediate->memory splitter): Remove.
26425 (DImode immediate->memory peephole2): Remove.
26426 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26427 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26428 alternative to (!o,*r).
26429 (*movtf_internal_sse): New pattern.
26430 (*movxf_internal_rex64): New pattern.
26431 (*movxf_internal): Disable for TARGET_64BIT.
26432 (*movdf_internal_rex64): Remove (!o,F) alternative.
26434 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26436 PR middle-end/56074
26437 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26438 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26439 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26440 stmt locations where LOCATION_LOCUS of the stmt location is
26441 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26444 * config/i386/i386.md (UNSPEC_STOS): New.
26445 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26446 *strsetqi_1): Add UNSPEC_STOS.
26448 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26451 * doc/invoke.texi: Remove left over -Wsynth example.
26453 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26455 PR tree-optimization/56051
26456 * fold-const.c (fold_binary_loc): Don't fold
26457 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26458 a narrowing conversion, or widening conversion from signed
26461 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26463 PR rtl-optimization/56023
26464 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26465 dependent on debug instruction.
26467 2013-01-21 Martin Jambor <mjambor@suse.cz>
26469 PR middle-end/56022
26470 * function.c (allocate_struct_function): Call
26471 invoke_set_current_function_hook earlier.
26473 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26475 * reload1.c (init_reload): Only initialize reload_obstack
26476 during the first call.
26478 2013-01-21 Marek Polacek <polacek@redhat.com>
26480 * cfgloop.c (verify_loop_structure): Fix up grammar.
26482 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26484 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26485 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26487 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26490 * config/arm/marvell-pj4.md: Update copyright year.
26491 Fix up use of alu to alu_reg and simple_alu_imm.
26493 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26495 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26497 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26500 * lra-constraints.c (curr_insn_transform): Don't reuse original
26501 insn for secondary memory move when memory mode should be different.
26503 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26505 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26506 atomic_storedi_1): New patterns.
26508 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26510 btver2 pipeline descriptions.
26511 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26513 * config/i386/i386.md (btver2_decode): New type attributes.
26514 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26516 * config/i386/btver2.md: New file describing btver2 pipelines.
26518 2013-01-19 Andrew Pinski <apinski@cavium.com>
26520 PR tree-optimization/52631
26521 * tree-ssa-sccvn (visit_use): Before looking up the original
26522 statement, try looking up the simplified expression.
26524 2013-01-19 Anthony Green <green@moxielogic.com>
26526 * config/moxie/moxie.c (moxie_expand_prologue): Set
26527 current_function_static_stack_size.
26529 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26531 PR tree-optimization/56029
26532 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26533 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26535 2013-01-18 Sharad Singhai <singhai@google.com>
26537 PR tree-optimization/55995
26538 * dumpfile.c (dump_loc): Print location only if available.
26539 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26541 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26544 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26545 secondary memory move.
26546 (inherit_reload_reg): Use rclass instead of cl for
26547 check_secondary_memory_needed_p.
26549 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26551 PR middle-end/56015
26552 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26553 the case where writing real complex part of target modifies op1.
26555 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26557 * config/aarch64/aarch64-simd.md
26558 (aarch64_vcond_internal<mode>): Handle unordered cases.
26559 * config/aarch64/iterators.md (v_cmp_result): New.
26561 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26562 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26564 * config/arm/marvell-pj4.md: New file.
26565 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26566 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26567 (generic_vfp): Likewise.
26568 * config/arm/arm-cores.def: Add marvell-pj4.
26569 * config/arm/arm-tune.md: Regenerate.
26570 * config/arm/arm-tables.opt: Regenerate.
26571 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26572 * doc/invoke.texi: Document marvell-pj4.
26574 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26576 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26578 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26583 * var-tracking.c (negative_power_of_two_p): New.
26584 (global_get_addr_cache, local_get_addr_cache): New.
26585 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26586 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26588 (vt_stack_offset_p): Remove.
26589 (vt_canon_true_dep): Always canonicalize loc's address.
26590 (clobber_overlapping_mems): Make sure we have a MEM.
26591 (local_get_addr_clear_given_value): New.
26592 (val_reset): Clear local cached entries.
26593 (compute_bb_dataflow): Create and release the local cache.
26594 Disable duplicate MEMs clobbering.
26595 (emit_notes_in_bb): Clobber MEMs likewise.
26596 (vt_emit_notes): Create and release the local cache.
26597 (vt_initialize, vt_finalize): Create and release the global
26598 cache, respectively.
26599 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26601 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26603 PR libmudflap/53359
26604 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26605 not found in the symtab.
26607 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26610 PR rtl-optimization/55547
26611 PR rtl-optimization/53827
26614 * alias.c (offset_overlap_p): New, factored out of...
26615 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26616 the conservative special case for symbolic constants. Don't
26617 adjust zero sizes on alignment.
26619 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26621 PR rtl-optimization/52573
26622 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26623 REG_UNUSED for the same register.
26625 2013-01-17 Richard Biener <rguenther@suse.de>
26626 Marek Polacek <polacek@redhat.com>
26628 PR rtl-optimization/55833
26629 * loop-unswitch.c (unswitch_loops): Move loop verification...
26630 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26631 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26632 Set it to true when we're removing a loop from hierarchy tree in
26633 an irreducible region.
26634 (fix_bb_placements): Adjust caller.
26635 (fix_loop_placements): Likewise.
26637 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26639 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26640 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26641 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26642 Remove ID. Adjust comments.
26643 * config/avr/avr-c.c (avr_builtin_name): Remove.
26644 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26645 * config/avr/avr.c (avr_tolower): New static function.
26646 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26647 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26648 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26651 2013-01-17 Jan Hubicka <jh@suse.cz>
26653 PR tree-optimization/55273
26654 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26656 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26659 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26660 store through atomic_store<mode>_1.
26661 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26663 2013-01-17 Martin Jambor <mjambor@suse.cz>
26665 PR tree-optimizations/55264
26666 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26667 for virtual methods.
26668 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26669 virtual methods before inlining is over.
26670 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26672 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26675 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26677 PR rtl-optimization/56005
26678 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26679 pending reads for prefetch.
26681 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26683 * config/aarch64/aarch64.md
26684 (*cstoresi_neg_uxtw): New pattern.
26685 (*cmovsi_insn_uxtw): New pattern.
26686 (*<optab>si3_uxtw): New pattern.
26687 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26688 (*<optab>si3_insn_uxtw): New pattern.
26689 (*bswapsi2_uxtw): New pattern.
26691 2013-01-16 Richard Biener <rguenther@suse.de>
26693 * tree-inline.c (tree_function_versioning): Remove set but
26694 never used variable.
26696 2013-01-16 Richard Biener <rguenther@suse.de>
26698 PR tree-optimization/55964
26699 * tree-flow.h (rename_variables_in_loop): Remove.
26700 (rename_variables_in_bb): Likewise.
26701 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26702 (copy_loop_before): Adjust and delete update-ssa status.
26703 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26704 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26705 (rename_variables_in_loop): Remove.
26706 (slpeel_update_phis_for_duplicate_loop): Likewise.
26707 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26708 use available cfg machinery instead of duplicating it.
26709 Update PHI nodes and perform poor-mans SSA update here.
26710 (slpeel_tree_peel_loop_to_edge): Adjust.
26712 2013-01-16 Richard Biener <rguenther@suse.de>
26714 PR tree-optimization/54767
26715 PR tree-optimization/53465
26716 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26717 (vrp_visit_phi_node): For PHI arguments coming via backedges
26718 drop all symbolical range information.
26719 (execute_vrp): Compute backedges.
26721 2013-01-16 Richard Biener <rguenther@suse.de>
26723 * doc/install.texi: Update CLooG and ISL requirements to
26726 2013-01-16 Christian Bruel <christian.bruel@st.com>
26729 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26730 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26731 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26733 2013-01-16 DJ Delorie <dj@redhat.com>
26735 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26736 (UNSPECV_SP_SWITCH_E): New.
26737 (sp_switch_1): Change to an unspec.
26738 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26741 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26743 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26744 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26745 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26746 (expand_mem_thread_fence): Ditto.
26747 (expand_mem_signal_fence): Ditto.
26748 (expand_atomic_load): Ditto.
26749 (expand_atomic_store): Ditto.
26751 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26753 PR rtl-optimization/55547
26754 PR rtl-optimization/53827
26757 * alias.c (memrefs_conflict_p): Set sizes to negative after
26760 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26763 * function.c (thread_prologue_and_epilogue_insns): Always
26764 add crtl->drap_reg to set_up_by_prologue.set, even if
26765 stack_realign_drap is false.
26767 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26769 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26770 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26771 *call): Fix indention.
26773 2013-01-15 Tom de Vries <tom@codesourcery.com>
26776 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26779 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26781 PR rtl-optimization/55153
26782 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26784 2013-01-15 Martin Jambor <mjambor@suse.cz>
26786 PR tree-optimization/55920
26787 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26788 accesses as grp_to_be_debug_replaced.
26790 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26792 PR tree-optimization/55920
26793 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26794 there is non-useless type conversion needed from debug rhs to lhs,
26795 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26797 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26798 Mikael Pettersson <mikpe@it.uu.se>
26801 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26803 (ASM_OUTPUT_CASE_LABEL): Remove.
26804 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26805 * final.c (shorten_branches): Update alignment of labels before
26806 jump tables if CASE_VECTOR_SHORTEN_MODE.
26808 2013-01-15 Richard Biener <rguenther@suse.de>
26811 * system.h: Do not include gmp.h for building host tools.
26813 2013-01-15 Richard Biener <rguenther@suse.de>
26815 PR middle-end/55882
26816 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26817 account for bitpos when computing alignment.
26819 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26821 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26822 (ix86_target_macros_internal): Likewise.
26824 * config/i386/i386.c (m_CORE2I7): Removed.
26825 (m_CORE_HASWELL): New macro.
26826 (m_CORE_ALL): Likewise.
26827 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26828 (initial_ix86_arch_features): Likewise.
26829 (processor_target_table): Initializations for Core avx2.
26830 (cpu_names): New names "core-avx2".
26831 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26832 PROCESSOR_CORE_HASWELL.
26833 (ix86_issue_rate): New case.
26834 (ia32_multipass_dfa_lookahead): Likewise.
26835 (ix86_sched_init_global): Likewise.
26837 * config/i386/i386.h (TARGET_HASWELL): New macro.
26838 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
26839 (processor_type): New PROCESSOR_HASWELL.
26841 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26843 PR tree-optimization/55955
26844 * tree-vect-loop.c (vectorizable_reduction): Give up early on
26845 *SHIFT_EXPR and *ROTATE_EXPR codes.
26847 PR tree-optimization/48766
26848 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
26849 -ftrapv disable -fwrapv.
26851 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26854 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
26855 etc. to 1 and not to __flash.
26856 Use LL suffix for __INT24_MAX__ with -mint8.
26857 Use ULL suffix for __UINT24_MAX__ with -mint8.
26859 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26861 * config/avr/avr-arch.h
26862 (struct base_arch_s): Use typedef avr_arch_t instead.
26863 (struct arch_info_s): Use typedef avr_arch_info_t instead.
26864 (struct mcu_type_s): Use typedef avr_mcu_t instead.
26865 * config/avr/avr.c: Same.
26866 * config/avr/avr-devices.c: Same.
26867 * config/avr/driver-avr.c: Same.
26868 * config/avr/gen-avr-mmcu-texi.c: Same.
26869 * config/avr/avr-mcus.def: Adjust comment.
26871 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
26873 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
26874 * config/aarch64/iterators.md (VALLDI): New.
26876 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
26877 Andi Kleen <ak@linux.intel.com>
26880 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
26881 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
26884 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26886 * config/avr/avr-stdint.h: Remove trailing blanks.
26887 * config/avr/avr-log.h: Same.
26888 * config/avr/avr-arch.h: Same.
26889 * config/avr/avr-devices.c: Same.
26890 * config/avr/avr-dimode.md: Same.
26891 * config/avr/predicates.md: Same.
26892 * config/avr/avr-c.c: Same. And fix typo.
26894 * config/avr/avr-protos.h: Same. And:
26895 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
26896 (init_cumulative_args): Rename to avr_init_cumulative_args.
26897 (expand_prologue): Rename to avr_expand_prologue.
26898 (expand_epilogue): Rename to avr_expand_epilogue.
26899 (adjust_insn_length): Rename to avr_adjust_insn_length.
26900 (notice_update_cc): Rename to avr_notice_update_cc.
26901 (final_prescan_insn): Rename to avr_final_prescan_insn.
26902 * config/avr/avr.c: Same.
26903 * config/avr/avr.h: Same.
26904 * config/avr/avr.md: Remove trailing blanks.
26905 (prologue): Use avr_expand_prologue.
26906 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
26908 2013-01-14 Richard Biener <rguenther@suse.de>
26910 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
26911 verify_location, collect_subblocks): New functions.
26912 (verify_gimple_in_cfg): Verify that locations only reference
26913 BLOCKs in the functions BLOCK tree.
26915 2013-01-14 Richard Biener <rguenther@suse.de>
26917 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
26919 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
26921 (insert_out_of_ssa_copy_on_edge): Likewise.
26922 (rewrite_close_phi_out_of_ssa): Likewise.
26923 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
26925 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
26926 propagated constants.
26927 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
26930 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
26932 * config/avr/avr-modes.def: Add GPL copyright notice.
26934 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
26936 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
26937 MEMMODEL_MASK to determine memory model.
26938 (atomic_store<mode>): Ditto from operands[2].
26939 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
26941 2013-01-13 Jakub Jelinek <jakub@redhat.com>
26944 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
26945 (fold_gimple_assign): Don't call unshare_expr here.
26946 (fold_ctor_reference): Call unshare_expr.
26948 2013-01-13 Terry Guo <terry.guo@arm.com>
26950 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
26951 * doc/fragments.texi: Document MULTILIB_REUSE.
26952 * gcc.c (multilib_reuse): New internal spec.
26953 (set_multilib_dir): Also search multilib from multilib_reuse.
26954 * genmultilib (tmpmultilib3): Refactor code.
26955 (tmpmultilib4): Ditto.
26956 (multilib_reuse): New multilib argument.
26958 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
26960 * Makefile.in: Update copyright.
26962 2013-01-12 Tom de Vries <tom@codesourcery.com>
26964 PR middle-end/55890
26965 * calls.c (expand_call): Check if arg_nr is valid.
26967 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
26969 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
26970 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
26971 documentation. Add missing '__' in front of
26972 __builtin_ia32_packssdw256.
26974 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26977 * config/s390/s390.c (s390_preferred_reload_class): Do not return
26978 NO_REGS for larl operands.
26979 (s390_reload_larl_operand): Use s390_load_address instead of
26982 2013-01-11 Richard Biener <rguenther@suse.de>
26984 * tree-cfg.c (verify_node_sharing_1): Split out from ...
26985 (verify_node_sharing): ... here.
26986 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
26988 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
26990 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
26991 Substitute TREECHECKING.
26992 * configure: Regenerate.
26993 * Makefile.in (TREECHECKING): New.
26995 2013-01-11 Richard Guenther <rguenther@suse.de>
26997 PR tree-optimization/44061
26998 * tree-vrp.c (extract_range_basic): Compute zero as
26999 value-range for __builtin_constant_p of function parameters.
27001 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
27003 Update copyright years.
27005 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
27007 PR rtl-optimization/55672
27008 * lra-eliminations.c (mark_not_eliminable): Permit addition with
27009 const to be eliminable.
27011 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
27013 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
27014 * configure: Regenerate.
27016 2013-01-10 Richard Biener <rguenther@suse.de>
27018 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
27020 2013-01-10 Richard Biener <rguenther@suse.de>
27023 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
27024 locations for virtual PHI arguments.
27025 (rewrite_update_phi_arguments): Likewise.
27027 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
27029 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
27032 2013-01-10 Jakub Jelinek <jakub@redhat.com>
27034 PR tree-optimization/55921
27035 * tree-complex.c (expand_complex_asm): New function.
27036 (expand_complex_operations_1): Call it for GIMPLE_ASM.
27038 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27041 * config/s390/s390.c (s390_symref_operand_p)
27042 (s390_loadrelative_operand_p): Merge the two functions.
27043 (s390_check_qrst_address, print_operand_address): Add parameters
27044 to s390_loadrelative_operand_p invokation.
27045 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
27046 (s390_reload_larl_operand, s390_secondary_reload): Use
27047 s390_loadrelative_operand_p instead of s390_symref_operand_p.
27048 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
27050 2013-01-09 Mike Stump <mikestump@comcast.net>
27052 * dse.c (record_store): Remove unnecessary assert.
27054 2013-01-09 Jan Hubicka <jh@suse.cz>
27056 PR tree-optimization/55569
27057 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
27058 * cfgloop.h (scale_loop_profile): Likewise.
27060 2013-01-09 Jan Hubicka <jh@suse.cz>
27063 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
27065 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
27067 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
27069 PR middle-end/55114
27070 * expr.h (maybe_emit_group_store): Declare.
27071 * expr.c (maybe_emit_group_store): New function.
27072 * builtins.c (expand_builtin_int_roundingfn): Call it.
27073 (expand_builtin_int_roundingfn_2): Likewise.
27075 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
27077 PR rtl-optimization/55829
27078 * lra-constraints.c (match_reload): Add code for absent output.
27079 (curr_insn_transform): Add code for reloads of matched inputs
27082 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
27084 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
27085 attribute of movddup insn to DF.
27086 (*vec_interleave_lowv2df): Ditto.
27087 (vec_dupv2df): Ditto.
27089 2013-01-09 Jan Hubicka <jh@suse.cz>
27091 PR tree-optimiation/55875
27092 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
27093 EVERY_ITERATION parameter.
27094 (number_of_iterations_exit): Check if exit is executed every iteration.
27095 (idx_infer_loop_bounds): Similarly here.
27096 (n_of_executions_at_most): Simplify
27097 to only test for cases where statement is dominated by the
27098 particular bound; handle correctly the "postdominance" test.
27099 (scev_probably_wraps_p): Use max loop iterations info
27100 as a global bound first.
27102 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
27103 Nick Clifton <nickc@redhat.com>
27105 * config/v850/v850.md (cbranchsf4): New pattern.
27106 (cstoresf4): New pattern.
27107 (cbranchdf4): New pattern.
27108 (cstoredf4): New pattern.
27109 (movsicc): Disallow floating point comparisons.
27110 (cmpsf_le_insn): Fix order of operators.
27111 (cmpsf_lt_insn): Likewise.
27112 (cmpsf_eq_insn): Likewise.
27113 (cmpdf_le_insn): Likewise.
27114 (cmpdf_lt_insn): Likewise.
27115 (cmpdf_eq_insn): Likewise.
27116 (cmpsf_ge_insn): Use LE comparison.
27117 (cmpdf_ge_insn): Likewise.
27118 (cmpsf_gt_insn): Use LT comparison.
27119 (cmpdf_gt_insn): Likewise.
27120 (cmpsf_ne_insn): Delete pattern.
27121 (cmpdf_ne_insn): Delete pattern.
27122 * config/v850/v850.c (v850_gen_float_compare): Use
27123 gen_cmpdf_eq_insn for NE comparison.
27124 (v850_float_z_comparison_operator)
27125 (v850_float_nz_comparison_operator): Move from here ...
27126 * config/v850/predicates.md: ... to here. Move GT and GE
27127 comparisons into v850_float_z_comparison_operator.
27128 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
27130 (v850_float_nz_comparison_operator): Likewise.
27132 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27134 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
27135 with calls to gen_insvsi/gen_insvdi.
27137 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27139 * config/i386/i386.c (initial_ix86_tune_features): Set up
27140 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
27142 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
27143 Jakub Jelinek <jakub@redhat.com>
27145 PR tree-optimization/48189
27146 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
27147 If nitercst is 0, don't predict the exit edge.
27149 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27151 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
27152 in asm_fprintf with reg_names.
27153 (aarch64_print_operand_address): Likewise.
27154 (aarch64_return_addr): Likewise.
27155 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
27157 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27159 * config/pa/pa.h (VAL_U6_BITS_P): Define.
27160 (INT_U6_BITS): Likewise.
27161 * config/pa/predicates.md (uint6_operand): New predicate.
27162 (shift5_operand, shift6_operand): Likewise.
27163 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
27165 (lshrdi3): Use shift6_operand.
27166 (shrpsi4, shrpdi4): New insn patterns.
27167 (extzv): Delete expander.
27168 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
27169 predicates in unamed zero extract patterns. Tighten common constraint.
27170 (extv): Delete expander.
27171 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
27172 predicates in unamed sign extract patterns. Tighten common constraint.
27173 (insv): Delete expander.
27174 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
27175 predicates in unamed insert patterns. Tighten common constraint.
27176 Change uint32_operand predicate to uint6_operand predicate in unamed
27177 DImode pattern to insert constant values of type 1...1xxxx.
27179 2013-01-04 Jan Hubicka <jh@suse.cz>
27181 PR tree-optimization/55823
27182 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27185 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27186 Uros Bizjak <ubizjak@gmail.com>
27188 PR rtl-optimization/55845
27189 * df-problems.c (can_move_insns_across): Stop scanning at
27190 volatile_insn_p source instruction or give up if
27191 across_from .. across_to range contains any volatile_insn_p
27194 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27196 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27197 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27199 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27200 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27202 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27205 * asan.c (asan_clear_shadow): New function.
27206 (asan_emit_stack_protection): Use it.
27208 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27210 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27211 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27212 with tab instead of space.
27214 2013-01-08 Nick Clifton <nickc@redhat.com>
27216 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27217 register bank 0 at the start of an interrupt handler.
27218 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27221 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27223 * config/aarch64/aarch64-simd.md
27224 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27225 (aarch64_simd_bsl): Likewise.
27226 (aarch64_vcond_internal<mode>): Likewise.
27227 (vcond<mode><mode>): Likewise.
27228 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27229 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27231 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27233 * config/aarch64/aarch64-builtins.c
27234 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27236 2013-01-08 Martin Jambor <mjambor@suse.cz>
27239 * tree-sra.c (analyze_access_subtree): Return true also after
27240 potentially creating a debug-only replacement.
27242 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27244 PR middle-end/55890
27245 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27247 PR tree-optimization/54120
27248 * tree-vrp.c (range_fits_type_p): Don't allow
27249 src_precision < precision from signed vr to unsigned_p
27250 if vr->min or vr->max is negative.
27251 (simplify_float_conversion_using_ranges): Test can_float_p
27252 against CODE_FOR_nothing.
27254 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27255 Richard Biener <rguenther@suse.de>
27257 PR middle-end/55851
27258 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27259 types instead of just INTEGER_TYPE types.
27261 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27263 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27266 2013-01-07 Steve Ellcey <sellcey@mips.com>
27269 * config/mips/mips.opt: Change mad to mmad to match documentation.
27271 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27274 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27275 .progmemx.data now.
27277 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27280 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27281 (avr_addrspace_t): Add .section_name field.
27282 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27284 (avr_addrspace): Same. Initialize .section_name. Remove last
27285 NULL entry. Put __memx into .progmemx.data.
27286 (progmem_section_prefix): Remove.
27287 (avr_asm_init_sections): No need to initialize progmem_section.
27288 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27289 section name prefix.
27290 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27291 retrieve the progmem section.
27292 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27293 boundary to run over avr_addrspace[].
27294 (avr_register_target_pragmas): Ditto.
27296 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27298 * varasm.c (output_constant_def_contents): For asan_protect_global
27299 protected strings, adjust DECL_ALIGN if needed, before testing for
27301 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27302 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27304 (output_object_block): For asan protected decls, emit asan padding
27305 after their contents.
27306 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27307 (asan_finish_file): Test it here instead.
27309 2013-01-07 Nick Clifton <nickc@redhat.com>
27310 Matthias Klose <doko@debian.org>
27311 Doug Kwan <dougkwan@google.com>
27312 H.J. Lu <hongjiu.lu@intel.com>
27315 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27317 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27319 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27321 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27323 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27325 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27328 * doc/install.texi (Cross-Compiler-Specific Options): Document
27331 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27333 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27334 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27335 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27336 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27337 vqmovun_high_s64): Fix source operand number and update copyright.
27339 2013-01-07 Richard Biener <rguenther@suse.de>
27341 PR middle-end/55890
27342 * gimple.h (gimple_call_builtin_p): New overload.
27343 * gimple.c (validate_call): New function.
27344 (gimple_call_builtin_p): Likewise.
27345 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27346 Use gimple_call_builtin_p.
27347 (find_func_clobbers): Likewise.
27348 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27349 (strlen_optimize_stmt): Likewise.
27351 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27353 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27354 (vld1q_dup_*): Likewise.
27355 (vld1_*): Likewise.
27356 (vld1q_*): Likewise.
27357 (vld1_lane_*): Likewise.
27358 (vld1q_lane_*): Likewise.
27360 2013-01-07 Richard Biener <rguenther@suse.de>
27362 * lto-streamer.h (LTO_minor_version): Bump to 2.
27364 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27366 * config/aarch64/aarch64-protos.h
27367 (aarch64_const_double_zero_rtx_p): Rename to...
27368 (aarch64_float_const_zero_rtx_p): ...this.
27369 (aarch64_float_const_representable_p): New.
27370 (aarch64_output_simd_mov_immediate): Likewise.
27371 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27372 move immediate case.
27373 * config/aarch64/aarch64.c
27374 (aarch64_const_double_zero_rtx_p): Rename to...
27375 (aarch64_float_const_zero_rtx_p): ...this.
27376 (aarch64_print_operand): Allow printing of new constants.
27377 (aarch64_valid_floating_const): New.
27378 (aarch64_legitimate_constant_p): Check for valid floating-point
27380 (aarch64_simd_valid_immediate): Likewise.
27381 (aarch64_vect_float_const_representable_p): New.
27382 (aarch64_float_const_representable_p): Likewise.
27383 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27384 (aarch64_output_simd_mov_immediate): New.
27385 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27386 (*movdf_aarch64): Likewise.
27387 * config/aarch64/constraints.md (Ufc): New.
27388 (Y): call aarch64_float_const_zero_rtx.
27389 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27391 2013-01-07 Richard Biener <rguenther@suse.de>
27393 PR tree-optimization/55888
27394 PR tree-optimization/55862
27395 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27396 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27397 not if it is contained therein.
27399 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27401 * config/avr/t-avr: Typo.
27403 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27406 * config/avr/t-avr: Don't automatically rebuild
27407 $(srcdir)/config/avr/t-multilib
27408 $(srcdir)/config/avr/avr-tables.opt
27409 $(srcdir)/doc/avr-mmcu.texi
27410 (avr-mcus): New phony target to build them on request.
27411 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27412 * avr/avr-mcus.def: Adjust comments.
27414 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27416 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27418 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27420 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27422 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27424 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27426 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27428 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27429 to generate profiling.
27430 * config/rs6000/aix64.h (LIB_SPEC): Same.
27432 2013-01-04 Andrew Pinski <apinski@cavium.com>
27434 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27436 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27438 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27440 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27442 (ix86_expand_move): Ditto.
27443 (ix86_zero_extend_to_Pmode): Ditto.
27444 (ix86_expand_call): Ditto.
27445 (ix86_expand_special_args_builtin): Ditto.
27446 (ix86_expand_builtin): Ditto.
27448 2013-01-04 Richard Biener <rguenther@suse.de>
27450 PR tree-optimization/55862
27451 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27452 translating them through PHI nodes.
27454 2013-01-04 Martin Jambor <mjambor@suse.cz>
27456 PR tree-optimization/55755
27457 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27458 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27460 2013-01-04 Marek Polacek <polacek@redhat.com>
27462 PR middle-end/55859
27463 * opts.c (default_options_optimization): Clarify error message.
27465 2013-01-04 Richard Biener <rguenther@suse.de>
27467 PR middle-end/55863
27468 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27471 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27474 * config/pa/pa.md (movsi): Revert previous change.
27475 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27478 2013-01-03 Richard Henderson <rth@redhat.com>
27480 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27481 after eliminating TLS symbols.
27483 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27486 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27487 * graphite-poly.c (debug_gmp_value): Likewise.
27489 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27492 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27493 selected code model, define __code_mode_small__, __code_model_medium__,
27494 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27495 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27496 xchg temporary register with %k. Declare temporary register as
27498 [__x86_64__]: For medium and large code models, preserve %rbx register.
27500 2013-01-03 Richard Biener <rguenther@suse.de>
27502 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27503 (dump_subscript): Adjust.
27504 (finalize_ddr_dependent): Do not dump redundant info.
27505 (analyze_siv_subscript): Adjust.
27506 (subscript_dependence_tester): Likewise.
27507 (compute_affine_dependence): Likewise.
27509 2013-01-03 Richard Biener <rguenther@suse.de>
27512 2013-01-03 Richard Biener <rguenther@suse.de>
27514 PR tree-optimization/55857
27515 * tree-vect-stmts.c (vectorizable_load): Do not setup
27516 re-alignment for invariant loads.
27518 2013-01-02 Richard Biener <rguenther@suse.de>
27520 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27521 invariant load do not generate a vector load from the scalar location.
27523 2013-01-03 Richard Biener <rguenther@suse.de>
27525 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27526 for not vectorizing.
27527 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27528 not build INDIRECT_REFs, call get_name once only.
27529 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27530 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27532 2013-01-03 Richard Biener <rguenther@suse.de>
27534 PR tree-optimization/55857
27535 * tree-vect-stmts.c (vectorizable_load): Do not setup
27536 re-alignment for invariant loads.
27538 2013-01-03 Richard Biener <rguenther@suse.de>
27541 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27542 prefer a built-in decl.
27544 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27546 * gcc.c (process_command): Update copyright notice dates.
27547 * gcov.c (print_version): Likewise.
27548 * gcov-dump.c (print_version): Likewise.
27550 PR rtl-optimization/55838
27551 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27552 iv0.step, iv1.step and step.
27554 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27555 Marc Glisse <marc.glisse@inria.fr>
27557 PR tree-optimization/55832
27558 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27559 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27560 integer_{one,zero}_node.
27562 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27565 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27566 * var-tracking.c (reverse_op): Don't add reverse ops to
27567 VALUEs that have already
27568 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27570 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27572 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27574 2013-01-02 Teresa Johnson <tejohnson@google.com>
27576 * dumpfile.c (dump_loc): Print filename with location.
27577 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27578 new location_t parameter to emit complete unroll message with
27579 new dump framework.
27580 (canonicalize_loop_induction_variables): Compute loops location
27581 and pass to try_unroll_loop_completely.
27582 * loop-unroll.c (report_unroll_peel): New function.
27583 (peel_loops_completely): Use new dump format with location
27584 for main dumpfile message, and invoke report_unroll_peel on success.
27585 (decide_unrolling_and_peeling): Ditto.
27586 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27587 by report_unroll_peel.
27588 (decide_peel_completely): Ditto.
27589 (decide_unroll_constant_iterations): Ditto.
27590 (decide_unroll_runtime_iterations): Ditto.
27591 (decide_peel_simple): Ditto.
27592 (decide_unroll_stupid): Ditto.
27593 * cfgloop.c (get_loop_location): New function.
27594 * cfgloop.h (get_loop_location): Declare.
27596 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27598 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27601 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27603 PR middle-end/55198
27604 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27605 BLKmode objects when EXPAND_MEMORY is specified.
27607 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27609 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27611 (fold_builtin_cpu): Do not share cpu model decls across statements.
27613 2013-01-02 Jason Merrill <jason@redhat.com>
27616 * tree.c (build_array_type_1): Revert earlier change.
27618 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27620 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27622 * config/aarch64/aarch64-tune.md: Re-generate.
27624 2013-01-02 Richard Biener <rguenther@suse.de>
27626 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27627 invariant load do not generate a vector load from the scalar location.
27629 2013-01-02 Richard Biener <rguenther@suse.de>
27632 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27633 * configure: Regenerate.
27635 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27637 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27638 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27639 (expand_builtin_int_roundingfn_2): Keep the original target around
27640 for the fallback case.
27642 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27644 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27645 to be clear for sign changes.
27647 2013-01-01 Jan Hubicka <jh@suse.cz>
27649 * ipa-inline-analysis.c: Fix formatting.
27651 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27653 PR tree-optimization/55831
27654 * tree-vect-loop.c (get_initial_def_for_induction): Use
27655 gsi_after_labels instead of gsi_start_bb.
27657 Copyright (C) 2013 Free Software Foundation, Inc.
27659 Copying and distribution of this file, with or without modification,
27660 are permitted in any medium without royalty provided the copyright
27661 notice and this notice are preserved.