1 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3 * config/i386/constraints.md (B): New.
6 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
7 * config/i386/i386-modes.def (BND32): New.
9 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
10 * config/i386/i386.c (isa_opts): Add mmpx.
11 (regclass_map): Add bound registers.
12 (dbx_register_map): Likewise.
13 (dbx64_register_map): Likewise.
14 (svr4_dbx_register_map): Likewise.
16 (ix86_option_override_internal): Support MPX ISA.
17 (ix86_conditional_register_usage): Support bound registers.
18 (print_reg): Likewise.
19 (ix86_code_end): Add MPX bnd prefix.
20 (output_set_got): Likewise.
21 (ix86_output_call_insn): Likewise.
22 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
23 (ix86_print_operand_punct_valid_p): Likewise.
24 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
26 (ix86_class_likely_spilled_p): Add bound regs support.
27 (ix86_hard_regno_mode_ok): Likewise.
28 (x86_order_regs_for_local_alloc): Likewise.
29 (ix86_bnd_prefixed_insn_p): New.
30 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
31 (FIXED_REGISTERS): Add bound registers.
32 (CALL_USED_REGISTERS): Likewise.
33 (REG_ALLOC_ORDER): Likewise.
34 (HARD_REGNO_NREGS): Likewise.
36 (VALID_BND_REG_MODE): New.
39 (reg_class): Add BND_REGS.
40 (REG_CLASS_NAMES): Likewise.
41 (REG_CLASS_CONTENTS): Likewise.
45 (HI_REGISTER_NAMES): Add bound registers.
46 * config/i386/i386.md (UNSPEC_BNDMK): New.
47 (UNSPEC_BNDMK_ADDR): New.
50 (UNSPEC_BNDLDX_ADDR): New.
54 (UNSPEC_MPX_FENCE): New.
57 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
58 (length_immediate): Likewise.
59 (prefix_0f): Likewise.
61 (prefix_rep): Check for bnd prefix.
63 (length): Use length_nobnd if specified.
68 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
71 (simple_return_internal): Likewise.
72 (simple_return_pop_internal): Likewise.
73 (*indirect_jump): Add MPX bnd prefix.
74 (*tablejump_1): Likewise.
75 (simple_return_internal_long): Likewise.
76 (simple_return_indirect_internal): Likewise.
80 (*mov<mode>_internal_mpx): New.
81 (<mode>_<bndcheck>): New.
82 (*<mode>_<bndcheck>): New.
87 * config/i386/predicates.md (lea_address_operand): Rename to...
88 (address_no_seg_operand): ... this.
89 (address_mpx_no_base_operand): New.
90 (address_mpx_no_index_operand): New.
91 (bnd_mem_operator): New.
92 * config/i386/i386.opt (mmpx): New.
93 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
94 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
96 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
98 * mode-classes.def (MODE_POINTER_BOUNDS): New.
99 * tree.def (POINTER_BOUNDS_TYPE): New.
100 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
101 (POINTER_BOUNDS_MODE): New.
102 (make_pointer_bounds_mode): New.
103 * machmode.h (POINTER_BOUNDS_MODE_P): New.
104 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
105 (layout_type): Support POINTER_BOUNDS_TYPE.
106 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
107 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
108 (type_contains_placeholder_1): Likewise.
109 * tree.h (POINTER_BOUNDS_TYPE_P): New.
110 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
111 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
113 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
115 * expr.c (expand_expr_real_1): Use mode of memory reference rather
116 than mode of address computation when calling memory_address_addr_space.
118 2013-08-24 Richard Henderson <rth@twiddle.net>
121 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
122 instead of create_convert_operand_to.
123 (maybe_emit_sync_lock_test_and_set): Likewise.
124 (expand_atomic_compare_and_swap): Likewise.
125 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
127 2013-08-24 Sriraman Tallam <tmsriram@google.com>
129 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
130 Change param_is to use the struct and not the pointer to the struct.
132 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
134 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
135 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
137 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
138 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
139 Relocate from gimplify.c.
140 * targhooks.h: Add 2 prototypes.
141 * tree.h: Delete 2 prototypes.
143 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
145 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
146 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
148 2013-10-24 Cong Hou <congh@google.com>
150 * convert.c (convert_to_real): Guard those unsafe math function
151 convertions with flag_unsafe_math_optimizations. Handle sqrt()
154 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
157 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
159 (cgraph_create_edge): Use the new argument.
160 (cgraph_create_indirect_edge): Likewise.
162 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
164 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
166 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
167 changes statep->state.
169 2013-10-24 Tobias Burnus <burnus@net-b.de>
172 * tree-cfg.c (replace_loop_annotate): New function.
173 (execute_build_cfg): Call it.
174 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
175 * internal-fn.c (expand_ANNOTATE): New function.
176 * internal-fn.def (ANNOTATE): Define as new internal function.
177 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
178 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
179 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
180 * doc/extend.texi (Pragmas): Document #pragma ivdep.
181 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
183 2013-10-17 Ian Bolton <ian.bolton@arm.com>
184 Marcus Shawcroft <marcus.shawcroft@arm.com>
186 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
187 Special case reload SP+C into none GENERAL_REGS.
189 2013-10-24 Michael Matz <matz@suse.de>
191 * gengtype.c (is_file_equal): Check that files will be same
194 2013-10-25 Christian Bruel <christian.bruel@st.com>
196 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
197 * gcc/config/sh/t-sh (sh-mem.o): New rule.
198 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
199 (sh4_expand_cmpstr): New function.
200 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c
201 * gcc/config/sh/sh-protos.h (sh4_expand_cmpstr): Declare.
202 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
205 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
207 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
208 * Makefile.in (CXXDEPMODE): Assign and change users.
210 * configure: Regenerate.
212 2013-10-23 David Malcolm <dmalcolm@redhat.com>
214 * gengtype-parse.c (require_without_advance): New.
215 (type): For GTY-marked types that are not GTY((user)), parse any
216 base classes, requiring them to be single-inheritance, and not
217 be templates. For non-GTY-marked types and GTY((user)),
218 continue to skip over any C++ inheritance specification.
219 * gengtype-state.c (state_writer::write_state_struct_type):
220 Write base class of type (if any).
221 (read_state_struct_type): Read base class of type (if any).
222 * gengtype.c (new_structure): Add a "base_class" parameter.
223 (create_optional_field_): Update for new parameter to
225 (adjust_field_rtx_def): Likewise.
226 (adjust_field_tree_exp): Likewise.
227 * gengtype.h (struct type): Add "base_class" field to the s
229 (new_structure): Add "base" parameter.
231 2013-10-23 Sriraman Tallam <tmsriram@google.com>
234 * config/i386/i386.c (ix86_option_override_internal):
235 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
236 (ix86_valid_target_attribute_tree):
237 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
238 Change TARGET_SSE to TARGET_SSE_P (opts->...)
240 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
242 * tree-ssa-loop.h: Remove include files.
243 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
244 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
245 * graphite-clast-to-gimple.c: Likewise.
246 * graphite-scop-detection.c: Likewise.
247 * graphite-sese-to-poly.c: Likewise.
248 * ipa-inline-analysis.c: Likewise.
249 * ipa-pure-const.c: Likewise.
250 * loop-init.c: Likewise.
251 * passes.c: Likewise.
252 * predict.c: Likewise.
253 * tree-cfg.c: Likewise.
254 * tree-cfgcleanup.c: Likewise.
255 * tree-chrec.c: Likewise.
256 * tree-data-ref.c: Likewise.
257 * tree-loop-distribution.c: Likewise.
258 * tree-parloops.c: Likewise.
259 * tree-predcom.c: Likewise.
260 * tree-scalar-evolution.c: Likewise.
261 * tree-ssa-address.c: Likewise.
262 * tree-ssa.c: Likewise.
263 * tree-ssa-dce.c: Likewise.
264 * tree-ssa-loop.c: Likewise.
265 * tree-ssa-loop-im.c: Likewise.
266 * tree-ssa-loop-ivcanon.c: Likewise.
267 * tree-ssa-loop-ivopts.c: Likewise.
268 * tree-ssa-loop-manip.c: Likewise.
269 * tree-ssa-loop-niter.c: Likewise.
270 * tree-ssa-loop-prefetch.c: Likewise.
271 * tree-ssa-loop-unswitch.c: Likewise.
272 * tree-ssa-reassoc.c: Likewise.
273 * tree-vect-data-refs.c: Likewise.
274 * tree-vect-loop.c: Likewise.
275 * tree-vect-loop-manip.c: Likewise.
276 * tree-vectorizer.c: Likewise.
277 * tree-vect-stmts.c: Likewise.
278 * tree-vrp.c: Likewise.
280 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
282 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
284 2013-10-23 Jakub Jelinek <jakub@redhat.com>
286 PR tree-optimization/58775
287 PR tree-optimization/58791
288 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
289 (insert_stmt_after): Rewritten, don't move the stmt, but really
291 (get_stmt_uid_with_default): Remove.
292 (build_and_add_sum): Use insert_stmt_after and
293 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only
295 (update_range_test): Set uid on stmts added by
296 force_gimple_operand_gsi. Don't immediately modify statements
297 in inter-bb optimization, just update oe->op values.
298 (optimize_range_tests): Return bool whether any changed have
300 (update_ops): New function.
301 (struct inter_bb_range_test_entry): New type.
302 (maybe_optimize_range_tests): Perform statement changes here.
303 (not_dominated_by, appears_later_in_bb, get_def_stmt,
304 ensure_ops_are_available): Remove.
305 (find_insert_point): Rewritten.
306 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
307 return LHS of the (new resp. old) stmt. Don't call
308 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
309 instead of last, move new stmt at the right place.
310 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
311 (negate_value): Likewise. Set uids.
312 (break_up_subtract_bb): Initialize uids.
313 (reassociate_bb): Adjust rewrite_expr_tree caller.
314 (do_reassoc): Don't call renumber_gimple_stmt_uids.
316 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
319 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
320 TARGET_32BIT final condition.
321 (mulsi3_internal2 and splitter): Same.
323 2013-10-23 Jeff Law <law@redhat.com>
325 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
326 through joiner blocks with abnormal outgoing edges.
328 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
329 Add parameter JOINERS, to allow/disallow threading through joiner
331 (thread_block): New. Call thread_block_1.
332 (mark_threaded_blocks): Remove code to filter out certain cases
333 of threading through joiner blocks.
334 (thread_through_all_blocks): Document how we can have a dangling
335 edge AUX field and clear it.
337 2013-10-23 Ian Lance Taylor <iant@google.com>
339 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
340 (C++ Dialect Options): Likewise.
341 (Optimize Options): Likewise.
343 2013-10-23 Tom de Vries <tom@codesourcery.com>
345 PR tree-optimization/58805
346 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
348 2013-10-23 Jakub Jelinek <jakub@redhat.com>
350 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
351 sequence based on get_range_info returned range.
353 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
355 * tree-ssa.h: Remove all #include's
356 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
357 * alias.c: Move required includes from tree-ssa.h.
359 * builtins.c: Likewise.
361 * cfgexpand.c: Likewise.
362 * cfghooks.c: Likewise.
363 * cfgloop.c: Likewise.
364 * cfgloopmanip.c: Likewise.
365 * cgraph.c: Likewise.
366 * cgraphbuild.c: Likewise.
367 * cgraphclones.c: Likewise.
368 * cgraphunit.c: Likewise.
370 * except.c: Likewise.
373 * fold-const.c: Likewise.
374 * ggc-page.c: Likewise.
375 * gimple-builder.c: Likewise.
376 * gimple-fold.c: Likewise.
377 * gimple-iterator.c: Likewise.
378 * gimple-low.c: Likewise.
379 * gimple-pretty-print.c: Likewise.
380 * gimple-ssa-strength-reduction.c: Likewise.
381 * gimple-streamer-in.c: Likewise.
382 * gimple-streamer-out.c: Likewise.
383 * gimplify.c: Likewise.
384 * graphite-blocking.c: Likewise.
385 * graphite-clast-to-gimple.c: Likewise.
386 * graphite-dependences.c: Likewise.
387 * graphite-interchange.c: Likewise.
388 * graphite-optimize-isl.c: Likewise.
389 * graphite-poly.c: Likewise.
390 * graphite-scop-detection.c: Likewise.
391 * graphite-sese-to-poly.c: Likewise.
392 * graphite.c: Likewise.
393 * ipa-cp.c: Likewise.
394 * ipa-inline-analysis.c: Likewise.
395 * ipa-inline-transform.c: Likewise.
396 * ipa-inline.c: Likewise.
397 * ipa-prop.c: Likewise.
398 * ipa-pure-const.c: Likewise.
399 * ipa-reference.c: Likewise.
400 * ipa-split.c: Likewise.
401 * ipa-utils.c: Likewise.
402 * loop-init.c: Likewise.
403 * lto-cgraph.c: Likewise.
404 * lto-section-in.c: Likewise.
405 * lto-section-out.c: Likewise.
406 * lto-streamer-in.c: Likewise.
407 * lto-streamer-out.c: Likewise.
408 * lto-streamer.c: Likewise.
409 * omp-low.c: Likewise.
410 * passes.c: Likewise.
411 * predict.c: Likewise.
412 * print-tree.c: Likewise.
413 * profile.c: Likewise.
415 * targhooks.c: Likewise.
416 * tracer.c: Likewise.
417 * trans-mem.c: Likewise.
418 * tree-call-cdce.c: Likewise.
419 * tree-cfg.c: Likewise.
420 * tree-cfgcleanup.c: Likewise.
421 * tree-chrec.c: Likewise.
422 * tree-complex.c: Likewise.
423 * tree-data-ref.c: Likewise.
424 * tree-dfa.c: Likewise.
425 * tree-eh.c: Likewise.
426 * tree-emutls.c: Likewise.
427 * tree-if-conv.c: Likewise.
428 * tree-inline.c: Likewise.
429 * tree-into-ssa.c: Likewise.
430 * tree-loop-distribution.c: Likewise.
431 * tree-mudflap.c: Likewise.
432 * tree-nested.c: Likewise.
433 * tree-nrv.c: Likewise.
434 * tree-object-size.c: Likewise.
435 * tree-outof-ssa.c: Likewise.
436 * tree-parloops.c: Likewise.
437 * tree-phinodes.c: Likewise.
438 * tree-predcom.c: Likewise.
439 * tree-pretty-print.c: Likewise.
440 * tree-profile.c: Likewise.
441 * tree-scalar-evolution.c: Likewise.
442 * tree-sra.c: Likewise.
443 * tree-ssa-address.c: Likewise.
444 * tree-ssa-alias.c: Likewise.
445 * tree-ssa-ccp.c: Likewise.
446 * tree-ssa-coalesce.c: Likewise.
447 * tree-ssa-copy.c: Likewise.
448 * tree-ssa-copyrename.c: Likewise.
449 * tree-ssa-dce.c: Likewise.
450 * tree-ssa-dom.c: Likewise.
451 * tree-ssa-dse.c: Likewise.
452 * tree-ssa-forwprop.c: Likewise.
453 * tree-ssa-ifcombine.c: Likewise.
454 * tree-ssa-live.c: Likewise.
455 * tree-ssa-loop-ch.c: Likewise.
456 * tree-ssa-loop-im.c: Likewise.
457 * tree-ssa-loop-ivcanon.c: Likewise.
458 * tree-ssa-loop-ivopts.c: Likewise.
459 * tree-ssa-loop-manip.c: Likewise.
460 * tree-ssa-loop-niter.c: Likewise.
461 * tree-ssa-loop-prefetch.c: Likewise.
462 * tree-ssa-loop-unswitch.c: Likewise.
463 * tree-ssa-loop.c: Likewise.
464 * tree-ssa-math-opts.c: Likewise.
465 * tree-ssa-operands.c: Likewise.
466 * tree-ssa-phiopt.c: Likewise.
467 * tree-ssa-phiprop.c: Likewise.
468 * tree-ssa-pre.c: Likewise.
469 * tree-ssa-propagate.c: Likewise.
470 * tree-ssa-reassoc.c: Likewise.
471 * tree-ssa-sccvn.c: Likewise.
472 * tree-ssa-sink.c: Likewise.
473 * tree-ssa-strlen.c: Likewise.
474 * tree-ssa-structalias.c: Likewise.
475 * tree-ssa-tail-merge.c: Likewise.
476 * tree-ssa-ter.c: Likewise.
477 * tree-ssa-threadedge.c: Likewise.
478 * tree-ssa-threadupdate.c: Likewise.
479 * tree-ssa-uncprop.c: Likewise.
480 * tree-ssa-uninit.c: Likewise.
481 * tree-ssa.c: Likewise.
482 * tree-ssanames.c: Likewise.
483 * tree-stdarg.c: Likewise.
484 * tree-streamer-in.c: Likewise.
485 * tree-switch-conversion.c: Likewise.
486 * tree-tailcall.c: Likewise.
487 * tree-vect-data-refs.c: Likewise.
488 * tree-vect-generic.c: Likewise.
489 * tree-vect-loop-manip.c: Likewise.
490 * tree-vect-loop.c: Likewise.
491 * tree-vect-patterns.c: Likewise.
492 * tree-vect-slp.c: Likewise.
493 * tree-vect-stmts.c: Likewise.
494 * tree-vectorizer.c: Likewise.
495 * tree-vrp.c: Likewise.
498 * value-prof.c: Likewise.
499 * var-tracking.c: Likewise.
500 * varpool.c: Likewise.
501 * vtable-verify.c: Likewise.
503 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
505 * config/tilegx/tilegx.c: Include "tree.h".
507 2013-10-23 Jakub Jelinek <jakub@redhat.com>
509 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
510 the info, not after it.
511 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
512 dump_ssaname_info call.
513 (pp_gimple_stmt_1): Adjust caller.
514 (dump_phi_nodes): Likewise. Don't print "# " here.
516 2013-10-22 Jan Hubicka <jh@suse.cz>
518 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New tuning flag.
519 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
520 * i386.c (expand_small_movmem_or_setmem): New function.
521 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New function
522 (alg_usable_p): Add support for value ranges; cleanup.
523 (ix86_expand_set_or_movmem): Add support for misaligned moves.
525 2013-10-22 Sterling Augustine <saugustine@google.com>
527 * doc/invoke.texi: Document -ggnu-pubnames.
528 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
530 * dwarf2out.c: Include gdb/gdb-index.h.
531 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
532 debug_generate_pub_sections.
533 (is_java, output_pubtables, output_pubname): New functions.
534 (include_pubname_in_output): Handle debug_generate_pub_sections at
536 (size_of_pubnames): Use new local space_for_flags based on
537 debug_generate_pub_sections.
538 (output_pubnames): Unify pubnames and pubtypes output logic.
539 Genericize comments. Call output_pubname.
540 (dwarf2out_finish): Move logic to output_pubnames and call it.
542 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
545 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
546 Remove CCCmode handling.
547 <case LTU>: Return 'c' suffix for CCCmode.
548 <case GEU>: Return 'nc' suffix for CCCmode.
549 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
550 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
551 (*sub<mode>3_cc_overflow): Ditto.
552 (*subsi3_zext_cc_overflow): Ditto.
554 2013-10-22 Steve Ellcey <sellcey@mips.com>
556 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
557 (AND (NOT OP1) (NOT OP2)).
559 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
561 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
562 meaning of merge-high and merge-low masks for little endian; avoid
563 use of vector-pack masks for little endian for mismatched modes.
565 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
567 * config/tilepro/tilepro.c: Include "tree.h".
569 2013-10-22 Andreas Schwab <schwab@suse.de>
571 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
574 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
576 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
577 [Ed Smith-Rowland]): New entries.
578 ([Stephen M. Webb]): Update.
580 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
582 * tree-ssa-ter.h: Remove duplicate copy of file contents.
584 2013-10-21 Marek Polacek <polacek@redhat.com>
587 * fold-const.c (fold_range_test): Return 0 if the type is not
590 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
592 * system.h: Move hwint.h include further down.
593 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
594 gcc_checking_asserts.
595 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
597 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
599 Fix volatile issues in optimize_bit_field_compare.
600 * fold-const.c (optimize_bit_field_compare): Bail out if
601 lvolatilep or rvolatilep.
603 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
605 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
606 and get_inner_reference returning different pmode for non-volatile
607 bit-field members dependent on flag_strict_volatile_bitfields.
608 * stor-layout.c (layout_decl): Remove special handling of
609 flag_strict_volatile_bitfields.
610 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
611 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
613 2013-10-21 Paulo Matos <pmatos@broadcom.com>
615 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
618 2013-10-21 Jeff Law <law@redhat.com>
620 * tree-ssa-threadedge.c (thread_through_normal_block): New
621 argument VISITED. Remove VISISTED as a local variable. When we
622 have a threadable jump, verify the destination of the jump has not
624 (thread_across_edge): Allocate VISITED bitmap once at function
625 scope and use it throughout. Make sure to set appropriate bits in
626 VISITED for E (start of jump thread path).
627 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
628 through a joiner if any edge on the path has a recorded jump thread.
630 2013-10-21 Ian Lance Taylor <iant@google.com>
632 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
633 don't imply that attributes can solve all problems.
634 (Directory Options): Fix typo.
636 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
638 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
639 extend and extend_add.
641 2013-10-21 Richard Biener <rguenther@suse.de>
643 PR tree-optimization/58794
644 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
645 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
647 2013-10-21 Richard Biener <rguenther@suse.de>
650 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
651 to (T) X for sign-changing conversions (or no conversion).
653 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
655 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
657 2013-10-20 Jan Hubicka <jh@suse.cz>
659 * config/i386/i386-tune.def: Add comment; organize into categories
661 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
663 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
664 argument. Update function comment.
665 (expand_set_or_movmem_via_rep): New function combining
666 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
667 (expand_movmem_via_rep_mov): Remove.
668 expand_setmem_via_rep_stos): Remove.
669 (expand_movmem_epilogue): Update calls correspondingly.
670 (expand_setmem_epilogue_via_loop): Likewise.
672 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
673 (expand_set_or_movmem_prologue): New function combining
674 expand_movmem_prologue and expand_setmem_prologue.
675 (expand_movmem_prologue): Remove.
676 (expand_setmem_prologue): Remove.
677 (expand_set_or_movmem_constant_prologue): New function combining
678 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
679 (expand_constant_movmem_prologue): Remove.
680 (expand_constant_setmem_prologue): Remove.
681 (promote_duplicated_reg): Allow vector-const0 value.
682 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
683 and ix86_expand_setmem.
684 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
685 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
687 2013-10-21 Diego Novillo <dnovillo@google.com>
689 * asan.c: Include tree.h
690 * bb-reorder.c: Likewise.
691 * cfgcleanup.c: Likewise.
692 * cfgloopmanip.c: Likewise.
693 * data-streamer-in.c: Likewise.
694 * data-streamer-out.c: Likewise.
695 * data-streamer.c: Likewise.
696 * dwarf2cfi.c: Likewise.
697 * graphite-blocking.c: Likewise.
698 * graphite-clast-to-gimple.c: Likewise.
699 * graphite-dependences.c: Likewise.
700 * graphite-interchange.c: Likewise.
701 * graphite-optimize-isl.c: Likewise.
702 * graphite-poly.c: Likewise.
703 * graphite-scop-detection.c: Likewise.
704 * graphite-sese-to-poly.c: Likewise.
705 * graphite.c: Likewise.
706 * ipa-devirt.c: Likewise.
707 * ipa-profile.c: Likewise.
710 * loop-init.c: Likewise.
711 * loop-unroll.c: Likewise.
712 * lower-subreg.c: Likewise.
713 * lto/lto-object.c: Likewise.
715 * reginfo.c: Likewise.
716 * tree-loop-distribution.c: Likewise.
717 * tree-parloops.c: Likewise.
718 * tree-ssa-strlen.c: Likewise.
719 * tree-streamer.c: Likewise.
720 * value-prof.c: Likewise.
721 * target-globals.c: Likewise.
722 * expr.h: Include tree-core.h instead of tree.h.
723 * gimple.h: Likewise.
724 * ipa-prop.h: Likewise.
725 * ipa-utils.h: Likewise.
726 * lto-streamer.h: Likewise.
727 * streamer-hooks.h: Likewise.
728 * ipa-reference.h: Include cgraph.h instead of tree.h.
729 * cgraph.h: Include basic-block.h instead of tree.h.
730 * tree-streamer.h: Do not include tree.h.
731 * genattrtab.c (write_header): Generate inclusion of
733 * genautomata.c (main): Likewise.
734 * genemit.c: Likewise.
735 * genopinit.c: Likewise.
736 * genoutput.c (output_prologue): Likewise.
737 * genpeep.c: Likewise.
739 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
741 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
743 (vec_unpacku_hi_v8hi): Likewise.
744 (vec_unpacku_lo_v16qi): Likewise.
745 (vec_unpacku_lo_v8hi): Likewise.
747 2013-10-20 Jan Hubicka <jh@suse.cz>
749 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
750 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
751 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
753 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
755 * config/mips/mips.h (ISA_HAS_WSBH): Define.
756 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
758 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
760 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
763 * system.h: Undef m_slot.
765 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
767 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
768 all elements for both endian flavors.
770 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
773 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
774 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
775 and SI_REG for 64bit SYSV ABI targets.
777 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
779 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
780 to multi_reg_return. Clarify that we are skipping USEs of multiple
781 return registers. Use bool type where appropriate.
783 2013-10-18 Jan Hubicka <jh@suse.cz>
785 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
787 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
788 (X86_TUNE_PUSH_MEMORY): Likewise.
789 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
790 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
791 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
793 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
794 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
796 (ix86_option_override_internal): Update to use tune features instead
799 2013-10-18 Cong Hou <congh@google.com>
801 PR tree-optimization/58508
802 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
803 statement that contains data refs with zero-step.
805 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
807 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
808 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
809 * gimple-low.c (gimple_check_call_arg,
810 gimple_check_call_matching_types): Move to cgraph.c.
811 * gimple-low.h: Remove prototype.
812 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
813 Relocate from gimple-low.c.
814 * cgraph.h: Add prototype. Don't include basic-block.h.
815 * gimplify.c: Add gimple-low to include list.
816 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
817 * tree-eh.c: Add gimple-low to include list.
818 * tree-nested.c: Likewise.
819 * cfgexpand.c: Add tree-ssa-address.h to include list.
821 * gimple-fold.c: Likewise.
822 * gimple-ssa-strength-reduction.c: Likewise.
823 * trans-mem.c: Likewise.
824 * tree-mudflap.c: Likewise.
825 * tree-ssa-loop-ivopts.c: Likewise.
826 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
827 (degenerate_phi_result): Move to tree-phinodes.c.
828 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
829 * tree-ssa-threadedge.c: Likewise.
830 * tree-vrp.c: Likewise.
831 * tree-phinodes.c (degenerate_phi_result): Relocate here.
832 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
833 * tree-phinodes.h (degenerate_phi_result): Add prototype.
834 * tree-ssa-copy.c: Include tree-ssa-dom.h.
835 * tree-ssa-forwprop.c: Likewise.
836 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
837 pass_data_cleanup_cfg_post_optimizing,
838 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
839 * tree-optimize.c: Delete File.
840 * graphite.c: Include tree-cfgcleanup.h.
841 * passes.c: Likewise.
842 * tree-cfg.c: Likewise.
843 * tree-profile.c: Likewise.
844 * tree-ssa-dse.c: Likewise.
845 * tree-ssa-loop-ivcanon.c: Likewise.
846 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
847 * tree-outof-ssa.c: Include sbitmap.h.
848 * tree-ssa-live.c: Likewise.
849 * tree-ssa-propagate.c: Likewise.
850 * tree-ssa-structalias.c: Likewise.
851 * tree-stdarg.c: Likewise.
852 * Makefile.in (OBJS): Delete tree-optimize.o.
853 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
854 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
855 * varasm.c: Include basic-block.h.
856 * cfgloop.h: Include function.h instead of basic-block.h
857 (bb_loop_depth): Move to cfgloop.c.
858 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
860 2013-10-18 Teresa Johnson <tejohnson@google.com>
862 * predict.c (probably_never_executed): Compare frequency-based
863 count to number of training runs.
864 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
866 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
868 * config/arm/arm.c (cortexa9_extra_costs): New table.
869 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
871 2013-10-18 Jeff Law <law@redhat.com>
873 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
875 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
876 (register_jump_thread): Add "registered_jump_thread" debug
878 * dbgcnt.def (registered_jump_thread): New debug counter.
880 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
882 * config/rs6000/rs6000.c: Include cgraph.h.
884 2013-10-18 Teresa Johnson <tejohnson@google.com>
886 * tree-ssa-tail-merge.c (replace_block_by): Update edge
887 weights during merging.
889 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
891 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
892 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
893 * tree-ssa-operands.h: Remove prototype.
894 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
895 * gimple.c (virtual_operand_p): Relocate from gimple.c.
896 * gimple.h: Add prototype.
897 * gimple-ssa.h: Include tree-ssa-operands.h.
898 * tree-dump.c: Add tree-cfg.h to include list.
899 * tree-ssa-alias.c: Add ipa-reference.h to include list.
900 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
901 * config/i386/i386.c: Don't include tree-flow.h.
902 * config/rs6000/rs6000.c: Likewise.
904 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
906 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
908 2013-10-18 Richard Biener <rguenther@suse.de>
910 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
911 or TYPE_UNSIGNED of integral types.
912 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
913 NULL_TREE for zero-precision integral types.
915 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
917 * config/aarch64/arm_neon.h
918 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
920 2013-10-17 Sriraman Tallam <tmsriram@google.com>
923 * opth-gen.awk: Define target_flags_explicit.
925 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
927 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
928 fields to the reg_addr array that describes the valid addressing
929 mode for any register, general purpose registers, floating point
930 registers, and Altivec registers.
931 (FIRST_RELOAD_REG_CLASS): Likewise.
932 (LAST_RELOAD_REG_CLASS): Likewise.
933 (struct reload_reg_map_type): Likewise.
934 (reload_reg_map_type): Likewise.
935 (RELOAD_REG_VALID): Likewise.
936 (RELOAD_REG_MULTIPLE): Likewise.
937 (RELOAD_REG_INDEXED): Likewise.
938 (RELOAD_REG_OFFSET): Likewise.
939 (RELOAD_REG_PRE_INCDEC): Likewise.
940 (RELOAD_REG_PRE_MODIFY): Likewise.
941 (reg_addr): Likewise.
942 (mode_supports_pre_incdec_p): New helper functions to say whether
943 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
944 (mode_supports_pre_modify_p): Likewise.
945 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
946 print the valid address mode bits for each mode.
947 (rs6000_debug_print_mode): Likewise.
948 (rs6000_debug_reg_global): Likewise.
949 (rs6000_setup_reg_addr_masks): New function to set up the address
950 mask bits for each type.
951 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
952 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
953 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
954 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
955 PRE_MODIFY are supported.
956 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
957 registers, instead of {src,dest}_av_p.
958 (rs6000_print_options_internal): Tweak the debug output slightly.
960 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
962 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
965 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
967 * tree-flow.h (struct omp_region): Move to omp-low.c.
968 Remove omp_ prototypes and variables.
969 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
970 (copy_var_decl): Relocate prototype from tree-flow.h.
971 * gimple.c (copy_var_decl): Relocate from omp-low.c.
972 * tree.h: Move prototype to omp-low.h.
973 * omp-low.h: New File. Relocate prototypes here.
974 * omp-low.c (struct omp_region): Make local here.
975 (root_omp_region): Make static.
976 (copy_var_decl) Move to gimple.c.
977 (new_omp_region): Make static.
978 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
979 * tree-cfg.c: Include omp-low.h.
980 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
981 * gimplify.c: Include omp-low.h.
982 * tree-parloops.c: Likewise.
984 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
986 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
987 operands are in memory, prefer to force non-matched operand 1 to
990 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
993 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
994 restrict TImode addresses to single indirect registers if both
995 -mquad-memory and -mvsx-timode are used.
996 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
997 we should emit load/store quad. Remove using %y for quad memory
1000 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
1001 constraints to allow load/store quad on machines where TImode is
1002 not allowed in VSX registers. Use 'n' instead of 'F' constraint
1003 for TImode to load integer constants.
1005 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1007 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
1009 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
1011 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
1012 handling of STACK_REG.
1014 2013-10-17 Richard Biener <rguenther@suse.de>
1016 PR tree-optimization/58143
1017 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
1019 (rewrite_to_defined_overflow): Likewise.
1020 (move_computations_dom_walker::before_dom): Rewrite stmts
1021 with undefined signed overflow that are not always executed
1022 into unsigned arithmetic.
1024 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
1027 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
1028 explicit isa flag to be an options variable, instead of using
1029 global_options_set. Remove define from rs6000.h.
1030 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
1032 * config/rs6000/rs6000.c (rs6000_option_override_internal):
1033 Initialize rs6000_isa_flags_explicit.
1034 (rs6000_function_specific_save): Add gcc_options* parameter, so
1035 that the powerpc builds after the 2013-10-15 changes.
1036 (rs6000_function_specific_restore): Likewise.
1038 2013-10-16 DJ Delorie <dj@redhat.com>
1040 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
1041 op is a REG before checking REGNO.
1042 (rl78_alloc_physical_registers): Verify pattern is a SET before
1045 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1047 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
1049 (vec_unpacks_lo_v4sf): Likewise.
1050 (vec_unpacks_float_hi_v4si): Likewise.
1051 (vec_unpacks_float_lo_v4si): Likewise.
1052 (vec_unpacku_float_hi_v4si): Likewise.
1053 (vec_unpacku_float_lo_v4si): Likewise.
1055 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1057 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
1058 (vsx_concat_v2sf): Likewise.
1060 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
1062 * config/aarch64/aarch64.md
1063 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
1065 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
1067 PR tree-optimization/58697
1068 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
1069 estimated_loop_iterations_int.
1070 (max_stmt_executions_int): Call get_max_loop_iterations_int.
1071 (get_max_loop_iterations_int): New. HWINT version of
1072 get_max_loop_iterations.
1073 * cfgloop.h: Add prototypes.
1074 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
1075 * loop-unroll.c (decide_peel_once_rolling): Call
1076 get_estimated_loop_iterations_int.
1077 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
1078 * tree-ssa-loop-niter.h: Tweak prototypes.
1080 2013-10-16 David Malcolm <dmalcolm@redhat.com>
1082 * gengtype-parse.c (struct_field_seq): Ignore access-control
1083 keywords ("public:" etc).
1085 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
1087 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
1088 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
1090 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
1092 * config/arm/arm.opt (mlra): New option.
1093 * config/arm/arm.c (arm_lra_p): New function.
1094 (TARGET_LRA_P): Define.
1096 2013-10-16 Paulo Matos <pmatos@broadcom.com>
1098 * tree-core.h (tree_code_name): Remove.
1099 * tree.h (get_tree_code_name): New prototype.
1100 * tree.c (tree_code_name): Make static.
1101 (get_tree_code_name): New function.
1102 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
1103 tree_class_check_failed, tree_range_check_failed,
1104 tree_not_class_check_failed, omp_clause_check_failed,
1105 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
1106 wrapper get_tree_code_name instead of calling tree_code_name directly.
1107 * tree-vrp.c (dump_asserts_for): Likewise.
1108 * tree-dump.c (dequeue_and_dump): Likewise.
1109 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
1110 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
1111 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
1112 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
1113 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
1114 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
1115 dump_gimple_omp_for): Likewise.
1116 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
1117 * tree-ssa-pre.c (print_pre_expr): Likewise.
1118 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
1119 * print-tree.c (print_node_brief, print_node): Likewise.
1120 * gimple.c (gimple_check_failed): Likewise.
1121 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
1122 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
1123 * config/mep/mep.c (mep_validate_vliw): Likewise.
1124 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
1125 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
1127 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
1129 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
1132 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
1134 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
1135 (MULTILIB_MATCHES): Add multilib for -march=v8.
1137 2013-10-15 Sriraman Tallam <tmsriram@google.com>
1140 * optc-save-gen.awk: Add extra parameter to the save and restore
1142 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
1143 * tree.c (build_optimization_node): New parameter. Add extra parameter
1144 to call to cl_optimization_save.
1145 (build_target_option_node): New parameter. Add extra parameter
1146 to call to cl_target_option_save.
1147 * tree.h (build_optimization_node): New parameter.
1148 (build_target_option_node): New parameter.
1149 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
1150 build_optimization_node and build_target_option_node.
1151 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
1152 (handle_pragma_push_options): Ditto.
1153 * toplev.c (process_options): Ditto.
1154 * opts.c (init_options_struct): Check for opts_set non-null.
1155 * target.def (target_option.save): New parameter.
1156 (target_option.restore): New parameter.
1157 * tm.texi: Generate.
1158 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
1159 (ix86_pragma_target_parse): Ditto.
1160 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
1162 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
1163 to build_optimization_node and build_target_option_node.
1164 (rs6000_valid_attribute_p): Ditto.
1165 (rs6000_pragma_target_parse): Ditto.
1166 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
1168 * config/i386/i386.h:
1169 TARGET_64BIT_P: New Macro
1170 TARGET_MMX_P: New Macro.
1171 TARGET_3DNOW_P: New Macro.
1172 TARGET_3DNOW_A_P: New Macro.
1173 TARGET_SSE_P: New Macro.
1174 TARGET_SSE2_P: New Macro.
1175 TARGET_SSE3_P: New Macro.
1176 TARGET_SSSE3_P: New Macro.
1177 TARGET_SSE4_1_P: New Macro.
1178 TARGET_SSE4_2_P: New Macro.
1179 TARGET_AVX_P: New Macro.
1180 TARGET_AVX2_P: New Macro.
1181 TARGET_AVX512F_P: New Macro.
1182 TARGET_AVX512PF_P: New Macro.
1183 TARGET_AVX512ER_P: New Macro.
1184 TARGET_AVX512CD_P: New Macro.
1185 TARGET_FMA_P: New Macro.
1186 TARGET_SSE4A_P: New Macro.
1187 TARGET_FMA4_P: New Macro.
1188 TARGET_XOP_P: New Macro.
1189 TARGET_LWP_P: New Macro.
1190 TARGET_ABM_P: New Macro.
1191 TARGET_BMI_P: New Macro.
1192 TARGET_BMI2_P: New Macro.
1193 TARGET_LZCNT_P: New Macro.
1194 TARGET_TBM_P: New Macro.
1195 TARGET_POPCNT_P: New Macro.
1196 TARGET_SAHF_P: New Macro.
1197 TARGET_MOVBE_P: New Macro.
1198 TARGET_CRC32_P: New Macro.
1199 TARGET_AES_P: New Macro.
1200 TARGET_PCLMUL_P: New Macro.
1201 TARGET_CMPXCHG16B_P: New Macro.
1202 TARGET_FSGSBASE_P: New Macro.
1203 TARGET_RDRND_P: New Macro.
1204 TARGET_F16C_P: New Macro.
1205 TARGET_RTM_P: New Macro.
1206 TARGET_HLE_P: New Macro.
1207 TARGET_RDSEED_P: New Macro.
1208 TARGET_PRFCHW_P: New Macro.
1209 TARGET_ADX_P: New Macro.
1210 TARGET_FXSR_P: New Macro.
1211 TARGET_XSAVE_P: New Macro.
1212 TARGET_XSAVEOPT_P: New Macro.
1213 TARGET_LP64_P: New Macro.
1214 TARGET_X32_P: New Macro.
1215 TARGET_FPMATH_DEFAULT_P: New Macro.
1216 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
1217 * config/i386/i386.c (ix86_option_override_internal): New parameters.
1219 Change ix86_tune_string to access opts->x_ix86_tune_string.
1220 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
1221 Change ix86_arch_string to access opts->x_ix86_arch_string.
1222 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
1223 Change ix86_pmode to access opts->x_ix86_pmode.
1224 Change ix86_abi to access opts->x_ix86_abi.
1225 Change ix86_cmodel to access opts->x_ix86_cmodel.
1226 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
1227 Change ix86_isa_flags_explicit to access
1228 opts->x_ix86_isa_flags_explicit.
1229 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
1230 Change ix86_regparm to access opts->x_ix86_regparm.
1231 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
1232 Change ix86_preferred_stack_boundary_arg to access
1233 opts->x_ix86_preferred_stack_boundary_arg.
1234 Change ix86_force_align_arg_pointer to access
1235 opts->x_ix86_force_align_arg_pointer.
1236 Change ix86_incoming_stack_boundar_arg to access
1237 opts->x_ix86_incoming_stack_boundar_arg.
1238 Change ix86_fpmath to access opts->x_ix86_fpmath.
1239 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
1240 Change ix86_recip_name to access opts->x_ix86_recip_name.
1241 Change ix86_stack_protector_guard to access
1242 opts->x_ix86_stack_protector_guard.
1243 Change ix86_tune_memcpy_strategy to access
1244 opts->x_ix86_tune_memcpy_strategy.
1245 Change ix86_tune_memset_strategy to access
1246 opts->x_ix86_tune_memset_strategy.
1247 Change global_options to access opts.
1248 Change global_options_set to access opts_set.
1249 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
1250 Change TARGET_MMX to TARGET_MMX_P (opts->...).
1251 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
1252 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
1253 Change TARGET_SSE to TARGET_SSE_P (opts->...).
1254 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
1255 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
1256 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
1257 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
1258 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
1259 Change TARGET_AVX to TARGET_AVX_P (opts->...).
1260 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
1261 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
1262 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
1263 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
1264 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
1265 Change TARGET_FMA to TARGET_FMA_P (opts->...).
1266 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
1267 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
1268 Change TARGET_XOP to TARGET_XOP_P (opts->...).
1269 Change TARGET_LWP to TARGET_LWP_P (opts->...).
1270 Change TARGET_ABM to TARGET_ABM_P (opts->...).
1271 Change TARGET_BMI to TARGET_BMI_P (opts->...).
1272 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
1273 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
1274 Change TARGET_TBM to TARGET_TBM_P (opts->...).
1275 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
1276 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
1277 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
1278 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
1279 Change TARGET_AES to TARGET_AES_P (opts->...).
1280 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
1281 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
1282 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
1283 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
1284 Change TARGET_F16C to TARGET_F16C_P (opts->...).
1285 Change TARGET_RTM to TARGET_RTM_P (opts->...).
1286 Change TARGET_HLE to TARGET_HLE_P (opts->...).
1287 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
1288 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
1289 Change TARGET_ADX to TARGET_ADX_P (opts->...).
1290 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
1291 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
1292 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
1293 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
1294 Change TARGET_X32 to TARGET_X32_P (opts->...).
1295 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
1296 Change TARGET_FLOAT_RETURNS_IN_80387 to
1297 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
1298 (ix86_function_specific_save): New parameter. Use opts-> fields
1299 to replace global fields.
1300 (ix86_function_specific_restore): Ditto.
1301 (ix86_valid_target_attribute_inner_p): New parameters.
1303 Fix call to ix86_handle_option and set_option.
1304 (ix86_valid_target_attribute_tree): New parameters.
1305 Change global_options to access opts.
1306 Change global_options_set to access opts_set.
1307 Fix call to ix86_valid_target_attribute_inner_p.
1308 Change ix86_tune_string to access opts->x_ix86_tune_string.
1309 Change ix86_arch_string to access opts->x_ix86_arch_string.
1310 Change ix86_fpmath to access opts->x_ix86_fpmath
1311 Fix call to ix86_option_override_internal.
1312 Fix call to ix86_add_new_builtins.
1313 Fix calls to build_optimization_node and build_target_option_node.
1314 (ix86_valid_target_attribute_p): Remove access to global_options.
1315 Use new gcc_options structure func_options.
1316 Fix call to ix86_valid_target_attribute_tree.
1317 Fix call to build_optimization_node.
1318 (get_builtin_code_for_version): Fix call to
1319 ix86_valid_target_attribute_tree.
1321 2013-10-15 David Malcolm <dmalcolm@redhat.com>
1323 * Makefile.in (PICFLAG): New.
1324 (enable_host_shared): New.
1325 (INTERNAL_CFLAGS): Use PICFLAG.
1326 (LIBIBERTY): Use pic build of libiberty.a if configured with
1327 --enable-host-shared.
1328 * configure.ac: Add --enable-host-shared, setting up new
1330 * configure: Regenerate.
1331 * doc/install.texi (--enable-shared): Add note contrasting it with ...
1332 (--enable-host-shared): New option.
1334 2013-10-15 Richard Biener <rguenther@suse.de>
1336 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
1337 for built-in functions.
1339 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
1341 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
1342 (optimize_range_tests_1): New function,
1343 extracted from optimize_range_tests.
1344 (optimize_range_tests_xor): Similarly.
1345 (optimize_range_tests_diff): New function.
1346 (optimize_range_tests): Use optimize_range_tests_1.
1348 2013-10-15 Cong Hou <congh@google.com>
1350 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
1351 requirement of the reduction pattern so that one operand of the
1352 reduction operation can come from outside of the loop.
1354 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1356 * config/arm/neon-schedgen.ml: Remove.
1357 * config/arm/cortex-a9-neon.md: Remove comment regarding
1360 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1362 * config/arm/types: Remove old neon types.
1364 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1366 * config/arm/cortex-a7.md
1367 (cortex_a7_neon_type): New.
1368 (cortex_a7_neon_mul): Update for new types.
1369 (cortex_a7_neon_mla): Likewise.
1370 (cortex_a7_neon): Likewise.
1372 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1374 * config/arm/cortex-a15-neon.md
1375 (cortex_a15_neon_type): New,
1377 (cortex_a15_neon_int_1): Remove.
1378 (cortex_a15_neon_int_2): Likewise.
1379 (cortex_a15_neon_int_3): Likewise.
1380 (cortex_a15_neon_int_4): Likewise.
1381 (cortex_a15_neon_int_5): Likewise.
1382 (cortex_a15_neon_vqneg_vqabs): Likewise.
1383 (cortex_a15_neon_vmov): Likewise.
1384 (cortex_a15_neon_vaba): Likewise.
1385 (cortex_a15_neon_vaba_qqq): Likewise.
1386 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1387 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1388 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1390 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1391 (cortex_a15_neon_mla_qqq_8_16): Likewise.
1392 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
1393 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1394 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1395 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
1396 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1397 (cortex_a15_neon_shift_1): Likewise.
1398 (cortex_a15_neon_shift_2): Likewise.
1399 (cortex_a15_neon_shift_3): Likewise.
1400 (cortex_a15_neon_vshl_ddd): Likewise.
1401 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1402 (cortex_a15_neon_vsra_vrsra): Likewise.
1403 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1404 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1405 (cortex_a15_neon_bp_3cycle): Likewise.
1406 (cortex_a15_neon_ldm_2): Likewise.
1407 (cortex_a15_neon_stm_2): Likewise.
1408 (cortex_a15_neon_mcr): Likewise.
1409 (cortex_a15_neon_mrc): Likewise.
1410 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
1411 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
1412 (cortex_a15_neon_fp_vmul_ddd): Likewise.
1413 (cortex_a15_neon_fp_vmul_qqd): Likewise.
1414 (cortex_a15_neon_fp_vmla_ddd): Likewise.
1415 (cortex_a15_neon_fp_vmla_qqq): Likewise.
1416 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
1417 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
1418 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1419 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1420 (cortex_a15_neon_bp_simple): Likewise.
1421 (cortex_a15_neon_bp_2cycle): Likewise.
1422 (cortex_a15_neon_bp_3cycle): Likewise.
1423 (cortex_a15_neon_vld1_1_2_regs): Likewise.
1424 (cortex_a15_neon_vld1_3_4_regs): Likewise.
1425 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1426 (cortex_a15_neon_vld2_4_regs): Likewise.
1427 (cortex_a15_neon_vld3_vld4): Likewise.
1428 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1429 (cortex_a15_neon_vst1_3_4_regs): Likewise.
1430 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
1431 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
1432 (cortex_a15_neon_vst3_vst4): Rename to...
1433 (cortex_a15_neon_vst4): This, update for new attributes.
1434 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
1435 (cortex_a15_neon_vld3_vld4_lane): Likewise.
1436 (cortex_a15_neon_vst1_vst2_lane): Likewise.
1437 (cortex_a15_neon_vst3_vst4_lane): Likewise.
1438 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
1439 (cortex_a15_neon_ldm_2): Likewise.
1440 (cortex_a15_neon_stm_2): Likewise.
1441 (cortex_a15_neon_mcr): Likewise.
1442 (cortex_a15_neon_mcr_2_mcrr): Likewise.
1443 (cortex_a15_neon_mrc): Likewise.
1444 (cortex_a15_neon_mrrc): Likewise.
1446 (cortex_a15_neon_abd): New.
1447 (cortex_a15_neon_abd_q): Likewise.
1448 (cortex_a15_neon_aba): Likewise.
1449 (cortex_a15_neon_aba_q): Likewise.
1450 (cortex_a15_neon_acc): Likewise.
1451 (cortex_a15_neon_acc_q): Likewise.
1452 (cortex_a15_neon_arith_basic): Likewise.
1453 (cortex_a15_neon_arith_complex): Likewise.
1454 (cortex_a15_neon_multiply): Likewise.
1455 (cortex_a15_neon_multiply_q): Likewise.
1456 (cortex_a15_neon_mla): Likewise.
1457 (cortex_a15_neon_mla_q): Likewise.
1458 (cortex_a15_neon_sat_mla_long): Likewise.
1459 (cortex_a15_neon_shift_acc): Likewise.
1460 (cortex_a15_neon_shift_imm_basic): Likewise.
1461 (cortex_a15_neon_shift_imm_complex): Likewise.
1462 (cortex_a15_neon_shift_reg_basic): Likewise.
1463 (cortex_a15_neon_shift_reg_basic_q): Likewise.
1464 (cortex_a15_neon_shift_reg_complex): Likewise.
1465 (cortex_a15_neon_shift_reg_complex_q): Likewise.
1466 (cortex_a15_neon_fp_negabs): Likewise
1467 (cortex_a15_neon_fp_arith): Likewise
1468 (cortex_a15_neon_fp_arith_q): Likewise
1469 (cortex_a15_neon_fp_cvt_int): Likewise
1470 (cortex_a15_neon_fp_cvt_int_q): Likewise
1471 (cortex_a15_neon_fp_cvt_16): Likewise
1472 (cortex_a15_neon_fp_mul): Likewise
1473 (cortex_a15_neon_fp_mul_q): Likewise
1474 (cortex_a15_neon_fp_mla): Likewise
1475 (cortex_a15_neon_fp_mla_q): Likewise
1476 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
1477 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
1478 (cortex_a15_neon_bitops): Likewise.
1479 (cortex_a15_neon_bitops_q): Likewise.
1480 (cortex_a15_neon_from_gp): Likewise.
1481 (cortex_a15_neon_from_gp_q): Likewise.
1482 (cortex_a15_neon_tbl3_tbl4): Likewise.
1483 (cortex_a15_neon_zip_q): Likewise.
1484 (cortex_a15_neon_to_gp): Likewise.
1485 (cortex_a15_neon_load_a): Likewise.
1486 (cortex_a15_neon_load_b): Likewise.
1487 (cortex_a15_neon_load_c): Likewise.
1488 (cortex_a15_neon_load_d): Likewise.
1489 (cortex_a15_neon_load_e): Likewise.
1490 (cortex_a15_neon_load_f): Likewise.
1491 (cortex_a15_neon_store_a): Likewise.
1492 (cortex_a15_neon_store_b): Likewise.
1493 (cortex_a15_neon_store_c): Likewise.
1494 (cortex_a15_neon_store_d): Likewise.
1495 (cortex_a15_neon_store_e): Likewise.
1496 (cortex_a15_neon_store_f): Likewise.
1497 (cortex_a15_neon_store_g): Likewise.
1498 (cortex_a15_neon_store_h): Likewise.
1499 (cortex_a15_vfp_to_from_gp): Likewise.
1501 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1503 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
1505 (cortex_a9_neon_vshl_ddd): Remove.
1506 (cortex_a9_neon_vst3_vst4): Likewise.
1507 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1509 (cortex_a9_neon_bit_ops_q): New.
1511 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
1512 (cortex_a9_neon_int_2): Likewise.
1513 (cortex_a9_neon_int_3): Likewise.
1514 (cortex_a9_neon_int_4): Likewise.
1515 (cortex_a9_neon_int_5): Likewise.
1516 (cortex_a9_neon_vqneg_vqabs): Likewise.
1517 (cortex_a9_neon_vmov): Likewise.
1518 (cortex_a9_neon_vaba): Likewise.
1519 (cortex_a9_neon_vaba_qqq): Likewise.
1520 (cortex_a9_neon_shift_1): Likewise.
1521 (cortex_a9_neon_shift_2): Likewise.
1522 (cortex_a9_neon_shift_3): Likewise.
1523 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1524 (cortex_a9_neon_vsra_vrsra): Likewise.
1525 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1526 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1527 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1529 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1530 (cortex_a9_neon_mla_qqq_8_16): Likewise.
1531 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1533 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1534 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1535 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
1536 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1537 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
1538 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
1539 (cortex_a9_neon_fp_vsum): Likewise.
1540 (cortex_a9_neon_fp_vmul_ddd): Likewise.
1541 (cortex_a9_neon_fp_vmul_qqd): Likewise.
1542 (cortex_a9_neon_fp_vmla_ddd): Likewise.
1543 (cortex_a9_neon_fp_vmla_qqq): Likewise.
1544 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
1545 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
1546 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1547 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1548 (cortex_a9_neon_bp_simple): Likewise.
1549 (cortex_a9_neon_bp_2cycle): Likewise.
1550 (cortex_a9_neon_bp_3cycle): Likewise.
1551 (cortex_a9_neon_ldr): Likewise.
1552 (cortex_a9_neon_str): Likewise.
1553 (cortex_a9_neon_vld1_1_2_regs): Likewise.
1554 (cortex_a9_neon_vld1_3_4_regs): Likewise.
1555 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1556 (cortex_a9_neon_vld2_4_regs): Likewise.
1557 (cortex_a9_neon_vld3_vld4): Likewise.
1558 (cortex_a9_neon_vld1_vld2_lane): Likewise.
1559 (cortex_a9_neon_vld3_vld4_lane): Likewise.
1560 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
1561 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1562 (cortex_a9_neon_vst1_3_4_regs): Likewise.
1563 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
1564 (cortex_a9_neon_vst1_vst2_lane): Likewise.
1565 (cortex_a9_neon_vst3_vst4_lane): Likewise.
1566 (cortex_a9_neon_mcr): Likewise.
1567 (cortex_a9_neon_mcr_2_mcrr): Likewise.
1568 (cortex_a9_neon_mrc): Likewise.
1569 (cortex_a9_neon_mrrc): Likewise.
1571 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1573 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
1575 (cortex_a8_neon_vshl_ddd): Remove.
1576 (cortex_a8_neon_vst3_vst4): Likewise.
1577 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
1579 (cortex_a8_neon_bit_ops_q): New.
1581 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
1582 (cortex_a8_neon_int_2): Likewise..
1583 (cortex_a8_neon_int_3): Likewise.
1584 (cortex_a8_neon_int_5): Likewise.
1585 (cortex_a8_neon_vqneg_vqabs): Likewise.
1586 (cortex_a8_neon_int_4): Likewise.
1587 (cortex_a8_neon_vaba): Likewise.
1588 (cortex_a8_neon_vaba_qqq): Likewise.
1589 (cortex_a8_neon_shift_1): Likewise.
1590 (cortex_a8_neon_shift_2): Likewise.
1591 (cortex_a8_neon_shift_3): Likewise.
1592 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
1593 (cortex_a8_neon_vsra_vrsra): Likewise.
1594 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1595 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
1596 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
1598 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
1599 (cortex_a8_neon_mla_qqq_8_16): Likewise.
1600 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
1602 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
1603 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
1604 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
1605 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
1606 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
1607 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
1608 (cortex_a8_neon_fp_vsum): Likewise.
1609 (cortex_a8_neon_fp_vmul_ddd): Likewise.
1610 (cortex_a8_neon_fp_vmul_qqd): Likewise.
1611 (cortex_a8_neon_fp_vmla_ddd): Likewise.
1612 (cortex_a8_neon_fp_vmla_qqq): Likewise.
1613 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
1614 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
1615 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
1616 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
1617 (cortex_a8_neon_bp_simple): Likewise.
1618 (cortex_a8_neon_bp_2cycle): Likewise.
1619 (cortex_a8_neon_bp_3cycle): Likewise.
1620 (cortex_a8_neon_ldr): Likewise.
1621 (cortex_a8_neon_str): Likewise.
1622 (cortex_a8_neon_vld1_1_2_regs): Likewise.
1623 (cortex_a8_neon_vld1_3_4_regs): Likewise.
1624 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
1625 (cortex_a8_neon_vld2_4_regs): Likewise.
1626 (cortex_a8_neon_vld3_vld4): Likewise.
1627 (cortex_a8_neon_vld1_vld2_lane): Likewise.
1628 (cortex_a8_neon_vld3_vld4_lane): Likewise.
1629 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
1630 (cortex_a8_neon_vst1_3_4_regs): Likewise.
1631 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
1632 (cortex_a8_neon_vst1_vst2_lane): Likewise.
1633 (cortex_a8_neon_vst3_vst4_lane): Likewise.
1634 (cortex_a8_neon_mcr): Likewise.
1635 (cortex_a8_neon_mcr_2_mcrr): Likewise.
1636 (cortex_a8_neon_mrc): Likewise.
1637 (cortex_a8_neon_mrrc): Likewise.
1639 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1641 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
1643 * config/aarch64/aarch64-simd.md (neon_type): Remove.
1644 (aarch64_simd_dup<mode>): Add "type" attribute.
1645 (aarch64_dup_lane<mode>): Likewise.
1646 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
1647 (*aarch64_simd_mov<mode>): Likewise.
1648 (aarch64_simd_mov_from_<mode>low): Likewise.
1649 (aarch64_simd_mov_from_<mode>high): Likewise.
1650 (orn<mode>3): Likewise.
1651 (bic<mode>3): Likewise.
1652 (add<mode>3): Likewise.
1653 (sub<mode>3): Likewise.
1654 (mul<mode>3): Likewise.
1655 (*aarch64_mul3_elt<mode>): Likewise.
1656 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
1657 (*aarch64_mul3_elt_to_128df): Likewise.
1658 (*aarch64_mul3_elt_to_64v2df): Likewise.
1659 (neg<mode>2): Likewise.
1660 (abs<mode>2): Likewise.
1661 (abd<mode>_3): Likewise.
1662 (aba<mode>_3): Likewise.
1663 (fabd<mode>_3): Likewise.
1664 (*fabd_scalar<mode>3): Likewise.
1665 (and<mode>3): Likewise.
1666 (ior<mode>3): Likewise.
1667 (xor<mode>3): Likewise.
1668 (one_cmpl<mode>2): Likewise.
1669 (aarch64_simd_vec_set<mode>): Likewise.
1670 (aarch64_simd_lshr<mode>): Likewise.
1671 (aarch64_simd_ashr<mode>): Likewise.
1672 (aarch64_simd_imm_shl<mode>): Likewise.
1673 (aarch64_simd_reg_sshl<mode): Likewise.
1674 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
1675 (aarch64_simd_reg_shl<mode>_signed): Likewise.
1676 (aarch64_simd_vec_setv2di): Likewise.
1677 (aarch64_simd_vec_set<mode>): Likewise.
1678 (aarch64_mla<mode>): Likewise.
1679 (*aarch64_mla_elt<mode>): Likewise.
1680 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
1681 (aarch64_mls<mode>): Likewise.
1682 (*aarch64_mls_elt<mode>): Likewise.
1683 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
1684 (<su><maxmin><mode>3): Likewise.
1685 (move_lo_quad_<mode>): Likewise.
1686 (aarch64_simd_move_hi_quad_<mode>): Likewise.
1687 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
1688 (vec_pack_trunc_<mode>): Likewise.
1689 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
1690 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
1691 (*aarch64_<su>mlal_lo<mode>): Likewise.
1692 (*aarch64_<su>mlal_hi<mode>): Likewise.
1693 (*aarch64_<su>mlsl_lo<mode>): Likewise.
1694 (*aarch64_<su>mlsl_hi<mode>): Likewise.
1695 (*aarch64_<su>mlal<mode>): Likewise.
1696 (*aarch64_<su>mlsl<mode>): Likewise.
1697 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
1698 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
1699 (add<mode>3): Likewise.
1700 (sub<mode>3): Likewise.
1701 (mul<mode>3): Likewise.
1702 (div<mode>3): Likewise.
1703 (neg<mode>2): Likewise.
1704 (abs<mode>2): Likewise.
1705 (fma<mode>4): Likewise.
1706 (*aarch64_fma4_elt<mode>): Likewise.
1707 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
1708 (*aarch64_fma4_elt_to_128df): Likewise.
1709 (*aarch64_fma4_elt_to_64v2df): Likewise.
1710 (fnma<mode>4): Likewise.
1711 (*aarch64_fnma4_elt<mode>): Likewise.
1712 (*aarch64_fnma4_elt_<vswap_width_name><mode>
1713 (*aarch64_fnma4_elt_to_128df): Likewise.
1714 (*aarch64_fnma4_elt_to_64v2df): Likewise.
1715 (<frint_pattern><mode>2): Likewise.
1716 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
1717 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
1718 (vec_unpacks_lo_v4sf): Likewise.
1719 (aarch64_float_extend_lo_v2df): Likewise.
1720 (vec_unpacks_hi_v4sf): Likewise.
1721 (aarch64_float_truncate_lo_v2sf): Likewise.
1722 (aarch64_float_truncate_hi_v4sf): Likewise.
1723 (aarch64_vmls<mode>): Likewise.
1724 (<su><maxmin><mode>3): Likewise.
1725 (<maxmin_uns><mode>3): Likewise.
1726 (reduc_<sur>plus_<mode>): Likewise.
1727 (reduc_<sur>plus_v2di): Likewise.
1728 (reduc_<sur>plus_v2si): Likewise.
1729 (reduc_<sur>plus_<mode>): Likewise.
1730 (aarch64_addpv4sf): Likewise.
1731 (clz<mode>2): Likewise.
1732 (reduc_<maxmin_uns>_<mode>): Likewise.
1733 (reduc_<maxmin_uns>_v2di): Likewise.
1734 (reduc_<maxmin_uns>_v2si): Likewise.
1735 (reduc_<maxmin_uns>_<mode>): Likewise.
1736 (reduc_<maxmin_uns>_v4sf): Likewise.
1737 (aarch64_simd_bsl<mode>_internal): Likewise.
1738 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
1739 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
1740 (aarch64_get_lane<mode>): Likewise.
1741 (*aarch64_combinez<mode>): Likewise.
1742 (aarch64_combine<mode>): Likewise.
1743 (aarch64_simd_combine<mode>): Likewise.
1744 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
1745 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
1746 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
1747 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
1748 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
1749 (aarch64_<sur>h<addsub><mode>): Likewise.
1750 (aarch64_<sur><addsub>hn<mode>): Likewise.
1751 (aarch64_<sur><addsub>hn2<mode>): Likewise.
1752 (aarch64_pmul<mode>): Likewise.
1753 (aarch64_<su_optab><optab><mode>): Likewise.
1754 (aarch64_<sur>qadd<mode>): Likewise.
1755 (aarch64_sqmovun<mode>): Likewise.
1756 (aarch64_<sur>qmovn<mode>): Likewise.
1757 (aarch64_s<optab><mode>): Likewise.
1758 (aarch64_sq<r>dmulh<mode>): Likewise.
1759 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1760 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
1761 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
1762 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
1763 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1764 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
1765 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
1766 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
1767 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
1768 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
1769 (aarch64_sqdmull<mode>): Likewise.
1770 (aarch64_sqdmull_lane<mode>_internal): Likewise.
1771 (aarch64_sqdmull_n<mode>): Likewise.
1772 (aarch64_sqdmull2<mode>_internal): Likewise.
1773 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
1774 (aarch64_sqdmull2_n<mode>_internal): Likewise.
1775 (aarch64_<sur>shl<mode>): Likewise.
1776 (aarch64_<sur>q<r>shl<mode>
1777 (aarch64_<sur>shll_n<mode>): Likewise.
1778 (aarch64_<sur>shll2_n<mode>): Likewise.
1779 (aarch64_<sur>shr_n<mode>): Likewise.
1780 (aarch64_<sur>sra_n<mode>): Likewise.
1781 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
1782 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
1783 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
1784 (aarch64_cm<optab><mode>): Likewise.
1785 (aarch64_cm<optab>di): Likewise.
1786 (aarch64_cm<optab><mode>): Likewise.
1787 (aarch64_cm<optab>di): Likewise.
1788 (aarch64_cmtst<mode>): Likewise.
1789 (aarch64_cmtstdi): Likewise.
1790 (aarch64_cm<optab><mode>): Likewise.
1791 (*aarch64_fac<optab><mode>): Likewise.
1792 (aarch64_addp<mode>): Likewise.
1793 (aarch64_addpdi): Likewise.
1794 (sqrt<mode>2): Likewise.
1795 (vec_load_lanesoi<mode>): Likewise.
1796 (vec_store_lanesoi<mode>): Likewise.
1797 (vec_load_lanesci<mode>): Likewise.
1798 (vec_store_lanesci<mode>): Likewise.
1799 (vec_load_lanesxi<mode>): Likewise.
1800 (vec_store_lanesxi<mode>): Likewise.
1801 (*aarch64_mov<mode>): Likewise.
1802 (aarch64_ld2<mode>_dreg): Likewise.
1803 (aarch64_ld2<mode>_dreg): Likewise.
1804 (aarch64_ld3<mode>_dreg): Likewise.
1805 (aarch64_ld3<mode>_dreg): Likewise.
1806 (aarch64_ld4<mode>_dreg): Likewise.
1807 (aarch64_ld4<mode>_dreg): Likewise.
1808 (aarch64_tbl1<mode>): Likewise.
1809 (aarch64_tbl2v16qi): Likewise.
1810 (aarch64_combinev16qi): Likewise.
1811 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
1812 (aarch64_st2<mode>_dreg): Likewise.
1813 (aarch64_st2<mode>_dreg): Likewise.
1814 (aarch64_st3<mode>_dreg): Likewise.
1815 (aarch64_st3<mode>_dreg): Likewise.
1816 (aarch64_st4<mode>_dreg): Likewise.
1817 (aarch64_st4<mode>_dreg): Likewise.
1818 (*aarch64_simd_ld1r<mode>): Likewise.
1819 (aarch64_frecpe<mode>): Likewise.
1820 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
1821 (aarch64_frecps<mode>): Likewise.
1823 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
1825 * config/arm/iterators.md (V_elem_ch): New.
1827 (VQH_type): Likewise.
1828 * config/arm/arm.md (is_neon_type): New.
1829 (conds): Use is_neon_type.
1830 (anddi3_insn): Update type attribute.
1831 (xordi3_insn): Likewise.
1832 (one_cmpldi2): Likewise.
1833 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
1834 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
1835 (*movmisalign<mode>_neon_store): Likewise.
1836 (*movmisalign<mode>_neon_load): Likewise.
1837 (vec_set<mode>_internal): Likewise.
1838 (vec_set<mode>_internal): Likewise.
1839 (vec_setv2di_internal): Likewise.
1840 (vec_extract<mode>): Likewise.
1841 (vec_extract<mode>): Likewise.
1842 (vec_extractv2di): Likewise.
1843 (*add<mode>3_neon): Likewise.
1844 (adddi3_neon): Likewise.
1845 (*sub<mode>3_neon): Likewise.
1846 (subdi3_neon): Likewise.
1847 (fma<VCVTF:mode>4): Likewise.
1848 (fma<VCVTF:mode>4_intrinsic): Likewise.
1849 (*fmsub<VCVTF:mode>4): Likewise.
1850 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
1851 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
1852 (ior<mode>3): Likewise.
1853 (and<mode>3): Likewise.
1854 (orn<mode>3_neon): Likewise.
1855 (orndi3_neon): Likewise.
1856 (bic<mode>3_neon): Likewise.
1857 (bicdi3_neon): Likewise.
1858 (xor<mode>3): Likewise.
1859 (one_cmpl<mode>2): Likewise.
1860 (abs<mode>2): Likewise.
1861 (neg<mode>2): Likewise.
1862 (negdi2_neon): Likewise.
1863 (*umin<mode>3_neon): Likewise.
1864 (*umax<mode>3_neon): Likewise.
1865 (*smin<mode>3_neon): Likewise.
1866 (*smax<mode>3_neon): Likewise.
1867 (vashl<mode>3): Likewise.
1868 (vashr<mode>3_imm): Likewise.
1869 (vlshr<mode>3_imm): Likewise.
1870 (ashl<mode>3_signed): Likewise.
1871 (ashl<mode>3_unsigned): Likewise.
1872 (neon_load_count): Likewise.
1873 (ashldi3_neon_noclobber): Likewise.
1874 (ashldi3_neon): Likewise.
1875 (signed_shift_di3_neon): Likewise.
1876 (unsigned_shift_di3_neon): Likewise.
1877 (ashrdi3_neon_imm_noclobber): Likewise.
1878 (lshrdi3_neon_imm_noclobber): Likewise.
1879 (<shift>di3_neon): Likewise.
1880 (widen_ssum<mode>3): Likewise.
1881 (widen_usum<mode>3): Likewise.
1882 (quad_halves_<code>v4si): Likewise.
1883 (quad_halves_<code>v4sf): Likewise.
1884 (quad_halves_<code>v8hi): Likewise.
1885 (quad_halves_<code>v16qi): Likewise.
1886 (reduc_splus_v2di): Likewise.
1887 (neon_vpadd_internal<mode>): Likewise.
1888 (neon_vpsmin<mode>): Likewise.
1889 (neon_vpsmax<mode>): Likewise.
1890 (neon_vpumin<mode>): Likewise.
1891 (neon_vpumax<mode>): Likewise.
1892 (*ss_add<mode>_neon): Likewise.
1893 (*us_add<mode>_neon): Likewise.
1894 (*ss_sub<mode>_neon): Likewise.
1895 (*us_sub<mode>_neon): Likewise.
1896 (neon_vadd<mode>_unspec): Likewise.
1897 (neon_vaddl<mode>): Likewise.
1898 (neon_vaddw<mode>): Likewise.
1899 (neon_vhadd<mode>): Likewise.
1900 (neon_vqadd<mode>): Likewise.
1901 (neon_vaddhn<mode>): Likewise.
1902 (neon_vmul<mode>): Likewise.
1903 (neon_vfms<VCVTF:mode>): Likewise.
1904 (neon_vmlal<mode>): Likewise.
1905 (neon_vmls<mode>): Likewise.
1906 (neon_vmlsl<mode>): Likewise.
1907 (neon_vqdmulh<mode>): Likewise.
1908 (neon_vqdmlal<mode>): Likewise.
1909 (neon_vqdmlsl<mode>): Likewise.
1910 (neon_vmull<mode>): Likewise.
1911 (neon_vqdmull<mode>): Likewise.
1912 (neon_vsub<mode>_unspec): Likewise.
1913 (neon_vsubl<mode>): Likewise.
1914 (neon_vsubw<mode>): Likewise.
1915 (neon_vqsub<mode>): Likewise.
1916 (neon_vhsub<mode>): Likewise.
1917 (neon_vsubhn<mode>): Likewise.
1918 (neon_vceq<mode>): Likewise.
1919 (neon_vcge<mode>): Likewise.
1920 (neon_vcgeu<mode>): Likewise.
1921 (neon_vcgt<mode>): Likewise.
1922 (neon_vcgtu<mode>): Likewise.
1923 (neon_vcle<mode>): Likewise.
1924 (neon_vclt<mode>): Likewise.
1925 (neon_vcage<mode>): Likewise.
1926 (neon_vcagt<mode>): Likewise.
1927 (neon_vtst<mode>): Likewise.
1928 (neon_vabd<mode>): Likewise.
1929 (neon_vabdl<mode>): Likewise.
1930 (neon_vaba<mode>): Likewise.
1931 (neon_vabal<mode>): Likewise.
1932 (neon_vmax<mode>): Likewise.
1933 (neon_vmin<mode>): Likewise.
1934 (neon_vpaddl<mode>): Likewise.
1935 (neon_vpadal<mode>): Likewise.
1936 (neon_vpmax<mode>): Likewise.
1937 (neon_vpmin<mode>): Likewise.
1938 (neon_vrecps<mode>): Likewise.
1939 (neon_vrsqrts<mode>): Likewise.
1940 (neon_vqabs<mode>): Likewise.
1941 (neon_vqneg<mode>): Likewise.
1942 (neon_vcls<mode>): Likewise.
1943 (clz<mode>2): Likewise.
1944 (popcount<mode>2): Likewise.
1945 (neon_vrecpe<mode>): Likewise.
1946 (neon_vrsqrte<mode>): Likewise.
1947 (neon_vget_lane<mode>_sext_internal): Likewise.
1948 (neon_vget_lane<mode>_zext_internal): Likewise.
1949 (neon_vdup_n<mode>): Likewise.
1950 (neon_vdup_n<mode>): Likewise.
1951 (neon_vdup_nv2di): Likewise.
1952 (neon_vdup_lane<mode>_interal): Likewise.
1953 (*neon_vswp<mode>): Likewise.
1954 (neon_vcombine<mode>): Likewise.
1955 (float<mode><V_cvtto>2): Likewise.
1956 (floatuns<mode><V_cvtto>2): Likewise.
1957 (fix_trunc<mode><V_cvtto>2): Likewise.
1958 (fixuns_trunc<mode><V_cvtto>2
1959 (neon_vcvt<mode>): Likewise.
1960 (neon_vcvt<mode>): Likewise.
1961 (neon_vcvtv4sfv4hf): Likewise.
1962 (neon_vcvtv4hfv4sf): Likewise.
1963 (neon_vcvt_n<mode>): Likewise.
1964 (neon_vcvt_n<mode>): Likewise.
1965 (neon_vmovn<mode>): Likewise.
1966 (neon_vqmovn<mode>): Likewise.
1967 (neon_vqmovun<mode>): Likewise.
1968 (neon_vmovl<mode>): Likewise.
1969 (neon_vmul_lane<mode>): Likewise.
1970 (neon_vmul_lane<mode>): Likewise.
1971 (neon_vmull_lane<mode>): Likewise.
1972 (neon_vqdmull_lane<mode>): Likewise.
1973 (neon_vqdmulh_lane<mode>): Likewise.
1974 (neon_vqdmulh_lane<mode>): Likewise.
1975 (neon_vmla_lane<mode>): Likewise.
1976 (neon_vmla_lane<mode>): Likewise.
1977 (neon_vmlal_lane<mode>): Likewise.
1978 (neon_vqdmlal_lane<mode>): Likewise.
1979 (neon_vmls_lane<mode>): Likewise.
1980 (neon_vmls_lane<mode>): Likewise.
1981 (neon_vmlsl_lane<mode>): Likewise.
1982 (neon_vqdmlsl_lane<mode>): Likewise.
1983 (neon_vext<mode>): Likewise.
1984 (neon_vrev64<mode>): Likewise.
1985 (neon_vrev32<mode>): Likewise.
1986 (neon_vrev16<mode>): Likewise.
1987 (neon_vbsl<mode>_internal): Likewise.
1988 (neon_vshl<mode>): Likewise.
1989 (neon_vqshl<mode>): Likewise.
1990 (neon_vshr_n<mode>): Likewise.
1991 (neon_vshrn_n<mode>): Likewise.
1992 (neon_vqshrn_n<mode>): Likewise.
1993 (neon_vqshrun_n<mode>): Likewise.
1994 (neon_vshl_n<mode>): Likewise.
1995 (neon_vqshl_n<mode>): Likewise.
1996 (neon_vqshlu_n<mode>): Likewise.
1997 (neon_vshll_n<mode>): Likewise.
1998 (neon_vsra_n<mode>): Likewise.
1999 (neon_vsri_n<mode>): Likewise.
2000 (neon_vsli_n<mode>): Likewise.
2001 (neon_vtbl1v8qi): Likewise.
2002 (neon_vtbl2v8qi): Likewise.
2003 (neon_vtbl3v8qi): Likewise.
2004 (neon_vtbl4v8qi): Likewise.
2005 (neon_vtbl1v16qi): Likewise.
2006 (neon_vtbl2v16qi): Likewise.
2007 (neon_vcombinev16qi): Likewise.
2008 (neon_vtbx1v8qi): Likewise.
2009 (neon_vtbx2v8qi): Likewise.
2010 (neon_vtbx3v8qi): Likewise.
2011 (neon_vtbx4v8qi): Likewise.
2012 (*neon_vtrn<mode>_insn): Likewise.
2013 (*neon_vzip<mode>_insn): Likewise.
2014 (*neon_vuzp<mode>_insn): Likewise.
2015 (neon_vld1<mode>): Likewise.
2016 (neon_vld1_lane<mode>): Likewise.
2017 (neon_vld1_lane<mode>): Likewise.
2018 (neon_vld1_dup<mode>): Likewise.
2019 (neon_vld1_dup<mode>): Likewise.
2020 (neon_vld1_dupv2di): Likewise.
2021 (neon_vst1<mode>): Likewise.
2022 (neon_vst1_lane<mode>): Likewise.
2023 (neon_vst1_lane<mode>): Likewise.
2024 (neon_vld2<mode>): Likewise.
2025 (neon_vld2<mode>): Likewise.
2026 (neon_vld2_lane<mode>): Likewise.
2027 (neon_vld2_lane<mode>): Likewise.
2028 (neon_vld2_dup<mode>): Likewise.
2029 (neon_vst2<mode>): Likewise.
2030 (neon_vst2<mode>): Likewise.
2031 (neon_vst2_lane<mode>): Likewise.
2032 (neon_vst2_lane<mode>): Likewise.
2033 (neon_vld3<mode>): Likewise.
2034 (neon_vld3qa<mode>): Likewise.
2035 (neon_vld3qb<mode>): Likewise.
2036 (neon_vld3_lane<mode>): Likewise.
2037 (neon_vld3_lane<mode>): Likewise.
2038 (neon_vld3_dup<mode>): Likewise.
2039 (neon_vst3<mode>): Likewise.
2040 (neon_vst3qa<mode>): Likewise.
2041 (neon_vst3qb<mode>): Likewise.
2042 (neon_vst3_lane<mode>): Likewise.
2043 (neon_vst3_lane<mode>): Likewise.
2044 (neon_vld4<mode>): Likewise.
2045 (neon_vld4qa<mode>): Likewise.
2046 (neon_vld4qb<mode>): Likewise.
2047 (neon_vld4_lane<mode>): Likewise.
2048 (neon_vld4_lane<mode>): Likewise.
2049 (neon_vld4_dup<mode>): Likewise.
2050 (neon_vst4<mode>): Likewise.
2051 (neon_vst4qa<mode>): Likewise.
2052 (neon_vst4qb<mode>): Likewise.
2053 (neon_vst4_lane<mode>): Likewise.
2054 (neon_vst4_lane<mode>): Likewise.
2055 (neon_vec_unpack<US>_lo_<mode>): Likewise.
2056 (neon_vec_unpack<US>_hi_<mode>): Likewise.
2057 (neon_vec_<US>mult_lo_<mode>): Likewise.
2058 (neon_vec_<US>mult_hi_<mode>): Likewise.
2059 (neon_vec_<US>shiftl_<mode>): Likewise.
2060 (neon_unpack<US>_<mode>): Likewise.
2061 (neon_vec_<US>mult_<mode>): Likewise.
2062 (vec_pack_trunc_<mode>): Likewise.
2063 (neon_vec_pack_trunc_<mode>): Likewise.
2064 (neon_vabd<mode>_2): Likewise.
2065 (neon_vabd<mode>_3): Likewise.
2067 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2069 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
2070 (load_pair): Update type attribute.
2071 (store_pair): Update type attribute.
2072 * config/aarch64/iterators.md (q): New.
2074 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2076 * config/arm/types.md: Add new types for Neon insns.
2078 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2079 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2080 Sergey Lega <sergey.s.lega@intel.com>
2081 Anna Tikhonova <anna.tikhonova@intel.com>
2082 Ilya Tocar <ilya.tocar@intel.com>
2083 Andrey Turetskiy <andrey.turetskiy@intel.com>
2084 Ilya Verbin <ilya.verbin@intel.com>
2085 Kirill Yukhin <kirill.yukhin@intel.com>
2086 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2088 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
2089 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
2090 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
2092 (srcp14<mode>): Ditto.
2093 (rsqrt14<mode>): Ditto.
2094 (rsqrt14<mode>): Ditto.
2095 (avx512f_vmscalef<mode>): Ditto.
2096 (avx512f_scalef<mode>): Ditto.
2097 (avx512f_getexp<mode>): Ditto.
2098 (avx512f_sgetexp<mode>): Ditto.
2099 (avx512f_fixupimm<mode>): Ditto.
2100 (avx512f_sfixupimm<mode>): Ditto.
2101 (avx512f_rndscale<mode>): Ditto.
2102 (*avx512er_exp2<mode>): Ditto.
2103 (*avx512er_rcp28<mode>): Ditto.
2104 (avx512er_rsqrt28<mode>): Ditto.
2105 (avx512f_getmant<mode>): Ditto.
2106 (avx512f_getmant<mode>): Ditto.
2107 (avx512f_rndscale<mode>): Fix formatting.
2109 2013-10-15 Martin Jambor <mjambor@suse.cz>
2111 * ipa-utils.h (ipa_edge_within_scc): Declare.
2112 * ipa-cp.c (edge_within_scc): Moved...
2113 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
2115 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2116 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2117 Sergey Lega <sergey.s.lega@intel.com>
2118 Anna Tikhonova <anna.tikhonova@intel.com>
2119 Ilya Tocar <ilya.tocar@intel.com>
2120 Andrey Turetskiy <andrey.turetskiy@intel.com>
2121 Ilya Verbin <ilya.verbin@intel.com>
2122 Kirill Yukhin <kirill.yukhin@intel.com>
2123 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2125 * config/i386/predicates.md (const_8_to_15_operand): New.
2126 (const_16_to_31_operand): Ditto.
2127 * config/i386/sse.md (V8FI): New.
2129 (reduc_splus_v8df): Ditto.
2130 (reduc_splus_v16sf): Ditto.
2131 (avx512f_vextract<shuffletype>32x4_1): Ditto.
2132 (vec_extract_hi_<mode>): Ditto.
2133 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
2134 (vec_set_lo_<mode>): Ditto.
2135 (vec_set_hi_<mode>): Ditto.
2136 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
2137 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
2138 (avx512f_pshufd_1): Ditto.
2139 (avx512f_broadcast<mode>): Ditto.
2140 (avx512f_broadcast<mode>): Ditto.
2141 (define_split): Split vec_extract_lo into move.
2142 (ssequartermode): Ditto.
2143 (ssedoublemode): Extened with wider modes.
2144 (vec_extract_lo_<mode>): Ditto.
2146 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2147 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2148 Sergey Lega <sergey.s.lega@intel.com>
2149 Anna Tikhonova <anna.tikhonova@intel.com>
2150 Ilya Tocar <ilya.tocar@intel.com>
2151 Andrey Turetskiy <andrey.turetskiy@intel.com>
2152 Ilya Verbin <ilya.verbin@intel.com>
2153 Kirill Yukhin <kirill.yukhin@intel.com>
2154 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2156 * config/i386/predicates.md (register_or_constm1_operand): New.
2157 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
2158 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
2159 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
2160 UNSPEC_SCATTER_PREFETCH
2162 (avx512f_ucmp<mode>3): Ditto.
2163 (avx512f_vternlog<mode>): Ditto.
2164 (avx512f_align<mode>): Ditto.
2165 (<shift_insn><mode>3): Ditto.
2166 (avx512f_<rotate>v<mode>): Ditto.
2167 (avx512f_<rotate><mode>): Ditto.
2168 (avx512f_eq<mode>3): Ditto.
2169 (avx512f_eq<mode>3_1): Ditto.
2170 (avx512f_gt<mode>3): Ditto.
2171 (avx512f_testm<mode>3): Ditto.
2172 (avx512f_testnm<mode>3): Ditto.
2173 (avx512pf_gatherpf<mode>): Ditto.
2174 (*avx512pf_gatherpf<mode>_mask): Ditto.
2175 (*avx512pf_gatherpf<mode>): Ditto.
2176 (avx512pf_scatterpf<mode>): Ditto.
2177 (*avx512pf_scatterpf<mode>_mask): Ditto.
2178 (*avx512pf_scatterpf<mode>): Ditto.
2179 (avx512f_vec_dup_gpr<mode>): Ditto.
2180 (clz<mode>2): Ditto.
2181 (conflict<mode>): Ditto.
2182 (REDUC_SMINMAX_MODE): Extened with wider modes.
2183 (reduc_<code>_<mode>): Ditto.
2184 (vlshr<mode>3): Ditto.
2185 (vashl<mode>3): Ditto.
2187 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
2188 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2189 Sergey Lega <sergey.s.lega@intel.com>
2190 Anna Tikhonova <anna.tikhonova@intel.com>
2191 Ilya Tocar <ilya.tocar@intel.com>
2192 Andrey Turetskiy <andrey.turetskiy@intel.com>
2193 Ilya Verbin <ilya.verbin@intel.com>
2194 Kirill Yukhin <kirill.yukhin@intel.com>
2195 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2197 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
2200 (avx512fmaskmode): Ditto.
2201 (bcstscalarsuff): Ditto.
2202 (avx512f_blendm<mode>): Ditto.
2203 (cmp_imm_predicate): Ditto.
2204 (avx512f_cmp<mode>3): Ditto.
2205 (avx512f_vec_dup<mode>): Ditto.
2206 (avx512f_vec_dup_mem<mode>): Ditto.
2207 (avx512f_vpermi2var<mode>3): Ditto.
2208 (avx512f_vpermt2var<mode>3): Ditto.
2209 (vec_init<mode>): Ditto.
2210 (avx512f_gathersi<mode>): Ditto.
2211 (*avx512f_gathersi<mode>): Ditto.
2212 (*avx512f_gathersi<mode>_2): Ditto.
2213 (avx512f_gatherdi<mode>): Ditto.
2214 (*avx512f_gatherdi<mode>): Ditto.
2215 (*avx512f_gatherdi<mode>_2): Ditto.
2216 (avx512f_scattersi<mode>): Ditto.
2217 (*avx512f_scattersi<mode>): Ditto.
2218 (avx512f_scatterdi<mode>): Ditto.
2219 (*avx512f_scatterdi<mode>): Ditto.
2220 (sseintprefix): Extened with wider modes.
2221 (VEC_GATHER_IDXSI): Ditto.
2222 (VEC_GATHER_IDXDI): Ditto.
2223 (VEC_GATHER_SRCDI): Ditto.
2225 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
2226 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
2228 * config/arm/t-aprofile: New file.
2229 * config.gcc: Handle --with-multilib-list option.
2231 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
2233 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
2236 2013-10-15 Richard Biener <rguenther@suse.de>
2238 * tree-loop-distribution.c (build_empty_rdg): Inline into
2240 (rdg_flag_vertex): Inline into single user.
2241 (rdg_flag_vertex_and_dependent): Likewise.
2242 (build_rdg_partition_for_vertex): Remove processed bitmap.
2243 (rdg_build_partitions): Simplify.
2245 2013-10-15 Richard Biener <rguenther@suse.de>
2247 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
2248 Restructure forwarding through conversions and copies to
2249 avoid performing copy-propagation the wrong way. Adjust
2250 recursion invocations.
2251 (forward_propagate_addr_expr): Add argument stating if we
2252 are recursing from a single-use.
2253 (ssa_forward_propagate_and_combine): Adjust.
2255 2013-10-14 David Malcolm <dmalcolm@redhat.com>
2257 * dumpfile.h (gcc::dump_manager): New class, to hold state
2258 relating to dumpfile management.
2259 (get_dump_file_name): Remove in favor of method of dump_manager.
2260 (dump_initialized_p): Likewise.
2261 (dump_start): Likewise.
2262 (dump_finish): Likewise.
2263 (dump_switch_p): Likewise.
2264 (dump_register): Likewise.
2265 (get_dump_file_info): Likewise.
2266 * context.c (gcc::context::context): Construct the dump_manager
2268 * context.h (gcc::context::get_dumps): New.
2269 (gcc::context::m_dumps): New.
2270 * coverage.c (coverage_init): Port to dump_manager API.
2271 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
2272 (extra_dump_files_in_use): Likewise.
2273 (extra_dump_files_alloced): Likewise.
2274 (gcc::dump_manager::dump_manager): New.
2275 (dump_register): Convert to...
2276 (gcc::dump_manager::dump_register): ...method, replacing
2277 function-static next_dump with m_next_dump field.
2278 (get_dump_file_info): Convert to...
2279 (gcc::dump_manager::get_dump_file_info): ...method.
2280 (get_dump_file_name): Convert to...
2281 (gcc::dump_manager::get_dump_file_name): ...method.
2282 (dump_start): Convert to...
2283 (gcc::dump_manager::dump_start): ...method.
2284 (dump_finish): Convert to...
2285 (gcc::dump_manager::dump_finish): ...method.
2286 (dump_begin): Replace body with...
2287 (gcc::dump_manager::dump_begin): ...new method.
2288 (dump_phase_enabled_p): Convert to...
2289 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2290 (dump_phase_enabled_p): Convert to...
2291 (gcc::dump_manager::dump_phase_enabled_p): ...method.
2292 (dump_initialized_p): Convert to...
2293 (gcc::dump_manager::dump_initialized_p): ...method.
2294 (dump_flag_name): Replace body with...
2295 (gcc::dump_manager::dump_flag_name): ...new method.
2296 (dump_enable_all): Convert to...
2297 (gcc::dump_manager::dump_enable_all): ...new method.
2298 (opt_info_enable_passes): Convert to...
2299 (gcc::dump_manager::opt_info_enable_passes): ...new method.
2300 (dump_switch_p_1): Convert to...
2301 (gcc::dump_manager::dump_switch_p_1): ...new method.
2302 (dump_switch_p): Convert to...
2303 (gcc::dump_manager::dump_switch_p): ...new method.
2304 (opt_info_switch_p): Port to dump_manager API.
2305 (enable_rtl_dump_file): Likewise.
2306 * opts-global.c (handle_common_deferred_options): Port to new
2308 * passes.c (pass_manager::finish_optimization_passes): Likewise.
2309 (pass_manager::register_one_dump_file): Likewise.
2310 (pass_manager::register_pass): Likewise.
2311 (pass_init_dump_file): Likewise.
2312 (pass_fini_dump_file): Likewise.
2313 * statistics.c (statistics_early_init): Likewise.
2315 2013-10-14 Richard Biener <rguenther@suse.de>
2317 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
2318 iterative_hash_canonical_type, gimple_canonical_type_hash,
2319 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
2320 gimple_register_canonical_type, print_gimple_types_stats,
2321 free_gimple_type_tables): Move to lto/lto.c
2322 (gt-gimple.h): Do not include.
2323 * gimple.h (gimple_register_canonical_type,
2324 print_gimple_types_stats, free_gimple_type_tables): Remove.
2325 * Makefile.in (GTFILES): Remove gimple.c.
2327 2013-10-14 Travis Snoozy <quandary@remstate.com>
2330 * config/msp430/msp430.c (msp430_option_override): Correct thinko
2331 scanning for msp430x targets.
2333 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
2336 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
2337 (registers_ok_for_ldd_peep): Move around.
2338 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
2339 * config/sparc/sparc.md (widening peepholes): Use it.
2341 2013-10-14 Richard Biener <rguenther@suse.de>
2345 * gimple.c (iterative_hash_canonical_type): Make sure to
2346 record the hash into the correct hashtable slot.
2348 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
2350 PR rtl-optimization/58662
2351 * combine.c (try_combine): Take into account death nodes on I2 when
2352 splitting a PARALLEL of two independent SETs. Fix dump message.
2354 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
2357 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
2358 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
2359 * config/sh/t-sh (sh_treg_combine.o): New entry.
2360 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
2361 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
2362 (register_sh_passes): New function. Register sh_treg_combine pass.
2363 (sh_option_override): Invoke it.
2364 (sh_canonicalize_comparison): Handle op0_preserve_value.
2365 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
2366 opportunities. Canonicalize branch condition.
2367 (nott): Allow only if pseudos can be created for non-SH2A.
2369 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
2372 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
2373 (ix86_expand_movmem): Replace copy_addr_to_reg with
2374 ix86_copy_addr_to_reg.
2375 (ix86_expand_setmem): Likewise.
2377 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
2379 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
2380 provided by ix86_fp_compare_mode instead of CCFPUmode.
2382 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
2384 * config/aarch64/arm_neon.h
2385 (vtbx<1,3>_<psu>8): Fix register constriants.
2387 2013-10-11 Jeff Law <law@redhat.com>
2389 PR tree-optimization/58640
2390 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
2391 threading paths that cross over two loop entry points.
2393 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2395 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
2396 handle vector float as well.
2397 (*vsx_le_perm_load_v4si): Likewise.
2398 (*vsx_le_perm_store_v2di): Likewise.
2399 (*vsx_le_perm_store_v4si): Likewise.
2401 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2403 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
2404 directly to circumvent subtract from splat{31} workaround.
2405 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
2407 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
2408 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
2409 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
2410 (altivec_vperm_<mode>): Convert to define_insn_and_split to
2411 separate big and little endian logic.
2412 (*altivec_vperm_<mode>_internal): New define_insn.
2413 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
2414 separate big and little endian logic.
2415 (*altivec_vperm_<mode>_uns_internal): New define_insn.
2416 (vec_permv16qi): Add little endian logic.
2418 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
2420 * doc/extend.texi (returns_nonnull): Remove arguments.
2422 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2423 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2424 Sergey Lega <sergey.s.lega@intel.com>
2425 Anna Tikhonova <anna.tikhonova@intel.com>
2426 Ilya Tocar <ilya.tocar@intel.com>
2427 Andrey Turetskiy <andrey.turetskiy@intel.com>
2428 Ilya Verbin <ilya.verbin@intel.com>
2429 Kirill Yukhin <kirill.yukhin@intel.com>
2430 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2432 * config/i386/sse.md (VI48F_256_512): New.
2433 (avx2_permvar<mode>): Change to ...
2434 (<avx2_avx512f>_permvar<mode>): This.
2436 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2437 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2438 Sergey Lega <sergey.s.lega@intel.com>
2439 Anna Tikhonova <anna.tikhonova@intel.com>
2440 Ilya Tocar <ilya.tocar@intel.com>
2441 Andrey Turetskiy <andrey.turetskiy@intel.com>
2442 Ilya Verbin <ilya.verbin@intel.com>
2443 Kirill Yukhin <kirill.yukhin@intel.com>
2444 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2446 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
2447 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
2448 * config/i386/sse.md (VI4_AVX): New.
2449 (sf2simodelower): Ditto.
2450 (sse2_cvtps2dq): Change to ...
2451 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
2453 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2454 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2455 Sergey Lega <sergey.s.lega@intel.com>
2456 Anna Tikhonova <anna.tikhonova@intel.com>
2457 Ilya Tocar <ilya.tocar@intel.com>
2458 Andrey Turetskiy <andrey.turetskiy@intel.com>
2459 Ilya Verbin <ilya.verbin@intel.com>
2460 Kirill Yukhin <kirill.yukhin@intel.com>
2461 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2463 * config/i386/sse.md (V_512): New.
2465 (vcond<V_512:mode><VF_512:mode>): Ditto.
2466 (vcond<V_512:mode><VI_512:mode>): Ditto.
2467 (vcondu<V_512:mode><VI_512:mode>): Ditto.
2469 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2470 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2471 Sergey Lega <sergey.s.lega@intel.com>
2472 Anna Tikhonova <anna.tikhonova@intel.com>
2473 Ilya Tocar <ilya.tocar@intel.com>
2474 Andrey Turetskiy <andrey.turetskiy@intel.com>
2475 Ilya Verbin <ilya.verbin@intel.com>
2476 Kirill Yukhin <kirill.yukhin@intel.com>
2477 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2479 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
2480 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
2482 (fma<mode>4): Removed condition.
2483 (fms<mode>4): Ditto.
2484 (fnma<mode>4): Ditto.
2485 (fnms<mode>4): Ditto.
2486 (fma4i_fmadd_<mode>): Ditto.
2487 (*fma_fmadd_<mode>): Ditto.
2488 (*fma_fmsub_<mode>): Ditto.
2489 (*fma_fnmadd_<mode>): Ditto.
2490 (*fma_fnmsub_<mode>): Ditto.
2491 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
2492 (*fma_fmaddsub_<mode>): Ditto.
2493 (*fma_fmsubadd_<mode>): Ditto.
2494 (*fmai_fmadd_<mode>): Ditto.
2495 (*fmai_fmsub_<mode>): Ditto.
2496 (*fmai_fnmadd_<mode>): Ditto.
2497 (*fmai_fnmsub_<mode>): Ditto.
2499 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2500 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2501 Sergey Lega <sergey.s.lega@intel.com>
2502 Anna Tikhonova <anna.tikhonova@intel.com>
2503 Ilya Tocar <ilya.tocar@intel.com>
2504 Andrey Turetskiy <andrey.turetskiy@intel.com>
2505 Ilya Verbin <ilya.verbin@intel.com>
2506 Kirill Yukhin <kirill.yukhin@intel.com>
2507 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2509 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
2510 (VI124_256): Changed to ...
2511 (VI124_256_48_512): This.
2512 (ssepackmode): Extended with wider modes.
2513 (<code><mode>3): Changed iterator.
2514 (*avx2_<code><mode>3): Ditto.
2515 (vec_pack_trunc_<mode>): Ditto.
2517 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2518 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2519 Sergey Lega <sergey.s.lega@intel.com>
2520 Anna Tikhonova <anna.tikhonova@intel.com>
2521 Ilya Tocar <ilya.tocar@intel.com>
2522 Andrey Turetskiy <andrey.turetskiy@intel.com>
2523 Ilya Verbin <ilya.verbin@intel.com>
2524 Kirill Yukhin <kirill.yukhin@intel.com>
2525 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2527 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
2528 (VI8F_256_512): Ditto.
2529 (abs<mode>2): Changed iterator.
2530 (avx2_perm<mode>): Changed to ...
2531 (<avx2_avx512f>_perm<mode>): This.
2532 (avx2_perm<mode>_1): Changed to ...
2533 (<avx2_avx512f>_perm<mode>_1): This.
2535 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2536 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2537 Sergey Lega <sergey.s.lega@intel.com>
2538 Anna Tikhonova <anna.tikhonova@intel.com>
2539 Ilya Tocar <ilya.tocar@intel.com>
2540 Andrey Turetskiy <andrey.turetskiy@intel.com>
2541 Ilya Verbin <ilya.verbin@intel.com>
2542 Kirill Yukhin <kirill.yukhin@intel.com>
2543 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2545 * config/i386/sse.md (VI48_AVX512F): New.
2546 (VI48_AVX2): Changed to ...
2547 (VI48_AVX2_48_AVX512F): This.
2548 (avx2_ashrv<mode>): Changed to ...
2549 (<avx2_avx512f>_ashrv<mode>): This.
2550 (avx2_<shift_insn>v<mode>): Changed to ...
2551 (<avx2_avx512f>_<shift_insn>v<mode>): This.
2553 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2554 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2555 Sergey Lega <sergey.s.lega@intel.com>
2556 Anna Tikhonova <anna.tikhonova@intel.com>
2557 Ilya Tocar <ilya.tocar@intel.com>
2558 Andrey Turetskiy <andrey.turetskiy@intel.com>
2559 Ilya Verbin <ilya.verbin@intel.com>
2560 Kirill Yukhin <kirill.yukhin@intel.com>
2561 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2563 * config/i386/sse.md (VI4_AVX512F): New.
2564 (VI8_AVX2_AVX512F): Ditto.
2565 (mul<mode>3): Extended with wider modes.
2566 (*<sse4_1_avx2>_mul<mode>3): Ditto.
2567 (mul<mode>3): Ditto.
2568 (vec_widen_<s>mult_odd_<mode>): Ditto.
2570 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2571 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2572 Sergey Lega <sergey.s.lega@intel.com>
2573 Anna Tikhonova <anna.tikhonova@intel.com>
2574 Ilya Tocar <ilya.tocar@intel.com>
2575 Andrey Turetskiy <andrey.turetskiy@intel.com>
2576 Ilya Verbin <ilya.verbin@intel.com>
2577 Kirill Yukhin <kirill.yukhin@intel.com>
2578 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2580 * config/i386/sse.md (VI2_AVX512F): New.
2581 (VI124_AVX512F): Ditto.
2582 (sseunpackmode): Extended with wider modes.
2583 (sseunpackfltmode): Ditto.
2584 (vec_unpacks_float_hi_<mode>): Ditto.
2585 (vec_unpacks_float_lo_<mode>): Ditto.
2586 (vec_unpacku_float_hi_<mode>): Ditto.
2587 (vec_unpacku_float_lo_<mode>): Ditto.
2588 (vec_unpacks_lo_<mode>): Ditto.
2589 (vec_unpacks_hi_<mode>): Ditto.
2590 (vec_unpacku_lo_<mode>): Ditto.
2591 (vec_unpacku_hi_<mode>): Ditto.
2593 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2594 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2595 Sergey Lega <sergey.s.lega@intel.com>
2596 Anna Tikhonova <anna.tikhonova@intel.com>
2597 Ilya Tocar <ilya.tocar@intel.com>
2598 Andrey Turetskiy <andrey.turetskiy@intel.com>
2599 Ilya Verbin <ilya.verbin@intel.com>
2600 Kirill Yukhin <kirill.yukhin@intel.com>
2601 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2603 * config/i386/i386.md (multdiv): New.
2604 (multdiv_mnemonic): Ditto.
2605 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
2606 (<sse>_vm<multdiv_mnemonic><mode>3): This.
2607 (<sse>_vmdiv<mode>3): Removed.
2609 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2610 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2611 Sergey Lega <sergey.s.lega@intel.com>
2612 Anna Tikhonova <anna.tikhonova@intel.com>
2613 Ilya Tocar <ilya.tocar@intel.com>
2614 Andrey Turetskiy <andrey.turetskiy@intel.com>
2615 Ilya Verbin <ilya.verbin@intel.com>
2616 Kirill Yukhin <kirill.yukhin@intel.com>
2617 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2619 * config/i386/sse.md (V): Extended with wider modes.
2621 (ssehalfvecmode): Ditto.
2623 (ssepackfltmode): Ditto.
2624 (avx_vec_concat<mode>): Ditto.
2625 (V_256_512): New iterator.
2626 (VF2_512_256): Ditto.
2627 (si2dfmode): New attribute.
2628 (si2dfmodelower): Ditto.
2630 (concat_tg_mode): Ditto.
2631 (floatv4siv4df2): Changed to ...
2632 (float<si2dfmodelower><mode>2): This.
2633 (avx_cvtps2pd256): Changed to ...
2634 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
2635 (vec_pack_trunc_v4df): Changed to ...
2636 (vec_pack_trunc_<mode>): This.
2637 (avx_vpermil<mode>): Changed to ...
2638 (<sse2_avx_avx512f>_vpermil<mode>): This.
2639 (<fixsuffix>fix_truncv8dfv8si2): New.
2640 (vec_pack_sfix_trunc_v8df): Ditto.
2641 (avx512f_rndscale<mode>): Ditto.
2642 (avx512f_roundpd512): Ditto.
2643 (vec_pack_ufix_trunc_<mode>): Updated iterator.
2645 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2646 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2647 Sergey Lega <sergey.s.lega@intel.com>
2648 Anna Tikhonova <anna.tikhonova@intel.com>
2649 Ilya Tocar <ilya.tocar@intel.com>
2650 Andrey Turetskiy <andrey.turetskiy@intel.com>
2651 Ilya Verbin <ilya.verbin@intel.com>
2652 Kirill Yukhin <kirill.yukhin@intel.com>
2653 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2655 * config/i386/i386.md (any_fix): New iterator.
2656 (fixsuffix): New attribute.
2657 * config/i386/sse.md (VF1): Extened with wider modes.
2661 (sseintvecmodelower): Ditto.
2662 (ssescalarmode): Ditto.
2663 (ssescalarnum): Ditto.
2665 (ssexmmmode): Ditto.
2666 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
2667 (<sse>_rcp<mode>2): Change iterator.
2668 (rsqrt<mode>2): Ditto.
2669 (<sse>_rsqrt<mode>2): Ditto.
2670 (avx2_vec_dup<mode>): Ditto.
2671 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
2672 (round<mode>2_sfix): Ditto.
2673 (avx2_pbroadcast<mode>): Ditto.
2674 (*andnot<mode>3): Handle XI mode.
2675 (*<code><mode>3): Ditto.
2676 (AVXTOSSEMODE): Removed.
2677 (avx_vpermil<mode>): Changed to ...
2678 (<sse2_avx_avx512f>_vpermil<mode>): This.
2680 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
2681 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2682 Sergey Lega <sergey.s.lega@intel.com>
2683 Anna Tikhonova <anna.tikhonova@intel.com>
2684 Ilya Tocar <ilya.tocar@intel.com>
2685 Andrey Turetskiy <andrey.turetskiy@intel.com>
2686 Ilya Verbin <ilya.verbin@intel.com>
2687 Kirill Yukhin <kirill.yukhin@intel.com>
2688 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2690 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
2691 (<sse>_comi): Ditto.
2692 (<sse>_ucomi): Ditto.
2693 (sse_cvtss2siq_2): Ditto.
2694 (sse2_cvtsd2si): Ditto.
2695 (sse2_cvtsd2siq): Ditto.
2696 (sse2_cvttsd2si): Ditto.
2697 (sse2_cvttsd2siq): Ditto.
2698 (<shift_insn><mode>3): Ditto.
2699 (sse2_cvtsi2sdq): Update constraint and prefix.
2700 (sse_cvtsi2ss): Update prefix.
2701 (sse_cvtsi2ssq): Ditto.
2703 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2705 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
2706 ignore internal calls.
2708 2013-10-11 Richard Biener <rguenther@suse.de>
2710 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
2711 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
2712 TREE_OVERFLOW is set.
2714 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
2716 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
2718 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
2719 * gimple.h: Likewise.
2721 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
2723 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
2726 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
2728 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
2729 * configure: Regenerate.
2731 2013-10-11 Jakub Jelinek <jakub@redhat.com>
2733 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
2734 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
2735 formatting fixes, use pp_colon instead of pp_character (..., ':'),
2736 similarly pp_right_paren.
2737 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
2738 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
2739 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
2740 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
2742 (omp_declare_simd_clauses_equal,
2743 omp_remove_redundant_declare_simd_attrs): New functions.
2744 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
2745 (walk_tree_1): Handle new OpenMP 4.0 clauses.
2746 * tree.h (OMP_LOOP_CHECK): Define.
2747 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
2748 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
2749 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
2750 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
2751 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
2752 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
2753 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
2754 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
2755 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
2756 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
2757 OMP_CLAUSE_SIMDLEN_EXPR): Define.
2758 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
2759 (omp_remove_redundant_declare_simd_attrs): New prototype.
2760 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
2761 GIMPLE_OMP_TEAMS): New codes.
2762 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
2763 * omp-low.c (struct omp_context): Add cancel_label and cancellable
2765 (target_nesting_level): New variable.
2766 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
2767 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
2768 for collapse > 1 static schedule unless ordered.
2769 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
2770 (determine_parallel_type): Adjust get_ws_args_for caller.
2771 (install_var_field): Handle mask & 4 for double indirection.
2772 (scan_sharing_clauses): Ignore shared clause on teams construct.
2773 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
2774 (create_omp_child_function): If inside target or declare target
2775 constructs, set "omp declare target" attribute on the child function.
2776 (find_combined_for): New function.
2777 (scan_omp_parallel): Handle combined loops.
2778 (scan_omp_target, scan_omp_teams): New functions.
2779 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
2780 restrictions and set ctx->cancellable for cancellable constructs.
2781 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
2782 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
2783 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
2784 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
2785 (omp_clause_aligned_alignment): New function.
2786 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
2787 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
2788 on teams constructs. Handle user defined reductions and new
2790 (lower_reduction_clauses): Don't set placeholder to address of ref
2791 if it has already the right type.
2792 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
2793 (expand_parallel_call): Use the new non-_start suffixed builtins,
2794 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
2795 and GOMP_parallel_end after the call.
2796 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
2797 (expand_omp_for_init_counts): Handle combined loops.
2798 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
2800 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
2801 end of cancellable loops.
2802 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
2803 Likewise. Handle collapse > 1 loops.
2804 (expand_omp_simd): Handle combined loops.
2805 (expand_omp_for): Add inner_stmt argument, adjust callers of
2806 expand_omp_for* functions, use expand_omp_for_static*chunk even
2807 for collapse > 1 unless ordered.
2808 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
2809 of cancellable sections.
2810 (expand_omp_single): Remove need_barrier variable, just rely on
2811 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
2812 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
2813 (expand_omp_atomic_load, expand_omp_atomic_store,
2814 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
2815 (expand_omp_target): New function.
2816 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
2817 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
2818 (build_omp_regions_1): Immediately close region for
2819 GF_OMP_TARGET_KIND_UPDATE.
2820 (maybe_add_implicit_barrier_cancel): New function.
2821 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
2823 (lower_omp_single): Likewise. Add clobber after the barrier.
2824 (lower_omp_taskgroup): New function.
2825 (lower_omp_for): Handle combined loops. Adjust
2826 lower_rec_input_clauses caller. Handle cancellation.
2827 (lower_depend_clauses): New function.
2828 (lower_omp_taskreg): Lower depend clauses. Adjust
2829 lower_rec_input_clauses caller. Add clobber after the call. Handle
2831 (lower_omp_target, lower_omp_teams): New functions.
2832 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
2833 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
2834 and GOMP_cancellation_point calls.
2835 (lower_omp): Fold stmts inside of target region.
2836 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
2837 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2838 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
2839 (BT_FN_VOID_OMPFN_PTR_UINT,
2840 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
2841 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
2842 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
2843 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
2844 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
2845 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
2846 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
2847 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
2848 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
2849 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
2850 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
2851 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
2852 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
2853 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
2854 BUILT_IN_GOMP_PARALLEL_END.
2855 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
2856 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
2857 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
2858 GF_OMP_FOR_KIND_DISTRIBUTE.
2859 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
2860 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
2861 (dump_gimple_omp_return): Print lhs if it has any.
2862 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
2863 gimple_omp_atomic_seq_cst_p.
2864 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
2865 and GIMPLE_OMP_TEAMS.
2866 * langhooks.c (lhd_omp_mappable_type): New function.
2867 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
2868 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
2870 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
2871 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
2872 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
2873 (struct gimplify_omp_ctx): Add combined_loop field.
2874 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
2875 on stmts inside of target region.
2876 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
2877 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
2878 ORT_TARGET and ORT_TARGET_DATA.
2879 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
2881 (omp_notice_threadprivate_variable): Complain about threadprivate
2882 variables in target region.
2883 (omp_notice_variable): Complain about vars with non-mappable type
2884 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
2885 (omp_check_private): Ignore ORT_TARGET* regions.
2886 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
2887 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
2888 (find_combined_omp_for): New function.
2889 (gimplify_omp_for): Handle gimplification of combined loops.
2890 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
2892 (gimplify_omp_target_update): New function.
2893 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
2894 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
2895 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
2896 (gimplify_body): If fndecl has "omp declare target" attribute, add
2897 implicit ORT_TARGET context around it.
2898 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
2899 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
2900 * tree-nested.c (convert_nonlocal_reference_stmt,
2901 convert_local_reference_stmt, convert_gimple_call): Handle
2902 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2903 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
2904 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
2905 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
2906 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
2907 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
2908 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
2909 BUILT_IN_GOMP_LOOP_END_CANCEL,
2910 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
2911 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
2912 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
2913 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
2914 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
2915 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
2916 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
2917 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
2918 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
2919 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
2920 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
2921 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
2922 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
2923 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
2924 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
2925 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2926 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
2927 gimple_build_omp_teams): New functions.
2928 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
2929 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
2930 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
2931 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2932 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
2933 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
2934 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
2935 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
2936 GF_OMP_ATOMIC_SEQ_CST): New.
2937 (gimple_build_omp_taskgroup, gimple_build_omp_target,
2938 gimple_build_omp_teams): New prototypes.
2939 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
2940 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2941 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
2942 GIMPLE_OMP_SINGLE as end of range.
2943 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
2944 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
2945 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
2946 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
2947 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
2948 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
2949 gimple_omp_target_kind, gimple_omp_target_set_kind,
2950 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
2951 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
2952 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
2953 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
2954 gimple_omp_teams_set_clauses): New inlines.
2955 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
2956 and GIMPLE_OMP_TASKGROUP.
2957 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
2958 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
2959 enum omp_clause_proc_bind_kind): New.
2960 (union omp_clause_subcode): Add depend_kind, map_kind and
2961 proc_bind_kind fields.
2962 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
2963 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
2964 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
2965 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
2966 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
2968 2013-10-10 Teresa Johnson <tejohnson@google.com>
2970 * predict.c (tree_estimate_probability): Add new parameter
2971 for estimate_bb_frequencies.
2972 (estimate_bb_frequencies): Add new parameter to force estimation.
2973 (rebuild_frequencies): When max frequency in function is small,
2974 recompute counts from frequencies.
2975 * predict.h (estimate_bb_frequencies): New parameter.
2977 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2979 * ipa-inline.c (ipa_inline): Fix leak of "order" when
2980 optimizations are disabled.
2982 2013-10-10 David Malcolm <dmalcolm@redhat.com>
2984 * coverage.c (coverage_finish): Fix leak of da_file_name.
2986 2013-10-10 Jan Hubicka <jh@suse.cz>
2988 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
2989 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
2992 2013-10-10 Jakub Jelinek <jakub@redhat.com>
2995 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
2996 if any labels are in FALLTHRU_BB, use a special label emitted
2997 immediately after the asm goto insn rather than label_rtx
2999 (expand_asm_stmt): Adjust caller.
3000 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
3001 edge if the last insn in predecessor is a jump with single successor,
3002 but it isn't simplejump_p.
3004 2013-10-10 Richard Biener <rguenther@suse.de>
3006 PR tree-optimization/58656
3007 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
3009 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3011 * gimplify.c: Include expr.h and tm_p.h for targets with special
3012 va-arg padding requirements.
3014 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3016 * tree-flow.h: Move some prototypes to gimple.h.
3017 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
3018 * gimple.h: Relocate some prototypes from tree-flow.h
3019 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
3021 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
3022 (build_va_arg_indirect_ref): Relocate and make static.
3023 (std_gimplify_va_arg_expr): Relocate here.
3024 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
3025 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
3027 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3029 * doc/md.texi: Document the mnemonic attribute.
3031 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3034 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
3035 (set_attr_alternative x ...) when searching for user defined
3038 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
3040 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
3042 2013-10-09 Easwaran Raman <eraman@google.com>
3044 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
3045 * cfgexpand.c (defer_stack_allocation): ...use here
3046 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
3048 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
3050 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
3051 (operand_equal_for_value_replacement): New function, extracted from
3052 value_replacement and enhanced to catch more cases.
3053 (value_replacement): Use operand_equal_for_value_replacement.
3055 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3057 * loop-doloop.c (doloop_modify, doloop_optimize): Use
3058 get_max_loop_iterations.
3060 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3062 * config/arm/aarch-common.c (arm_early_load_addr_dep):
3063 Place comment above function.
3065 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3067 * tree-flow.h: Remove all remaining prototypes, enums and structs that
3068 are not related to tree-cfg.c.
3069 * tree-ssa-address.h: New file. Relocate prototypes.
3070 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
3071 (addr_for_mem_ref): New. Combine call to get_address_description and
3072 return addr_for_mem_ref.
3073 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
3074 * tree-ssa-live.h: Adjust prototypes.
3075 * passes.c: Include tree-ssa-live.h.
3076 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
3077 * graphite.c (graphite_transform_loops): Make static.
3078 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
3079 make_pass_graphite, pass_data_graphite_transforms,
3080 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
3081 * ipa-pure-const.c (warn_function_noreturn): Make static.
3082 (execute_warn_function_noreturn, gate_warn_function_noreturn,
3083 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
3084 Relocate from tree-cfg.c
3085 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
3087 (execute_warn_function_noreturn, gate_warn_function_noreturn,
3088 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
3089 Move to ipa-pure-const.c.
3090 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
3091 Relocate from tree-optimize.c.
3092 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
3093 make_pass_fixup_cfg): Move to tree-cfg.c.
3094 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
3095 Relocate some prototypes.
3096 * tree-data-ref.h (tree_check_data_deps) Add prototype.
3097 * tree-dump.c (dump_function_to_file): Remove prototype.
3098 Add tree-flow.h to the include file.
3099 * tree-dump.h: Remove prototype.
3100 * tree-parloops.h: New File. Add prototypes.
3101 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
3102 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
3103 from tree-ssa-loop.c.
3104 * tree-predcom.c (run_tree_predictive_commoning,
3105 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
3106 Relocate here from tree-ssa-loop.c.
3107 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
3108 ssa_name_values.release ().
3109 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
3110 (ssa_name_values): Relocate from tree-flow.h.
3111 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
3112 * tree-ssa-loop.c (run_tree_predictive_commoning,
3113 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
3114 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
3115 make_pass_graphite, pass_data_graphite_transforms,
3116 make_pass_graphite_transforms, gate_tree_parallelize_loops,
3117 tree_parallelize_loops, pass_data_parallelize_loops,
3118 make_pass_parallelize_loops): Move to other files.
3119 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
3121 * tree.h: Remove prototypes from tree-address.c.
3123 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3125 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
3126 (struct int_tree_map): Move to tree-hasher.h
3127 (SCALE, LABEL, PERCENT): Move to gimple.h
3128 * tree-flow-inline.h: Delete. Move functions to other files.
3129 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
3130 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
3131 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
3132 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
3133 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
3134 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
3135 * tree-hasher.h: Don't include tree-flow.h.
3136 (struct int_tree_map): Relocate from tree-flow.h.
3137 * tree-sra.c (contains_view_convert_expr_p): Relocate from
3138 tree-flow-inline.h and make static.
3139 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
3141 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
3142 tree-flow-inline.h and make static.
3143 * tree.h (is_global_var, may_be_aliased): Relocate from
3145 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
3146 * value-prof.c: No longer include tree-flow-inline.h.
3147 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
3149 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
3151 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
3152 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
3153 (enum move_pos): Move to tree-ssa-loop-im.h
3154 * cfgloop.h: Move some prototypes.
3155 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
3156 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
3157 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
3158 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
3159 (loop_containing_stmt): Relocate from tree-flow-inline.h.
3160 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
3161 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
3162 (enum move_pos): Relocate here.
3163 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
3165 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
3166 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
3167 make_pass_lim): Relocate here from tree-ssa-loop.c.
3168 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
3170 (loop_edge_to_cancel, unloop_loops): Make static.
3171 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3172 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
3173 (tree_complete_unroll, gate_tree_complete_unroll,
3174 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
3175 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3176 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
3177 * tree-ssa-loop-ivopts.c: Remove local prototypes.
3178 (stmt_invariant_in_loop_p): Remove unused function.
3179 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
3180 * tree-ssa-loop-manip.h: New file. Add prototypes.
3181 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
3182 (gcov_type_to_double_int): Move to cfgloop.h.
3183 (double_int_cmp, bound_index,
3184 estimate_numbers_of_iterations_loop): Make static.
3185 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
3186 (max_loop_iterations): Factor out get_max_loop_iterations.
3187 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
3189 * tree-ssa-loop-niter.h: New file. Add prototypes.
3190 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
3191 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
3192 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
3193 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
3194 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
3195 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
3196 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
3197 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
3198 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
3199 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
3200 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
3201 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
3202 pass_data_complete_unroll, make_pass_complete_unroll,
3203 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
3204 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
3205 tree-ssa-loop-ivcanon.c.
3206 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
3207 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
3208 tree-ssa-loop-prefetch.c.
3209 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
3211 (get_lsm_tmp_name): Relocate and add suffix parameter.
3212 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
3213 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
3214 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
3215 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
3216 (get_estimated_loop_iterations): Factor out accessor from
3217 estimated_loop_iterations in tree-ssa-loop-niter.c.
3218 (get_max_loop_iterations): Factor out accessor from
3219 _max_loop_iterations in tree-ssa-niter.c.
3220 * loop-unroll.c (decide_unroll_constant_iterations,
3221 decide_unroll_runtime_iterations, decide_peel_simple,
3222 decide_unroll_stupid): Use new get_* accessors.
3224 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
3226 PR tree-optimization/20318
3227 * doc/extend.texi (returns_nonnull): New function attribute.
3228 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
3230 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
3231 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
3233 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
3236 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
3237 false if both components are bitfields.
3239 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3241 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
3242 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
3243 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
3244 * config/aarch64/aarch64.h
3245 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
3246 * config/aarch64/aarch64-simd-builtins.def
3247 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
3249 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3251 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
3253 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3255 * config/aarch64/arm_neon.h (vdiv_f64): Added.
3257 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
3259 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
3260 (vneg_f64): New intrinsic.
3261 (vneg_s8): Asm replaced with C.
3262 (vneg_s16): Likewise.
3263 (vneg_s32): Likewise.
3264 (vneg_s64): New intrinsic.
3265 (vnegq_f32): Asm replaced with C.
3266 (vnegq_f64): Likewise.
3267 (vnegq_s8): Likewise.
3268 (vnegq_s16): Likewise.
3269 (vnegq_s32): Likewise.
3270 (vnegq_s64): Likewise.
3272 2013-10-09 Renlin Li <Renlin.Li@arm.com>
3274 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
3276 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3278 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
3279 packed stack special handling.
3280 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
3281 back to fixed stack slots for FPRs saved due to stdarg.
3283 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3285 * config/s390/s390.c (s390_frame_info): Restructure function.
3287 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3289 * config/s390/s390.c (struct s390_frame_layout): New field
3291 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
3292 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
3293 regs_ever_clobbered to char*.
3294 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
3295 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
3296 (s390_register_info): Enable FPR save slots. Move/Copy some
3297 functionality into ...
3298 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
3299 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
3301 (s390_frame_info): Do gpr slot allocation here now. stdarg does
3302 not imply a stack frame.
3303 (s390_init_frame_layout): Remove variable clobbered_regs.
3304 (s390_update_register_info): Remove function.
3305 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
3306 cannot be used for register renaming.
3307 (s390_hard_regno_scratch_ok): New function.
3308 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
3309 (s390_initial_elimination_offset): Change offset calculation of
3310 the return address pointer.
3311 (save_gprs): Deal with only r6 being saved from the call-saved regs.
3312 (restore_gprs): Set frame related flag.
3313 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
3314 (s390_emit_prologue): Call s390_register_info instead of
3315 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
3316 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
3317 (s390_optimize_prologue): Call s390_optimize_register_info.
3318 Try to remove also FPR slot save/restore INSNs. Remove frame
3319 related flags from restore INSNs.
3321 2013-10-08 DJ Delorie <dj@redhat.com>
3323 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
3326 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
3327 avoid conflict with the MI use of %c.
3328 * config/rl78/rl78-real.md: change %c to %C throughout.
3329 * config/rl78/rl78-virt.md: Likewise.
3331 2013-10-08 Jan Hubicka <jh@suse.cz>
3333 * config/i386/i386.c (ix86_option_override_internal): Switch
3334 to SSE math for -ffast-math when target ISA supports SSE2.
3336 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
3338 * tree-flow.h: Remove some prototypes.
3339 * tree.h: Remove some protypes, add a couple.
3340 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
3341 using_eh_for_cleanups_p): Add interface routines for front ends.
3342 * tree-eh.h: New file. Add protoptyes.
3343 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
3344 (add_stmt_to_eh_lp_fn): Make static.
3345 (lower_try_finally): Use new using_eh_for_cleanups_p.
3346 * emit-rtl.c: Include tree-eh.h.
3347 * gimple.h: Include tree-eh.h.
3349 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
3351 PR tree-optimization/58480
3352 * tree-vrp.c (infer_nonnull_range): New function.
3353 (infer_value_range): Call infer_nonnull_range.
3355 2013-10-08 Dehao Chen <dehao@google.com>
3357 PR tree-optimization/58619
3358 * tree-inline.c (copy_phis_for_bb): Combine location data
3361 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
3364 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
3365 RTX_FRAME_RELATED_P on INSN.
3367 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3369 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
3370 (altivec_expand_vec_perm_const): Call it.
3372 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3374 * config/rs6000/vector.md (mov<mode>): Emit permuted move
3375 sequences for LE VSX loads and stores at expand time.
3376 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
3378 * config/rs6000/rs6000.c (rs6000_const_vec): New.
3379 (rs6000_gen_le_vsx_permute): New.
3380 (rs6000_gen_le_vsx_load): New.
3381 (rs6000_gen_le_vsx_store): New.
3382 (rs6000_gen_le_vsx_move): New.
3383 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
3384 (*vsx_le_perm_load_v4si): New.
3385 (*vsx_le_perm_load_v8hi): New.
3386 (*vsx_le_perm_load_v16qi): New.
3387 (*vsx_le_perm_store_v2di): New.
3388 (*vsx_le_perm_store_v4si): New.
3389 (*vsx_le_perm_store_v8hi): New.
3390 (*vsx_le_perm_store_v16qi): New.
3391 (*vsx_xxpermdi2_le_<mode>): New.
3392 (*vsx_xxpermdi4_le_<mode>): New.
3393 (*vsx_xxpermdi8_le_V8HI): New.
3394 (*vsx_xxpermdi16_le_V16QI): New.
3395 (*vsx_lxvd2x2_le_<mode>): New.
3396 (*vsx_lxvd2x4_le_<mode>): New.
3397 (*vsx_lxvd2x8_le_V8HI): New.
3398 (*vsx_lxvd2x16_le_V16QI): New.
3399 (*vsx_stxvd2x2_le_<mode>): New.
3400 (*vsx_stxvd2x4_le_<mode>): New.
3401 (*vsx_stxvd2x8_le_V8HI): New.
3402 (*vsx_stxvd2x16_le_V16QI): New.
3404 2013-10-07 Renlin Li <Renlin.Li@arm.com>
3406 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
3408 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3410 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
3411 loop to work also for 31bit ABI.
3412 Save the stack pointer for frame_size > 0.
3414 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
3416 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
3417 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
3418 constraint letters from expanders.
3419 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
3420 retry count to general_operand.
3421 ("tabort"): Give operand 0 a mode.
3422 ("tabort_1"): Add mode and constraint letter for operand 0.
3423 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
3425 2013-10-04 Jeff Law <law@redhat.com>
3427 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
3429 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
3431 (thread_across_edge): Here. Call it.
3433 2013-10-04 Cary Coutant <ccoutant@google.com>
3435 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
3436 discarding a location list expression (or a piece of one).
3438 2013-10-03 Jan Hubicka <jh@suse.cz>
3440 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
3441 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3442 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
3444 2013-10-03 Jan Hubicka <jh@suse.cz>
3446 * config/i386/i386.c (ix86_option_override_internal): Do not enable
3447 accumulate-outgoing-args when producing unwind info.
3449 2013-10-03 Wei Mi <wmi@google.com>
3451 * lra-constraints.c (insert_move_for_subreg): New function
3452 extracted from simplify_operand_subreg.
3453 (simplify_operand_subreg): Add reload for paradoxical subreg.
3455 2013-10-03 Rong Xu <xur@google.com>
3457 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
3458 the candidate of builtin_expect such that we should fix the
3459 size/time estimation.
3460 (estimate_function_body_sizes): Do the acutally size/time fix-up
3463 2013-10-03 Rong Xu <xur@google.com>
3465 * predict.c (tree_predict_by_opcode): Get the probability
3466 for builtin_expect from param builtin_expect_probability.
3467 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
3468 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
3469 * doc/invoke.texi: Add documentation for builtin-expect-probability.
3471 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
3474 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
3475 * calls.c (alloca_call_p): Use get_callee_fndecl.
3476 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
3477 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
3479 (vrp_visit_stmt): Remove duplicated code.
3481 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
3483 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
3484 ceildf2, btruncdf2, instead of vsx_* name.
3486 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
3487 iterators to only do V2DF and V4SF here. Move the DF code to
3488 rs6000.md where it is combined with SF mode. Replace <VSv> with
3489 just 'v' since only vector operations are handled with these insns
3490 after moving the DF support to rs6000.md.
3491 (vsx_sub<mode>3): Likewise.
3492 (vsx_mul<mode>3): Likewise.
3493 (vsx_div<mode>3): Likewise.
3494 (vsx_fre<mode>2): Likewise.
3495 (vsx_neg<mode>2): Likewise.
3496 (vsx_abs<mode>2): Likewise.
3497 (vsx_nabs<mode>2): Likewise.
3498 (vsx_smax<mode>3): Likewise.
3499 (vsx_smin<mode>3): Likewise.
3500 (vsx_sqrt<mode>2): Likewise.
3501 (vsx_rsqrte<mode>2): Likewise.
3502 (vsx_fms<mode>4): Likewise.
3503 (vsx_nfma<mode>4): Likewise.
3504 (vsx_copysign<mode>3): Likewise.
3505 (vsx_btrunc<mode>2): Likewise.
3506 (vsx_floor<mode>2): Likewise.
3507 (vsx_ceil<mode>2): Likewise.
3508 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
3509 (vsx_sminsf3): Likewise.
3510 (vsx_fmadf4): Likewise.
3511 (vsx_fmsdf4): Likewise.
3512 (vsx_nfmadf4): Likewise.
3513 (vsx_nfmsdf4): Likewise.
3514 (vsx_cmpdf_internal1): Likewise.
3516 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
3517 simpler to select whether a target has SPE or traditional floating
3518 point support in iterators.
3519 (TARGET_DF_SPE): Likewise.
3520 (TARGET_SF_FPR): Likewise.
3521 (TARGET_DF_FPR): Likewise.
3522 (TARGET_SF_INSN): Macros to say whether floating point support
3523 exists for a given operation for expanders.
3524 (TARGET_DF_INSN): Likewise.
3526 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
3527 combining of SF/DF mode operations, using both traditional and VSX
3535 (abs<mode>2): Combine SF/DF modes using traditional floating point
3536 instructions. Add support for using the upper DF registers with
3537 VSX support, and SF registers with power8-vector support. Update
3538 expanders for operations supported by both the SPE and traditional
3539 floating point units.
3540 (abs<mode>2_fpr): Likewise.
3541 (nabs<mode>2): Likewise.
3542 (nabs<mode>2_fpr): Likewise.
3543 (neg<mode>2): Likewise.
3544 (neg<mode>2_fpr): Likewise.
3545 (add<mode>3): Likewise.
3546 (add<mode>3_fpr): Likewise.
3547 (sub<mode>3): Likewise.
3548 (sub<mode>3_fpr): Likewise.
3549 (mul<mode>3): Likewise.
3550 (mul<mode>3_fpr): Likewise.
3551 (div<mode>3): Likewise.
3552 (div<mode>3_fpr): Likewise.
3553 (sqrt<mode>3): Likewise.
3554 (sqrt<mode>3_fpr): Likewise.
3555 (fre<Fs>): Likewise.
3556 (rsqrt<mode>2): Likewise.
3557 (cmp<mode>_fpr): Likewise.
3558 (smax<mode>3): Likewise.
3559 (smin<mode>3): Likewise.
3560 (smax<mode>3_vsx): Likewise.
3561 (smin<mode>3_vsx): Likewise.
3562 (negsf2): Delete SF operations that are merged with DF.
3569 (fmasf4_fpr): Likewise.
3570 (fmssf4_fpr): Likewise.
3571 (nfmasf4_fpr): Likewise.
3572 (nfmssf4_fpr): Likewise.
3573 (sqrtsf2): Likewise.
3574 (rsqrtsf_internal1): Likewise.
3575 (smaxsf3): Likewise.
3576 (sminsf3): Likewise.
3577 (cmpsf_internal1): Likewise.
3578 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
3579 (negdf2): Delete DF operations that are merged with SF.
3581 (nabsdf2): Likewise.
3587 (rsqrtdf_internal1): Likewise.
3588 (fmadf4_fpr): Likewise.
3589 (fmsdf4_fpr): Likewise.
3590 (nfmadf4_fpr): Likewise.
3591 (nfmsdf4_fpr): Likewise.
3592 (sqrtdf2): Likewise.
3593 (smaxdf3): Likewise.
3594 (smindf3): Likewise.
3595 (cmpdf_internal1): Likewise.
3596 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
3597 (btrunc<mode>2): Delete separate expander, and combine with the
3598 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
3599 (btrunc<mode>2_fpr): Likewise.
3600 (ceil<mode>2): Likewise.
3601 (ceil<mode>2_fpr): Likewise.
3602 (floor<mode>2): Likewise.
3603 (floor<mode>2_fpr): Likewise.
3604 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
3605 Add support for using the upper registers with VSX and
3606 power8-vector. Move insns to be closer to the define_expands. On
3607 VSX systems, prefer the traditional form of FMA over the VSX
3608 version, since the traditional form allows the target not to
3609 overlap with the inputs.
3610 (fms<mode>4_fpr): Likewise.
3611 (nfma<mode>4_fpr): Likewise.
3612 (nfms<mode>4_fpr): Likewise.
3614 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3615 Richard Earnshaw <richard.earnshaw@arm.com>
3617 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
3618 (struct mult_cost_table): Likewise.
3619 (struct mem_cost_table): Likewise.
3620 (struct fp_cost_table): Likewise.
3621 (struct vector_cost_table): Likewise.
3622 (cpu_cost_table): Likewise.
3623 * config/arm/arm.opt (mold-rts-costs): New option.
3624 (mnew-generic-costs): Likewise.
3625 * config/arm/arm.c (generic_extra_costs): New table.
3626 (cortexa15_extra_costs): Likewise.
3627 (arm_slowmul_tune): Use NULL as new costs.
3628 (arm_fastmul_tune): Likewise.
3629 (arm_strongarm_tune): Likewise.
3630 (arm_xscale_tune): Likewise.
3631 (arm_9e_tune): Likewise.
3632 (arm_v6t2_tune): Likewise.
3633 (arm_cortex_a5_tune): Likewise.
3634 (arm_cortex_a9_tune): Likewise.
3635 (arm_v6m_tune): Likewise.
3636 (arm_fa726te_tune): Likewise.
3637 (arm_cortex_a15_tune): Use cortex15_extra_costs.
3638 (arm_cortex_tune): Use generict_extra_costs.
3639 (shifter_op_p): New function.
3640 (arm_unspec_cost): Likewise.
3641 (LIBCALL_COST): Define.
3642 (arm_new_rtx_costs): New function.
3643 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
3644 table is available. Use old costs otherwise unless mnew-generic-costs
3646 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
3647 (cpu_cost_table): Declare.
3649 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
3652 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
3653 (*subs_mul_imm_<mode>)
3654 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
3655 (*sub_<shift>_<mode>)
3656 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
3657 Remove k constraint.
3659 2013-10-03 Ian Bolton <ian.bolton@arm.com>
3661 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
3663 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
3665 2013-10-02 Teresa Johnson <tejohnson@google.com>
3667 * predict.c (probably_never_executed): New function.
3668 (probably_never_executed_bb_p): Invoke probably_never_executed.
3669 (probably_never_executed_edge_p): Ditto.
3670 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
3671 Treat profile insanities conservatively.
3673 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
3675 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
3677 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
3679 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
3680 use it. Use smaller increase for scratch. Don't increase reject
3681 for early clobber scratch.
3682 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
3683 setting eliminated regs except setting fp from hfp.
3684 (lra_eliminate): Check lra_insn_recog_data on NULL.
3686 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
3689 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
3690 setting -mvsx-timode by default until the underlying problem is fixed.
3691 (RS6000_CPU, power7 defaults): Likewise.
3693 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3695 * config/x-linux (host-linux.o): Remove header dependencies.
3696 Use $(COMPILE) and $(POSTCOMPILE).
3697 * config/t-linux-android (linux-android.o): Ditto.
3699 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3701 * Makefile.in (expmed.o-warn): Remove.
3703 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3705 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
3706 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
3708 2013-10-02 Teresa Johnson <tejohnson@google.com>
3710 * dojump.c (do_jump_1): Divide probability between
3711 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
3713 2013-10-02 Tom Tromey <tromey@redhat.com>
3715 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
3717 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3719 * tree-flow.h: Remove some prototypes.
3720 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
3721 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
3722 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
3723 mark_virtual_phi_result_for_renaming): Relocate here.
3724 * tree-into-ssa.h: Add prototypes.
3725 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3726 single_pred_before_succ_order.
3727 (blocks_in_phiopt_order): Rename and move to cfganal.c.
3728 (nonfreeing_call_p) Move to gimple.c.
3729 * cfganal.c (single_pred_before_succ_order): Move and renamed from
3731 * basic-block.h (single_pred_before_succ_order): Add prototype.
3732 * gimple.c (nonfreeing_call_p): Relocate here.
3733 * gimple.h: Add prototype.
3734 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
3735 * tree-ssa-dom.h: New file. Relocate prototypes here.
3736 * tree-ssa.h: Include tree-ssa-dom.h.
3738 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
3740 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
3741 Use $(COMPILE) and $(POSTCOMPILE).
3743 * config/alpha/x-alpha (driver-alpha.o): Ditto.
3745 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3747 * tree-flow.h: Remove some prototypes.
3748 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
3749 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3750 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
3751 propagate_tree_value*): Move from here to...
3752 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
3753 propagate_tree_value*): Relocate here.
3754 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
3755 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
3756 * gimple-fold.c: Remove gimple-fold.h from include list.
3757 * tree-vrp.c: Remove gimple-fold.h from include list.
3758 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
3759 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
3760 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
3761 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
3762 * sese.c: Add tree-ssa-propagate.h to include list.
3764 2013-10-02 Richard Biener <rguenther@suse.de>
3766 * tree-loop-distribution.c: Include tree-vectorizer.h for
3768 (enum partition_kind): Remove PKIND_REDUCTION.
3769 (struct partition_s): Remove has_writes member, add reduction_p member.
3770 (partition_alloc): Adjust.
3771 (partition_builtin_p): Likewise.
3772 (partition_has_writes): Remove.
3773 (partition_reduction_p): New function.
3774 (partition_merge_into): Likewise.
3775 (generate_code_for_partition): Commonize builtin partition
3777 (rdg_cannot_recompute_vertex_p): Remove.
3778 (already_processed_vertex_p): Likewise.
3779 (rdg_flag_vertex): Do not set has_writes.
3780 (classify_partition): Adjust.
3781 (rdg_build_partitions): Do not set has_writes, treat all
3782 partitions as useful.
3783 (distribute_loop): Record number of library calls generated. Adjust.
3784 (tree_loop_distribution): Report number of loops and library
3785 calls generated as opt-info.
3787 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
3789 * tree-flow.h: Include new .h files. Move prototypes.
3790 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
3791 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
3792 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
3793 * tree-pretty-print.h: Add prototypes from tree-flow.h.
3794 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
3795 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
3796 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
3797 * tree.h (get_ref_base_and_extent): Move prototype out.
3798 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
3800 * gimple-low.h: New File. Add prototypes from tree-flow.h.
3801 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
3802 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
3803 * tree-scalar-evolution.c: Include tree.h.
3804 * sese.c: Include tree.h.
3805 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
3806 * dwarf2out.c: Include tree-dfa.h.
3807 * tree-chrec.c: Include tree.h.
3808 * tree-data-ref.c: Include tree.h.
3810 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
3812 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
3815 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
3817 * config/t-sol2 (sol2-c.o): Remove header dependencies.
3818 Use $(COMPILE) and $(POSTCOMPILE).
3819 (sol2-cxx.o): Likewise.
3820 (sol2-stubs.o): Likewise.
3822 * config/x-solaris (host-solaris.o): Likewise.
3824 * config/sparc/t-sparc (sparc.o): Remove.
3825 (sparc-c.o): Remove header dependencies.
3826 Use $(COMPILE) and $(POSTCOMPILE).
3827 * config/sparc/x-sparc: Likewise.
3829 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
3831 * config/arc/arc-opts.h: Add 2013 to Copyright years.
3832 * config/arc/arc700.md: Likewise.
3833 * config/arc/arc-modes.def: Likewise.
3834 * config/arc/arc-simd.h: Likewise.
3835 * config/arc/t-arc-uClibc: Likewise.
3836 * config/arc/t-arc-newlib: Likewise.
3838 2013-10-02 Renlin Li <renlin.li@arm.com>
3840 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3842 (aarch64_expand_epilogue): Likewise.
3844 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3845 Yufeng Zhang <yufeng.zhang@arm.com>
3847 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
3849 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
3850 'base_in' represent a conversion and legal_cast_p_1 holds; set
3851 'base_in' with the returned value from get_unwidened.
3853 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3855 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
3856 plus_constant is not used.
3858 2013-10-01 Wei Mi <wmi@google.com>
3860 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
3861 * config/i386/i386.md: Add define_peephole2 to
3862 break partial reg stall for cvtss2sd/cvtsd2ss.
3864 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3866 * config/arc/arc.c (pass_arc_ifcvt::clone):
3867 Update for ctxt_ -> m_ctxt change.
3869 2013-10-01 Jeff Law <law@redhat.com>
3871 * tree-ssa-threadupdate.c (struct redirection_data): Delete
3872 outgoing_edge and intermediate_edge fields. Instead store the path.
3873 (redirection_data::hash): Hash on the last edge's destination index.
3874 (redirection_data::equal): Check the entire thread path.
3875 (lookup_redirectio_data): Corresponding changes.
3876 (create_edge_and_update_destination_phis): Likewise.
3877 (thread_single_edge): Likewise.
3879 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3880 Diego Novillo <dnovillo@google.com>
3882 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
3883 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
3884 (vld32wh_insn, vld32wl_insn): Delete commented-out old
3885 versions of these patterns.
3887 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
3888 (__builtin_arc_aligned): Likewise.
3890 * config/arc/arc.md: Expand adc_0 comment stating the intended
3891 purpose and why it isn't ready.
3892 Replace commented out call_value_via_label_mixed with a
3893 plain comment about bl_s.
3895 * config/arc/arc.c (stdio.h): Don't include directly.
3896 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
3897 Remove if (1) condition.
3898 (arc_encode_section_info): Fix comment.
3900 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3902 * config/arc/arc.c (arc_conditional_register_usage):
3903 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
3904 Also set reg_alloc_order for DMA config regs.
3906 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3907 Jeremy Bennett <jeremy.bennett@embecosm.com>
3909 * doc/install.texi (--with-cpu): Mention ARC.
3910 (arc-*-elf32): New paragraph.
3911 (arc-linux-uclibc): Likewise.
3912 * doc/md.texi (Machine Constraints): Add ARC part.
3913 * doc/invoke.texi: (menu): Add ARC Options.
3914 (Machine Dependent Options) <ARC Options>: Add synopsis.
3915 (node ARC Options): Add.
3916 * doc/extend.texi (long_call / short_call attribute): Add ARC.
3917 (ARC Built-in Functions): New section defining
3918 generic ARC built-in functions.
3919 (ARC SIMD Built-in Functions): New section defining SIMD specific
3921 (Declaring Attributes of Functions): Extended
3922 description of short_call and long_call attributes for ARC and
3923 added index entries.
3925 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
3926 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
3927 Joern Rennecke <joern.rennecke@embecosm.com>
3928 Muhammad Khurram Riaz <khurram.riaz@arc.com>
3929 Brendan Kehoe <brendan@zen.org>
3930 Michael Eager <eager@eagercon.com>
3931 Simon Cook <simon.cook@embecosm.com>
3932 Jeremy Bennett <jeremy.bennett@embecosm.com>
3934 * config/arc, common/config/arc: New directories.
3936 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
3937 Brendan Kehoe <brendan@zen.org>
3938 Simon Cook <simon.cook@embecosm.com>
3940 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
3942 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3944 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
3945 * tree-ssa-coalesce.h: New. Move prototype to here.
3946 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
3947 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
3948 (gimple_can_coalesce_p): Move to...
3949 * gimple.c (gimple_can_coalesce_p): Here.
3951 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
3953 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
3954 (dump_decl_set): Move to gimple.c.
3955 * gimple.h: Don't include tree-ssa-operands.h.
3956 (dump_decl_set): Add prototype.
3957 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3958 Move to gimple-ssa.h.
3959 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
3960 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
3961 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
3962 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
3963 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
3965 (dump_decl_set): Relocate here.
3966 * gimple-ssa.h: New file.
3967 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
3968 Relocate from gimple.h.
3969 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
3970 * tree-ssa-operands.c (swap_ssa_operands): Rename from
3971 swap_tree_operands and remove non-ssa path.
3972 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
3973 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
3975 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
3976 vect_is_simple_reduction_1): Use swap_ssa_operands.
3977 * tree-flow.h: Move various prototypes to tree-phinodes.h.
3978 (enum need_phi_state): Move to tree-into-ssa.c.
3979 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
3980 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
3981 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
3982 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
3983 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
3984 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
3985 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
3986 num_imm_uses): Move to ssa-iterators.h.
3987 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
3988 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
3990 (op_iter_done, op_iter_next_def, op_iter_next_tree,
3991 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
3992 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
3993 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
3994 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
3995 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
3996 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
3997 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
3998 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
3999 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
4000 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
4001 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
4002 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
4003 (set_phi_nodes): Move to tree-phinodes.h.
4004 * tree-ssa-operands.h (enum ssa_op_iter_type,
4005 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
4006 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
4007 (dump_decl_set): Remove prototype.
4008 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
4009 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
4010 (set_phi_nodes): Relocate from tree-flow-inline.h.
4011 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
4013 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
4014 include list. Temporarily add gimple.h to include list.
4015 * ssa-iterators.h: New file.
4016 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
4017 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
4018 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
4019 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
4020 Relocate from tree-ssa-operands.h.
4021 (delink_imm_use, link_imm_use_to_list, link_imm_use,
4022 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
4023 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
4024 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
4025 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
4026 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
4027 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
4028 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
4029 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
4030 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
4031 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
4032 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
4033 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
4034 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
4035 Relocate from tree-flow-inline.h.
4036 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
4038 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
4041 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
4042 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
4043 Insert '\t' to output template.
4044 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
4045 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
4046 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
4047 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
4049 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
4051 * doc/install.texi (Host/target specific installation notes for GCC):
4052 Put @anchor before @heading.
4053 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
4054 Use @email for email addresses.
4056 2013-10-01 Jeff Law <law@redhat.com>
4058 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
4059 a vec. Only delete the path if we create one without successfully
4060 registering a jump thread.
4061 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
4063 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
4064 (paths): New vector of jump threading paths.
4065 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
4066 (THREAD_PATH): New accessor macro for the entire thread path.
4067 (lookup_redirection_data): Get intermediate and final outgoing edge
4068 from the thread path.
4069 (create_edge_and_update_destination_phis): Copy the threading path.
4070 (ssa_fix_duplicate_block_edges): Get edges and block types from the
4071 jump threading path.
4072 (ssa_redirect_edges): Get edges and block types from the jump threading
4073 path. Free the path vector.
4074 (thread_block): Get edges from the jump threading path. Look at the
4075 entire path to see if we thread to a loop exit. If we cancel a jump
4076 thread request, then free the path vector.
4077 (thread_single_edge): Get edges and block types from the jump threading
4078 path. Free the path vector.
4079 (thread_through_loop_header): Get edges and block types from the jump
4080 threading path. Free the path vector.
4081 (mark_threaded_blocks): Iterate over the vector of paths and store
4082 the path on the appropriate edge. Get edges and block types from the
4083 jump threading path.
4084 (mark_threaded_blocks): Get edges and block types from the jump
4085 threading path. Free the path vector.
4086 (thread_through_all_blocks): Use the vector of paths rather than
4087 a vector of 3-edge sets.
4088 (register_jump_thread): Accept pointer to a path vector rather
4089 than the path vector itself. Store the path vector for later use.
4092 2013-10-01 Jakub Jelinek <jakub@redhat.com>
4093 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4096 * config/s390/s390.c (s390_split_branches): Modify check for table
4098 (s390_chunkify_start): Rearrange table jump insn check in order to
4099 deal with compare and branch insns correctly.
4101 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
4105 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
4106 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
4107 define_insn_and_split.
4108 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
4109 (shiftsi3_compare): New pattern.
4111 * config/arm/unspecs.md (UNSPEC_RRX): New.
4113 2013-10-01 Alan Modra <amodra@gmail.com>
4115 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
4116 casing inout operands.
4118 2013-10-01 Richard Biener <rguenther@suse.de>
4120 PR tree-optimization/58553
4121 * tree-loop-distribution.c (struct partition_s): Add niter member.
4122 (classify_partition): Populate niter member for the partition
4123 and properly identify whether the relevant store happens before
4124 or after the loop exit.
4125 (generate_memset_builtin): Use niter member from the partition.
4126 (generate_memcpy_builtin): Likewise.
4128 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
4130 * vec.h (vec_prefix, vec): Prefix member names with "m_".
4131 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
4133 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
4135 * basic-block.h (edge_list): Prefix member names with "m_".
4136 * context.h (context): Likewise.
4137 * domwalk.h (dom_walker): Likewise.
4138 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
4139 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
4140 * hash-table.h (hash_table): Likewise.
4141 * machmode.h (bit_field_mode_iterator): Likewise.
4142 * pass_manager.h (pass_list): Likewise.
4143 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
4144 * tree-pass.h (pass_data): Likewise.
4145 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
4146 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
4147 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
4148 * asan.c (pass_data_asan): Update accordingly.
4149 * cfganal.c (control_dependences::find_control_dependence): Likewise.
4150 (control_dependences::control_dependences): Likewise.
4151 (control_dependences::~control_dependences): Likewise.
4152 (control_dependences::~control_dependences): Likewise.
4153 (control_dependences::get_edges_dependent_on): Likewise.
4154 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
4155 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
4156 * context.c (gcc::context::context): Likewise.
4157 * cprop.c (pass_rtl_cprop::clone): Likewise.
4158 * domwalk.c (dom_walker::walk): Likewise.
4159 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
4160 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
4161 * mode-switching.c (pass_mode_switching::clone): Likewise.
4162 * passes.c (opt_pass::opt_pass): Likewise.
4163 (pass_manager::pass_manager): Likewise.
4164 * predict.c (pass_strip_predict_hints::clone): Likewise.
4165 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
4166 (pass_split_all_insns::clone): Likewise.
4167 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
4169 (bit_field_mode_iterator::next_mode): Likewise.
4170 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
4171 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
4172 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
4173 * tree-complex.c (pass_lower_complex::clone): Likewise.
4174 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
4175 * tree-object-size.c (pass_object_sizes::clone): Likewise.
4176 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
4177 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
4178 (pass_fold_builtins::clone): Likewise.
4179 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
4180 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
4181 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
4182 (pass_cd_dce::clone): Likewise.
4183 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
4184 (pass_phi_only_cprop::clone): Likewise.
4185 * tree-ssa-dse.c (pass_dse::clone): Likewise.
4186 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
4187 * tree-ssa-loop.c (pass_lim::clone): Likewise.
4188 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
4189 * tree-ssa-pre.c (pass_fre::clone): Likewise.
4190 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
4191 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
4192 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
4193 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
4194 * tree-vrp.c (pass_vrp::clone): Likewise.
4195 * tsan.c (pass_tsan::clone): Likewise.
4197 2013-09-30 Jakub Jelinek <jakub@redhat.com>
4200 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
4201 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
4204 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
4205 optimization, punt if sign_bit_p looked through any zero extension.
4207 2013-09-30 Teresa Johnson <tejohnson@google.com>
4209 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
4210 Update redirected out edge count in joiner case.
4211 (ssa_redirect_edges): Common the joiner and non-joiner cases
4212 so that joiner case gets profile updates.
4214 2013-09-30 Richard Biener <rguenther@suse.de>
4216 PR tree-optimization/58554
4217 * tree-loop-distribution.c (classify_partition): Require
4218 unconditionally executed stores for memcpy and memset recognition.
4219 (tree_loop_distribution): Calculate dominance info.
4221 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
4223 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
4224 (NO_PROFILE_COUNTERS): Likewise.
4225 (PROFILE_HOOK): Likewise.
4226 (FUNCTION_PROFILER): Likewise.
4227 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
4229 2013-09-30 Iain Sandoe <iain@codesourcery.com>
4231 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
4232 calls and defines in TARGET_MACHO conditional.
4233 (load_macho_picbase_di): Likewise.
4234 (reload_macho_picbase): Likewise.
4235 (reload_macho_picbase_si): Likewise.
4236 (reload_macho_picbase_di): Likewise.
4237 (nonlocal_goto_receiver): Likewise.
4239 2013-09-30 Nick Clifton <nickc@redhat.com>
4241 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
4242 that use the MSP430X ISA.
4243 (msp430_option_override): Scan -mmcu command line option for any
4244 MCU name that supports the MSP430X ISA.
4245 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
4246 -mmcu options which enable the MSP430X ISA.
4248 2013-09-30 Richard Biener <rguenther@suse.de>
4251 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
4252 before looking for setjmp-like calls.
4254 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4257 * config/darwin-protos.h (machopic_get_function_picbase): New.
4258 * config/darwin.c (machopic_get_function_picbase): New.
4259 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
4260 label for a new func. (load_macho_picbase_di): Likewise.
4261 (reload_macho_picbase): New expand.
4262 (reload_macho_picbase_si): New insn.
4263 (reload_macho_picbase_di): New insn.
4264 (nonlocal_goto_receiver): New define and split.
4265 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
4266 (unspecv enum): Add UNSPECV_NLGR.
4268 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4270 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
4271 that altivec registers are correctly sized on Darwin.
4273 2013-09-29 Iain Sandoe <iain@codesourcery.com>
4275 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
4276 darwin-driver.o): Use COMPILE and POSTCOMPILE.
4277 * config/x-darwin (host-darwin.o): Likewise.
4278 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
4279 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
4280 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
4282 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
4284 * doc/invoke.texi: Fix usage of @tie{} command.
4286 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
4288 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
4290 2013-09-28 Jan Hubicka <jh@suse.cz>
4292 * config/i386/x86-tune.def: Add documentation for each of the options;
4295 2013-09-28 Jan Hubicka <jh@suse.cz>
4297 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
4299 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
4300 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
4301 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
4302 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
4303 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
4305 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
4307 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
4308 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
4309 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
4310 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
4311 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
4312 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
4313 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
4314 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
4315 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
4316 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
4317 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
4318 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
4319 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
4320 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
4321 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
4322 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
4323 gimple.h, godump.c, graphite-clast-to-gimple.c,
4324 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
4325 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
4326 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
4327 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
4328 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
4329 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
4330 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
4331 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
4332 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
4333 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
4334 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
4335 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
4336 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
4337 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
4338 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
4339 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
4340 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
4341 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
4342 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
4343 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
4344 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
4345 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
4346 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
4347 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
4348 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
4349 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
4350 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
4351 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
4352 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
4353 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
4354 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
4355 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
4356 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
4357 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
4358 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4359 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
4360 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
4361 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
4362 whitespace before "(".
4364 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
4366 * expr.h (extract_bit_field): Remove packedp parameter.
4367 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
4368 from forward declaration.
4369 (store_split_bit_field): Remove packedp arg from calls to
4370 extract_fixed_bit_field.
4371 (extract_bit_field_1): Remove packedp parameter and packedp
4372 argument from recursive calls and calls to extract_fixed_bit_field.
4373 (extract_bit_field): Remove packedp parameter and corresponding
4374 arg to extract_bit_field_1.
4375 (extract_fixed_bit_field): Remove packedp parameter. Remove code
4377 (extract_split_bit_field): Remove packedp arg from call to
4378 extract_fixed_bit_field.
4379 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
4380 (copy_blkmode_from_reg): Likewise.
4381 (copy_blkmode_to_reg): Likewise.
4382 (read_complex_part): Likewise.
4383 (store_field): Likewise.
4384 (expand_expr_real_1): Likewise.
4385 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
4386 to extract_bit_field.
4387 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
4388 call to extract_bit_field.
4389 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
4390 call to extract_bit_field.
4391 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
4392 and special packedp behavior from -fstrict-volatile-bitfields
4395 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4397 * lra-eliminations.c (init_elim_table): Guard value_p.
4399 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
4401 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
4402 DFmode, DImode, and SFmode in the upper VSX registers based on the
4403 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
4404 if -mpower8-vector. Combine -mvsx-timode handling with the rest
4405 of the VSX register handling.
4407 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
4409 (zero_extendsidi2_lfiwzx): Add support for loading into the
4410 Altivec registers with -mpower8-vector. Use wu/wv constraints to
4411 only do VSX memory options on Altivec registers.
4412 (extendsidi2_lfiwax): Likewise.
4413 (extendsfdf2_fpr): Likewise.
4414 (mov<mode>_hardfloat, SF/SD modes): Likewise.
4415 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
4416 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
4417 (movdi_internal64): Likewise.
4419 2013-09-27 Xinliang David Li <davidxl@google.com>
4421 * opts.c (finish_options): Adjust parameters
4422 according to vect cost model.
4423 (common_handle_option): Set dynamic vect cost
4425 targhooks.c (default_add_stmt_cost): Compute stmt cost
4427 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
4428 Use helper function.
4429 * tree-vectorizer.h (unlimited_cost_model): New function.
4430 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
4431 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
4433 (vect_enhance_data_refs_alignment): Ditto.
4434 * flag-types.h: New enum.
4435 * common/config/i386/i386-common.c (ix86_option_init_struct):
4436 No need to initialize vect_cost_model flag.
4437 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
4440 2013-09-27 Diego Novillo <dnovillo@google.com>
4442 * gimple.h (enum ssa_mode): Remove.
4444 2013-09-27 Paulo Matos <pmatos@broadcom.com>
4446 * cfgloop.h (number_of_loops): Fix typo in check for null.
4448 2013-09-27 Jakub Jelinek <jakub@redhat.com>
4451 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
4452 are children of outermost saved_cfun's loop, and set it up to
4453 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
4454 if loop != loop0 and SESE region contains bbs that belong to loop0.
4456 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4458 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
4459 (binary_scale_code_p, get_base_term, get_index_term): New functions.
4460 (set_address_segment, set_address_base, set_address_index)
4461 (set_address_disp): Accept the argument unconditionally.
4462 (baseness): Remove must_be_base_p and must_be_index_p checks.
4463 (decompose_normal_address): Classify as much as possible in the
4466 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
4468 * cse.c (count_reg_usage): Handle INT_LIST.
4469 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
4470 * reginfo.c (reg_scan_mark_refs): Likewise.
4471 * reload1.c (eliminate_regs_1): Likewise.
4473 2013-09-27 Iain Sandoe <iain@codesourcery.com>
4476 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
4479 2013-09-27 Richard Biener <rguenther@suse.de>
4481 PR tree-optimization/58459
4482 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
4483 restriction not propagating into loops.
4485 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
4487 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
4488 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
4489 * doc/tree-ssa.texi (Walking use-def chains): Delete.
4491 2013-09-26 Richard Biener <rguenther@suse.de>
4493 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
4495 2013-09-26 Richard Biener <rguenther@suse.de>
4497 * alias.h (component_uses_parent_alias_set): Rename to ...
4498 (component_uses_parent_alias_set_from): ... this.
4499 * alias.c (component_uses_parent_alias_set): Rename to ...
4500 (component_uses_parent_alias_set_from): ... this and return
4502 (reference_alias_ptr_type_1): Use the result from
4503 component_uses_parent_alias_set_from instead of stripping
4504 components one at a time.
4505 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
4507 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4509 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
4510 Move prototypes to...
4511 * tree-ssa-ter.h: New File. Move prototypes here.
4512 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
4513 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
4514 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
4515 from is_replaceable_p.
4516 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
4517 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
4518 (process_replaceable): Use ter_is_replaceable_p.
4519 (find_replaceable_in_bb): Use ter_is_replaceable_p.
4520 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
4521 newly refactored ssa_is_replaceable_p.
4522 * cfgexpand.c: Include tree-outof-ssa.h.
4523 * ssaexpand.h: Delete.
4525 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
4527 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
4528 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
4530 (create_gimple_tmp): Delete.
4531 (get_expr_type, build_assign, build_type_cast): Move to...
4532 * gimple-builder.c: New File.
4533 (get_expr_type): Relocate from gimple.c.
4534 (build_assign, build_type_cast): Change to only create ssanames.
4535 * gimple.h: Move prototypes to...
4536 * gimple-builder.h: New File. Here.
4537 * tree-ssa.h: And here.
4538 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
4539 count_uses_and_derefs): Relocate from gimple.c.
4540 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
4541 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
4542 * tree-ssa-math-opts (execute_cse_reciprocals): Use
4543 gimple_replace_ssa_lhs.
4544 * asan.c: Include gimple-builder.h.
4545 * Makefile.in: Add gimple-builder.o.
4547 2013-09-26 Richard Biener <rguenther@suse.de>
4549 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
4550 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
4551 (loe_visit_block): Use gcc_checking_assert.
4552 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
4553 gimple_assign_ssa_name_copy_p.
4554 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
4557 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
4559 * config/rs6000/t-rs6000 (rs6000.o): Remove.
4560 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
4562 2013-09-26 Richard Biener <rguenther@suse.de>
4564 PR tree-optimization/58539
4565 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
4566 the fact that debug statements are not taking part in loop-closed
4569 2013-09-26 Nick Clifton <nickc@redhat.com>
4571 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
4572 time warning message.
4573 (msp430_print_operand_raw): Delete unused letter parameter.
4574 (TARGET_PRINT_OPERAND_ADDRESS): Define.
4575 (msp430_print_operand_address): New function.
4576 (msp430_print_operand): Move address printing code from here to
4578 * config/msp430/msp430.md (movsipsi2): Add comment in generated
4580 (zero_extendpsisi2): Likewise.
4581 (extendpsisi2): New pattern.
4582 (andneghi3): New pattern.
4584 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
4586 * config/aarch64/aarch64.opt (mlra): New option.
4587 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
4588 (TARGET_LRA_P): Define.
4590 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
4592 * expr.c (expand_assignment): Remove obsolete comment.
4594 2013-09-25 Jeff Law <law@redhat.com>
4596 * tree-flow.h (thread_through_all_blocks): Prototype moved into
4597 tree-ssa-threadupdate.h.
4598 (register_jump_thread): Similarly.
4599 * tree-ssa-threadupdate.h: New header file.
4600 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
4601 * tree-vrp.c: Likewise.
4602 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
4603 (thread_around_empty_blocks): Change type of path vector argument to
4604 an edge,type pair from just an edge. Initialize both elements when
4605 appending to a jump threading path. Tweak references to elements
4607 (thread_across_edge): Similarly. Release memory for the elements
4609 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
4610 (dump_jump_thread_path): New function broken out from
4611 register_jump_thread.
4612 (register_jump_thread): Use dump_jump_thread_path. Change type of
4613 path vector entries. Search the path for NULL edges and dump
4614 the path if one is found. Tweak the conversion of path to 3-edge
4615 form to use the block copy type information embedded in the path.
4617 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4619 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
4621 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
4622 Vladimir Makarov <vmakarov@redhat.com>
4624 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
4625 from the least significant bit.
4626 (strip_address_mutations): Add bitfield operations handling.
4627 (must_be_index_p): Add shifting and rotate operations handling.
4628 (set_address_base): Use must_be_base_p predicate.
4629 (set_address_index): Use must_be_index_p predicate.
4631 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
4632 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4633 Sergey Lega <sergey.s.lega@intel.com>
4634 Anna Tikhonova <anna.tikhonova@intel.com>
4635 Ilya Tocar <ilya.tocar@intel.com>
4636 Andrey Turetskiy <andrey.turetskiy@intel.com>
4637 Ilya Verbin <ilya.verbin@intel.com>
4638 Kirill Yukhin <kirill.yukhin@intel.com>
4639 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4641 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
4643 (ix86_expand_vector_move_misalign): Support new unaligned load and
4644 stores and use new names.
4645 (CODE_FOR_sse2_storedqu): Rename to ...
4646 (CODE_FOR_sse2_storedquv16qi): ... this.
4647 (CODE_FOR_sse2_loaddqu): Rename to ...
4648 (CODE_FOR_sse2_loaddquv16qi): ... this.
4649 (CODE_FOR_avx_loaddqu256): Rename to ...
4650 (CODE_FOR_avx_loaddquv32qi): ... this.
4651 (CODE_FOR_avx_storedqu256): Rename to ...
4652 (CODE_FOR_avx_storedquv32qi): ... this.
4653 * config/i386/i386.md (fpint_logic): New.
4654 * config/i386/sse.md (VMOVE): Extend for AVX512.
4658 (VI_UNALIGNED_LOADSTORE): Ditto.
4659 (sse2_avx_avx512f): Ditto.
4660 (sse2_avx2): Extend for AVX512.
4661 (sse4_1_avx2): Ditto.
4662 (avx2_avx512f): New.
4663 (sse): Extend for AVX512.
4666 (avxsizesuffix): Ditto.
4667 (sseintvecmode): Ditto.
4669 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
4670 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
4671 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4672 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
4673 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
4674 (<sse2_avx_avx512f>_storedqu<mode): ... this.
4675 (<sse>_movnt<mode>): Replace constraint "x" with "v".
4676 (STORENT_MODE): Extend for AVX512.
4677 (*absneg<mode>2): Replace constraint "x" with "v".
4678 (*mul<mode>3): Ditto.
4679 (*ieee_smin<mode>3): Ditto.
4680 (*ieee_smax<mode>3): Ditto.
4681 (avx_cmp<mode>3): Replace VF with VF_128_256.
4682 (*<sse>_maskcmp<mode>3_comm): Ditto.
4683 (<sse>_maskcmp<mode>3): Ditto.
4684 (<sse>_andnot<mode>3): Extend for AVX512.
4685 (<code><mode>3, anylogic): Replace VF with VF_128_256.
4686 (<code><mode>3, fpint_logic): New.
4687 (*<code><mode>3): Extend for AVX512.
4688 (avx512flogicsuff): New.
4689 (avx512f_<logic><mode>): Ditto.
4690 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
4692 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
4693 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
4694 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
4695 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
4696 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
4697 (xop_vpermil2<mode>3): Ditto.
4698 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
4699 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
4700 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
4701 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
4703 2013-09-25 Tom Tromey <tromey@redhat.com>
4705 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
4706 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
4707 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
4708 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
4709 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
4710 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
4711 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
4712 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
4713 (GRAPHITE_HTAB_H): Remove.
4715 2013-09-25 Tom Tromey <tromey@redhat.com>
4717 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
4719 2013-09-25 Tom Tromey <tromey@redhat.com>
4721 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
4723 2013-09-25 Tom Tromey <tromey@redhat.com>
4725 * config/i386/t-i386 (i386.o): Remove.
4726 (i386-c.o): Use COMPILE and POSTCOMPILE.
4728 2013-09-25 Tom Tromey <tromey@redhat.com>
4730 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
4732 2013-09-25 Tom Tromey <tromey@redhat.com>
4734 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
4735 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
4736 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
4737 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
4738 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
4739 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
4740 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
4741 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
4742 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
4743 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
4744 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
4745 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
4746 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
4747 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
4748 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
4749 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
4750 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
4751 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
4752 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
4753 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
4754 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
4755 (tree-ssa-pre.o, tree-ssa-sccvn.o)
4756 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
4757 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
4758 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
4759 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
4760 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
4761 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
4762 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
4763 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
4764 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
4765 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
4766 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
4767 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
4768 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
4769 (graphite-clast-to-gimple.o, graphite-dependences.o)
4770 (graphite-interchange.o, graphite-poly.o)
4771 (graphite-scop-detection.o, graphite-sese-to-poly.o)
4772 (graphite-optimize-isl.o, tree-vect-loop.o)
4773 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
4774 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
4775 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
4776 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
4777 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
4778 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
4779 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
4780 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
4781 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
4782 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
4783 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
4784 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
4785 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
4786 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
4787 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
4788 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
4789 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
4790 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
4791 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
4792 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
4793 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
4794 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
4795 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
4796 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
4797 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
4798 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
4799 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
4800 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
4801 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
4802 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
4803 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
4804 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
4805 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
4806 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
4807 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
4808 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
4809 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
4810 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
4811 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
4812 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
4813 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
4814 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
4815 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
4816 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
4817 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
4818 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
4819 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
4820 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
4821 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
4822 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
4823 (gcov-dump.o): Remove.
4824 (default-c.o): Use COMPILE and POSTCOMPILE.
4825 (CFLAGS-gcc.o): New variable.
4826 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
4828 2013-09-25 Tom Tromey <tromey@redhat.com>
4830 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
4831 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
4832 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
4833 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
4834 (c-family/c-pragma.o, c-family/c-pretty-print.o)
4835 (c-family/c-semantics.o, c-family/c-ada-spec.o)
4836 (c-family/array-notation-common.o, c-family/stub-objc.o)
4837 (c-family/c-ubsan.o): Remove.
4839 2013-09-25 Tom Tromey <tromey@redhat.com>
4841 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
4843 2013-09-25 Tom Tromey <tromey@redhat.com>
4845 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
4846 to add -DENABLE_SHARED_LIBGCC.
4847 (gcc.o): Don't use subshell.
4849 2013-09-25 Tom Tromey <tromey@redhat.com>
4851 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
4852 * configure.ac: Don't invoke AM_PROG_CC_C_O.
4853 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
4854 * configure, config.in: Rebuild.
4856 2013-09-25 Tom Tromey <tromey@redhat.com>
4858 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
4859 (COMPILE, POSTCOMPILE): New variables.
4860 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
4861 (DEPFILES): New variable.
4862 Include ".Po" files.
4863 * configure.ac: Add checks for dependency checking.
4864 * configure, aclocal.m4: Regenerate.
4866 2013-09-25 Tom Tromey <tromey@redhat.com>
4868 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
4869 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
4871 2013-09-25 Tom Tromey <tromey@redhat.com>
4873 * Makefile.in (generated_files): Add options.h,
4874 target-hooks-def.h, insn-opinit.h,
4875 common/common-target-hooks-def.h, pass-instances.def,
4876 c-family/c-target-hooks-def.h.
4878 2013-09-25 Jeff Law <law@redhat.com>
4880 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
4881 than foo[foo.length () - 1] to access last member in a vec.
4882 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
4884 2013-09-25 Richard Biener <rguenther@suse.de>
4887 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
4889 2013-09-25 Jan Hubicka <jh@suse.cz>
4891 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
4894 2013-09-25 Marek Polacek <polacek@redhat.com>
4897 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
4898 when determining the type name.
4900 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
4902 * config/sh/sh.md: Fix formatting.
4904 2013-09-24 Xinliang David Li <davidxl@google.com>
4906 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
4907 max peel iterations parameter.
4908 * param.def: New parameter.
4909 * doc/invoke.texi: Document New parameter.
4911 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
4913 * gimple-pretty-print.c: Various whitespace tweaks.
4914 * tree-core.h: Likewise.
4915 * tree-pretty-print.c: Likewise.
4916 * tree-ssa-alias.c: Likewise.
4917 * tree-ssa-copy.c: Likewise.
4918 * tree-ssanames.c: Likewise.
4919 * tree-ssanames.h: Likewise.
4920 * tree-vrp.c: Likewise.
4922 2013-09-24 Alan Modra <amodra@gmail.com>
4926 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
4927 for output operands that disallow regs. Don't use EXPAND_WRITE on
4930 2013-09-24 Richard Biener <rguenther@suse.de>
4933 * tree.c (reference_alias_ptr_type): Move ...
4934 * alias.c (reference_alias_ptr_type): ... here and implement
4935 in terms of the new reference_alias_ptr_type_1.
4936 (ref_all_alias_ptr_type_p): New helper.
4937 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
4938 use ref_all_alias_ptr_type_p.
4939 (get_deref_alias_set): Add flag_strict_aliasing check here.
4940 (reference_alias_ptr_type_1): New function, split out from ...
4941 (get_alias_set): ... here.
4942 (alias_ptr_types_compatible_p): New function.
4943 * alias.h (reference_alias_ptr_type): Declare.
4944 (alias_ptr_types_compatible_p): Likewise.
4945 * tree.h (reference_alias_ptr_type): Remove.
4946 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
4947 to compare MEM_REF alias types.
4949 2013-09-24 Richard Biener <rguenther@suse.de>
4951 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
4953 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
4955 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
4956 reload helper function arrays into a single array reg_addr.
4957 (reload_fpr_gpr): Likewise.
4958 (reload_gpr_vsx): Likewise.
4959 (reload_vsx_gpr): Likewise.
4960 (struct rs6000_reg_addr): Likewise.
4961 (reg_addr): Likewise.
4962 (rs6000_debug_reg_global): Change rs6000_vector_reload,
4963 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
4964 (rs6000_init_hard_regno_mode_ok): Likewise.
4965 (rs6000_secondary_reload_direct_move): Likewise.
4966 (rs6000_secondary_reload): Likewise.
4968 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
4969 constraints: wu, ww, and wy. Repurpose wv constraint added during
4970 power8 changes. Put wg constraint in alphabetical order.
4972 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
4973 for future work to add ISA 2.07 VSX single precision support.
4974 (-mvsx-scalar-double): Change default from -1 to 1, update
4975 documentation comment.
4976 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
4977 (-mupper-regs-df): New debug switch to control whether DF values
4978 can go in the traditional Altivec registers.
4979 (-mupper-regs-sf): New debug switch to control whether SF values
4980 can go in the traditional Altivec registers.
4982 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
4984 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
4985 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
4986 Add new constraints, wu/ww/wy. Repurpose wv constraint.
4987 (rs6000_debug_legitimate_address_p): Print if we are running
4988 before, during, or after reload.
4989 (rs6000_secondary_reload): Add a comment.
4990 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
4992 * config/rs6000/constraints.md (wa constraint): Sort w<x>
4993 constraints. Update documentation string.
4994 (wd constraint): Likewise.
4995 (wf constraint): Likewise.
4996 (wg constraint): Likewise.
4997 (wn constraint): Likewise.
4998 (ws constraint): Likewise.
4999 (wt constraint): Likewise.
5000 (wx constraint): Likewise.
5001 (wz constraint): Likewise.
5002 (wu constraint): New constraint for ISA 2.07 SFmode scalar
5004 (ww constraint): Likewise.
5005 (wy constraint): Likewise.
5006 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
5007 in the previous submissions.
5008 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
5010 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
5012 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
5013 (REG_BR_PROB): Say that the probability is stored in an int_list.
5014 * reg-notes.def: Update commentary to mention INT_LIST.
5015 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
5016 (INT_LIST): New rtx.
5017 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
5018 * rtlanal.c (int_reg_note_p): New function.
5019 (alloc_reg_note): Assert that the note does not have an int argument.
5020 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
5021 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
5022 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
5023 rather than an INSN_LIST. Handle INT_LIST.
5024 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
5025 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
5026 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
5027 Manipulate them as ints rather than rtxes.
5028 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
5029 * regmove.c (copy_src_to_dest): Likewise.
5030 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
5032 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
5033 into the cases that need it.
5034 * config/arm/arm.c (arm_unwind_emit): Likewise.
5036 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
5037 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
5038 * loop-doloop.c (add_test, doloop_modify): Likewise.
5039 * loop-unswitch.c (compare_and_jump_seq): Likewise.
5040 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
5041 * predict.c (combine_predictions_for_insn): Likewise.
5042 * print-rtl.c (print_rtx): Handle INT_LIST.
5043 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
5044 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
5045 * config/arm/arm.c (emit_unlikely_jump): Likewise.
5046 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
5047 (ix86_split_fp_branch, predict_jump): Likewise.
5048 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
5049 * config/sh/sh.c (expand_cbranchsi4): Likewise.
5050 * config/spu/spu.c (ea_load_store_inline): Likewise.
5052 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
5053 value of a REG_BR_PROB note.
5054 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
5055 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
5056 * emit-rtl.c (try_split): Likewise.
5057 * predict.c (br_prob_note_reliable_p): Likewise.
5058 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
5059 * reorg.c (mostly_true_jump): Likewise.
5060 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
5061 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
5062 * config/i386/i386.c (ix86_print_operand): Likewise.
5063 * config/ia64/ia64.c (ia64_print_operand): Likewise.
5064 * config/mmix/mmix.c (mmix_print_operand): Likewise.
5065 * config/rs6000/rs6000.c (output_cbranch): Likewise.
5066 * config/s390/s390.c (s390_expand_tbegin): Likewise.
5067 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
5068 * config/sparc/sparc.c (output_cbranch): Likewise.
5069 * config/spu/spu.c (get_branch_target): Likewise.
5070 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
5071 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
5073 2013-09-23 Jan Hubicka <jh@suse.cz>
5075 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
5077 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
5078 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
5080 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
5081 (gimple_fold_call): Dump inconsistent devirtualizations; add
5082 sanity check for type based devirtualizations.
5083 * ipa-prop.c: Include ipa-utils.h
5084 (ipa_intraprocedural_devirtualization): Add sanity check.
5085 (try_make_edge_direct_virtual_call): Likewise.
5087 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
5089 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
5090 assignment statements.
5092 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
5094 * gimple-pretty-print.c (dump_ssaname_info): New function.
5095 (dump_gimple_phi): Call it.
5096 (pp_gimple_stmt_1): Likewise.
5097 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
5098 (range_info_def): Declare.
5099 * tree-pretty-print.c (pp_double_int): New function.
5100 (dump_generic_node): Call it.
5101 * tree-pretty-print.h (pp_double_int): Declare.
5102 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
5103 * tree-ssanames.h (range_info_def): New structure.
5104 (value_range_type): Move definition here.
5105 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
5107 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
5109 (set_range_info): New function.
5110 (get_range_info): Likewise.
5111 (duplicate_ssa_name_range_info): Likewise.
5112 (duplicate_ssa_name_fn): Check pointer type and call
5113 duplicate_ssa_name_range_info.
5114 * tree-ssa-copy.c (fini_copy_prop): Likewise.
5115 * tree-vrp.c (value_range_type): Remove definition, now in
5117 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
5118 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
5119 (SSA_NAME_RANGE_INFO): New macro.
5121 2013-09-23 Richard Biener <rguenther@suse.de>
5123 PR tree-optimization/58464
5124 * tree-ssa-pre.c (phi_trans_lookup): Remove.
5125 (phi_trans_add): Change to add conditionally on being not present.
5126 (phi_translate_1): Remove recursion detection here.
5127 (phi_translate): Pre-seed the cache with NULL to catch
5128 recursion here in a more generic way.
5129 (bitmap_find_leader): Adjust comment.
5130 (get_representative_for): Dump value-numbers.
5131 (create_expression_by_pieces): Likewise.
5132 (insert_into_preds_of_block): Likewise.
5134 2013-09-23 Christian Bruel <christian.bruel@st.com>
5137 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
5138 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
5140 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
5143 2013-09-20 Renlin Li <renlin.li@arm.com>
5145 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5147 (aarch64_expand_epilogue): Likewise.
5148 (aarch64_legitimize_reload_address): Likewise.
5150 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
5152 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
5153 NULL_TREE before pushing them onto the vector. Likewise for labels.
5155 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
5157 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
5158 comparison operators when -fno-trapping-math is in effect.
5159 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
5160 comparison operators in TFmode and assert that unsupported operators
5162 (ia64_print_operand): Likewise.
5164 2013-09-21 Jan Hubicka <jh@suse.cz>
5166 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
5167 (sse_typeless_stores): Enable for core
5168 (sse_load0_by_pxor): Likewise.
5169 (four_jump_limit): Disable for core.
5170 (pad_returns): Likewise.
5171 (avoid_vector_decode): Likewise.
5172 (fuse_cmp_and_branch): Enable for cores.
5173 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
5175 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5178 * config/pa/pa.c (pa_option_override): Disable auto increment and
5179 decrement instructions until reload is completed.
5181 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
5182 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
5183 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
5185 2013-09-20 DJ Delorie <dj@redhat.com>
5186 Nick Clifton <nickc@redhat.com>
5188 * config/rl78/rl78.c: Various whitespace and comment tweaks.
5189 (need_to_save): Save bank 0 on interrupts.
5190 (characterize_address): Strip far address wrappers.
5191 (rl78_as_legitimate_address): Likewise.
5192 (transcode_memory_rtx): Likewise.
5193 (rl78_peep_movhi_p): Disable this peephole after devirt.
5194 (rl78_propogate_register_origins): Forget all origins when a
5196 * config/rl78/rl78-virt.md: Various whitespace tweaks.
5197 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
5199 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
5200 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
5201 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
5203 * config/rl78/constraints.md: Various whitespace and paren tweaks.
5205 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
5207 * config/pa/pa.md: In "scc" insn patterns, change output template to
5208 handle const0_rtx in reg_or_0_operand operands.
5210 2013-09-20 Martin Husemann <martin@NetBSD.org>
5213 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
5214 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
5216 2013-09-20 Richard Biener <rguenther@suse.de>
5219 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
5220 remembering SSA name version and block index.
5221 (new_scev_info_str): Adjust.
5222 (hash_scev_info): Likewise. Also hash the block index.
5223 (eq_scev_info): Adjust.
5224 (find_var_scev_info): Likewise.
5225 (struct instantiate_cache_entry): Remove.
5226 (struct instantiate_cache_type): Use a htab to map name, block
5228 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5229 (get_instantiated_value_entry): Likewise.
5230 (hash_idx_scev_info, eq_idx_scev_info): New functions.
5231 (instantiate_scev_name): Adjust.
5233 2013-09-20 Jeff Law <law@redhat.com>
5235 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
5237 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
5239 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
5240 Call aarch64_simd_expand_args to update op[argc].
5242 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5244 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
5247 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
5249 * gengtype.c (file_rules): Added rule for *.cc files.
5250 (get_output_file_with_visibility): Give fatal message when no
5253 2013-09-20 Renlin Li <renlin.li@arm.com>
5255 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
5256 (aarch64_expand_epilogue): Likewise.
5257 (aarch64_legitimize_reload_address): Likewise.
5259 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
5262 * expr.c (expand_assignment): Remove misalignp code path.
5264 2013-09-20 Marek Polacek <polacek@redhat.com>
5267 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
5268 TYPE_PRECISION. Add asserts.
5270 2013-09-20 Richard Biener <rguenther@suse.de>
5272 PR tree-optimization/58453
5273 * tree-loop-distribution.c (distribute_loop): Apply the cost
5274 model for -ftree-loop-distribute-patterns, too.
5276 2013-09-20 Richard Biener <rguenther@suse.de>
5279 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
5280 make type comparison less strict.
5282 2013-09-20 Alan Modra <amodra@gmail.com>
5284 * configure: Regenerate.
5285 * aclocal.m4: Regenerate.
5287 2013-09-20 Marek Polacek <polacek@redhat.com>
5290 * doc/extend.texi: Document that attribute used is meant to be used
5291 on variables with static storage duration.
5293 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5295 PR tree-optimization/58472
5296 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
5297 simd_lane_access set inv_p = false.
5298 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
5299 the simduid magic VAR_DECL.
5301 2013-09-19 Jan Hubicka <jh@suse.cz>
5303 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
5305 2013-09-17 Jeff Law <law@redhat.com>
5307 * tree-ssa-dom.c (record_temporary_equivalences): New function
5308 split out of dom_opt_dom_walker::after_dom_children.
5309 (dom_opt_dom_walker::thread_across_edge): Move common code
5310 in here from dom_opt_dom_walker::after_dom_children.
5311 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
5313 2013-09-19 Jan Hubicka <jh@suse.cz>
5315 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
5316 (TARGET_GENERIC): Use PROCESOR_GENERIC
5317 (enum processor_type): Unify generic32 and 64.
5318 * i386.md (cpu): Likewise.
5319 * x86-tune.def (use_leave): Enable for generic32.
5320 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
5321 * athlon.md: Change generic64 to generic in all occurences.
5322 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
5323 (ix86_target_macros_internal): Likewise.
5324 * driver-i386.c (host_detect_local_cpu): Likewise.
5325 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
5327 (generic_memcpy, generic_memset, generic_cost): This one.
5328 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
5329 (m_GENERIC32, m_GENERIC64): Remove.
5330 (m_GENERIC): Turn into one flag.
5331 (processor_target): Unify generic tunnings.
5332 (ix86_option_override_internal): Replace generic32/64 by generic.
5333 (ix86_issue_rate): Likewise.
5334 (ix86_adjust_cost): Likewise.
5336 2013-09-19 Jan Hubicka <jh@suse.cz>
5338 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
5339 of speculative flag.
5341 2013-09-19 Jakub Jelinek <jakub@redhat.com>
5343 * omp-low.c (expand_omp_sections): Always pass len - 1 to
5344 GOMP_sections_start, even if !exit_reachable.
5346 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5348 * lra-constraints.c (need_for_all_save_p): Use macro
5349 HARD_REGNO_CALL_PART_CLOBBERED.
5350 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
5351 set up pseudo conflict hard regs.
5353 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
5356 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
5359 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
5361 PR rtl-optimization/58438
5362 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
5363 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
5364 from previous subpasses.
5366 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
5368 * arm.c (arm_get_frame_offsets): Validate architecture supports
5369 LDRD/STRD before accepting the tuning preference.
5370 (arm_expand_prologue): Likewise.
5371 (arm_expand_epilogue): Likewise.
5373 2013-09-18 Richard Biener <rguenther@suse.de>
5375 PR tree-optimization/58417
5376 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
5377 have chrecs with symbols defined in the loop as operands.
5378 (chrec_fold_multiply): Likewise.
5379 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
5380 parameters before folding binary operations.
5381 (struct instantiate_cache_entry_hasher): Remove.
5382 (struct instantiate_cache_type): Use a pointer-map.
5383 (instantiate_cache_type::instantiate_cache_type): New function.
5384 (instantiate_cache_type::get): Likewise.
5385 (instantiate_cache_type::set): Likewise.
5386 (instantiate_cache_type::~instantiate_cache_type): Adjust.
5387 (get_instantiated_value_entry): Likewise.
5388 (global_cache): New global.
5389 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
5390 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
5391 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
5392 (instantiate_scev_name): Adjust.
5393 (instantiate_scev): Construct global instead of local cache.
5394 (resolve_mixers): Likewise.
5396 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
5397 Paolo Carlini <paolo.carlini@oracle.com>
5400 * doc/implement-cxx.texi: Fix references to the C++ standards.
5402 2013-09-18 Jakub Jelinek <jakub@redhat.com>
5404 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
5405 * tree-vect-data-refs.c (vect_analyze_data_refs): For
5406 simd_lane_access drs, update also DR_ALIGNED_TO.
5408 2013-09-18 Marek Polacek <polacek@redhat.com>
5411 * doc/extend.texi: Document no_sanitize_undefined attribute.
5412 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
5413 no_sanitize_undefined attribute.
5415 2013-09-18 Nick Clifton <nickc@redhat.com>
5417 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
5418 (ASM_DECLARE_FUNCTION_NAME): Define.
5420 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
5422 * compare-elim.c (find_comparison_dom_walker): New class
5423 (find_comparisons_in_bb): Rename to
5424 find_comparison_dom_walker::before_dom_children
5425 (find_comparisons): Adjust
5426 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
5428 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
5429 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
5430 (init_walk_dominator_tree): Remove declaration.
5431 (fini_walk_dominator_tree): Remove declaration.
5432 * fwprop.c (single_def_use_dom_walker): New class
5433 (single_def_use_enter_block): Convert to
5434 single_def_use_dom_walker::before_dom_children.
5435 (single_def_use_leave_block): Convert to
5436 single_def_use_dom_walker::after_dom_children.
5437 (build_single_def_use_links): Adjust.
5438 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
5440 (find_candidates_in_block): Convert to
5441 find_candidates_dom_walker::before_dom_children.
5442 (execute_strength_reduction): Adjust.
5443 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
5444 (sese_dom_walker): New class.
5445 (sese_dom_walker::sese_dom_walker): New constructor.
5446 (sese_dom_walker::~sese_dom_walker): New destructor.
5447 (build_sese_conditions_before): Convert to
5448 sese_dom_walker::before_dom_children.
5449 (build_sese_conditions_after): Convert to
5450 sese_dom_walker::after_dom_children.
5451 (build_poly_scop): Adjust
5452 * tree-into-ssa.c (rewrite_dom_walker): New class
5453 (rewrite_enter_block): Convert to
5454 rewrite_dom_walker::before_dom_children.
5455 (rewrite_leave_block): Convert to
5456 rewrite_dom_walker::after_dom_children.
5457 (rewrite_update_dom_walker): New class.
5458 (rewrite_update_enter_block): Convert to
5459 rewrite_update_dom_walker::before_dom_children.
5460 (rewrite_update_leave_block): Convert to
5461 rewrite_update_dom_walker::after_dom_children.
5462 (rewrite_blocks, rewrite_into_ssa): Adjust.
5463 (mark_def_dom_walker): New class.
5464 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
5465 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
5466 (mark_def_sites_blocks): Convert to
5467 mark_def_dom_walker::before_dom_children.
5468 (mark_def_site_blocks): Remove.
5469 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
5470 (tree_ssa_dominator_optimize): Adjust.
5471 (dom_thread_across_edge): Convert to method
5472 dom_opt_dom_walker::thread_across_edge.
5473 (dom_opt_enter_block): Convert to member function
5474 dom_opt_dom_walker::before_dom_children.
5475 (dom_opt_leave_block): Convert to member function
5476 dom_opt_dom_walker::after_dom_children.
5477 * tree-ssa-dse.c (dse_dom_walker): New class.
5478 (dse_enter_block): Convert to member function
5479 dse_dom_walker::before_dom_children.
5480 (tree_ssa_dse): Adjust.
5481 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
5482 (determine_invariantness_stmt): Convert to method
5483 invariantness_dom_walker::before_dom_children.
5484 (determine_invariantness): Remove
5485 (move_computations_dom_walker): New class.
5486 (move_computations_stmt): Convert to method
5487 move_computations_dom_walker::before_dom_children.
5488 (move_computations, tree_ssa_lim): Adjust.
5489 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
5490 (nt_init_block): Convert to method
5491 notrappping_dom_walker::before_dom_children.
5492 (nt_fini_block): Convert to method
5493 method nontrapping_dom_walker::after_dom_children.
5494 (get_non_trapping): Adjust.
5495 * tree-ssa-pre.c (eliminate_dom_walker): New class.
5496 (eliminate_bb): Convert to method
5497 eliminate_dom_walker::before_dom_children.
5498 (eliminate_leave_block): Convert to method
5499 eliminate_dom_walker::after_dom_children.
5500 (eliminate): Adjust.
5501 * tree-ssa-strlen.c (strlen_dom_walker): New class.
5502 (strlen_enter_block): Convert to method
5503 strlen_dom_walker::before_dom_children.
5504 (strlen_leave_block): Convert to method
5505 method strlen_dom_walker::after_dom_children.
5506 (tree_ssa_strlen): Adjust.
5507 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
5508 (tree_ssa_uncprop): Adjust.
5509 (uncprop_leave_block): Convert to method
5510 uncprop_dom_walker::after_dom_children.
5511 (uncprop_leave_block): Convert to method
5512 uncprop_dom_walker::before_dom_children.
5514 2013-09-18 Bin Cheng <bin.cheng@arm.com>
5516 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
5517 branch in same basic block. Check both src and dest of the move insn.
5519 2013-09-17 Nick Clifton <nickc@redhat.com>
5521 * config/rl78/rl78-real.md (bf): New pattern.
5523 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
5524 (rl78_print_operand): Do not put a # before a %B.
5525 * config/rl78/rl78.opt: Tweak doc strings.
5527 2013-09-17 DJ Delorie <dj@redhat.com>
5529 * config/rl78/constraints.md (Wcv): Allow up to $r31.
5530 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
5531 (rl78_option_override): Likewise, if -mallregs.
5532 (is_virtual_register): Likewise.
5533 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
5534 (REGNO_OK_FOR_BASE_P): Likewise.
5535 * config/rl78/rl78.opt (-mallregs): New.
5537 2013-09-17 Nick Clifton <nickc@redhat.com>
5539 * config/rl78/rl78.c (need_to_save): Change return type to bool.
5540 For interrupt functions: save all call clobbered registers if the
5541 interrupt handler is not a leaf function.
5542 (rl78_expand_prologue): Always recompute the frame information.
5543 For interrupt functions: only select bank 0 if one of the bank 0
5544 registers is going to be psuhed.
5546 2013-09-17 DJ Delorie <dj@redhat.com>
5548 * config/rl78/constraints.md: For each W* constraint, rename to C*
5549 and create a W* constraint that checks for an optional ES: prefix
5551 * config/rl78/rl78.md (UNS_ES_ADDR): New.
5552 (es_addr): New. Used to wrap far addresses.
5553 * config/rl78/rl78-protos.h (rl78_es_addr): New.
5554 (rl78_es_base): New.
5555 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
5556 wrapped far addresses.
5557 (rl78_print_operand_1): Unwrap far addresses before processing.
5558 (rl78_lo16): Wrap far addresses in unspecs.
5559 (rl78_es_addr): New.
5560 (rl78_es_base): New.
5561 (insn_ok_now): Check for not-yet-wrapped far addresses.
5562 (transcode_memory_rtx): Properly re-wrap far addresses.
5564 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
5566 * config/sparc/t-rtems: Add leon3 multilibs.
5568 2013-09-17 Cong Hou <congh@google.com>
5570 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
5571 when checking the dot production pattern. The type of rhs operand
5572 of multiply is now checked correctly.
5574 2013-09-17 Jeff Law <law@redhat.com>
5576 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5577 edge implied equivalences into successor phis.
5578 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
5579 here from tree-ssa-threadedge.c.
5580 (mark_threaded_blocks): When threading through a joiner, if both
5581 successors of the joiner's clone reach the same block, verify the
5582 PHI arguments are equal. If not, cancel the jump threading request.
5583 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
5584 tree-ssa-threadupdate.c
5585 (thread_across_edge): Don't check PHI argument equality when
5586 threading through joiner block here.
5588 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
5590 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
5591 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
5592 (warn_uninit, warn_uninitialized_vars,
5593 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5594 Move to tree-ssa-uninit.c.
5595 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
5596 (has_undefined_value_p): New. Pass dependant parts of
5597 ssa_undefined_value_p.
5598 (uninit_undefined_value_p): Use has_undefined_value_p.
5599 (warn_uninit, warn_uninitialized_vars,
5600 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
5601 Move from tree-ssa.c.
5602 * tree-ssa.h: Adjust prototypes.
5604 2013-09-17 Jan Hubicka <jh@suse.cz>
5607 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
5608 * ipa-inline.c (can_inline_edge_p): Do not downgrade
5609 FUNCTION_NOT_OPTIMIZED.
5610 * ipa-inline-analysis.c (compute_inline_parameters): Function
5611 not optimized is not inlinable unless it is alwaysinline.
5612 (inline_analyze_function): Force calls in not optimized
5613 function not inlinable.
5615 2013-09-17 Jan Hubicka <jh@suse.cz>
5618 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
5620 * ipa.c (function_and_variable_visibility): Likewise.
5621 * ipa-profile.c (ipa_profile): Likewise.
5623 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
5626 * cgraph.c (cgraph_function_body_availability): Check for ifunc
5627 attribute, and don't inline the resolver in this case.
5629 2013-09-17 Teresa Johnson <tejohnson@google.com>
5631 * coverage.c (get_coverage_counts): Add missing newline.
5633 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5635 PR tree-optimization/58088
5636 * fold-const.c (mask_with_trailing_zeros): New function.
5637 (fold_binary_loc): Make sure we don't recurse infinitely
5638 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
5639 Use mask_with_trailing_zeros where appropriate.
5641 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
5643 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
5644 of 'prev' var to get better distance estimation.
5646 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
5648 * tree-inline.h (struct copy_body_data): Add transform_parameter.
5649 * tree-inline.c (is_parameter_of): New predicate.
5650 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
5651 a parameter has been remapped.
5652 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
5653 (optimize_inline_calls): Initialize transform_parameter.
5654 (copy_gimple_seq_and_replace_locals): Likewise.
5655 (tree_function_versioning): Likewise.
5656 (maybe_inline_call_in_expr): Likewise.
5658 2013-09-17 Nick Clifton <nickc@redhat.com>
5660 * config/msp430/msp430-protos.h: Add prototypes for new functions.
5661 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
5663 (is_attr_func): New function.
5664 (msp430_is_interrupt_func): New function.
5665 (is_naked_func): New function.
5666 (is_reentrant_func): New function.
5667 (is_critical_func): New function.
5668 (msp430_start_function): Add annotations for function attributes.
5669 (msp430_attr): New function.
5670 (msp430_attribute_table): New.
5671 (msp430_function_section): New function.
5672 (TARGET_ASM_FUNCTION_SECTION): Define.
5673 (msp430_builtin): New enum.
5674 (msp430_init_builtins): New function.
5675 (msp430_builtin_devl): New function.
5676 (msp430_expand_builtin): New function.
5677 (TARGET_INIT_BUILTINS): Define.
5678 (TARGET_EXPAND_BUILTINS): Define.
5679 (TARGET_BUILTIN_DECL): Define.
5680 (msp430_expand_prologue): Add support for naked, interrupt,
5681 critical and reentrant functions.
5682 (msp430_expand_epilogue): Likewise.
5683 (msp430_print_operand): Handle 'O' character.
5684 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
5686 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
5687 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
5688 (pushm): Use a 'n' rather than an 'i' constraint.
5689 (msp_return): Add generation of the interrupt return instruction.
5690 (disable_interrupts): New pattern.
5691 (enable_interrupts): New pattern.
5692 (push_intr_state): New pattern.
5693 (pop_intr_state): New pattern.
5694 (bic_SR): New pattern.
5695 (bis_SR): New pattern.
5696 * doc/extend.texi: Document MSP430 function attributes and builtin
5699 2013-09-17 Richard Biener <rguenther@suse.de>
5701 PR tree-optimization/58432
5702 * tree-loop-distribution.c (tree_loop_distribution): Also
5703 scan PHIs for outside loop uses and seed a partition from them.
5705 2013-09-17 Bin Cheng <bin.cheng@arm.com>
5707 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
5708 (restructure_reference): Call backtrace_base_for_ref.
5710 2013-09-17 Alan Modra <amodra@gmail.com>
5713 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
5716 2013-09-16 DJ Delorie <dj@redhat.com>
5718 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
5719 vregs location for RL78/G10.
5720 (rl78_expand_prologue): Avoid SEL on G10.
5721 (rl78_expand_epilogue): Likewise.
5722 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
5723 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
5724 __RL78_G10__ when appropriate.
5725 (ASM_SPEC): Pass -mg10 along to the assembler.
5726 * config/rl78/rl78.md (sel_rb): Disable for G10.
5727 * config/rl78/rl78.opt: Add -mg10 option.
5728 * config/rl78/t-rl78: Add -mg10 multilib.
5730 2013-09-16 Xinliang David Li <davidxl@google.com>
5732 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
5733 * omp-low.c (omp_max_vf): Ditto.
5734 (expand_omp_simd): Ditto.
5735 * tree-vectorizer.c (vectorize_loops): Ditto.
5736 (gate_vect_slp): Ditto.
5737 (gate_increase_alignment): Ditto.
5738 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
5739 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
5740 (gate_tree_vectorize): Name change.
5741 (tree_vectorize): Ditto.
5742 (pass_vectorize::gate): Call new function.
5743 (pass_vectorize::execute): Ditto.
5744 * opts.c: O3 default setting change.
5745 (finish_options): Check new flag.
5746 * doc/invoke.texi: Document new flags.
5747 * common.opt: New flags.
5749 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
5751 * doc/tm.texi.in (Cond Exec Macros): Remove node.
5752 (Condition Code): Don't reference it.
5753 * doc/tm.texi: Regenerate.
5755 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
5758 * lra-constraints.c (undo_optional_reloads): Consider all optional
5759 reload even if it did not get a hard reg.
5761 2013-09-16 Teresa Johnson <tejohnson@google.com>
5763 * dumpfile.c (dump_loc): Remove newline emission.
5764 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
5765 emission to dump_printf_loc calls where missing.
5766 (vect_mark_for_runtime_alias_test): Ditto.
5767 (vect_analyze_data_ref_dependence): Ditto.
5768 (vect_analyze_data_ref_dependences): Ditto.
5769 (vect_slp_analyze_data_ref_dependence): Ditto.
5770 (vect_slp_analyze_data_ref_dependences): Ditto.
5771 (vect_compute_data_ref_alignment): Ditto.
5772 (vect_update_misalignment_for_peel): Ditto.
5773 (vect_verify_datarefs_alignment): Ditto.
5774 (vector_alignment_reachable_p): Ditto.
5775 (vect_get_data_access_cost): Ditto.
5776 (vect_enhance_data_refs_alignment): Ditto.
5777 (vect_find_same_alignment_drs): Ditto.
5778 (vect_analyze_data_refs_alignment): Ditto.
5779 (vect_analyze_group_access): Ditto.
5780 (vect_analyze_data_ref_access): Ditto.
5781 (vect_analyze_data_ref_accesses): Ditto.
5782 (vect_prune_runtime_alias_test_list): Ditto.
5783 (vect_analyze_data_refs): Ditto.
5784 (vect_create_addr_base_for_vector_ref): Ditto.
5785 (vect_create_data_ref_ptr): Ditto.
5786 (vect_grouped_store_supported): Ditto.
5787 (vect_grouped_load_supported): Ditto.
5788 * value-prof.c (check_counter): Ditto.
5789 (check_ic_target): Ditto.
5790 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
5791 (vect_recog_widen_mult_pattern): Ditto.
5792 (vect_recog_widen_sum_pattern): Ditto.
5793 (vect_recog_over_widening_pattern): Ditto.
5794 (vect_recog_widen_shift_pattern): Ditto.
5795 (vect_recog_rotate_pattern): Ditto.
5796 (vect_recog_vector_vector_shift_pattern): Ditto.
5797 (vect_recog_divmod_pattern): Ditto.
5798 (vect_recog_mixed_size_cond_pattern): Ditto.
5799 (vect_recog_bool_pattern): Ditto.
5800 (vect_pattern_recog_1): Ditto.
5801 (vect_pattern_recog): Ditto.
5802 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
5803 (vect_is_simple_iv_evolution): Ditto.
5804 (vect_analyze_scalar_cycles_1): Ditto.
5805 (vect_get_loop_niters): Ditto.
5806 (vect_analyze_loop_1): Ditto.
5807 (vect_analyze_loop_form): Ditto.
5808 (vect_analyze_loop_operations): Ditto.
5809 (vect_analyze_loop_2): Ditto.
5810 (vect_analyze_loop): Ditto.
5811 (report_vect_op): Ditto.
5812 (vect_is_slp_reduction): Ditto.
5813 (vect_is_simple_reduction_1): Ditto.
5814 (vect_get_known_peeling_cost): Ditto.
5815 (vect_estimate_min_profitable_iters): Ditto.
5816 (vect_model_reduction_cost): Ditto.
5817 (vect_model_induction_cost): Ditto.
5818 (get_initial_def_for_induction): Ditto.
5819 (vect_create_epilog_for_reduction): Ditto.
5820 (vectorizable_reduction): Ditto.
5821 (vectorizable_induction): Ditto.
5822 (vectorizable_live_operation): Ditto.
5823 (vect_loop_kill_debug_uses): Ditto.
5824 (vect_transform_loop): Ditto.
5825 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
5826 (vect_stmt_relevant_p): Ditto.
5827 (process_use): Ditto.
5828 (vect_mark_stmts_to_be_vectorized): Ditto.
5829 (vect_model_simple_cost): Ditto.
5830 (vect_model_promotion_demotion_cost): Ditto.
5831 (vect_model_store_cost): Ditto.
5832 (vect_get_store_cost): Ditto.
5833 (vect_model_load_cost): Ditto.
5834 (vect_get_load_cost): Ditto.
5835 (vect_init_vector_1): Ditto.
5836 (vect_get_vec_def_for_operand): Ditto.
5837 (vect_finish_stmt_generation): Ditto.
5838 (vectorizable_call): Ditto.
5839 (vectorizable_conversion): Ditto.
5840 (vectorizable_assignment): Ditto.
5841 (vectorizable_shift): Ditto.
5842 (vectorizable_operation): Ditto.
5843 (vectorizable_store): Ditto.
5844 (vectorizable_load): Ditto.
5845 (vectorizable_condition): Ditto.
5846 (vect_analyze_stmt): Ditto.
5847 (vect_transform_stmt): Ditto.
5848 (vect_is_simple_use): Ditto.
5849 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
5850 (vect_can_advance_ivs_p): Ditto.
5851 (vect_update_ivs_after_vectorizer): Ditto.
5852 (vect_do_peeling_for_loop_bound): Ditto.
5853 (vect_gen_niters_for_prolog_loop): Ditto.
5854 (vect_update_inits_of_drs): Ditto.
5855 (vect_create_cond_for_alias_checks): Ditto.
5856 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
5857 (vect_build_slp_tree_1): Ditto.
5858 (vect_supported_load_permutation_p): Ditto.
5859 (vect_analyze_slp_instance): Ditto.
5860 (vect_analyze_slp): Ditto.
5861 (vect_make_slp_decision): Ditto.
5862 (vect_detect_hybrid_slp): Ditto.
5863 (vect_bb_vectorization_profitable_p): Ditto.
5864 (vect_slp_analyze_bb_1): Ditto.
5865 (vect_update_slp_costs_according_to_vf): Ditto.
5866 (vect_get_mask_element): Ditto.
5867 (vect_transform_slp_perm_load): Ditto.
5868 (vect_schedule_slp_instance): Ditto.
5869 (vect_schedule_slp): Ditto.
5870 (vect_slp_transform_bb): Ditto.
5871 * profile.c (read_profile_edge_counts): Ditto.
5872 (compute_branch_probabilities): Ditto.
5873 * coverage.c (get_coverage_counts): Ditto.
5875 2013-09-16 Diego Novillo <dnovillo@google.com>
5877 * tree-core.h: Add missing comment lines from refactoring of tree.h.
5879 2013-09-16 Jan Hubicka <jh@suse.cz>
5881 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
5882 abstract functions; for static functions check the presence of body.
5884 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5886 * config/aarch64/aarch64-simd-builtins.def (fma): New.
5887 * config/aarch64/aarch64-simd.md
5888 (aarch64_mla_elt<mode>): New.
5889 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
5890 (aarch64_mls_elt<mode>): Likewise.
5891 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
5892 (aarch64_fma4_elt<mode>): Likewise.
5893 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
5894 (aarch64_fma4_elt_to_128v2df): Likewise.
5895 (aarch64_fma4_elt_to_64df): Likewise.
5896 (fnma<mode>4): Likewise.
5897 (aarch64_fnma4_elt<mode>): Likewise.
5898 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
5899 (aarch64_fnma4_elt_to_128v2df): Likewise.
5900 (aarch64_fnma4_elt_to_64df): Likewise.
5901 * config/aarch64/iterators.md (VDQSF): New.
5902 * config/aarch64/arm_neon.h
5903 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
5904 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
5906 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5908 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
5909 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
5910 (aarch64_mul3_elt_to_128df): Likewise.
5911 (aarch64_mul3_elt_to_64v2df): Likewise.
5912 * config/aarch64/iterators.md (VEL): Also handle DFmode.
5914 (VMUL_CHANGE_NLANES) Likewise.
5917 * config/aarch64/arm_neon.h
5918 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
5920 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
5922 * config/aarch64/arm_neon.h
5923 (vcvtx_high_f32_f64): Fix parameters.
5925 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5926 Uros Bizjak <ubizjak@gmail.com>
5928 * config/alpha.c: Include tree-ssanames.h.
5930 2013-09-16 Richard Biener <rguenther@suse.de>
5932 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
5933 (dot_rdg_1): Handle control_dd.
5934 (create_edge_for_control_dependence): New function.
5935 (create_rdg_edges): Add control dependences if asked for.
5936 (build_rdg): Likewise.
5937 (generate_loops_for_partition): If there are not necessary
5938 control stmts remove all their dependencies.
5939 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
5940 (distribute_loop): Pass on control dependences.
5941 (tree_loop_distribution): Compute control dependences and remove
5942 restriction on number of loop nodes.
5944 2013-09-16 Jakub Jelinek <jakub@redhat.com>
5946 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
5949 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
5951 * cse.c (try_const_anchors): Punt on CC modes.
5953 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5955 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
5957 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
5960 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
5963 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5966 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
5968 (darwin_objc1_section): Likewise.
5969 (darwin_file_end): Emit Image Info section when required.
5971 2013-09-14 Jan Hubicka <jh@suse.cz>
5973 * tree-into-ssa.c (gate_into_ssa): New.
5974 (pass_data_build_ssa): Use it.
5975 * cgraph.h (expand_thunk): Update prototype.
5976 * cgraphunit.c (analyze_function): Expand thunks early.
5977 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
5978 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
5979 set lowered flag; do not add new function.
5980 (assemble_thunks_and_aliases): Update.
5981 * tree-ssa.c (gate_init_datastructures): New gate.
5982 (pass_data_init_datastructures): Use it.
5984 2013-09-14 Iain Sandoe <iain@codesourcery.com>
5987 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
5988 xmm register set described in the psABI.
5990 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
5992 * dwarf2out.c (should_emit_struct_debug): Add check
5993 for type_decl variable is not NULL.
5995 2013-09-13 Jacek Caban <jacek@codeweavers.com>
5997 * config.gcc: Use new winnt-c.c target hooks
5998 * config/t-winnt: New file
5999 * config/winnt-c.c: New file
6000 * doc/tm.texi.in: Document new hook
6001 * doc/tm.texi: Regenerated
6003 2013-09-13 Jan Hubicka <jh@suse.cz>
6006 * ipa-inline.c (check_callers): New function.
6007 (check_caller_edge): Remove.
6008 (want_inline_function_to_all_callers_p): Also permit alises that are
6010 (inline_to_all_callers): Terminate the walk when devirtualization
6011 introduce new calls.
6013 2013-09-13 Jan Hubicka <jh@suse.cz>
6015 * ipa-inline-analysis.c (struct growth_data): Add node.
6016 (do_estimate_growth_1): Fix detection of recursion.
6018 2013-09-13 Jakub Jelinek <jakub@redhat.com>
6020 PR tree-optimization/58392
6021 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
6022 to avoid shadowing of outer loop variable. If
6023 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
6024 replace_by_duplicate_decl simduid of loops that have it set and
6025 set dest_cfun->has_simduid_loops and/or
6026 dest_cfun->has_force_vect_loops.
6027 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
6028 instead of maybe_lookup_decl.
6029 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
6030 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
6031 call. Copy over force_vect and copy and remap simduid. Set
6032 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
6033 (copy_cfg_body): Remove blocks_to_copy argument. Use
6034 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
6035 caller. Don't set cfun->has_simduid_loops and/or
6036 cfun->has_force_vect_loops here.
6037 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
6039 (expand_call_inline, tree_function_versioning): Adjust copy_body
6042 2013-09-13 Martin Jambor <mjambor@suse.cz>
6045 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
6046 the assert if the edge was a speculative one.
6048 2013-09-13 Richard Biener <rguenther@suse.de>
6050 * tree-data-ref.h (known_dependences_p): Move here ...
6051 * tree-loop-distribution.c (known_dependences_p): ... from here.
6052 (dump_rdg_component, debug_rdg_component): Remove.
6054 (generate_loops_for_partition): Use gimple_uid instead of
6055 relying on matching stmt visit order.
6056 (rdg_build_partitions): Take starting stmt vector.
6057 (ldist_gen): Merge into ...
6058 (distribute_loop): ... this function. Do not compute starting
6060 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
6062 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6064 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
6065 Set type attribute correctly. Set predicable_short_it attribute.
6066 (cmpsi_shiftsi): Remove %? from output template.
6068 2013-09-13 Richard Biener <rguenther@suse.de>
6070 * tree-loop-distribution.c (struct rdg_component,
6071 rdg_defs_used_in_other_loops_p, free_rdg_components,
6072 rdg_build_components): Remove.
6073 (stmts_from_loop): Do not record virtual PHIs.
6074 (generate_loops_for_partition): Skip virtual PHIs.
6075 (build_rdg_partition_for_component): Rename to ...
6076 (build_rdg_partition_for_vertex): ... this and adjust.
6077 (rdg_build_partitions): Take a vector of starting vertices
6078 instead of components. Remove unnecessary leftover handling.
6079 (ldist_gen): Do not build components or record other stores.
6080 (distribute_loop): Do not distribute loops containing stmts
6083 2013-09-13 Christian Bruel <christian.bruel@st.com>
6086 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
6088 2013-09-13 Kai Tietz <ktietz@redhat.com>
6090 * config.gcc: Separate cases for mingw and cygwin targets,
6091 and add 64-bit cygwin target case.
6093 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
6094 dll-export inline-functions.
6095 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
6098 2013-09-13 Jeff Law <law@redhat.com>
6102 2013-09-06 Jeff Law <law@redhat.com>
6104 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6105 edge implied equivalences into successor phis.
6107 2013-09-12 DJ Delorie <dj@redhat.com>
6109 * config/rl78/rl78-virt.md: Change from | to \; for asm line
6112 2013-09-12 Brooks Moses <bmoses@google.com>
6115 * Makefile.in: Do not install driver binaries in $(target)/bin.
6117 2013-09-12 DJ Delorie <dj@redhat.com>
6119 * config/rl78/rl78.opt (mrelax): New.
6120 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
6121 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
6123 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
6125 (rl78_expand_epilogue): Likewise.
6127 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
6130 * lra-eliminations.c (remove_reg_equal_offset_note): New.
6131 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
6132 pointer elimination with using remove_reg_equal_offset_note.
6134 2013-09-12 DJ Delorie <dj@redhat.com>
6136 * config/msp430/: New port.
6137 * config.gcc (msp430): Added.
6138 * doc/invoke.texi: Document MSP430 options.
6139 * doc/install.texi: Document msp430-elf
6140 * doc/md.texi: Document msp430-elf
6141 * doc/contrib.texi: Document msp430-elf
6143 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
6146 2013-09-12 Martin Jambor <mjambor@suse.cz>
6149 * ipa-prop.c (remove_described_reference): Give up if the edge in the
6150 reference descriptor is NULL.
6151 (ipa_edge_removal_hook): If owning a reference descriptor, set its
6154 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6156 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
6157 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
6158 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
6159 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
6160 make_temp_ssa_name): move to tree-ssanames.h
6161 * tree-ssa-alias.h: Move prototype.
6162 * tree-ssa.h: Include tree-ssanames.h.
6163 * tree-ssanames.c (FREE_SSANAMES): Move to here.
6164 * tree-ssanames.h: New. Move items from tree-flow*.h
6165 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
6167 2013-09-12 Richard Biener <rguenther@suse.de>
6169 PR tree-optimization/58404
6170 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
6171 propagate non-invariant addresses into dereferences wrapped
6172 in component references.
6174 2013-09-12 Richard Biener <rguenther@suse.de>
6176 PR tree-optimization/58402
6177 * passes.def: Move pass_late_warn_uninitialized later.
6179 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
6181 * tree-ssa.h: New. Move content from tree-flow.h and
6183 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
6184 Move prototypes belonging to tree-ssa.c.
6185 * tree-flow-inline.h (redirect_edge_var_map_def,
6186 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
6188 * gimple.h: Adjust prototypes.
6189 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
6191 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
6192 * tree.h: Move prototype to tree-ssa.h.
6193 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
6194 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
6195 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
6196 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
6197 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
6198 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
6199 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
6200 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
6201 graphite-blocking.c, graphite-clast-to-gimple.c,
6202 graphite-dependences.c, graphite-interchange.c,
6203 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
6204 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
6205 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
6206 ipa-reference.c, ipa-split.c, ipa-utils.c,
6207 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
6208 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
6209 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
6210 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
6211 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
6212 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
6213 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
6214 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
6215 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
6216 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
6217 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
6218 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6219 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
6220 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
6221 value-prof.c, var-tracking.c,
6222 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
6224 2013-09-12 Richard Biener <rguenther@suse.de>
6226 PR tree-optimization/58396
6227 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
6228 (build_rdg): Take a loop-nest parameter, fix memleaks.
6229 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
6231 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
6233 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
6236 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
6238 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
6239 The FMA instruction names should have a 'v' prefix.
6241 2013-09-12 Richard Biener <rguenther@suse.de>
6243 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
6244 (dot_rdg): Use popen instead of system in optional code.
6245 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
6246 (already_processed_vertex_p): Adjust.
6247 (has_anti_or_output_dependence, predecessor_has_mem_write,
6248 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
6249 rdg_flag_uses): Remove.
6250 (rdg_flag_vertex): Simplify.
6251 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
6253 (build_rdg_partition_for_component): Process the first vertex
6254 of a component only.
6255 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
6257 2013-09-12 Alan Modra <amodra@gmail.com>
6259 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
6261 2013-09-11 DJ Delorie <dj@redhat.com>
6262 Nick Clifton <nickc@redhat.com>
6264 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
6265 (rl78_stack_based_mem): New.
6266 * config/rl78/constraints.md (Iv08): New.
6275 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
6277 (movsi): Change from expand to insn-and-split.
6278 (ashrsi3): Clobber AX.
6282 * config/rl78/rl78.md (CC_REG): Fix.
6283 (addsi3): Allow memory and immediate operands.
6284 (addsi3_internal): Split into...
6285 (addsi3_internal_virt): ...new, and ...
6286 (addsi3_internal_real): ...new.
6288 (subsi3_internal_virt): New.
6289 (subsi3_internal_real): New.
6290 (mulsi3): Add memory operand.
6291 (mulsi3_rl78): Likewise.
6292 (mulsi3_g13): Likewise.
6293 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
6294 (cbranchqi4_real): Add more constraint options.
6295 (cbranchhi4_real): Expand pattern.
6296 (cbranchhi4_real_signed): New.
6297 (cbranchhi4_real_inverted): New.
6298 (cbranchsi4_real_lt): New.
6299 (cbranchsi4_real_ge): New.
6300 (cbranchsi4_real_signed): New.
6301 (cbranchsi4_real): New.
6303 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
6305 (lshrsi3_virt): Likewise.
6306 (ashlsi3_virt): Likewise.
6307 (cbranchqi4_virt_signed): New.
6308 (cbranchhi4_virt_signed): New.
6309 (cbranchsi4_virt): New.
6310 * config/rl78/rl78.c: Whitespace fixes throughout.
6311 (move_elim_pass): New.
6312 (pass_data_rl78_move_elim): New.
6313 (pass_rl78_move_elim): New.
6314 (make_pass_rl78_move_elim): New.
6315 (rl78_devirt_info): Run devirt earlier.
6316 (rl78_move_elim_info): New.
6317 (rl78_asm_file_start): Register it.
6318 (rl78_split_movsi): New.
6319 (rl78_as_legitimate_address): Allow virtual base registers when
6321 (rl78_addr_space_convert): Remove spurious debug stuff.
6322 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
6323 (rl78_print_operand): More cases for not printing '#'.
6324 (rl78_expand_compare): Remove most of the logic.
6325 (content_memory): New.
6326 (clear_content_memory): New.
6327 (get_content_index): New.
6328 (get_content_name): New.
6329 (display_content_memory): New.
6330 (update_content): New.
6331 (record_content): New.
6332 (already_contains): New.
6333 (insn_ok_now): Re-recog insns with virtual registers.
6334 (add_postponed_content_update): New.
6335 (process_postponed_content_update): New.
6336 (gen_and_emit_move): New.
6337 (transcode_memory_rtx): Record new location content.
6338 Use gen_and_emit_move.
6339 (force_into_acc): New.
6340 (move_to_acc): Use gen_and_emit_move.
6341 (move_from_acc): Likewise.
6342 (move_acc_to_reg): Likewise.
6343 (move_to_x): Likewise.
6344 (move_to_hl): Likewise.
6345 (move_to_de): Likewise.
6346 (rl78_alloc_physical_registers_op1): Record location content.
6347 (has_constraint): New.
6348 (rl78_alloc_physical_registers_op2): Record location content.
6350 (rl78_alloc_physical_registers_ro1): Likewise.
6351 (rl78_alloc_physical_registers_cmp): Likewise.
6352 (rl78_alloc_physical_registers_umul): Likewise.
6353 (rl78_alloc_address_registers_macax): New.
6354 (rl78_alloc_physical_registers): Initialize and set location
6355 content memory as needed.
6356 (rl78_reorg): Make sure split2 is called.
6357 (rl78_rtx_costs): New.
6359 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
6361 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
6362 for (not (neg ...)) and (neg (not ...)) cases.
6364 2013-09-11 Richard Biener <rguenther@suse.de>
6367 * passes.def: Split critical edges before late uninit warning passes.
6368 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
6370 2013-09-11 Jakub Jelinek <jakub@redhat.com>
6372 PR tree-optimization/58385
6373 * fold-const.c (build_range_check): If both low and high are NULL,
6374 use omit_one_operand_loc to preserve exp side-effects.
6376 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6378 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
6380 2013-09-11 Richard Biener <rguenther@suse.de>
6382 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
6383 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
6384 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
6385 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
6386 stmts_from_loop, known_dependences_p, build_empty_rdg,
6387 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
6388 * tree-loop-distribution.c: ... here.
6389 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
6390 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
6391 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
6392 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
6393 * tree-loop-distribution.c: ... here.
6394 * tree-loop-distribution.c: Include gimple-pretty-print.h.
6395 (struct partition_s): Add loops member.
6396 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
6397 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
6398 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
6400 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6401 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6402 Sergey Lega <sergey.s.lega@intel.com>
6403 Anna Tikhonova <anna.tikhonova@intel.com>
6404 Ilya Tocar <ilya.tocar@intel.com>
6405 Andrey Turetskiy <andrey.turetskiy@intel.com>
6406 Ilya Verbin <ilya.verbin@intel.com>
6407 Kirill Yukhin <kirill.yukhin@intel.com>
6408 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6410 * config/i386/constraints.md (k): New.
6412 * config/i386/i386.c (const regclass_map): Add new mask registers.
6413 (dbx_register_map): Ditto.
6414 (dbx64_register_map): Ditto.
6415 (svr4_dbx_register_map): Ditto.
6416 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
6418 (ix86_preferred_reload_class): Disable constants for mask registers.
6419 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
6420 (ix86_hard_regno_mode_ok): Support new mask registers.
6421 (x86_order_regs_for_local_alloc): Ditto.
6422 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
6423 (FIXED_REGISTERS): Add new mask registers.
6424 (CALL_USED_REGISTERS): Ditto.
6425 (REG_ALLOC_ORDER): Ditto.
6426 (VALID_MASK_REG_MODE): New.
6427 (FIRST_MASK_REG): Ditto.
6428 (LAST_MASK_REG): Ditto.
6429 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
6430 (MAYBE_MASK_CLASS_P): New.
6431 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
6432 (REG_CLASS_CONTENTS): Ditto.
6433 (MASK_REGNO_P): New.
6434 (ANY_MASK_REG_P): Ditto.
6435 (HI_REGISTER_NAMES): Add new mask registers.
6436 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
6437 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
6439 (attribute "type"): Add mskmov, msklog.
6440 (attribute "length_immediate"): Support them.
6441 (attribute "memory"): Ditto.
6442 (attribute "prefix_0f"): Ditto.
6443 (*movhi_internal): Support new mask registers.
6444 (*movqi_internal): Ditto.
6445 (define_split): Split out clobber pattern is a logic
6446 insn on mask registers.
6447 (*k<logic><mode>): New.
6448 (*andhi_1): Extend to support mask regs.
6449 (*andqi_1): Extend to support mask regs.
6451 (define_split): Split and-not to and and not if operands
6453 (*<code><mode>_1): Separate HI mode to new pattern...
6454 (*<code>hi_1): This.
6455 (*<code>qi_1): Extend to support mask regs.
6457 (kortestzhi): Ditto.
6458 (kortestchi): Ditto.
6460 (*one_cmpl<mode>2_1): Remove HImode and handle it...
6461 (*one_cmplhi2_1): ...Here, now with mask registers support.
6462 (*one_cmplqi2_1): Support new mask registers.
6463 (HI/QImode arithmetics splitter): Don't split if mask registers
6465 (HI/QImode not splitter): Ditto.
6466 * config/i386/predicated.md (mask_reg_operand): New.
6467 (general_reg_operand): Ditto.
6469 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
6471 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
6472 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
6474 2013-09-10 Jeff Law <law@redhat.com>
6476 PR tree-optimization/58380
6477 * tree-ssa-threadupdate.c (thread_block): Recognize another case
6478 of threading through a buried loop header.
6480 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
6481 return value for single successor case.
6483 2013-09-10 Jan Hubicka <jh@suse.cz>
6485 * ipa-devirt.c (ipa_devirt): Enable with LTO.
6487 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
6490 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
6491 support conditional execution.
6492 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
6494 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
6496 * lra.c (lra): Clear lra_optional_reload_pseudos before every
6498 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
6499 Check destination too to check move insn.
6500 (undo_optional_reloads): Add check that the original peudo did not
6501 changed its allocation and the optional reload was inherited on last
6502 inheritance pass. Break loop after deciding to keep optional reload.
6503 (lra_undo_inheritance): Add check that inherited pseudo still in
6506 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
6508 * config/aarch64/aarch64.md (generic_sched): New.
6509 * config/aarch64/aarch64-generic.md (load): Make conditional
6510 on generic_sched attribute.
6511 (nonload): Likewise.
6513 2013-09-10 Jan Hubicka <jh@suse.cz>
6515 * lto-cgraph.c: Include ipa-utils.h.
6516 (compute_ltrans_boundary): Also add possible targets into the boundary.
6518 2013-09-10 Jan Hubicka <jh@suse.cz>
6520 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
6521 VAR_DECL of vtable rather than full expression.
6523 2013-09-10 Jan Hubicka <jh@suse.cz>
6524 Paolo Carlini <paolo.carlini@oracle.com>
6526 * cgraphunit.c (analyze_functions): Save input_location, set it
6527 to UNKNOWN_LOCATION and restore it at the end.
6529 2013-09-10 Martin Jambor <mjambor@suse.cz>
6531 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
6532 represented by a thunk.
6534 2013-09-10 Jeff Law <law@redhat.com>
6536 PR tree-optimization/58343
6537 * tree-ssa-threadupdate.c (thread_block): Identify and disable
6538 jump threading requests through loop headers buried in the middle
6539 of a jump threading path.
6541 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
6542 in return value/type.
6544 2013-09-10 Jakub Jelinek <jakub@redhat.com>
6546 PR rtl-optimization/58365
6547 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
6548 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
6551 2013-09-10 Richard Biener <rguenther@suse.de>
6553 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
6554 * tree-data-ref.c (create_rdg_vertices): Collect all data
6555 references, signal failure to the caller, use data-ref API.
6556 (build_rdg): Compute data references only once. Maintain lifetime
6557 of data references and data dependences from within RDG.
6558 (free_rdg): Free dependence relations.
6559 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
6560 inventing extra dependences.
6561 (distribute_loop): Update for RDG API changes.
6563 2013-09-10 Kai Tietz <ktietz@redhat.com>
6565 * doc/invoke.texi (fms-extensions): Document changed
6566 behavior for ms-abi targets.
6567 * config/i386/i386.c (ix86_option_override_internal):
6568 Set default value of option -fms-extension for ms-abi targets.
6570 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
6572 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
6574 2013-09-10 Alan Modra <amodra@gmail.com>
6577 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
6579 2013-09-10 Alan Modra <amodra@gmail.com>
6581 * config/rs6000/predicates.md (add_cint_operand): New.
6582 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
6583 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
6584 using add_cint_operand.
6585 (largetoc_high_plus_aix): Likewise.
6587 2013-09-09 Jakub Jelinek <jakub@redhat.com>
6589 PR tree-optimization/58364
6590 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
6591 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
6592 the current range can't be an unconditional true or false.
6594 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
6596 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
6598 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
6600 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
6602 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
6605 * doc/invoke.texi (-Wdelete-incomplete): Document it.
6607 2013-09-09 Ian Bolton <ian.bolton@arm.com>
6609 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
6610 NO_REGS for immediate that can't be moved directly into FP_REGS.
6612 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6614 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
6615 comparison with negated operand.
6616 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
6619 2013-09-09 Richard Biener <rguenther@suse.de>
6622 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
6623 of a subloop record all its block as affecting loop-closed SSA form.
6625 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6627 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
6628 of an rtx/bitpos pair.
6629 (store_fixed_bit_field): Update accordingly.
6631 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6633 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
6635 * builtins.c (expand_errno_check): Likewise.
6636 * dwarf2cfi.c (init_return_column_size): Likewise.
6637 * except.c (sjlj_mark_call_sites): Likewise.
6638 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
6639 * lra-constraints.c (emit_inc): Likewise.
6640 * ree.c (combine_set_extension): Likewise.
6641 * regmove.c (fixup_match_2): Likewise.
6642 * reload1.c (inc_for_reload): Likewise.
6644 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6646 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
6647 (force_to_mode, simplify_shift_const_1, simplify_comparison):
6648 Use gen_int_mode with the mode of the associated simplify_* call.
6649 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
6650 * expmed.c (expand_shift_1): Likewise.
6651 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6652 * loop-iv.c (iv_number_of_iterations): Likewise.
6653 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6654 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
6656 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6658 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
6659 of the associated expand_* call.
6660 (asan_emit_stack_protection): Likewise.
6661 * builtins.c (round_trampoline_addr): Likewise.
6662 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
6663 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
6664 (emit_store_flag): Likewise.
6665 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
6666 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
6668 * function.c (instantiate_virtual_regs_in_insn): Likewise.
6669 * ifcvt.c (noce_try_store_flag_constants): Likewise.
6670 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
6671 * modulo-sched.c (generate_prolog_epilog): Likewise.
6672 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
6673 (expand_ctz, expand_ffs, expand_unop): Likewise.
6675 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6677 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
6678 of the associated gen_rtx_* call.
6679 * caller-save.c (init_caller_save): Likewise.
6680 * combine.c (find_split_point, make_extraction): Likewise.
6681 (make_compound_operation): Likewise.
6682 * dwarf2out.c (mem_loc_descriptor): Likewise.
6683 * explow.c (plus_constant, probe_stack_range): Likewise.
6684 * expmed.c (expand_mult_const): Likewise.
6685 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
6686 * reload1.c (init_reload): Likewise.
6687 * valtrack.c (cleanup_auto_inc_dec): Likewise.
6688 * var-tracking.c (adjust_mems): Likewise.
6689 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
6690 rather than gen_rtx_fmt_ee.
6692 2013-09-09 Jan Hubicka <jh@suse.cz>
6695 * value-prof.c (gimple_ic): Copy also abnormal edges.
6697 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
6699 * asan.c (asan_shadow_cst): Use gen_int_mode.
6701 2013-09-08 Jan Hubicka <jh@suse.cz>
6703 * ipa-profile.c: Add toplevel comment.
6704 (ipa_propagate_frequency_1): Be more conservative when profile is read.
6705 (contains_hot_call_p): New function.
6706 (ipa_propagate_frequency): Set frequencies based on counts when
6708 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
6709 profile; do not tamper with profile after inlining if it is read.
6711 2013-09-08 Jan Hubicka <jh@suse.cz>
6713 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
6716 2013-09-08 Jan Hubicka <jh@suse.cz>
6718 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
6721 2013-09-08 Jeff Law <law@redhat.com>
6724 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
6727 2013-09-08 Andi Kleen <ak@linux.intel.com>
6729 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
6731 2013-09-08 Jan Hubicka <jh@suse.cz>
6733 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
6735 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
6737 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
6738 for non-debug insns.
6739 * lra.c (new_insn_reg): Take the containing insn as a parameter.
6740 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
6741 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
6744 2013-09-08 Jan Hubicka <jh@suse.cz>
6746 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
6747 targets and devirtualize to BUILT_IN_UNREACHABLE.
6748 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
6749 * ipa.c (walk_polymorphic_call_targets): New function.
6750 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
6751 functions; use the new timevar.
6752 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
6753 was removed from the program.
6754 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
6755 not consider it in the walk when its vtable is dead.
6756 (possible_polymorphic_call_targets_1): Pass anonymous flag to
6758 (devirt_variable_node_removal_hook): New function.
6759 (possible_polymorphic_call_targets): Also register
6760 devirt_variable_node_removal_hook.
6761 (ipa_devirt): Do not do non-speculative devirtualization.
6762 (gate_ipa_devirt): One execute if devirtualizing speculatively.
6764 2013-09-08 Jan Hubicka <jh@suse.cz>
6766 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
6767 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
6768 varpool_remove_variable_insertion_hook): Declare.
6769 * varpool.c (varpool_node_hook_list): New structure.
6770 (first_varpool_node_removal_hook,
6771 first_varpool_variable_insertion_hook): New variables.
6772 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
6773 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
6774 varpool_remove_variable_insertion_hook,
6775 varpool_call_variable_insertion_hooks): New functions.
6776 (varpool_remove_node): Use it.
6778 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6781 * diagnostic.c (diagnostic_build_prefix): When s.file is
6782 "<built-in>" don't output line and column numbers.
6784 2013-09-06 Jan Hubicka <jh@suse.cz>
6786 * cgraphunit.c (expand_thunk): Get body before touching arguments.
6787 * lto-streamer-out.c: Stream thunks, too.
6788 * lto-streamer-in.c (input_function): Pop cfun here
6789 (lto_read_body): Instead of here.
6791 2013-09-06 Caroline Tice <cmtice@google.com>
6793 * doc/install.texi: Add documentation for the --enable-vtable-verify
6794 and the --disable-libvtv configure options.
6796 2013-09-06 Jeff Law <law@redhat.com>
6798 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
6799 edge implied equivalences into successor phis.
6801 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
6803 * resource.c (mark_referenced_resources): Handle COND_EXEC.
6805 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
6807 * resource.c (mark_target_live_regs): Compute resources taking
6808 into account if a call is predicated or not.
6810 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
6812 * toplev.c (output_stack_usage): Be prepared for suffixes created by
6813 the compiler in the function names.
6815 2013-09-06 Jan Hubicka <jh@suse.cz>
6818 * ipa-inline.c (has_caller_p): New function.
6819 (want_inline_function_to_all_callers_p): Use it.
6820 (sum_callers, inline_to_all_callers): Break out from ...
6821 (ipa_inline): ... here.
6823 2013-09-06 Jan Hubicka <jh@suse.cz>
6825 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
6826 only when AVX is enabled.
6828 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6830 * config/aarch64/aarch64.md
6831 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
6832 is fpsimd_<load/store>2.
6833 (load_pair<mode>): Likewise.
6834 (store_pair<mode>): Likewise.
6836 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6838 * config/arm/types.md (type): Add "mrs" type.
6839 * config/aarch64/aarch64.md
6840 (aarch64_load_tp_hard): Make type "mrs".
6842 (load_tp_hard): Make type "mrs".
6843 * config/arm/cortex-a15.md: Update with new attributes.
6844 * config/arm/cortex-a5.md: Update with new attributes.
6845 * config/arm/cortex-a53.md: Update with new attributes.
6846 * config/arm/cortex-a7.md: Update with new attributes.
6847 * config/arm/cortex-a8.md: Update with new attributes.
6848 * config/arm/cortex-a9.md: Update with new attributes.
6849 * config/arm/cortex-m4.md: Update with new attributes.
6850 * config/arm/cortex-r4.md: Update with new attributes.
6851 * config/arm/fa526.md: Update with new attributes.
6852 * config/arm/fa606te.md: Update with new attributes.
6853 * config/arm/fa626te.md: Update with new attributes.
6854 * config/arm/fa726te.md: Update with new attributes.
6856 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6858 * config/aarch64/aarch64.md
6859 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
6860 (*movtf_aarch64): Likewise.
6862 (thumb1_movdi_insn): Use "multiple" for type where more than one
6863 instruction is used for a move.
6864 (*arm32_movhf): Likewise.
6865 (*thumb_movdf_insn): Likewise.
6867 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6869 * config/arm/types.md (type): Rename fcpys to fmov.
6871 (*arm_movsi_vfp): Rename type fcpys as fmov.
6872 (*thumb2_movsi_vfp): Likewise
6873 (*movhf_vfp_neon): Likewise
6874 (*movhf_vfp): Likewise
6875 (*movsf_vfp): Likewise
6876 (*thumb2_movsf_vfp): Likewise
6877 (*movsfcc_vfp): Likewise
6878 (*thumb2_movsfcc_vfp): Likewise
6879 * config/aarch64/aarch64-simd.md
6880 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
6881 * config/aarch64/aarch64.md
6882 (*movsi_aarch64): Replace type mov_reg with fmovs.
6883 (*movdi_aarch64): Likewise
6884 (*movsf_aarch64): Likewise
6885 (*movdf_aarch64): Likewise
6887 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
6888 * config/arm/iwmmxt.md
6889 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
6890 * config/arm/arm1020e.md: Update with new attributes.
6891 * config/arm/cortex-a15-neon.md: Update with new attributes.
6892 * config/arm/cortex-a5.md: Update with new attributes.
6893 * config/arm/cortex-a53.md: Update with new attributes.
6894 * config/arm/cortex-a7.md: Update with new attributes.
6895 * config/arm/cortex-a8-neon.md: Update with new attributes.
6896 * config/arm/cortex-a9.md: Update with new attributes.
6897 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6898 * config/arm/cortex-r4f.md: Update with new attributes.
6899 * config/arm/marvell-pj4.md: Update with new attributes.
6900 * config/arm/vfp11.md: Update with new attributes.
6902 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6904 * config/aarch64/aarch64.md
6905 (*madd<mode>): Fix type attribute.
6906 (*maddsi_uxtw): Likewise.
6907 (*msub<mode>): Likewise.
6908 (*msubsi_uxtw): Likewise.
6909 (<su_optab>maddsidi4): Likewise.
6910 (<su_optab>msubsidi4): Likewise.
6912 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6914 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
6915 * config/arm/arm.md (core_cycles): Remove fdiv.
6916 * config/arm/vfp.md:
6917 (*sqrtsf2_vfp): Update for attribute changes.
6918 (*sqrtdf2_vfp): Likewise.
6919 * config/aarch64/aarch64.md:
6920 (sqrt<mode>2): Update for attribute changes.
6921 * config/arm/arm1020e.md: Update with new attributes.
6922 * config/arm/cortex-a15-neon.md: Update with new attributes.
6923 * config/arm/cortex-a5.md: Update with new attributes.
6924 * config/arm/cortex-a53.md: Update with new attributes.
6925 * config/arm/cortex-a7.md: Update with new attributes.
6926 * config/arm/cortex-a8-neon.md: Update with new attributes.
6927 * config/arm/cortex-a9.md: Update with new attributes.
6928 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6929 * config/arm/cortex-r4f.md: Update with new attributes.
6930 * config/arm/marvell-pj4.md: Update with new attributes.
6931 * config/arm/vfp11.md: Update with new attributes.
6933 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6935 * config/arm/types.md
6936 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
6937 * config/aarch64/aarch64.md
6938 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
6940 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
6941 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
6942 (float<GPI:mode><GPF:mode>2): Likewise.
6944 (*truncsisf2_vfp): Update with new attributes.
6945 (*truncsidf2_vfp): Likewise.
6946 (fixuns_truncsfsi2): Likewise.
6947 (fixuns_truncdfsi2): Likewise.
6948 (*floatsisf2_vfp): Likewise.
6949 (*floatsidf2_vfp): Likewise.
6950 (floatunssisf2): Likewise.
6951 (floatunssidf2): Likewise.
6952 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
6953 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
6954 * config/arm/arm1020e.md: Update with new attributes.
6955 * config/arm/cortex-a15-neon.md: Update with new attributes.
6956 * config/arm/cortex-a5.md: Update with new attributes.
6957 * config/arm/cortex-a53.md: Update with new attributes.
6958 * config/arm/cortex-a7.md: Update with new attributes.
6959 * config/arm/cortex-a8-neon.md: Update with new attributes.
6960 * config/arm/cortex-a9.md: Update with new attributes.
6961 * config/arm/cortex-m4-fpu.md: Update with new attributes.
6962 * config/arm/cortex-r4f.md: Update with new attributes.
6963 * config/arm/marvell-pj4.md: Update with new attributes.
6964 * config/arm/vfp11.md: Update with new attributes.
6966 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6968 * config/aarch64/arm_neon.h
6969 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
6970 (vqtbx<1,2,3,4><q>_s8): Likewise.
6972 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
6974 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
6975 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
6976 (add<mode>3): Add type attribute.
6977 (add<mode>3): Likewise.
6978 (usadd<mode>3): Likewise.
6979 (ssadd<mode>3): Likewise.
6980 (sub<mode>3): Likewise.
6981 (sub<mode>3): Likewise.
6982 (ussub<mode>3): Likewise.
6983 (sssub<mode>3): Likewise.
6984 (ssmulsa3): Likewise.
6985 (usmulusa3): Likewise.
6986 (arm_usatsihi): Likewise.
6988 (*movdi_vfp): Add types for all instructions.
6989 (*movdi_vfp_cortexa8): Likewise.
6990 (*movhf_vfp_neon): Likewise.
6991 (*movhf_vfp): Likewise.
6992 (*movdf_vfp): Likewise.
6993 (*thumb2_movdf_vfp): Likewise.
6994 (*thumb2_movdfcc_vfp): Likewise.
6995 * config/arm/arm.md: Add type attribute to all insn patterns.
6996 (*thumb1_adddi3): Add type attribute.
6997 (*arm_adddi3): Likewise.
6998 (*adddi_sesidi_di): Likewise.
6999 (*adddi_zesidi_di): Likewise.
7000 (*thumb1_addsi3): Likewise.
7001 (addsi3_compare0): Likewise.
7002 (*addsi3_compare0_scratch): Likewise.
7003 (*compare_negsi_si): Likewise.
7004 (cmpsi2_addneg): Likewise.
7005 (*addsi3_carryin_<optab>): Likewise.
7006 (*addsi3_carryin_alt2_<optab>): Likewise.
7007 (*addsi3_carryin_clobercc_<optab>): Likewise.
7008 (*subsi3_carryin): Likewise.
7009 (*subsi3_carryin_const): Likewise.
7010 (*subsi3_carryin_compare): Likewise.
7011 (*subsi3_carryin_compare_const): Likewise.
7012 (*arm_subdi3): Likewise.
7013 (*thumb_subdi3): Likewise.
7014 (*subdi_di_zesidi): Likewise.
7015 (*subdi_di_sesidi): Likewise.
7016 (*subdi_zesidi_di): Likewise.
7017 (*subdi_sesidi_di): Likewise.
7018 (*subdi_zesidi_ze): Likewise.
7019 (thumb1_subsi3_insn): Likewise.
7020 (*arm_subsi3_insn): Likewise.
7021 (*anddi3_insn): Likewise.
7022 (*anddi_zesidi_di): Likewise.
7023 (*anddi_sesdi_di): Likewise.
7024 (*ne_zeroextracts): Likewise.
7025 (*ne_zeroextracts): Likewise.
7026 (*ite_ne_zeroextr): Likewise.
7027 (*ite_ne_zeroextr): Likewise.
7028 (*anddi_notdi_di): Likewise.
7029 (*anddi_notzesidi): Likewise.
7030 (*anddi_notsesidi): Likewise.
7031 (andsi_notsi_si): Likewise.
7032 (thumb1_bicsi3): Likewise.
7033 (*iordi3_insn): Likewise.
7034 (*iordi_zesidi_di): Likewise.
7035 (*iordi_sesidi_di): Likewise.
7036 (*thumb1_iorsi3_insn): Likewise.
7037 (*xordi3_insn): Likewise.
7038 (*xordi_zesidi_di): Likewise.
7039 (*xordi_sesidi_di): Likewise.
7040 (*arm_xorsi3): Likewise.
7041 (*andsi_iorsi3_no): Likewise.
7042 (*smax_0): Likewise.
7043 (*smax_m1): Likewise.
7044 (*arm_smax_insn): Likewise.
7045 (*smin_0): Likewise.
7046 (*arm_smin_insn): Likewise.
7047 (*arm_umaxsi3): Likewise.
7048 (*arm_uminsi3): Likewise.
7049 (*minmax_arithsi): Likewise.
7050 (*minmax_arithsi_): Likewise.
7051 (*satsi_<SAT:code>): Likewise.
7052 (arm_ashldi3_1bit): Likewise.
7053 (arm_ashrdi3_1bit): Likewise.
7054 (arm_lshrdi3_1bit): Likewise.
7055 (*arm_negdi2): Likewise.
7056 (*thumb1_negdi2): Likewise.
7057 (*arm_negsi2): Likewise.
7058 (*thumb1_negsi2): Likewise.
7059 (*negdi_extendsid): Likewise.
7060 (*negdi_zero_extend): Likewise.
7061 (*arm_abssi2): Likewise.
7062 (*thumb1_abssi2): Likewise.
7063 (*arm_neg_abssi2): Likewise.
7064 (*thumb1_neg_abss): Likewise.
7065 (one_cmpldi2): Likewise.
7066 (extend<mode>di2): Likewise.
7067 (*compareqi_eq0): Likewise.
7068 (*arm_extendhisi2addsi): Likewise.
7069 (*arm_movdi): Likewise.
7070 (*thumb1_movdi_insn): Likewise.
7071 (*arm_movt): Likewise.
7072 (*thumb1_movsi_insn): Likewise.
7073 (pic_add_dot_plus_four): Likewise.
7074 (pic_add_dot_plus_eight): Likewise.
7075 (tls_load_dot_plus_eight): Likewise.
7076 (*thumb1_movhi_insn): Likewise.
7077 (*thumb1_movsf_insn): Likewise.
7078 (*movdf_soft_insn): Likewise.
7079 (*thumb_movdf_insn): Likewise.
7080 (cbranchsi4_insn): Likewise.
7081 (cbranchsi4_scratch): Likewise.
7082 (*negated_cbranchsi4): Likewise.
7083 (*tbit_cbranch): Likewise.
7084 (*tlobits_cbranch): Likewise.
7085 (*tstsi3_cbranch): Likewise.
7086 (*cbranchne_decr1): Likewise.
7087 (*addsi3_cbranch): Likewise.
7088 (*addsi3_cbranch_scratch): Likewise.
7089 (*arm_cmpdi_insn): Likewise.
7090 (*arm_cmpdi_unsig): Likewise.
7091 (*arm_cmpdi_zero): Likewise.
7092 (*thumb_cmpdi_zero): Likewise.
7093 (*deleted_compare): Likewise.
7094 (*mov_scc): Likewise.
7095 (*mov_negscc): Likewise.
7096 (*mov_notscc): Likewise.
7097 (*cstoresi_eq0_thumb1_insn): Likewise.
7098 (cstoresi_nltu_thumb1): Likewise.
7099 (cstoresi_ltu_thu): Likewise.
7100 (thumb1_addsi3_addgeu): Likewise.
7101 (*arm_jump): Likewise.
7102 (*thumb_jump): Likewise.
7103 (*check_arch2): Likewise.
7104 (arm_casesi_internal): Likewise.
7105 (thumb1_casesi_dispatch): Likewise.
7106 (*arm_indirect_jump): Likewise.
7107 (*thumb1_indirect_jump): Likewise.
7109 (*and_scc): Likewise.
7110 (*ior_scc): Likewise.
7111 (*compare_scc): Likewise.
7112 (*cond_move): Likewise.
7113 (*cond_arith): Likewise.
7114 (*cond_sub): Likewise.
7115 (*cmp_ite0): Likewise.
7116 (*cmp_ite1): Likewise.
7117 (*cmp_and): Likewise.
7118 (*cmp_ior): Likewise.
7119 (*ior_scc_scc): Likewise.
7120 (*ior_scc_scc_cmp): Likewise.
7121 (*and_scc_scc): Likewise.
7122 (*and_scc_scc_cmp): Likewise.
7123 (*and_scc_scc_nod): Likewise.
7124 (*negscc): Likewise.
7125 (movcond_addsi): Likewise.
7126 (movcond): Likewise.
7127 (*ifcompare_plus_move): Likewise.
7128 (*if_plus_move): Likewise.
7129 (*ifcompare_move_plus): Likewise.
7130 (*if_move_plus): Likewise.
7131 (*ifcompare_arith_arith): Likewise.
7132 (*if_arith_arith): Likewise.
7133 (*ifcompare_arith_move): Likewise.
7134 (*if_arith_move): Likewise.
7135 (*ifcompare_move_arith): Likewise.
7136 (*if_move_arith): Likewise.
7137 (*ifcompare_move_not): Likewise.
7138 (*if_move_not): Likewise.
7139 (*ifcompare_not_move): Likewise.
7140 (*if_not_move): Likewise.
7141 (*ifcompare_shift_move): Likewise.
7142 (*if_shift_move): Likewise.
7143 (*ifcompare_move_shift): Likewise.
7144 (*if_move_shift): Likewise.
7145 (*ifcompare_shift_shift): Likewise.
7146 (*ifcompare_not_arith): Likewise.
7147 (*ifcompare_arith_not): Likewise.
7148 (*if_arith_not): Likewise.
7149 (*ifcompare_neg_move): Likewise.
7150 (*if_neg_move): Likewise.
7151 (*ifcompare_move_neg): Likewise.
7152 (*if_move_neg): Likewise.
7153 (prologue_thumb1_interwork): Likewise.
7154 (*cond_move_not): Likewise.
7155 (*sign_extract_onebit): Likewise.
7156 (*not_signextract_onebit): Likewise.
7157 (stack_tie): Likewise.
7158 (align_4): Likewise.
7159 (align_8): Likewise.
7160 (consttable_end): Likewise.
7161 (consttable_1): Likewise.
7162 (consttable_2): Likewise.
7163 (consttable_4): Likewise.
7164 (consttable_8): Likewise.
7165 (consttable_16): Likewise.
7166 (*thumb1_tablejump): Likewise.
7167 (prefetch): Likewise.
7168 (force_register_use): Likewise.
7169 (thumb_eh_return): Likewise.
7170 (load_tp_hard): Likewise.
7171 (load_tp_soft): Likewise.
7172 (tlscall): Likewise.
7173 (*arm_movtas_ze): Likewise.
7174 (*arm_rev): Likewise.
7175 (*arm_revsh): Likewise.
7176 (*arm_rev16): Likewise.
7177 * config/arm/thumb2.md
7178 (*thumb2_smaxsi3): Likewise.
7179 (*thumb2_sminsi3): Likewise.
7180 (*thumb32_umaxsi3): Likewise.
7181 (*thumb2_uminsi3): Likewise.
7182 (*thumb2_negdi2): Likewise.
7183 (*thumb2_abssi2): Likewise.
7184 (*thumb2_neg_abss): Likewise.
7185 (*thumb2_movsi_insn): Likewise.
7186 (tls_load_dot_plus_four): Likewise.
7187 (*thumb2_movhi_insn): Likewise.
7188 (*thumb2_mov_scc): Likewise.
7189 (*thumb2_mov_negs): Likewise.
7190 (*thumb2_mov_negs): Likewise.
7191 (*thumb2_mov_nots): Likewise.
7192 (*thumb2_mov_nots): Likewise.
7193 (*thumb2_movsicc_): Likewise.
7194 (*thumb2_movsfcc_soft_insn): Likewise.
7195 (*thumb2_indirect_jump): Likewise.
7196 (*thumb2_and_scc): Likewise.
7197 (*thumb2_ior_scc): Likewise.
7198 (*thumb2_ior_scc_strict_it): Likewise.
7199 (*thumb2_cond_move): Likewise.
7200 (*thumb2_cond_arith): Likewise.
7201 (*thumb2_cond_ari): Likewise.
7202 (*thumb2_cond_sub): Likewise.
7203 (*thumb2_negscc): Likewise.
7204 (*thumb2_movcond): Likewise.
7205 (thumb2_casesi_internal): Likewise.
7206 (thumb2_casesi_internal_pic): Likewise.
7207 (*thumb2_alusi3_short): Likewise.
7208 (*thumb2_mov<mode>_shortim): Likewise.
7209 (*thumb2_addsi_short): Likewise.
7210 (*thumb2_subsi_short): Likewise.
7211 (thumb2_addsi3_compare0): Likewise.
7212 (*thumb2_cbz): Likewise.
7213 (*thumb2_cbnz): Likewise.
7214 (*thumb2_one_cmplsi2_short): Likewise.
7215 (*thumb2_negsi2_short): Likewise.
7216 (*orsi_notsi_si): Likewise.
7217 * config/arm/arm1020e.md: Update with new attributes.
7218 * config/arm/arm1026ejs.md: Update with new attributes.
7219 * config/arm/arm1136jfs.md: Update with new attributes.
7220 * config/arm/arm926ejs.md: Update with new attributes.
7221 * config/arm/cortex-a15.md: Update with new attributes.
7222 * config/arm/cortex-a5.md: Update with new attributes.
7223 * config/arm/cortex-a53.md: Update with new attributes.
7224 * config/arm/cortex-a7.md: Update with new attributes.
7225 * config/arm/cortex-a8.md: Update with new attributes.
7226 * config/arm/cortex-a9.md: Update with new attributes.
7227 * config/arm/cortex-m4.md: Update with new attributes.
7228 * config/arm/cortex-r4.md: Update with new attributes.
7229 * config/arm/fa526.md: Update with new attributes.
7230 * config/arm/fa606te.md: Update with new attributes.
7231 * config/arm/fa626te.md: Update with new attributes.
7232 * config/arm/fa726te.md: Update with new attributes.
7234 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
7236 * config/aarch64/aarch64-simd.md
7237 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
7238 <vwx> iterator to ensure correct register choice.
7239 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
7240 (aarch64_sqdmull_n<mode>): Likewise.
7241 (aarch64_sqdmull2_n<mode>_internal): Likewise.
7242 * config/aarch64/arm_neon.h
7243 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
7244 (vml<as><q>_n_<su>16): Likewise.
7245 (vml<as>l_high_lane<q>_<su>16): Likewise.
7246 (vml<as>l_high_n_<su>16): Likewise.
7247 (vml<as>l_lane<q>_<su>16): Likewise.
7248 (vml<as>l_n_<su>16): Likewise.
7249 (vmul<q>_lane<q>_<su>16): Likewise.
7250 (vmul<q>_n_<su>16): Likewise.
7251 (vmull_lane<q>_<su>16): Likewise.
7252 (vmull_n_<su>16): Likewise.
7253 (vmull_high_lane<q>_<su>16): Likewise.
7254 (vmull_high_n_<su>16): Likewise.
7255 (vqrdmulh<q>_n_s16): Likewise.
7257 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
7259 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
7260 have the correct lane parameter.
7262 2013-09-06 Richard Biener <rguenther@suse.de>
7264 * cfganal.c (control_dependences::~control_dependences):
7265 Properly free all of the vector.
7267 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
7270 * config/i386/i386.c (ix86_conditional_register_usage):
7271 Proper initialize extended SSE registers.
7273 2013-09-06 Jan Hubicka <jh@suse.cz>
7275 PR tree-optimization/58311
7276 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
7278 2013-09-06 Jan Hubicka <jh@suse.cz>
7280 * Makefile.in (tree-sra.o): Update dependencies.
7281 * tree-sra.c: Include ipa-utils.h
7282 (scan_function): Use recursive_call_p.
7283 (has_caller_p): New function.
7284 (cgraph_for_node_and_aliases): Count also callers of aliases.
7286 2013-09-06 Jan Hubicka <jh@suse.cz>
7289 * cgraph.h (symtab_semantically_equivalent_p): Declare.
7290 * tree-tailcall.c: Include ipa-utils.h.
7291 (find_tail_calls): Use it.
7292 * ipa-pure-const.c (check_call): Likewise.
7293 * ipa-utils.c (recursive_call_p): New function.
7294 * ipa-utils.h (recursive_call_p): Dclare.
7295 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
7296 (symtab_semantically_equivalent_p): New function.
7297 * Makefile.in (tree-tailcall.o): Update dependencies.
7299 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
7301 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
7304 2013-09-06 Richard Biener <rguenther@suse.de>
7306 * lto-streamer.h (lto_global_var_decls): Remove.
7307 * Makefile.in (OBJS): Remove lto-symtab.o.
7308 (lto-symtab.o): Remove.
7309 (GTFILES): Remove lto-symtab.c
7310 * lto-symtab.c: Move to lto/
7312 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7314 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
7315 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
7316 (UNSPEC_FPINT_RINT): New constant definitions.
7317 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
7318 definition with 2 attributes.
7319 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
7320 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
7323 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7325 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
7326 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
7327 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
7328 attribute to "z196_alone".
7329 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
7332 2013-09-06 Richard Biener <rguenther@suse.de>
7334 * basic-block.h (class control_dependences): New.
7335 * tree-ssa-dce.c (control_dependence_map): Remove.
7337 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
7338 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
7339 find_pdom, find_control_dependence, find_all_control_dependences):
7341 (mark_control_dependent_edges_necessary,
7342 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
7343 tree_dce_done, perform_tree_ssa_dce): Adjust.
7344 * cfganal.c (set_control_dependence_map_bit,
7345 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
7346 find_all_control_dependences): Move from tree-ssa-dce.c and
7347 implement as methods of control_dependences class.
7348 (control_dependences::control_dependences): New.
7349 (control_dependences::~control_dependences): Likewise.
7350 (control_dependences::get_edges_dependent_on): Likewise.
7351 (control_dependences::get_edge): Likewise.
7353 2013-09-04 Jan Hubicka <jh@suse.cz>
7355 * tree.c (types_same_for_odr): Drop overactive check.
7356 * ipa-devirt.c (hash_type_name): Likewise.
7358 2013-09-04 Jan Hubicka <jh@suse.cz>
7360 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
7361 (analyze_functions): ... here.
7363 2013-09-04 Jan Hubicka <jh@suse.cz>
7366 * cgraphunit.c (analyze_functions): Clear AUX fields
7367 after processing; initialize assembler name has.
7369 2013-09-05 Jeff Law <law@redhat.com>
7371 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
7372 from thread_around_empty_block. Record threading path into PATH.
7373 Recurse if threading through the initial block is successful.
7374 (thread_across_edge): Corresponding changes to slightly simplify.
7376 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7378 * config/aarch64/aarch64.md
7379 (type): Remove frecpe, frecps, frecpx.
7380 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
7381 fix to be a TARGET_SIMD instruction.
7382 (aarch64_frecps): Remove.
7383 * config/aarch64/aarch64-simd.md
7384 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
7385 (aarch64_frecps<mode>): Handle all float/vector of float modes.
7387 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7388 Sofiane Naci <sofiane.naci@arm.com>
7390 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
7391 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
7392 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
7393 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
7394 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
7395 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
7396 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
7397 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
7398 into "clz, "rbit". Rename "shift" to "shift_imm".
7399 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
7400 changes. Update for attribute changes all occurrences of arlo_* and
7402 * config/arm/arm-fixed.md: Update for attribute changes
7403 all occurrences of arlo_* types.
7404 * config/arm/thumb2.md: Update for attribute changes all occurrences
7406 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
7407 (cortexa7_older_only): Likewise.
7408 (cortexa7_younger): Likewise.
7409 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
7410 (1020alu_shift_op): Likewise.
7411 (1020alu_shift_reg_op): Likewise.
7412 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
7413 (alu_shift_op): Likewise.
7414 (alu_shift_reg_op): Likewise.
7415 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
7416 (11_alu_shift_op): Likewise.
7417 (11_alu_shift_reg_op): Likewise.
7418 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
7419 (9_alu_shift_reg_op): Likewise.
7420 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
7422 (cortex_a15_alu_shift): Likewise.
7423 (cortex_a15_alu_shift_reg): Likewise.
7424 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
7426 (cortex_a5_alu_shift): Likewise.
7427 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
7429 (cortex_a53_alu_shift): Likewise.
7430 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
7432 (cortex_a7_alu_reg): Likewise.
7433 (cortex_a7_alu_shift): Likewise.
7434 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
7436 (cortex_a8_alu_shift): Likewise.
7437 (cortex_a8_alu_shift_reg): Likewise.
7438 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
7439 (cortex_a9_dp_shift): Likewise.
7440 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
7442 * config/arm/cortex-r4.md
7443 (cortex_r4_alu): Update for attribute changes.
7444 (cortex_r4_mov): Likewise.
7445 (cortex_r4_alu_shift_reg): Likewise.
7446 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
7447 (526_alu_shift_op): Likewise.
7448 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
7449 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
7450 (626te_alu_shift_op): Likewise.
7451 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
7452 (726te_alu_shift_op): Likewise.
7453 (726te_alu_shift_reg_op): Likewise.
7454 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
7455 (mp626_alu_shift_op): Likewise.
7456 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
7457 (pj4_alu_conds): Likewise.
7458 (pj4_shift): Likewise.
7459 (pj4_shift_conds): Likewise.
7460 (pj4_alu_shift): Likewise.
7461 (pj4_alu_shift_conds): Likewise.
7462 * config/aarch64/aarch64.md: Update for attribute change
7463 all occurrences of arlo_* and shift* types.
7465 2013-09-05 Mike Stump <mikestump@comcast.net>
7467 * tree.h: Move documentation for tree_function_decl to tree-core.h
7468 with the declaration.
7470 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
7473 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
7474 looking for widest mode.
7476 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
7478 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
7480 2013-09-05 Richard Biener <rguenther@suse.de>
7482 PR tree-optimization/58137
7483 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
7484 Do not create vectors of pointers.
7485 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
7486 types for the components of the vector initializer.
7487 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
7488 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
7490 2013-09-05 Martin Jambor <mjambor@suse.cz>
7492 * ipa-prop.c (remove_described_reference): Accept missing references,
7493 return false if that hppens, otherwise return true.
7494 (cgraph_node_for_jfunc): New function.
7495 (try_decrement_rdesc_refcount): Likewise.
7496 (try_make_edge_direct_simple_call): Use them.
7497 (ipa_edge_removal_hook): Remove references from rdescs.
7498 (ipa_edge_duplication_hook): Clone rdescs and their references
7499 when the new edge has the same caller as the old one.
7500 * cgraph.c (cgraph_resolve_speculation): Remove speculative
7501 reference before removing any edges.
7503 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
7505 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
7507 * thumb2.md (thumb2_storewb_parisi): New pattern.
7509 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
7511 * config/aarch64/aarch64-option-extensions.def: Add
7512 AARCH64_OPT_EXTENSION of 'crc'.
7513 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
7514 (AARCH64_ISA_CRC): Ditto.
7515 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
7516 description of the CRC extension.
7518 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
7520 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
7521 * config/rs6000/linux.h: Ditto.
7522 * alpha/linux.h: Ditto.
7523 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
7524 no_c99_libc_has_function.
7525 * config/c6x/uclinux-elf.h: Ditto.
7526 * config/lm32/uclinux-elf.h: Ditto.
7527 * config/m68k/uclinux.h: Ditto.
7528 * config/moxie/uclinux.h: Ditto.
7529 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
7530 (crisv32-*-linux*, cris-*-linux*): Ditto.
7531 * config/bfin/bfin.c: Include "tm_p.h".
7533 2013-09-05 Richard Biener <rguenther@suse.de>
7535 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
7536 check for a definition without a basic-block.
7538 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
7539 Sofiane Naci <sofiane.naci@arm.com>
7541 * config/aarch64/aarch64.md
7542 (*movti_aarch64): Rename r_2_f and f_2_r.
7543 (*movsf_aarch64): Likewise.
7544 (*movdf_aarch64): Likewise.
7545 (*movtf_aarch64): Likewise.
7546 (aarch64_movdi_<mode>low): Likewise.
7547 (aarch64_movdi_<mode>high): Likewise.
7548 (aarch64_mov<mode>high_di): Likewise.
7549 (aarch64_mov<mode>low_di): Likewise.
7550 (aarch64_movtilow_tilow): Likewise.
7551 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
7552 values to config/arm/types.md
7553 (attribute "conds"): Update for attribute change.
7554 (anddi3_insn): Likewise.
7555 (iordi3_insn): Likewise.
7556 (xordi3_insn): Likewise.
7557 (one_cmpldi2): Likewise.
7558 * config/arm/types.md (type): Add Neon types.
7559 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
7560 use "type" attribute.
7561 (movmisalign<mode>_neon_store): Likewise.
7562 (movmisalign<mode>_neon_load): Likewise.
7563 (vec_set<mode>_internal): Likewise.
7564 (vec_setv2di_internal): Likewise.
7565 (vec_extract<mode>): Likewise.
7566 (vec_extractv2di): Likewise.
7567 (add<mode>3_neon): Likewise.
7568 (adddi3_neon): Likewise.
7569 (sub<mode>3_neon): Likewise.
7570 (subdi3_neon): Likewise.
7571 (mul<mode>3_neon): Likewise.
7572 (mul<mode>3add<mode>_neon): Likewise.
7573 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
7574 (fma<VCVTF:mode>4)): Likewise.
7575 (fma<VCVTF:mode>4_intrinsic): Likewise.
7576 (fmsub<VCVTF:mode>4)): Likewise.
7577 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
7578 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
7579 (ior<mode>3): Likewise.
7580 (and<mode>3): Likewise.
7581 (anddi3_neon): Likewise.
7582 (orn<mode>3_neon): Likewise.
7583 (orndi3_neon): Likewise.
7584 (bic<mode>3_neon): Likewise.
7585 (bicdi3_neon): Likewise.
7586 (xor<mode>3): Likewise.
7587 (one_cmpl<mode>2): Likewise.
7588 (abs<mode>2): Likewise.
7589 (neg<mode>2): Likewise.
7590 (umin<mode>3_neon): Likewise.
7591 (umax<mode>3_neon): Likewise.
7592 (smin<mode>3_neon): Likewise.
7593 (smax<mode>3_neon): Likewise.
7594 (vashl<mode>3): Likewise.
7595 (vashr<mode>3_imm): Likewise.
7596 (vlshr<mode>3_imm): Likewise.
7597 (ashl<mode>3_signed): Likewise.
7598 (ashl<mode>3_unsigned): Likewise.
7599 (neon_load_count): Likewise.
7600 (ashldi3_neon_noclobber): Likewise.
7601 (signed_shift_di3_neon): Likewise.
7602 (unsigned_shift_di3_neon): Likewise.
7603 (ashrdi3_neon_imm_noclobber): Likewise.
7604 (lshrdi3_neon_imm_noclobber): Likewise.
7605 (widen_ssum<mode>3): Likewise.
7606 (widen_usum<mode>3): Likewise.
7607 (quad_halves_<code>v4si): Likewise.
7608 (quad_halves_<code>v4sf): Likewise.
7609 (quad_halves_<code>v8hi): Likewise.
7610 (quad_halves_<code>v16qi): Likewise.
7611 (reduc_splus_v2di): Likewise.
7612 (neon_vpadd_internal<mode>): Likewise.
7613 (neon_vpsmin<mode>): Likewise.
7614 (neon_vpsmax<mode>): Likewise.
7615 (neon_vpumin<mode>): Likewise.
7616 (neon_vpumax<mode>): Likewise.
7617 (ss_add<mode>_neon): Likewise.
7618 (us_add<mode>_neon): Likewise.
7619 (ss_sub<mode>_neon): Likewise.
7620 (us_sub<mode>_neon): Likewise.
7621 (neon_vadd<mode>_unspec): Likewise.
7622 (neon_vaddl<mode>): Likewise.
7623 (neon_vaddw<mode>): Likewise.
7624 (neon_vhadd<mode>): Likewise.
7625 (neon_vqadd<mode>): Likewise.
7626 (neon_vaddhn<mode>): Likewise.
7627 (neon_vmul<mode>): Likewise.
7628 (neon_vmla<mode>): Likewise.
7629 (neon_vmlal<mode>): Likewise.
7630 (neon_vmls<mode>): Likewise.
7631 (neon_vmlsl<mode>): Likewise.
7632 (neon_vqdmulh<mode>): Likewise.
7633 (neon_vqdmlal<mode>): Likewise.
7634 (neon_vqdmlsl<mode>): Likewise.
7635 (neon_vmull<mode>): Likewise.
7636 (neon_vqdmull<mode>): Likewise.
7637 (neon_vsub<mode>_unspec): Likewise.
7638 (neon_vsubl<mode>): Likewise.
7639 (neon_vsubw<mode>): Likewise.
7640 (neon_vqsub<mode>): Likewise.
7641 (neon_vhsub<mode>): Likewise.
7642 (neon_vsubhn<mode>): Likewise.
7643 (neon_vceq<mode>): Likewise.
7644 (neon_vcge<mode>): Likewise.
7645 (neon_vcgeu<mode>): Likewise.
7646 (neon_vcgt<mode>): Likewise.
7647 (neon_vcgtu<mode>): Likewise.
7648 (neon_vcle<mode>): Likewise.
7649 (neon_vclt<mode>): Likewise.
7650 (neon_vcage<mode>): Likewise.
7651 (neon_vcagt<mode>): Likewise.
7652 (neon_vtst<mode>): Likewise.
7653 (neon_vabd<mode>): Likewise.
7654 (neon_vabdl<mode>): Likewise.
7655 (neon_vaba<mode>): Likewise.
7656 (neon_vabal<mode>): Likewise.
7657 (neon_vmax<mode>): Likewise.
7658 (neon_vmin<mode>): Likewise.
7659 (neon_vpaddl<mode>): Likewise.
7660 (neon_vpadal<mode>): Likewise.
7661 (neon_vpmax<mode>): Likewise.
7662 (neon_vpmin<mode>): Likewise.
7663 (neon_vrecps<mode>): Likewise.
7664 (neon_vrsqrts<mode>): Likewise.
7665 (neon_vqabs<mode>): Likewise.
7666 (neon_vqneg<mode>): Likewise.
7667 (neon_vcls<mode>): Likewise.
7668 (clz<mode>2): Likewise.
7669 (popcount<mode>2): Likewise.
7670 (neon_vrecpe): Likewise.
7671 (neon_vrsqrte): Likewise.
7672 (neon_vget_lane<mode>_sext_internal): Likewise.
7673 (neon_vget_lane<mode>_zext_internal): Likewise.
7674 (neon_vdup_n<mode>): Likewise.
7675 (neon_vdup_nv2di): Likewise.
7676 (neon_vdpu_lane<mode>_internal): Likewise.
7677 (neon_vswp<mode>): Likewise.
7678 (float<mode><V_cvtto>2): Likewise.
7679 (floatuns<mode><V_cvtto>2): Likewise.
7680 (fix_trunc<mode><V_cvtto>)2): Likewise
7681 (fixuns_trunc<mode><V_cvtto)2): Likewise.
7682 (neon_vcvt<mode>): Likewise.
7683 (neon_vcvtv4sfv4hf): Likewise.
7684 (neon_vcvtv4hfv4sf): Likewise.
7685 (neon_vcvt_n<mode>): Likewise.
7686 (neon_vmovn<mode>): Likewise.
7687 (neon_vqmovn<mode>): Likewise.
7688 (neon_vqmovun<mode>): Likewise.
7689 (neon_vmovl<mode>): Likewise.
7690 (neon_vmul_lane<mode>): Likewise.
7691 (neon_vmull_lane<mode>): Likewise.
7692 (neon_vqdmull_lane<mode>): Likewise.
7693 (neon_vqdmulh_lane<mode>): Likewise.
7694 (neon_vmla_lane<mode>): Likewise.
7695 (neon_vmlal_lane<mode>): Likewise.
7696 (neon_vqdmlal_lane<mode>): Likewise.
7697 (neon_vmls_lane<mode>): Likewise.
7698 (neon_vmlsl_lane<mode>): Likewise.
7699 (neon_vqdmlsl_lane<mode>): Likewise.
7700 (neon_vext<mode>): Likewise.
7701 (neon_vrev64<mode>): Likewise.
7702 (neon_vrev32<mode>): Likewise.
7703 (neon_vrev16<mode>): Likewise.
7704 (neon_vbsl<mode>_internal): Likewise.
7705 (neon_vshl<mode>): Likewise.
7706 (neon_vqshl<mode>): Likewise.
7707 (neon_vshr_n<mode>): Likewise.
7708 (neon_vshrn_n<mode>): Likewise.
7709 (neon_vqshrn_n<mode>): Likewise.
7710 (neon_vqshrun_n<mode>): Likewise.
7711 (neon_vshl_n<mode>): Likewise.
7712 (neon_vqshl_n<mode>): Likewise.
7713 (neon_vqshlu_n<mode>): Likewise.
7714 (neon_vshll_n<mode>): Likewise.
7715 (neon_vsra_n<mode>): Likewise.
7716 (neon_vsri_n<mode>): Likewise.
7717 (neon_vsli_n<mode>): Likewise.
7718 (neon_vtbl1v8qi): Likewise.
7719 (neon_vtbl2v8qi): Likewise.
7720 (neon_vtbl3v8qi): Likewise.
7721 (neon_vtbl4v8qi): Likewise.
7722 (neon_vtbx1v8qi): Likewise.
7723 (neon_vtbx2v8qi): Likewise.
7724 (neon_vtbx3v8qi): Likewise.
7725 (neon_vtbx4v8qi): Likewise.
7726 (neon_vtrn<mode>_internal): Likewise.
7727 (neon_vzip<mode>_internal): Likewise.
7728 (neon_vuzp<mode>_internal): Likewise.
7729 (neon_vld1<mode>): Likewise.
7730 (neon_vld1_lane<mode>): Likewise.
7731 (neon_vld1_dup<mode>): Likewise.
7732 (neon_vld1_dupv2di): Likewise.
7733 (neon_vst1<mode>): Likewise.
7734 (neon_vst1_lane<mode>): Likewise.
7735 (neon_vld2<mode>): Likewise.
7736 (neon_vld2_lane<mode>): Likewise.
7737 (neon_vld2_dup<mode>): Likewise.
7738 (neon_vst2<mode>): Likewise.
7739 (neon_vst2_lane<mode>): Likewise.
7740 (neon_vld3<mode>): Likewise.
7741 (neon_vld3qa<mode>): Likewise.
7742 (neon_vld3qb<mode>): Likewise.
7743 (neon_vld3_lane<mode>): Likewise.
7744 (neon_vld3_dup<mode>): Likewise.
7745 (neon_vst3<mode>): Likewise.
7746 (neon_vst3qa<mode>): Likewise.
7747 (neon_vst3qb<mode>): Likewise.
7748 (neon_vst3_lane<mode>): Likewise.
7749 (neon_vld4<mode>): Likewise.
7750 (neon_vld4qa<mode>): Likewise.
7751 (neon_vld4qb<mode>): Likewise.
7752 (neon_vld4_lane<mode>): Likewise.
7753 (neon_vld4_dup<mode>): Likewise.
7754 (neon_vst4<mode>): Likewise.
7755 (neon_vst4qa<mode>): Likewise.
7756 (neon_vst4qb<mode>): Likewise.
7757 (neon_vst4_lane<mode>): Likewise.
7758 (neon_vec_unpack<US>_lo_<mode>): Likewise.
7759 (neon_vec_unpack<US>_hi_<mode>): Likewise.
7760 (neon_vec_<US>mult_lo_<mode>): Likewise.
7761 (neon_vec_<US>mult_hi_<mode>): Likewise.
7762 (neon_vec_<US>shiftl_<mode>): Likewise.
7763 (neon_unpack<US>_<mode>): Likewise.
7764 (neon_vec_<US>mult_<mode>): Likewise.
7765 (vec_pack_trunc_<mode>): Likewise.
7766 (neon_vec_pack_trunk_<mode>): Likewise.
7767 (neon_vabd<mode>_2): Likewise.
7768 (neon_vabd<mode>_3): Likewise.
7769 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
7770 (thumb2_movsi_vfp): Likewise.
7771 (movdi_vfp): Likewise.
7772 (movdi_vfp_cortexa8): Likewise.
7773 (movhf_vfp_neon): Likewise.
7774 (movhf_vfp): Likewiwse.
7775 (movsf_vfp): Likewiwse.
7776 (thumb2_movsf_vfp): Likewiwse.
7777 (movdf_vfp): Likewise.
7778 (thumb2_movdf_vfp): Likewise.
7779 (movsfcc_vfp): Likewise.
7780 (thumb2_movsfcc_vfp): Likewise.
7781 (movdfcc_vfp): Likewise.
7782 (thumb2_movdfcc_vfp): Likewise.
7783 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
7784 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
7785 (v10_v2c): Likewise.
7786 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
7788 (cortex_a15_neon_int_2): Likewise.
7789 (cortex_a15_neon_int_3): Likewise.
7790 (cortex_a15_neon_int_4): Likewise.
7791 (cortex_a15_neon_int_5): Likewise.
7792 (cortex_a15_neon_vqneg_vqabs): Likewise.
7793 (cortex_a15_neon_vmov): Likewise.
7794 (cortex_a15_neon_vaba): Likewise.
7795 (cortex_a15_neon_vaba_qqq): Likewise.
7796 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7797 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7798 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
7799 scalar_64_32_long_scalar): Likewise.
7800 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7801 (cortex_a15_neon_mla_qqq_8_16): Likewise.
7802 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7803 lotype_qdd_64_32_long): Likewise.
7804 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7805 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7806 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
7807 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7808 (cortex_a15_neon_shift_1): Likewise.
7809 (cortex_a15_neon_shift_2): Likewise.
7810 (cortex_a15_neon_shift_3): Likewise.
7811 (cortex_a15_neon_vshl_ddd): Likewise.
7812 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7813 (cortex_a15_neon_vsra_vrsra): Likewise.
7814 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
7815 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
7816 (cortex_a15_neon_fp_vmul_ddd): Likewise.
7817 (cortex_a15_neon_fp_vmul_qqd): Likewise.
7818 (cortex_a15_neon_fp_vmla_ddd): Likewise.
7819 (cortex_a15_neon_fp_vmla_qqq): Likewise.
7820 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
7821 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
7822 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7823 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7824 (cortex_a15_neon_bp_simple): Likewise.
7825 (cortex_a15_neon_bp_2cycle): Likewise.
7826 (cortex_a15_neon_bp_3cycle): Likewise.
7827 (cortex_a15_neon_vld1_1_2_regs): Likewise.
7828 (cortex_a15_neon_vld1_3_4_regs): Likewise.
7829 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7830 (cortex_a15_neon_vld2_4_regs): Likewise.
7831 (cortex_a15_neon_vld3_vld4): Likewise.
7832 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7833 (cortex_a15_neon_vst1_3_4_regs): Likewise.
7834 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
7835 (cortex_a15_neon_vst3_vst4): Likewise.
7836 (cortex_a15_neon_vld1_vld2_lane): Likewise.
7837 (cortex_a15_neon_vld3_vld4_lane" 10
7838 (cortex_a15_neon_vst1_vst2_lane): Likewise.
7839 (cortex_a15_neon_vst3_vst4_lane): Likewise.
7840 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
7841 (cortex_a15_neon_ldm_2): Likewise.0
7842 (cortex_a15_neon_stm_2): Likewise.
7843 (cortex_a15_neon_mcr): Likewise.
7844 (cortex_a15_neon_mcr_2_mcrr): Likewise.
7845 (cortex_a15_neon_mrc): Likewise.
7846 (cortex_a15_neon_mrrc): Likewise.
7847 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
7849 (cortex_a15_alu_shift): Likewise.
7850 (cortex_a15_alu_shift_reg): Likewise.
7851 (cortex_a15_mult32): Likewise.
7852 (cortex_a15_mult64): Likewise.
7853 (cortex_a15_block): Likewise.
7854 (cortex_a15_branch): Likewise.
7855 (cortex_a15_load1): Likewise.
7856 (cortex_a15_load3): Likewise.
7857 (cortex_a15_store1): Likewise.
7858 (cortex_a15_store3): Likewise.
7859 (cortex_a15_call): Likewise.
7860 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
7861 (cortex_a5_f2r): Likewise.
7862 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
7864 (cortex_a53_f2r): Likewise.
7865 * config/arm/cortex-a7.md
7866 (cortex_a7_branch): Update for attribute change.
7867 (cortex_a7_call): Likewise.
7868 (cortex_a7_alu_imm): Likewise.
7869 (cortex_a7_alu_reg): Likewise.
7870 (cortex_a7_alu_shift): Likewise.
7871 (cortex_a7_mul): Likewise.
7872 (cortex_a7_load1): Likewise.
7873 (cortex_a7_store1): Likewise.
7874 (cortex_a7_load2): Likewise.
7875 (cortex_a7_store2): Likewise.
7876 (cortex_a7_load3): Likewise.
7877 (cortex_a7_store3): Likewise.
7878 (cortex_a7_load4): Likewise.
7879 (cortex_a7_store4): Likewise.
7880 (cortex_a7_fpalu): Likewise.
7881 (cortex_a7_fconst): Likewise.
7882 (cortex_a7_fpmuls): Likewise.
7883 (cortex_a7_neon_mul): Likewise.
7884 (cortex_a7_fpmacs): Likewise.
7885 (cortex_a7_neon_mla: Likewise.
7886 (cortex_a7_fpmuld: Likewise.
7887 (cortex_a7_fpmacd: Likewise.
7888 (cortex_a7_fpfmad: Likewise.
7889 (cortex_a7_fdivs: Likewise.
7890 (cortex_a7_fdivd: Likewise.
7891 (cortex_a7_r2f: Likewise.
7892 (cortex_a7_f2r: Likewise.
7893 (cortex_a7_f_flags: Likewise.
7894 (cortex_a7_f_loads: Likewise.
7895 (cortex_a7_f_loadd: Likewise.
7896 (cortex_a7_f_stores: Likewise.
7897 (cortex_a7_f_stored: Likewise.
7898 (cortex_a7_neon): Likewise.
7899 * config/arm/cortex-a8-neon.md
7900 (cortex_a8_neon_mrc): Update for attribute change.
7901 (cortex_a8_neon_mrrc): Likewise.
7902 (cortex_a8_neon_int_1): Likewise.
7903 (cortex_a8_neon_int_2): Likewise.
7904 (cortex_a8_neon_int_3): Likewise.
7905 (cortex_a8_neon_int_4): Likewise.
7906 (cortex_a8_neon_int_5): Likewise.
7907 (cortex_a8_neon_vqneg_vqabs): Likewise.
7908 (cortex_a8_neon_vmov): Likewise.
7909 (cortex_a8_neon_vaba): Likewise.
7910 (cortex_a8_neon_vaba_qqq): Likewise.
7911 (cortex_a8_neon_vsma): Likewise.
7912 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7913 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7914 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7915 long_scalar): Likewise.
7916 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7917 (cortex_a8_neon_mla_qqq_8_16): Likewise.
7918 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7919 long_scalar_qdd_64_32_long): Likewise.
7920 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7921 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7922 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
7923 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7924 (cortex_a8_neon_shift_1): Likewise.
7925 (cortex_a8_neon_shift_2): Likewise.
7926 (cortex_a8_neon_shift_3): Likewise.
7927 (cortex_a8_neon_vshl_ddd): Likewise.
7928 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7929 (cortex_a8_neon_vsra_vrsra): Likewise.
7930 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
7931 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
7932 (cortex_a8_neon_fp_vsum): Likewise.
7933 (cortex_a8_neon_fp_vmul_ddd): Likewise.
7934 (cortex_a8_neon_fp_vmul_qqd): Likewise.
7935 (cortex_a8_neon_fp_vmla_ddd): Likewise.
7936 (cortex_a8_neon_fp_vmla_qqq): Likewise.
7937 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
7938 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
7939 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
7940 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
7941 (cortex_a8_neon_bp_simple): Likewise.
7942 (cortex_a8_neon_bp_2cycle): Likewise.
7943 (cortex_a8_neon_bp_3cycle): Likewise.
7944 (cortex_a8_neon_ldr): Likewise.
7945 (cortex_a8_neon_str): Likewise.
7946 (cortex_a8_neon_vld1_1_2_regs): Likewise.
7947 (cortex_a8_neon_vld1_3_4_regs): Likewise.
7948 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
7949 (cortex_a8_neon_vld2_4_regs): Likewise.
7950 (cortex_a8_neon_vld3_vld4): Likewise.
7951 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
7952 (cortex_a8_neon_vst1_3_4_regs): Likewise.
7953 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
7954 (cortex_a8_neon_vst3_vst4): Likewise.
7955 (cortex_a8_neon_vld1_vld2_lane): Likewise.
7956 (cortex_a8_neon_vld3_vld4_lane): Likewise.
7957 (cortex_a8_neon_vst1_vst2_lane): Likewise.
7958 (cortex_a8_neon_vst3_vst4_lane): Likewise.
7959 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
7960 (cortex_a8_neon_mcr): Likewise.
7961 (cortex_a8_neon_mcr_2_mcrr): Likewise.
7962 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
7963 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
7965 (ca9_neon_mrrc): Likewise.
7966 (cortex_a9_neon_int_1): Likewise.
7967 (cortex_a9_neon_int_2): Likewise.
7968 (cortex_a9_neon_int_3): Likewise.
7969 (cortex_a9_neon_int_4): Likewise.
7970 (cortex_a9_neon_int_5): Likewise.
7971 (cortex_a9_neon_vqneg_vqabs): Likewise.
7972 (cortex_a9_neon_vmov): Likewise.
7973 (cortex_a9_neon_vaba): Likewise.
7974 (cortex_a9_neon_vaba_qqq): Likewise.
7975 (cortex_a9_neon_vsma): Likewise.
7976 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7977 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
7978 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
7979 long_scalar): Likewise.
7980 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
7981 (cortex_a9_neon_mla_qqq_8_16): Likewise.
7982 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
7983 long_scalar_qdd_64_32_long): Likewise.
7984 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
7985 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
7986 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
7987 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
7988 (cortex_a9_neon_shift_1): Likewise.
7989 (cortex_a9_neon_shift_2): Likewise.
7990 (cortex_a9_neon_shift_3): Likewise.
7991 (cortex_a9_neon_vshl_ddd): Likewise.
7992 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
7993 (cortex_a9_neon_vsra_vrsra): Likewise.
7994 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
7995 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
7996 (cortex_a9_neon_fp_vsum): Likewise.
7997 (cortex_a9_neon_fp_vmul_ddd): Likewise.
7998 (cortex_a9_neon_fp_vmul_qqd): Likewise.
7999 (cortex_a9_neon_fp_vmla_ddd): Likewise.
8000 (cortex_a9_neon_fp_vmla_qqq): Likewise.
8001 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
8002 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
8003 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
8004 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
8005 (cortex_a9_neon_bp_simple): Likewise.
8006 (cortex_a9_neon_bp_2cycle): Likewise.
8007 (cortex_a9_neon_bp_3cycle): Likewise.
8008 (cortex_a9_neon_ldr): Likewise.
8009 (cortex_a9_neon_str): Likewise.
8010 (cortex_a9_neon_vld1_1_2_regs): Likewise.
8011 (cortex_a9_neon_vld1_3_4_regs): Likewise.
8012 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
8013 (cortex_a9_neon_vld2_4_regs): Likewise.
8014 (cortex_a9_neon_vld3_vld4): Likewise.
8015 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
8016 (cortex_a9_neon_vst1_3_4_regs): Likewise.
8017 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
8018 (cortex_a9_neon_vst3_vst4): Likewise.
8019 (cortex_a9_neon_vld1_vld2_lane): Likewise.
8020 (cortex_a9_neon_vld3_vld4_lane): Likewise.
8021 (cortex_a9_neon_vst1_vst2_lane): Likewise.
8022 (cortex_a9_neon_vst3_vst4_lane): Likewise.
8023 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
8024 (cortex_a9_neon_mcr): Likewise.
8025 (cortex_a9_neon_mcr_2_mcrr): Likewise.
8026 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
8027 (cortex_a9_fps): Likewise.
8028 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
8030 (cortex_m4_fmuls): Likewise.
8031 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
8033 (cortex_r4_mrc): Likewise.
8034 * config/arm/iterators.md: Update comment referring to neon_type.
8035 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
8036 (iwmmxt_movsi_insn): Likewise.
8037 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
8039 (pj4_core_to_vfp): Likewise.
8040 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
8042 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
8043 (vfp_fstore): Likewise.
8044 * doc/md.texi: Change references to neon_type to refer to type.
8046 2013-09-04 Dodji Seketeli <dodji@redhat.com>
8048 * tree.h (DECL_BUILT_IN): Fix typo in comment.
8050 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
8052 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
8055 2013-09-04 Easwaran Raman <eraman@google.com>
8058 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
8059 (build_and_add_sum): Use it.
8060 (appears_later_in_bb): Simplify code.
8062 2013-09-04 Teresa Johnson <tejohnson@google.com>
8064 * dumpfile.c (dump_finish): Don't close stderr/stdout.
8066 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
8068 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
8070 2013-09-04 Jan Hubicka <jh@suse.cz>
8072 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
8073 * ipa-devirt.c: Include diganostic.h
8074 (odr_type_d): Add types and types_set.
8075 (hash_type_name): Work for types with vtables during LTO.
8076 (odr_hasher::remove): Fix comment; destroy types_set.
8077 (add_type_duplicate): New function,
8078 (get_odr_type): Use it.
8079 (dump_type_inheritance_graph): Dump type duplicates.
8080 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
8082 * tree.c (types_same_for_odr): Give exact answers on types with
8085 2013-09-04 Dodji Seketeli <dodji@redhat.com>
8087 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
8088 explaining their differences.
8090 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
8092 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
8094 2013-09-03 Jeff Law <law@redhat.com>
8096 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
8097 when not threading through a joiner block. Pass joiner/no joiner
8098 state to register_jump_thread.
8099 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
8100 state from argument rather than implying on path length.
8101 Dump the entire jump thread path into debugging dump.
8102 * tree-flow.h (register_jump_thread): Update prototype.
8104 2013-08-29 Xinliang David Li <davidxl@google.com>
8106 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
8107 Remove a trivial gcc_assert.
8109 2013-08-29 Xinliang David Li <davidxl@google.com>
8111 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
8112 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
8113 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
8114 Delay base decl alignment adjustment.
8115 * tree-vectorizer.c (vect_destroy_datarefs): New function.
8116 * tree-vectorizer.h: New data structure.
8117 (set_dr_misalignment): New function.
8118 (dr_misalignment): Ditto.
8119 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
8120 (vectorizable_load): Ditto.
8121 (ensure_base_align): New function.
8122 (vectorize_loops): Add dbg_cnt support.
8123 (execute_vect_slp): Ditto.
8124 * dbgcnt.def: New debug counter.
8125 * Makefile: New dependency.
8127 2013-09-03 Meador Inge <meadori@codesourcery.com>
8131 2013-08-30 Meador Inge <meadori@codesourcery.com>
8133 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8135 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
8137 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
8138 function descriptor.
8140 2013-09-03 Richard Biener <rguenther@suse.de>
8142 * tree-affine.c (add_elt_to_tree): Fix association issue,
8143 avoid useless converts and make sure to always return a
8144 properly typed result.
8146 2013-09-03 Richard Biener <rguenther@suse.de>
8149 * fold-const.c (negate_expr_p): Fix division case.
8150 (negate_expr): Likewise.
8152 2013-09-03 Richard Biener <rguenther@suse.de>
8155 * tree-streamer-out.c: Include tm.h.
8156 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
8158 2013-09-03 Jan Hubicka <jh@suse.cz>
8160 * tree-profile.c (tree_profiling): Cleanup CFG when done.
8162 2013-09-03 Alan Modra <amodra@gmail.com>
8164 * config.gcc (powerpc*-*-linux*): Add support for little-endian
8165 multilibs to big-endian target and vice versa.
8166 * config/rs6000/t-linux64: Use := assignment on all vars.
8167 (MULTILIB_EXTRA_OPTS): Remove fPIC.
8168 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
8169 * config/rs6000/t-linux64le: New file.
8170 * config/rs6000/t-linux64bele: New file.
8171 * config/rs6000/t-linux64lebe: New file.
8173 2013-09-02 Jan Hubicka <jh@suse.cz>
8175 * ipa-inline-transform.c (inline_transform): Do not
8176 optimize_inline_calls when not optimizing.
8178 2013-09-02 Jan Hubicka <jh@suse.cz>
8180 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
8181 duplicated nodes for assembler names.
8182 * symtab.c (symtab_unregister_node): Do not attempt to unlink
8183 hard registers from assembler name hash.
8185 2013-09-02 Jan Hubicka <jh@suse.cz>
8187 * ipa-split.c (execute_split_functions): Split externally visible
8188 functions called once.
8190 2013-09-02 Martin Jambor <mjambor@suse.cz>
8193 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
8194 linked list. When finding the correct duplicate, also consider also
8195 the caller in additon to its inlined_to node.
8197 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
8199 * config/aarch64/aarch64-simd-builtins.def
8200 (dup_lane_scalar): Remove.
8201 * config/aarch64/aarch64-simd.md
8202 (aarch64_simd_dup): Add 'w->w' alternative.
8203 (aarch64_dup_lane<mode>): Allow for VALL.
8204 (aarch64_dup_lane_scalar<mode>): Remove.
8205 (aarch64_dup_lane_<vswap_width_name><mode>): New.
8206 (aarch64_get_lane_signed<mode>): Add w->w altenative.
8207 (aarch64_get_lane_unsigned<mode>): Likewise.
8208 (aarch64_get_lane<mode>): Likewise.
8209 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
8210 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
8211 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
8212 (VCON): Change container of V2SF.
8213 (vswap_width_name): Likewise.
8214 * config/aarch64/arm_neon.h
8215 (__aarch64_vdup_lane_any): New.
8216 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8217 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
8218 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
8220 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
8223 * expr.c (emit_move_complex): Do not move complex FP values as parts if
8224 the source or the destination is a single hard register.
8226 2013-09-02 Richard Biener <rguenther@suse.de>
8229 * tree-scalar-evolution.c (instantiate_scev_name): Allow
8232 2013-09-02 Richard Biener <rguenther@suse.de>
8234 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
8235 arithmetic to sizetype.
8237 2013-09-02 Bin Cheng <bin.cheng@arm.com>
8239 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
8240 Find auto-increment use both before and after candidate.
8242 2013-09-02 Marek Polacek <polacek@redhat.com>
8244 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
8246 2013-09-01 Jan Hubicka <jh@suse.cz>
8248 * Makefile.in: Add ipa-profile.o
8249 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
8250 * cgraph.c (struct cgraph_propagate_frequency_data,
8251 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
8252 ipa-profile.c; replace cgraph_ by ipa_ prefix.
8253 * cgraph.h (cgraph_propagate_frequency): Remove.
8254 * ipa-inline-analysis.c: Include ipa-utils.h;
8255 drop duplicated cfgloop.h.
8256 (inline_update_callee_summaries): Update.
8257 * ipa-profile.c: New file.
8258 * ipa-utils.h (ipa_propagate_frequency): Declare.
8259 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
8260 data-streamer.h, value-prof.h.
8261 (symtab_remove_unreachable_nodes): Update profile.
8262 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
8263 account_time_size, cmp_counts, dump_histogram,
8264 ipa_profile_generate_summary, ipa_profile_write_summary,
8265 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
8266 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
8267 Move to ipa-profile.c.
8269 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
8271 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
8273 2013-09-01 Jan Hubicka <jh@suse.cz>
8275 * common.opt (fdevirtualize-speculatively): New function.
8276 * invoke.texi (fdevirtualize-speculatively): Document.
8277 * ipa-devirt.c: Include ipa-inline.h
8278 (likely_target_p): New function.
8279 (ipa_devirt): New function.
8280 (gate_ipa_devirt): New function.
8281 (pass_data_ipa_devirt): New static var.
8282 (pass_ipa_devirt): Likewise.
8283 (make_pass_ipa_devirt): New function.
8284 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
8285 (common_handle_option): Disable devirtualization when
8286 value range profiling is available.
8287 * passes.def (pass_ipa_devirt): Add.
8288 * timever.def (TV_IPA_DEVIRT): New timevar.
8289 * tree-pass.h (make_pass_ipa_devirt):
8291 2013-09-01 Iain Sandoe <iain@codesourcery.com>
8293 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
8294 include sanitize(undefined).
8296 2013-08-31 Diego Novillo <dnovillo@google.com>
8298 * Makefile.in (TREE_CORE_H): Define.
8300 (GTFILES): Add tree-core.h.
8301 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
8303 * tree-core.h: New file.
8304 Move all data structures, enum, typedefs, global
8305 declarations and constants from ...
8308 2013-08-31 Jan Hubicka <jh@suse.cz>
8310 * bulitins.c (expand_builtin): Do not early exit for gcov
8311 instrumented functions.
8313 2013-08-31 Marek Polacek <polacek@redhat.com>
8315 * ubsan.c: Include tm_p.h.
8317 2013-08-31 Jan Hubicka <jh@suse.cz>
8319 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
8322 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
8323 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
8324 * tree-cfg.c (verify_gimple_label): ... here.
8325 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
8326 (ipa_merge_profiles): New function.
8327 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
8328 (lto_input_function_body): Likewise.
8329 * ipa-utils.h (ipa_merge_profiles): Declare.
8330 * lto-streamer.h (lto_input_function_body): Update prototype.
8331 (emit_label_in_global_context_p): Remove.
8332 * lto-symtab.c: Include ipa-utils.h
8333 (lto_cgraph_replace_node): Use ipa_merge_profiles.
8335 2013-08-31 Jan Hubicka <jh@suse.cz>
8337 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
8339 2013-08-31 Jan Hubicka <jh@suse.cz>
8341 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
8343 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
8345 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
8346 "cmp" RTX before signed_comparison_operator check to account
8349 2013-08-30 Jan Hubicka <jh@suse.cz>
8351 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
8352 (detect_type_change_1): Rename to ...
8353 (detect_type_change): ... this one; early return on non-polymorphic
8355 (detect_type_change_ssa): Add comp_type parameter; update
8356 use of detect_type_change.
8357 (compute_complex_assign_jump_func): Add param_type parameter;
8358 update use of detect_type_change_ssa.
8359 (compute_complex_ancestor_jump_func): Likewise.
8360 (ipa_get_callee_param_type): New function.
8361 (ipa_compute_jump_functions_for_edge): Compute parameter type;
8362 update calls to the jump function computation functions.
8364 2013-08-30 Teresa Johnson <tejohnson@google.com>
8365 Steven Bosscher <steven@gcc.gnu.org>
8367 * cfgrtl.c (fixup_new_cold_bb): New routine.
8368 (commit_edge_insertions): Invoke fixup_partitions.
8369 (find_partition_fixes): New routine.
8370 (fixup_partitions): Ditto.
8371 (verify_hot_cold_block_grouping): Update comments.
8372 (rtl_verify_edges): Invoke find_partition_fixes.
8373 (rtl_verify_bb_pointers): Update comments.
8374 (rtl_verify_bb_layout): Ditto.
8375 * basic-block.h (probably_never_executed_edge_p): Declare.
8376 (fixup_partitions): Ditto.
8377 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
8378 * bb-reorder.c (sanitize_hot_paths): New function.
8379 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
8381 * predict.c (probably_never_executed_edge_p): New routine.
8382 * cfg.c (check_bb_profile): Add partition insanity warnings.
8384 2013-08-30 Meador Inge <meadori@codesourcery.com>
8386 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
8388 2013-08-30 Marek Polacek <polacek@redhat.com>
8390 * Makefile.in (ubsan.o): Add.
8391 (c-family/c-ubsan.o): Add.
8392 (builtins.o): Add ubsan.h dependency.
8393 * ubsan.h: New file.
8394 * ubsan.c: New file.
8395 * common.opt: Add -fsanitize=undefined option.
8396 (flag_sanitize): Add variable.
8397 (fsanitize=): Add option. Add Driver.
8398 (fsanitize=thread): Remove option.
8399 (fsanitize=address): Likewise.
8400 (static-libubsan): New option.
8401 * doc/invoke.texi: Document the new flag and -static-libubsan.
8402 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
8403 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
8404 * builtin-attrs.def (ATTR_COLD): Define.
8405 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8406 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
8407 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
8408 * flag-types.h (sanitize_code): New enum.
8409 * opts.c (common_handle_option): Parse command line arguments
8410 of -fsanitize=. Add -fsanitize=unreachable option.
8411 * varasm.c (get_variable_section): Adjust.
8412 (assemble_noswitch_variable): Likewise.
8413 (assemble_variable): Likewise.
8414 (output_constant_def_contents): Likewise.
8415 (categorize_decl_for_section): Likewise.
8416 (place_block_symbol): Likewise.
8417 (output_object_block): Likewise.
8418 * builtins.def: Likewise.
8419 * toplev.c (compile_file): Likewise.
8420 (process_options): Likewise.
8421 * cppbuiltin.c: Likewise.
8422 * tsan.c (tsan_pass): Likewise.
8423 (tsan_gate): Likewise.
8424 (tsan_gate_O0): Likewise.
8425 * cfgexpand.c (partition_stack_vars): Likewise.
8426 (expand_stack_vars): Likewise.
8427 (defer_stack_allocation): Likewise.
8428 (expand_used_vars): Likewise.
8429 * cfgcleanup.c (old_insns_match_p): Likewise.
8430 * asan.c (asan_finish_file): Likewise.
8431 (asan_instrument): Likewise.
8432 (gate_asan): Likewise.
8433 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
8434 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
8435 (asan_global_struct): Use pointer_sized_int_node instead
8436 calling build_nonstandard_integer_type.
8437 (initialize_sanitizer_builtins): Likewise.
8438 (asan_finish_file): Likewise.
8439 * gcc.c: Document %{%:function(args):X}.
8440 (static_spec_functions): Add sanitize.
8441 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
8442 store funcval != NULL there.
8443 (do_spec_1): Adjust handle_spec_function caller.
8444 (handle_braces): Allow %:function(args) as condition.
8445 (sanitize_spec_function): New function.
8446 (ADD_STATIC_LIBUBSAN_LIBS): Define.
8447 (LIBUBSAN_SPEC): Likewise.
8448 (LIBUBSAN_EARLY_SPEC): Likewise.
8449 (SANITIZER_SPEC): Handle libubsan.
8450 (SANITIZER_EARLY_SPEC): Likewise.
8451 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
8452 instead of fsanitize=address.
8453 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
8454 instead of fsanitize=address*.
8455 * builtins.c: Include ubsan.h.
8456 (fold_builtin_0): Instrument __builtin_unreachable.
8457 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
8458 instead of flag_asan.
8459 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
8460 (pointer_sized_int_node): Define.
8461 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
8463 2013-08-30 Mike Stump <mikestump@comcast.net>
8465 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
8468 2013-08-29 Jan Hubicka <jh@suse.cz>
8470 * cgraph.c (cgraph_function_body_availability): Handle weakref
8472 * passes.def: Remove pass_fixup_cfg.
8473 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
8474 track when we need to remove functions.
8475 (gate_ipa_inline): Execute inlining always; add comment why.
8476 (pass_data_ipa_inline): Remove TODO_remove_functions.
8477 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
8478 do not produce summaries.
8479 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
8480 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
8481 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
8482 constant pool and vtable.
8484 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
8486 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
8487 New arm_neon.h's internal macros to specify 64-bit constants.
8488 Avoid using stdint.h's macros.
8490 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
8492 * recog.c (verify_changes): Verify that changes[i].old is non-zero
8493 before applying REG_P.
8495 2013-08-30 Jakub Jelinek <jakub@redhat.com>
8497 PR tree-optimization/58277
8498 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
8499 after seeing too many stmts with vdef in between dombb and current
8500 bb, invalidate everything.
8502 2013-08-30 Richard Biener <rguenther@suse.de>
8504 * fold-const.c (fold_single_bit_test): Fix overflow test.
8506 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
8508 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
8509 and which can live in a register, always retrieve the value on entry.
8510 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
8511 passed by invisible reference specially.
8512 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
8513 (vt_add_function_parameter): Correctly deal with a parameter passed by
8514 invisible reference.
8516 2013-08-30 Jan Hubicka <jh@suse.cz>
8518 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
8520 2013-08-30 Richard Biener <rguenther@suse.de>
8522 PR tree-optimization/58228
8523 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
8524 allow invariant loads in nested loop vectorization.
8526 2013-08-30 Richard Biener <rguenther@suse.de>
8528 PR tree-optimization/58223
8529 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
8530 (has_anti_or_output_dependence): ... this and adjust to also
8531 look for output dependences.
8532 (mark_nodes_having_upstream_mem_writes): Adjust.
8533 (rdg_flag_uses): Likewise.
8535 2013-08-30 Richard Biener <rguenther@suse.de>
8537 PR tree-optimization/58010
8538 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
8539 assert that we have a loop-closed PHI.
8541 2013-08-29 Jan Hubicka <jh@suse.cz>
8543 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
8544 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
8545 * lto-section-in.c (lto_free_function_in_decl_state): New function.
8546 (lto_free_function_in_decl_state_for_node): New function.
8548 2013-08-29 Xinliang David Li <davidxl@google.com>
8550 * loop-unroll.c (report_unroll_peel): Minor message change.
8551 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
8552 Emit alignment peeling message with default -fopt-info.
8553 (vect_loop_versioning): Emit loop version info message.
8554 * tree-vectorizer.c (vectorize_loops): Minor message change.
8555 (execute_vect_slp): Ditto.
8557 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
8559 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
8560 of the clone from the DECL_NAME of the original function.
8562 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
8564 * passes.c (register_pass): Add overload.
8565 * tree-pass.h (register_pass): Forward declare it. Add comment.
8567 2013-08-29 Jan Hubicka <jh@suse.cz>
8569 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
8570 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
8571 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
8572 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8573 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
8574 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
8575 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
8576 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
8578 2013-08-29 Teresa Johnson <tejohnson@google.com>
8580 * dumpfile.c (dump_loc): Output column number.
8581 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
8582 * doc/invoke.texi: Document optall -fopt-info flag.
8583 * profile.c (read_profile_edge_counts): Use new dump framework.
8584 (compute_branch_probabilities): Ditto.
8585 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
8586 when pass not in any opt group.
8587 * pass_manager.h (pass_manager::get_pass_profile): New method.
8588 * value-prof.c (check_counter): Use new dump framework.
8589 (check_ic_target): Ditto.
8590 * coverage.c (get_coverage_counts): Ditto.
8591 (coverage_init): Setup new dump framework.
8593 2013-08-29 Richard Biener <rguenther@suse.de>
8595 PR tree-optimization/58246
8596 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
8597 handle the dominance check inside a basic-block.
8599 2013-08-29 Richard Biener <rguenther@suse.de>
8602 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
8603 of default defs that appear in abnormal PHI nodes.
8605 2013-08-29 Richard Biener <rguenther@suse.de>
8607 PR tree-optimization/57685
8608 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
8609 single-use operands to avoid exponential complexity.
8611 2013-08-28 Dehao Chen <dehao@google.com>
8613 * ipa-inline.c (edge_badness): Fix integer underflow.
8615 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
8617 * gtm-builtins.def (_ITM_free): Declare leaf.
8619 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8622 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
8623 (*tls_local_dynamic_base_64_largepic): Likewise.
8624 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
8625 Remove predicate from call operand.
8626 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
8627 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
8629 2013-08-28 Jeff Law <law@redhat.com>
8631 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
8632 checks for the number of predecessors and successors allowed.
8633 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
8634 which require copying a joiner block if there is a request which
8635 is a subpath that requires no joiner block copying.
8637 2013-08-28 Jan Hubicka <jh@suse.cz>
8639 * lto-streamer-out.c (DFS_write_tree_body): Drop
8640 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
8641 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
8642 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
8643 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
8644 Do not read DECL_ERROR_ISSUED.
8645 (unpack_ts_decl_with_vis_value_fields): Do not read
8647 (lto_input_ts_binfo_tree_pointers): Do not read
8648 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
8649 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
8650 write DECL_ERROR_ISSUED..
8651 (pack_ts_decl_with_vis_value_fields): Do not write
8653 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
8654 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
8655 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
8656 * tree.h (tree_decl_common): Update comment.
8657 (DECL_ERROR_ISSUED): Remove.
8659 2013-08-28 Jakub Jelinek <jakub@redhat.com>
8662 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
8664 2013-08-28 Jan Hubicka <jh@suse.cz>
8666 * builtins.def (free): Declare leaf.
8668 2013-08-27 David Malcolm <dmalcolm@redhat.com>
8671 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
8672 * configure: Regenerate.
8674 2013-08-27 Martin Jambor <mjambor@suse.cz>
8676 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
8677 (ipa_ancestor_jf_data): Likewise.
8678 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
8679 (ipa_get_jf_pass_through_type_preserved): New function.
8680 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
8681 (ipa_get_jf_ancestor_type_preserved): New function.
8682 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
8683 (ipa_get_jf_ancestor_result): Likewise.
8684 (propagate_vals_accross_pass_through): Use
8685 ipa_get_jf_pass_through_result to do all the value mappings.
8686 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
8687 type_preserved flag.
8688 (ipa_set_jf_cst_copy): New function.
8689 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
8690 (ipa_set_jf_arith_pass_through): Likewise.
8691 (ipa_set_ancestor_jf): Likewise.
8692 (compute_complex_assign_jump_func): Set type_preserved instead of
8694 (ipa_compute_jump_functions_for_edge): Likewise.
8695 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
8696 (update_jump_functions_after_inlining): Update type_preserved.
8697 Explicitely create jump functions when combining one with pass_through.
8698 (ipa_write_jump_function): Stream the type_preserved flags.
8699 (ipa_read_jump_function): Likewise.
8701 2013-08-27 Jakub Jelinek <jakub@redhat.com>
8702 Aldy Hernandez <aldyh@redhat.com>
8704 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
8705 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
8706 * function.h (struct function): Add has_force_vect_loops and
8708 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
8709 * gimple.c (gimple_build_omp_critical): Add KIND argument and
8711 * gimple.def: Update CLAUSES comments.
8712 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
8713 (gimple_build_omp_for): Add argument to prototype.
8714 (gimple_omp_for_kind): New.
8715 (gimple_omp_for_set_kind): New.
8716 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
8717 GOVD_DATA_SHARE_CLASS.
8718 (enum omp_region_type): Add ORT_SIMD.
8719 (gimple_add_tmp_var): Handle ORT_SIMD.
8720 (gimplify_var_or_parm_decl): Same.
8721 (is_gimple_stmt): Same.
8722 (omp_firstprivatize_variable): Same.
8723 (omp_add_variable): Only use splay_tree_insert if lookup failed.
8724 (omp_notice_variable): Handle ORT_SIMD.
8725 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
8726 (omp_check_private): Handle ORT_SIMD.
8727 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8729 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
8730 Handle OMP_CLAUSE_LASTPRIVATE.
8731 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
8733 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
8734 (gimplify_expr): Handle OMP_SIMD.
8735 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
8736 (expand_GOMP_SIMD_VF): New.
8737 (expand_GOMP_SIMD_LAST_LANE): New.
8738 * internal-fn.def (GOMP_SIMD_LANE): New.
8739 (GOMP_SIMD_VF): New.
8740 (GOMP_SIMD_LAST_LANE): New.
8741 * omp-low.c: Include target.h.
8742 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
8744 (check_omp_nesting_restrictions): Same.
8746 (lower_rec_simd_input_clauses): New.
8747 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
8749 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
8750 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
8751 (expand_omp_build_assign): New.
8752 (expand_omp_for_init_counts): New.
8753 (expand_omp_for_init_vars): New.
8754 (extract_omp_for_update_vars): New.
8755 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
8756 and rewrite accordingly.
8757 (expand_omp_simd): New.
8758 (expand_omp_for): Use expand_omp_simd.
8759 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
8760 (lower_omp_for): Do not lower the body.
8761 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
8763 * tree-flow.h (find_omp_clause): Remove prototype.
8764 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
8765 forcing vectorization of the loop, or if flag_tree_vectorize.
8766 (gate_tree_if_conversion): Similarly.
8767 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
8768 gimple_build_omp_for.
8769 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
8770 * tree-parloops (create_parallel_loop): Pass kind argument to
8771 gimple_build_omp_for.
8772 * tree-pretty-print.c (dump_omp_clause): Add cases for
8773 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
8774 OMP_CLAUSE__SIMDUID_.
8775 (dump_generic_node): Handle OMP_SIMD.
8776 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
8777 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
8778 unroll OMP_SIMD loops here.
8779 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
8780 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
8782 (vect_analyze_data_refs): Handle simd loops.
8783 * tree-vect-loop.c (vectorizable_live_operation): Handle
8785 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
8786 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
8787 (vectorizable_load): Same.
8788 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
8789 (struct simduid_to_vf): New.
8790 (simduid_to_vf::hash): New.
8791 (simduid_to-vf::equal): New.
8792 (struct simd_array_to_simduid): New.
8793 (simd_array_to_simduid::hash): New.
8794 (simd_array_to_simduid::equal): New.
8795 (adjust_simduid_builtins): New.
8796 (struct note_simd_array_uses_struct): New.
8797 (note_simd_array_uses_cb): New.
8798 (note_simd_array_uses): New.
8799 (vectorize_loops): Handle simd hints and adjust simd builtins
8801 * tree-vectorizer.h (struct _stmt_vec_info): Add
8802 simd_lane_access_p field.
8803 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
8804 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
8805 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
8806 (omp_clause_code_name): Same.
8807 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
8808 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
8809 * tree.def (OMP_SIMD): New entry.
8810 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
8811 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
8812 (OMP_CLAUSE_DECL): Adjust range for new clauses.
8813 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
8814 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
8815 (OMP_CLAUSE_LINEAR_STEP): New.
8816 (OMP_CLAUSE_SAFELEN_EXPR): New.
8817 (OMP_CLAUSE__SIMDUID__DECL): New.
8818 (find_omp_clause): New prototype.
8820 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
8822 * config/i386/driver-i386.c (host_detect_local_cpu): Update
8823 Haswell processor detection.
8825 2013-08-27 Christian Widmer <shadow@umbrox.de>
8828 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
8829 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
8830 AVX2 capable processors.
8832 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
8834 * config/aarch64/arm_neon.h: Replace all inline asm implementations
8835 of vget_low_* with implementations in terms of other intrinsics.
8837 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
8840 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
8841 values to -1, 0 and 1.
8843 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
8845 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
8846 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
8847 (<optab><mode>3_insn): Remove.
8848 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
8849 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
8850 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
8851 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
8852 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
8853 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
8854 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
8855 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
8856 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
8857 (ror<mode>3_insn): Likewise.
8858 * config/aarch64/predicates.md (aarch64_simd_register): New.
8860 2013-08-27 Richard Biener <rguenther@suse.de>
8862 PR tree-optimization/57521
8863 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
8864 one edge is non-critical.
8865 (find_phi_replacement_condition): Make sure to use a non-critical
8866 edge. Cleanup and remove old bug workarounds.
8867 (bb_postdominates_preds): Remove.
8868 (if_convertible_loop_p_1): Do not compute post-dominators.
8869 (combine_blocks): Do not free post-dominators.
8870 (main_tree_if_conversion): Likewise.
8871 (pass_data_if_conversion): Add TODO_verify_ssa.
8873 2013-08-27 DJ Delorie <dj@redhat.com>
8875 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
8877 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
8879 * function.c (assign_parm_find_data_types): Set passed_mode and
8880 nominal_mode to the TYPE_MODE of nominal_type for the built
8881 pointer type in case of the struct-pass-by-reference.
8883 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8885 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
8886 (UINT16_TYPE): Change default to "unsigned int".
8888 * config/avr/avr.opt (mfract-convert-truncate): New option.
8889 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
8890 is set, round negative fractional integers according to n1169
8891 when converting to integer types.
8893 2013-08-26 Jan Hubicka <jh@suse.cz>
8895 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
8896 methods can not be called indirectly when their address is not taken.
8898 2013-08-26 Jan Hubicka <jh@suse.cz>
8900 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
8903 2013-08-26 Jan Hubicka <jh@suse.cz>
8905 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
8908 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8910 * reload.c (find_valid_class): Allow classes that do not include
8911 FIRST_PSEUDO_REGISTER - 1.
8913 2013-08-26 Jan Hubicka <jh@suse.cz>
8915 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
8916 fix edge count/frequency when speculation failed; fix type check
8917 for the direct call.
8919 2013-08-26 Jan Hubicka <jh@suse.cz>
8921 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
8923 2013-08-26 Jan Hubicka <jh@suse.cz>
8925 * ipa-inline-transform.c (inline_transform): Be ready for basic block
8926 to be changed by edge redirection.
8928 2013-08-26 Jan Hubicka <jh@suse.cz>
8930 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
8931 formating; add sanity check.
8932 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
8933 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
8934 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
8935 (resolve_noninline_speculation): Update callee keys, too.
8937 2013-08-26 Jan Hubicka <jh@suse.cz>
8939 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
8940 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
8942 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
8944 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
8947 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
8949 * config/i386/i386.c (ix86_debug_options): Remove prototype.
8950 (x86_64_elf_select_section): Ditto.
8951 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
8953 (ix86_pass_by_reference): Ditto.
8954 (output_set_got): Ditto.
8955 (ix86_unary_operator_ok): Ditto.
8956 (ix86_expand_builtin): Ditto.
8958 2013-08-23 Jan Hubicka <jh@suse.cz>
8960 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
8962 2013-08-23 Jan Hubicka <jh@suse.cz>
8964 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
8965 (tree_decl_with_vis): Add FINAL field.
8967 2013-08-23 Jeff Law <law@redhat.com>
8969 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
8970 the debugging dump when the expression is fully redundant.
8972 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
8974 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
8975 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
8976 * pretty-print.c (pp_formatted_text_data): Likewise.
8977 (pp_write_text_to_stream): Likewise.
8978 (pp_write_text_as_dot_label_to_stream): Likewise.
8979 (pp_append_r): Likewise.
8980 (pp_format): Likewise.
8981 (pp_flush): Likewise.
8982 (pp_clear_output_area): Likewise.
8983 (pp_append_text): Likewise.
8984 (pp_formatted_text): Likewise.
8985 (pp_remaining_character_count_for_line): Likewise.
8986 (pp_newline): Likewise.
8987 (pp_character): Likewise.
8988 (output_buffer::~output_buffer): Define.
8989 (pretty_printer::~pretty_printer): Destruct output buffer.
8990 * pretty-print.h (output_buffer::~output_buffer): Declare.
8991 (pretty_printer::~pretty_printer): Declare virtual.
8993 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
8996 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
8997 HOST_WIDE_INT_M1U): New macros.
8998 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
8999 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
9000 unsigned -1 for lshift.
9001 * cse.c (cse_insn): Likewise.
9002 * double-int.c (rshift_double, lshift_double): Likewise.
9003 * builtins.c (fold_builtin_bitop): Likewise.
9004 * combine.c (force_to_mode): Likewise.
9005 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
9006 * simplify-rtx.c (simplify_const_unary_operation,
9007 simplify_const_binary_operation): Likewise.
9008 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
9009 check_va_list_escapes): Likewise.
9010 * rtlanal.c (nonzero_bits1): Likewise.
9011 * expmed.c (expand_smod_pow2): Likewise.
9012 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
9014 2013-08-23 Jan Hubicka <jh@suse.cz>
9016 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
9017 as having address taken.
9019 2013-08-23 Jan Hubicka <jh@suse.cz>
9021 * ipa-utils.h (method_class_type): Declare.
9022 * ipa-devirt.c (method_class_type): Export.
9024 * cgraphunit.c (analyze_functions): Do basic devirtualization;
9025 do not walk base classes of anonymous types.
9027 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
9029 PR rtl-optimization/58220
9031 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
9032 handle SEQUENCE insns properly.
9034 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
9036 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
9038 (pp_newline_and_indent): Likewise.
9039 (pp_separate_with): Likewise.
9040 * pretty-print.c (pp_newline_and_flush): Define.
9041 (pp_newline_and_indent): Likewise.
9042 (pp_separate_with): Likewise.
9044 2013-08-23 Jakub Jelinek <jakub@redhat.com>
9047 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
9048 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
9050 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
9052 * config/i386/predicates.md (ext_sse_reg_operand): New.
9053 * config/i386/i386.md (*movti_internal): Use
9054 predicate to determine if EVEX is needed.
9055 (*movsi_internal): Ditto.
9056 (*movdf_internal): Ditto.
9057 (*movsf_internal): Ditto.
9058 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
9060 2013-08-23 Jakub Jelinek <jakub@redhat.com>
9062 PR tree-optimization/58209
9063 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
9064 (find_tail_calls): Give up for pointer result types if m is non-NULL.
9065 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
9066 emit POINTER_PLUS_EXPR.
9067 (create_tailcall_accumulator): For pointer result type accumulate in
9070 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
9072 * configure.ac: Add backslashes missing from the last change.
9073 * configure: Regenerate.
9075 2013-08-22 Jan Hubicka <jh@susue.cz>
9077 * ipa.c (function_and_variable_visibility): First remember function
9078 was global and then make it local.
9080 2013-08-22 Julian Brown <julian@codesourcery.com>
9082 * configure.ac: Add aarch64 to list of arches which use "nop" in
9084 * configure: Regenerate.
9086 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9088 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
9089 gnu_libc_has_function.
9090 * config/s390/tpf.h: Likewise.
9092 2013-08-22 Jan Hubicka <jh@susue.cz>
9094 * timevar.c (validate_phases): Add cast.
9096 2013-08-22 Jan Hubicka <jh@susue.cz>
9098 * timevar.c (validate_phases): Use size_t for memory.
9099 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
9101 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
9103 * pretty-print.h (output_buffer::output_buffer): Declare.
9104 (pretty_printer::pretty_printer): Likewise.
9105 (pp_construct): Remove.
9106 * pretty-print.c (output_buffer::output_buffer): Define.
9107 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
9108 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
9109 (print_gimple_expr): Likewise.
9110 (print_gimple_seq): Likewise.
9111 (gimple_dump_bb): Likewise.
9112 * sched-vis.c (dump_value_slim): Likewise.
9113 (dump_insn_slim): Likewise.
9114 (dump_rtl_slim): Likewise.
9115 (str_pattern_slim): Likewise.
9116 * tree-mudflap.c (mf_varname_tree): Likewise.
9117 * graph.c (print_graph_cfg): Likewise.
9118 (start_graph_dump): Likewise.
9119 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
9121 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
9122 pretty printer initialization.
9123 * coretypes.h (diagnostic_context): Remove superflous type alias
9125 (pretty_printer): Likewise. Declare directly as a class.
9126 (pretty_print_info): Remove declaration as class.
9127 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
9128 and pp_clear_output_area.
9129 (asan_add_global): Likewise.
9131 2013-08-22 Jan Hubicka <jh@suse.cz>
9133 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
9134 * ipa-utils.h (update_type_inheritance_graph): Declare.
9135 (possible_polymorphic_call_target_p): Declare.
9136 (possible_polymorphic_call_target_p): New.
9137 * ipa-devirt.c: Update toplevel comments.
9138 (cached_polymorphic_call_targets): Move up.
9139 (odr_type_d): Move ID down.
9140 (polymorphic_type_binfo_p): Update comment.
9141 (odr_hasher::remove): Likewise;
9142 (get_odr_type): Set anonymous_namespace.
9143 (dump_odr_type): Dump it.
9144 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
9145 (maybe_record_node): Record node in cached_polymorphic_call_targets.
9146 (record_binfo): Add comment.
9147 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
9149 (devirt_node_removal_hook): Do not iCE when cache is freed.
9150 (possible_polymorphic_call_target_p): New predicate.
9151 (update_type_inheritance_graph): New function.
9153 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
9154 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9155 Sergey Lega <sergey.s.lega@intel.com>
9156 Anna Tikhonova <anna.tikhonova@intel.com>
9157 Ilya Tocar <ilya.tocar@intel.com>
9158 Andrey Turetskiy <andrey.turetskiy@intel.com>
9159 Ilya Verbin <ilya.verbin@intel.com>
9160 Kirill Yukhin <kirill.yukhin@intel.com>
9161 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9163 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
9164 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
9165 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
9166 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
9167 (OPTION_MASK_ISA_AVX2_UNSET): Update.
9168 (OPTION_MASK_ISA_AVX512F_UNSET): New.
9169 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
9170 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
9171 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
9172 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
9173 OPT_mavx512pf, OPT_mavx512er cases.
9174 * config/i386/constraints.md (v): New constraint.
9175 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
9176 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
9177 (bit_AVX512CD): New.
9178 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
9179 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
9180 * config/i386/i386-c.c (ix86_target_macros_internal):
9181 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
9183 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
9184 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
9185 * config/i386/i386.c (regclass_map, dbx_register_map)
9186 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
9187 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
9188 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
9190 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
9191 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
9192 -mavx512pf options. Fix formatting.
9193 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
9194 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
9195 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
9196 -mavx512cd, -mavx512pf options.
9197 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
9198 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
9199 (ix86_preferred_output_reload_class): Replace SSE_REGS with
9201 (ix86_hard_regno_mode_ok): Support 512-bit registers.
9202 (ix86_set_reg_reg_cost): Ditto.
9203 (x86_order_regs_for_local_alloc): Ditto.
9204 (MAX_VECT_LEN): Extend to 64-byte.
9205 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
9206 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
9207 (TARGET_AVX512ER, TARGET_AVX512CD): New.
9208 (BIGGEST_ALIGNMENT): Extend to 512-bits.
9209 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
9210 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
9211 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
9212 (SSE_REG_MODE_P): Support new modes.
9213 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
9214 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
9215 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
9216 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
9217 (REG_CLASS_CONTENTS): Add new registers.
9218 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
9219 (EXT_REX_SSE_REGNO_P): New.
9220 (HI_REGISTER_NAMES): Add new registers.
9221 * config/i386/i386.md: Define constants for new registers.
9222 (mode): Add new 512-bit modes.
9223 (prefix): Support evex prefix.
9224 (isa): Support avx512f, noavx512f, fma_avx512f.
9225 (ssemodesuffix): Add new 512-bit modes.
9227 (*movxi_internal_avx512f): Ditto.
9228 (*movdi_internal): Replace constraint "x" with the new constraint "v".
9230 (*movsi_internal): Likewise.
9231 (*movdf_internal): Likewise.
9232 (*movsf_internal): Likewise.
9233 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
9234 (<code><mode>3): Likewise.
9235 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
9237 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
9238 with the new constraint "v".
9239 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
9241 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
9242 with the new constraint "v".
9243 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
9244 (<sse2>_storedqu<avxsizesuffix>): Likewise.
9245 (*<plusminus_insn><mode>3): Likewise.
9246 (<sse>_vm<plusminus_insn><mode>3): Likewise.
9247 (*mul<mode>3): Likewise.
9248 (<sse>_vmmul<mode>3): Likewise.
9249 (<sse>_div<mode>3): Likewise.
9250 (<sse>_vmdiv<mode>3): Likewise.
9251 (<sse>_sqrt<mode>2): Likewise.
9252 (<sse>_vmsqrt<mode>2): Likewise.
9253 (*<code><mode>3_finite): Likewise.
9254 (*<code><mode>3) <smaxmin>: Likewise.
9255 (<sse>_vm<code><mode>3): Likewise.
9256 (*<code><mode>3) <any_logic>: Likewise.
9257 (*fma_fmadd_<mode>): Likewise.
9258 (*fma_fmsub_<mode>): Likewise.
9259 (*fma_fnmadd_<mode>): Likewise.
9260 (*fma_fnmsub_<mode>): Likewise.
9261 (*fma_fmaddsub_<mode>): Likewise.
9262 (*fma_fmsubadd_<mode>): Likewise.
9263 (*fmai_fmadd_<mode>): Likewise.
9264 (*fmai_fmsub_<mode>): Likewise.
9265 (*fmai_fnmadd_<mode>): Likewise.
9266 (*fmai_fnmsub_<mode>): Likewise.
9267 (sse_cvtsi2ss): Likewise.
9268 (sse_cvtsi2ssq): Likewise.
9269 (sse_cvtss2si): Likewise.
9270 (sse_cvtss2si_2): Likewise.
9271 (sse_cvtss2siq): Likewise.
9272 (sse_cvtss2siq_2): Likewise.
9273 (sse_cvttss2si): Likewise.
9274 (sse_cvtss2siq_2): Likewise.
9275 (float<sseintvecmodelower><mode>2): Likewise.
9276 (sse2_cvtsd2si_2): Likewise.
9277 (sse2_cvtsd2siq_2): Likewise.
9278 (*<plusminus_insn><mode>3): Likewise.
9279 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
9280 (*<sse4_1_avx2>_mul<mode>3): Likewise.
9281 (ashr<mode>3): Likewise.
9282 (<shift_insn><mode>3): Likewise.
9283 (avx2_<code><mode>3): Likewise.
9284 (*avx2_<code><mode>3): Likewise.
9285 (*andnot<mode>3): Likewise.
9286 (*<code><mode>3) <any_logic>: Likewise.
9287 (abs<mode>2): Likewise.
9288 (avx2_permvar<mode>): Likewise.
9289 (avx2_perm<mode>_1): Likewise.
9290 (*avx_vpermilp<mode>): Likewise.
9291 (avx_vpermilvar<mode>3): Likewise.
9292 (avx2_ashrv<mode>): Likewise.
9293 (avx2_<shift_insn>v<mode>): Likewise.
9294 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
9296 * doc/rtl.texi: Document XImode.
9298 2013-08-21 Jeff Law <law@redhat.com>
9300 * tree-flow.h (register_jump_thread): Pass vector of edges
9301 instead of each important edge.
9302 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
9303 thread path into a vector and pass that to register_jump_thread.
9304 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
9305 passed in edge vector to the current 3-edge form.
9308 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9310 * dce.c (fini_dce): Call df_analyze again just in case
9311 delete_unmarked_insns removed anything.
9313 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
9315 * reload.h (struct reg_equivs): Rename to ..
9316 (struct reg_equivs_s): .. this.
9318 2013-08-20 Martin Liska <marxin.liska@gmail.com>
9320 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
9322 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
9324 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
9326 2013-08-21 Jeff Law <law@redhat.com>
9328 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
9329 simplify assignments too. If the RHS collapses to a singleton
9330 range, then return the value for the range.
9332 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
9334 * config/i386/sse.md (V16): Rename to...
9336 (mov<mode>): Update iterator name.
9337 (*mov<mode>_internal): Ditto.
9338 (push<mode>1): Ditto.
9339 (movmisalign<mode>): Ditto.
9341 2013-08-20 Jan Hubicka <jh@suse.cz>
9344 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
9345 entry for direct edges.
9346 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
9348 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9350 Revert my last two changes, r201865 and r201864:
9353 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9355 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9356 instances can own GC refs.
9358 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9359 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9360 (gcc::context::gt_pch_nx): Likewise.
9361 (gcc::context::gt_pch_nx): Likewise.
9362 * ggc.h (gt_ggc_mx <T>): New.
9363 (gt_pch_nx_with_op <T>): New.
9364 (gt_pch_nx <T>): New.
9365 * passes.c (opt_pass::gt_ggc_mx): New.
9366 (opt_pass::gt_pch_nx): New.
9367 (opt_pass::gt_pch_nx_with_op): New.
9368 (pass_manager::gt_ggc_mx): New.
9369 (pass_manager::gt_pch_nx): New.
9370 (pass_manager::gt_pch_nx_with_op): New.
9371 (pass_manager::operator new): Use
9372 ggc_internal_cleared_alloc_stat rather than xcalloc.
9373 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9374 (pass_manager::gt_ggc_mx): New.
9375 (pass_manager::gt_pch_nx): New.
9376 (pass_manager::gt_pch_nx_with_op): New.
9377 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9378 (opt_pass::operator new): New.
9379 (opt_pass::gt_ggc_mx): New.
9380 (opt_pass::gt_pch_nx): New.
9381 (opt_pass::gt_pch_nx_with_op): New.
9384 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9386 * Makefile.in (GTFILES): Add context.h.
9387 * context.c (gcc::context::operator new): New.
9388 (gcc::context::gt_ggc_mx): New.
9389 (gcc::context::gt_pch_nx): New.
9390 (gcc::context::gt_pch_nx): New.
9391 * context.h (gcc::context): Add GTY((user)) marking.
9392 (gcc::context::operator new): New.
9393 (gcc::context::gt_ggc_mx): New.
9394 (gcc::context::gt_pch_nx): New.
9395 (gcc::context::gt_pch_nx): New.
9396 (g): Add GTY marking.
9397 (gt_ggc_mx (gcc::context *)): New.
9398 (gt_pch_nx (gcc::context *)): New.
9399 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9400 void *cookie)): New.
9401 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9403 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
9405 * dce.c (fini_dce): Call df_analyze again just in case
9406 delete_unmarked_insns removed anything.
9408 2013-08-20 Teresa Johnson <tejohnson@google.com>
9410 PR rtl-optimizations/57451
9411 * final.c (reemit_insn_block_notes): Prevent lexical blocks
9412 from crossing split section boundaries.
9414 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
9416 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
9417 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
9418 with MULTLIB_DEFAULTS.
9420 2013-08-20 Nick Clifton <nickc@redhat.com>
9422 * target.def (narrow_volatile_bitfield): Note that the default
9423 value is false, not !TARGET_STRICT_ALIGN.
9424 * doc/tm.texi: Regenerate.
9426 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
9428 Fix LIB_SPEC for systems without libpthread.
9430 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
9431 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
9433 * config/i386/linux-common.h: Likewise.
9434 * config/mips/linux-common.h: Likewise.
9436 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
9438 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
9441 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9443 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
9444 instances can own GC refs.
9446 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
9447 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
9448 (gcc::context::gt_pch_nx): Likewise.
9449 (gcc::context::gt_pch_nx): Likewise.
9450 * ggc.h (gt_ggc_mx <T>): New.
9451 (gt_pch_nx_with_op <T>): New.
9452 (gt_pch_nx <T>): New.
9453 * passes.c (opt_pass::gt_ggc_mx): New.
9454 (opt_pass::gt_pch_nx): New.
9455 (opt_pass::gt_pch_nx_with_op): New.
9456 (pass_manager::gt_ggc_mx): New.
9457 (pass_manager::gt_pch_nx): New.
9458 (pass_manager::gt_pch_nx_with_op): New.
9459 (pass_manager::operator new): Use
9460 ggc_internal_cleared_alloc_stat rather than xcalloc.
9461 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
9462 (pass_manager::gt_ggc_mx): New.
9463 (pass_manager::gt_pch_nx): New.
9464 (pass_manager::gt_pch_nx_with_op): New.
9465 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
9466 (opt_pass::operator new): New.
9467 (opt_pass::gt_ggc_mx): New.
9468 (opt_pass::gt_pch_nx): New.
9469 (opt_pass::gt_pch_nx_with_op): New.
9471 2013-08-20 David Malcolm <dmalcolm@redhat.com>
9473 * Makefile.in (GTFILES): Add context.h.
9474 * context.c (gcc::context::operator new): New.
9475 (gcc::context::gt_ggc_mx): New.
9476 (gcc::context::gt_pch_nx): New.
9477 (gcc::context::gt_pch_nx): New.
9478 * context.h (gcc::context): Add GTY((user)) marking.
9479 (gcc::context::operator new): New.
9480 (gcc::context::gt_ggc_mx): New.
9481 (gcc::context::gt_pch_nx): New.
9482 (gcc::context::gt_pch_nx): New.
9483 (g): Add GTY marking.
9484 (gt_ggc_mx (gcc::context *)): New.
9485 (gt_pch_nx (gcc::context *)): New.
9486 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
9487 void *cookie)): New.
9488 * gengtype.c (open_base_files) <ifiles>: Add context.h.
9490 2013-08-20 Alan Modra <amodra@gmail.com>
9493 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
9494 (rs6000_emit_epilogue): Likewise.
9496 2013-08-19 Dehao Chen <dehao@google.com>
9498 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
9500 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
9501 Jakub Jelinek <jakub@redhat.com>
9503 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
9504 (BUILT_IN_FABSD64): Likewise.
9505 (BUILT_IN_FABSD128): Likewise.
9506 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
9507 (fold_builtin_1): Likewise.
9508 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
9509 destination and source operands.
9510 (*abstd2_fpr): Likewise.
9511 (*nabstd2_fpr): Likewise.
9513 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
9515 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
9518 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
9520 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
9523 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
9525 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
9526 * builtins.c (default_libc_has_function): New.
9527 (gnu_libc_has_function): Ditto.
9528 (no_c99_libc_has_function): Ditto.
9529 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
9530 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
9531 (fold_builtin_sincos): Likewise.
9532 (fold_builtin_cexp): Likewise.
9533 * builtins.def (DEF_C94_BUILTIN): Likewise.
9534 (DEF_C99_BUILTIN): Likewise.
9535 (DEF_C99_C90RES_BUILTIN): Likewise.
9536 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
9537 definitions to using this define.
9538 * config/darwin-protos.h (darwin_libc_has_function): New.
9539 * config/darwin.c (darwin_libc_has_function): Ditto.
9540 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
9541 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
9542 * config/darwin.h: Ditto.
9543 * config/elfos.h: Ditto.
9544 * config/freebsd.h: Ditto.
9545 * config/i386/cygming.h: Ditto.
9546 * config/i386/djgpp.h: Ditto.
9547 * config/i386/i386-interix.h: Ditto.
9548 * config/microblaze/microblaze.h: Ditto.
9549 * config/mmix/mmix.h: Ditto.
9550 * config/gnu-user.h: Ditto.
9551 * config/ia64/hpux.h: Ditto.
9552 * config/pa/pa-hpux.h: Ditto.
9553 * config/pdp11/pdp11.h: Ditto.
9554 * config/picochip/picochip.h: Ditto.
9555 * config/linux.h: Ditto.
9556 * config/netbsd.h: Ditto.
9557 * config/openbsd.h: Ditto.
9558 * config/rs6000/aix43.h: Ditto.
9559 * config/rs6000/aix51.h: Ditto.
9560 * config/rs6000/aix52.h: Ditto.
9561 * config/rs6000/aix53.h: Ditto.
9562 * config/rs6000/aix61.h: Ditto.
9563 * config/rs6000/darwin.h: Ditto.
9564 * config/rs6000/linux.h: Ditto.
9565 * config/rs6000/linux64.h: Ditto.
9566 * config/s390/tpf.h: Ditto.
9567 * config/sol2-10.h: Ditto.
9568 * config/sol2.h: Ditto.
9569 * config/vms/vms.h: Ditto.
9570 * config/vxworks.h: Ditto.
9571 * config/linux-android.c (linux_android_libc_has_function):
9572 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
9573 * config/linux-protos.h (linux_android_libc_has_function):
9575 * config/i386/i386.c (ix86_libc_has_function): New.
9576 * config/i386/i386-protos.h
9577 (ix86_libc_has_function): New declaration.
9578 * config/i386/i386.md
9579 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
9580 ("isinf<mode>2): Likewise.
9581 * convert.c (convert_to_integer): Using new target hook
9582 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
9583 TARGET_C99_FUNCTIONS.
9584 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
9585 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
9586 * coretypes.h (function_class): New enum for different
9587 classes of functions.
9588 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
9589 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
9590 (TARGET_HAS_SINCOS): Likewise.
9591 (TARGET_LIBC_HAS_FUNCTION): New.
9592 * doc/tm.texi: Regenerated.
9593 * targhooks.h (default_libc_has_function): New declaration.
9594 (no_c99_libc_has_function): Ditto.
9595 (gnu_libc_has_function): Ditto.
9596 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
9597 and TARGET_HAS_SINCOS.
9599 2013-08-18 Jan Hubicka <jh@suse.cz>
9601 * Makeifle-in (ipa-devirt.o): New.
9602 (GTFILES): Add ipa-utils.h and ipa-devirt.c
9603 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
9604 (analyze_functions): Look into possible targets of polymorphic call.
9605 * dumpfile.c (dump_files): Add type-inheritance dump.
9606 * dumpfile.h (TDI_inheritance): New.
9607 * ipa-devirt.c: New file.
9608 * ipa-utils.h (odr_type_d): Forward declare.
9609 (odr_type): New type.
9610 (build_type_inheritance_graph): Declare.
9611 (possible_polymorphic_call_targets): Declare and introduce inline
9612 variant when only edge is pased.
9613 (dump_possible_polymorphic_call_targets): Likewise.
9614 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
9615 * tree.c (type_in_anonymous_namespace_p): Break out from ...
9616 (types_same_for_odr): ... here.
9617 * tree.h (type_in_anonymous_namespace_p): Declare.
9619 2013-08-18 Jakub Jelinek <jakub@redhat.com>
9621 PR tree-optimization/58006
9622 * tree-parloops.c (take_address_of): Don't ICE if get_name
9624 (eliminate_local_variables_stmt): Remove clobber stmts.
9626 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
9628 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
9629 error message is issued for an alias to undefined symbol.
9631 2013-08-18 Jan Hubicka <jh@suse.cz>
9633 * cgraph.c (cgraph_create_indirect_edge): Discover
9634 polymorphic calls and record basic info into indirect_info.
9635 * gimple-fold.c (gimple_fold_call): When doing BINFO based
9636 devirtualization, ignore objc function calls.
9637 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
9638 call with no parm index info.
9639 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
9640 * tree.c (virtual_method_call_p): New function.
9641 * tree.h (virtual_method_call_p): Declare.
9643 2013-08-16 Jan Hubicka <jh@suse.cz>
9646 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
9648 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
9650 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
9653 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9655 * gengtype.c (type_for_name): Add special-case support for
9656 locating types within the "gcc::" namespace.
9657 (open_base_files): Emit a "using namespace gcc" directive.
9659 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
9662 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
9663 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
9665 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
9666 array instead of each individual operand as a separate argument.
9667 (emit_fusion_gpr_load): Likewise.
9668 (expand_fusion_gpr_load): Add new function declaration.
9670 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
9671 signature to have the operands passed as an array, instead of as
9672 separate arguments. Allow ZERO_EXTEND to be in the memory
9673 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
9674 depend on the register live/dead flags when peepholes are run.
9675 (expand_fusion_gpr_load): New function to be called from the
9676 peephole2 pass, to change the register that addis sets to be the
9678 (emit_fusion_gpr_load): Change the calling signature to have the
9679 operands passed as an array, instead of as separate arguments.
9680 Allow ZERO_EXTEND to be in the memory address, and also
9681 SIGN_EXTEND if -mpower8-fusion-sign.
9683 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
9685 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
9686 adjust the register addis loads up in the peephole2 pass. Do not
9687 depend on the register live/dead state when the peephole pass is done.
9689 2013-08-16 David Malcolm <dmalcolm@redhat.com>
9691 * gengtype.c (create_user_defined_type): Ensure that the kind
9692 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
9693 declaration is seen before the GTY((user)) marking.
9695 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
9698 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
9700 2013-08-16 Jan Hubicka <jh@suse.cz>
9702 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
9703 arugment expected_type.
9704 (gimple_fold_call): Use it.
9705 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
9706 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
9707 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
9708 (try_make_edge_direct_virtual_call): Likewise.
9709 * tree.c (obj_type_ref_class): New.
9710 * tree.h (obj_type_ref_class): Use it.
9712 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
9714 * sched-vis.c (rtl_slim_pp_initialized): Remove.
9715 (rtl_slim_pp): Likewise.
9716 (init_rtl_slim_pretty_print): Likewise.
9717 (dump_value_slim): Don't call it. Use local pretty printer.
9718 (dump_insn_slim): Likewise.
9719 (dump_rtl_slim): Likewise.
9720 (str_pattern_slim): Likewise.
9721 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
9724 2013-08-16 Jakub Jelinek <jakub@redhat.com>
9726 PR tree-optimization/58164
9727 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
9728 walk gimple_goto_dest of GIMPLE_GOTO.
9730 PR tree-optimization/58165
9731 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
9732 bi_call must be the last stmt in a bb, don't split_block, instead
9733 use fallthru edge from it and give up if there is none.
9734 Release conds vector when returning early.
9736 2013-08-14 Xinliang David Li <davidxl@google.com>
9738 * config/i386/i386.c (ix86_option_override_internal):
9739 Remove unused variable and field.
9741 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
9744 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
9745 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
9746 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
9747 and Linux, correct BLKmode alignment when 128-bit alignment is
9748 required and compatibility flag is not set.
9749 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
9750 for zero-size arguments when compatibility flag is not set.
9752 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9754 PR tree-optimization/58145
9755 * tree-sra.c (build_ref_for_offset): If prev_base has
9756 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
9758 2013-08-14 Xinliang David Li <davidxl@google.com>
9760 * config/i386/i386.c (ix86_option_override_internal):
9761 Fix uninitialized variable error.
9763 2013-08-14 Xinliang David Li <davidxl@google.com>
9765 * config/i386/i386.opt: Define two new options.
9766 * config/i386/x86-tune.def: Add arch selector field in macros.
9767 * config/i386/i386.h: Adjust macro definition.
9768 * config/i386/i386.c (ix86_option_override_internal):
9770 (parse_mtune_ctrl_str): New function.
9771 (set_ix86_tune_features): New function.
9772 (ix86_function_specific_restore): Call the new helper function.
9774 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
9776 PR rtl-optimization/57662
9777 * sel-sched.c (code_motion_process_successors): When the current insn
9778 is removed after the recursive traversal, break from the loop.
9779 Add comments and debug printouts.
9781 2013-08-14 Jakub Jelinek <jakub@redhat.com>
9782 Alexandre Oliva <aoliva@redhat.com>
9785 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
9786 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
9787 there also UNSPEC_PLTOFF.
9789 2013-08-14 Marek Polacek <polacek@redhat.com>
9791 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
9792 NUM_CONDITIONS bit positions.
9794 2013-08-13 Cary Coutant <ccoutant@google.com>
9796 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
9797 (attr_checksum): Hash vector contents instead of pointer.
9798 (attr_checksum_ordered): Likewise.
9800 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
9802 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
9803 when Pmode != word_mode. Add length_address attribute.
9804 (sse3_monitor_<mode>): Merge from sse3_monitor and
9805 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
9806 Pmode != word_mode. Update insn length attribute.
9807 * config/i386/i386.c (ix86_option_override_internal): Update
9808 ix86_gen_monitor selection for merged sse3_monitor insn.
9810 2013-08-13 Julian Brown <julian@codesourcery.com>
9812 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
9813 perform invalid legitimization on greater-than-word-size modes for
9816 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
9818 * ira.c (setup_class_translate_array): Use aclass instead of cl
9819 for classes not fully covered by allocno classes.
9821 2013-08-13 Jakub Jelinek <jakub@redhat.com>
9823 PR tree-optimization/57661
9824 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
9825 * tree-inline.c (tree_function_versioning): Initialize it.
9826 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
9827 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
9828 that is not being copied.
9831 * asan.c (instrument_strlen_call): Fix typo in comment.
9832 Use char * type even for the lhs of POINTER_PLUS_EXPR.
9834 2013-08-13 Steve Ellcey <sellcey@mips.com>
9836 * config/mips/mips.md (prefetch): Use lw instead of ld on
9837 loongson in 32bit mode.
9839 2013-08-13 Nick Clifton <nickc@redhat.com>
9841 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
9843 2013-08-13 Jan Hubicka <jh@suse.cz>
9845 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
9846 introduced edge; fix typo in sanity check.
9847 (cgraph_resolve_speculation): Export; improve diagnostic.
9848 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
9849 speculation at type mismatch.
9850 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
9851 (cgraph_resolve_speculation): Declare.
9852 (symtab_can_be_discarded): New function.
9853 * value-prof.c (gimple_ic_transform): Remove actual transform code.
9854 * ipa-inline-transform.c (speculation_removed): New global var.
9855 (clone_inlined_nodes): See if speculation can be removed.
9856 (inline_call): If speculations was removed, we growths may not match.
9857 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
9858 (speculation_useful_p): New function.
9859 (resolve_noninline_speculation): New function.
9860 (inline_small_functions): Resolve useless speculations.
9861 * ipa-inline.h (speculation_useful_p): Declare
9862 * ipa.c (can_replace_by_local_alias): Simplify.
9863 (ipa_profile): Produce speculative calls in non-lto, too;
9864 add simple cost model; produce local aliases.
9866 2013-08-13 David Malcolm <dmalcolm@redhat.com>
9868 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
9871 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
9873 * config/i386/i386.c (ix86_function_versions): Use error + inform.
9875 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
9877 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
9878 iterator instead of X87MODEF.
9880 2013-08-12 Perez Read <netfirewall@gmail.com>
9883 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
9884 operand 0 for intel asm alternative.
9885 (*movabs<mode>_2): Ditto for operand 1.
9887 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
9889 * config/aarch64/arm_none.h
9890 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
9892 2013-08-12 Nick Clifton <nickc@redhat.com>
9894 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
9896 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9898 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
9899 expand for QI/HImode operand to produce more effictive code for
9900 unsigned char(short) --> float(double) conversion.
9902 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
9904 * doc/invoke.texi: Mention that -ftls-model does not force the final
9907 2013-08-12 Marek Polacek <polacek@redhat.com>
9908 Marc Glisse <marc.glisse@inria.fr>
9910 PR tree-optimization/57980
9911 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
9912 when creating -1 constant.
9914 2013-08-10 Jan Hubicka <jh@suse.cz>
9916 Workaround binutils PR14342.
9917 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
9918 (gimple_init_edge_profiler): Likewise.
9919 (gimple_gen_ic_func_profiler): Likewise.
9921 2013-08-09 Jan Hubicka <jh@suse.cz>
9923 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
9925 2013-08-09 Xinliang David Li <davidxl@google.com>
9927 * config/i386/stringop.def: New file.
9928 * config/i386/stringop.opt: New file.
9929 * config/i386/i386-opts.h: Include stringopt.def.
9930 * config/i386/i386.opt: Include stringopt.opt.
9931 * config/i386/i386.c (ix86_option_override_internal):
9932 Override default size based stringop inline strategies with options.
9933 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
9936 2013-08-09 Jan Hubicka <jh@suse.cz>
9938 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
9940 2013-08-09 Jan Hubicka <jh@suse.cz>
9942 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
9944 (dump_cgraph_node): Dump profile-id.
9945 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
9946 and common_target_probability.
9947 * lto-cgraph.c (lto_output_edge): Stream common targets.
9948 (lto_output_node): Stream profile ids.
9949 (input_node): Stream profile ids.
9950 (input_edge): Stream common targets.
9951 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
9952 * ipa.c: Include value-prof.h
9953 (ipa_profile_generate_summary): Turn indirect call statement histograms
9954 into common targets.
9955 (ipa_profile): Turn common targets into speculative edges.
9957 2013-08-09 Jan Hubicka <jh@suse.cz>
9959 * cgraph.h (cgraph_node): Add profile_id.
9960 * value-prof.c (cgraph_node_map): Turn into pointer_map.
9961 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
9962 (del_node_map): Update.
9963 (find_func_by_funcdef_no): Replace by ...
9964 (find_func_by_profile_id): ... this one.
9965 (gimple_ic_transform): Do not remove useful histograms when
9966 speculation is not done; dump info when indirect call removal
9968 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
9969 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
9970 (__gcov_indirect_call_profiler_v2): .. this one.
9971 * profile.h (init_node_map): Update.
9972 * coverage.c (coverage_compute_profile_id): New function.
9973 * coverage.h (coverage_compute_profile_id): Declare.
9974 * tree-profile.c (init_ic_make_global_vars): Make
9975 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
9976 (gimple_init_edge_profiler): Update prototype of
9977 __gcov_indirect_call_profiler.
9978 (gimple_gen_ic_func_profiler): Simplify.
9979 (tree_profiling): Use init_node_map
9981 2013-08-09 Jan Hubicka <jh@suse.cz>
9983 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
9984 non-speculative refs.
9985 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
9986 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
9987 (cgraph_set_call_stmt): Likewise.
9988 (cgraph_create_edge_1): Fix release checking compilatoin;
9990 (cgraph_free_edge): Free indirect info.
9991 (cgraph_turn_edge_to_speculative): New function.
9992 (cgraph_speculative_call_info): New function.
9993 (cgraph_make_edge_direct): Return direct edge; handle speculation.
9994 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
9995 (dump_cgraph_node): Dump speculation.
9996 (verify_edge_count_and_frequency): Accept speculative edges.
9997 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
9998 (verify_cgraph_node): Handle speculation.
9999 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
10000 (cgraph_set_call_stmt): Update prototype.
10001 (cgraph_make_edge_direct): Update prototype.
10002 (cgraph_speculative_call_info): Declare.
10003 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
10004 to change; update call of ipa_find_references.
10005 * ipa-ref.c (ipa_record_reference): Fix return value; clear
10006 lto_stmt_uid and speculative flags.
10007 (ipa_dump_references): Dump speculation.
10008 (ipa_clone_references): Clone speculative flag.
10009 (ipa_clone_referring): Likewise.
10010 (ipa_clone_ref): New function.
10011 (ipa_find_reference): Look into lto_stmt_uids
10012 (ipa_clear_stmts_in_references): Do not clear speculative calls.
10013 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
10014 (ipa_find_reference): Update declaration.
10015 (ipa_clone_ref): Declare.
10016 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
10017 stream speculative flag.
10018 (lto_output_ref): Stream statements uids and speculation.
10019 (input_ref): Likewise.
10020 (input_edge): Stream speuclation.
10021 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
10022 (cgraph_set_call_stmt_including_clones): Handle speculation.
10023 * ipa-inline.c (heap_edge_removal_hook): New function.
10024 (inline_small_functions): Register it.
10025 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
10026 also initialize refs.
10027 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
10029 (try_make_edge_direct_simple_call): Likewise.
10030 (try_make_edge_direct_simple_call): Likewise.
10031 (update_indirect_edges_after_inlining): Likewise.
10032 (remove_described_reference): Look proper lto_stmt_uid.
10033 (propagate_controlled_uses): Likewise.
10034 (propagate_controlled_uses): Liekwise.
10035 * tree-inline.c (copy_bb): Copy speculative edges.
10036 (redirect_all_calls): New function.
10037 (copy_cfg_body): Do redirection after loop info is updated.
10038 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
10040 2013-08-09 Jan Hubicka <jh@suse.cz>
10042 * lto-streamer-out.c (output_function): Renumber PHIs.
10043 * lto-streamer-in.c (input_function): Likewise.
10045 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
10047 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
10048 (get_lane_unsigned): Likewise.
10049 (dup_lane_scalar): Likewise.
10050 (get_lane): enable for VALL.
10051 * config/aarch64/aarch64-simd.md
10052 (aarch64_dup_lane_scalar<mode>): Remove.
10053 (aarch64_get_lane_signed<mode>): Likewise.
10054 (aarch64_get_lane_unsigned<mode>): Likewise.
10055 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
10056 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
10057 (aarch64_get_lane<mode>): Enable for all vector modes.
10058 (aarch64_get_lanedi): Remove misleading constraints.
10059 * config/aarch64/arm_neon.h
10060 (__aarch64_vget_lane_any): Define.
10061 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
10062 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
10063 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
10064 * config/aarch64/iterators.md (VDQQH): New.
10065 (VDQQHS): Likewise.
10066 (vwcore): Likewise.
10068 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
10070 * configure.ac: Add GAS check for LEON instructions on SPARC.
10071 * configure: Regenerate.
10072 * config.in: Likewise.
10073 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
10075 * config/sparc/sparc.opt (LEON, LEON3): New masks.
10076 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
10078 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
10079 (AS_LEON_FLAG): New macro.
10080 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
10081 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
10082 Deal with LEON and LEON3 for the memory model.
10083 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
10084 (atomic_compare_and_swap<mode>_1): Likewise.
10085 (*atomic_compare_and_swap<mode>_1): Likewise.
10087 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10089 * config/arm/neon.md (vcond): Fix floating-point vector
10090 comparisons against 0.
10092 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
10094 * lra-constraints.c (emit_spill_move): Remove assert.
10095 (process_alt_operands): Add more debugging
10096 output. Increase reject for spilling into memory. Decrease
10097 reject for reloading scratch.
10098 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
10100 2013-08-08 Steve Ellcey <sellcey@mips.com>
10102 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
10103 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
10104 micromips incompatible. Add nan2008.
10105 (MULTILIB_DIRNAMES): Add nan2008.
10106 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
10107 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
10108 and micromips incompatible. Add nan2008.
10109 (MULTILIB_DIRNAMES): Add nan2008.
10110 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
10112 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
10114 PR rtl-optimization/58079
10115 * combine.c (combine_simplify_rtx): Avoid using SUBST if
10116 simplify_comparison has widened a comparison with an integer.
10118 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10120 * config/arm/neon.md (movmisalign<mode>): Disable when we
10121 don't allow unaligned accesses.
10122 (*movmisalign<mode>_neon_store): Likewise.
10123 (*movmisalign<mode>_neon_load): Likewise.
10124 (*movmisalign<mode>_neon_store): Likewise.
10125 (*movmisalign<mode>_neon_load): Likewise.
10127 2013-08-08 Jan Hubicka <jh@suse.cz>
10129 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
10130 (make_pass_rebuild_cgraph_edges): Also clear references.
10131 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
10132 * ipa-inline-transform.c (inline_transform): Remove all references
10134 * cgraphunit.c (expand_function): Remove all references after
10136 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
10137 (ipa_find_reference): Rewrite to iterator.
10138 (remove_stmt_references): Likewise.
10139 (ipa_clear_stmts_in_references): New function.
10140 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
10141 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
10143 * ipa-split.c (split_function): Remove references in split function.
10145 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
10148 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
10149 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
10151 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
10154 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
10155 suggested mode for the assignment isn't compatible with the
10156 registers required.
10158 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
10161 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
10163 2013-08-07 Xinliang David Li <davidxl@google.com>
10165 * config/i386/i386.opt: New option -mtune-ctrl=.
10166 * config/i386/x86-tune.def: New file.
10167 * config/i386/i386.h: include x86-tune.def.
10168 * config/i386/i386.c (ix86_option_override_internal):
10169 Parsing -mtune-ctrl= option and set tune features.
10171 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
10174 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
10175 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
10178 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10180 * rtl.h (update_alignments): Declare.
10181 * final.c (grow_label_align): New function extracted from...
10182 (shorten_branches): ...here. Call it.
10183 (update_alignments): New function.
10184 * reorg.c (sibling_labels): New variable.
10185 (get_label_before): Add SIBLING parameter. If it is non-zero, push
10186 the new label along with it onto the sibling_labels vector.
10187 (fill_simple_delay_slots): Adjust call to get_label_before.
10188 (fill_slots_from_thread): Likewise.
10189 (relax_delay_slots): Likewise.
10190 (make_return_insns): Likewise.
10191 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
10193 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
10195 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
10196 document its semantics.
10197 (diagnostic_report_diagnostic): Adjust accordingly.
10199 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10201 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
10202 (sparc_option_override): ...and port to new C++ pass API.
10203 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
10205 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
10207 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
10209 2013-08-06 Caroline Tice <cmtice@google.com>
10211 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
10212 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
10213 * tree-pass.h: Add pass_vtable_verify.
10214 * varasm.c (assemble_variable): Add code to properly set the comdat
10215 section and name for the .vtable_map_vars section.
10216 (assemble_vtyv_preinit_initializer): New function.
10217 (default_sectin_type_flags): Make sure .vtable_map_vars section has
10219 * output.h: Add function decl for assemble_vtv_preinit_initializer.
10220 * vtable-verify.c: New file.
10221 * vtable-verify.h: New file.
10222 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
10223 initialiation levels.
10224 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
10225 * passes.def: Insert pass_vtable_verify.
10226 * aclocal.m4: Reorder includes.
10227 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
10228 -fvtv-counts options.
10229 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
10230 as appropriate, if -fvtable-verify=... is used.
10231 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
10232 -fvtable-verify=... is used.
10233 * Makefile.in (OBJS): Add vtable-verify.o to list.
10234 (vtable-verify.o): Add new build rule.
10235 (GTFILES): Add vtable-verify.c to list.
10236 * common.opt (fvtable-verify=): New flag.
10237 (vtv_priority): Values for fvtable-verify= flag.
10238 (fvtv-counts): New flag.
10239 (fvtv-debug): New flag.
10240 * tree.h (save_vtable_map_decl): New extern function decl.
10242 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10244 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
10245 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
10246 (pass_data_rl78_devirt): ...new pass_data instance and...
10247 (make_pass_rl78_devirt): ...new function.
10248 (rl78_asm_file_start): Port pass registration to new C++ API.
10250 2013-08-07 David Malcolm <dmalcolm@redhat.com>
10252 * coretypes.h (rtl_opt_pass): Add.
10253 (gcc::context): Add.
10254 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
10255 (epiphany_init): Port to new C++ pass API.
10256 (epiphany_optimize_mode_switching): Likewise.
10257 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
10258 (pass_manager::get_pass_mode_switching): New.
10259 (pass_manager::get_pass_peephole2): New.
10260 * mode-switching.c (pass_mode_switching): Add clone method.
10261 * recog.c (pass_peephole2): Add clone method.
10262 (pass_split_all_insns): Add clone method.
10264 2013-08-06 David Malcolm <dmalcolm@redhat.com>
10266 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
10267 (mips_option_override): ...here, porting to new C++ API for passes.
10269 2013-08-06 Jan Hubicka <jh@suse.cz>
10271 * cgraph.c (cgraph_get_body): New function based on lto.c
10273 * cgraph.h (cgraph_get_body): Declare.
10274 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
10276 * cgraphunit.c (expand_function): Get body prior expanding.
10277 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
10278 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
10280 * passes.c (do_per_function_toporder): Get body.
10281 * tree-inline.c (expand_call_inline): Get body prior inlining it.
10282 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
10284 2013-08-06 Martin Jambor <mjambor@suse.cz>
10287 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
10288 is not re-finalized. Rename second parameter to no_collect.
10290 2013-08-06 Martin Jambor <mjambor@suse.cz>
10292 PR middle-end/58041
10293 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
10294 MEM_REF has proper alignment information.
10296 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
10299 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
10301 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
10302 rtx (*) (rtx, ...) with insn_gen_fn.
10303 * genoutput.c (output_insn_data): Cast gen_? function pointers to
10304 insn_gen_fn::stored_funcptr. Add initializer braces.
10306 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10308 Rewrite how instances of passes are cloned to remove assumptions
10309 about their sizes (thus allowing pass subclasses to have
10310 additional data fields, albeit non-GC-managed ones at this point).
10312 * passes.c (make_pass_instance): Now that passes have clone
10313 methods, rewrite this function to eliminate XNEW and memcpy
10314 calls that used hardcoded sizes. Since this function no longer
10315 creates pass instances, rename it to...
10316 (add_pass_instance): ...this. Document the old way that passes were
10317 numbered and flagged, and rework this function to continue using it.
10318 (next_pass_1): Add an initial_pass argument for use by
10320 (position_pass): When adding multiple instances of a pass, use
10321 the pass's clone method, rather than relying on the XNEW/memcpy
10322 within the former make_pass_instance (now add_pass_instance).
10323 (pass_manager::pass_manager): When invoking next_pass_1, also supply
10324 the initial instance of the current pass within the pass manager.
10326 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10328 This is the automated part of the conversion of passes from C
10329 structs to C++ classes.
10331 Patch autogenerated by refactor_passes.py from
10332 https://github.com/davidmalcolm/gcc-refactoring-scripts
10333 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
10335 * asan.c (pass_asan): Convert from a global struct to a subclass of
10336 gimple_opt_pass along with...
10337 (pass_data_asan): ...new pass_data instance and...
10338 (make_pass_asan): ...new function.
10339 (pass_asan_O0): Convert from a global struct to a subclass of
10340 gimple_opt_pass along with...
10341 (pass_data_asan_O0): ...new pass_data instance and...
10342 (make_pass_asan_O0): ...new function.
10343 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
10344 subclass of rtl_opt_pass along with...
10345 (pass_data_inc_dec): ...new pass_data instance and...
10346 (make_pass_inc_dec): ...new function.
10347 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
10348 a subclass of rtl_opt_pass along with...
10349 (pass_data_reorder_blocks): ...new pass_data instance and...
10350 (make_pass_reorder_blocks): ...new function.
10351 (pass_duplicate_computed_gotos): Convert from a global struct to a
10352 subclass of rtl_opt_pass along with...
10353 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
10354 (make_pass_duplicate_computed_gotos): ...new function.
10355 (pass_partition_blocks): Convert from a global struct to a subclass of
10356 rtl_opt_pass along with...
10357 (pass_data_partition_blocks): ...new pass_data instance and...
10358 (make_pass_partition_blocks): ...new function.
10359 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
10360 struct to a subclass of rtl_opt_pass along with...
10361 (pass_data_branch_target_load_optimize1): ...new pass_data instance
10363 (make_pass_branch_target_load_optimize1): ...new function.
10364 (pass_branch_target_load_optimize2): Convert from a global struct to a
10365 subclass of rtl_opt_pass along with...
10366 (pass_data_branch_target_load_optimize2): ...new pass_data instance
10368 (make_pass_branch_target_load_optimize2): ...new function.
10369 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
10370 of rtl_opt_pass along with...
10371 (pass_data_jump): ...new pass_data instance and...
10372 (make_pass_jump): ...new function.
10373 (pass_jump2): Convert from a global struct to a subclass of
10374 rtl_opt_pass along with...
10375 (pass_data_jump2): ...new pass_data instance and...
10376 (make_pass_jump2): ...new function.
10377 * cfgexpand.c (pass_expand): Convert from a global struct to a
10378 subclass of rtl_opt_pass along with...
10379 (pass_data_expand): ...new pass_data instance and...
10380 (make_pass_expand): ...new function.
10381 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
10382 of rtl_opt_pass along with...
10383 (pass_data_free_cfg): ...new pass_data instance and...
10384 (make_pass_free_cfg): ...new function.
10385 (pass_into_cfg_layout_mode): Convert from a global struct to a
10386 subclass of rtl_opt_pass along with...
10387 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
10388 (make_pass_into_cfg_layout_mode): ...new function.
10389 (pass_outof_cfg_layout_mode): Convert from a global struct to a
10390 subclass of rtl_opt_pass along with...
10391 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
10392 (make_pass_outof_cfg_layout_mode): ...new function.
10393 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
10394 struct to a subclass of gimple_opt_pass along with...
10395 (pass_data_build_cgraph_edges): ...new pass_data instance and...
10396 (make_pass_build_cgraph_edges): ...new function.
10397 (pass_rebuild_cgraph_edges): Convert from a global struct to a
10398 subclass of gimple_opt_pass along with...
10399 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
10400 (make_pass_rebuild_cgraph_edges): ...new function.
10401 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
10402 subclass of gimple_opt_pass along with...
10403 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
10405 (make_pass_remove_cgraph_callee_edges): ...new function.
10406 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
10407 struct to a subclass of rtl_opt_pass along with...
10408 (pass_data_stack_adjustments): ...new pass_data instance and...
10409 (make_pass_stack_adjustments): ...new function.
10410 * combine.c (pass_combine): Convert from a global struct to a subclass
10411 of rtl_opt_pass along with...
10412 (pass_data_combine): ...new pass_data instance and...
10413 (make_pass_combine): ...new function.
10414 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
10415 global struct to a subclass of rtl_opt_pass along with...
10416 (pass_data_compare_elim_after_reload): ...new pass_data instance
10418 (make_pass_compare_elim_after_reload): ...new function.
10419 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
10420 of rtl_opt_pass along with...
10421 (pass_data_rtl_cprop): ...new pass_data instance and...
10422 (make_pass_rtl_cprop): ...new function.
10423 * cse.c (pass_cse): Convert from a global struct to a subclass of
10424 rtl_opt_pass along with...
10425 (pass_data_cse): ...new pass_data instance and...
10426 (make_pass_cse): ...new function.
10427 (pass_cse2): Convert from a global struct to a subclass of
10428 rtl_opt_pass along with...
10429 (pass_data_cse2): ...new pass_data instance and...
10430 (make_pass_cse2): ...new function.
10431 (pass_cse_after_global_opts): Convert from a global struct to a
10432 subclass of rtl_opt_pass along with...
10433 (pass_data_cse_after_global_opts): ...new pass_data instance and...
10434 (make_pass_cse_after_global_opts): ...new function.
10435 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
10436 of rtl_opt_pass along with...
10437 (pass_data_ud_rtl_dce): ...new pass_data instance and...
10438 (make_pass_ud_rtl_dce): ...new function.
10439 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
10440 rtl_opt_pass along with...
10441 (pass_data_fast_rtl_dce): ...new pass_data instance and...
10442 (make_pass_fast_rtl_dce): ...new function.
10443 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
10444 a subclass of rtl_opt_pass along with...
10445 (pass_data_df_initialize_opt): ...new pass_data instance and...
10446 (make_pass_df_initialize_opt): ...new function.
10447 (pass_df_initialize_no_opt): Convert from a global struct to a
10448 subclass of rtl_opt_pass along with...
10449 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
10450 (make_pass_df_initialize_no_opt): ...new function.
10451 (pass_df_finish): Convert from a global struct to a subclass of
10452 rtl_opt_pass along with...
10453 (pass_data_df_finish): ...new pass_data instance and...
10454 (make_pass_df_finish): ...new function.
10455 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
10456 rtl_opt_pass along with...
10457 (pass_data_rtl_dse1): ...new pass_data instance and...
10458 (make_pass_rtl_dse1): ...new function.
10459 (pass_rtl_dse2): Convert from a global struct to a subclass of
10460 rtl_opt_pass along with...
10461 (pass_data_rtl_dse2): ...new pass_data instance and...
10462 (make_pass_rtl_dse2): ...new function.
10463 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
10464 subclass of rtl_opt_pass along with...
10465 (pass_data_dwarf2_frame): ...new pass_data instance and...
10466 (make_pass_dwarf2_frame): ...new function.
10467 * except.c (pass_set_nothrow_function_flags): Convert from a global
10468 struct to a subclass of rtl_opt_pass along with...
10469 (pass_data_set_nothrow_function_flags): ...new pass_data instance
10471 (make_pass_set_nothrow_function_flags): ...new function.
10472 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
10473 subclass of rtl_opt_pass along with...
10474 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
10476 (make_pass_convert_to_eh_region_ranges): ...new function.
10477 * final.c (pass_compute_alignments): Convert from a global struct to a
10478 subclass of rtl_opt_pass along with...
10479 (pass_data_compute_alignments): ...new pass_data instance and...
10480 (make_pass_compute_alignments): ...new function.
10481 (pass_final): Convert from a global struct to a subclass of
10482 rtl_opt_pass along with...
10483 (pass_data_final): ...new pass_data instance and...
10484 (make_pass_final): ...new function.
10485 (pass_shorten_branches): Convert from a global struct to a subclass of
10486 rtl_opt_pass along with...
10487 (pass_data_shorten_branches): ...new pass_data instance and...
10488 (make_pass_shorten_branches): ...new function.
10489 (pass_clean_state): Convert from a global struct to a subclass of
10490 rtl_opt_pass along with...
10491 (pass_data_clean_state): ...new pass_data instance and...
10492 (make_pass_clean_state): ...new function.
10493 * function.c (pass_instantiate_virtual_regs): Convert from a global
10494 struct to a subclass of rtl_opt_pass along with...
10495 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
10496 (make_pass_instantiate_virtual_regs): ...new function.
10497 (pass_leaf_regs): Convert from a global struct to a subclass of
10498 rtl_opt_pass along with...
10499 (pass_data_leaf_regs): ...new pass_data instance and...
10500 (make_pass_leaf_regs): ...new function.
10501 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
10502 subclass of rtl_opt_pass along with...
10503 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
10505 (make_pass_thread_prologue_and_epilogue): ...new function.
10506 (pass_match_asm_constraints): Convert from a global struct to a
10507 subclass of rtl_opt_pass along with...
10508 (pass_data_match_asm_constraints): ...new pass_data instance and...
10509 (make_pass_match_asm_constraints): ...new function.
10510 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
10511 subclass of rtl_opt_pass along with...
10512 (pass_data_rtl_fwprop): ...new pass_data instance and...
10513 (make_pass_rtl_fwprop): ...new function.
10514 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
10515 rtl_opt_pass along with...
10516 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
10517 (make_pass_rtl_fwprop_addr): ...new function.
10518 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
10519 rtl_opt_pass along with...
10520 (pass_data_rtl_pre): ...new pass_data instance and...
10521 (make_pass_rtl_pre): ...new function.
10522 (pass_rtl_hoist): Convert from a global struct to a subclass of
10523 rtl_opt_pass along with...
10524 (pass_data_rtl_hoist): ...new pass_data instance and...
10525 (make_pass_rtl_hoist): ...new function.
10526 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
10527 subclass of gimple_opt_pass along with...
10528 (pass_data_lower_cf): ...new pass_data instance and...
10529 (make_pass_lower_cf): ...new function.
10530 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
10531 from a global struct to a subclass of gimple_opt_pass along with...
10532 (pass_data_strength_reduction): ...new pass_data instance and...
10533 (make_pass_strength_reduction): ...new function.
10534 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
10535 of rtl_opt_pass along with...
10536 (pass_data_rtl_ifcvt): ...new pass_data instance and...
10537 (make_pass_rtl_ifcvt): ...new function.
10538 (pass_if_after_combine): Convert from a global struct to a subclass of
10539 rtl_opt_pass along with...
10540 (pass_data_if_after_combine): ...new pass_data instance and...
10541 (make_pass_if_after_combine): ...new function.
10542 (pass_if_after_reload): Convert from a global struct to a subclass of
10543 rtl_opt_pass along with...
10544 (pass_data_if_after_reload): ...new pass_data instance and...
10545 (make_pass_if_after_reload): ...new function.
10546 * init-regs.c (pass_initialize_regs): Convert from a global struct to
10547 a subclass of rtl_opt_pass along with...
10548 (pass_data_initialize_regs): ...new pass_data instance and...
10549 (make_pass_initialize_regs): ...new function.
10550 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
10551 of ipa_opt_pass_d along with...
10552 (pass_data_ipa_cp): ...new pass_data instance and...
10553 (make_pass_ipa_cp): ...new function.
10554 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
10555 global struct to a subclass of gimple_opt_pass along with...
10556 (pass_data_inline_parameters): ...new pass_data instance and...
10557 (make_pass_inline_parameters): ...new function.
10558 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
10559 subclass of gimple_opt_pass along with...
10560 (pass_data_early_inline): ...new pass_data instance and...
10561 (make_pass_early_inline): ...new function.
10562 (pass_ipa_inline): Convert from a global struct to a subclass of
10563 ipa_opt_pass_d along with...
10564 (pass_data_ipa_inline): ...new pass_data instance and...
10565 (make_pass_ipa_inline): ...new function.
10566 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
10567 struct to a subclass of gimple_opt_pass along with...
10568 (pass_data_local_pure_const): ...new pass_data instance and...
10569 (make_pass_local_pure_const): ...new function.
10570 (pass_ipa_pure_const): Convert from a global struct to a subclass of
10571 ipa_opt_pass_d along with...
10572 (pass_data_ipa_pure_const): ...new pass_data instance and...
10573 (make_pass_ipa_pure_const): ...new function.
10574 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
10575 to a subclass of ipa_opt_pass_d along with...
10576 (pass_data_ipa_reference): ...new pass_data instance and...
10577 (make_pass_ipa_reference): ...new function.
10578 * ipa-split.c (pass_split_functions): Convert from a global struct to
10579 a subclass of gimple_opt_pass along with...
10580 (pass_data_split_functions): ...new pass_data instance and...
10581 (make_pass_split_functions): ...new function.
10582 (pass_feedback_split_functions): Convert from a global struct to a
10583 subclass of gimple_opt_pass along with...
10584 (pass_data_feedback_split_functions): ...new pass_data instance and...
10585 (make_pass_feedback_split_functions): ...new function.
10586 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
10587 global struct to a subclass of simple_ipa_opt_pass along with...
10588 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
10590 (make_pass_ipa_function_and_variable_visibility): ...new function.
10591 (pass_ipa_free_inline_summary): Convert from a global struct to a
10592 subclass of simple_ipa_opt_pass along with...
10593 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
10594 (make_pass_ipa_free_inline_summary): ...new function.
10595 (pass_ipa_whole_program_visibility): Convert from a global struct to a
10596 subclass of ipa_opt_pass_d along with...
10597 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
10599 (make_pass_ipa_whole_program_visibility): ...new function.
10600 (pass_ipa_profile): Convert from a global struct to a subclass of
10601 ipa_opt_pass_d along with...
10602 (pass_data_ipa_profile): ...new pass_data instance and...
10603 (make_pass_ipa_profile): ...new function.
10604 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
10605 ipa_opt_pass_d along with...
10606 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
10607 (make_pass_ipa_cdtor_merge): ...new function.
10608 * ira.c (pass_ira): Convert from a global struct to a subclass of
10609 rtl_opt_pass along with...
10610 (pass_data_ira): ...new pass_data instance and...
10611 (make_pass_ira): ...new function.
10612 (pass_reload): Convert from a global struct to a subclass of
10613 rtl_opt_pass along with...
10614 (pass_data_reload): ...new pass_data instance and...
10615 (make_pass_reload): ...new function.
10616 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
10617 subclass of rtl_opt_pass along with...
10618 (pass_data_cleanup_barriers): ...new pass_data instance and...
10619 (make_pass_cleanup_barriers): ...new function.
10620 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
10621 of rtl_opt_pass along with...
10622 (pass_data_loop2): ...new pass_data instance and...
10623 (make_pass_loop2): ...new function.
10624 (pass_rtl_loop_init): Convert from a global struct to a subclass of
10625 rtl_opt_pass along with...
10626 (pass_data_rtl_loop_init): ...new pass_data instance and...
10627 (make_pass_rtl_loop_init): ...new function.
10628 (pass_rtl_loop_done): Convert from a global struct to a subclass of
10629 rtl_opt_pass along with...
10630 (pass_data_rtl_loop_done): ...new pass_data instance and...
10631 (make_pass_rtl_loop_done): ...new function.
10632 (pass_rtl_move_loop_invariants): Convert from a global struct to a
10633 subclass of rtl_opt_pass along with...
10634 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
10635 (make_pass_rtl_move_loop_invariants): ...new function.
10636 (pass_rtl_unswitch): Convert from a global struct to a subclass of
10637 rtl_opt_pass along with...
10638 (pass_data_rtl_unswitch): ...new pass_data instance and...
10639 (make_pass_rtl_unswitch): ...new function.
10640 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
10641 subclass of rtl_opt_pass along with...
10642 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
10644 (make_pass_rtl_unroll_and_peel_loops): ...new function.
10645 (pass_rtl_doloop): Convert from a global struct to a subclass of
10646 rtl_opt_pass along with...
10647 (pass_data_rtl_doloop): ...new pass_data instance and...
10648 (make_pass_rtl_doloop): ...new function.
10649 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
10650 a subclass of rtl_opt_pass along with...
10651 (pass_data_lower_subreg): ...new pass_data instance and...
10652 (make_pass_lower_subreg): ...new function.
10653 (pass_lower_subreg2): Convert from a global struct to a subclass of
10654 rtl_opt_pass along with...
10655 (pass_data_lower_subreg2): ...new pass_data instance and...
10656 (make_pass_lower_subreg2): ...new function.
10657 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
10658 struct to a subclass of ipa_opt_pass_d along with...
10659 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
10660 (make_pass_ipa_lto_gimple_out): ...new function.
10661 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
10662 of ipa_opt_pass_d along with...
10663 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
10664 (make_pass_ipa_lto_finish_out): ...new function.
10665 * mode-switching.c (pass_mode_switching): Convert from a global struct
10666 to a subclass of rtl_opt_pass along with...
10667 (pass_data_mode_switching): ...new pass_data instance and...
10668 (make_pass_mode_switching): ...new function.
10669 * modulo-sched.c (pass_sms): Convert from a global struct to a
10670 subclass of rtl_opt_pass along with...
10671 (pass_data_sms): ...new pass_data instance and...
10672 (make_pass_sms): ...new function.
10673 * omp-low.c (pass_expand_omp): Convert from a global struct to a
10674 subclass of gimple_opt_pass along with...
10675 (pass_data_expand_omp): ...new pass_data instance and...
10676 (make_pass_expand_omp): ...new function.
10677 (pass_lower_omp): Convert from a global struct to a subclass of
10678 gimple_opt_pass along with...
10679 (pass_data_lower_omp): ...new pass_data instance and...
10680 (make_pass_lower_omp): ...new function.
10681 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
10682 of gimple_opt_pass along with...
10683 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
10684 (make_pass_diagnose_omp_blocks): ...new function.
10685 * passes.c (pass_early_local_passes): Convert from a global struct to
10686 a subclass of simple_ipa_opt_pass along with...
10687 (pass_data_early_local_passes): ...new pass_data instance and...
10688 (make_pass_early_local_passes): ...new function.
10689 (pass_all_early_optimizations): Convert from a global struct to a
10690 subclass of gimple_opt_pass along with...
10691 (pass_data_all_early_optimizations): ...new pass_data instance and...
10692 (make_pass_all_early_optimizations): ...new function.
10693 (pass_all_optimizations): Convert from a global struct to a subclass
10694 of gimple_opt_pass along with...
10695 (pass_data_all_optimizations): ...new pass_data instance and...
10696 (make_pass_all_optimizations): ...new function.
10697 (pass_all_optimizations_g): Convert from a global struct to a subclass
10698 of gimple_opt_pass along with...
10699 (pass_data_all_optimizations_g): ...new pass_data instance and...
10700 (make_pass_all_optimizations_g): ...new function.
10701 (pass_rest_of_compilation): Convert from a global struct to a subclass
10702 of rtl_opt_pass along with...
10703 (pass_data_rest_of_compilation): ...new pass_data instance and...
10704 (make_pass_rest_of_compilation): ...new function.
10705 (pass_postreload): Convert from a global struct to a subclass of
10706 rtl_opt_pass along with...
10707 (pass_data_postreload): ...new pass_data instance and...
10708 (make_pass_postreload): ...new function.
10709 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
10710 subclass of rtl_opt_pass along with...
10711 (pass_data_gcse2): ...new pass_data instance and...
10712 (make_pass_gcse2): ...new function.
10713 * postreload.c (pass_postreload_cse): Convert from a global struct to
10714 a subclass of rtl_opt_pass along with...
10715 (pass_data_postreload_cse): ...new pass_data instance and...
10716 (make_pass_postreload_cse): ...new function.
10717 * predict.c (pass_profile): Convert from a global struct to a subclass
10718 of gimple_opt_pass along with...
10719 (pass_data_profile): ...new pass_data instance and...
10720 (make_pass_profile): ...new function.
10721 (pass_strip_predict_hints): Convert from a global struct to a subclass
10722 of gimple_opt_pass along with...
10723 (pass_data_strip_predict_hints): ...new pass_data instance and...
10724 (make_pass_strip_predict_hints): ...new function.
10725 * recog.c (pass_peephole2): Convert from a global struct to a subclass
10726 of rtl_opt_pass along with...
10727 (pass_data_peephole2): ...new pass_data instance and...
10728 (make_pass_peephole2): ...new function.
10729 (pass_split_all_insns): Convert from a global struct to a subclass of
10730 rtl_opt_pass along with...
10731 (pass_data_split_all_insns): ...new pass_data instance and...
10732 (make_pass_split_all_insns): ...new function.
10733 (pass_split_after_reload): Convert from a global struct to a subclass
10734 of rtl_opt_pass along with...
10735 (pass_data_split_after_reload): ...new pass_data instance and...
10736 (make_pass_split_after_reload): ...new function.
10737 (pass_split_before_regstack): Convert from a global struct to a
10738 subclass of rtl_opt_pass along with...
10739 (pass_data_split_before_regstack): ...new pass_data instance and...
10740 (make_pass_split_before_regstack): ...new function.
10741 (pass_split_before_sched2): Convert from a global struct to a subclass
10742 of rtl_opt_pass along with...
10743 (pass_data_split_before_sched2): ...new pass_data instance and...
10744 (make_pass_split_before_sched2): ...new function.
10745 (pass_split_for_shorten_branches): Convert from a global struct to a
10746 subclass of rtl_opt_pass along with...
10747 (pass_data_split_for_shorten_branches): ...new pass_data instance
10749 (make_pass_split_for_shorten_branches): ...new function.
10750 * ree.c (pass_ree): Convert from a global struct to a subclass of
10751 rtl_opt_pass along with...
10752 (pass_data_ree): ...new pass_data instance and...
10753 (make_pass_ree): ...new function.
10754 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
10755 subclass of rtl_opt_pass along with...
10756 (pass_data_stack_regs): ...new pass_data instance and...
10757 (make_pass_stack_regs): ...new function.
10758 (pass_stack_regs_run): Convert from a global struct to a subclass of
10759 rtl_opt_pass along with...
10760 (pass_data_stack_regs_run): ...new pass_data instance and...
10761 (make_pass_stack_regs_run): ...new function.
10762 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
10763 subclass of rtl_opt_pass along with...
10764 (pass_data_cprop_hardreg): ...new pass_data instance and...
10765 (make_pass_cprop_hardreg): ...new function.
10766 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
10767 subclass of rtl_opt_pass along with...
10768 (pass_data_reginfo_init): ...new pass_data instance and...
10769 (make_pass_reginfo_init): ...new function.
10770 * regmove.c (pass_regmove): Convert from a global struct to a subclass
10771 of rtl_opt_pass along with...
10772 (pass_data_regmove): ...new pass_data instance and...
10773 (make_pass_regmove): ...new function.
10774 * regrename.c (pass_regrename): Convert from a global struct to a
10775 subclass of rtl_opt_pass along with...
10776 (pass_data_regrename): ...new pass_data instance and...
10777 (make_pass_regrename): ...new function.
10778 * reorg.c (pass_delay_slots): Convert from a global struct to a
10779 subclass of rtl_opt_pass along with...
10780 (pass_data_delay_slots): ...new pass_data instance and...
10781 (make_pass_delay_slots): ...new function.
10782 (pass_machine_reorg): Convert from a global struct to a subclass of
10783 rtl_opt_pass along with...
10784 (pass_data_machine_reorg): ...new pass_data instance and...
10785 (make_pass_machine_reorg): ...new function.
10786 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
10787 of rtl_opt_pass along with...
10788 (pass_data_sched): ...new pass_data instance and...
10789 (make_pass_sched): ...new function.
10790 (pass_sched2): Convert from a global struct to a subclass of
10791 rtl_opt_pass along with...
10792 (pass_data_sched2): ...new pass_data instance and...
10793 (make_pass_sched2): ...new function.
10794 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
10795 to a subclass of rtl_opt_pass along with...
10796 (pass_data_stack_ptr_mod): ...new pass_data instance and...
10797 (make_pass_stack_ptr_mod): ...new function.
10798 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
10799 to a subclass of rtl_opt_pass along with...
10800 (pass_data_rtl_store_motion): ...new pass_data instance and...
10801 (make_pass_rtl_store_motion): ...new function.
10802 * tracer.c (pass_tracer): Convert from a global struct to a subclass
10803 of gimple_opt_pass along with...
10804 (pass_data_tracer): ...new pass_data instance and...
10805 (make_pass_tracer): ...new function.
10806 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
10807 to a subclass of gimple_opt_pass along with...
10808 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
10809 (make_pass_diagnose_tm_blocks): ...new function.
10810 (pass_lower_tm): Convert from a global struct to a subclass of
10811 gimple_opt_pass along with...
10812 (pass_data_lower_tm): ...new pass_data instance and...
10813 (make_pass_lower_tm): ...new function.
10814 (pass_tm_init): Convert from a global struct to a subclass of
10815 gimple_opt_pass along with...
10816 (pass_data_tm_init): ...new pass_data instance and...
10817 (make_pass_tm_init): ...new function.
10818 (pass_tm_mark): Convert from a global struct to a subclass of
10819 gimple_opt_pass along with...
10820 (pass_data_tm_mark): ...new pass_data instance and...
10821 (make_pass_tm_mark): ...new function.
10822 (pass_tm_edges): Convert from a global struct to a subclass of
10823 gimple_opt_pass along with...
10824 (pass_data_tm_edges): ...new pass_data instance and...
10825 (make_pass_tm_edges): ...new function.
10826 (pass_tm_memopt): Convert from a global struct to a subclass of
10827 gimple_opt_pass along with...
10828 (pass_data_tm_memopt): ...new pass_data instance and...
10829 (make_pass_tm_memopt): ...new function.
10830 (pass_ipa_tm): Convert from a global struct to a subclass of
10831 simple_ipa_opt_pass along with...
10832 (pass_data_ipa_tm): ...new pass_data instance and...
10833 (make_pass_ipa_tm): ...new function.
10834 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
10835 subclass of gimple_opt_pass along with...
10836 (pass_data_call_cdce): ...new pass_data instance and...
10837 (make_pass_call_cdce): ...new function.
10838 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
10839 subclass of gimple_opt_pass along with...
10840 (pass_data_build_cfg): ...new pass_data instance and...
10841 (make_pass_build_cfg): ...new function.
10842 (pass_split_crit_edges): Convert from a global struct to a subclass of
10843 gimple_opt_pass along with...
10844 (pass_data_split_crit_edges): ...new pass_data instance and...
10845 (make_pass_split_crit_edges): ...new function.
10846 (pass_warn_function_return): Convert from a global struct to a
10847 subclass of gimple_opt_pass along with...
10848 (pass_data_warn_function_return): ...new pass_data instance and...
10849 (make_pass_warn_function_return): ...new function.
10850 (pass_warn_function_noreturn): Convert from a global struct to a
10851 subclass of gimple_opt_pass along with...
10852 (pass_data_warn_function_noreturn): ...new pass_data instance and...
10853 (make_pass_warn_function_noreturn): ...new function.
10854 (pass_warn_unused_result): Convert from a global struct to a subclass
10855 of gimple_opt_pass along with...
10856 (pass_data_warn_unused_result): ...new pass_data instance and...
10857 (make_pass_warn_unused_result): ...new function.
10858 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
10859 a subclass of gimple_opt_pass along with...
10860 (pass_data_merge_phi): ...new pass_data instance and...
10861 (make_pass_merge_phi): ...new function.
10862 * tree-complex.c (pass_lower_complex): Convert from a global struct to
10863 a subclass of gimple_opt_pass along with...
10864 (pass_data_lower_complex): ...new pass_data instance and...
10865 (make_pass_lower_complex): ...new function.
10866 (pass_lower_complex_O0): Convert from a global struct to a subclass of
10867 gimple_opt_pass along with...
10868 (pass_data_lower_complex_O0): ...new pass_data instance and...
10869 (make_pass_lower_complex_O0): ...new function.
10870 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
10871 subclass of gimple_opt_pass along with...
10872 (pass_data_lower_eh): ...new pass_data instance and...
10873 (make_pass_lower_eh): ...new function.
10874 (pass_refactor_eh): Convert from a global struct to a subclass of
10875 gimple_opt_pass along with...
10876 (pass_data_refactor_eh): ...new pass_data instance and...
10877 (make_pass_refactor_eh): ...new function.
10878 (pass_lower_resx): Convert from a global struct to a subclass of
10879 gimple_opt_pass along with...
10880 (pass_data_lower_resx): ...new pass_data instance and...
10881 (make_pass_lower_resx): ...new function.
10882 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
10883 of gimple_opt_pass along with...
10884 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
10885 (make_pass_lower_eh_dispatch): ...new function.
10886 (pass_cleanup_eh): Convert from a global struct to a subclass of
10887 gimple_opt_pass along with...
10888 (pass_data_cleanup_eh): ...new pass_data instance and...
10889 (make_pass_cleanup_eh): ...new function.
10890 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
10891 to a subclass of simple_ipa_opt_pass along with...
10892 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
10893 (make_pass_ipa_lower_emutls): ...new function.
10894 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
10895 a subclass of gimple_opt_pass along with...
10896 (pass_data_if_conversion): ...new pass_data instance and...
10897 (make_pass_if_conversion): ...new function.
10898 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
10899 subclass of gimple_opt_pass along with...
10900 (pass_data_build_ssa): ...new pass_data instance and...
10901 (make_pass_build_ssa): ...new function.
10902 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
10903 global struct to a subclass of gimple_opt_pass along with...
10904 (pass_data_loop_distribution): ...new pass_data instance and...
10905 (make_pass_loop_distribution): ...new function.
10906 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
10907 subclass of gimple_opt_pass along with...
10908 (pass_data_mudflap_1): ...new pass_data instance and...
10909 (make_pass_mudflap_1): ...new function.
10910 (pass_mudflap_2): Convert from a global struct to a subclass of
10911 gimple_opt_pass along with...
10912 (pass_data_mudflap_2): ...new pass_data instance and...
10913 (make_pass_mudflap_2): ...new function.
10914 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
10915 subclass of gimple_opt_pass along with...
10916 (pass_data_mudflap_1): ...new pass_data instance and...
10917 (make_pass_mudflap_1): ...new function.
10918 (pass_mudflap_2): Convert from a global struct to a subclass of
10919 gimple_opt_pass along with...
10920 (pass_data_mudflap_2): ...new pass_data instance and...
10921 (make_pass_mudflap_2): ...new function.
10922 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
10923 gimple_opt_pass along with...
10924 (pass_data_nrv): ...new pass_data instance and...
10925 (make_pass_nrv): ...new function.
10926 (pass_return_slot): Convert from a global struct to a subclass of
10927 gimple_opt_pass along with...
10928 (pass_data_return_slot): ...new pass_data instance and...
10929 (make_pass_return_slot): ...new function.
10930 * tree-object-size.c (pass_object_sizes): Convert from a global struct
10931 to a subclass of gimple_opt_pass along with...
10932 (pass_data_object_sizes): ...new pass_data instance and...
10933 (make_pass_object_sizes): ...new function.
10934 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
10935 global struct to a subclass of gimple_opt_pass along with...
10936 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
10938 (make_pass_cleanup_cfg_post_optimizing): ...new function.
10939 (pass_fixup_cfg): Convert from a global struct to a subclass of
10940 gimple_opt_pass along with...
10941 (pass_data_fixup_cfg): ...new pass_data instance and...
10942 (make_pass_fixup_cfg): ...new function.
10943 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
10944 (make_pass_mudflap_1): ...new function.
10945 (pass_mudflap_2): Replace declaration with that of...
10946 (make_pass_mudflap_2): ...new function.
10947 (pass_asan): Replace declaration with that of...
10948 (make_pass_asan): ...new function.
10949 (pass_asan_O0): Replace declaration with that of...
10950 (make_pass_asan_O0): ...new function.
10951 (pass_tsan): Replace declaration with that of...
10952 (make_pass_tsan): ...new function.
10953 (pass_tsan_O0): Replace declaration with that of...
10954 (make_pass_tsan_O0): ...new function.
10955 (pass_lower_cf): Replace declaration with that of...
10956 (make_pass_lower_cf): ...new function.
10957 (pass_refactor_eh): Replace declaration with that of...
10958 (make_pass_refactor_eh): ...new function.
10959 (pass_lower_eh): Replace declaration with that of...
10960 (make_pass_lower_eh): ...new function.
10961 (pass_lower_eh_dispatch): Replace declaration with that of...
10962 (make_pass_lower_eh_dispatch): ...new function.
10963 (pass_lower_resx): Replace declaration with that of...
10964 (make_pass_lower_resx): ...new function.
10965 (pass_build_cfg): Replace declaration with that of...
10966 (make_pass_build_cfg): ...new function.
10967 (pass_early_tree_profile): Replace declaration with that of...
10968 (make_pass_early_tree_profile): ...new function.
10969 (pass_cleanup_eh): Replace declaration with that of...
10970 (make_pass_cleanup_eh): ...new function.
10971 (pass_sra): Replace declaration with that of...
10972 (make_pass_sra): ...new function.
10973 (pass_sra_early): Replace declaration with that of...
10974 (make_pass_sra_early): ...new function.
10975 (pass_early_ipa_sra): Replace declaration with that of...
10976 (make_pass_early_ipa_sra): ...new function.
10977 (pass_tail_recursion): Replace declaration with that of...
10978 (make_pass_tail_recursion): ...new function.
10979 (pass_tail_calls): Replace declaration with that of...
10980 (make_pass_tail_calls): ...new function.
10981 (pass_tree_loop): Replace declaration with that of...
10982 (make_pass_tree_loop): ...new function.
10983 (pass_tree_loop_init): Replace declaration with that of...
10984 (make_pass_tree_loop_init): ...new function.
10985 (pass_lim): Replace declaration with that of...
10986 (make_pass_lim): ...new function.
10987 (pass_tree_unswitch): Replace declaration with that of...
10988 (make_pass_tree_unswitch): ...new function.
10989 (pass_predcom): Replace declaration with that of...
10990 (make_pass_predcom): ...new function.
10991 (pass_iv_canon): Replace declaration with that of...
10992 (make_pass_iv_canon): ...new function.
10993 (pass_scev_cprop): Replace declaration with that of...
10994 (make_pass_scev_cprop): ...new function.
10995 (pass_empty_loop): Replace declaration with that of...
10996 (make_pass_empty_loop): ...new function.
10997 (pass_record_bounds): Replace declaration with that of...
10998 (make_pass_record_bounds): ...new function.
10999 (pass_graphite): Replace declaration with that of...
11000 (make_pass_graphite): ...new function.
11001 (pass_graphite_transforms): Replace declaration with that of...
11002 (make_pass_graphite_transforms): ...new function.
11003 (pass_if_conversion): Replace declaration with that of...
11004 (make_pass_if_conversion): ...new function.
11005 (pass_loop_distribution): Replace declaration with that of...
11006 (make_pass_loop_distribution): ...new function.
11007 (pass_vectorize): Replace declaration with that of...
11008 (make_pass_vectorize): ...new function.
11009 (pass_slp_vectorize): Replace declaration with that of...
11010 (make_pass_slp_vectorize): ...new function.
11011 (pass_complete_unroll): Replace declaration with that of...
11012 (make_pass_complete_unroll): ...new function.
11013 (pass_complete_unrolli): Replace declaration with that of...
11014 (make_pass_complete_unrolli): ...new function.
11015 (pass_parallelize_loops): Replace declaration with that of...
11016 (make_pass_parallelize_loops): ...new function.
11017 (pass_loop_prefetch): Replace declaration with that of...
11018 (make_pass_loop_prefetch): ...new function.
11019 (pass_iv_optimize): Replace declaration with that of...
11020 (make_pass_iv_optimize): ...new function.
11021 (pass_tree_loop_done): Replace declaration with that of...
11022 (make_pass_tree_loop_done): ...new function.
11023 (pass_ch): Replace declaration with that of...
11024 (make_pass_ch): ...new function.
11025 (pass_ccp): Replace declaration with that of...
11026 (make_pass_ccp): ...new function.
11027 (pass_phi_only_cprop): Replace declaration with that of...
11028 (make_pass_phi_only_cprop): ...new function.
11029 (pass_build_ssa): Replace declaration with that of...
11030 (make_pass_build_ssa): ...new function.
11031 (pass_build_alias): Replace declaration with that of...
11032 (make_pass_build_alias): ...new function.
11033 (pass_build_ealias): Replace declaration with that of...
11034 (make_pass_build_ealias): ...new function.
11035 (pass_dominator): Replace declaration with that of...
11036 (make_pass_dominator): ...new function.
11037 (pass_dce): Replace declaration with that of...
11038 (make_pass_dce): ...new function.
11039 (pass_dce_loop): Replace declaration with that of...
11040 (make_pass_dce_loop): ...new function.
11041 (pass_cd_dce): Replace declaration with that of...
11042 (make_pass_cd_dce): ...new function.
11043 (pass_call_cdce): Replace declaration with that of...
11044 (make_pass_call_cdce): ...new function.
11045 (pass_merge_phi): Replace declaration with that of...
11046 (make_pass_merge_phi): ...new function.
11047 (pass_split_crit_edges): Replace declaration with that of...
11048 (make_pass_split_crit_edges): ...new function.
11049 (pass_pre): Replace declaration with that of...
11050 (make_pass_pre): ...new function.
11051 (pass_profile): Replace declaration with that of...
11052 (make_pass_profile): ...new function.
11053 (pass_strip_predict_hints): Replace declaration with that of...
11054 (make_pass_strip_predict_hints): ...new function.
11055 (pass_lower_complex_O0): Replace declaration with that of...
11056 (make_pass_lower_complex_O0): ...new function.
11057 (pass_lower_complex): Replace declaration with that of...
11058 (make_pass_lower_complex): ...new function.
11059 (pass_lower_vector): Replace declaration with that of...
11060 (make_pass_lower_vector): ...new function.
11061 (pass_lower_vector_ssa): Replace declaration with that of...
11062 (make_pass_lower_vector_ssa): ...new function.
11063 (pass_lower_omp): Replace declaration with that of...
11064 (make_pass_lower_omp): ...new function.
11065 (pass_diagnose_omp_blocks): Replace declaration with that of...
11066 (make_pass_diagnose_omp_blocks): ...new function.
11067 (pass_expand_omp): Replace declaration with that of...
11068 (make_pass_expand_omp): ...new function.
11069 (pass_expand_omp_ssa): Replace declaration with that of...
11070 (make_pass_expand_omp_ssa): ...new function.
11071 (pass_object_sizes): Replace declaration with that of...
11072 (make_pass_object_sizes): ...new function.
11073 (pass_strlen): Replace declaration with that of...
11074 (make_pass_strlen): ...new function.
11075 (pass_fold_builtins): Replace declaration with that of...
11076 (make_pass_fold_builtins): ...new function.
11077 (pass_stdarg): Replace declaration with that of...
11078 (make_pass_stdarg): ...new function.
11079 (pass_early_warn_uninitialized): Replace declaration with that of...
11080 (make_pass_early_warn_uninitialized): ...new function.
11081 (pass_late_warn_uninitialized): Replace declaration with that of...
11082 (make_pass_late_warn_uninitialized): ...new function.
11083 (pass_cse_reciprocals): Replace declaration with that of...
11084 (make_pass_cse_reciprocals): ...new function.
11085 (pass_cse_sincos): Replace declaration with that of...
11086 (make_pass_cse_sincos): ...new function.
11087 (pass_optimize_bswap): Replace declaration with that of...
11088 (make_pass_optimize_bswap): ...new function.
11089 (pass_optimize_widening_mul): Replace declaration with that of...
11090 (make_pass_optimize_widening_mul): ...new function.
11091 (pass_warn_function_return): Replace declaration with that of...
11092 (make_pass_warn_function_return): ...new function.
11093 (pass_warn_function_noreturn): Replace declaration with that of...
11094 (make_pass_warn_function_noreturn): ...new function.
11095 (pass_cselim): Replace declaration with that of...
11096 (make_pass_cselim): ...new function.
11097 (pass_phiopt): Replace declaration with that of...
11098 (make_pass_phiopt): ...new function.
11099 (pass_forwprop): Replace declaration with that of...
11100 (make_pass_forwprop): ...new function.
11101 (pass_phiprop): Replace declaration with that of...
11102 (make_pass_phiprop): ...new function.
11103 (pass_tree_ifcombine): Replace declaration with that of...
11104 (make_pass_tree_ifcombine): ...new function.
11105 (pass_dse): Replace declaration with that of...
11106 (make_pass_dse): ...new function.
11107 (pass_nrv): Replace declaration with that of...
11108 (make_pass_nrv): ...new function.
11109 (pass_rename_ssa_copies): Replace declaration with that of...
11110 (make_pass_rename_ssa_copies): ...new function.
11111 (pass_sink_code): Replace declaration with that of...
11112 (make_pass_sink_code): ...new function.
11113 (pass_fre): Replace declaration with that of...
11114 (make_pass_fre): ...new function.
11115 (pass_check_data_deps): Replace declaration with that of...
11116 (make_pass_check_data_deps): ...new function.
11117 (pass_copy_prop): Replace declaration with that of...
11118 (make_pass_copy_prop): ...new function.
11119 (pass_vrp): Replace declaration with that of...
11120 (make_pass_vrp): ...new function.
11121 (pass_uncprop): Replace declaration with that of...
11122 (make_pass_uncprop): ...new function.
11123 (pass_return_slot): Replace declaration with that of...
11124 (make_pass_return_slot): ...new function.
11125 (pass_reassoc): Replace declaration with that of...
11126 (make_pass_reassoc): ...new function.
11127 (pass_rebuild_cgraph_edges): Replace declaration with that of...
11128 (make_pass_rebuild_cgraph_edges): ...new function.
11129 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
11130 (make_pass_remove_cgraph_callee_edges): ...new function.
11131 (pass_build_cgraph_edges): Replace declaration with that of...
11132 (make_pass_build_cgraph_edges): ...new function.
11133 (pass_local_pure_const): Replace declaration with that of...
11134 (make_pass_local_pure_const): ...new function.
11135 (pass_tracer): Replace declaration with that of...
11136 (make_pass_tracer): ...new function.
11137 (pass_warn_unused_result): Replace declaration with that of...
11138 (make_pass_warn_unused_result): ...new function.
11139 (pass_diagnose_tm_blocks): Replace declaration with that of...
11140 (make_pass_diagnose_tm_blocks): ...new function.
11141 (pass_lower_tm): Replace declaration with that of...
11142 (make_pass_lower_tm): ...new function.
11143 (pass_tm_init): Replace declaration with that of...
11144 (make_pass_tm_init): ...new function.
11145 (pass_tm_mark): Replace declaration with that of...
11146 (make_pass_tm_mark): ...new function.
11147 (pass_tm_memopt): Replace declaration with that of...
11148 (make_pass_tm_memopt): ...new function.
11149 (pass_tm_edges): Replace declaration with that of...
11150 (make_pass_tm_edges): ...new function.
11151 (pass_split_functions): Replace declaration with that of...
11152 (make_pass_split_functions): ...new function.
11153 (pass_feedback_split_functions): Replace declaration with that of...
11154 (make_pass_feedback_split_functions): ...new function.
11155 (pass_strength_reduction): Replace declaration with that of...
11156 (make_pass_strength_reduction): ...new function.
11157 (pass_ipa_lower_emutls): Replace declaration with that of...
11158 (make_pass_ipa_lower_emutls): ...new function.
11159 (pass_ipa_function_and_variable_visibility): Replace declaration with
11161 (make_pass_ipa_function_and_variable_visibility): ...new function.
11162 (pass_ipa_tree_profile): Replace declaration with that of...
11163 (make_pass_ipa_tree_profile): ...new function.
11164 (pass_early_local_passes): Replace declaration with that of...
11165 (make_pass_early_local_passes): ...new function.
11166 (pass_ipa_whole_program_visibility): Replace declaration with that
11168 (make_pass_ipa_whole_program_visibility): ...new function.
11169 (pass_ipa_lto_gimple_out): Replace declaration with that of...
11170 (make_pass_ipa_lto_gimple_out): ...new function.
11171 (pass_ipa_increase_alignment): Replace declaration with that of...
11172 (make_pass_ipa_increase_alignment): ...new function.
11173 (pass_ipa_inline): Replace declaration with that of...
11174 (make_pass_ipa_inline): ...new function.
11175 (pass_ipa_free_lang_data): Replace declaration with that of...
11176 (make_pass_ipa_free_lang_data): ...new function.
11177 (pass_ipa_free_inline_summary): Replace declaration with that of...
11178 (make_pass_ipa_free_inline_summary): ...new function.
11179 (pass_ipa_cp): Replace declaration with that of...
11180 (make_pass_ipa_cp): ...new function.
11181 (pass_ipa_reference): Replace declaration with that of...
11182 (make_pass_ipa_reference): ...new function.
11183 (pass_ipa_pure_const): Replace declaration with that of...
11184 (make_pass_ipa_pure_const): ...new function.
11185 (pass_ipa_pta): Replace declaration with that of...
11186 (make_pass_ipa_pta): ...new function.
11187 (pass_ipa_lto_finish_out): Replace declaration with that of...
11188 (make_pass_ipa_lto_finish_out): ...new function.
11189 (pass_ipa_tm): Replace declaration with that of...
11190 (make_pass_ipa_tm): ...new function.
11191 (pass_ipa_profile): Replace declaration with that of...
11192 (make_pass_ipa_profile): ...new function.
11193 (pass_ipa_cdtor_merge): Replace declaration with that of...
11194 (make_pass_ipa_cdtor_merge): ...new function.
11195 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
11197 (make_pass_cleanup_cfg_post_optimizing): ...new function.
11198 (pass_init_datastructures): Replace declaration with that of...
11199 (make_pass_init_datastructures): ...new function.
11200 (pass_fixup_cfg): Replace declaration with that of...
11201 (make_pass_fixup_cfg): ...new function.
11202 (pass_expand): Replace declaration with that of...
11203 (make_pass_expand): ...new function.
11204 (pass_instantiate_virtual_regs): Replace declaration with that of...
11205 (make_pass_instantiate_virtual_regs): ...new function.
11206 (pass_rtl_fwprop): Replace declaration with that of...
11207 (make_pass_rtl_fwprop): ...new function.
11208 (pass_rtl_fwprop_addr): Replace declaration with that of...
11209 (make_pass_rtl_fwprop_addr): ...new function.
11210 (pass_jump): Replace declaration with that of...
11211 (make_pass_jump): ...new function.
11212 (pass_jump2): Replace declaration with that of...
11213 (make_pass_jump2): ...new function.
11214 (pass_lower_subreg): Replace declaration with that of...
11215 (make_pass_lower_subreg): ...new function.
11216 (pass_cse): Replace declaration with that of...
11217 (make_pass_cse): ...new function.
11218 (pass_fast_rtl_dce): Replace declaration with that of...
11219 (make_pass_fast_rtl_dce): ...new function.
11220 (pass_ud_rtl_dce): Replace declaration with that of...
11221 (make_pass_ud_rtl_dce): ...new function.
11222 (pass_rtl_dce): Replace declaration with that of...
11223 (make_pass_rtl_dce): ...new function.
11224 (pass_rtl_dse1): Replace declaration with that of...
11225 (make_pass_rtl_dse1): ...new function.
11226 (pass_rtl_dse2): Replace declaration with that of...
11227 (make_pass_rtl_dse2): ...new function.
11228 (pass_rtl_dse3): Replace declaration with that of...
11229 (make_pass_rtl_dse3): ...new function.
11230 (pass_rtl_cprop): Replace declaration with that of...
11231 (make_pass_rtl_cprop): ...new function.
11232 (pass_rtl_pre): Replace declaration with that of...
11233 (make_pass_rtl_pre): ...new function.
11234 (pass_rtl_hoist): Replace declaration with that of...
11235 (make_pass_rtl_hoist): ...new function.
11236 (pass_rtl_store_motion): Replace declaration with that of...
11237 (make_pass_rtl_store_motion): ...new function.
11238 (pass_cse_after_global_opts): Replace declaration with that of...
11239 (make_pass_cse_after_global_opts): ...new function.
11240 (pass_rtl_ifcvt): Replace declaration with that of...
11241 (make_pass_rtl_ifcvt): ...new function.
11242 (pass_into_cfg_layout_mode): Replace declaration with that of...
11243 (make_pass_into_cfg_layout_mode): ...new function.
11244 (pass_outof_cfg_layout_mode): Replace declaration with that of...
11245 (make_pass_outof_cfg_layout_mode): ...new function.
11246 (pass_loop2): Replace declaration with that of...
11247 (make_pass_loop2): ...new function.
11248 (pass_rtl_loop_init): Replace declaration with that of...
11249 (make_pass_rtl_loop_init): ...new function.
11250 (pass_rtl_move_loop_invariants): Replace declaration with that of...
11251 (make_pass_rtl_move_loop_invariants): ...new function.
11252 (pass_rtl_unswitch): Replace declaration with that of...
11253 (make_pass_rtl_unswitch): ...new function.
11254 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
11255 (make_pass_rtl_unroll_and_peel_loops): ...new function.
11256 (pass_rtl_doloop): Replace declaration with that of...
11257 (make_pass_rtl_doloop): ...new function.
11258 (pass_rtl_loop_done): Replace declaration with that of...
11259 (make_pass_rtl_loop_done): ...new function.
11260 (pass_web): Replace declaration with that of...
11261 (make_pass_web): ...new function.
11262 (pass_cse2): Replace declaration with that of...
11263 (make_pass_cse2): ...new function.
11264 (pass_df_initialize_opt): Replace declaration with that of...
11265 (make_pass_df_initialize_opt): ...new function.
11266 (pass_df_initialize_no_opt): Replace declaration with that of...
11267 (make_pass_df_initialize_no_opt): ...new function.
11268 (pass_reginfo_init): Replace declaration with that of...
11269 (make_pass_reginfo_init): ...new function.
11270 (pass_inc_dec): Replace declaration with that of...
11271 (make_pass_inc_dec): ...new function.
11272 (pass_stack_ptr_mod): Replace declaration with that of...
11273 (make_pass_stack_ptr_mod): ...new function.
11274 (pass_initialize_regs): Replace declaration with that of...
11275 (make_pass_initialize_regs): ...new function.
11276 (pass_combine): Replace declaration with that of...
11277 (make_pass_combine): ...new function.
11278 (pass_if_after_combine): Replace declaration with that of...
11279 (make_pass_if_after_combine): ...new function.
11280 (pass_ree): Replace declaration with that of...
11281 (make_pass_ree): ...new function.
11282 (pass_partition_blocks): Replace declaration with that of...
11283 (make_pass_partition_blocks): ...new function.
11284 (pass_match_asm_constraints): Replace declaration with that of...
11285 (make_pass_match_asm_constraints): ...new function.
11286 (pass_regmove): Replace declaration with that of...
11287 (make_pass_regmove): ...new function.
11288 (pass_split_all_insns): Replace declaration with that of...
11289 (make_pass_split_all_insns): ...new function.
11290 (pass_fast_rtl_byte_dce): Replace declaration with that of...
11291 (make_pass_fast_rtl_byte_dce): ...new function.
11292 (pass_lower_subreg2): Replace declaration with that of...
11293 (make_pass_lower_subreg2): ...new function.
11294 (pass_mode_switching): Replace declaration with that of...
11295 (make_pass_mode_switching): ...new function.
11296 (pass_sms): Replace declaration with that of...
11297 (make_pass_sms): ...new function.
11298 (pass_sched): Replace declaration with that of...
11299 (make_pass_sched): ...new function.
11300 (pass_ira): Replace declaration with that of...
11301 (make_pass_ira): ...new function.
11302 (pass_reload): Replace declaration with that of...
11303 (make_pass_reload): ...new function.
11304 (pass_clean_state): Replace declaration with that of...
11305 (make_pass_clean_state): ...new function.
11306 (pass_branch_prob): Replace declaration with that of...
11307 (make_pass_branch_prob): ...new function.
11308 (pass_value_profile_transformations): Replace declaration with that
11310 (make_pass_value_profile_transformations): ...new function.
11311 (pass_postreload_cse): Replace declaration with that of...
11312 (make_pass_postreload_cse): ...new function.
11313 (pass_gcse2): Replace declaration with that of...
11314 (make_pass_gcse2): ...new function.
11315 (pass_split_after_reload): Replace declaration with that of...
11316 (make_pass_split_after_reload): ...new function.
11317 (pass_branch_target_load_optimize1): Replace declaration with that
11319 (make_pass_branch_target_load_optimize1): ...new function.
11320 (pass_thread_prologue_and_epilogue): Replace declaration with that
11322 (make_pass_thread_prologue_and_epilogue): ...new function.
11323 (pass_stack_adjustments): Replace declaration with that of...
11324 (make_pass_stack_adjustments): ...new function.
11325 (pass_peephole2): Replace declaration with that of...
11326 (make_pass_peephole2): ...new function.
11327 (pass_if_after_reload): Replace declaration with that of...
11328 (make_pass_if_after_reload): ...new function.
11329 (pass_regrename): Replace declaration with that of...
11330 (make_pass_regrename): ...new function.
11331 (pass_cprop_hardreg): Replace declaration with that of...
11332 (make_pass_cprop_hardreg): ...new function.
11333 (pass_reorder_blocks): Replace declaration with that of...
11334 (make_pass_reorder_blocks): ...new function.
11335 (pass_branch_target_load_optimize2): Replace declaration with that
11337 (make_pass_branch_target_load_optimize2): ...new function.
11338 (pass_leaf_regs): Replace declaration with that of...
11339 (make_pass_leaf_regs): ...new function.
11340 (pass_split_before_sched2): Replace declaration with that of...
11341 (make_pass_split_before_sched2): ...new function.
11342 (pass_compare_elim_after_reload): Replace declaration with that of...
11343 (make_pass_compare_elim_after_reload): ...new function.
11344 (pass_sched2): Replace declaration with that of...
11345 (make_pass_sched2): ...new function.
11346 (pass_stack_regs): Replace declaration with that of...
11347 (make_pass_stack_regs): ...new function.
11348 (pass_stack_regs_run): Replace declaration with that of...
11349 (make_pass_stack_regs_run): ...new function.
11350 (pass_df_finish): Replace declaration with that of...
11351 (make_pass_df_finish): ...new function.
11352 (pass_compute_alignments): Replace declaration with that of...
11353 (make_pass_compute_alignments): ...new function.
11354 (pass_duplicate_computed_gotos): Replace declaration with that of...
11355 (make_pass_duplicate_computed_gotos): ...new function.
11356 (pass_variable_tracking): Replace declaration with that of...
11357 (make_pass_variable_tracking): ...new function.
11358 (pass_free_cfg): Replace declaration with that of...
11359 (make_pass_free_cfg): ...new function.
11360 (pass_machine_reorg): Replace declaration with that of...
11361 (make_pass_machine_reorg): ...new function.
11362 (pass_cleanup_barriers): Replace declaration with that of...
11363 (make_pass_cleanup_barriers): ...new function.
11364 (pass_delay_slots): Replace declaration with that of...
11365 (make_pass_delay_slots): ...new function.
11366 (pass_split_for_shorten_branches): Replace declaration with that of...
11367 (make_pass_split_for_shorten_branches): ...new function.
11368 (pass_split_before_regstack): Replace declaration with that of...
11369 (make_pass_split_before_regstack): ...new function.
11370 (pass_convert_to_eh_region_ranges): Replace declaration with that
11372 (make_pass_convert_to_eh_region_ranges): ...new function.
11373 (pass_shorten_branches): Replace declaration with that of...
11374 (make_pass_shorten_branches): ...new function.
11375 (pass_set_nothrow_function_flags): Replace declaration with that of...
11376 (make_pass_set_nothrow_function_flags): ...new function.
11377 (pass_dwarf2_frame): Replace declaration with that of...
11378 (make_pass_dwarf2_frame): ...new function.
11379 (pass_final): Replace declaration with that of...
11380 (make_pass_final): ...new function.
11381 (pass_rtl_seqabstr): Replace declaration with that of...
11382 (make_pass_rtl_seqabstr): ...new function.
11383 (pass_release_ssa_names): Replace declaration with that of...
11384 (make_pass_release_ssa_names): ...new function.
11385 (pass_early_inline): Replace declaration with that of...
11386 (make_pass_early_inline): ...new function.
11387 (pass_inline_parameters): Replace declaration with that of...
11388 (make_pass_inline_parameters): ...new function.
11389 (pass_update_address_taken): Replace declaration with that of...
11390 (make_pass_update_address_taken): ...new function.
11391 (pass_convert_switch): Replace declaration with that of...
11392 (make_pass_convert_switch): ...new function.
11393 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
11394 to a subclass of simple_ipa_opt_pass along with...
11395 (pass_data_ipa_tree_profile): ...new pass_data instance and...
11396 (make_pass_ipa_tree_profile): ...new function.
11397 * tree-sra.c (pass_sra_early): Convert from a global struct to a
11398 subclass of gimple_opt_pass along with...
11399 (pass_data_sra_early): ...new pass_data instance and...
11400 (make_pass_sra_early): ...new function.
11401 (pass_sra): Convert from a global struct to a subclass of
11402 gimple_opt_pass along with...
11403 (pass_data_sra): ...new pass_data instance and...
11404 (make_pass_sra): ...new function.
11405 (pass_early_ipa_sra): Convert from a global struct to a subclass of
11406 gimple_opt_pass along with...
11407 (pass_data_early_ipa_sra): ...new pass_data instance and...
11408 (make_pass_early_ipa_sra): ...new function.
11409 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
11410 subclass of gimple_opt_pass along with...
11411 (pass_data_ccp): ...new pass_data instance and...
11412 (make_pass_ccp): ...new function.
11413 (pass_fold_builtins): Convert from a global struct to a subclass of
11414 gimple_opt_pass along with...
11415 (pass_data_fold_builtins): ...new pass_data instance and...
11416 (make_pass_fold_builtins): ...new function.
11417 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
11418 subclass of gimple_opt_pass along with...
11419 (pass_data_copy_prop): ...new pass_data instance and...
11420 (make_pass_copy_prop): ...new function.
11421 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
11422 global struct to a subclass of gimple_opt_pass along with...
11423 (pass_data_rename_ssa_copies): ...new pass_data instance and...
11424 (make_pass_rename_ssa_copies): ...new function.
11425 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
11426 subclass of gimple_opt_pass along with...
11427 (pass_data_dce): ...new pass_data instance and...
11428 (make_pass_dce): ...new function.
11429 (pass_dce_loop): Convert from a global struct to a subclass of
11430 gimple_opt_pass along with...
11431 (pass_data_dce_loop): ...new pass_data instance and...
11432 (make_pass_dce_loop): ...new function.
11433 (pass_cd_dce): Convert from a global struct to a subclass of
11434 gimple_opt_pass along with...
11435 (pass_data_cd_dce): ...new pass_data instance and...
11436 (make_pass_cd_dce): ...new function.
11437 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
11438 subclass of gimple_opt_pass along with...
11439 (pass_data_dominator): ...new pass_data instance and...
11440 (make_pass_dominator): ...new function.
11441 (pass_phi_only_cprop): Convert from a global struct to a subclass of
11442 gimple_opt_pass along with...
11443 (pass_data_phi_only_cprop): ...new pass_data instance and...
11444 (make_pass_phi_only_cprop): ...new function.
11445 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
11446 subclass of gimple_opt_pass along with...
11447 (pass_data_dse): ...new pass_data instance and...
11448 (make_pass_dse): ...new function.
11449 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
11450 a subclass of gimple_opt_pass along with...
11451 (pass_data_forwprop): ...new pass_data instance and...
11452 (make_pass_forwprop): ...new function.
11453 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
11454 struct to a subclass of gimple_opt_pass along with...
11455 (pass_data_tree_ifcombine): ...new pass_data instance and...
11456 (make_pass_tree_ifcombine): ...new function.
11457 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
11458 subclass of gimple_opt_pass along with...
11459 (pass_data_ch): ...new pass_data instance and...
11460 (make_pass_ch): ...new function.
11461 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
11462 subclass of gimple_opt_pass along with...
11463 (pass_data_tree_loop): ...new pass_data instance and...
11464 (make_pass_tree_loop): ...new function.
11465 (pass_tree_loop_init): Convert from a global struct to a subclass of
11466 gimple_opt_pass along with...
11467 (pass_data_tree_loop_init): ...new pass_data instance and...
11468 (make_pass_tree_loop_init): ...new function.
11469 (pass_lim): Convert from a global struct to a subclass of
11470 gimple_opt_pass along with...
11471 (pass_data_lim): ...new pass_data instance and...
11472 (make_pass_lim): ...new function.
11473 (pass_tree_unswitch): Convert from a global struct to a subclass of
11474 gimple_opt_pass along with...
11475 (pass_data_tree_unswitch): ...new pass_data instance and...
11476 (make_pass_tree_unswitch): ...new function.
11477 (pass_predcom): Convert from a global struct to a subclass of
11478 gimple_opt_pass along with...
11479 (pass_data_predcom): ...new pass_data instance and...
11480 (make_pass_predcom): ...new function.
11481 (pass_vectorize): Convert from a global struct to a subclass of
11482 gimple_opt_pass along with...
11483 (pass_data_vectorize): ...new pass_data instance and...
11484 (make_pass_vectorize): ...new function.
11485 (pass_graphite): Convert from a global struct to a subclass of
11486 gimple_opt_pass along with...
11487 (pass_data_graphite): ...new pass_data instance and...
11488 (make_pass_graphite): ...new function.
11489 (pass_graphite_transforms): Convert from a global struct to a subclass
11490 of gimple_opt_pass along with...
11491 (pass_data_graphite_transforms): ...new pass_data instance and...
11492 (make_pass_graphite_transforms): ...new function.
11493 (pass_check_data_deps): Convert from a global struct to a subclass of
11494 gimple_opt_pass along with...
11495 (pass_data_check_data_deps): ...new pass_data instance and...
11496 (make_pass_check_data_deps): ...new function.
11497 (pass_iv_canon): Convert from a global struct to a subclass of
11498 gimple_opt_pass along with...
11499 (pass_data_iv_canon): ...new pass_data instance and...
11500 (make_pass_iv_canon): ...new function.
11501 (pass_scev_cprop): Convert from a global struct to a subclass of
11502 gimple_opt_pass along with...
11503 (pass_data_scev_cprop): ...new pass_data instance and...
11504 (make_pass_scev_cprop): ...new function.
11505 (pass_record_bounds): Convert from a global struct to a subclass of
11506 gimple_opt_pass along with...
11507 (pass_data_record_bounds): ...new pass_data instance and...
11508 (make_pass_record_bounds): ...new function.
11509 (pass_complete_unroll): Convert from a global struct to a subclass of
11510 gimple_opt_pass along with...
11511 (pass_data_complete_unroll): ...new pass_data instance and...
11512 (make_pass_complete_unroll): ...new function.
11513 (pass_complete_unrolli): Convert from a global struct to a subclass of
11514 gimple_opt_pass along with...
11515 (pass_data_complete_unrolli): ...new pass_data instance and...
11516 (make_pass_complete_unrolli): ...new function.
11517 (pass_parallelize_loops): Convert from a global struct to a subclass
11518 of gimple_opt_pass along with...
11519 (pass_data_parallelize_loops): ...new pass_data instance and...
11520 (make_pass_parallelize_loops): ...new function.
11521 (pass_loop_prefetch): Convert from a global struct to a subclass of
11522 gimple_opt_pass along with...
11523 (pass_data_loop_prefetch): ...new pass_data instance and...
11524 (make_pass_loop_prefetch): ...new function.
11525 (pass_iv_optimize): Convert from a global struct to a subclass of
11526 gimple_opt_pass along with...
11527 (pass_data_iv_optimize): ...new pass_data instance and...
11528 (make_pass_iv_optimize): ...new function.
11529 (pass_tree_loop_done): Convert from a global struct to a subclass of
11530 gimple_opt_pass along with...
11531 (pass_data_tree_loop_done): ...new pass_data instance and...
11532 (make_pass_tree_loop_done): ...new function.
11533 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
11534 struct to a subclass of gimple_opt_pass along with...
11535 (pass_data_cse_reciprocals): ...new pass_data instance and...
11536 (make_pass_cse_reciprocals): ...new function.
11537 (pass_cse_sincos): Convert from a global struct to a subclass of
11538 gimple_opt_pass along with...
11539 (pass_data_cse_sincos): ...new pass_data instance and...
11540 (make_pass_cse_sincos): ...new function.
11541 (pass_optimize_bswap): Convert from a global struct to a subclass of
11542 gimple_opt_pass along with...
11543 (pass_data_optimize_bswap): ...new pass_data instance and...
11544 (make_pass_optimize_bswap): ...new function.
11545 (pass_optimize_widening_mul): Convert from a global struct to a
11546 subclass of gimple_opt_pass along with...
11547 (pass_data_optimize_widening_mul): ...new pass_data instance and...
11548 (make_pass_optimize_widening_mul): ...new function.
11549 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
11550 subclass of gimple_opt_pass along with...
11551 (pass_data_phiopt): ...new pass_data instance and...
11552 (make_pass_phiopt): ...new function.
11553 (pass_cselim): Convert from a global struct to a subclass of
11554 gimple_opt_pass along with...
11555 (pass_data_cselim): ...new pass_data instance and...
11556 (make_pass_cselim): ...new function.
11557 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
11558 subclass of gimple_opt_pass along with...
11559 (pass_data_phiprop): ...new pass_data instance and...
11560 (make_pass_phiprop): ...new function.
11561 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
11562 subclass of gimple_opt_pass along with...
11563 (pass_data_pre): ...new pass_data instance and...
11564 (make_pass_pre): ...new function.
11565 (pass_fre): Convert from a global struct to a subclass of
11566 gimple_opt_pass along with...
11567 (pass_data_fre): ...new pass_data instance and...
11568 (make_pass_fre): ...new function.
11569 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
11570 subclass of gimple_opt_pass along with...
11571 (pass_data_reassoc): ...new pass_data instance and...
11572 (make_pass_reassoc): ...new function.
11573 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
11574 subclass of gimple_opt_pass along with...
11575 (pass_data_sink_code): ...new pass_data instance and...
11576 (make_pass_sink_code): ...new function.
11577 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
11578 subclass of gimple_opt_pass along with...
11579 (pass_data_strlen): ...new pass_data instance and...
11580 (make_pass_strlen): ...new function.
11581 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
11582 struct to a subclass of gimple_opt_pass along with...
11583 (pass_data_build_alias): ...new pass_data instance and...
11584 (make_pass_build_alias): ...new function.
11585 (pass_build_ealias): Convert from a global struct to a subclass of
11586 gimple_opt_pass along with...
11587 (pass_data_build_ealias): ...new pass_data instance and...
11588 (make_pass_build_ealias): ...new function.
11589 (pass_ipa_pta): Convert from a global struct to a subclass of
11590 simple_ipa_opt_pass along with...
11591 (pass_data_ipa_pta): ...new pass_data instance and...
11592 (make_pass_ipa_pta): ...new function.
11593 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
11594 subclass of gimple_opt_pass along with...
11595 (pass_data_uncprop): ...new pass_data instance and...
11596 (make_pass_uncprop): ...new function.
11597 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
11598 global struct to a subclass of gimple_opt_pass along with...
11599 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
11600 (make_pass_late_warn_uninitialized): ...new function.
11601 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
11602 to a subclass of gimple_opt_pass along with...
11603 (pass_data_init_datastructures): ...new pass_data instance and...
11604 (make_pass_init_datastructures): ...new function.
11605 (pass_early_warn_uninitialized): Convert from a global struct to a
11606 subclass of gimple_opt_pass along with...
11607 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
11608 (make_pass_early_warn_uninitialized): ...new function.
11609 (pass_update_address_taken): Convert from a global struct to a
11610 subclass of gimple_opt_pass along with...
11611 (pass_data_update_address_taken): ...new pass_data instance and...
11612 (make_pass_update_address_taken): ...new function.
11613 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
11614 struct to a subclass of gimple_opt_pass along with...
11615 (pass_data_release_ssa_names): ...new pass_data instance and...
11616 (make_pass_release_ssa_names): ...new function.
11617 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
11618 subclass of gimple_opt_pass along with...
11619 (pass_data_stdarg): ...new pass_data instance and...
11620 (make_pass_stdarg): ...new function.
11621 * tree-switch-conversion.c (pass_convert_switch): Convert from a
11622 global struct to a subclass of gimple_opt_pass along with...
11623 (pass_data_convert_switch): ...new pass_data instance and...
11624 (make_pass_convert_switch): ...new function.
11625 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
11626 to a subclass of gimple_opt_pass along with...
11627 (pass_data_tail_recursion): ...new pass_data instance and...
11628 (make_pass_tail_recursion): ...new function.
11629 (pass_tail_calls): Convert from a global struct to a subclass of
11630 gimple_opt_pass along with...
11631 (pass_data_tail_calls): ...new pass_data instance and...
11632 (make_pass_tail_calls): ...new function.
11633 * tree-vect-generic.c (pass_lower_vector): Convert from a global
11634 struct to a subclass of gimple_opt_pass along with...
11635 (pass_data_lower_vector): ...new pass_data instance and...
11636 (make_pass_lower_vector): ...new function.
11637 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
11638 gimple_opt_pass along with...
11639 (pass_data_lower_vector_ssa): ...new pass_data instance and...
11640 (make_pass_lower_vector_ssa): ...new function.
11641 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
11642 to a subclass of gimple_opt_pass along with...
11643 (pass_data_slp_vectorize): ...new pass_data instance and...
11644 (make_pass_slp_vectorize): ...new function.
11645 (pass_ipa_increase_alignment): Convert from a global struct to a
11646 subclass of simple_ipa_opt_pass along with...
11647 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
11648 (make_pass_ipa_increase_alignment): ...new function.
11649 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
11650 gimple_opt_pass along with...
11651 (pass_data_vrp): ...new pass_data instance and...
11652 (make_pass_vrp): ...new function.
11653 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
11654 subclass of simple_ipa_opt_pass along with...
11655 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
11656 (make_pass_ipa_free_lang_data): ...new function.
11657 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
11658 gimple_opt_pass along with...
11659 (pass_data_tsan): ...new pass_data instance and...
11660 (make_pass_tsan): ...new function.
11661 (pass_tsan_O0): Convert from a global struct to a subclass of
11662 gimple_opt_pass along with...
11663 (pass_data_tsan_O0): ...new pass_data instance and...
11664 (make_pass_tsan_O0): ...new function.
11665 * var-tracking.c (pass_variable_tracking): Convert from a global
11666 struct to a subclass of rtl_opt_pass along with...
11667 (pass_data_variable_tracking): ...new pass_data instance and...
11668 (make_pass_variable_tracking): ...new function.
11669 * web.c (pass_web): Convert from a global struct to a subclass of
11670 rtl_opt_pass along with...
11671 (pass_data_web): ...new pass_data instance and...
11672 (make_pass_web): ...new function.
11673 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
11674 declaration with that of...
11675 (make_pass_mode_switch_use): ...new function.
11676 (pass_resolve_sw_modes): Replace declaration with that of...
11677 (make_pass_resolve_sw_modes): ...new function.
11678 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
11679 from a global struct to a subclass of rtl_opt_pass along with...
11680 (pass_data_mode_switch_use): ...new pass_data instance and...
11681 (make_pass_mode_switch_use): ...new function.
11682 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
11683 from a global struct to a subclass of rtl_opt_pass along with...
11684 (pass_data_resolve_sw_modes): ...new pass_data instance and...
11685 (make_pass_resolve_sw_modes): ...new function.
11686 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
11687 struct to a subclass of rtl_opt_pass along with...
11688 (pass_data_insert_vzeroupper): ...new pass_data instance and...
11689 (make_pass_insert_vzeroupper): ...new function.
11690 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
11691 global struct to a subclass of rtl_opt_pass along with...
11692 (pass_data_work_around_errata): ...new pass_data instance and...
11693 (make_pass_work_around_errata): ...new function.
11694 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
11695 struct to a subclass of rtl_opt_pass along with...
11696 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
11697 (make_pass_mips_machine_reorg2): ...new function.
11699 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11701 * passes.c (pass_manager::operator new): New.
11703 2013-08-05 David Malcolm <dmalcolm@redhat.com>
11705 Handwritten part of conversion of passes to C++ classes.
11707 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
11708 (toplev.o): Add dep on PASS_MANAGER_H.
11709 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
11710 of early local pases to reflect this moving from a global to a
11711 member of gcc::pass_manager.
11712 (cgraph_add_new_function): Likewise.
11713 * lto-cgraph.c (lto_output_node): Update for conversion of
11714 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
11715 * passes.c (opt_pass::clone): New.
11716 (opt_pass::gate): New.
11717 (opt_pass::execute): New.
11718 (opt_pass::opt_pass): New.
11719 (pass_manager::execute_early_local_passes): New.
11720 (pass_manager::execute_pass_mode_switching): new.
11721 (finish_optimization_passes): Convert to...
11722 (pass_manager::finish_optimization_passes): ...this.
11723 (finish_optimization_passes): Update for conversion of passes to
11725 (register_dump_files_1): Use has_gate since we cannot portably
11726 check a vtable entry against NULL.
11727 (dump_one_pass): Likewise.
11728 (ipa_write_summaries_2): Likewise.
11729 (ipa_write_optimization_summaries_1): Likewise.
11730 (ipa_read_summaries_1): Likewise.
11731 (ipa_read_optimization_summaries_1): Likewise.
11732 (execute_ipa_stmt_fixups): Likewise.
11733 (pass_manager::pass_manager): Rewrite pass-creation, invoking
11734 pass-creation functions rather than wiring up globals, and
11735 storing the results in fields of pass_manager generated using
11736 pass-instances.def.
11737 (pass_manager::dump_profile_report): Update for conversion of
11738 passes to C++ classes.
11739 (pass_manager::execute_ipa_summary_passes): Likewise.
11740 (execute_one_ipa_transform_pass): Likewise.
11741 (execute_one_pass): Use has_gate and has_execute since we cannot
11742 portably check a vtable entry against NULL.
11743 * pass_manager.h (pass_manager::finish_optimization_passes): New.
11744 (pass_manager): Use pass-instances.def to add fields for the
11745 various pass instances.
11746 * toplev.c (finalize): Update for move of
11747 finish_optimization_passes to a method of gcc::pass_manager.
11748 * toplev.h (finish_optimization_passes): Move to method of class
11750 * tree-pass.h (struct pass_data): New.
11751 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
11752 (opt_pass::gate): Convert to virtual function.
11753 (opt_pass::~opt_pass): New.
11754 (opt_pass::clone): New.
11755 (opt_pass::execute): Convert to virtual function.
11756 (opt_pass::opt_pass): New.
11757 (opt_pass::ctxt_): new.
11758 (gimple_opt_pass): Convert to subclass of opt_pass.
11759 (gimple_opt_pass::gimple_opt_pass): New.
11760 (rtl_opt_pass): Convert to subclass of opt_pass.
11761 (rtl_opt_pass::rtl_opt_pass): New.
11762 (ipa_opt_pass_d): Convert to subclass of opt_pass.
11763 (ipa_opt_pass_d::ipa_opt_pass_d): New.
11764 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
11765 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
11766 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
11767 invocation of pass_mode_switching to reflect this moving from a
11768 global to a member of gcc::pass_manager.
11769 (ix86_option_override): Rework how pass_insert_vzeroupper is
11770 added to the pass_manager to reflect autogenerated changes.
11771 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
11773 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
11775 PR rtl-optimization/57708
11776 * recog.c (peep2_find_free_register): Validate all regs in a
11779 2013-08-05 Jan Hubicka <jh@suse.cz>
11782 * cgraph.c (verify_cgraph_node): Accept local flags from other
11784 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
11785 (function_and_variable_visibility): Likewise.
11786 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
11788 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11790 * graph.c (init_graph_slim_pretty_print): Remove.
11791 (print_graph_cfg): Do not call it. Use local pretty printer.
11792 (start_graph_dump): Likewise.
11794 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11796 * gimple-pretty-print.c (buffer): Remove.
11797 (initialized): Likewise.
11798 (maybe_init_pretty_print): Likewise.
11799 (print_gimple_stmt): Do not call it. Use non-static local
11800 pretty_printer variable.
11801 (print_gimple_expr): Likewise.
11802 (print_gimple_seq): Likewise.
11803 (gimple_dump_bb): Likewise.
11805 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
11807 * asan.c (asan_pp): Remove.
11808 (asan_pp_initialized): Likewise.
11809 (asan_pp_initialize): Likewise.
11810 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
11811 (asan_emit_stack_protection): Tidy. Use local pretty printer.
11812 (asan_add_global): Likewise.
11814 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
11816 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
11817 * diagnostic.h (diagnostic_flush_buffer): Adjust.
11818 * pretty-print.c (pp_formatted_text_data): Likewise.
11819 (pp_indent): Rename from pp_base_indent.
11820 (pp_format): Rename from pp_base_format.
11821 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
11822 (pp_format_verbatim): Rename from pp_base_format_verbatim.
11823 (pp_flush): Rename from pp_base_flush.
11824 (pp_set_line_maximum_length): Rename from
11825 pp_base_set_line_maximum_length.
11826 (pp_clear_output_area): Rename from pp_base_clear_output_area.
11827 (pp_set_prefix): Rename from pp_base_set_prefix.
11828 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
11829 (pp_emit_prefix): Rename from pp_base_emit_prefix.
11830 (pp_append_text): Rename from pp_base_append_text.
11831 (pp_formatted_text): Rename from pp_base_formatted_text.
11832 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
11833 (pp_remaining_character_count_for_line): Rename from
11834 pp_base_remaining_character_count_for_line.
11835 (pp_newline): Rename from pp_base_newline.
11836 (pp_character): Rename from pp_base_character.
11837 (pp_string): Rename from pp_base_string.
11838 (pp_maybe_space): Rename from pp_base_maybe_space.
11839 * asan.c (asan_pp_string): Adjust.
11840 (asan_emit_stack_protection): Likewise.
11841 (asan_add_global): Likewise.
11842 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
11843 * tree-mudflap.c (mf_varname_tree): Likewise.
11844 * tree-pretty-print.c (pp_tree_identifier): Rename from
11845 pp_base_tree_identifier.
11846 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
11847 Declare as function.
11849 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11851 * pretty-print.h (pp_bar_bar): New.
11852 (pp_ampersand_ampersand): Likewise.
11853 (pp_less_equal): Likewise.
11854 (pp_greater_equal): Likewise.
11855 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
11856 printer functions instead of pp_string or operators and punctuators.
11857 (dump_gimple_call): Likewise.
11858 (dump_gimple_omp_for): Likewise.
11859 (dump_gimple_transaction): Likewise.
11860 (dump_gimple_phi): Likewise.
11861 (pp_gimple_stmt_1): Likewise.
11862 * sched-vis.c (print_insn): Likewise.
11863 * tree-mudflap.c (mf_varname_tree): Likewise.
11864 * tree-pretty-print.c (dump_block_node): Likewise.
11865 (dump_generic_node): Likewise.
11867 2013-08-02 Jan Hubicka <jh@suse.cz>
11869 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
11871 * lto-streamer-out.c (tree_is_indexable): Results decls and
11872 parm decls are not indexable.
11873 (DFS_write_tree_body): Do not follow args and results.
11874 (hash_tree): Likewise.
11875 (output_functions): Rearrange so struct function is needed
11876 only when real body is output; be able to also ouptut abstract
11877 functions; output DECL_ARGUMENTS and DECL_RESULT.
11878 (lto_output): When not in WPA, ale store abstract functions.
11879 (write_symbol): Do not care about RESULT_DECL.
11880 (output_symbol_p): Handle correctly sbtract decls.
11881 * lto-streamer-in.c (input_function): Rearrange so struct
11882 function can be NULL at entry; allow streaming of
11883 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
11884 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
11885 sanity check during LTO.
11886 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
11887 RESULT_DECl and DECL_ARGUMENTS.
11888 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
11891 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
11893 * pretty-print.h (pp_underscore): New.
11895 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
11896 printer functions instead of pp_character.
11897 (dump_binary_rhs): Likewise.
11898 (dump_ternary_rhs): Likewise.
11899 (dump_gimple_call_args): Likewise.
11900 (pp_points_to_solution): Likewise.
11901 (dump_gimple_call): Likewise.
11902 (dump_gimple_switch): Likewise.
11903 (dump_gimple_cond): Likewise.
11904 (dump_gimple_bind): Likewise.
11905 (dump_gimple_try): Likewise.
11906 (dump_gimple_omp_for): Likewise.
11907 (dump_gimple_omp_continue): Likewise.
11908 (dump_gimple_omp_single): Likewise.
11909 (dump_gimple_omp_sections): Likewise.
11910 (dump_gimple_omp_block): Likewise.
11911 (dump_gimple_omp_critical): Likewise.
11912 (dump_gimple_transaction): Likewise.
11913 (dump_gimple_asm): Likewise.
11914 (dump_gimple_phi): Likewise.
11915 (dump_gimple_omp_parallel): Likewise.
11916 (dump_gimple_omp_task): Likewise.
11917 (dump_gimple_omp_atomic_load): Likewise.
11918 (dump_gimple_omp_atomic_store): Likewise.
11919 (dump_gimple_mem_ops): Likewise.
11920 (pp_gimple_stmt_1): Likewise.
11921 (pp_cfg_jump): Likewise.
11922 (dump_implicit_edges): Likewise.
11923 (gimple_dump_bb_for_graph): Likewise.
11924 * graph.c (draw_cfg_node): Likewise.
11925 * langhooks.c (lhd_print_error_function): Likewise.
11926 * sched-vis.c (print_exp): Likewise.
11927 (print_value): Likewise.
11928 (print_pattern): Likewise.
11929 (print_insn): Likewise.
11930 (rtl_dump_bb_for_graph): Likewise.
11931 * tree-pretty-print.c (dump_function_declaration): Likewise.
11932 (dump_array_domain): Likewise.
11933 (dump_omp_clause): Likewise.
11934 (dump_location): Likewise.
11935 (dump_generic_node): Likewise.
11936 (print_struct_decl): Likewise.
11937 * diagnostic.c (diagnostic_show_locus): Use pp_space.
11939 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11941 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
11942 candidate table when replacing a candidate statement.
11943 (replace_rhs_if_not_dup): Likewise.
11944 (replace_one_candidate): Likewise.
11946 2013-08-02 Jan Hubicka <jh@suse.cz>
11947 Martin Liska <marxin.liska@gmail.com>
11949 * cgraphunit.c (add_new_function): Fix logic when adding from
11951 (assemble_thunk): Rename to ...
11952 (expand_thunk); .. this one; export; get it working with
11953 general functions; make produced gimple valid.
11954 * cgraph.h (expand_thunk): Declare.
11956 2013-08-02 Jan Hubicka <jh@suse.cz>
11958 * ipa-cp.c (gather_context_independent_values): Use
11959 ipa_get_param_move_cost.
11960 (get_replacement_map): Remove PARAM; move parameter folding
11962 (create_specialized_node): Update.
11963 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
11964 assert that we have gimple body; update move_cost.
11965 (count_formal_params): Assert that we have gimple body.
11966 (ipa_dump_param): New function.
11967 (ipa_alloc_node_params): Break out from ...
11968 (ipa_initialize_node_params): ... here.
11969 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
11970 (ipa_write_node_info): Stream move costs.
11971 (ipa_read_node_info): Read move costs.
11972 (ipa_update_after_lto_read): Do not recompute node params.
11973 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
11974 (ipa_get_param): Check we are not in WPA.
11975 (ipa_get_param_move_cost): New.
11976 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
11977 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
11978 parm numbers to be present.
11980 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
11982 PR rtl-optimization/58048
11983 * lra-constraints.c (process_alt_operands): Don't check asm
11984 operand on register.
11986 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
11988 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
11989 the implied StoreLoad barrier for atomic operations if before.
11991 2013-08-02 Jan Hubicka <jh@suse.cz>
11992 Martin Liska <marxin.liska@gmail.com>
11994 * cgraph.c (cgraph_function_body_availability): Do not check
11996 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
11997 symtab_node_availability): Declare.
11998 * ipa.c (can_replace_by_local_alias): New.
11999 (function_and_variable_visibility): Use it.
12000 * symtab.c (symtab_for_node_and_aliases,
12001 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
12003 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
12005 PR rtl-optimization/57963
12006 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
12007 (lra_constraints): Use them.
12009 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
12011 * config/arm/types.md (define_attr "type"): Add "load_acq"
12013 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
12015 (cortex_a53_store1): Likewise.
12017 2013-08-01 Jan Hubicka <jh@suse.cz>
12019 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
12020 partitions are not needed.
12022 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
12024 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
12025 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
12026 MAYBE_NON_Q_CLASS_P where appropriate.
12028 2013-08-01 Jan Hubicka <jh@suse.cz>
12030 * cgraph.h (release_function_body): Declare.
12031 * tree.c (free_lang_data_in_decl): Free, parameters and return values
12032 of unused delcarations.
12034 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12036 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
12037 RTL form when subtracting a constant.
12039 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12041 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
12042 Generate canonical plus rtx with negated immediate instead of minus
12044 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
12046 2013-08-01 Jan Hubicka <jh@suse.cz>
12048 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
12049 (cgraph_release_function_body): Likewise.
12050 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
12051 * cgraph.h (cgrpah_node): Rename abstract_and_needed
12052 to used_as_abstract_origin.
12053 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
12054 symbols used as abstract origins.
12055 * cgraphunit.c (analyze_functions): Update.
12056 * ipa.c (symtab_remove_unreachable_nodes): Recompute
12057 used_as_abstract_origin.
12058 * tree-inline.c (tree_function_versioning): Update
12059 used_as_abstract_origin; be ready for DECL_RESULT and
12060 DECL_ARGUMENTS to be NULL.
12062 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
12063 for abstract functions.
12064 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
12067 2013-08-01 Jan Hubicka <jh@suse.cz>
12069 * profile.c (compute_value_histograms): Fix thinko.
12071 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
12073 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
12074 aarch-common-protos.h to extra_headers.
12075 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
12076 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
12077 * config/aarch64/t-aarch64 (aarch-common.o): Define.
12079 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
12081 * config/aarch64/aarch64.md (define_attr "type"): Delete.
12082 Include "../arm/types.md". Define "type" attribute for all patterns.
12083 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
12086 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
12088 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
12089 to support power8 load fusion.
12090 (fusion_gpr_mem_load): Likewise.
12092 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
12094 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
12095 declarations for power8 load fusion.
12096 (emit_fusion_gpr_load): Likewise.
12098 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
12099 tuning for power8, turn on fusion mode by default. Turn on sign
12100 extending fusion mode if normal fusion mode is on, and we are at
12102 (fusion_gpr_load_p): New function, return true if we can fuse an
12103 addis instruction with a dependent load to a GPR.
12104 (emit_fusion_gpr_load): Emit the instructions for power8 load
12107 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
12108 (VSX load fusion peepholes): New peepholes to fuse together an
12109 addi instruction with a VSX load instruction.
12111 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
12112 peepholes to fuse an addis instruction with a load to a GPR base
12113 register. If we are supporting sign extending fusions, convert
12114 sign extending loads to zero extending loads and add an explicit
12117 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
12119 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
12120 aarch-common-protos.h to extra_headers.
12121 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
12122 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
12123 (arm_early_store_addr_dep): Likewise.
12124 (arm_no_early_alu_shift_dep): Likewise.
12125 (arm_no_early_alu_shift_value_dep): Likewise.
12126 (arm_no_early_mul_dep): Likewise.
12127 (arm_no_early_store_addr_dep): Likewise.
12128 (arm_mac_accumulator_is_mul_result): Likewise.
12129 (arm_mac_accumulator_is_result): Likewise.
12130 * config/arm/aarch-common.c: ... here. New file.
12131 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
12133 (arm_early_store_addr_dep): Likewise.
12134 (arm_no_early_alu_shift_dep): Likewise.
12135 (arm_no_early_alu_shift_value_dep): Likewise.
12136 (arm_no_early_mul_dep): Likewise.
12137 (arm_no_early_store_addr_dep): Likewise.
12138 (arm_mac_accumulator_is_mul_result): Likewise.
12139 (arm_mac_accumulator_is_result): Likewise.
12140 * config/arm/aarch-common-protos.h: ... here. New file.
12141 * config/arm/t-arm (aarch-common.o): Define.
12143 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
12145 * config/arm/arm.md: Include new file "types.md".
12146 (define_attr "type"): Move from here to ...
12147 (define_attr "mul32"): Likewise.
12148 (define_attr "mul64"): Likewise.
12149 * config/arm/types.md: ... here. New file.
12151 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
12153 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
12154 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
12156 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12158 * gen-pass-instances.awk: Fix offset of substr().
12160 2013-07-31 David Malcolm <dmalcolm@redhat.com>
12162 * Makefile.in (pass-instances.def): New.
12163 (passes.o): Replace dependency on passes.def with one on
12166 * gen-pass-instances.awk: New.
12168 * passes.c (pass_manager::pass_manager): Use pass-instances.def
12169 rather than passes.def, updating local definition of NEXT_PASS
12170 macro to add an extra NUM parameter (currently unused).
12172 2013-07-30 David Malcolm <dmalcolm@redhat.com>
12174 * Makefile.in (PASS_MANAGER_H): New.
12175 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
12176 (passes.o): Likewise.
12177 (statistics.o): Likewise.
12178 (cgraphunit.o): Likewise.
12179 (context.o): Depend on PASS_MANAGER_H.
12181 * pass_manager.h: New.
12183 * cgraphunit.c (cgraph_add_new_function): Update for moves
12184 of globals to fields of pass_manager.
12185 (analyze_function): Likewise.
12186 (expand_function): Likewise.
12187 (ipa_passes): Likewise.
12188 (compile): Likewise.
12190 * context.c (context::context): New.
12191 * context.h (context::context): New.
12192 (context::get_passes): New.
12193 (context::passes_): New.
12195 * lto-cgraph.c (input_node): Update for moves of globals to
12196 fields of pass_manager.
12198 * passes.c (all_passes): Remove, in favor of a field of the
12199 same name within the new class pass_manager.
12200 (all_small_ipa_passes): Likewise.
12201 (all_lowering_passes): Likewise.
12202 (all_regular_ipa_passes): Likewise.
12203 (all_late_ipa_passes): Likewise.
12204 (all_lto_gen_passes): Likewise.
12205 (passes_by_id): Likewise.
12206 (passes_by_id_size): Likewise.
12207 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12208 the new class pass_manager.
12209 (set_pass_for_id): Convert to...
12210 (pass_manager::set_pass_for_id): ...method.
12211 (get_pass_for_id): Convert to...
12212 (pass_manager::get_pass_for_id): ...method.
12213 (register_one_dump_file): Move body of implementation into...
12214 (pass_manager::register_one_dump_file): ...here.
12215 (register_dump_files_1): Convert to...
12216 (pass_manager::register_dump_files_1): ...method.
12217 (register_dump_files): Convert to...
12218 (pass_manager::register_dump_files): ...method.
12219 (create_pass_tab): Update for moves of globals to fields of
12221 (dump_passes): Move body of implementation into...
12222 (pass_manager::dump_passes): ...here.
12223 (register_pass): Move body of implementation into...
12224 (pass_manager::register_pass): ...here.
12225 (init_optimization_passes): Convert into...
12226 (pass_manager::pass_manager): ...constructor for new
12227 pass_manager class, and initialize the pass_lists array.
12228 (check_profile_consistency): Update for moves of globals to
12229 fields of pass_manager.
12230 (dump_profile_report): Move body of implementation into...
12231 (pass_manager::dump_profile_report): ...here.
12232 (ipa_write_summaries_1): Update for moves of pass lists from
12233 being globals to fields of pass_manager.
12234 (ipa_write_optimization_summaries): Likewise.
12235 (ipa_read_summaries): Likewise.
12236 (ipa_read_optimization_summaries): Likewise.
12237 (execute_all_ipa_stmt_fixups): Likewise.
12239 * statistics.c (statistics_fini): Update for moves of globals to
12240 fields of pass_manager.
12242 * toplev.c (general_init): Replace call to
12243 init_optimization_passes with construction of the pass_manager
12246 * tree-pass.h (all_passes): Remove, in favor of a field of the
12247 same name within the new class pass_manager.
12248 (all_small_ipa_passes): Likewise.
12249 (all_lowering_passes): Likewise.
12250 (all_regular_ipa_passes): Likewise.
12251 (all_lto_gen_passes): Likewise.
12252 (all_late_ipa_passes): Likewise.
12253 (passes_by_id): Likewise.
12254 (passes_by_id_size): Likewise.
12255 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
12256 the new class pass_manager.
12257 (get_pass_for_id): Remove.
12259 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12261 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
12264 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
12266 * arm.md (mulhi3): New expand pattern.
12268 2013-07-30 Jan Hubicka <jh@suse.cz>
12269 Martin Liska <marxin.liska@gmail.com>
12271 * profile.c (compute_value_histograms): Do not ICE when
12272 there is mismatch only on some counters.
12274 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
12276 PR rtl-optimization/57637
12277 * function.c (move_insn_for_shrink_wrap): Also check the
12278 GEN set of the LIVE problem for the liveness analysis
12279 if it exists, otherwise give up.
12281 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12283 PR tree-optimization/57993
12284 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
12285 replaced statement in the candidate table.
12286 (phi_add_costs): Return infinite cost when the hidden basis does
12287 not dominate all phis on which the candidate is dependent.
12288 (replace_one_candidate): Record replaced statement in the
12291 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
12293 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
12294 (ashlv2si3): New expander.
12295 (*ashlv2si3_i): New define_insn_and_split.
12296 * predicates.md (float_operation): Allow patterns with three
12297 basic sub-patterns.
12299 PR rtl-optimization/58021
12300 * mode-switching.c (create_pre_exit): Always split off preceding
12301 insns if we are not at the basic block head.
12303 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
12305 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
12306 (UCLIBC_DYNAMIC_LINKER): New macro.
12307 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
12309 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
12310 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
12312 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
12313 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
12314 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
12315 for SF and DF modes. Use ieee_quad_format for TF mode.
12316 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
12317 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
12318 (mips_option_override): Handle `-mnan=legacy'.
12319 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
12320 `-mabs=2008' and `-mnan=2008'.
12321 (OPTION_DEFAULT_SPECS): Add "nan" default.
12322 (ASM_SPEC): Handle `-mnan='.
12323 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
12324 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
12325 comment accordingly.
12326 (neg<mode>2): Likewise.
12327 * config/mips/mips.opt (mabs, mnan): New options.
12328 * doc/install.texi (Configuration): Document `--with-nan=' option.
12329 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
12331 (MIPS Options): Document them.
12332 * config.gcc <mips*-*-*>: Handle `--with-nan='.
12333 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
12334 * configure: Regenerate.
12335 * config.in: Regenerate.
12337 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12339 * config/i386/i386.md (float post-reload splitters): Do not check
12340 for subregs of SSE registers.
12342 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
12343 H.J. Lu <hongjiu.lu@intel.com>
12347 * config/i386/i386.md (post-reload splitter
12348 to avoid partial SSE reg dependency stalls): New pattern.
12350 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
12352 * config/s390/s390.md ("movcc"): Swap load and store instructions.
12354 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
12356 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
12357 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
12359 2013-07-26 Cary Coutant <ccoutant@google.com>
12361 * dwarf2out.c (die_checksum_ordered): Don't include template
12362 instantiations in signature.
12363 (is_template_parameter): New function.
12364 (is_template_instantiation): New function.
12365 (generate_skeleton_bottom_up): Don't include template instantiations
12367 (generate_skeleton): Likewise.
12368 (break_out_comdat_types): Move recursive call to break out nested
12370 (prune_unused_types_mark_generic_parms_dies): Call
12371 is_template_parameter.
12373 2013-07-26 Ian Bolton <ian.bolton@arm.com>
12375 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
12376 uses vector registers.
12377 * config/aarch64/iterators.md: Add attributes rtn and vas.
12379 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12380 Richard Earnshaw <richard.earnshaw@arm.com>
12382 * combine.c (simplify_comparison): Re-canonicalize operands
12384 * config/arm/arm.md (movcond_addsi): New splitter.
12386 2013-07-25 Sterling Augustine <saugustine@google.com>
12388 * dwarf2out.c (size_of_pubnames): Move code to...
12389 (include_pubname_in_output): ...here. New.
12390 (want_pubnames): Rearrange.
12391 (output_pubnames): Call include_pubname_in_output. Move assertion.
12393 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12395 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
12397 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
12400 * doc/extend.texi: Remove obsolete builtins. Fix
12401 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
12403 2013-07-25 Jan Hubicka <jh@suse.cz>
12405 * cgraph.c (release_function_body): Break out from ...
12406 (cgraph_release_function_body): ... this one; also release DECL_RESULT
12407 and DECL_ARGUMENTS.
12408 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
12409 old_tree in the map.
12410 (create_specialized_node): Update.
12411 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
12413 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
12414 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
12415 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
12416 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
12419 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12421 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
12422 addsi3_carryin_alt2_<optab>): Correct output template.
12424 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12426 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
12427 Adjust for arm_restrict_it.
12428 Remove trailing whitespace.
12430 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
12432 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
12433 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
12435 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
12437 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
12439 PR rtl-optimization/57960
12440 * lra-constraints.c (process_alt_operands): Use the right mode
12441 when checking strict_low.
12443 2013-07-25 Jan Hubicka <jh@suse.cz>
12445 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
12446 * cgraph.c (cgraph_remove_node): Do not release function body
12447 when in cgraph streaming.
12448 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
12449 in other partitions are not considered reachable; fix handling of
12452 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12454 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
12456 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
12461 * config/arm/arm.md ("*sibcall_insn): Replace use of
12462 Ss with US. Adjust output for v5 and v4t.
12463 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
12465 * config/arm/constraints.md ("Ss"): Rename to US.
12467 2013-07-25 Terry Guo <terry.guo@arm.com>
12469 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
12470 shift_add/shift_sub0/shift_sub1 RTXs.
12472 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12473 Anton Blanchard <anton@au1.ibm.com>
12475 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
12476 (altivec_vpks<VI_char>ss): Likewise.
12477 (altivec_vpks<VI_char>us): Likewise.
12478 (altivec_vpku<VI_char>us): Likewise.
12479 (altivec_vpku<VI_char>um): Likewise.
12481 2013-07-24 David Malcolm <dmalcolm@redhat.com>
12483 Introduce context class.
12485 * Makefile.in (CONTEXT_H): New.
12486 (OBJS): Add context.o.
12487 (toplev.o): Add CONTEXT_H to dependencies.
12490 * toplev.c (general_init): Create the singleton gcc::context instance.
12496 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
12498 PR rtl-optimization/57968
12499 * mode-switching.c (create_pre_exit): Allow instructions that
12500 don't set a return register to need a non-exit mode.
12502 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12503 Anton Blanchard <anton@au1.ibm.com>
12505 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
12506 operands to vperm for little endian.
12507 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
12508 of lvsl to create the control mask for a vperm for little endian.
12510 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12511 Anton Blanchard <anton@au1.ibm.com>
12513 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
12514 two operands for little-endian.
12516 2013-07-23 Steve Ellcey <sellcey@mips.com>
12518 * config/mips/mips.c (mips_case_values_threshold): New.
12519 (TARGET_CASE_VALUES_THRESHOLD): Define.
12521 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12522 Anton Blanchard <anton@au1.ibm.com>
12524 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
12525 selection of field for vector splat in little endian mode.
12527 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
12529 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
12530 expanders to rs6000.md.
12531 (ior<mode>3): Likewise.
12532 (and<mode>3): Likewise.
12533 (one_cmpl<mode>2): Likewise.
12534 (nor<mode>3): Likewise.
12535 (andc<mode>3): Likewise.
12536 (eqv<mode>3): Likewise.
12537 (nand<mode>3): Likewise.
12538 (orc<mode>3): Likewise.
12540 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
12543 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
12544 to split multi-word logical operations.
12545 (rs6000_split_logical_di): Likewise.
12546 (rs6000_split_logical): Likewise.
12548 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
12549 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
12550 and allow TImode operations in 32-bit.
12551 (vsx_and<mode>3_64bit): Likewise.
12552 (vsx_ior<mode>3_32bit): Likewise.
12553 (vsx_ior<mode>3_64bit): Likewise.
12554 (vsx_xor<mode>3_32bit): Likewise.
12555 (vsx_xor<mode>3_64bit): Likewise.
12556 (vsx_one_cmpl<mode>2_32bit): Likewise.
12557 (vsx_one_cmpl<mode>2_64bit): Likewise.
12558 (vsx_nor<mode>3_32bit): Likewise.
12559 (vsx_nor<mode>3_64bit): Likewise.
12560 (vsx_andc<mode>3_32bit): Likewise.
12561 (vsx_andc<mode>3_64bit): Likewise.
12562 (vsx_eqv<mode>3_32bit): Likewise.
12563 (vsx_eqv<mode>3_64bit): Likewise.
12564 (vsx_nand<mode>3_32bit): Likewise.
12565 (vsx_nand<mode>3_64bit): Likewise.
12566 (vsx_orc<mode>3_32bit): Likewise.
12567 (vsx_orc<mode>3_64bit): Likewise.
12569 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
12570 logical types in GPRs.
12572 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
12573 logical insns to rs6000.md, and allow TImode operations in
12575 (altivec_ior<mode>3): Likewise.
12576 (altivec_xor<mode>3): Likewise.
12577 (altivec_one_cmpl<mode>2): Likewise.
12578 (altivec_nor<mode>3): Likewise.
12579 (altivec_andc<mode>3): Likewise.
12581 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
12582 attributes for moving the 128-bit logical operations into
12584 (BOOL_REGS_OUTPUT): Likewise.
12585 (BOOL_REGS_OP1): Likewise.
12586 (BOOL_REGS_OP2): Likewise.
12587 (BOOL_REGS_UNARY): Likewise.
12588 (BOOL_REGS_AND_CR0): Likewise.
12589 (one_cmpl<mode>2): Add support for DI logical operations on
12590 32-bit, splitting the operations to 32-bit.
12591 (anddi3): Likewise.
12592 (iordi3): Likewise.
12593 (xordi3): Likewise.
12594 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
12595 changes to combine the 32/64-bit code, allow logical operations on
12596 TI mode in 32-bit, and to use similar match_operator patterns like
12597 scalar mode uses. Combine the Altivec and VSX code for logical
12598 operations, and move it here.
12599 (ior<mode>3, 128-bit types): Likewise.
12600 (xor<mode>3, 128-bit types): Likewise.
12601 (one_cmpl<mode>3, 128-bit types): Likewise.
12602 (nor<mode>3, 128-bit types): Likewise.
12603 (andc<mode>3, 128-bit types): Likewise.
12604 (eqv<mode>3, 128-bit types): Likewise.
12605 (nand<mode>3, 128-bit types): Likewise.
12606 (orc<mode>3, 128-bit types): Likewise.
12607 (and<mode>3_internal): Likewise.
12608 (bool<mode>3_internal): Likewise.
12609 (boolc<mode>3_internal1): Likewise.
12610 (boolc<mode>3_internal2): Likewise.
12611 (boolcc<mode>3_internal1): Likewise.
12612 (boolcc<mode>3_internal2): Likewise.
12613 (eqv<mode>3_internal1): Likewise.
12614 (eqv<mode>3_internal2): Likewise.
12615 (one_cmpl1<mode>3_internal): Likewise.
12617 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12619 * config/microblaze/microblaze.c (microblaze_expand_prologue):
12620 Rename flag_stack_usage to flag_stack_usage_info.
12622 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
12624 * config/microblaze/sync.md: New file.
12625 * config/microblaze/microblaze.md: Include sync.md
12626 * config/microblaze/microblaze.c: Add print_operand 'y'.
12627 * config/microblaze/constraints.md: Add memory_contraint
12628 'Q' which is a single register.
12630 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
12632 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
12634 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12636 * reload.c (find_reloads): Exit loop once we find this operand
12637 cannot be reloaded somehow for this alternative.
12639 * reload.c (find_reloads): Exit loop once we find a hard register.
12641 * rtlanal.c (computed_jump_p): Exit loop once we find label
12644 * i386.c (ix86_pad_returns): Exit loop after setting replace.
12646 * cfgloopmanip.c (remove_path): Exit loop after setting
12649 * gensupport.c (subst_dup): Avoid loop if code is not
12650 MATCH_DUP nor MATCH_OP_DUP.
12652 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
12654 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
12656 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12658 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
12659 true for SP_REGNUM if mode == ptr_mode.
12660 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
12661 with value R0_REGNUM + 31.
12663 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12665 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
12666 pad pointer-typed argument downward.
12668 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12670 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
12671 and __ILP32__ when the ILP32 model is in use.
12673 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12675 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
12676 (aarch64_load_symref_appropriately): In the case of
12677 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
12678 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
12679 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
12680 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
12681 if 'mode' doesn't equal to 'ptr_mode'.
12682 (aarch64_output_mi_thunk): Add an assertion on the alignment of
12683 'vcall_offset'; change to call aarch64_emit_move differently depending
12684 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
12685 to calculate the upper bound of 'vcall_offset'.
12686 (aarch64_cannot_force_const_mem): Change to also return true if
12688 (aarch64_legitimize_reload_address): In the case of large
12689 displacements, add new local variable 'xmode' and an assertion
12690 based on it; change to use 'xmode' to generate the new rtx and
12692 (aarch64_asm_trampoline_template): Change to generate the template
12693 differently depending on TARGET_ILP32 or not; change to use
12694 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
12695 (aarch64_trampoline_size): Removed.
12696 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
12697 and replace immediate literals with it. Change to use 'ptr_mode'
12698 instead of 'DImode' and call convert_memory_address if the mode
12699 of 'fnaddr' doesn't equal to 'ptr_mode'.
12700 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
12702 (aarch64_elf_asm_destructor): Likewise.
12703 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
12704 on TARGET_ILP32 instead of aarch64_trampoline_size.
12705 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
12706 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
12707 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
12708 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
12709 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
12710 (storewb_pair<GPI:mode>_<P:mode>): ... this.
12711 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
12712 depending on the value of 'mode'.
12713 (add_losym_<mode>): New.
12714 (ldr_got_small_<mode>): New, based on ldr_got_small.
12715 (ldr_got_small): Remove.
12716 (ldr_got_small_sidi): New.
12717 * config/aarch64/iterators.md (P): New.
12718 (PTR): Change to 'ptr_mode' in the condition.
12720 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
12722 * config.gcc (aarch64*-*-*): Support --with-abi.
12723 (aarch64*-*-elf): Support --with-multilib-list.
12724 (aarch64*-*-linux*): Likewise.
12725 (supported_defaults): Add abi to aarch64*-*-*.
12726 * configure.ac: Mention AArch64 for --with-multilib-list.
12727 * configure: Re-generated.
12728 * config/aarch64/biarchilp32.h: New file.
12729 * config/aarch64/biarchlp64.h: New file.
12730 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
12732 (MULTILIB_DEFAULTS): Ditto.
12733 (DRIVER_SELF_SPECS): Ditto.
12734 (ASM_SPEC): Update to also substitute -mabi.
12735 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
12736 file whose name depends on -mabi= and -mbig-endian.
12737 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
12739 (POINTER_SIZE): New define.
12740 (POINTERS_EXTEND_UNSIGNED): Ditto.
12741 (enum aarch64_abi_type): New enumeration tag.
12742 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
12743 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
12744 (TARGET_ILP32): New define.
12745 * config/aarch64/aarch64.opt (mabi): New.
12746 (aarch64_abi): New.
12747 (ilp32, lp64): New values for -mabi.
12748 * config/aarch64/t-aarch64 (comma): New define.
12749 (MULTILIB_OPTIONS): Ditto.
12750 (MULTILIB_DIRNAMES): Ditto.
12751 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
12752 * doc/invoke.texi: Document -mabi for AArch64.
12754 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
12756 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
12758 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12759 Anton Blanchard <anton@au1.ibm.com>
12761 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
12762 endianness when selecting field to splat.
12764 2013-07-22 Eric Christopher <echristo@gmail.com>
12766 * dwarf2out.c (die_odr_checksum): New function to use
12767 CHECKSUM_ macros and ULEB128 for DIE tag.
12768 (generate_type_signature): Use.
12770 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
12772 * config.gcc (sparc*-*-*): Accept leon3 processor.
12773 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
12774 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
12775 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
12776 * config/sparc/sparc.opt (enum processor_type): Add leon3.
12777 (mfix-ut699): Adjust comment.
12778 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
12779 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
12780 (CPP_CPU_SPEC): Likewise.
12781 (ASM_CPU_SPEC): Likewise.
12782 * config/sparc/sparc.c (leon3_cost): New constant.
12783 (sparc_option_override): Add leon3 support.
12784 (mem_ref): New function.
12785 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
12786 (sparc_do_work_around_errata): Look into the instruction in the delay
12787 slot and adjust accordingly. Add fix for the data cache nullify issues
12788 of the UT699. Change insertion position for the NOP.
12789 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
12790 (leon3_load): New reservation.
12791 (leon_store): Bump latency to 2.
12792 (grfpu): New automaton.
12793 (grfpu_alu): New unit.
12794 (grfpu_ds): Likewise.
12795 (leon_fp_alu): Adjust.
12796 (leon_fp_mult): Delete.
12797 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
12798 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
12799 * config/sparc/sparc.md (cpu): Add leon3.
12800 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
12801 (swapsi): Likewise.
12802 (atomic_test_and_set): Likewise.
12803 (ldstub): Likewise.
12805 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
12807 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
12808 default for R5900 targets.
12809 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
12810 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
12811 * config/mips/mips.c (mips_option_override): Report an error for
12812 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
12813 for -march=r5900 -mhard-float.
12815 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
12817 * df-problems.c (can_move_insns_across): Exit loop once we
12818 find a non-fixed, non-global register.
12820 * ipa-pure-const.c (propagate_nothrow): Exit loop after
12823 * omega.c (omega_eliminate_red): Break after setting red_found.
12824 (omega_problem_has_red_equations): Similarly after setting found.
12825 (omega_query_variable): Similarly after setting coupled.
12827 2013-07-22 Marek Polacek <polacek@redhat.com>
12829 * gimplify.c: Don't include gimple.h twice.
12831 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12833 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
12834 instead of TARGET_THUMB1.
12835 (Pz): New constraint.
12836 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
12838 (compare_negsi_si): Likewise.
12839 (compare_addsi2_op0): Likewise.
12840 (compare_addsi2_op1): Likewise.
12841 (addsi3_carryin_<optab>): Likewise.
12842 (addsi3_carryin_alt2_<optab>): Likewise.
12843 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
12844 for arm_restrict_it.
12845 (subsi3_carryin): Likewise.
12846 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
12847 (minmax_arithsi): Disable for arm_restrict_it.
12848 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
12849 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
12850 (satsi_<SAT:code>_shift): Likewise.
12851 (arm_shiftsi3): Add alternative for 16-bit encoding.
12852 (arm32_movhf): Disable for arm_restrict_it.
12853 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
12854 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
12856 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
12858 * config/arm/arm.md (attribute "insn"): Delete.
12859 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
12860 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
12861 (not_shiftsi): Update for attribute change.
12862 (not_shiftsi_compare0): Likewise.
12863 (not_shiftsi_compare0_scratch): Likewise.
12864 (arm_one_cmplsi2): Likewise.
12865 (thumb1_one_cmplsi2): Likewise.
12866 (notsi_compare0): Likewise.
12867 (notsi_compare0_scratch): Likewise.
12868 (thumb1_movdi_insn): Likewise.
12869 (arm_movsi_insn): Likewise.
12870 (movhi_insn_arch4): Likewise.
12871 (movhi_bytes): Likewise.
12872 (arm_movqi_insn): Likewise.
12873 (thumb1_movqi_insn): Likewise.
12874 (arm32_movhf): Likewise.
12875 (thumb1_movhf): Likewise.
12876 (arm_movsf_soft_insn): Likewise.
12877 (thumb1_movsf_insn): Likewise.
12878 (thumb_movdf_insn): Likewise.
12879 (movsicc_insn): Likewise.
12880 (movsfcc_soft_insn): Likewise.
12881 (and_scc): Likewise.
12882 (cond_move): Likewise.
12883 (if_move_not): Likewise.
12884 (if_not_move): Likewise.
12885 (if_shift_move): Likewise.
12886 (if_move_shift): Likewise.
12887 (if_shift_shift): Likewise.
12888 (if_not_arith): Likewise.
12889 (if_arith_not): Likewise.
12890 (cond_move_not): Likewise.
12891 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
12892 (neon_mov<mode>): Likewise.
12893 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
12894 (thumb2_movsi_vfp): Likewise.
12895 (movsf_vfp): Likewise.
12896 (thumb2_movsf_vfp): Likewise.
12897 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
12899 (cortexa7_older_only): Likewise.
12900 (cortexa7_younger): Likewise.
12901 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
12902 (1020alu_shift_op): Likewise.
12903 (1020alu_shift_reg_op): Likewise.
12904 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
12905 (alu_shift_op): Likewise.
12906 (alu_shift_reg_op): Likewise.
12907 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
12908 (11_alu_shift_op): Likewise.
12909 (11_alu_shift_reg_op): Likewise.
12910 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
12911 (9_alu_shift_reg_op): Likewise.
12912 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
12914 (cortex_a15_alu_shift): Likewise.
12915 (cortex_a15_alu_shift_reg): Likewise.
12916 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
12917 (cortex_a5_alu_shift): Likewise.
12918 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
12920 (cortex_a53_alu_shift): Likewise.
12921 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
12923 (cortex_a7_alu_reg): Likewise.
12924 (cortex_a7_alu_shift): Likewise.
12925 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
12926 (cortex_a8_alu_shift): Likewise.
12927 (cortex_a8_alu_shift_reg): Likewise.
12928 (cortex_a8_mov): Likewise.
12929 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
12930 (cortex_a9_dp_shift): Likewise.
12931 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
12932 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
12933 (cortex_r4_mov): Likewise.
12934 (cortex_r4_alu_shift): Likewise.
12935 (cortex_r4_alu_shift_reg): Likewise.
12936 * config/arm/fa526.md (526_alu_op): Update for attribute change.
12937 (526_alu_shift_op): Likewise.
12938 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
12939 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
12940 (626te_alu_shift_op): Likewise.
12941 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
12942 (726te_alu_op): Likewise.
12943 (726te_alu_shift_op): Likewise.
12944 (726te_alu_shift_reg_op): Likewise.
12945 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
12946 (mp626_alu_shift_op): Likewise.
12947 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
12948 (pj4_alu_e1_conds): Likewise.
12949 (pj4_alu): Likewise.
12950 (pj4_alu_conds): Likewise.
12951 (pj4_shift): Likewise.
12952 (pj4_shift_conds): Likewise.
12953 (pj4_alu_shift): Likewise.
12954 (pj4_alu_shift_conds): Likewise.
12956 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12958 * config/arm/predicates.md (shiftable_operator_strict_it):
12960 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
12961 Disable cond_exec version for arm_restrict_it.
12962 (thumb2_smaxsi3): Convert to generate cond_exec.
12963 (thumb2_sminsi3): Likewise.
12964 (thumb32_umaxsi3): Likewise.
12965 (thumb2_uminsi3): Likewise.
12966 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
12967 (thumb2_neg_abssi2): Likewise.
12968 (thumb2_mov_scc): Add alternative for 16-bit encoding.
12969 (thumb2_movsicc_insn): Adjust alternatives.
12970 (thumb2_mov_negscc): Disable for arm_restrict_it.
12971 (thumb2_mov_negscc_strict_it): New pattern.
12972 (thumb2_mov_notscc_strict_it): New pattern.
12973 (thumb2_mov_notscc): Disable for arm_restrict_it.
12974 (thumb2_ior_scc): Likewise.
12975 (thumb2_ior_scc_strict_it): New pattern.
12976 (thumb2_cond_move): Adjust for arm_restrict_it.
12977 (thumb2_cond_arith): Disable for arm_restrict_it.
12978 (thumb2_cond_arith_strict_it): New pattern.
12979 (thumb2_cond_sub): Adjust for arm_restrict_it.
12980 (thumb2_movcond): Likewise.
12981 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
12982 (thumb2_zero_extendhisi2_v6): Likewise.
12983 (thumb2_zero_extendqisi2_v6): Likewise.
12984 (orsi_notsi_si): Likewise.
12985 (orsi_not_shiftsi_si): Likewise.
12987 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
12989 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
12990 instruction sequence is 1 byte shorter.
12992 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
12994 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
12995 it is not needed after split.
12997 2013-07-20 Iain Sandoe <iain@codesourcery.com>
13000 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
13001 second label for nonlocal goto receivers. Don't output pic base labels
13002 unless we're producing PIC; mark that action unreachable().
13003 (ix86_save_reg): If the function contains a nonlocal label, save the
13005 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
13006 * config/darwin.c (emitted_pic_label_num): New GTY.
13007 (update_pic_label_number_if_needed): New.
13008 (machopic_output_function_base_name): Adjust for nonlocal receiver
13010 (machopic_should_output_picbase_label): New.
13011 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
13012 (nonlocal_goto_receiver): New insn and split.
13014 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
13016 * config/aarch64/aarch64-builtins.c
13017 (aarch64_fold_builtin): Fold abs in all modes.
13018 * config/aarch64/aarch64-simd-builtins.def
13019 (abs): Enable for all modes.
13020 * config/aarch64/arm_neon.h
13021 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
13022 (vabs_f64): Add missing intrinsic.
13024 2013-07-19 Ian Bolton <ian.bolton@arm.com>
13026 * config/aarch64/arm_neon.h (vabs_s64): New function
13028 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
13031 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
13032 * config/avr/avr.md (adjust_len): Add `round'.
13033 * config/avr/avr-protos.h (avr_out_round): New prototype.
13034 (avr_out_plus): Add `out_label' argument.
13035 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
13036 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
13037 Handle the case where `insn' is just a pattern.
13038 (avr_out_bitop): Handle the case where `insn' is just a pattern.
13039 (avr_out_round): New function.
13040 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
13042 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
13044 * config/microblaze/microblaze.c (microblaze_expand_prologue):
13045 Add check for flag_stack_usage to handle -fstack-usage support
13047 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
13049 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
13050 interaction for new Power8 flags and VSX.
13052 2013-07-18 Sriraman Tallam <tmsriram@google.com>
13054 PR middle-end/57698
13055 * tree-inline.c (expand_call_inline): Emit errors during
13056 early_inlining only if optimization is not turned on.
13058 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13062 * passes.c (init_optimization_passes): Move the construction of
13063 the pass hierarchy into a new passes.def file.
13065 * Makefile.in (passes.o): Add dependency on passes.def.
13067 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13069 * passes.c (init_optimization_passes): Introduce macros for
13070 constructing the tree of passes (INSERT_PASSES_AFTER,
13071 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
13072 TERMINATE_PASS_LIST).
13074 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
13075 Wei Mi <wmi@google.com>
13077 PR rtl-optimization/57878
13078 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
13080 (reload_pseudo_compare_func): Check nregs first for reload
13083 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13085 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
13087 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
13089 * read-rtl.c (validate_const_int): Once an invalid character is
13090 seen, quit the loop.
13092 * gengtype.c (write_roots): Similarly once we find the "deletable"
13093 or "if_marked" option.
13095 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
13097 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
13098 "xtab" and "sat". Move value "clz" from here to ...
13099 (attriubte "type"): ... here.
13100 (satsi_<SAT:code>): Delete "insn" attribute.
13101 (satsi_<SAT:code>_shift): Likewise.
13102 (arm_zero_extendqisi2addsi): Likewise.
13103 (arm_extendqisi2addsi): Likewise.
13104 (clzsi2): Update for attribute changes.
13105 (rbitsi2): Likewise.
13106 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
13108 (arm_usatsihi): Likewise.
13109 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
13111 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
13113 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
13114 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
13115 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
13116 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
13117 in alphabetical order.
13118 (attribute "core_cycles"): Update for attribute changes.
13119 (arm_addsi3): Likewise.
13120 (addsi3_compare0): Likewise.
13121 (addsi3_compare0_scratch): Likewise.
13122 (addsi3_compare_op1): Likewise.
13123 (addsi3_compare_op2): Likewise.
13124 (compare_addsi2_op0): Likewise.
13125 (compare_addsi2_op1): Likewise.
13126 (addsi3_carryin_shift_<optab>): Likewise.
13127 (subsi3_carryin_shift): Likewise.
13128 (rsbsi3_carryin_shift): Likewise.
13129 (arm_subsi3_insn): Likewise.
13130 (subsi3_compare0): Likewise.
13131 (subsi3_compare): Likewise.
13132 (arm_andsi3_insn): Likewise.
13133 (thumb1_andsi3_insn): Likewise.
13134 (andsi3_compare0): Likewise.
13135 (andsi3_compare0_scratch): Likewise.
13136 (zeroextractsi_compare0_scratch
13137 (andsi_not_shiftsi_si): Likewise.
13138 (iorsi3_insn): Likewise.
13139 (iorsi3_compare0): Likewise.
13140 (iorsi3_compare0_scratch): Likewise.
13141 (arm_xorsi3): Likewise.
13142 (thumb1_xorsi3_insn): Likewise.
13143 (xorsi3_compare0): Likewise.
13144 (xorsi3_compare0_scratch): Likewise.
13145 (satsi_<SAT:code>_shift): Likewise.
13147 (arm_shiftsi3): Likewise.
13148 (shiftsi3_compare0): Likewise.
13149 (not_shiftsi): Likewise.
13150 (not_shiftsi_compare0): Likewise.
13151 (not_shiftsi_compare0_scratch): Likewise.
13152 (arm_one_cmplsi2): Likewise.
13153 (thumb_one_complsi2): Likewise.
13154 (notsi_compare0): Likewise.
13155 (notsi_compare0_scratch): Likewise.
13156 (thumb1_zero_extendhisi2): Likewise.
13157 (arm_zero_extendhisi2): Likewise.
13158 (arm_zero_extendhisi2_v6): Likewise.
13159 (arm_zero_extendhisi2addsi): Likewise.
13160 (thumb1_zero_extendqisi2): Likewise.
13161 (thumb1_zero_extendqisi2_v6): Likewise.
13162 (arm_zero_extendqisi2): Likewise.
13163 (arm_zero_extendqisi2_v6): Likewise.
13164 (arm_zero_extendqisi2addsi): Likewise.
13165 (thumb1_extendhisi2): Likewise.
13166 (arm_extendhisi2): Likewise.
13167 (arm_extendhisi2_v6): Likewise.
13168 (arm_extendqisi): Likewise.
13169 (arm_extendqisi_v6): Likewise.
13170 (arm_extendqisi2addsi): Likewise.
13171 (thumb1_extendqisi2): Likewise.
13172 (thumb1_movdi_insn): Likewise.
13173 (arm_movsi_insn): Likewise.
13174 (movsi_compare0): Likewise.
13175 (movhi_insn_arch4): Likewise.
13176 (movhi_bytes): Likewise.
13177 (arm_movqi_insn): Likewise.
13178 (thumb1_movqi_insn): Likewise.
13179 (arm32_movhf): Likewise.
13180 (thumb1_movhf): Likewise.
13181 (arm_movsf_soft_insn): Likewise.
13182 (thumb1_movsf_insn): Likewise.
13183 (movdf_soft_insn): Likewise.
13184 (thumb_movdf_insn): Likewise.
13185 (arm_cmpsi_insn): Likewise.
13186 (cmpsi_shiftsi): Likewise.
13187 (cmpsi_shiftsi_swp): Likewise.
13188 (arm_cmpsi_negshiftsi_si): Likewise.
13189 (movsicc_insn): Likewise.
13190 (movsfcc_soft_insn): Likewise.
13191 (arith_shiftsi): Likewise.
13192 (arith_shiftsi_compare0
13193 (arith_shiftsi_compare0_scratch
13194 (sub_shiftsi): Likewise.
13195 (sub_shiftsi_compare0
13196 (sub_shiftsi_compare0_scratch
13197 (and_scc): Likewise.
13198 (cond_move): Likewise.
13199 (if_plus_move): Likewise.
13200 (if_move_plus): Likewise.
13201 (if_move_not): Likewise.
13202 (if_not_move): Likewise.
13203 (if_shift_move): Likewise.
13204 (if_move_shift): Likewise.
13205 (if_shift_shift): Likewise.
13206 (if_not_arith): Likewise.
13207 (if_arith_not): Likewise.
13208 (cond_move_not): Likewise.
13209 (thumb1_ashlsi3): Set type attribute.
13210 (thumb1_ashrsi3): Likewise.
13211 (thumb1_lshrsi3): Likewise.
13212 (thumb1_rotrsi3): Likewise.
13213 (shiftsi3_compare0_scratch): Likewise.
13214 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
13215 (neon_mov<mode>): Likewise.
13216 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
13218 (thumb2_movsi_insn): Likewise.
13219 (thumb2_cmpsi_neg_shiftsi): Likewise.
13220 (thumb2_extendqisi_v6): Likewise.
13221 (thumb2_zero_extendhisi2_v6): Likewise.
13222 (thumb2_zero_extendqisi2_v6): Likewise.
13223 (thumb2_shiftsi3_short): Likewise.
13224 (thumb2_addsi3_compare0_scratch): Likewise.
13225 (orsi_not_shiftsi_si): Likewise.
13226 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
13227 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
13229 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
13230 (1020alu_shift_op): Likewise.
13231 (1020alu_shift_reg_op): Likewise.
13232 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
13233 (alu_shift_op): Likewise.
13234 (alu_shift_reg_op): Likewise.
13235 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
13236 (11_alu_shift_op): Likewise.
13237 (11_alu_shift_reg_op): Likewise.
13238 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
13239 (9_alu_shift_reg_op): Likewise.
13240 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
13242 (cortex_a15_alu_shift): Likewise.
13243 (cortex_a15_alu_shift_reg): Likewise.
13244 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
13246 (cortex_a5_alu_shift): Likewise.
13247 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
13249 (cortex_a53_alu_shift): Likewise.
13250 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
13252 (cortex_a7_alu_reg): Likewise.
13253 (cortex_a7_alu_shift): Likewise.
13254 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
13256 (cortex_a8_alu_shift): Likewise.
13257 (cortex_a8_alu_shift_reg): Likewise.
13258 (cortex_a8_mov): Likewise.
13259 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
13260 (cortex_a9_dp_shift): Likewise.
13261 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
13263 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
13265 (cortex_r4_mov): Likewise.
13266 (cortex_r4_alu_shift): Likewise.
13267 (cortex_r4_alu_shift_reg): Likewise.
13268 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
13269 (526_alu_shift_op): Likewise.
13270 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
13271 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
13272 (626te_alu_shift_op): Likewise.
13273 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
13274 (726te_alu_op): Likewise.
13275 (726te_alu_shift_op): Likewise.
13276 (726te_alu_shift_reg_op): Likewise.
13277 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
13278 (mp626_alu_shift_op): Likewise.
13279 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
13280 (pj4_alu_e1_conds): Likewise.
13281 (pj4_alu): Likewise.
13282 (pj4_alu_conds): Likewise.
13283 (pj4_shift): Likewise.
13284 (pj4_shift_conds): Likewise.
13285 (pj4_alu_shift): Likewise.
13286 (pj4_alu_shift_conds): Likewise.
13287 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
13289 (cortexa7_older_only): Likewise.
13290 (cortexa7_younger): Likewise.
13292 2013-07-18 David Malcolm <dmalcolm@redhat.com>
13294 * ipa-pure-const.c (generate_summary): Rename to...
13295 (pure_const_generate_summary): ... this.
13297 2013-07-17 Iain Sandoe <iain@codesourcery.com>
13299 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
13301 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
13304 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
13307 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13309 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
13310 enabled without -march=zEC12.
13311 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
13314 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13316 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
13317 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
13318 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
13319 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
13320 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
13321 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
13322 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
13324 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
13325 and ISA_HAS_NMADD3_NMSUB3.
13326 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
13327 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
13328 (nmsub4<mode>, nmsub3<mode>): Likewise.
13329 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
13331 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13333 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
13334 TARGET_MIPS5400 checking.
13336 2013-07-16 Jakub Jelinek <jakub@redhat.com>
13337 Peter Bergner <bergner@vnet.ibm.com>
13339 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
13340 registers in the comment.
13341 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
13342 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
13343 rather than FIRST_PSEUDO_REGISTERS.
13345 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
13347 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
13348 enable extra ISA flags with TARGET_HTM.
13350 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
13352 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
13355 2013-07-15 Cong Hou <congh@google.com>
13357 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
13358 in compare function for sorting.
13360 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
13362 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
13363 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
13364 * config/rs6000/rs6000.opt: Add -mhtm option.
13365 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
13366 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
13367 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
13368 __HTM__ if the HTM instructions are available.
13369 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
13370 htm_spr_reg_operand): New define_predicates.
13371 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
13372 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
13374 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
13375 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
13376 HTM builtin functions.
13377 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
13378 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
13379 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
13380 (rs6000_builtin_mask_calculate): Likewise.
13381 (rs6000_option_override_internal): Likewise.
13382 (bdesc_htm): Add new HTM builtin support.
13383 (htm_spr_num): New function.
13384 (htm_spr_regno): Likewise.
13385 (rs6000_htm_spr_icode): Likewise.
13386 (htm_expand_builtin): Likewise.
13387 (htm_init_builtins): Likewise.
13388 (rs6000_expand_builtin): Add support for HTM builtin functions.
13389 (rs6000_init_builtins): Likewise.
13390 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
13392 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
13393 (TARGET_HTM, MASK_HTM): Define macros.
13394 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
13395 (FIXED_REGISTERS): Likewise.
13396 (CALL_USED_REGISTERS): Likewise.
13397 (CALL_REALLY_USED_REGISTERS): Likewise.
13398 (REG_ALLOC_ORDER): Likewise.
13399 (enum reg_class): Likewise.
13400 (REG_CLASS_NAMES): Likewise.
13401 (REG_CLASS_CONTENTS): Likewise.
13402 (REGISTER_NAMES): Likewise.
13403 (ADDITIONAL_REGISTER_NAMES): Likewise.
13404 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
13405 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
13406 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
13407 * config/rs6000/htm.md: New file.
13408 * config/rs6000/htmintrin.h: New file.
13409 * config/rs6000/htmxlintrin.h: New file.
13411 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
13413 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
13414 Define SYMBOL_TINY_GOT, update comment.
13415 * config/aarch64/aarch64.c
13416 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
13417 (aarch64_expand_mov_immediate): Likewise.
13418 (aarch64_print_operand): Likewise.
13419 (aarch64_classify_symbol): Likewise.
13420 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
13421 (ldr_got_tiny): Define.
13423 2013-07-13 Tobias Grosser <tobias@grosser.es>
13425 PR tree-optimization/54094
13426 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
13427 scheduling dimension for the parallelism check from the polyhedral
13428 information in the AST.
13429 * graphite-dependences.c (carries_deps): Do not assume the schedule is
13432 2013-07-13 Jason Merrill <jason@redhat.com>
13434 * print-tree.c (debug_vec_tree): Use debug_raw.
13435 (debug_raw (vec<tree, va_gc> &)): New.
13436 (debug_raw (vec<tree, va_gc> *)): New.
13437 * tree.h: Declare them.
13439 2013-07-13 Bin Cheng <bin.cheng@arm.com>
13441 * ifcvt.c (ifcvt_after_combine): New static variable.
13442 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
13444 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
13445 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
13446 rest_of_handle_if_after_reload): Pass new argument for if_convert.
13448 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
13450 * config/mips/mips.c (mips_expand_call): Remove empty statement.
13452 2013-07-12 Michael Matz <matz@suse.de>
13454 PR middle-end/55771
13455 * convert.c (convert_to_real): Reject non-float inner types.
13457 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
13459 * config/aarch64/aarch64-protos.h
13460 (aarch64_simd_immediate_valid_for_move): Remove.
13461 * config/aarch64/aarch64.c (simd_immediate_info): New member.
13462 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
13464 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
13466 2013-07-11 Steve Ellcey <sellcey@mips.com>
13468 * config/mips/mips.c (mips_conditional_register_usage): Do not
13469 use t[0-7] registers in MIPS16 mode when optimizing for size.
13471 2013-07-11 Sriraman Tallam <tmsriram@google.com>
13473 * config/i386/i386.c (dispatch_function_versions): Fix array
13474 indexing of function_version_info to match actual_versions.
13476 2013-07-11 Teresa Johnson <tejohnson@google.com>
13478 * vec.h (struct va_gc): Move release out-of-line.
13479 (va_gc::release): Call ggc_free on released vec.
13481 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
13483 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
13484 Require GOT register as additional operand in UNSPEC.
13485 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
13486 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13487 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
13488 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
13489 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
13490 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
13491 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
13493 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
13496 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
13497 name seen by assembler/linker rather if available.
13499 2013-07-11 Andreas Schwab <schwab@suse.de>
13501 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
13503 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
13505 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
13507 2013-07-10 Joseph Myers <joseph@codesourcery.com>
13509 * doc/tm.texi.in: Move hook documentation to ....
13510 * target.def: ... here.
13512 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
13513 text on @hook line.
13514 * doc/tm.texi: Regenerate.
13516 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
13519 * doc/invoke.texi: Document -Wconditionally-supported.
13521 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13524 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
13527 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13530 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
13531 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
13532 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
13533 Remove duplicate devices.
13534 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
13535 * config/avr/t-multilib: Regenerate.
13536 * config/avr/avr-tables.opt: Regenerate.
13537 * doc/avr-mmcu.texi: Regenerate.
13539 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
13542 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
13544 2013-07-10 Graham Stott <graham.stott@btinternet.com>
13546 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
13547 the cost of MULT when optimizing for size.
13549 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
13551 * config/cr16/cr16-protos.h: Don't include target.h.
13553 2013-07-09 Joseph Myers <joseph@codesourcery.com>
13555 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
13556 adjust register size for TDmode and TFmode for VSX registers.
13558 2013-07-08 Kai Tietz <ktietz@redhat.com>
13561 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
13562 hook_bool_const_tree_true.
13564 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13566 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
13567 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
13568 * config/s390/s390.md: Define FPR*_REGNUM constants.
13569 Fix FPR2_REGNUM constant (18 -> 17).
13570 ("*trunc<BFP:mode><DFP_ALL:mode>2")
13571 ("*trunc<DFP_ALL:mode><BFP:mode>2")
13572 ("trunc<BFP:mode><DFP_ALL:mode>2")
13573 ("trunc<DFP_ALL:mode><BFP:mode>2")
13574 ("*extend<BFP:mode><DFP_ALL:mode>2")
13575 ("*extend<DFP_ALL:mode><BFP:mode>2")
13576 ("extend<BFP:mode><DFP_ALL:mode>2")
13577 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
13580 2013-07-08 Graham Stott <graham.stott@btinternet.com>
13582 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
13584 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13586 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
13587 and cfun_fpr_bit_p to cfun_fpr_save_p.
13588 (s390_frame_area, s390_register_info, s390_frame_info)
13589 (s390_emit_prologue, s390_emit_epilogue)
13590 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
13592 * config/s390/s390.h: Define *_REGNUM macros for floating point
13595 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
13597 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
13599 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
13601 PR rtl-optimization/57786
13602 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
13603 and break out of the loop when it is set to false.
13605 2013-07-08 Jakub Jelinek <jakub@redhat.com>
13608 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
13609 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
13610 (const_int 63)) 0)).
13611 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
13612 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
13613 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
13615 PR rtl-optimization/57829
13616 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
13617 mask bits outside of mode are just sign-extension from mode to HWI.
13619 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
13621 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
13622 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
13623 adjust_address instead of change_address to keep info about alignment.
13624 (emit_strmov): Remove.
13625 (emit_memmov): New function.
13626 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
13627 (expand_movmem_epilogue): Likewise and return updated rtx for
13629 (expand_constant_movmem_prologue): Likewise and return updated rtx for
13630 destination and source.
13631 (decide_alignment): Refactor, handle vector_loop.
13632 (ix86_expand_movmem): Likewise.
13633 (ix86_expand_setmem): Likewise.
13634 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
13636 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
13638 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
13639 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
13641 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13643 * config/i386/sse.md (sse_movlhps): Change alternative 3
13644 of operand 2 to "m".
13646 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
13649 * config/i386/sse.md (iptr): New mode attribute.
13650 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
13651 (<sse>_vm<plusminus_insn><mode>3): Ditto.
13652 (<sse>_vmmul<mode>3): Ditto.
13653 (<sse>_vmdiv<mode>3): Ditto.
13654 (sse_vmrcpv4sf2): Ditto.
13655 (<sse>_vmsqrt<mode>2): Ditto.
13656 (sse_vmrsqrtv4sf2): Ditto.
13657 (<sse>_vm<code><mode>3): Ditto.
13658 (avx_vmcmp<mode>3): Ditto.
13659 (<sse>_vmmaskcmp<mode>3): Ditto.
13660 (<sse>_comi): Ditto.
13661 (<sse>_ucomi): Ditto.
13662 (*xop_vmfrcz_<mode>): Ditto.
13663 (*fmai_fmadd_<mode>): Ditto.
13664 (*fmai_fmsub_<mode>): Ditto.
13665 (*fmai_fnmadd_<mode>): Ditto.
13666 (*fmai_fnmsub_<mode>): Ditto.
13667 (*fma4i_vmfmadd_<mode>): Ditto.
13668 (*fma4i_vmfmsub_<mode>): Ditto.
13669 (*fma4i_vmfnmadd_<mode>): Ditto.
13670 (*fma4i_vmfnmsub_<mode>): Ditto.
13671 (*xop_vmfrcz_<mode>): Ditto.
13672 (sse_cvtps2pi): Ditto.
13673 (sse_cvttps2pi): Ditto.
13674 (sse_cvtss2si): Ditto.
13675 (sse_cvtss2si_2): Ditto.
13676 (sse_cvtss2siq_2): Ditto.
13677 (sse_cvttss2si): Ditto.
13678 (sse_cvttss2siq): Ditto.
13679 (sse_cvtsd2si): Ditto.
13680 (sse_cvtsd2si_2): Ditto.
13681 (sse_cvtsd2siq_2): Ditto.
13682 (sse_cvttsd2si): Ditto.
13683 (sse_cvttsd2siq): Ditto.
13684 (sse_cvtsd2ss): Ditto.
13685 (sse_cvtss2sd): Ditto.
13686 (avx2_pbroadcast<mode>): Ditto.
13687 (avx2_pbroadcast<mode>_1): Ditto.
13688 (*avx_vperm_broadcast_v4sf): Ditto.
13690 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
13691 (sse_movlhps): Ditto.
13692 (sse_storehps): Ditto.
13693 (sse_loadhps): Ditto.
13694 (sse_storelps): Ditto.
13695 (sse_loadlps): Ditto.
13696 (*vec_concatv4sf): Ditto.
13697 (*vec_interleave_highv2df): Ditto.
13698 (*vec_interleave_lowv2df): Ditto.
13699 (*vec_extractv2df_1_sse): Ditto.
13700 (*vec_extractv2df_0_sse): Ditto.
13701 (sse2_storelpd): Ditto.
13702 (sse2_loadlpd): Ditto.
13703 (sse2_movsd): Ditto.
13704 (*vec_concatv4si): Ditto.
13705 (vec_concatv2di): Ditto.
13707 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
13708 for Intel asm dialect.
13709 (mmx_punpcklwd): Ditto.
13710 (mmx_punpckldq): Ditto.
13712 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
13713 for intel assembler dialect.
13715 2013-07-06 Jakub Jelinek <jakub@redhat.com>
13718 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
13719 for BUILT_IN_C{LZ,LRSB}*.
13720 * tree.h (CASE_INT_FN): Add FN##IMAX case.
13721 * tree-vrp.c (extract_range_basic): Handle
13722 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
13723 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
13724 fall thru to code calling set_value*.
13725 * builtins.c (expand_builtin): Remove *IMAX cases.
13726 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
13727 if width is bigger than 2*HWI.
13729 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
13731 PR rtl-optimization/55342
13732 * lra-int.h (lra_subreg_reload_pseudos): New.
13733 * lra.c: Add undoing optional reloads to the block diagram.
13734 (lra_subreg_reload_pseudos): New.
13735 (lra_optional_reload_pseudos): Change comments.
13736 (lra): Init and clear lra_subreg_reload_pseudos. Clear
13737 lra_optional_reload_pseudos after undo transformations.
13738 * lra-assigns.c (pseudo_prefix_title): New.
13739 (lra_setup_reg_renumber): Use it.
13740 (spill_for): Ditto. Check subreg reload pseudos too.
13741 (assign_by_spills): Consider subreg reload pseudos too.
13742 * lra-constraints.c (simplify_operand_subreg): Use
13743 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
13744 (curr_insn_transform): Recognize and do optional reloads.
13745 (undo_optional_reloads): New.
13746 (lra_undo_inheritance): Call undo_optional_reloads.
13748 2013-07-05 Thomas Quinot <quinot@adacore.com>
13750 * tree-complex.c (expand_complex_operations_1): Fix typo.
13752 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
13754 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
13755 (tune_params): New member 'const vec_costs'.
13756 * config/aarch64/aarch64.c (generic_vector_cost): New.
13757 (generic_tunings): New member 'generic_vector_cost'.
13758 (aarch64_builtin_vectorization_cost): New.
13759 (aarch64_add_stmt_cost): New.
13760 (TARGET_VECTORIZE_ADD_STMT_COST): New.
13761 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
13763 2013-07-03 Jakub Jelinek <jakub@redhat.com>
13766 * config/i386/predicates.md (vsib_address_operand): Disallow
13767 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
13769 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
13771 PR middle-end/55030
13772 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
13773 expand_builtin_setjmp_receiver.
13774 (expand_label): Adjust, call expand_builtin_setjmp_receiver
13775 with NULL for the label parameter.
13776 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
13777 the frame-pointer. Adjust comments.
13778 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
13779 only if LABEL is non-NULL.
13781 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
13783 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
13784 (ARM_ABI_AAPCS64): Ditto.
13786 (ARM_DEFAULT_ABI): Ditto.
13788 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
13790 * config/aarch64/aarch64-builtins.c
13791 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
13792 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
13794 * config/aarch64/aarch64-simd.md
13795 (aarch64_ld1<VALL:mode>): New.
13796 (aarch64_st1<VALL:mode>): Likewise.
13797 * config/aarch64/arm_neon.h
13798 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
13800 2013-07-02 Sriraman Tallam <tmsriram@google.com>
13802 * config/i386/i386.c (gate_insert_vzeroupper): Check if
13804 (ix86_option_override_internal):Turn on all -mavx target flags by
13805 default as they are dependent on AVX anyway.
13807 2013-07-02 Cary Coutant <ccoutant@google.com>
13809 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
13810 deterministic hash.
13811 (loc_checksum_ordered): Likewise.
13812 (hash_loc_operands): Remove inline keyword.
13814 2013-07-02 Jakub Jelinek <jakub@redhat.com>
13816 PR tree-optimization/57741
13817 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
13818 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
13819 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
13820 Allow REAL_CST step_exprs if flag_associative_math.
13821 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
13823 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13825 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
13827 2013-07-02 Ian Bolton <ian.bolton@arm.com>
13829 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
13831 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13833 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
13835 (iorsi3_insn): Likewise.
13836 (arm_xorsi3): Likewise.
13838 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
13840 * arm.md (attribute "wtype"): Delete. Move attribute values from here
13842 (attribute "type"): ... here, and prefix with "wmmx_".
13843 (attribute "core_cycles"): Update for attribute changes.
13844 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
13845 (tbcstv4hi): Likewise.
13846 (tbcstv2si): Likewise.
13847 (iwmmxt_iordi3): Likewise.
13848 (iwmmxt_xordi3): Likewise.
13849 (iwmmxt_anddi3): Likewise.
13850 (iwmmxt_nanddi3): Likewise.
13851 (iwmmxt_arm_movdi): Likewise.
13852 (iwmmxt_movsi_insn): Likewise.
13853 (mov<mode>_internal): Likewise.
13854 (and<mode>3_iwmmxt): Likewise.
13855 (ior<mode>3_iwmmxt): Likewise.
13856 (xor<mode>3_iwmmxt): Likewise.
13857 (add<mode>3_iwmmxt): Likewise.
13858 (ssaddv8qi3): Likewise.
13859 (ssaddv4hi3): Likewise.
13860 (ssaddv2si3): Likewise.
13861 (usaddv8qi3): Likewise.
13862 (usaddv4hi3): Likewise.
13863 (usaddv2si3): Likewise.
13864 (sub<mode>3_iwmmxt): Likewise.
13865 (sssubv8qi3): Likewise.
13866 (sssubv4hi3): Likewise.
13867 (sssubv2si3): Likewise.
13868 (ussubv8qi3): Likewise.
13869 (ussubv4hi3): Likewise.
13870 (ussubv2si3): Likewise.
13871 (mulv4hi3_iwmmxt): Likewise.
13872 (smulv4hi3_highpart): Likewise.
13873 (umulv4hi3_highpart): Likewise.
13874 (iwmmxt_wmacs): Likewise.
13875 (iwmmxt_wmacsz): Likewise.
13876 (iwmmxt_wmacu): Likewise.
13877 (iwmmxt_wmacuz): Likewise.
13878 (iwmmxt_clrdi): Likewise.
13879 (iwmmxt_clrv8qi): Likewise.
13880 (iwmmxt_clr4hi): Likewise.
13881 (iwmmxt_clr2si): Likewise.
13882 (iwmmxt_uavgrndv8qi3): Likewise.
13883 (iwmmxt_uavgrndv4hi3): Likewise.
13884 (iwmmxt_uavgv8qi3): Likewise.
13885 (iwmmxt_uavgv4hi3): Likewise.
13886 (iwmmxt_tinsrb): Likewise.
13887 (iwmmxt_tinsrh): Likewise.
13888 (iwmmxt_tinsrw): Likewise.
13889 (iwmmxt_textrmub): Likewise.
13890 (iwmmxt_textrmsb): Likewise.
13891 (iwmmxt_textrmuh): Likewise.
13892 (iwmmxt_textrmsh): Likewise.
13893 (iwmmxt_textrmw): Likewise.
13894 (iwmxxt_wshufh): Likewise.
13895 (eqv8qi3): Likewise.
13896 (eqv4hi3): Likewise.
13897 (eqv2si3): Likewise.
13898 (gtuv8qi3): Likewise.
13899 (gtuv4hi3): Likewise.
13900 (gtuv2si3): Likewise.
13901 (gtv8qi3): Likewise.
13902 (gtv4hi3): Likewise.
13903 (gtv2si3): Likewise.
13904 (smax<mode>3_iwmmxt): Likewise.
13905 (umax<mode>3_iwmmxt): Likewise.
13906 (smin<mode>3_iwmmxt): Likewise.
13907 (umin<mode>3_iwmmxt): Likewise.
13908 (iwmmxt_wpackhss): Likewise.
13909 (iwmmxt_wpackwss): Likewise.
13910 (iwmmxt_wpackdss): Likewise.
13911 (iwmmxt_wpackhus): Likewise.
13912 (iwmmxt_wpackwus): Likewise.
13913 (iwmmxt_wpackdus): Likewise.
13914 (iwmmxt_wunpckihb): Likewise.
13915 (iwmmxt_wunpckihh): Likewise.
13916 (iwmmxt_wunpckihw): Likewise.
13917 (iwmmxt_wunpckilb): Likewise.
13918 (iwmmxt_wunpckilh): Likewise.
13919 (iwmmxt_wunpckilw): Likewise.
13920 (iwmmxt_wunpckehub): Likewise.
13921 (iwmmxt_wunpckehuh): Likewise.
13922 (iwmmxt_wunpckehuw): Likewise.
13923 (iwmmxt_wunpckehsb): Likewise.
13924 (iwmmxt_wunpckehsh): Likewise.
13925 (iwmmxt_wunpckehsw): Likewise.
13926 (iwmmxt_wunpckelub): Likewise.
13927 (iwmmxt_wunpckeluh): Likewise.
13928 (iwmmxt_wunpckeluw): Likewise.
13929 (iwmmxt_wunpckelsb): Likewise.
13930 (iwmmxt_wunpckelsh): Likewise.
13931 (iwmmxt_wunpckelsw): Likewise.
13932 (ror<mode>3): Likewise.
13933 (ashr<mode>3_iwmmxt): Likewise.
13934 (lshr<mode>3_iwmmxt): Likewise.
13935 (ashl<mode>3_iwmmxt): Likewise.
13936 (ror<mode>3_di): Likewise.
13937 (ashr<mode>3_di): Likewise.
13938 (lshr<mode>3_di): Likewise.
13939 (ashl<mode>3_di): Likewise.
13940 (iwmmxt_wmadds): Likewise.
13941 (iwmmxt_wmaddu): Likewise.
13942 (iwmmxt_tmia): Likewise.
13943 (iwmmxt_tmiaph): Likewise.
13944 (iwmmxt_tmiabb): Likewise.
13945 (iwmmxt_tmiatb): Likewise.
13946 (iwmmxt_tmiabt): Likewise.
13947 (iwmmxt_tmiatt): Likewise.
13948 (iwmmxt_tmovmskb): Likewise.
13949 (iwmmxt_tmovmskh): Likewise.
13950 (iwmmxt_tmovmskw): Likewise.
13951 (iwmmxt_waccb): Likewise.
13952 (iwmmxt_wacch): Likewise.
13953 (iwmmxt_waccw): Likewise.
13954 (iwmmxt_waligni): Likewise.
13955 (iwmmxt_walignr): Likewise.
13956 (iwmmxt_walignr0): Likewise.
13957 (iwmmxt_walignr1): Likewise.
13958 (iwmmxt_walignr2): Likewise.
13959 (iwmmxt_walignr3): Likewise.
13960 (iwmmxt_wsadb): Likewise.
13961 (iwmmxt_wsadh): Likewise.
13962 (iwmmxt_wsadbz): Likewise.
13963 (iwmmxt_wsadhz): Likewise.
13964 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
13965 (iwmmxt_wabsdiffb): Likewise.
13966 (iwmmxt_wabsdiffh): Likewise.
13967 (iwmmxt_wabsdiffw): Likewise.
13968 (iwmmxt_waddsubhx): Likewise
13969 (iwmmxt_wsubaddhx): Likewise.
13970 (addc<mode>3): Likewise.
13971 (iwmmxt_avg4): Likewise.
13972 (iwmmxt_avg4r): Likewise.
13973 (iwmmxt_wmaddsx): Likewise.
13974 (iwmmxt_wmaddux): Likewise.
13975 (iwmmxt_wmaddsn): Likewise.
13976 (iwmmxt_wmaddun): Likewise.
13977 (iwmmxt_wmulwsm): Likewise.
13978 (iwmmxt_wmulwum): Likewise.
13979 (iwmmxt_wmulsmr): Likewise.
13980 (iwmmxt_wmulumr): Likewise.
13981 (iwmmxt_wmulwsmr): Likewise.
13982 (iwmmxt_wmulwumr): Likewise.
13983 (iwmmxt_wmulwl): Likewise.
13984 (iwmmxt_wqmulm): Likewise.
13985 (iwmmxt_wqmulwm): Likewise.
13986 (iwmmxt_wqmulmr): Likewise.
13987 (iwmmxt_wqmulwmr): Likewise.
13988 (iwmmxt_waddbhusm): Likewise.
13989 (iwmmxt_waddbhusl): Likewise.
13990 (iwmmxt_wqmiabb): Likewise.
13991 (iwmmxt_wqmiabt): Likewise.
13992 (iwmmxt_wqmiatb): Likewise.
13993 (iwmmxt_wqmiatt): Likewise.
13994 (iwmmxt_wqmiabbn): Likewise.
13995 (iwmmxt_wqmiabtn): Likewise.
13996 (iwmmxt_wqmiatbn): Likewise.
13997 (iwmmxt_wqmiattn): Likewise.
13998 (iwmmxt_wmiabb): Likewise.
13999 (iwmmxt_wmiabt): Likewise.
14000 (iwmmxt_wmiatb): Likewise.
14001 (iwmmxt_wmiatt): Likewise.
14002 (iwmmxt_wmiabbn): Likewise.
14003 (iwmmxt_wmiabtn): Likewise.
14004 (iwmmxt_wmiatbn): Likewise.
14005 (iwmmxt_wmiattn): Likewise.
14006 (iwmmxt_wmiawbb): Likewise.
14007 (iwmmxt_wmiawbt): Likewise.
14008 (iwmmxt_wmiawtb): Likewise.
14009 (iwmmxt_wmiawtt): Likewise.
14010 (iwmmxt_wmiawbbn): Likewise.
14011 (iwmmxt_wmiawbtn): Likewise.
14012 (iwmmxt_wmiawtbn): Likewise.
14013 (iwmmxt_wmiawttn): Likewise.
14014 (iwmmxt_wmerge): Likewise.
14015 (iwmmxt_tandc<mode>3): Likewise.
14016 (iwmmxt_torc<mode>3): Likewise.
14017 (iwmmxt_torvsc<mode>3): Likewise.
14018 (iwmmxt_textrc<mode>3): Likewise.
14019 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
14020 (wmmxt_pack): Likewise.
14021 (wmmxt_mult_c1): Likewise.
14022 (wmmxt_mult_c2): Likewise.
14023 (wmmxt_alu_c1): Likewise.
14024 (wmmxt_alu_c2): Likewise.
14025 (wmmxt_alu_c3): Likewise.
14026 (wmmxt_transfer_c1): Likewise.
14027 (wmmxt_transfer_c2): Likewise.
14028 (wmmxt_transfer_c3): Likewise.
14029 (marvell_f_iwmmxt_wstr): Likewise.
14030 (marvell_f_iwmmxt_wldr): Likewise.
14032 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
14034 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
14036 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14039 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14040 * lra-constraints.c (need_for_split_p): Check call used hard regs
14041 living through calls.
14043 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
14044 call used regs for call insn.
14046 2013-06-28 Jakub Jelinek <jakub@redhat.com>
14049 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
14050 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
14052 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
14054 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
14056 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
14058 * lra-constraints.c (need_for_split_p): Check call used hard regs
14059 living through calls.
14061 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
14064 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
14065 to tie with any other modes. Eliminate Altivec vector mode tests,
14066 since these are a subset of ALTIVEC or VSX vector modes. Simplify
14067 code, to return 0 if testing MODE2 for a condition, if we've
14068 already tested MODE1 for the same condition.
14070 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14072 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
14075 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14077 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
14078 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
14080 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
14082 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
14084 (aarch64_symbolic_constant_p): Remove.
14085 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
14086 static. Fix line length and white space.
14087 (aarch64_symbolic_constant_p): Remove.
14088 * config/aarch64/predicates.md (aarch64_valid_symref):
14089 Use aarch64_classify_symbol_expression.
14091 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14093 * config/arm/constraints.md (Ts): New constraint.
14094 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
14096 (compare_scc): Use "Ts" constraint for operand 0.
14097 (ior_scc_scc): Likewise.
14098 (and_scc_scc): Likewise.
14099 (and_scc_scc_nodom): Likewise.
14100 (ior_scc_scc_cmp): Likewise for operand 7.
14101 (and_scc_scc_cmp): Likewise.
14102 * config/arm/thumb2.md (thumb2_movsi_insn):
14103 Add alternatives for 16-bit encodings.
14104 (thumb2_movhi_insn): Likewise.
14105 (thumb2_movsicc_insn): Likewise.
14106 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
14107 (thumb2_negscc): Use "Ts" constraint.
14108 Move mvn instruction outside cond_exec block.
14109 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
14110 for 16-bit encodings.
14112 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14114 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
14116 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
14117 (mulsi3subsi): Likewise.
14118 (mulsidi3adddi): Likewise.
14119 (mulsidi3_v6): Likewise.
14120 (umulsidi3_v6): Likewise.
14121 (umulsidi3adddi_v6): Likewise.
14122 (smulsi3_highpart_v6): Likewise.
14123 (umulsi3_highpart_v6): Likewise.
14124 (mulhisi3tb): Likewise.
14125 (mulhisi3bt): Likewise.
14126 (mulhisi3tt): Likewise.
14127 (maddhisi4): Likewise.
14128 (maddhisi4tb): Likewise.
14129 (maddhisi4tt): Likewise.
14130 (maddhidi4): Likewise.
14131 (maddhidi4tb): Likewise.
14132 (maddhidi4tt): Likewise.
14133 (zeroextractsi_compare0_scratch): Likewise.
14134 (insv_zero): Likewise.
14135 (insv_t2): Likewise.
14136 (anddi_notzesidi_di): Likewise.
14137 (anddi_notsesidi_di): Likewise.
14138 (andsi_notsi_si): Likewise.
14139 (iordi_zesidi_di): Likewise.
14140 (xordi_zesidi_di): Likewise.
14141 (andsi_iorsi3_notsi): Likewise.
14142 (smax_0): Likewise.
14143 (smax_m1): Likewise.
14144 (smin_0): Likewise.
14145 (not_shiftsi): Likewise.
14146 (unaligned_loadsi): Likewise.
14147 (unaligned_loadhis): Likewise.
14148 (unaligned_loadhiu): Likewise.
14149 (unaligned_storesi): Likewise.
14150 (unaligned_storehi): Likewise.
14151 (extv_reg): Likewise.
14152 (extzv_t2): Likewise.
14153 (divsi3): Likewise.
14154 (udivsi3): Likewise.
14155 (arm_zero_extendhisi2addsi): Likewise.
14156 (arm_zero_extendqisi2addsi): Likewise.
14157 (compareqi_eq0): Likewise.
14158 (arm_extendhisi2_v6): Likewise.
14159 (arm_extendqisi2addsi): Likewise.
14160 (arm_movt): Likewise.
14161 (thumb2_ldrd): Likewise.
14162 (thumb2_ldrd_base): Likewise.
14163 (thumb2_ldrd_base_neg): Likewise.
14164 (thumb2_strd): Likewise.
14165 (thumb2_strd_base): Likewise.
14166 (thumb2_strd_base_neg): Likewise.
14167 (arm_negsi2): Add alternative for 16-bit encoding.
14168 (arm_one_cmplsi2): Likewise.
14170 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14172 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
14173 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
14174 (movdfcc): Likewise.
14175 * config/arm/vfp.md (*thumb2_movsf_vfp):
14176 Disable predication for arm_restrict_it.
14177 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
14178 (*thumb2_movdfcc_vfp): Likewise.
14179 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
14180 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
14181 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
14182 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
14183 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
14184 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
14185 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
14186 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
14187 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
14188 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
14189 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
14190 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
14191 Disable predication for arm_restrict_it.
14193 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
14195 * config/i386/bmiintrin.h (_bextr_u32): New.
14196 (_bextr_u64): Ditto.
14198 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
14200 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
14201 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
14202 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
14203 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
14204 * config/mips/n32-elf.h: ...this new file.
14206 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
14209 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
14210 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
14212 2013-06-27 Catherine Moore <clm@codesourcery.com>
14214 * config/mips/mips-tables.opt: Regenerate.
14215 * config/mips/mips-cpus.def: Add m14ke and m14kec.
14216 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
14217 * doc/invoke.texi: Add -m14kc.
14219 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14222 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
14223 constraints of operand 1 and 2.
14226 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
14227 to match RTL canonicalization. Swap predicates and
14228 constraints of operand 1 and 2.
14230 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
14232 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
14233 Process OP_INOUT regs for splitting too.
14235 2013-06-27 Jakub Jelinek <jakub@redhat.com>
14237 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
14238 decl before the loop, initialize to NULL.
14239 (vectorizable_load): Initialize ptr_incr to NULL.
14241 2013-06-27 Martin Jambor <mjambor@suse.cz>
14244 * ipa-ref.h (ipa_maybe_record_reference): Declare.
14245 * ipa-ref.c (ipa_maybe_record_reference): New function.
14246 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
14247 * ipa-cp.c (create_specialized_node): Record potential references from
14249 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
14251 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14253 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
14254 parameter 'mode' of type 'enum machine_mode mode'; change to pass
14255 'mode' to force_reg.
14256 (aarch64_add_offset): Update calls to aarch64_force_temporary.
14257 (aarch64_expand_mov_immediate): Likewise.
14259 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
14261 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
14262 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
14264 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14266 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
14267 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14268 (struct machine_function): Add tbegin_p.
14269 (s390_canonicalize_comparison): Fold CC mode compares to
14270 conditional jump if possible.
14271 (s390_emit_jump): Return the emitted jump.
14272 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
14273 Handle CCRAWmode compares.
14274 (s390_option_override): Default to -mhtm if available.
14275 (s390_reg_clobbered_rtx): Handle floating point regs as well.
14276 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
14277 FPRs instead of df_regs_ever_live_p.
14278 (s390_optimize_nonescaping_tx): New function.
14279 (s390_init_frame_layout): Extend clobbered_regs array to cover
14281 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
14282 (s390_expand_tbegin): New function.
14283 (enum s390_builtin): New enum definition.
14284 (code_for_builtin): New array definition.
14285 (s390_init_builtins): New function.
14286 (s390_expand_builtin): New function.
14287 (TARGET_INIT_BUILTINS): Define.
14288 (TARGET_EXPAND_BUILTIN): Define.
14289 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
14290 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
14291 (s390_alc_comparison): Likewise.
14292 * config/s390/s390-modes.def: Add CCRAWmode.
14293 * config/s390/s390.h (processor_flags): Add PF_TX.
14294 (TARGET_CPU_HTM): Define macro.
14295 (TARGET_HTM): Define macro.
14296 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
14297 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
14298 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
14299 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
14300 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
14301 (TBEGIN_MASK, TBEGINC_MASK): New constants.
14302 ("*cc_to_int"): Move up.
14303 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
14304 constants other than 0.
14305 ("*ccraw_to_int"): New insn and splitter definition.
14306 ("tbegin", "tbegin_nofloat", "tbegin_retry")
14307 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
14308 ("tx_assist"): New expander.
14309 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
14310 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
14311 * config/s390/s390.opt: Add -mhtm option.
14312 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
14313 * config/s390/htmxlintrin.h: New file.
14314 * config/s390/htmintrin.h: New file.
14315 * config/s390/s390intrin.h: New file.
14316 * doc/extend.texi: Document htm builtins.
14317 * config.gcc: Add the new header files to extra_headers.
14319 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
14321 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
14322 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
14324 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
14325 Pat Haugen <pthaugen@us.ibm.com>
14326 Peter Bergner <bergner@vnet.ibm.com>
14328 * config/rs6000/power8.md: New.
14329 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
14330 setting for power8 entry.
14331 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
14332 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
14333 test for Power4/Power5 only.
14334 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
14336 (force_new_group): Adjust comment.
14337 * config/rs6000/rs6000.md: Include power8.md.
14339 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
14341 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
14342 * config/arm/arm-protos.h (arm_max_conditional_execute): New
14344 (tune_params): Update comment.
14345 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
14346 (arm_max_conditional_execute): New function.
14347 (thumb2_final_prescan_insn): Use max_insn_skipped and
14348 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
14350 2013-06-25 Jakub Jelinek <jakub@redhat.com>
14352 PR tree-optimization/57705
14353 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
14354 SSA_NAME step, provided that it is not defined inside the loop.
14355 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
14356 (get_initial_def_for_induction): Handle SSA_NAME IV step.
14358 2013-06-25 Martin Jambor <mjambor@suse.cz>
14360 PR middle-end/57670
14361 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
14362 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
14364 (ipa_note_param_call): Initialize member_ptr flag.
14365 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
14366 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
14367 (ipa_write_indirect_edge_info): Stream member_ptr flag.
14368 (ipa_read_indirect_edge_info): Likewise.
14370 2013-06-25 Richard Biener <rguenther@suse.de>
14372 PR middle-end/56977
14373 * passes.c (init_optimization_passes): Move pass_fold_builtins
14374 and pass_dce earlier with -Og.
14376 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
14378 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
14379 <BIT_FIELD_REF>: Remove trailing TAB.
14380 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
14383 2013-06-24 Martin Jambor <mjambor@suse.cz>
14385 PR tree-optimization/57358
14386 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
14387 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
14388 (ipa_analyze_params_uses): Generate pessimistic info when true.
14390 2013-06-24 Martin Jambor <mjambor@suse.cz>
14392 PR tree-optimization/57539
14393 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
14394 global.inlined_to of the new node to it. All callers changed.
14395 * ipa-inline-transform.c (clone_inlined_nodes): New variable
14396 inlining_into, pass it to cgraph_clone_node.
14397 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
14398 ipa_free_edge_args_substructures.
14399 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
14400 rdesc linked list. Do not assert rdesc edges have inlined caller.
14401 Assert we have found an rdesc in the rdesc list.
14403 2013-06-24 Richard Biener <rguenther@suse.de>
14405 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
14406 (pointer_set_lookup): Declare.
14407 (class pointer_map): New template class implementing a
14408 generic pointer to T map.
14409 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
14410 pointer_map<T>::contains, pointer_map<T>::insert,
14411 pointer_map<T>::traverse): New functions.
14412 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
14413 (pointer_set_lookup): New function.
14414 (pointer_set_contains): Use pointer_set_lookup.
14415 (pointer_set_insert): Likewise.
14416 (insert_aux): Remove.
14417 (struct pointer_map_t): Embed a pointer_set_t.
14418 (pointer_map_create): Adjust.
14419 (pointer_map_destroy): Likewise.
14420 (pointer_map_contains): Likewise.
14421 (pointer_map_insert): Likewise.
14422 (pointer_map_traverse): Likewise.
14423 * tree-streamer.h (struct streamer_tree_cache_d): Use a
14424 pointer_map<unsigned> instead of a pointer_map_t.
14425 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
14426 (streamer_tree_cache_lookup): Likewise.
14427 (streamer_tree_cache_create): Likewise.
14428 (streamer_tree_cache_delete): Likewise.
14429 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
14430 pointer_map<unsigned> instead of a pointer_map_t.
14431 (lto_init_tree_ref_encoder): Adjust.
14432 (lto_destroy_tree_ref_encoder): Likewise.
14433 * lto-section-out.c (lto_output_decl_index): Likewise.
14434 (lto_record_function_out_decl_state): Likewise.
14435 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
14437 2013-06-24 Richard Biener <rguenther@suse.de>
14439 PR tree-optimization/57488
14440 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
14442 2013-06-24 Alan Modra <amodra@gmail.com>
14444 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
14445 (gen_easy_altivec_constant): Likewise.
14446 * config/rs6000/predicates.md (easy_vector_constant_add_self,
14447 easy_vector_constant_msb): Likewise.
14449 2013-06-23 Jakub Jelinek <jakub@redhat.com>
14452 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
14453 add missing return true.
14455 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
14458 * config/sh/predicates.md (general_extend_operand): Invoke
14459 general_movsrc_operand for memory operands.
14460 (general_movsrc_operand): Allow reg+reg addressing, do not use
14461 general_operand for memory operands.
14463 2013-06-23 Sriraman Tallam <tmsriram@google.com>
14465 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
14466 when current target options does not apply.
14467 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
14468 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
14469 * config/i386/bmiintrin.h: Pass appropriate target
14470 attributes to header.
14471 * config/i386/mmintrin.h: Ditto.
14472 * config/i386/nmmintrin.h: Ditto.
14473 * config/i386/avx2intrin.h: Ditto.
14474 * config/i386/fxsrintrin.h: Ditto.
14475 * config/i386/tbmintrin.h: Ditto.
14476 * config/i386/xsaveintrin.h: Ditto.
14477 * config/i386/f16cintrin.h: Ditto.
14478 * config/i386/xtestintrin.h: Ditto.
14479 * config/i386/xsaveoptintrin.h: Ditto.
14480 * config/i386/bmi2intrin.h: Ditto.
14481 * config/i386/lzcntintrin.h: Ditto.
14482 * config/i386/smmintrin.h: Ditto.
14483 * config/i386/wmmintrin.h: Ditto.
14484 * config/i386/x86intrin.h: Remove all header include guards.
14485 * config/i386/prfchwintrin.h: Ditto.
14486 * config/i386/pmmintrin.h: Ditto.
14487 * config/i386/tmmintrin.h: Ditto.
14488 * config/i386/xmmintrin.h: Ditto.
14489 * config/i386/popcntintrin.h: Ditto.
14490 * config/i386/rdseedintrin.h: Ditto.
14491 * config/i386/ammintrin.h: Ditto.
14492 * config/i386/emmintrin.h: Ditto.
14493 * config/i386/immintrin.h: Remove all header include guards.
14494 * config/i386/fma4intrin.h: Ditto.
14495 * config/i386/lwpintrin.h: Ditto.
14496 * config/i386/xopintrin.h: Ditto.
14497 * config/i386/ia32intrin.h: Ditto.
14498 * config/i386/avxintrin.h: Ditto.
14499 * config/i386/rtmintrin.h: Ditto.
14500 * config/i386/fmaintrin.h: Ditto.
14501 * config/i386/mm3dnow.h: Ditto.
14503 2013-06-22 Sriraman Tallam <tmsriram@google.com>
14505 * common/config/i386/i386-common.c: Handle LZCNT.
14507 2013-06-22 Andi Kleen <ak@linux.intel.com>
14509 * doc/extend.texi: Use __atomic_store_n instead of
14510 __atomic_store in HLE example.
14512 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
14514 * config/sh/sh.c: Remove <cstdlib> workaround.
14516 2013-06-21 Andi Kleen <ak@linux.intel.com>
14518 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
14520 2013-06-21 Andi Kleen <ak@linux.intel.com>
14522 * doc/extend.texi: Document that __atomic_clear and
14523 __atomic_test_and_set should only be used with bool.
14525 2013-06-20 Jan Hubicka <jh@suse.cz>
14527 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
14528 types_same_for_odr.
14529 * tree.c (decls_same_for_odr): New function.
14530 (same_for_edr): New function.
14531 (types_same_for_odr): New function.
14532 (get_binfo_at_offset): Use it.
14533 * tree.h (types_same_for_odr): Declare.
14535 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
14536 Jason Merrill <jason@redhat.com>
14538 * system.h: Include <cstdlib> as well as <stdlib.h>.
14540 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
14543 * config/i386/i386.c (construct_container): Report error if
14544 long double is used with disabled x87 float returns.
14546 2013-06-20 Jan Hubicka <jh@suse.cz>
14548 * lto-cgraph.c (input_symtab): Do not set cgraph state.
14550 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
14552 PR rtl-optimization/57425
14553 PR rtl-optimization/57569
14554 * alias.c (write_dependence_p): Remove parameters mem_mode and
14555 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
14556 Changed all callers.
14557 (canon_anti_dependence): Get comments and semantics in sync.
14558 Add parameter mem_canonicalized. Changed all callers.
14559 * rtl.h (canon_anti_dependence): Update prototype.
14561 2013-06-20 Richard Biener <rguenther@suse.de>
14563 * data-streamer-in.c (streamer_read_uhwi): Optimize single
14564 byte case, inline streamer_read_uchar and defer section
14567 2013-06-20 Richard Biener <rguenther@suse.de>
14569 PR tree-optimization/57584
14570 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
14571 SSA names into the expanded expression that take part in
14572 abnormal coalescing.
14574 2013-06-19 Sharad Singhai <singhai@google.com>
14576 * gcov.c (print_usage): Handle new option.
14577 (process_args): Ditto.
14578 (get_gcov_intermediate_filename): New function.
14579 (output_intermediate_file): New function.
14580 (output_gcov_file): New function
14581 (generate_results): Handle new option.
14582 (release_function): Relase demangled name.
14583 (read_graph_file): Handle demangled name.
14584 (output_lines): Ditto.
14585 * doc/gcov.texi: Document gcov intermediate format.
14587 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
14590 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
14591 (lra_emit_add): Use the functions. Add comment about Y as an
14594 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
14597 * collect2.c (collect_atexit): New.
14598 (collect_exit): Delete.
14599 (main): Register collect_atexit with atexit.
14600 (collect_wait): Change collect_exit to exit.
14602 * collect2.h (collect_exit): Delete.
14603 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
14605 2013-06-19 Wei Mi <wmi@google.com>
14607 PR rtl-optimization/57518
14608 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
14609 if regno is used in paradoxical subreg.
14610 (update_equiv_regs): Check pdx_subregs[regno] before
14611 set a reg to be equivalent with a mem.
14613 2013-06-19 Matthias Klose <doko@ubuntu.com>
14616 * file-find.h (find_a_file): Add a mode parameter.
14617 * file-find.c (find_a_file): Likewise.
14618 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
14619 with X_OK for the executables.
14620 * collect2.c (main): Call find_a_file with X_OK.
14622 2013-06-19 Steve Ellcey <sellcey@mips.com>
14625 * config/mips/mips.md (casesi_internal_mips16_<mode>):
14626 Use NEXT_INSN instead of next_real_insn.
14628 2013-06-19 Jan Hubicka <jh@suse.cz>
14630 * cgraph.h (const_value_known_p): Replace by ...
14631 (ctor_for_folding): .. this one.
14632 * cgraphunit.c (process_function_and_variable_attributes): Use it.
14633 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
14634 * expr.c (expand_expr_real_1): Likewise.
14635 (string_constant): Likewise.
14636 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
14637 * ipa.c (process_references): Likewise.
14638 (symtab_remove_unreachable_nodes): Likewise.
14639 * ipa-inline-analysis.c (param_change_prob): Likewise.
14640 * gimple-fold.c (canonicalize_constructor_val): Likewise.
14641 (get_base_constructor): Likwise.
14642 * varpool.c (varpool_remove_node): Likewise.
14643 (varpool_remove_initializer): LIkewise.
14644 (dump_varpool_node): LIkwise.
14645 (const_value_known_p): Rewrite to ...
14646 (ctor_for_folding): ... this one.
14648 2013-06-19 Jakub Jelinek <jakub@redhat.com>
14651 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
14652 PERSONALITY in $PATH derived prefixes.
14654 2013-06-19 Jeff Law <law@redhat.com>
14656 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
14659 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
14660 (simplify_bitwise_binary): Use it to simpify certain binary ops on
14663 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
14665 * config/arm/vfp.md: Move VFP instruction classification documentation
14667 * config/arm/arm.md: ... here. Update instruction classification
14670 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
14672 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
14673 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
14674 pattern. Use more efficient sequences on ARMv5 and Thumb2.
14676 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
14679 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14680 with NEXT_INSN. Use tablejump_p to check for jump table data
14683 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
14686 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
14687 that now in C++ the value is correct per the C++ standards.
14689 2013-06-19 Richard Biener <rguenther@suse.de>
14691 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
14692 for global context.
14694 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14697 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14700 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14701 with next_active_insn.
14703 2013-06-18 Sriraman Tallam <tmsriram@google.com>
14705 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
14706 functions are inlined during failures to flag an error.
14707 * tree-inline.c (expand_call_inline): Allow the error to be flagged
14708 in early inline pass.
14710 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
14712 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
14715 2013-06-18 Julian Brown <julian@codesourcery.com>
14717 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
14718 Permit virtual register pre-reload if !strict.
14719 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
14721 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
14723 * config/arm/neon.md (movmisalign<mode>): Use
14724 neon_perm_struct_or_reg_operand instead of
14725 neon_struct_or_register_operand.
14726 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
14727 neon_permissive_struct_operand instead of neon_struct_operand.
14728 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
14729 neon_vector_mem_operand.
14730 * config/arm/predicates.md (neon_struct_operand): Adjust call to
14731 neon_vector_mem_operand.
14732 (neon_permissive_struct_operand): New.
14733 (neon_struct_or_register_operand): Rename to...
14734 (neon_perm_struct_or_reg_operand): This. Adjust call to
14735 neon_vector_mem_operand.
14737 2013-06-18 Richard Biener <rguenther@suse.de>
14739 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
14740 * lto-streamer.h: Include pointer-set.h.
14741 (struct lto_decl_slot): Remove.
14742 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
14743 Remove next_index entry.
14744 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14745 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14746 (lto_init_tree_ref_encoder): Adjust.
14747 (lto_destroy_tree_ref_encoder): Likewise.
14748 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
14749 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
14750 (lto_output_decl_index): Adjust.
14751 (lto_new_out_decl_state): Likewise.
14752 (lto_record_function_out_decl_state): Likewise.
14753 * lto-streamer-out.c (copy_function): Likewise.
14755 2013-06-18 Richard Biener <rguenther@suse.de>
14757 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
14758 * cgraphunit.c: Include cfgloop.h.
14759 (init_lowered_empty_function): Initialize the loop tree.
14760 (assemble_thunk): Insert new BBs into loops.
14762 2013-06-18 Richard Biener <rguenther@suse.de>
14764 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
14765 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
14766 the map from cache entry to cache index optional.
14767 (streamer_tree_cache_replace_tree): Adjust accordingly.
14768 (streamer_tree_cache_append): Likewise.
14769 (streamer_tree_cache_delete): Likewise.
14770 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
14771 streamer cache map from cache entry to cache index.
14772 * lto-streamer-out.c (create_output_block): Adjust.
14774 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
14776 * config/arm/arm.md (attribute "insn"): Move multiplication and
14777 division attributes to...
14778 (attribute "type"): ... here. Remove mult.
14779 (attribute "mul32"): New attribute.
14780 (attribute "mul64"): Add umaal.
14781 (*arm_mulsi3): Update attributes.
14782 (*arm_mulsi3_v6): Likewise.
14783 (*thumb_mulsi3): Likewise.
14784 (*thumb_mulsi3_v6): Likewise.
14785 (*mulsi3_compare0): Likewise.
14786 (*mulsi3_compare0_v6): Likewise.
14787 (*mulsi_compare0_scratch): Likewise.
14788 (*mulsi_compare0_scratch_v6): Likewise.
14789 (*mulsi3addsi): Likewise.
14790 (*mulsi3addsi_v6): Likewise.
14791 (*mulsi3addsi_compare0): Likewise.
14792 (*mulsi3addsi_compare0_v6): Likewise.
14793 (*mulsi3addsi_compare0_scratch): Likewise.
14794 (*mulsi3addsi_compare0_scratch_v6): Likewise.
14795 (*mulsi3subsi): Likewise.
14796 (*mulsidi3adddi): Likewise.
14797 (*mulsi3addsi_v6): Likewise.
14798 (*mulsidi3adddi_v6): Likewise.
14799 (*mulsidi3_nov6): Likewise.
14800 (*mulsidi3_v6): Likewise.
14801 (*umulsidi3_nov6): Likewise.
14802 (*umulsidi3_v6): Likewise.
14803 (*umulsidi3adddi): Likewise.
14804 (*umulsidi3adddi_v6): Likewise.
14805 (*smulsi3_highpart_nov6): Likewise.
14806 (*smulsi3_highpart_v6): Likewise.
14807 (*umulsi3_highpart_nov6): Likewise.
14808 (*umulsi3_highpart_v6): Likewise.
14809 (mulhisi3): Likewise.
14810 (*mulhisi3tb): Likewise.
14811 (*mulhisi3bt): Likewise.
14812 (*mulhisi3tt): Likewise.
14813 (maddhisi4): Likewise.
14814 (*maddhisi4tb): Likewise.
14815 (*maddhisi4tt): Likewise.
14816 (maddhidi4): Likewise.
14817 (*maddhidi4tb): Likewise.
14818 (*maddhidi4tt): Likewise.
14819 (divsi3): Likewise.
14820 (udivsi3): Likewise.
14821 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
14822 (thumb2_mulsi_short_compare0): Likewise.
14823 (thumb2_mulsi_short_compare0_scratch): Likewise.
14824 * config/arm/arm1020e.md (1020mult1): Update attribute change.
14825 (1020mult2): Likewise.
14826 (1020mult3): Likewise.
14827 (1020mult4): Likewise.
14828 (1020mult5): Likewise.
14829 (1020mult6): Likewise.
14830 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
14832 (cortex_a15_mult64): Likewise.
14833 (cortex_a15_sdiv): Likewise.
14834 (cortex_a15_udiv): Likewise.
14835 * config/arm/arm1026ejs.md (mult1): Update attribute change.
14841 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
14842 (pj4_ir_div): Likewise.
14843 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
14844 (11_mult2): Likewise.
14845 (11_mult3): Likewise.
14846 (11_mult4): Likewise.
14847 (11_mult5): Likewise.
14848 (11_mult6): Likewise.
14849 (11_mult7): Likewise.
14850 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
14851 (cortex_a8_mla): Likewise.
14852 (cortex_a8_mull): Likewise.
14853 (cortex_a8_smulwy): Likewise.
14854 (cortex_a8_smlald): Likewise.
14855 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
14856 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
14857 (cortex_r4_mul_3): Likewise.
14858 (cortex_r4_mla_4): Likewise.
14859 (cortex_r4_mla_3): Likewise.
14860 (cortex_r4_smlald): Likewise.
14861 (cortex_r4_mull): Likewise.
14862 (cortex_r4_sdiv): Likewise.
14863 (cortex_r4_udiv): Likewise.
14864 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
14865 (cortex_a7_idiv): Likewise.
14866 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
14867 (9_mult2): Likewise.
14868 (9_mult3): Likewise.
14869 (9_mult4): Likewise.
14870 (9_mult5): Likewise.
14871 (9_mult6): Likewise.
14872 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
14873 (cortex_a53_sdiv): Likewise.
14874 (cortex_a53_udiv): Likewise.
14875 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
14876 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
14877 (mp626_mult2): Likewise.
14878 (mp626_mult3): Likewise.
14879 (mp626_mult4): Likewise.
14880 * config/arm/fa526.md (526_mult1): Update attribute change.
14881 (526_mult2): Likewise.
14882 * config/arm/arm-generic.md (mult): Update attribute change.
14883 (mult_ldsched_strongarm): Likewise.
14884 (mult_ldsched): Likewise.
14885 (multi_cycle): Likewise.
14886 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
14887 * config/arm/fa606te.md (606te_mult1): Update attribute change.
14888 (606te_mult2): Likewise.
14889 (606te_mult3): Likewise.
14890 (606te_mult4): Likewise.
14891 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
14892 (cortex_a9_mac16): Likewise.
14893 (cortex_a9_multiply): Likewise.
14894 (cortex_a9_mac): Likewise.
14895 (cortex_a9_multiply_long): Likewise.
14896 * config/arm/fa626te.md (626te_mult1): Update attribute change.
14897 (626te_mult2): Likewise.
14898 (626te_mult3): Likewise.
14899 (626te_mult4): Likewise.
14901 2013-06-18 Richard Biener <rguenther@suse.de>
14904 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
14906 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14909 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
14910 with next_active_insn.
14912 2013-06-18 Alan Modra <amodra@gmail.com>
14914 * config/rs6000/rs6000.h (enum data_align): New.
14915 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
14916 (DATA_ABI_ALIGNMENT): Define.
14917 (CONSTANT_ALIGNMENT): Correct comment.
14918 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
14919 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
14921 2013-06-17 David Malcolm <dmalcolm@redhat.com>
14923 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
14924 ATTRIBUTE_UNUSED marking.
14926 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
14928 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
14929 alternative and update.
14930 (aarch64_dup_lanedi): Delete.
14931 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
14932 * config/aarch64/aarch64-simd-builtins.def: Update.
14934 2013-06-17 Richard Biener <rguenther@suse.de>
14936 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
14937 (lto_input_scc): Declare.
14938 (lto_input_tree_1): Likewise.
14939 (struct lto_stats_d): Add num_tree_bodies_output and
14940 num_pickle_refs_output.
14941 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
14942 (lto_read_tree_1): Split out from ...
14943 (lto_read_tree): ... this.
14944 (lto_input_scc): New function.
14945 (lto_input_tree_1): Split out from ...
14946 (lto_input_tree): ... this. Handle LTO_tree_scc.
14947 (lto_data_in_create): Create the streamer cache without hashes.
14948 * lto-streamer-out.c (create_output_block): Create the streamer
14949 cache with hashes when not doing WPA.
14950 (lto_write_tree_1): Split out from ...
14951 (lto_write_tree): ... this.
14952 (get_symbol_initial_value): New function.
14953 (lto_output_tree_1): Split out from ...
14954 (lto_output_tree): ... this. Write trees as series of SCCs
14955 using a DFS walk via DFS_write_tree.
14956 (struct sccs, struct scc_entry): New types.
14957 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
14958 (DFS_write_tree_body): New function.
14959 (DFS_write_tree): Likewise.
14960 (hash_tree): Likewise.
14961 (scc_entry_compare): Likewise.
14962 (hash_scc): Likewise.
14963 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
14964 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
14965 TREE_CHAIN as regular reference.
14966 (streamer_read_integer_cst): Remove.
14967 (streamer_get_pickled_tree): Adjust.
14968 * tree-streamer-out.c (streamer_write_chain): Disable streaming
14969 of DECL_EXTERNALs in BLOCK_VARS for now.
14970 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
14972 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
14973 Add hash value argument and record that if hashes are recorded
14975 (streamer_tree_cache_insert_1): Adjust.
14976 (streamer_tree_cache_insert): Likewise.
14977 (streamer_tree_cache_insert_at): Rename to ...
14978 (streamer_tree_cache_replace_tree): ... this and adjust.
14979 (streamer_tree_cache_append): Adjust.
14980 (record_common_node): Likewise.
14981 (streamer_tree_cache_create): Add argument whether to
14982 record hash values together with trees.
14983 (streamer_tree_cache_delete): Adjust.
14984 * tree-streamer.h (struct streamer_tree_cache_d): Add
14986 (streamer_read_integer_cst): Remove.
14987 (streamer_tree_cache_insert): Adjust.
14988 (streamer_tree_cache_append): Likewise.
14989 (streamer_tree_cache_insert_at): Rename to ...
14990 (streamer_tree_cache_replace_tree): ... this and adjust.
14991 (streamer_tree_cache_create): Add argument whether to record hashes.
14992 (streamer_tree_cache_get): Rename to ...
14993 (streamer_tree_cache_get_tree): ... this.
14994 (streamer_tree_cache_get_hash): New function.
14995 * tree.c (cache_integer_cst): New function.
14996 * tree.h (cache_integer_cst): Declare.
14997 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
14998 * lto-symtab.c (lto_varpool_replace_node): Only release
14999 DECL_INITIAL of non-prevailing decls.
15000 * varpool.c (varpool_remove_initializer): Do not release
15001 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
15003 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
15005 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
15006 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
15007 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
15008 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
15009 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
15010 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
15011 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
15012 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
15013 instead of TARGET_64BIT.
15014 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
15015 Require ISA_HAS_<D>DIV.
15017 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
15019 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
15020 (mips*-*-linux*): Move default with_llsc setting to where other
15022 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
15024 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
15025 Likewise. Remove default with_tune setting. Move default float
15026 setting to its own block. Handle with_llsc in the same block as above.
15028 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
15030 PR rtl-optimization/57425
15031 PR rtl-optimization/57569
15032 * alias.c (write_dependence_p): Add new parameters mem_mode,
15033 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
15034 Changed all callers.
15035 (canon_anti_dependence): New function.
15036 * cse.c (check_dependence): Use canon_anti_dependence.
15037 * cselib.c (cselib_invalidate_mem): Likewise.
15038 * rtl.h (canon_anti_dependence): Declare.
15040 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
15042 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
15043 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
15044 ".set mips3" for 64-bit targets.
15046 2013-06-15 Dehao Chen <dehao@google.com>
15048 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
15049 * gimple-low.c (gimple_check_call_matching_types): Likewise.
15050 (gimple_check_call_args): Likewise.
15051 * value-prof.c (check_ic_target): Likewise.
15052 * ipa-inline.c (early_inliner): Likewise.
15053 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
15054 * cgraph.c (cgraph_create_edge_1): Likewise.
15055 (cgraph_make_edge_direct): Likewise.
15057 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
15060 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
15061 rs6000_output_move_128bit to handle emitting quad memory
15062 operations. Set attribute length to 8 bytes.
15064 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
15066 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
15068 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
15069 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
15070 (aarch64_<su>mlsl<mode>): Likewise.
15072 2013-06-14 Mike Stump <mikestump@comcast.net>
15074 * Makefile.in (TARGET_H): Add insn-codes.h.
15076 2013-06-14 Alan Modra <amodra@gmail.com>
15078 PR middle-end/57134
15079 PR middle-end/57586
15080 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
15081 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
15082 bitfield expansion when EXPAND_MEMORY.
15083 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
15085 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
15087 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
15088 test for clearing quad memory on 32-bit later.
15090 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
15092 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
15093 (fold_negate_expr): Likewise.
15094 (fold_real_zero_addition_p): Handle vectors.
15095 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
15097 2013-06-14 Alan Modra <amodra@gmail.com>
15099 * varasm.c (force_const_mem): Revert 2013-06-07 change.
15101 2013-06-13 Jan Hubicka <jh@suse.cz>
15103 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
15104 Local comdats are not externally visible.
15105 * symtab.c (dump_symtab_base): Dump externally visible.
15106 (verify_symtab_base): Verify back links in the symtab hash.
15108 2013-06-13 Bin Cheng <bin.cheng@arm.com>
15110 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
15111 CONVERT_EXPR as equal nodes.
15113 2013-06-13 Bin Cheng <bin.cheng@arm.com>
15115 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
15117 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
15119 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
15120 Generalize to complex and vector.
15121 * tree.c (build_all_ones_cst): New function.
15122 * tree.h (build_all_ones_cst): Declare it.
15124 2013-06-13 Alan Modra <amodra@gmail.com>
15126 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
15127 * config/rs6000/rs6000.md (signbittf2): New insn.
15128 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
15129 (abstf2_internal, cmptf_internal2): Likewise.
15130 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
15132 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
15133 Pat Haugen <pthaugen@us.ibm.com>
15134 Peter Bergner <bergner@vnet.ibm.com>
15136 * config/rs6000/rs6000.c (emit_load_locked): Add support for
15137 power8 byte, half-word, and quad-word atomic instructions.
15138 (emit_store_conditional): Likewise.
15139 (rs6000_expand_atomic_compare_and_swap): Likewise.
15140 (rs6000_expand_atomic_op): Likewise.
15142 * config/rs6000/sync.md (larx): Add new modes for power8.
15144 (AINT): New mode iterator to include TImode as well as normal
15145 integer modes on power8.
15146 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
15147 that VSX registers are not considered. Use AINT mode iterator
15148 instead of INT1 to allow inclusion of quad word atomic operations
15150 (load_locked<mode>): Likewise.
15151 (store_conditional<mode>): Likewise.
15152 (atomic_compare_and_swap<mode>): Likewise.
15153 (atomic_exchange<mode>): Likewise.
15154 (atomic_nand<mode>): Likewise.
15155 (atomic_fetch_<fetchop_name><mode>): Likewise.
15156 (atomic_nand_fetch<mode>): Likewise.
15157 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
15159 (ATOMIC): On power8, add QImode, HImode modes.
15160 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
15161 modes that promote to SImode.
15162 (load_lockedti): Convert TImode arguments to PTImode, so that we
15163 get a guaranteed even/odd register pair.
15164 (load_lockedpti): Likewise.
15165 (store_conditionalti): Likewise.
15166 (store_conditionalpti): Likewise.
15168 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
15169 atomic load/store instructions.
15172 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
15174 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
15176 (insn_count): New attribute, with most cases extracted from...
15177 (length): ...here. Redefine most cases in terms of insn_count.
15178 (single_insn): Delete.
15179 (can_delay): Use insn_count to check for single instructions.
15180 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
15181 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
15182 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
15183 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
15184 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
15185 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
15186 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
15187 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
15188 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
15189 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
15190 rather than "length".
15191 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
15192 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
15193 Use "insn_count" rather than "length".
15194 * config/mips/mips-dsp.md
15195 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
15196 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
15199 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
15201 PR tree-optimization/57361
15202 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
15204 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
15206 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
15208 (aarch64_simd_combine<mode>): New instruction expansion.
15209 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
15210 function prototype.
15211 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
15212 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
15214 2013-06-12 Jan Hubicka <jh@suse.cz>
15216 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
15217 decl has when in streaming stage.
15218 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
15219 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
15221 2013-06-12 Roland Stigge <stigge@antcom.de>
15224 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
15226 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15228 PR tree-optimization/57537
15229 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
15230 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
15232 2013-06-12 Richard Biener <rguenther@suse.de>
15234 * data-streamer.h (streamer_write_char_stream): CSE
15235 obs->current_pointer.
15236 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
15237 streamer_write_char_stream manually and optimize the resulting loop.
15238 (streamer_write_hwi_stream): Likewise.
15240 2013-06-12 Jan Hubicka <jh@suse.cz>
15242 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
15243 * cgraph.h (varpool_create_empty_node): Declare.
15244 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
15246 * symtab.c (symtab_unregister_node): Be lax about missin entries
15248 (symtab_get_node): Update comment.
15249 * varpool.c (varpool_create_empty_node): Break out from ...
15250 (varpool_node_for_decl): ... here.
15251 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
15253 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
15255 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
15256 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
15257 part. Use straight-line flow at the end.
15258 <COMPONENT_REF>: Remove superfluous else.
15259 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
15261 2013-06-12 Jakub Jelinek <jakub@redhat.com>
15264 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
15265 target hook even for !TREE_PUBLIC decls. If no resolution info
15266 is available, return false for common and external decls.
15268 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15270 * config/rl78/constraints.md (U): New constraint.
15271 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
15274 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
15277 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
15278 to allow returning address to AT_PLATFORM name.
15280 2013-06-11 Jan Hubicka <jh@suse.cz>
15282 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
15283 * cgraph.h (symtab_node_base): Add weakref flag.
15284 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
15285 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
15286 (output_weakrefs): Use weakref flag.
15287 * fold-const.c (simple_operand_p): Handle WEAK.
15288 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
15289 * ipa.c (varpool_externally_visible_p): Drop weakref.
15290 (function_and_variable_visibility): Update comment; fix weakref
15291 sanity checks; do not clear DECL_WEAK on them.
15292 * lto-cgraph.c (lto_output_node): update.
15293 (lto_output_varpool_node): Update.
15294 (input_overwrite_node): Update.
15295 (input_node): Update.
15296 (input_varpool_node): Update.
15297 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
15298 (lto_symtab_merge_symbols): Add sanity check.
15299 (lto_symtab_prevailing_decl): Do not special case weakrefs.
15300 * passes.c (rest_of_decl_compilation): Set static flag, too.
15301 * symtab.c (dump_symtab_base): Dump weakref.
15302 (verify_symtab_base): Sanity check weakrefs.
15303 (symtab_make_decl_local): Remove duplicated code.
15304 (symtab_alias_ultimate_target): Simplify.
15305 * varpool.c (varpool_create_variable_alias): Set weakref flag.
15307 2013-06-11 Tom de Vries <tom@codesourcery.com>
15309 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
15310 sequence_vect == NULL.
15312 2013-06-11 DJ Delorie <dj@redhat.com>
15314 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
15315 (rl78_unwind_word_mode): New.
15317 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15319 * final.c (debug_prefix_maps): Make static.
15321 2013-06-11 David Malcolm <dmalcolm@redhat.com>
15323 * function.c (initial_trampoline): Remove stray copy.
15325 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
15327 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
15329 2013-06-11 Martin Jambor <mjambor@suse.cz>
15331 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
15332 within bounds at the beginning of the function.
15334 2013-06-11 Alan Modra <amodra@gmail.com>
15336 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
15338 (get_named_section): Don't NULL !DECL_P decl.
15340 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
15342 * doc/invoke.texi (core-avx2): Document.
15344 (atom): Updated with MOVBE.
15346 2013-06-11 Richard Biener <rguenther@suse.de>
15348 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
15350 2013-06-11 Anton Blanchard <anton@samba.org>
15352 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
15353 correct shift value in little-endian mode.
15355 2013-06-11 Jakub Jelinek <jakub@redhat.com>
15358 * varasm.c (get_variable_align): Move #endif to the right place.
15360 2013-06-10 Cary Coutant <ccoutant@google.com>
15362 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
15363 for hash so that hash table traversal order is deterministic.
15365 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
15366 Pat Haugen <pthaugen@us.ibm.com>
15367 Peter Bergner <bergner@vnet.ibm.com>
15369 * config/rs6000/vector.md (GPR move splitter): Do not split moves
15370 of vectors in GPRS if they are direct moves or quad word load or
15373 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
15375 (direct_move_p): Likewise.
15376 (quad_load_store_p): Likewise.
15378 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
15379 classes into bins based on the physical register type.
15380 (reg_class_to_reg_type): Likewise.
15381 (IS_STD_REG_TYPE): Likewise.
15382 (IS_FP_VECT_REG_TYPE): Likewise.
15383 (reload_fpr_gpr): Arrays to determine what insn to use if we can
15384 use direct move instructions.
15385 (reload_gpr_vsx): Likewise.
15386 (reload_vsx_gpr): Likewise.
15387 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
15388 information that is a simplification of register classes. Also
15389 precalculate direct move reload helpers.
15390 (direct_move_p): New function to return true if the operation can
15391 be done as a direct move instruciton.
15392 (quad_load_store_p): New function to return true if the operation
15393 is a quad memory operation.
15394 (rs6000_legitimize_address): If quad memory, only allow register
15395 indirect for TImode addresses.
15396 (rs6000_legitimate_address_p): Likewise.
15397 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
15398 (rs6000_reload_register_type): Likewise.
15399 (register_to_reg_type): Return register type.
15400 (rs6000_secondary_reload_simple_move): New helper function for
15401 secondary reload and secondary memory needed to identify anything
15402 that is a simple move, and does not need reloading.
15403 (rs6000_secondary_reload_direct_move): New helper function for
15404 secondary reload to identify cases that can be done with several
15405 instructions via the direct move instructions.
15406 (rs6000_secondary_reload_move): New helper function for secondary
15407 reload to identify moves between register types that can be done.
15408 (rs6000_secondary_reload): Add support for quad memory operations
15409 and for direct move.
15410 (rs6000_secondary_memory_needed): Likewise.
15411 (rs6000_debug_secondary_memory_needed): Change argument names.
15412 (rs6000_output_move_128bit): New function to return the move to
15413 use for 128-bit moves, including knowing about the various
15414 limitations of quad memory operations.
15416 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
15417 memory operations. call rs6000_output_move_128bit for the actual
15418 instruciton(s) to generate.
15419 (vsx_movti_64bit): Likewise.
15421 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
15422 (UNSPEC_P8V_MTVSRWZ): Likewise.
15423 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
15424 (UNSPEC_P8V_MTVSRD): Likewise.
15425 (UNSPEC_P8V_XXPERMDI): Likewise.
15426 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
15427 (UNSPEC_FUSION_GPR): Likewise.
15428 (FMOVE128_GPR): New iterator for direct move.
15429 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
15430 (f32_sv): Likewise.
15431 (f32_dm): Likewise.
15432 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
15433 loads and direct move instructions.
15434 (zero_extendsidi2_lfiwzx): Likewise.
15435 (extendsidi2_lfiwax): Likewise.
15436 (extendsidi2_nocell): Likewise.
15437 (floatsi<mode>2_lfiwax): Likewise.
15438 (lfiwax): Likewise.
15439 (floatunssi<mode>2_lfiwzx): Likewise.
15440 (lfiwzx): Likewise.
15441 (fix_trunc<mode>_stfiwx): Likewise.
15442 (fixuns_trunc<mode>_stfiwx): Likewise.
15443 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
15444 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
15445 (parity<mode>2_cmpb): Set length/type attr.
15446 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
15447 for 'mr.' to fast_compare.
15448 (bpermd_<mode>): Change type attr to popcnt.
15449 (p8_fmrgow_<mode>): New insns for power8 direct move support.
15450 (p8_mtvsrwz_1): Likewise.
15451 (p8_mtvsrwz_2): Likewise.
15452 (reload_fpr_from_gpr<mode>): Likewise.
15453 (p8_mtvsrd_1): Likewise.
15454 (p8_mtvsrd_2): Likewise.
15455 (p8_xxpermdi_<mode>): Likewise.
15456 (reload_vsx_from_gpr<mode>): Likewise.
15457 (reload_vsx_from_gprsf): Likewise.
15458 (p8_mfvsrd_3_<mode>): LIkewise.
15459 (reload_gpr_from_vsx<mode>): Likewise.
15460 (reload_gpr_from_vsxsf): Likewise.
15461 (p8_mfvsrd_4_disf): Likewise.
15462 (multi-word GPR splits): Do not split direct moves or quad memory
15465 2013-06-10 David Malcolm <dmalcolm@redhat.com>
15467 * tree-into-ssa.c (interesting_blocks): Make static.
15469 2013-06-10 Jakub Jelinek <jakub@redhat.com>
15472 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
15473 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
15474 Use DATA_ABI_ALIGNMENT for that case instead if defined.
15475 (get_variable_align): New function.
15476 (get_variable_section, emit_bss, emit_common,
15477 assemble_variable_contents, place_block_symbol): Use
15478 get_variable_align instead of DECL_ALIGN.
15479 (assemble_noswitch_variable): Add align argument, use it
15480 instead of DECL_ALIGN.
15481 (assemble_variable): Adjust caller. Use get_variable_align
15482 instead of DECL_ALIGN.
15483 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
15485 (DATA_ABI_ALIGNMENT): Define.
15486 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
15487 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
15488 opt is false, only return the psABI mandated alignment increase.
15489 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
15490 (DATA_ABI_ALIGNMENT): ... this.
15491 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
15492 (DATA_ABI_ALIGNMENT): ... this.
15493 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
15494 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
15495 (DATA_ABI_ALIGNMENT): ... this.
15496 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
15497 * doc/tm.texi: Regenerated.
15499 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
15501 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
15502 cmp_code to construct REG_EQUAL note.
15504 2013-06-09 Jakub Jelinek <jakub@redhat.com>
15507 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
15508 that operands[2] doesn't overlap with operands[0].
15510 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
15511 Jan Hubicka <jh@suse.cz>
15513 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
15514 hack to mark symbols as used.
15516 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
15518 PR rtl-optimization/57559
15519 * lra-constraints.c (process_alt_operands): Don't discourage
15520 memory with known offset for offsetable memory constraint.
15521 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
15523 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
15525 * varasm.c (struct oc_local_state): Reorder fields.
15526 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
15527 and adjust accordingly.
15528 (output_constructor): Reorder initialization code and adjust call to
15529 output_constructor_bitfield.
15531 2013-06-07 Jan Hubicka <jh@suse.cz>
15533 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
15535 2013-06-07 David Malcolm <dmalcolm@redhat.com>
15537 * tree-object-size.c (unknown): Make const.
15539 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15541 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
15542 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
15543 for last alternative in the cpu_facility attribute.
15545 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15548 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
15549 (xordi3): Change operand 2 constraint to arm_xordi_operand.
15550 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
15551 * config/arm/constraints.md (Dg): New constraint.
15552 * config/arm/neon.md (xordi3_neon): Remove.
15553 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
15554 * config/arm/predicates.md (arm_xordi_operand): New predicate.
15556 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15558 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
15559 Clean up alternatives.
15561 2013-06-07 Alan Modra <amodra@gmail.com>
15563 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
15566 2013-06-07 Alan Modra <amodra@gmail.com>
15568 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
15570 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15572 * config/arm/constraints.md (Df): New constraint.
15573 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
15574 Correct length attribute for last two alternatives.
15576 2013-06-07 Alan Modra <amodra@gmail.com>
15578 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
15579 override user -mfp-in-toc.
15580 (offsettable_ok_by_alignment): Consider just the current access
15581 rather than the whole object, unless BLKmode. Handle
15582 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
15583 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
15584 for -mcmodel=medium.
15585 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
15586 override user -mfp-in-toc or -msum-in-toc. Default to
15587 -mno-fp-in-toc for -mcmodel=medium.
15589 2013-06-06 DJ Delorie <dj@redhat.com>
15591 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
15592 TARGET_VALID_POINTER_MODE.
15594 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
15595 Pat Haugen <pthaugen@us.ibm.com>
15596 Peter Bergner <bergner@vnet.ibm.com>
15598 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
15599 Document new power8 builtins.
15601 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
15602 condition code register, to allow 128-bit logical operations to be
15603 done in the VSX or GPR registers.
15604 (nor<mode>3): Use the canonical form for nor.
15605 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
15606 vclz*, and vpopcnt* vector instructions.
15607 (nand<mode>3): Likewise.
15608 (orc<mode>3): Likewise.
15609 (clz<mode>2): LIkewise.
15610 (popcount<mode>2): Likewise.
15612 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
15613 that only the GPRs are recognized.
15615 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15616 support for new power8 builtins.
15618 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
15620 (xscvdpspn): Likewise.
15626 (vpopcnt): Likewise.
15627 (vpopcntb): Likewise.
15628 (vpopcnth): Likewise.
15629 (vpopcntw): Likewise.
15630 (vpopcntd): Likewise.
15632 (vmrgew): Likewise.
15633 (vmrgow): Likewise.
15635 (eqv_v16qi3): Likewise.
15636 (eqv_v8hi3): Likewise.
15637 (eqv_v4si3): Likewise.
15638 (eqv_v2di3): Likewise.
15639 (eqv_v4sf3): Likewise.
15640 (eqv_v2df3): Likewise.
15642 (nand_v16qi3): Likewise.
15643 (nand_v8hi3): Likewise.
15644 (nand_v4si3): Likewise.
15645 (nand_v2di3): Likewise.
15646 (nand_v4sf3): Likewise.
15647 (nand_v2df3): Likewise.
15649 (orc_v16qi3): Likewise.
15650 (orc_v8hi3): Likewise.
15651 (orc_v4si3): Likewise.
15652 (orc_v2di3): Likewise.
15653 (orc_v4sf3): Likewise.
15654 (orc_v2df3): Likewise.
15656 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
15657 allow power8 quad mode in 64-bit.
15658 (rs6000_builtin_vectorized_function): Add support to vectorize
15659 ISA 2.07 count leading zeros, population count builtins.
15660 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
15661 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
15662 (builtin_function_type): Add vgbbd builtin function which takes an
15664 (altivec_expand_vec_perm_const): Add support for new power8 merge
15667 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
15668 that does not include TImdoe for use with 32-bit.
15669 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
15671 (UNSPEC_VSX_CVDPSPN): Likewise.
15672 (vsx_xscvdpspn): Likewise.
15673 (vsx_xscvspdpn): Likewise.
15674 (vsx_xscvdpspn_scalar): Likewise.
15675 (vsx_xscvspdpn_directmove): Likewise.
15676 (vsx_and<mode>3): Split logical operations into 32-bit and
15677 64-bit. Add support to do logical operations on TImode as well as
15678 VSX vector types. Allow logical operations to be done in either
15679 VSX registers or in general purpose registers in 64-bit mode. Add
15680 splitters if GPRs were used. For AND, add clobber of CCmode to
15681 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
15683 (vsx_and<mode>3_32bit): Likewise.
15684 (vsx_and<mode>3_64bit): Likewise.
15685 (vsx_ior<mode>3): Likewise.
15686 (vsx_ior<mode>3_32bit): Likewise.
15687 (vsx_ior<mode>3_64bit): Likewise.
15688 (vsx_xor<mode>3): Likewise.
15689 (vsx_xor<mode>3_32bit): Likewise.
15690 (vsx_xor<mode>3_64bit): Likewise.
15691 (vsx_one_cmpl<mode>2): Likewise.
15692 (vsx_one_cmpl<mode>2_32bit): Likewise.
15693 (vsx_one_cmpl<mode>2_64bit): Likewise.
15694 (vsx_nor<mode>3): Likewise.
15695 (vsx_nor<mode>3_32bit): Likewise.
15696 (vsx_nor<mode>3_64bit): Likewise.
15697 (vsx_andc<mode>3): Likewise.
15698 (vsx_andc<mode>3_32bit): Likewise.
15699 (vsx_andc<mode>3_64bit): Likewise.
15700 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
15701 and xxlorc instructions.
15702 (vsx_eqv<mode>3_64bit): Likewise.
15703 (vsx_nand<mode>3_32bit): Likewise.
15704 (vsx_nand<mode>3_64bit): Likewise.
15705 (vsx_orc<mode>3_32bit): Likewise.
15706 (vsx_orc<mode>3_64bit): Likewise.
15708 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
15710 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
15712 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
15713 (p8_vmrgow): Likewise.
15714 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
15715 GPRs to be split under VSX.
15716 (p8v_clz<mode>2): Add power8 count leading zero support.
15717 (p8v_popcount<mode>2): Add power8 population count support.
15718 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
15721 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
15724 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
15726 (vec_nand): Likewise.
15727 (vec_vclz): Likewise.
15728 (vec_vclzb): Likewise.
15729 (vec_vclzd): Likewise.
15730 (vec_vclzh): Likewise.
15731 (vec_vclzw): Likewise.
15732 (vec_vgbbd): Likewise.
15733 (vec_vmrgew): Likewise.
15734 (vec_vmrgow): Likewise.
15735 (vec_vpopcnt): Likewise.
15736 (vec_vpopcntb): Likewise.
15737 (vec_vpopcntd): Likewise.
15738 (vec_vpopcnth): Likewise.
15739 (vec_vpopcntw): Likewise.
15741 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15743 PR rtl-optimization/57468
15744 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
15747 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15749 PR rtl-optimization/57459
15750 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
15751 type when setting live regs.
15753 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
15755 * config/s390/s390.opt (mlra): New option.
15756 * config/s390/s390.c (s390_decompose_address): Check displacement
15757 for all registers for LRA.
15758 (s390_secondary_reload): Don't used secondary reloads for LRA.
15759 (s390_lra_p): New function.
15760 (TARGET_LRA_P): Define.
15761 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
15762 of attribute cpu_facility to zarch for the last alternative.
15763 (*cmpmem_short): Ditto.
15765 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
15767 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
15768 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
15769 (arm_expand_prologue): Likewise.
15771 2013-06-06 Teresa Johnson <tejohnson@google.com>
15774 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
15775 as this is now done by redirect_edge_and_branch_force.
15776 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
15777 barriers, and fix interaction with splitting.
15778 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
15779 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
15780 reflect changes made in the routine.
15781 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
15782 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
15783 since this is called in cfglayout mode, and replace partition fixup
15784 with assert as that is now done by force_nonfallthru_and_redirect.
15785 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
15786 already be marked with region crossing note.
15787 (insert_section_boundary_note): Make non-static, gate on flag
15788 has_bb_partition, rewrite to also check for multiple partitions.
15789 (rest_of_handle_reorder_blocks): Remove call to
15790 insert_section_boundary_note, now done later during free_cfg.
15791 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
15792 * bb-reorder.h (insert_section_boundary_note): Declare.
15793 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
15794 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
15795 invoke insert_section_boundary_note.
15796 (try_redirect_by_replacing_jump): Remove unnecessary
15797 check for region crossing note.
15798 (fixup_partition_crossing): New function.
15799 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
15800 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
15801 in non-cfglayout mode.
15802 (force_nonfallthru_and_redirect): Fixup partition boundaries,
15803 remove old code that tried to do this. Emit barrier correctly
15804 when we are in cfglayout mode.
15805 (last_bb_in_partition): New function.
15806 (rtl_split_edge): Correctly fixup partition boundaries.
15807 (commit_one_edge_insertion): Remove old code that tried to
15808 fixup region crossing edge since this is now handled in
15809 split_block, and set up insertion point correctly since
15810 block may now end in a jump.
15811 (verify_hot_cold_block_grouping): Guard against checking when not in
15812 linearized RTL mode.
15813 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
15815 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
15816 rtl_verify_flow_info, so not called in cfglayout mode.
15817 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
15818 (fixup_reorder_chain): Remove old code that attempted to fixup region
15819 crossing note as this is now handled in force_nonfallthru_and_redirect.
15820 (duplicate_insn_chain): Don't duplicate switch section notes.
15821 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
15823 * basic-block.h (emit_barrier_after_bb): Declare.
15825 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15827 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
15828 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
15829 arm_usatsihi): Adjust alternatives for arm_restrict_it.
15831 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15833 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
15835 * config/arm/ldmstm.md: Regenerate.
15837 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15839 * config/arm/sync.md (atomic_loaddi_1):
15840 Disable predication for arm_restrict_it.
15841 (arm_load_exclusive<mode>): Likewise.
15842 (arm_load_exclusivesi): Likewise.
15843 (arm_load_exclusivedi): Likewise.
15844 (arm_load_acquire_exclusive<mode>): Likewise.
15845 (arm_load_acquire_exclusivesi): Likewise.
15846 (arm_load_acquire_exclusivedi): Likewise.
15847 (arm_store_exclusive<mode>): Likewise.
15848 (arm_store_exclusive<mode>): Likewise.
15849 (arm_store_release_exclusivedi): Likewise.
15850 (arm_store_release_exclusive<mode>): Likewise.
15852 2013-06-06 Richard Biener <rguenther@suse.de>
15854 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
15856 (lto_tag_is_tree_code_p): Adjust.
15857 (lto_tag_is_gimple_code_p): Likewise.
15858 (lto_gimple_code_to_tag): Likewise.
15859 (lto_tag_to_gimple_code): Likewise.
15860 (lto_tree_code_to_tag): Likewise.
15861 (lto_tag_to_tree_code): Likewise.
15862 * data-streamer.h (streamer_write_hwi_in_range): Use
15863 uhwi streaming to stream the normalized range.
15864 (streamer_read_hwi_in_range): Likewise.
15866 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15868 * config/arm/arm.md (enabled_for_depr_it): New attribute.
15869 (predicable_short_it): Likewise.
15870 (predicated): Likewise.
15871 (enabled): Handle above.
15872 (define_cond_exec): Set predicated attribute to yes.
15874 2013-06-05 Mike Stump <mikestump@comcast.net>
15876 * gdbinit.in (__FUNCTION__): Add.
15878 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
15880 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
15881 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
15883 2013-06-05 Jan Hubicka <jh@suse.cz>
15885 * varasm.c (mark_decl_referenced): Revert the removal until targets
15888 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
15890 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
15891 instead of mark_decl_referenced.
15893 2013-06-05 Jan Hubicka <jh@suse.cz>
15895 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
15896 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
15897 and symtab_used_from_object_file_p.
15898 (cgraph_make_node_local_1): Clear forced_by_abi.
15899 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
15900 * cgraph.h (symtab_node_base): Add forced_by_abi.
15901 (decide_is_variable_needed): Remove.
15902 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
15903 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
15904 (decide_is_symbol_needed): ... this one; handle symbols in general;
15905 always analyze virtuals; honnor forced_by_abi.
15906 (cgraph_finalize_function): Update.
15907 (varpool_finalize_decl): Update.
15908 (symbol_defined_and_needed): Remove.
15909 (analyze_functions): Update.
15910 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
15911 output_refs, input_overwrite_node): Handle forced_by_abi.
15912 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
15913 (address_taken_from_non_vtable_p): ... this one.
15914 (comdat_can_be_unshared_p_1): New function.
15915 (cgraph_comdat_can_be_unshared_p): Rename to ...
15916 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
15917 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
15918 (function_and_variable_visibility): Clear forced_by_abi as needed.
15919 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
15920 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
15921 * symtab.c (dump_symtab_base): Dump forced_by_abi.
15922 * varpool.c (decide_is_variable_needed): Remove.
15924 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15926 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
15927 (arm_option_override): Override arm_restrict_it where appropriate.
15928 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
15929 * config/arm/arm.opt (mrestrict-it): New command-line option.
15930 * doc/invoke.texi: Document -mrestrict-it.
15932 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15934 * tsan.c (tsan_atomic_table): Make const.
15936 2013-06-05 Richard Biener <rguenther@suse.de>
15938 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
15939 index associated with the tree we are supposed to replace.
15940 * tree-streamer-out.c (pack_ts_base_value_fields): Output
15941 TREE_ASM_WRITTEN as zero for everything but SSA names.
15943 2013-06-05 David Malcolm <dmalcolm@redhat.com>
15945 * tree-ssa-structalias.c (call_stmt_vars): Make static.
15947 2013-06-04 Jan Hubicka <jh@suse.cz>
15949 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
15950 (input_node, input_varpool_node): Handle correctly external same
15952 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
15953 nodes at ltrans stage.
15955 2013-06-04 Jan Hubicka <jh@suse.cz>
15957 * ipa-inline.c (update_caller_keys): Fix availability test.
15958 (update_callee_keys): Likewise.
15959 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
15960 to follow ELF standard.
15962 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
15964 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
15965 (mips64r5900el-*-elf*): New configurations.
15966 * config/mips/mips-cpus.def (r5900): New processor.
15967 * config/mips/mips-tables.opt: Regenerate.
15968 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
15969 (mips_issue_rate): Handle PROCESSOR_R5900.
15970 (mips_reorg_process_insns): Force reorder mode for the R5900.
15971 * config/mips/mips.h (TARGET_MIPS5900): Define.
15972 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
15974 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
15976 * config/mips/mips.md (processor): Add r5900.
15977 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
15979 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15981 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
15982 into function to generate MOVI instruction.
15983 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
15984 (aarch64_preferred_simd_mode): Turn into wrapper.
15985 (aarch64_output_scalar_simd_mov_immediate): New function.
15986 * config/aarch64/aarch64-protos.h: Add prototype for above.
15988 2013-06-04 Ian Bolton <ian.bolton@arm.com>
15990 * config/aarch64/aarch64.c (simd_immediate_info): Remove
15991 element_char member.
15992 (sizetochar): Return signed char.
15993 (aarch64_simd_valid_immediate): Remove elchar and other
15994 unnecessary variables.
15995 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
15996 Calculate element_char as required.
15997 * config/aarch64/aarch64-protos.h: Update and move prototype
15998 for aarch64_output_simd_mov_immediate.
15999 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
16002 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16004 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
16005 information completed by aarch64_simd_valid_immediate.
16006 (aarch64_legitimate_constant_p): Update arguments.
16007 (aarch64_simd_valid_immediate): Work with struct rather than many
16009 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
16010 (aarch64_simd_make_constant): Update arguments.
16011 (aarch64_output_simd_mov_immediate): Work with struct rather than
16012 many pointers. Output immediate directly rather than as operand.
16013 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
16015 * config/aarch64/constraints.md (Dn): Update arguments.
16017 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16019 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
16021 (aarch64_simd_immediate_valid_for_move): Remove.
16022 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
16023 (aarch64_simd_make_constant): Update call.
16024 (aarch64_output_simd_mov_immediate): Update call.
16025 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
16027 * config/aarch64/constraints.md (Dn): Update call.
16029 2013-06-04 Ian Bolton <ian.bolton@arm.com>
16031 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
16032 return type to bool for prototype.
16033 (aarch64_legitimate_constant_p): Check for true instead of not -1.
16034 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
16035 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
16037 2013-06-04 Catherine Moore <clm@codesourcery.com>
16039 * config/mips/mips.opt (meva): New.
16040 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
16041 (ASM_SPEC): Handle -meva.
16042 * doc/invoke.texi (meva): Document.
16044 2013-06-04 Alan Modra <amodra@gmail.com>
16046 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
16049 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16051 * rtl.def: Add extra fourth optional field to define_cond_exec.
16052 * gensupport.c (process_one_cond_exec): Process attributes from
16054 * doc/md.texi: Document fourth field in define_cond_exec.
16056 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
16058 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
16059 out the processing order as in store_bit_field_1.
16061 2013-06-04 Jan Hubicka <jh@suse.cz>
16063 PR middle-end/57500
16064 * cgraphunit.c (cgraph_process_same_body_aliases): Create
16065 non-VAR_DECL node if it does not exist yet.
16067 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
16069 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
16070 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
16071 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
16072 target_cpu_default setting.
16074 2013-06-03 Teresa Johnson <tejohnson@google.com>
16076 * dumpfile.c (opt_info_switch_p): Change -fopt-info
16077 default to -fopt-info=optimized instead of all.
16078 * doc/invoke.texi: Ditto.
16079 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
16080 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
16081 (execute_vect_slp): Emit BB vectorization success under
16082 MSG_OPTIMIZED_LOCATIONS.
16083 * tree-vect-slp.c (vect_slp_transform_bb): Change
16084 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
16085 * tree-vect-loop.c (vect_transform_loop): Ditto.
16087 2013-06-03 Jason Merrill <jason@redhat.com>
16090 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
16091 Use TARGET_EXPR for C++.
16093 2013-06-03 Jakub Jelinek <jakub@redhat.com>
16095 PR rtl-optimization/57268
16096 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
16097 if DEBUG_INSN_P (insn).
16100 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16102 PR rtl-optimization/57268
16103 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16104 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16106 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16108 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
16109 (ix86_avoid_lea_for_addr): Likewise.
16110 (exact_dependency_1): Likewise.
16111 (ix86_adjust_cost): Likewise.
16112 (swap_top_of_ready_list): Fix formatting and !reload_completed check
16114 (do_reorder_for_imul): Fix typo, formatting and
16115 !reload_completed check removed.
16116 (ix86_sched_reorder): Fix typo and formatting.
16117 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
16120 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
16122 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
16124 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
16126 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
16127 <STRING_CST>: Likewise.
16128 <VECTOR_CST>: Likewise.
16130 2013-06-01 Janus Weil <janus@gcc.gnu.org>
16131 Mikael Morin <mikael@gcc.gnu.org>
16133 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
16134 * config.in: Regenerated.
16135 * configure: Regenerated.
16137 2013-06-01 Jan Hubicka <jh@suse.cz>
16139 PR middle-end/57366
16140 * cgraphunit.c (compile): When weakref is not supported,
16141 set up transparent aliases before final output pass.
16142 * varasm.c (assemble_alias): Do not try to do it here.
16144 2013-06-01 Jan Hubicka <jh@suse.cz>
16146 PR middle-end/57467
16147 * passes.c (for_per_function): Skip unanalyzed functions.
16149 2013-06-01 Jan Hubicka <jh@suse.cz>
16151 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
16152 (lto_symtab_merge_symbols_1): ... this one.
16153 (lto_symtab_merge_cgraph_nodes): Rename to ...
16154 (lto_symtab_merge_symbols): ... this one; simplify.
16155 * cgraph.c (same_body_aliases_done): Rename to ...
16156 (cpp_implicit_aliases_done): ... this one.
16157 (cgraph_create_function_alias): Update.
16158 (cgraph_same_body_alias): Update.
16159 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
16160 (verify_edge_corresponds_to_fndecl): Simplify.
16161 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
16162 (cgraph_node): Remove same_body_alias.
16163 (varpool_node): Remove alias_of and extra_name_alias.
16164 (same_body_aliases_done): Rename to ..
16165 (cpp_implicit_aliases_done): ... this one.
16166 (symtab_alias_ultimate_target): Add default parameter.
16167 (symtab_resolve_alias): New function.
16168 (fixup_same_cpp_alias_visibility): Declare.
16169 (cgraph_function_node): Add default parameter.
16170 (cgraph_node_asm_name): Likewise.
16171 (cgraph_function_or_thunk_node): Add default parameter; do
16172 not ICE when it is NULL.
16173 (varpool_variable_node): Likewise.
16174 * tree-emutls.c (create_emultls_var): Update.
16175 (ipa_lower_emutls): Update.
16176 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
16177 (cgraph_reset_node): Reset alias info.
16178 (cgraph_finalize_function): Update.
16179 (fixup_same_cpp_alias_visibility): Move to symtab.c.
16180 (analyze_function): Simplify.
16181 (cgraph_process_same_body_aliases): Simplify.
16182 (analyze_functions): Fixup same body aliases.
16183 (handle_alias_pairs): Simplify.
16184 (assemble_thunk): Update.
16185 (assemble_thunks_and_aliases): Update.
16186 (output_weakrefs): Rewrite.
16187 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
16188 (lto_output_varpool_node): Likewise.
16189 (compute_ltrans_boundary): Remve assert.
16190 (get_alias_symbol): New functoin.
16191 (input_node): Rewrite alias handling.
16192 (input_varpool_node): Likewise.
16193 * ipa-pure-const.c (propagate_pure_const): Fix formating.
16194 * ipa.c (process_references): Handle weakrefs correctly.
16195 (symtab_remove_unreachable_nodes): Likewise.
16196 * trans-mem.c (get_cg_data): Update.
16197 (ipa_tm_create_version_alias): Update.
16198 (ipa_tm_execute): Update.
16199 * symtab.c (dump_symtab_base): Dump aliases.
16200 (verify_symtab_base): Verify aliases.
16201 (symtab_node_availability): New function.
16202 (symtab_alias_ultimate_target): Simplify.
16203 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
16204 handle all the fixup cases.
16205 (symtab_resolve_alias): New function.
16206 * passes.c (ipa_write_summaries): Handle weakrefs.
16207 * varpool.c (varpool_analyze_node): Simplify.
16208 (assemble_aliases): Update.
16209 (varpool_create_variable_alias): Simplify.
16210 (varpool_extra_name_alias): Simplify.
16211 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
16212 (lto_symtab_merge_symbols): ... this one.
16214 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
16217 PR rtl-optimization/57268
16218 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16219 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16221 2013-06-01 Tobias Burnus <burnus@net-b.de>
16223 Partially reverted:
16224 2013-05-31 Tobias Burnus <burnus@net-b.de>
16226 PR middle-end/57073
16227 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
16230 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
16232 PR rtl-optimization/57268
16233 * sched-deps.c (sched_analyze_2): Flush dependence lists if
16234 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
16236 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16238 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
16239 unordered comparison operators when -fno-trapping-math is in effect
16241 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
16242 and implement unordered comparison operators properly on the e500.
16244 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
16246 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
16247 for constant scalar integers.
16248 (simplify_relational_operation_1): Likewise.
16250 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
16252 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
16253 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
16256 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16257 Igor Zamyatin <igor.zamyatin@intel.com>
16259 Silvermont (SLM) architecture performance tuning.
16260 * config/i386/i386.h (enum ix86_tune_indices): Add
16261 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
16262 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
16264 * config/i386/i386.c (initial_ix86_tune_features)
16265 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
16266 (ix86_lea_outperforms): Handle Silvermont tuning.
16267 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
16269 (ix86_use_lea_for_mov): Likewise.
16270 (ix86_avoid_lea_for_addr): Likewise.
16271 (ix86_lea_for_add_ok): Likewise.
16272 (exact_dependency_1): New function.
16273 (exact_store_load_dependency): Likewise.
16274 (ix86_adjust_cost): Handle Silvermont tuning.
16275 (do_reoder_for_imul): Likewise.
16276 (swap_top_of_ready_list): New function.
16277 (ix86_sched_reorder): Changed to handle Silvermont tuning.
16279 * config/i386/i386.md (peepholes that split memory operand in fp
16282 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16284 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
16285 Remove un-necessary braces.
16287 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
16289 * config/aarch64/aarch64.c (aarch64_classify_symbol):
16290 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
16292 2013-05-31 Tobias Burnus <burnus@net-b.de>
16294 PR middle-end/57073
16295 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
16297 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16300 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
16301 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
16302 * config/arm/neon.md (iordi3_neon): Remove.
16303 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
16304 * config/arm/predicates.md (imm_for_neon_logic_operand):
16305 Move to earlier in the file.
16306 (neon_logic_op2): Likewise.
16307 (arm_iordi_operand_neon): New predicate.
16309 2013-05-31 Richard Biener <rguenther@suse.de>
16311 PR tree-optimization/57478
16312 PR tree-optimization/57453
16313 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
16316 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16318 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
16319 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
16321 2013-05-30 Tobias Burnus <burnus@net-b.de>
16322 Thomas Koenig <tkoenig@gcc.gnu.org>
16324 PR middle-end/57073
16325 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
16326 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
16328 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
16330 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
16332 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
16334 * target.def (register_usage_leveling_p): New hook.
16335 * targhooks.c (default_register_usage_leveling_p): New.
16336 * targhooks.h (default_register_usage_leveling_p): New prototype.
16337 * lra-assigns.c (register_usage_leveling_p): Use the hook.
16338 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
16339 * doc/tm.texi: Update.
16340 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
16342 2013-05-30 Ian Bolton <ian.bolton@arm.com>
16344 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
16345 (*insv_reg<mode>): New define_insn.
16347 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
16349 PR rtl-optimization/57439
16350 * postreload.c (move2add_valid_value_p): Check that we have
16351 a zero subreg_regno_offset when accessing the register in
16352 the requested mode.
16354 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
16355 Igor Zamyatin <igor.zamyatin@intel.com>
16357 Silvermont (SLM) architecture pipeline model, tuning and
16359 * config.gcc: Add slm config options and target.
16361 * config/i386/slm.md: New.
16363 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
16365 * config/i386/i386-c.c (ix86_target_macros_internal): New case
16367 (ix86_target_macros_internal): Likewise.
16369 * config/i386/i386.c (slm_cost): New cost.
16370 (m_SLM): New macro flag.
16371 (initial_ix86_tune_features): Set m_SLM.
16372 (x86_accumulate_outgoing_args): Likewise.
16373 (x86_arch_always_fancy_math_387): Likewise.
16374 (processor_target_table): Add slm cost.
16375 (cpu_names): Add slm cpu name.
16376 (x86_option_override_internal): Set SLM ISA.
16377 (ix86_issue_rate): New case PROCESSOR_SLM.
16378 (ia32_multipass_dfa_lookahead): Likewise.
16379 (fold_builtin_cpu): Add slm.
16381 * config/i386/i386.h (TARGET_SLM): New target macro.
16382 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
16383 (processor_type): Add PROCESSOR_SLM.
16385 * config/i386/i386.md (cpu): Add new value "slm".
16386 (slm.md): Include slm.md.
16388 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
16389 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16391 * config/arm/arm-protos.h: Add and update function protos.
16392 * config/arm/arm.c (use_simple_return_p): New added.
16393 (thumb2_expand_return): Check simple_return flag.
16394 * config/arm/arm.md: Add simple_return and conditional simple_return.
16395 * config/arm/iterators.md: Add iterator for return and simple_return.
16397 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
16399 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
16400 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
16401 (arm_emit_vfp_multi_reg_pop): Likewise.
16402 (thumb2_emit_ldrd_pop): Likewise.
16403 (arm_expand_epilogue): Add misc REG_CFA notes.
16404 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
16406 2013-05-29 Lawrence Crowl <crowl@google.com>
16408 * config/arm/t-arm: Update for below.
16410 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
16411 Change type to hash_table. Update dependent calls and types.
16413 * config/i386/t-cygming: Update for below.
16415 * config/i386/t-interix: Update for below.
16417 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
16418 Change type to hash_table. Update dependent calls and types.
16419 (i386_find_on_wrapper_list::wrappers): Likewise.
16421 * config/ia64/t-ia64: Update for below.
16423 * config/ia64/ia64.c (bundle_state_table):
16424 Change type to hash_table. Update dependent calls and types.
16426 * config/mips/mips.c (mips_reorg_process_insns::htab):
16427 Change type to hash_table. Update dependent calls and types.
16429 * config/sol2.c (solaris_comdat_htab):
16430 Change type to hash_table. Update dependent calls and types.
16432 * config/t-sol2: Update for above.
16434 2013-05-29 Teresa Johnson <tejohnson@google.com>
16436 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
16437 functions are not yet marked as defined.
16439 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
16440 Pat Haugen <pthaugen@us.ibm.com>
16441 Peter Bergner <bergner@vnet.ibm.com>
16443 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
16447 (vrotl<mode>3): Likewise.
16448 (vashl<mode>3): Likewise.
16449 (vlshr<mode>3): Likewise.
16450 (vashr<mode>3): Likewise.
16452 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
16453 support for power8 V2DI builtins.
16455 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
16456 power8 V2DI builtins.
16457 (vupkhsw): Likewise.
16458 (vupklsw): Likewise.
16459 (vaddudm): Likewise.
16460 (vminsd): Likewise.
16461 (vmaxsd): Likewise.
16462 (vminud): Likewise.
16463 (vmaxud): Likewise.
16464 (vpkudum): Likewise.
16465 (vpksdss): Likewise.
16466 (vpkudus): Likewise.
16467 (vpksdus): Likewise.
16472 (vsubudm): Likewise.
16473 (vcmpequd): Likewise.
16474 (vcmpgtsd): Likewise.
16475 (vcmpgtud): Likewise.
16476 (vcmpequd_p): Likewise.
16477 (vcmpgtsd_p): Likewise.
16478 (vcmpgtud_p): Likewise.
16479 (vupkhsw): Likewise.
16480 (vupklsw): Likewise.
16481 (vaddudm): Likewise.
16482 (vmaxsd): Likewise.
16483 (vmaxud): Likewise.
16484 (vminsd): Likewise.
16485 (vminud): Likewise.
16486 (vpksdss): Likewise.
16487 (vpksdus): Likewise.
16488 (vpkudum): Likewise.
16489 (vpkudus): Likewise.
16494 (vsubudm): Likewise.
16496 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
16497 support for power8 V2DI instructions.
16499 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
16500 power8 V2DI instructions. Combine pack and unpack insns to use an
16501 iterator for each mode. Check whether a particular mode supports
16502 Altivec instructions instead of just checking TARGET_ALTIVEC.
16503 (UNSPEC_VPKUWUM): Likewise.
16504 (UNSPEC_VPKSHSS): Likewise.
16505 (UNSPEC_VPKSWSS): Likewise.
16506 (UNSPEC_VPKUHUS): Likewise.
16507 (UNSPEC_VPKSHUS): Likewise.
16508 (UNSPEC_VPKUWUS): Likewise.
16509 (UNSPEC_VPKSWUS): Likewise.
16510 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
16511 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
16512 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
16513 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
16514 (UNSPEC_VUPKHSB): Likewise.
16515 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
16516 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
16517 (UNSPEC_VUPKHSH): Likewise.
16518 (UNSPEC_VUPKLSB): Likewise.
16519 (UNSPEC_VUPKLSH): Likewise.
16521 (VI_char): Likewise.
16522 (VI_scalar): Likewise.
16523 (VI_unit): Likewise.
16525 (VP_small): Likewise.
16526 (VP_small_lc): Likewise.
16527 (VU_char): Likewise.
16528 (add<mode>3): Likewise.
16529 (altivec_vaddcuw): Likewise.
16530 (altivec_vaddu<VI_char>s): Likewise.
16531 (altivec_vadds<VI_char>s): Likewise.
16532 (sub<mode>3): Likewise.
16533 (altivec_vsubcuw): Likewise.
16534 (altivec_vsubu<VI_char>s): Likewise.
16535 (altivec_vsubs<VI_char>s): Likewise.
16536 (altivec_vavgs<VI_char>): Likewise.
16537 (altivec_vcmpbfp): Likewise.
16538 (altivec_eq<mode>): Likewise.
16539 (altivec_gt<mode>): Likewise.
16540 (altivec_gtu<mode>): Likewise.
16541 (umax<mode>3): Likewise.
16542 (smax<mode>3): Likewise.
16543 (umin<mode>3): Likewise.
16544 (smin<mode>3): Likewise.
16545 (altivec_vpkuhum): Likewise.
16546 (altivec_vpkuwum): Likewise.
16547 (altivec_vpkshss): Likewise.
16548 (altivec_vpkswss): Likewise.
16549 (altivec_vpkuhus): Likewise.
16550 (altivec_vpkshus): Likewise.
16551 (altivec_vpkuwus): Likewise.
16552 (altivec_vpkswus): Likewise.
16553 (altivec_vpks<VI_char>ss): Likewise.
16554 (altivec_vpks<VI_char>us): Likewise.
16555 (altivec_vpku<VI_char>us): Likewise.
16556 (altivec_vpku<VI_char>um): Likewise.
16557 (altivec_vrl<VI_char>): Likewise.
16558 (altivec_vsl<VI_char>): Likewise.
16559 (altivec_vsr<VI_char>): Likewise.
16560 (altivec_vsra<VI_char>): Likewise.
16561 (altivec_vsldoi_<mode>): Likewise.
16562 (altivec_vupkhsb): Likewise.
16563 (altivec_vupkhs<VU_char>): Likewise.
16564 (altivec_vupkls<VU_char>): Likewise.
16565 (altivec_vupkhsh): Likewise.
16566 (altivec_vupklsb): Likewise.
16567 (altivec_vupklsh): Likewise.
16568 (altivec_vcmpequ<VI_char>_p): Likewise.
16569 (altivec_vcmpgts<VI_char>_p): Likewise.
16570 (altivec_vcmpgtu<VI_char>_p): Likewise.
16571 (abs<mode>2): Likewise.
16572 (vec_unpacks_hi_v16qi): Likewise.
16573 (vec_unpacks_hi_v8hi): Likewise.
16574 (vec_unpacks_lo_v16qi): Likewise.
16575 (vec_unpacks_hi_<VP_small_lc>): Likewise.
16576 (vec_unpacks_lo_v8hi): Likewise.
16577 (vec_unpacks_lo_<VP_small_lc>): Likewise.
16578 (vec_pack_trunc_v8h): Likewise.
16579 (vec_pack_trunc_v4si): Likewise.
16580 (vec_pack_trunc_<mode>): Likewise.
16582 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
16584 (vec_vmaxsd): Likewise.
16585 (vec_vmaxud): Likewise.
16586 (vec_vminsd): Likewise.
16587 (vec_vminud): Likewise.
16588 (vec_vpksdss): Likewise.
16589 (vec_vpksdus): Likewise.
16590 (vec_vpkudum): Likewise.
16591 (vec_vpkudus): Likewise.
16592 (vec_vrld): Likewise.
16593 (vec_vsld): Likewise.
16594 (vec_vsrad): Likewise.
16595 (vec_vsrd): Likewise.
16596 (vec_vsubudm): Likewise.
16597 (vec_vupkhsw): Likewise.
16598 (vec_vupklsw): Likewise.
16600 2013-05-29 Jan Hubicka <jh@suse.cz>
16602 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
16603 flags; reorder rest of fields in more consistent way.
16604 (varpool_node): Remove analyzed, finalized and alias.
16605 (cgraph_ndoe): Likewise.
16606 (symtab_alias_ultimate_target): New function.
16607 (cgraph_function_node): Move offline.
16608 (cgraph_reset_node): Declare.
16609 (cgraph_comdat_can_be_unshared_p): Remove.
16610 (varpool_remove_initializer): Declare.
16611 (varpool_first_defined_variable, varpool_next_defined_variable
16612 cgraph_first_defined_function, cgraph_next_defined_function): Update.
16613 (cgraph_function_with_gimple_body_p): Update.
16614 (varpool_all_refs_explicit_p): Update.
16615 (symtab_alias_target): New function.
16616 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
16617 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
16618 (cgraph_function_or_thunk_node): Simplify using
16619 symtab_alias_ultimate_target.
16620 (varpool_variable_node): Likewise.
16621 * cgraph.c (cgraph_create_function_alias): Update.
16622 (cgraph_add_thunk): Update.
16623 (cgraph_remove_node): Update.
16624 (dump_cgraph_node): Do not dump removed flags.
16625 (cgraph_function_body_availability): Update.
16626 (cgraph_propagate_frequency): Update.
16627 (verify_cgraph_node): Check sanity of local flag.
16628 (cgraph_function_node): Move here from cgraph.h; revamp for
16629 cgraph_function_or_thunk_node.
16630 * lto-symtab.c (lto_varpool_replace_node): Update.
16631 (lto_symtab_resolve_can_prevail_p): Update.
16632 (lto_symtab_merge_cgraph_nodes): Update.
16633 * ipa-cp.c (determine_versionability, initialize_node_lattices,
16634 propagate_constants_accross_call, devirtualization_time_bonus,
16635 ipcp_propagate_stage): Update.
16636 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
16637 * ipa-inline-transform.c (clone_inlined_nodes,
16638 preserve_function_body_p): Update.
16639 * ipa-reference.c (propagate): Update.
16640 (write_node_summary_p): Update.
16641 * toplev.c (wrapup_global_declaration_2): Update.
16642 * cgraphunit.c (cgraph_analyze_function): Rename to ...
16643 (analyze_function) ... this one.
16644 (cgraph_process_new_functions): Update.
16645 (cgraph_reset_node): Export.
16646 (cgraph_finalize_function): Update.
16647 (cgraph_add_new_function): Update.
16648 (process_function_and_variable_attributes): Update.
16649 (varpool_finalize_decl): Update.
16650 (symbol_finalized): Remove.
16651 (symbol_finalized_and_needed): Rename to ...
16652 (symbol_defined_and_needed): ... update.
16653 (cgraph_analyze_functions): Update.
16654 (handle_alias_pairs): Update.
16655 (mark_functions_to_output): Update.
16656 (assemble_thunk): Update.
16657 (output_in_order): Update.
16658 (output_weakrefs): Update.
16659 (finalize_compilation_unit): Update.
16660 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
16661 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
16662 input_node, input_varpool_node): Update.
16663 * dbxout.c (dbxout_expand_expr): Update.
16664 * cgraphclones.c (cgraph_clone_node): Update.
16665 (cgraph_copy_node_for_versioning): Update.
16666 (cgraph_materialize_clone): Update.
16667 (cgraph_materialize_all_clones): Update.
16668 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
16669 propagate_pure_const, propagate_nothrow): Update.
16670 * lto-streamer-out.c (lto_output, write_symbol): Update.
16671 * ipa-utils.c (ipa_reverse_postorder): Update.
16672 * ipa-inline.c (can_inline_edge_p): Update.
16673 (update_caller_keys, ipa_inline): Update.
16674 * dwarf2out.c (reference_to_unused,
16675 premark_types_used_by_global_vars_helper): Update.
16676 * tree-eh.c (tree_could_trap_p): Update.
16677 * ipa-split.c (consider_split, execute_split_functions): Update.
16678 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
16679 has_addr_references_p): Update; move ahead in file for better
16681 (process_references): Simplify.
16682 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
16683 bodies are removed.
16684 (cgraph_comdat_can_be_unshared_p): Make static.
16685 (cgraph_externally_visible_p): Update.
16686 (varpool_externally_visible_p): Update.
16687 (function_and_variable_visibility): Update.
16688 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
16689 ipa_tm_mark_force_output_node): Update.
16690 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
16691 estimate_edge_devirt_benefit, inline_generate_summary,
16692 inline_write_summary): Update.
16693 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
16694 * ipa-prop.c (ipa_compute_jump_functions): Update.
16695 (ipa_print_node_params, ipa_prop_read_section,
16696 ipa_update_after_lto_read, read_replacements_section): Update.
16697 * varasm.c (mark_decl_referenced): Update.
16698 (assemble_alias, dump_tm_clone_pairs): Update.
16699 * tree-inline.c (copy_bb): Update.
16700 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
16702 * symtab.c (dump_symtab_base): Print new flags.
16703 (verify_symtab_base): Verify new flags.
16704 (symtab_alias_ultimate_target): New function.
16705 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
16706 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
16708 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
16710 * i386.c (ix86_get_function_versions_dispatcher,
16711 ix86_generate_version_dispatcher_body): Update.
16712 (fold_builtin_cpu): Use varpool_add_new_variable.
16713 * varpool.c (varpool_remove_initializer): Break out from ...
16714 (varpool_remove_node): ... this one.
16715 (dump_varpool_node, varpool_node_for_asm,
16716 cgraph_variable_initializer_availability, varpool_analyze_node,
16717 varpool_assemble_decl, varpool_remove_unreferenced_decls,
16718 varpool_finalize_named_section_flags, varpool_create_variable_alias):
16721 2013-05-29 Jan Hubicka <jh@suse.cz>
16723 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
16725 2013-05-29 Easwaran Raman <eraman@google.com>
16727 PR tree-optimization/57442
16728 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
16729 when control exits the main loop.
16731 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
16733 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
16735 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
16736 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
16737 * rx/t-rx: Add rx100 under multi library matches option for nofpu
16740 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16742 PR tree-optimization/57441
16743 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
16744 Don't limit size of incr_vec to number of candidates.
16746 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
16748 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
16749 and mips16 directories.
16750 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
16751 (MULTILIB_DIRNAMES): Ditto.
16752 (MULTILIB_EXCEPTIONS): Add new exceptions.
16753 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
16754 (MULTILIB_DIRNAMES): Ditto.
16755 (MULTILIB_EXCEPTIONS): Add new exceptions.
16757 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16758 Marcus Shawcroft <marcus.shawcroft@arm.com>
16760 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
16761 SYMBOL_TINY_ABSOLUTE.
16762 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
16763 SYMBOL_TINY_ABSOLUTE.
16764 (aarch64_expand_mov_immediate): Likewise.
16765 (aarch64_classify_symbol): Likewise.
16766 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
16767 Permit SYMBOL_TINY_ABSOLUTE.
16768 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
16770 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
16771 Marcus Shawcroft <marcus.shawcroft@arm.com>
16773 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
16774 Refactor if/switch. Replace gcc_assert with if.
16776 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16778 * config/i386/i386.c (initial_ix86_tune_features): Enable
16779 FP Reassociation for AMD bdver1 and bdver2.
16781 2013-05-29 Martin Jambor <mjambor@suse.cz>
16783 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
16784 and IMAGPART_EXPR do not occur within other handled_components.
16786 2013-05-29 Richard Biener <rguenther@suse.de>
16788 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
16789 access on whether the use is in the BB we currently try to
16791 (vect_bb_vectorization_profitable_p): Pass the BB we currently
16792 vectorize to vect_bb_slp_scalar_cost.
16794 2013-05-29 Richard Biener <rguenther@suse.de>
16796 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
16797 computing scalar cost offsetted by stmts that are kept live
16799 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
16800 for computation of scalar cost.
16802 2013-05-28 Steve Ellcey <sellcey@mips.com>
16804 * config/mips/mips-cpus.def (mips32r2): Change processor type.
16806 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
16808 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
16809 array notation built-in reduction functions.
16810 * doc/passes.texi (Passes): Added documentation about changes done
16812 * doc/invoke.texi (C Dialect Options): Added documentation about
16813 the -fcilkplus flag.
16814 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
16815 (BUILTINS_DEF): Depend on cilkplus.def.
16816 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
16817 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
16818 * cilkplus.def: New file.
16820 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
16822 PR rtl-optimization/57439
16823 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
16825 2013-05-28 Easwaran Raman <eraman@google.com>
16827 PR tree-optimization/57337
16828 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
16829 (find_insert_point): Correctly identify the insertion point
16830 when two statements with the same UID is compared.
16832 2013-05-28 Richard Biener <rguenther@suse.de>
16834 PR tree-optimization/56787
16835 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
16836 from the list of data references.
16837 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
16839 (vect_analyze_loop_operations): Likewise.
16840 (vect_transform_loop): Remove clobbers.
16842 2013-05-28 Martin Jambor <mjambor@suse.cz>
16844 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
16845 and REALPART_EXPRs have scalar type.
16847 2013-05-28 Richard Biener <rguenther@suse.de>
16849 PR tree-optimization/57411
16850 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
16852 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
16853 virtual operand propagation.
16855 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16857 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
16858 destination register for bmasksi_vis.
16859 (vector_init_bshuffle): Likewise.
16860 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
16862 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
16864 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
16865 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
16866 mode if the instruction isn't available in the original mode.
16867 * config/sparc/sparc.opt (mfix-ut699): New option.
16868 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
16869 (divdf3): Turn into expander.
16870 (divdf3_nofix): New insn.
16871 (divdf3_fix): Likewise.
16872 (divsf3): Disable if -mfix-ut699.
16873 (sqrtdf2): Turn into expander.
16874 (sqrtdf2_nofix): New insn.
16875 (sqrtdf2_fix): Likewise.
16876 (sqrtsf2): Disable if -mfix-ut699.
16878 2013-05-27 Richard Biener <rguenther@suse.de>
16880 PR middle-end/57412
16881 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
16882 block for the new loop.
16884 2013-05-27 Richard Biener <rguenther@suse.de>
16886 PR tree-optimization/57343
16887 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
16888 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
16889 (number_of_iterations_cond): Do not build the folded tree.
16891 2013-05-27 Richard Biener <rguenther@suse.de>
16894 PR middle-end/57381
16895 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
16896 OEP_CONSTANT_ADDRESS_OF retained.
16898 PR tree-optimization/57417
16899 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
16900 for unchanged base.
16901 (set_ssa_val_to): Compare addresses using
16902 get_addr_base_and_unit_offset.
16904 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
16906 PR rtl-optimization/56833
16907 * postreload.c (move2add_record_mode): New function.
16908 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
16909 (move2add_use_add2_insn): Use move2add_record_sym_value.
16910 (move2add_use_add3_insn): Likewise.
16911 (reload_cse_move2add): Use move2add_valid_value_p and
16912 move2add_record_mode. Invalidate call-clobbered and REG_INC
16913 affected regs by setting reg_mode to VOIDmode.
16914 (move2add_note_store): Don't pretend the inside of a SUBREG is
16915 the actual destination. Invalidate single/leading registers by
16916 setting reg_mode to VOIDmode.
16917 Use move2add_record_sym_value, move2add_valid_value_p and
16918 move2add_record_mode.
16920 2013-05-27 Richard Biener <rguenther@suse.de>
16922 PR tree-optimization/57396
16923 * tree-affine.c (double_int_constant_multiple_p): Properly
16924 return false for val == 0 and div != 0.
16926 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16928 * config/mips/mips.h: Use #elif in preprocessor conditions.
16930 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16933 * config/mips/constraints.md (kl): New constraint.
16934 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
16935 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
16936 constraint for operand 0. Split after CSE for MIPS16. Emit a move
16937 from LO for MIPS16.
16938 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
16940 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
16943 * config/mips/mips.c (mips_can_inline_p): New function.
16944 (TARGET_CAN_INLINE_P): Define.
16946 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
16948 * sched-int.h (ds_t, dw_t): Make unsigned int.
16949 Fix documentation that describes how all the ds_t bits are used.
16950 Reserve the last bit for delayed-branch scheduling.
16951 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
16952 (BITS_PER_DEP_WEAK): Fix definition and documentation.
16953 (gen_dep_weak_1): Remove prototype.
16954 * sched-deps.c (get_dep_weak_1): Make static.
16955 * target.def (speculate_insn, needs_block_p, gen_spec_check,
16956 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
16957 * doc/tm.texi: Regenerate.
16958 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
16960 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
16963 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
16965 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
16966 Sandra Loosemore <sandra@codesourcery.com>
16968 * config.gcc (powerpc-*): Allow native for with-cpu.
16970 2013-05-24 Jeff Law <law@redhat.com>
16972 PR tree-optimization/57124
16973 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
16974 conversion feeding a condition if the range has an overflow
16975 if -fstrict-overflow. Add warnings for when we do make the
16978 2013-05-24 Dehao Chen <dehao@google.com>
16980 * tree-cfg.c (locus_discrim_map): Fix the typo.
16981 (locus_discrim_hasher): Likewise.
16982 (locus_discrim_hasher::hash): Likewise.
16983 (locus_discrim_hasher::equal): Likewise.
16985 2013-05-24 Martin Jambor <mjambor@suse.cz>
16987 PR tree-optimization/57294
16988 * cgraph.h (ipa_record_stmt_references): Declare.
16989 * cgraphbuild.c (ipa_record_stmt_references): New function.
16990 (build_cgraph_edges): Use ipa_record_stmt_references.
16991 (rebuild_cgraph_edges): Likewise.
16992 (cgraph_rebuild_references): Likewise.
16993 * ipa-prop.c (ipa_modify_call_arguments): Discard references
16994 associated with the old statement and build references from the
16995 newly built statements.
16996 * ipa-ref.c (ipa_remove_stmt_references): New function.
16997 * ipa-ref.h (ipa_remove_stmt_references): Declare.
16999 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
17001 * lra-constraints.c (emit_spill_move): Use smaller mode for
17003 (check_and_process_move): Consider mem-reg moves for secondary
17005 (curr_insn_transform): Don't lose insns emitted before for
17006 secondary memory moves.
17007 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
17008 reg set up in the current insn.
17010 2013-05-24 Dehao Chen <dehao@google.com>
17012 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
17014 (locus_descrim_hasher::equal): Likewise.
17015 (build_gimple_cfg): New discriminator assignment algorithm.
17016 (make_edges): Likewise.
17017 (next_discriminator_for_locus): Likewise.
17018 (same_line_p): Likewise.
17019 (assign_discriminators): Likewise.
17020 (make_cond_expr_edges): Likewise.
17021 (make_gimple_switch_edges): Likewise.
17022 (make_goto_expr_edges): Likewise.
17023 (make_gimple_asm_edges): Likewise.
17025 2013-05-24 Ian Bolton <ian.bolton@arm.com>
17027 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
17028 X format specifier to only display bottom 16 bits.
17029 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
17030 immediate to match for operand 2, since it will be masked.
17032 2013-05-24 Richard Biener <rguenther@suse.de>
17034 PR tree-optimization/57287
17035 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
17036 all SSA names that occur in abnormal PHIs.
17038 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
17041 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
17042 that index is not negative.
17044 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
17046 PR rtl-optimization/55177
17047 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
17048 (simplify_byte_swapping_operation): New.
17049 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
17050 (simplify_relational_operation_1): Deal with BSWAP.
17052 2013-05-23 Richard Henderson <rth@redhat.com>
17055 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
17056 (ix86_reorg): Call it.
17058 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
17061 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
17062 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
17063 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
17065 2013-05-23 Christian Bruel <christian.bruel@st.com>
17068 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
17070 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17071 Marcus Shawcroft <marcus.shawcroft@arm.com>
17073 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
17074 * config/aarch64/constraints.md (Usa): Remove.
17075 * doc/md.texi (AArch64 Usa): Remove.
17077 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17078 Marcus Shawcroft <marcus.shawcroft@arm.com>
17080 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
17081 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
17082 * config/aarch64/predicates.md (aarch64_const_address): Remove.
17083 (aarch64_mov_operand): Use aarch64_mov_operand_p.
17085 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
17087 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
17088 instruction (AdvSIMD).
17089 * config/aarch64/aarch64-builtins.c
17090 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
17091 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
17093 2013-05-23 Martin Jambor <mjambor@suse.cz>
17095 PR middle-end/57347
17096 * tree.h (contains_bitfld_component_ref_p): Declare.
17097 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
17098 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
17100 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
17101 not access a bit-field. Assert all final offsets are byte-aligned.
17103 2013-05-23 Richard Biener <rguenther@suse.de>
17105 PR tree-optimization/57380
17106 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
17107 least one invariant or re-used load.
17108 * passes.c (init_optimization_passes): Move pass_phiprop before
17111 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
17113 * config/aarch64/aarch64-simd.md
17114 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
17116 2013-05-23 Richard Biener <rguenther@suse.de>
17118 PR middle-end/57381
17119 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
17120 OEP_CONSTANT_ADDRESS_OF retained.
17122 2013-05-23 Jakub Jelinek <jakub@redhat.com>
17124 PR middle-end/57344
17125 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
17126 don't lower unit. Handle unit not being always BITS_PER_WORD.
17128 2013-05-23 Richard Biener <rguenther@suse.de>
17130 PR rtl-optimization/57341
17131 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
17132 instead of true_dependence.
17134 2013-05-22 David Malcolm <dmalcolm@redhat.com>
17136 * bb-reorder.c (branch_threshold): Make const.
17137 (exec_threshold): Ditto.
17139 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17140 Pat Haugen <pthaugen@us.ibm.com>
17141 Peter Bergner <bergner@vnet.ibm.com>
17143 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
17144 documentation for the power8 crypto builtins.
17146 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
17148 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
17149 macros for defining power8 builtin functions.
17150 (BU_P8V_AV_2): Likewise.
17151 (BU_P8V_AV_P): Likewise.
17152 (BU_P8V_VSX_1): Likewise.
17153 (BU_P8V_OVERLOAD_1): Likewise.
17154 (BU_P8V_OVERLOAD_2): Likewise.
17155 (BU_CRYPTO_1): Likewise.
17156 (BU_CRYPTO_2): Likewise.
17157 (BU_CRYPTO_3): Likewise.
17158 (BU_CRYPTO_OVERLOAD_1): Likewise.
17159 (BU_CRYPTO_OVERLOAD_2): Likewise.
17160 (XSCVSPDP): Fix typo, point to the correct instruction.
17161 (VCIPHER): Add power8 crypto builtins.
17162 (VCIPHERLAST): Likewise.
17163 (VNCIPHER): Likewise.
17164 (VNCIPHERLAST): Likewise.
17165 (VPMSUMB): Likewise.
17166 (VPMSUMH): Likewise.
17167 (VPMSUMW): Likewise.
17168 (VPERMXOR_V2DI): Likewise.
17169 (VPERMXOR_V4SI: Likewise.
17170 (VPERMXOR_V8HI: Likewise.
17171 (VPERMXOR_V16QI: Likewise.
17172 (VSHASIGMAW): Likewise.
17173 (VSHASIGMAD): Likewise.
17174 (VPMSUM): Likewise.
17175 (VPERMXOR): Likewise.
17176 (VSHASIGMA): Likewise.
17178 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
17179 __CRYPTO__ if the crypto instructions are available.
17180 (altivec_overloaded_builtins): Add support for overloaded power8
17183 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
17184 support for power8 crypto builtins.
17185 (builtin_function_type): Likewise.
17186 (altivec_init_builtins): Add support for builtins that take vector
17187 long long (V2DI) arguments.
17189 * config/rs6000/crypto.md: New file, define power8 crypto
17192 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
17193 Pat Haugen <pthaugen@us.ibm.com>
17194 Peter Bergner <bergner@vnet.ibm.com>
17196 * doc/invoke.texi (Option Summary): Add power8 options.
17197 (RS/6000 and PowerPC Options): Likewise.
17199 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
17200 constraints.md instead of rs6000.h. Reorder w* constraints. Add
17201 wm, wn, wr documentation.
17203 * config/rs6000/constraints.md (wm): New constraint for VSX
17204 registers if direct move instructions are enabled.
17205 (wn): New constraint for no registers.
17206 (wq): New constraint for quad word even GPR registers.
17207 (wr): New constraint if 64-bit instructions are enabled.
17208 (wv): New constraint if power8 vector instructions are enabled.
17209 (wQ): New constraint for quad word memory locations.
17211 * config/rs6000/predicates.md (const_0_to_15_operand): New
17212 constraint for 0..15 for crypto instructions.
17213 (gpc_reg_operand): If VSX allow registers in VSX registers as well
17214 as GPR and floating point registers.
17215 (int_reg_operand): New predicate to match only GPR registers.
17216 (base_reg_operand): New predicate to match base registers.
17217 (quad_int_reg_operand): New predicate to match even GPR registers
17218 for quad memory operations.
17219 (vsx_reg_or_cint_operand): New predicate to allow vector logical
17220 operations in both GPR and VSX registers.
17221 (quad_memory_operand): New predicate for quad memory operations.
17222 (reg_or_indexed_operand): New predicate for direct move support.
17224 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
17225 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
17226 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
17227 (POWERPC_MASKS): Add power8 options.
17228 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
17231 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
17232 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
17234 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
17235 (-mpower8-fusion): New power8 options.
17236 (-mpower8-fusion-sign): Likewise.
17237 (-mpower8-vector): Likewise.
17238 (-mcrypto): Likewise.
17239 (-mdirect-move): Likewise.
17240 (-mquad-memory): Likewise.
17242 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
17243 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
17245 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
17246 (rs6000_debug_vector_unit): Add p8_vector.
17247 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
17248 definitions. Also print fusion state.
17249 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
17250 (rs6000_builtin_mask_calculate): Add power8 builtin support.
17251 (rs6000_option_override_internal): Add support for power8.
17252 (rs6000_common_init_builtins): Add debugging for skipped builtins
17253 if -mdebug=builtin.
17254 (rs6000_adjust_cost): Add power8 support.
17255 (rs6000_issue_rate): Likewise.
17256 (insn_must_be_first_in_group): Likewise.
17257 (insn_must_be_last_in_group): Likewise.
17258 (force_new_group): Likewise.
17259 (rs6000_register_move_cost): Likewise.
17260 (rs6000_opt_masks): Likewise.
17262 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
17263 power8 capable assembler, default to power7 options.
17264 (TARGET_DIRECT_MOVE): Likewise.
17265 (TARGET_CRYPTO): Likewise.
17266 (TARGET_P8_VECTOR): Likewise.
17267 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
17268 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
17269 (VECTOR_MEM_P8_VECTOR_P): Likewise.
17270 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
17271 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
17272 (TARGET_XSCVDPSPN): Likewise.
17273 (TARGET_XSCVSPDPN): Likewsie.
17274 (TARGET_SYNC_HI_QI): Likewise.
17275 (TARGET_SYNC_TI): Likewise.
17276 (MASK_CRYPTO): Likewise.
17277 (MASK_DIRECT_MOVE): Likewise.
17278 (MASK_P8_FUSION): Likewise.
17279 (MASK_P8_VECTOR): Likewise.
17280 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
17281 temporary used by some of the direct move instructions to get two FP
17282 temporary registers does not force creation of a stack frame.
17283 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
17284 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
17285 that any VSX registers are tieable, even if they are also an
17286 Altivec vector mode.
17287 (r6000_reg_class_enum): Add wm, wr, wv constraints.
17288 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
17289 (RS6000_BTM_CRYPTO): Likewise.
17290 (RS6000_BTM_COMMON): Likewise.
17292 * config/rs6000/rs6000.md (cpu attribute): Add power8.
17293 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
17294 (enum rs6000_vector): Add power8 vector support.
17296 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17300 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
17301 (any_sibcall_could_use_r3): this and handle indirect calls.
17302 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
17304 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17306 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
17308 2013-05-22 Richard Biener <rguenther@suse.de>
17310 PR middle-end/57349
17311 * profile.c (branch_prob): Do not split blocks that are
17312 abnormally receiving from ECF_RETURNS_TWICE functions.
17314 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17316 * recog.c (offsettable_address_addr_space_p): Fix calculation of
17317 address mode. Move pointer mode initialization to the same place.
17319 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
17321 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
17322 while it has any effect.
17324 2013-05-21 Easwaran Raman <eraman@google.com>
17326 PR tree-optimization/57322
17327 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
17328 UID of the statement added to the BB to be 1.
17330 2013-05-21 Jakub Jelinek <jakub@redhat.com>
17332 PR tree-optimization/57331
17333 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
17334 of conversion from pointer type to integral type with integer.
17336 2013-05-21 Martin Jambor <mjambor@suse.cz>
17339 * ipa-prop.c (ipa_read_node_info): Process param_used and
17340 controlled_uses in the same order as when writing.
17342 2013-05-21 Magnus Granberg <baldrick@free.fr>
17345 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
17347 2013-05-21 Richard Biener <rguenther@suse.de>
17349 PR tree-optimization/57318
17350 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
17351 estimate stmts with side-effects as likely eliminated.
17353 2013-05-21 Richard Biener <rguenther@suse.de>
17355 PR tree-optimization/57330
17356 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
17357 preserve the call stmts fntype.
17359 2013-05-21 Richard Biener <rguenther@suse.de>
17361 PR tree-optimization/57303
17362 * tree-ssa-sink.c (statement_sink_location): Improve killing
17363 stmt detection and properly handle self-assignments.
17365 2013-05-21 Christian Bruel <christian.bruel@st.com>
17367 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
17368 spanning registers. LEAF_REG_REMAP is supported only for contiguous
17369 registers. Set register size out of the PARALLEL loop.
17371 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
17374 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
17375 (*fmasf4, *fmasf4_media): New insns.
17377 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17379 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
17380 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
17381 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
17382 (mips_idiv_insns): Update the comments to say that the returned
17383 instruction counts are in units of BASE_INSN_LENGTH.
17384 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
17385 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
17386 using 2 rather than 4 as the length of indirect MIPS16 and
17387 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
17388 length of a NOP. Don't divide MIPS16 lengths by 2.
17389 (mips16_split_long_branches): Assume a branch is long if the
17390 length is greater than 4 rather than 8.
17391 * config/mips/mips.md (length): Give MIPS16 lengths directly,
17392 rather than multiplying them by 2. Multiply instruction counts
17393 by BASE_INSN_LENGTH rather than 4.
17394 (*jump_mips16, tls_get_tp_mips16_<mode>)
17395 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
17397 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17399 * config/mips/mips.md (extended_mips16): Remove branch case.
17400 (length): Remove duplicated extended_mips16 test.
17402 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
17404 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
17406 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
17408 * recog.h (Recog_data): Rename to...
17409 (recog_data_d): ...this.
17410 (recog_data): Update accordingly.
17411 * recog.c (recog_data): Likewise.
17412 * reload.c (save_recog_data): Likewise.
17413 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
17414 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
17416 2013-05-17 Julian Brown <julian@codesourcery.com>
17418 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
17419 found in a REG_EQUAL note, invalidate it.
17421 2013-05-17 Easwaran Raman <eraman@google.com>
17423 * tree-ssa-reassoc.c (find_insert_point): New function.
17424 (insert_stmt_after): Likewise.
17425 (get_def_stmt): Likewise.
17426 (ensure_ops_are_available): Likewise.
17427 (not_dominated_by): Likewise.
17428 (rewrite_expr_tree): Do not move statements beyond what is
17429 necessary. Remove call to swap_ops_for_binary_stmt...
17430 (reassociate_bb): ... and move it here.
17431 (build_and_add_sum): Assign UIDs for new statements.
17432 (linearize_expr): Likewise.
17433 (do_reassoc): Renumber gimple statement UIDs.
17435 2013-05-17 Jan Hubicka <jh@suse.cz>
17437 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
17439 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
17440 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
17442 (output_weakrefs): Update.
17444 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
17445 Martin Jambor <mjambor@suse.cz>
17447 PR middle-end/57276
17448 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
17449 value that corresponds to the given aggval is found in values vector.
17451 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
17453 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
17454 sse, sse2, sse3, ssse3 and sse4a flags to options.
17456 2013-05-17 David Malcolm <dmalcolm@redhat.com>
17458 * gengtype-state.c: (s_expr_writer): New class, to handle
17459 prettifying of output layout of s-expressions.
17460 (state_writer): New class, to write out gtype.state.
17461 (state_written_type_count): Move this variable into member data of
17463 (s_expr_writer::s_expr_writer): New code: constructor for new class
17464 (state_writer::state_writer(): ditto
17465 (s_expr_writer::write_new_line): New function
17466 (s_expr_writer::write_any_indent): ditto
17467 (s_expr_writer::begin_s_expr): ditto
17468 (s_expr_writer::end_s_expr): ditto
17469 (write_state_fileloc): convert to method of state_writer...
17470 (state_writer:: write_state_fileloc): ...and use methods of
17471 s_expr_writer to write indentation into the gtype.state output file
17472 to visually represent the hierarchical structure of the list
17474 (write_state_fields): ditto, renaming to...
17475 (state_writer::write_state_fields)
17476 (write_state_a_string): ditto, renaming to...
17477 (state_writer::write_state_a_string)
17478 (write_state_string_option): ditto, renaming to...
17479 (state_writer::write_state_string_option)
17480 (write_state_type_option): ditto, renaming to...
17481 (state_writer::write_state_type_option)
17482 (write_state_nested_option): ditto, renaming to...
17483 (state_writer::write_state_nested_option)
17484 (write_state_option): ditto, renaming to...
17485 (state_writer::write_state_option)
17486 (write_state_options): ditto, renaming to...
17487 (state_writer::write_state_options)
17488 (write_state_lang_bitmap): ditto, renaming to...
17489 (state_writer::write_state_lang_bitmap)
17490 (write_state_version): ditto, renaming to...
17491 (state_writer::write_state_version)
17492 (write_state_scalar_type): ditto, renaming to...
17493 (state_writer::write_state_scalar_type)
17494 (write_state_string_type): ditto, renaming to...
17495 (state_writer::write_state_string_type)
17496 (write_state_undefined_type): ditto, renaming to...
17497 (state_writer::write_state_undefined_type)
17498 (write_state_struct_union_type): ditto, renaming to...
17499 (state_writer::write_state_struct_union_type)
17500 (write_state_struct_type): ditto, renaming to...
17501 (state_writer::write_state_struct_type)
17502 (write_state_user_struct_type): ditto, renaming to...
17503 (state_writer::write_state_user_struct_type)
17504 (write_state_lang_struct_type): ditto, renaming to...
17505 (state_writer::write_state_lang_struct_type)
17506 (write_state_param_struct_type): ditto, renaming to...
17507 (state_writer::write_state_param_struct_type)
17508 (write_state_pointer_type): ditto, renaming to...
17509 (state_writer::write_state_pointer_type)
17510 (write_state_array_type): ditto, renaming to...
17511 (state_writer::write_state_array_type)
17512 (write_state_gc_used): ditto, renaming to...
17513 (state_writer::write_state_gc_used)
17514 (write_state_common_type_content): ditto, renaming to...
17515 (state_writer::write_state_common_type_content)
17516 (write_state_type): ditto, renaming to...
17517 (state_writer::write_state_type)
17518 (write_state_pair_list): ditto, renaming to...
17519 (state_writer::write_state_pair_list)
17520 (write_state_pair): ditto, renaming to...
17521 (state_writer::write_state_pair)
17522 (write_state_typedefs): ditto, renaming to...
17523 (state_writer::write_state_typedefs)
17524 (write_state_structures): ditto, renaming to...
17525 (state_writer::write_state_structures)
17526 (write_state_param_structs): ditto, renaming to...
17527 (state_writer::write_state_param_structs)
17528 (write_state_variables): ditto, renaming to...
17529 (state_writer::write_state_variables)
17530 (write_state_srcdir): ditto, renaming to...
17531 (state_writer::write_state_srcdir)
17532 (write_state_files_list): ditto, renaming to...
17533 (state_writer::write_state_files_list)
17534 (write_state_languages): ditto, renaming to...
17535 (state_writer::write_state_languages)
17536 (write_state): create a state_writer instance and use it when
17537 writing out the state file
17539 2013-05-17 Mike Stump <mikestump@comcast.net>
17541 PR rtl-optimization/57304
17542 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
17543 accessing DF_REF_REAL_LOC.
17545 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17547 PR rtl-optimization/57281
17548 PR rtl-optimization/57300
17549 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
17550 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
17551 what the other splitter did if the registers are dead.
17553 2013-05-17 Richard Biener <rguenther@suse.de>
17555 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
17558 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17560 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
17563 2013-05-17 Marek Polacek <polacek@redhat.com>
17565 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
17566 length when doing non-zero store of storing '\0' to '\0'.
17568 2013-05-17 Jakub Jelinek <jakub@redhat.com>
17570 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
17571 vect_external_def oprnd1 with loop_vinfo, try to emit
17572 optional cast, negation and and stmts on the loop preheader
17573 edge instead of into the pattern def seq.
17575 PR tree-optimization/57051
17576 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
17577 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
17579 2013-05-16 Nick Clifton <nickc@redhat.com>
17581 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
17582 (rl78_is_naked_func): New function.
17583 (rl78_expand_prologue): Skip prologue generation for naked functions.
17584 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17585 * doc/extend.texi (naked): Add RL78 to the list of processors
17586 that supports this attribute.
17588 2013-05-16 Jeff Law <law@redhat.com>
17590 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
17592 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17594 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
17595 cache parameters using detect_caches_amd also for CYRIX,
17596 NSC and TM2 signatures.
17598 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
17599 Dzianis Kahanovich <mahatma@eu.by>
17603 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
17604 VIA/Centaur processors and determine their cache parameters
17605 using detect_caches_amd.
17607 2013-05-16 Teresa Johnson <tejohnson@google.com>
17609 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
17610 (rtl_verify_edges): New function.
17611 (rtl_verify_bb_insns): Ditto.
17612 (rtl_verify_bb_pointers): Ditto.
17613 (rtl_verify_bb_insn_chain): Ditto.
17614 (rtl_verify_fallthru): Ditto.
17615 (rtl_verify_bb_layout): Ditto.
17616 (rtl_verify_flow_info_1): Outline checks into new functions.
17617 (rtl_verify_flow_info): Ditto.
17619 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
17621 * cfghooks.c (copy_bbs): Add update_dominance argument.
17622 * cfghooks.h (copy_bbs): Update prototype.
17623 * tree-cfg.c (gimple_duplicate_sese_region):
17624 Add update_dominance argument.
17625 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
17626 * tree-ssa-loop-ch.c (copy_loop_headers): Update
17627 gimple_duplicate_sese_region call.
17628 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
17629 Update copy_bbs call.
17630 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
17631 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
17633 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17635 * tree-vectorizer.h (NUM_PATTERNS): Increment.
17636 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
17637 vect_recog_rotate_pattern.
17638 (vect_recog_rotate_pattern): New function.
17640 2013-05-16 Jason Merrill <jason@redhat.com>
17642 * Makefile.in (LLINKER): New variable.
17643 (mostlyclean): Remove link mutex.
17644 * configure.ac: Handle --enable-link-mutex.
17645 * lock-and-run.sh: New script.
17647 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17650 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
17653 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17655 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
17657 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
17659 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
17660 * config/arm/arm.c (next_consecutive_mem): New function.
17661 (gen_movmem_ldrd_strd): Likewise.
17662 * config/arm/arm.md (movmemqi): Update condition and code.
17663 (unaligned_loaddi, unaligned_storedi): New patterns.
17665 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17667 * config.gcc: Obsolete *-*-solaris2.9*.
17668 * doc/install.texi (Specific, *-*-solaris2*): Document it.
17670 2013-05-16 Richard Biener <rguenther@suse.de>
17672 * passes.c (init_optimization_passes): Move pass_parallelize_loops
17673 earlier, after GRAPHITE transforms and IV canonicalization.
17675 2013-05-16 Jakub Jelinek <jakub@redhat.com>
17677 * omp-low.c (extract_omp_for_data): For collapsed loops,
17678 if at least one of the loops is known at compile time to
17679 iterate zero times, set count to 0.
17680 (expand_omp_regimplify_p): New function.
17681 (expand_omp_for_generic): For collapsed loops, if at least
17682 one of the loops isn't known to iterate at least once,
17683 add runtime check with setting count to 0.
17684 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
17685 For unsigned types if it isn't known at compile time that
17686 the loop will iterate at least once, add runtime check to bypass
17687 the whole loop if initial condition isn't true.
17689 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
17691 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
17693 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
17695 PR middle-end/57286
17696 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
17697 transformations to avoid an infinite loop.
17699 2013-05-16 Marek Polacek <polacek@redhat.com>
17701 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
17703 2013-05-15 Leif Ekblad <leif@rdos.net>
17705 * config/i386/i386.c (ix86_decompose_address): Use
17706 DEFAULT_TLS_SEG_REG to access TLS segment register.
17707 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
17708 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
17709 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
17711 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
17714 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
17715 sibling calls to functions that would normally be lazily bound,
17716 unless $gp is call-clobbered.
17718 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
17720 * config/i386/i386.c (ix86_option_override_internal): Update
17721 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
17722 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
17723 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
17724 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
17726 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
17728 2013-05-15 Andreas Schwab <schwab@suse.de>
17730 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
17731 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
17734 2013-05-15 Teresa Johnson <tejohnson@google.com>
17736 * loop-unroll.c (report_unroll_peel): Check decision before
17737 emitting unroll/peel message.
17739 2013-05-15 Teresa Johnson <tejohnson@google.com>
17741 * function.h (has_bb_partition): New rtl_data flag.
17742 (bb_reorder_complete): Ditto.
17743 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
17744 instead of flag_reorder_blocks_and_partition.
17745 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
17746 with some enhancements.
17747 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
17748 * bb-reorder.c (connect_traces): Check for has_bb_partition
17749 instead of flag_reorder_blocks_and_partition.
17750 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
17751 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
17752 verify_hot_cold_block_grouping.
17753 (partition_hot_cold_basic_blocks): Set has_bb_partition.
17755 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
17758 * config/arm/predicates.md (call_insn_operand): New predicate.
17759 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
17760 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
17761 if insn is not a tail call.
17762 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
17764 * config/arm/arm.h (enum reg_class): New caller save register class.
17765 (REG_CLASS_NAMES): Likewise.
17766 (REG_CLASS_CONTENTS): Likewise.
17767 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
17770 2013-05-15 Richard Biener <rguenther@suse.de>
17772 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
17773 of MSG_OPTIMIZED_LOCATIONS.
17774 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
17775 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
17777 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
17778 of MSG_OPTIMIZED_LOCATIONS.
17779 (execute_vect_slp): Likewise.
17780 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
17781 (vect_create_cond_for_alias_checks): Likewise.
17782 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
17783 (vect_recog_widen_mult_pattern): Likewise.
17784 (vect_recog_widen_sum_pattern): Likewise.
17785 (vect_recog_over_widening_pattern): Likewise.
17786 (vect_recog_widen_shift_pattern): Likewise.
17787 (vect_recog_vector_vector_shift_pattern): Likewise.
17788 (vect_recog_divmod_pattern): Likewise.
17789 (vect_recog_mixed_size_cond_pattern): Likewise.
17790 (vect_recog_bool_pattern): Likewise.
17791 (vect_pattern_recog_1): Likewise.
17793 2013-05-15 Martin Jambor <mjambor@suse.cz>
17795 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
17796 non-functions to builtin_unreachable.
17797 * ipa-inline-transform.c (inline_call): Do not assert estimates were
17798 correct when new direct edges were discovered.
17800 2013-05-15 Martin Jambor <mjambor@suse.cz>
17802 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
17803 header, print symbol order instead of node uid, print more information
17804 about indirect edge targets.
17805 (ipa_make_edge_direct_to_target): Print symbol order instead of node
17807 (ipa_make_edge_direct_to_target): Likewise.
17808 (remove_described_reference): Likewise.
17809 (propagate_controlled_uses): Likewise.
17810 (ipa_print_node_params): Also print symbol order.
17811 (ipcp_transform_function): Print symbol order instead of node uids.
17812 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
17813 (cgraph_get_create_real_symbol_node): Likewise.
17814 * ipa-cp.c (print_lattice): Likewise.
17815 (print_all_lattices): Likewise.
17816 (determine_versionability): Likewise.
17817 (initialize_node_lattices): Likewise.
17818 (estimate_local_effects): Likewise.
17819 (update_profiling_info): Likewise.
17820 (create_specialized_node): Likewise.
17821 (perhaps_add_new_callers): Likewise.
17822 (decide_about_value): Likewise.
17823 (decide_whether_version_node): Likewise.
17824 (identify_dead_nodes): Likewise.
17825 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
17826 (dump_inline_summary): Likewise.
17827 (estimate_node_size_and_time): Likewise.
17828 (inline_analyze_function): Likewise.
17829 * ipa-inline.c (report_inline_failed_reason): Likewise.
17830 (want_early_inline_function_p): Likewise.
17831 (edge_badness): Likewise.
17832 (update_edge_key): Likewise.
17833 (inline_small_functions): Likewise. Add dumping of order to two other
17835 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
17836 instead of node uids.
17837 (propagate_pure_const): Likewise.
17838 (propagate_pure_const): Likewise.
17839 * ipa-utils.c (dump_cgraph_node_set): Likewise.
17840 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
17841 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
17843 * tree-pretty-print.c (dump_function_header): Likewise.
17844 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
17845 Print symbol order instead of node uids.
17847 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17849 * config/s390/s390.c (s390_register_move_cost): Don't impose the
17850 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
17852 2013-05-15 Richard Biener <rguenther@suse.de>
17854 PR tree-optimization/57275
17855 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
17856 return value for fail to do runtime alias checks for gather loads.
17858 2013-05-15 Jan Hubicka <jh@suse.cz>
17862 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
17863 weakrefs are not external.
17864 (lto_symtab_merge_decls): Fix thinko when dealing with
17865 non-lto_symtab decls.
17866 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
17867 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
17868 * varpool.c (dump_varpool_node): Dump more flags.
17870 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
17872 * config/i386/i386.c (processor_alias_table): Add instruction
17873 FSGSBASE for AMD bdver3 architecture.
17875 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17877 * tree.c (warn_deprecated_use): Print file:line using locus color.
17878 * diagnostic.c (diagnostic_report_current_module): Print file:line
17879 and file:line:column using locus color.
17881 2013-05-14 Mike Stump <mikestump@comcast.net>
17883 * gdbinit.in: Add __null.
17885 2013-05-14 Mike Stump <mikestump@comcast.net>
17887 * recog.h: Rename struct recog_data to Recog_data.
17888 * recog.c: Likewise.
17889 * reload.c (can_reload_into): Likewise.
17890 * config/picochip/picochip.c: Likewise.
17892 2013-05-14 Mike Stump <mikestump@comcast.net>
17894 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
17896 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
17898 * resource.h (struct resources): Remove unch_memory member.
17899 (CLEAR_RESOURCE): Don't clear unch_memory.
17900 * resource.c (mark_referenced_resources): Don't set it.
17901 (mark_set_resources): Likewise.
17902 (mark_target_live_regs): Don't clear it.
17903 (init_resource_info): Likewise.
17904 * reorg.c (resource_conflicts_p): Don't compare it.
17905 (redundant_insn): Don't set it.
17907 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
17909 * emit-rtl.c (next_label): Remove unused function.
17910 (skip_consecutive_labels, link_cc0_insns): Move to ...
17911 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
17912 only place where these functions are used, and make them static.
17914 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17916 * fold-const.c (fold_negate_expr): Handle vectors.
17917 (fold_truth_not_expr): Make it static.
17918 (fold_invert_truthvalue): New static function.
17919 (invert_truthvalue_loc): Handle vectors. Do not call
17920 fold_truth_not_expr directly.
17921 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
17922 <TRUTH_NOT_EXPR>: Do not cast to boolean.
17923 (fold_comparison): Handle vector constants.
17924 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
17925 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
17926 * tree.h (fold_truth_not_expr): Remove declaration.
17928 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
17930 * config/aarch64/aarch64-simd.md
17931 (aarch64_vcond_internal<mode>): Rename to...
17932 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
17933 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
17934 float modes. Clarify all iterator modes.
17935 (vcond<mode><mode>): Use new name for vcond expanders.
17936 (vcond<v_cmp_result><mode>): Likewise.
17937 (vcondu<mode><mode>: Likewise.
17938 * config/aarch64/iterators.md (VDQF_COND): New.
17940 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
17943 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
17944 variable for the shift amount. Check that we shift by non-negative
17947 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
17950 * config/arm/arm.h (EPILOGUE_USES): Only return true
17951 for LR_REGNUM after epilogue_completed.
17953 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
17955 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
17956 is error_mark_node.
17958 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
17961 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
17962 and Solaris 11+/x86 with gld.
17963 * configure: Regenerate.
17965 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17967 * expmed.c (expand_shift_1): Canonicalize rotates by
17968 constant bitsize / 2 to bitsize - 1.
17969 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
17970 case ROTATERT>: Likewise.
17973 2013-05-10 Jakub Jelinek <jakub@redhat.com>
17975 * config/i386/i386.md (rotateinv): New code attr.
17976 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
17977 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
17978 roll $31, %eax, etc.
17980 2013-05-14 Richard Biener <rguenther@suse.de>
17982 PR middle-end/57235
17983 * tree-eh.c (sink_clobbers): Give up for successors with
17984 multiple predecessors and no virtual uses.
17986 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
17988 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
17989 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
17991 2013-05-14 Jakub Jelinek <jakub@redhat.com>
17993 PR middle-end/57251
17994 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
17995 the case when both op0 and op1 have VOIDmode.
17997 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17999 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
18000 in multiply-accumulate mode.
18002 2013-05-13 Guozhi Wei <carrot@google.com>
18004 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
18006 2013-05-13 Kai Tietz <ktietz@redhat.com>
18009 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
18010 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
18011 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
18012 * config/i386/i386.c (ix86_option_override_internal): Likewise.
18013 (ix86_expand_prologue): Likewise.
18014 (ix86_expand_split_stack_prologue): Likewise.
18015 (legitimate_pic_address_disp_p): Likewise.
18016 (legitimize_pic_address): Likewise.
18017 (legitimize_tls_address): Likewise.
18018 (legitimize_pe_coff_symbol): Likewise.
18019 (output_pic_addr_const): Likewise.
18020 (construct_plt_address): Likewise.
18021 (ix86_expand_call): Likewise.
18022 (x86_output_mi_thunk): Likewise.
18023 (x86_function_profiler): Likewise.
18025 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
18027 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
18028 similar switch cases.
18029 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
18030 (aarch64_simd_mov_to_<mode>low): Delete.
18031 (aarch64_simd_mov_to_<mode>high): Delete.
18032 (move_lo_quad_<mode>): Add w<-r alternative.
18033 (aarch64_simd_move_hi_quad_<mode>): Likewise.
18034 (aarch64_simd_mov_from_*): Update type attribute.
18035 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
18038 2013-05-13 Jan Hubicka <jh@suse.cz>
18040 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
18041 * config/i386/i386.c (ix86_compute_frame_layout,
18042 ix86_expand_epilogue, emit_i387_cw_initialization,
18043 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
18044 ix86_local_alignment): Fix use of size/speed predicates.
18046 2013-05-13 Jakub Jelinek <jakub@redhat.com>
18048 PR tree-optimization/45216
18049 PR tree-optimization/57157
18050 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
18051 the (-Y) & (B - 1) variant if OP is |.
18052 * expmed.c (expand_shift_1): For rotations by const0_rtx just
18053 return shifted. Use (-op1) & (prec - 1) as other_amount
18054 instead of prec - op1.
18056 2013-05-13 Martin Jambor <mjambor@suse.cz>
18058 PR middle-end/42371
18059 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
18060 (ipa_constant_data): New type.
18061 (ipa_jump_func): Use ipa_constant_data to hold information about
18062 constant jump functions.
18063 (ipa_get_jf_constant): Adjust to jump function type changes.
18064 (ipa_get_jf_constant_rdesc): New function.
18065 (ipa_param_descriptor): New field controlled_uses.
18066 (ipa_get_controlled_uses): New function.
18067 (ipa_set_controlled_uses): Likewise.
18068 * ipa-ref.h (ipa_find_reference): Declare.
18069 * ipa-prop.c (ipa_cst_ref_desc): New type.
18070 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
18072 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
18073 New parameter cs. Adjust all callers.
18074 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
18075 (remove_described_reference): New function.
18076 (jfunc_rdesc_usable): Likewise.
18077 (try_make_edge_direct_simple_call): Decrement controlled use count,
18078 attempt to remove reference if it hits zero.
18079 (combine_controlled_uses_counters): New function.
18080 (propagate_controlled_uses): Likewise.
18081 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
18082 (ipa_edge_duplication_hook): Duplicate reference descriptions.
18083 (ipa_print_node_params): Print described use counter.
18084 (ipa_write_jump_function): Adjust to jump function type changes.
18085 (ipa_read_jump_function): New parameter CS, pass it to
18086 ipa_set_jf_constant. Adjust caller.
18087 (ipa_write_node_info): Stream controlled use count
18088 (ipa_read_node_info): Likewise.
18089 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
18091 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
18092 count. Remove cloning-added reference if it reaches zero.
18093 * ipa-ref.c (ipa_find_reference): New function.
18095 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18097 * config/i386/i386.c (processor_target_table): Modified default
18098 alignment values for AMD BD and BT architectures.
18100 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
18102 * tree-vect-generic.c (uniform_vector_p): Move ...
18103 * tree.c (uniform_vector_p): ... here.
18104 * tree.h (uniform_vector_p): Declare it.
18105 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
18108 2013-05-13 Jakub Jelinek <jakub@redhat.com>
18110 PR tree-optimization/57230
18111 * tree-ssa-strlen.c (handle_char_store): Record length for
18112 array store from STRING_CST.
18114 PR tree-optimization/57230
18115 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
18118 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
18120 * config/epiphany/epiphany.c (epiphany_init): Check size of
18121 NUM_MODES_FOR_MODE_SWITCHING.
18122 (epiphany_expand_prologue):
18123 Remove CONFIG_REGNUM initial value handling code.
18124 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
18125 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
18126 (emit_set_fp_mode, epiphany_mode_after): Likewise.
18127 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
18128 Don't return 1 for FP_MODE_NONE.
18129 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
18130 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
18131 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
18132 * config/epiphany/epiphany.md (save_config): New pattern.
18134 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
18136 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
18138 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
18140 * config/i386/i386.md (memory): Handle sseishft1.
18141 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
18142 (*vec_extractv2di_1): Ditto.
18144 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
18146 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
18149 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
18151 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
18152 Add mthumb/march=armv7-a multilib.
18153 Add mthumb/march=armv7-r multilib.
18154 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
18156 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
18158 * config/v850/t-rtems: Add more multilibs.
18160 2013-05-10 Richard Biener <rguenther@suse.de>
18162 PR tree-optimization/57214
18163 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
18164 not propagate from SSA names that occur in abnormal PHI nodes.
18166 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
18168 * stor-layout.c (element_precision): New function.
18169 * machmode.h (element_precision): Declare it.
18170 * tree.c (build_minus_one_cst): New function.
18171 (element_precision): Likewise.
18172 * tree.h (build_minus_one_cst): Declare new function.
18173 (element_precision): Likewise.
18174 * fold-const.c (operand_equal_p): Use element_precision.
18175 (fold_binary_loc): Handle vector types.
18176 * convert.c (convert_to_integer): Use element_precision.
18177 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
18180 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
18182 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
18183 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
18184 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
18185 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18186 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
18187 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
18188 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
18189 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
18190 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
18191 (Uuw8): New constraints.
18192 (Usb4): Move into alphabetical order.
18193 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
18194 (sd8_operand, ub8_operand, uw8_operand): New predicates.
18195 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
18196 previously unnamed patterns.
18197 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
18198 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
18199 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
18200 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
18201 of set_attr_alternative/if_then_else. Use extended_mips16 instead
18202 of specific lengths.
18204 2013-05-10 Jakub Jelinek <jakub@redhat.com>
18206 * config/i386/i386.md (rotateinv): New code attr.
18207 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
18208 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
18209 roll $31, %eax, etc.
18211 PR tree-optimization/45216
18212 PR tree-optimization/57157
18213 * tree-ssa-forwprop.c (simplify_rotate): New function.
18214 (ssa_forward_propagate_and_combine): Call it.
18216 2013-05-10 Richard Biener <rguenther@suse.de>
18218 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
18219 disable peeling when we version for aliasing.
18220 (vector_alignment_reachable_p): Honor explicit user alignment.
18221 (vect_supportable_dr_alignment): Likewise.
18222 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
18223 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
18224 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
18225 then peeling to arrange for the cost-model check to come first.
18227 2013-05-10 Alan Modra <amodra@gmail.com>
18229 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
18230 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
18231 * configure: Regenerate.
18233 2013-05-10 Alan Modra <amodra@gmail.com>
18236 * varasm.c (default_elf_select_section): Move !DECL_P check..
18237 (get_named_section): ..to here before calling get_section_name.
18239 (default_section_type_flags): Add DECL_P check.
18240 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
18241 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
18243 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
18245 * config/epiphany/epiphany.c (epiphany_expand_prologue):
18246 When using gen_stack_adjust_str with a register offset, add a
18247 REG_FRAME_RELATED_EXPR note.
18249 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
18251 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
18252 (*vec_extractv4si_zext_mem): Ditto.
18253 (*vec_extractv2di): Add 0->x and x->x alternatives.
18254 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
18255 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
18257 2013-05-09 Jason Merrill <jason@redhat.com>
18259 N3639 C++1y VLA support
18260 * gimplify.c (gimplify_vla_decl): Don't touch an existing
18263 * tree.c (build_constructor_va): New.
18264 * tree.h: Declare it.
18266 2013-05-09 Martin Jambor <mjambor@suse.cz>
18269 * gimple-fold.c (canonicalize_constructor_val): Call
18270 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
18272 2013-05-09 Jan Hubicka <jh@suse.cz>
18273 Richard Biener <rguenther@suse.de>
18276 * symtab.c (symtab_make_decl_local): Do not add private names.
18278 2013-05-09 Jan Hubicka <jh@suse.cz>
18281 * symtab.c (insert_to_assembler_name_hash): Handle clones.
18282 (unlink_from_assembler_name_hash): Likewise.
18283 (symtab_prevail_in_asm_name_hash, symtab_register_node,
18284 symtab_unregister_node, symtab_initialize_asm_name_hash,
18285 change_decl_assembler_name): Update.
18287 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
18289 * config/aarch64/aarch64.md: New movtf split.
18290 (*movtf_aarch64): Update.
18291 (aarch64_movdi_tilow): Handle TF modes and rename to
18292 aarch64_movdi_<mode>low.
18293 (aarch64_movdi_tihigh): Handle TF modes and rename to
18294 aarch64_movdi_<mode>high
18295 (aarch64_movtihigh_di): Handle TF modes and rename to
18296 aarch64_mov<mode>high_di
18297 (aarch64_movtilow_di): Handle TF modes and rename to
18298 aarch64_mov<mode>low_di
18299 (aarch64_movtilow_tilow): Remove spurious whitespace.
18300 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
18302 (aarch64_print_operand): Update.
18304 2013-05-09 Alan Modra <amodra@gmail.com>
18306 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
18308 * configure: Regenerate.
18310 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18312 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
18313 splitter preparation statements.
18314 * config/i386/sse.md (*vec_extract* splitters): Ditto.
18315 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
18318 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18320 * gimple-ssa-strength-reduction.c (count_candidates): Change
18321 return value to int.
18322 (analyze_candidates_and_replace): Change type of length to int.
18324 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
18326 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
18327 (*vec_extract<mode>): Use VI12_128 mode iterator.
18328 (*vec_extract<mode>_mem): Ditto.
18329 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
18332 2013-05-08 Diego Novillo <dnovillo@google.com>
18337 2012-08-17 Diego Novillo <dnovillo@google.com>
18340 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
18341 * config.in: Regenerate.
18342 * configure: Regenerate.
18343 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
18345 2013-05-08 Jan Hubicka <jh@suse.cz>
18348 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
18349 * cgraph.h (symtab_node_base): Add unique_name.
18350 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18351 input_overwrite_node, input_varpool_node): Stream unique_name.
18352 * cgraphclones.c (cgraph_create_virtual_clone,
18353 cgraph_function_versioning): Set unique_name.
18354 * ipa.c (function_and_variable_visibility): Set unique_name.
18356 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18358 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
18359 (alloc_cand_and_find_basis): Restrict conditional candidate
18360 processing to CAND_MULTs.
18362 2013-05-08 Jan Hubicka <jh@suse.cz>
18365 lto-symtab.c (lto_symtab_symbol_p): New function.
18366 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
18367 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
18368 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
18369 Skip static symbols.
18371 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
18373 PR tree-optimization/57200
18374 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
18375 Only call inform if the preceding warning_at returns true.
18377 2013-05-07 Han Shen <shenhan@google.com>
18379 * cfgexpand.c (record_or_union_type_has_array_p): New function.
18380 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
18381 * common.opt (fstack-protector-strong): New option.
18382 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
18383 * doc/invoke.texi (Optimization Options): Document
18384 "-fstack-protector-strong".
18385 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
18387 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18389 * config/mips/mips.c (mips_machine_reorg2): Return 0.
18391 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
18393 * ira.c (update_equiv_regs): Add insn having equiv memory even if
18394 it is not lhs of the insn.
18395 (setup_reg_equiv): Remove insn having equiv memory which it is not
18397 * lra-constraints.c (process_address): Try to improve generation
18398 code for address base + disp.
18399 (lra_constraints): Make correct the code for checking insn setting
18400 up backward equivalence. Remove insn only if it is in the init
18402 * lra-eliminations.c (update_reg_eliminate): Change return value.
18403 (lra_eliminate): Use the result.
18405 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
18407 * config/i386/sse.md (ssescalarnummask): New mode attribute.
18408 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
18409 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
18410 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
18411 register target operands.
18412 (*vec_extractv8hi_sse2): New pattern.
18413 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
18414 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
18415 (*vec_extract<mode>_mem): New insn and split pattern.
18417 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
18419 * config/arm/arm.c (arm_asan_shadow_offset): New function.
18420 (TARGET_ASAN_SHADOW_OFFSET): Define.
18421 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
18422 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
18424 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18426 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
18427 (incr_vec_index): Return -1 if increment not found.
18428 (create_add_on_incoming_edge): Assert if increment not found.
18429 (record_increment): Limit number of increments recorded.
18430 (all_phi_incrs_profitable): Return false if an increment not found.
18431 (replace_profitable_candidates): Don't process increments that were
18433 (analyze_candidates_and_replace): Limit size of incr_vec.
18435 2013-05-07 Richard Biener <rguenther@suse.de>
18437 * calls.c (special_function_p): setjmp-like functions are leaf.
18438 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
18439 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
18441 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
18443 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
18444 (aarch64_simd_mov<mode>): New expander.
18445 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
18446 (aarch64_simd_mov_to_<mode>high): Likewise.
18447 (aarch64_simd_mov_from_<mode>low): Likewise.
18448 (aarch64_simd_mov_from_<mode>high): Likewise.
18449 (aarch64_dup_lane<mode>): Update.
18450 (aarch64_dup_lanedi): New instruction pattern.
18451 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
18452 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
18454 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18456 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
18457 (replace_mult_candidate): Remove unnecessary argument; remove
18458 unnecessary parameter from call to introduce_cast_before_cand.
18459 (replace_unconditional_candidate): Remove unnecessary parameter
18460 from call to replace_mult_candidate.
18461 (replace_conditional_candidate): Likewise.
18462 (insert_initializers): Use make_temp_ssa_name.
18463 (introduce_cast_before_cand): Remove unnecessary argument; use
18464 make_temp_ssa_name.
18465 (replace_one_candidate): Remove unnecessary argument; remove
18466 unnecessary parameter from calls to introduce_cast_before_cand.
18467 (replace_profitable_candidates): Remove unnecessary parameters
18468 from calls to replace_one_candidate.
18470 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18472 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
18473 phi def as possibly hiding a basis for a CAND_ADD whose operands
18474 have been commuted in the analysis.
18475 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
18477 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18479 * config/aarch64/aarch64.md
18480 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
18481 shift value between 0-4.
18483 2013-05-07 Richard Biener <rguenther@suse.de>
18485 * double-int.h (rshift): New overload.
18486 * double-int.c (rshift): New function.
18487 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
18488 (create_reference_ops_from_ref): Remove.
18489 (vn_reference_insert): Use shared ops for constructing the
18490 reference and copy it.
18492 2013-05-07 Richard Biener <rguenther@suse.de>
18494 PR middle-end/57190
18495 * tree-eh.c (sink_clobbers): Properly propagate
18496 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
18498 2013-05-07 Jakub Jelinek <jakub@redhat.com>
18500 PR tree-optimization/57149
18501 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
18502 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
18503 collect_phi_def_edges, execute_late_warn_uninitialized): Use
18504 uninit_undefined_value_p instead of ssa_undefined_value_p.
18507 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
18508 for modifier == EXPAND_INITIALIZER.
18510 2013-05-07 Anton Blanchard <anton@samba.org>
18512 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
18513 for powerpc64 little endian.
18514 * configure: Regenerate.
18516 2013-05-06 Graham Stott <grahams@btinternet.com>
18518 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
18519 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
18520 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
18521 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
18523 2013-05-06 Graham Stott <grahams@btinternet.com>
18525 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
18526 codes which allow non-lvalues.
18528 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
18530 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
18531 components are all 1s.
18532 (integer_minus_onep): New function.
18533 * tree.h (integer_minus_onep): Declare it.
18534 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
18535 integer_minus_onep instead of integer_all_onesp.
18537 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18540 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
18541 variations of these patterns.
18543 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18545 * config/i386/i386.md (isa): Add x64_sse4 member.
18546 (enabled): Handle x64_sse4.
18547 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
18548 instruction for 64bit SSE4_1 targets. Update insn attributes.
18549 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
18550 instruction for SSE4_1 targets. Update insn attributes.
18551 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18552 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
18553 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
18555 (*vec_extractv4si): Rename from *sse4_1_pextrd.
18556 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
18557 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
18559 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18562 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
18564 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
18566 * final.c (do_assembler_dialects): Don't handle curly braces and
18567 vertical bar escaped by % as dialect delimiters.
18568 (output_asm_insn): Print curly braces and vertical bar if escaped
18569 by % and ASSEMBLER_DIALECT defined.
18570 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
18571 * doc/tm.texi: Regenerated.
18573 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
18575 * config/mips/mips.c: Include tree-pass.h.
18576 (mips_reorg): Split in pre- and post-dbr_schedule parts.
18577 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
18578 (pass_mips_machine_reorg2): New machine specific pass.
18579 (insert_pass_mips_machine_reorg2): New pass plugin definition.
18580 (mips_option_override): Register the new pass.
18581 * rtl.h (cleanup_barriers): Remove prototype.
18582 (dbr_schedule): Likewise.
18583 * jump.c (cleanup_barriers): Make static.
18584 * reorg.c (dbr_schedule): Likewise.
18586 2013-05-06 Richard Biener <rguenther@suse.de>
18588 PR tree-optimization/57185
18589 * tree-parloops.c (add_field_for_reduction): Handle anonymous
18590 SSA names properly.
18592 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
18595 * config/i386/i386.c (add_parameter_dependencies): Add dependence
18596 between "first_arg" and "insn", not "last" and "insn".
18598 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18600 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
18601 (find_candidates_in_block): Re-enable slsr_process_phi.
18602 (create_phi_basis): Fix double counting of candidate adjustment.
18604 2013-05-06 Richard Biener <rguenther@suse.de>
18606 PR middle-end/57147
18607 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
18608 the edge is also fallthru, preserve it and just clear the
18610 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
18611 also complex, preserve that and just clear the fallthru flag.
18612 * tree-inline.c (update_ssa_across_abnormal_edges): Also
18613 update virtual operands.
18615 2013-05-06 Alan Modra <amodra@gmail.com>
18617 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
18618 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
18619 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
18620 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
18621 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18622 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
18624 2013-05-06 Alan Modra <amodra@gmail.com>
18626 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
18627 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
18628 (DEFAULT_ASM_ENDIAN): Define.
18629 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
18630 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
18631 Update -K PIC clause from sysv4.h.
18632 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
18633 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
18635 2013-05-06 Alan Modra <amodra@gmail.com>
18637 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
18638 twice for little-endian.
18639 (ashrdi3_no_power, ashrdi3): Support little-endian.
18641 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
18644 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
18645 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
18647 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
18648 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
18649 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
18652 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
18653 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
18655 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
18656 * config/pa/pa.opt: Make mbig-switch a no-op.
18657 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
18658 (CASE_VECTOR_MODE): Always return SImode.
18659 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
18660 for the !TARGET_BIG_SWITCH case.
18661 * config/pa/pa-linux.h: Likewise.
18662 * config/pa/pa-openbsd.h: Likewise.
18663 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
18664 * config/pa/pa.md (short_jump): Remove define_insn.
18665 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
18666 (casesi0): Remove define_insn.
18667 (type): Remove btable_branch.
18668 (pa_combine_type): Likewise.
18669 (in_nullified_branch_delay): Likewise.
18670 (in_call_delay): Likewise.
18671 (define_delay): Likewise.
18672 (define_insn_reservation "Z3"): Likewise.
18673 (define_insn_reservation "Z4"): Likewise.
18674 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
18675 (pa_adjust_insn_length): Remove adjustment for btable branches.
18676 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
18679 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
18681 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
18682 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
18683 Add m->r,x alternatives.
18684 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
18685 splitters using SWI48x mode iterator.
18686 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
18687 TARGET_64BIT. Add m->x alternative.
18688 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
18689 Add o->x alternative. Enable for TARGET_SSE.
18690 (sse_storeq): Remove expander.
18691 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
18692 with memory input operand.
18693 (*vec_extractv2di_1 splitter): New.
18694 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
18695 * config/i386/i386.md (ssevecmodelower): New mode attribute.
18697 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18699 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
18700 (INT_LOWPART): Delete.
18701 (extract_MB): Adjust.
18702 (extract_ME): Adjust.
18703 (print_operand): Adjust.
18705 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
18707 * config/rs6000/predicates.md (reg_or_add_cint_operand,
18708 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18709 (reg_or_logical_cint_operand, easy_fp_constant,
18710 logical_const_operand): Delete "CONST_DOUBLE" case.
18711 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
18712 "HOST_BITS_PER_WIDE_INT == 64" test.
18713 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
18714 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
18715 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
18716 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
18718 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
18719 CONST_DOUBLE DImode/VOIDmode case.
18720 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
18721 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18722 CONST_DOUBLE VOIDmode case.
18723 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
18724 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
18725 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
18726 Delete CONST_DOUBLE case.
18727 (splitters for mov FMOVE64 const_double): Delete
18728 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
18729 "HOST_BITS_PER_WIDE_INT >= 64" test.
18730 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
18732 (mov DI const_double): Delete.
18734 2013-05-04 Jakub Jelinek <jakub@redhat.com>
18736 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
18737 on op shows all bits zero in mode of a lowpart subreg, return zero.
18739 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
18742 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
18743 to save TFmode registers and DImode to save TImode registers for
18744 caller save operations.
18745 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
18746 mark being partially clobbered since they only use the first
18749 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
18750 and TDmode only use the upper 64-bits of each VSX register.
18752 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18754 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
18755 (find_candidates_in_block): Disable slsr_process_phi.
18757 2013-05-03 Guozhi Wei <carrot@google.com>
18759 * coverage.c (coverage_obj_init): Move the construction of gcov
18761 (build_init_ctor): ... here.
18763 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18765 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
18766 (slsr_cand_d): Redefine def_phi.
18767 (stride_status, phi_adjust_status, count_phis_status): New enums.
18768 (find_phi_def): New.
18769 (find_basis_for_base_expr): New.
18770 (find_basis_for_candidate): Handle hidden bases.
18771 (alloc_cand_and_find_basis): Handle phi candidates.
18772 (slsr_process_phi): New.
18773 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
18774 (create_mul_imm_cand): Likewise.
18775 (create_add_ssa_cand): Exclude phi base candidates.
18776 (create_add_imm_cand): Likewise.
18777 (slsr_process_cast): Likewise.
18778 (slsr_process_copy): Likewise.
18779 (find_candidates_in_block): Handle phi candidates.
18780 (dump_candidate): Likewise.
18781 (unconditional_cands): Delete.
18782 (unconditional_cands_with_known_stride_p): Delete.
18783 (phi_dependent_cand_p): New.
18784 (cand_increment): Handle phi-dependent candidates.
18785 (replace_dependent): Delete.
18786 (replace_mult_candidate): New.
18787 (replace_unconditional_candidate): New.
18788 (incr_vec_index): Move to avoid forward reference.
18789 (create_add_on_incoming_edge): New.
18790 (create_phi_basis): New.
18791 (replace_dependents): Delete.
18792 (replace_conditional_candidate): New.
18793 (phi_add_costs): New.
18794 (replace_uncond_cands_and_profitable_phis): New.
18795 (record_increment): Handle phi adjustments.
18796 (record_phi_increments): New.
18797 (record_increments): Handle phi adjustments.
18798 (phi_incr_cost): New.
18799 (lowest_cost_path): Handle phis.
18800 (total_savings): Likewise.
18801 (analyze_increments): Likewise.
18802 (ncd_with_phi): New.
18803 (ncd_of_cand_and_phis): New.
18804 (nearest_common_dominator_for_cands): Handle phi increments.
18805 (all_phi_incrs_profitable): New.
18806 (replace_profitable_candidates): Handle phi-dependent candidates.
18807 (analyze_candidates_and_replace): Likewise.
18809 2013-05-03 Teresa Johnson <tejohnson@google.com>
18812 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
18813 do not exceed REG_BR_PROB_BASE.
18815 2013-05-03 Jeff Law <law@redhat.com>
18817 PR tree-optimization/57144
18818 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
18819 operand of the condition will bit into the new type when eliminating
18820 a cast feeding a condition.
18822 2013-05-03 Jakub Jelinek <jakub@redhat.com>
18824 PR rtl-optimization/57130
18825 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
18826 of COMPARE as in_code to the recursive call if needed.
18828 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
18830 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
18831 (enabled): Handle new members.
18832 * config/i386/sse.md (*vec_concatv2si): Merge from
18833 *vec_concatv2si_sse2 and vec_concatv2si_sse.
18834 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
18836 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
18838 PR tree-optimization/57027
18839 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
18840 for fnms opportunity, check we got the prerequisite kind
18841 of tree / gimple before using accessor functions.
18843 2013-05-03 Richard Biener <rguenther@suse.de>
18845 * double-int.h (lshift): New overload without precision
18846 and arith argument.
18847 (operator *=, operator +=, operator -=): Move ...
18848 * double-int.c (operator *=, operator +=, operator -=): ... here
18849 and implement more efficiently.
18850 (mul_double_with_sign): Remove.
18851 (lshift_double): Adjust to take unsinged shift argument, push
18852 dispatching code to callers.
18853 (mul_double_wide_with_sign): Add early out for callers that
18854 are not interested in high parts or overflow.
18855 (lshift): New function.
18856 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
18857 dispatch code here.
18858 (lrotate, rrotate): Use logical shifts.
18859 * expr.c (get_inner_reference): Use lshift.
18860 * fixed-value.c (do_fixed_divide): Likewise.
18861 * tree-dfa.c (get_ref_base_and_extent): Likewise.
18862 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
18863 (indirect_refs_may_alias_p): Likewise.
18864 (stmt_kills_ref_p_1): Likewise.
18866 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18868 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
18870 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
18872 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
18873 scalar form of FABD instruction.
18875 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
18877 * lra-constraints.c (process_alt_operands): Add checking alt
18878 number to choose the best alternative.
18880 2013-05-02 Richard Biener <rguenther@suse.de>
18882 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
18883 bitmap and its handling.
18884 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
18886 2013-05-02 Richard Biener <rguenther@suse.de>
18888 PR middle-end/57140
18889 * tree-inline.c (copy_loops): Properly handle removed loops.
18890 (copy_cfg_body): Mark destination loops for fixup if source
18891 loops needed fixup.
18893 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
18896 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
18897 generating simple_return for naked functions.
18899 2013-05-02 Martin Jambor <mjambor@suse.cz>
18901 PR middle-end/56988
18902 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
18903 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
18905 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
18906 ipa_agg_replacement_value structures.
18907 (known_aggs_to_agg_replacement_list): Likewise.
18908 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
18909 (read_agg_replacement_chain): Likewise.
18910 (ipcp_transform_function): Also check that by_ref flags match.
18912 2013-05-02 Richard Biener <rguenther@suse.de>
18914 * graphds.h (struct graph): Add obstack member.
18915 * graphds.c (new_graph): Initialize obstack and allocate
18917 (add_edge): Allocate edge from the obstack.
18918 (free_graph): Free the obstack instead of all edges and vertices.
18920 2013-05-02 Teresa Johnson <tejohnson@google.com>
18922 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
18924 * cfg.c (update_bb_profile_for_threading): Ditto.
18925 * tree-inline.c (copy_bb): Ditto.
18926 (copy_edges_for_bb): Ditto.
18927 (initialize_cfun): Ditto.
18928 (copy_cfg_body): Ditto.
18929 (expand_call_inline): Ditto.
18930 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
18931 (estimate_node_size_and_time): Ditto.
18932 (inline_merge_summary): Ditto.
18933 * cgraphclones.c (cgraph_clone_edge): Ditto.
18934 (cgraph_clone_node): Ditto.
18935 * sched-rgn.c (compute_dom_prob_ps): Ditto.
18936 (compute_trg_info): Ditto.
18938 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18940 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
18941 S reg when fp attribute set.
18942 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
18944 2013-05-02 Ian Bolton <ian.bolton@arm.com>
18946 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
18948 (*and_one_cmplsi3_compare0_uxtw): Likewise.
18949 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
18950 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
18952 2013-05-02 Richard Biener <rguenther@suse.de>
18954 * tree-scalar-evolution.c (scev_info_hasher): Remove.
18955 (struct instantiate_cache_entry): New type.
18956 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
18957 (struct instantiate_cache_type): New type.
18958 (set_instantiated_value, get_instantiated_value): Remove.
18959 (get_instantiated_value_entry): New function.
18960 (instantiate_scev_name): Use the new cache and adjust.
18961 (instantiate_scev_poly): Adjust.
18962 (instantiate_scev_binary): Likewise.
18963 (instantiate_array_ref): Likewise.
18964 (instantiate_scev_convert): Likewise.
18965 (instantiate_scev_not): Likewise.
18966 (instantiate_scev_3): Likewise.
18967 (instantiate_scev_2): Likewise.
18968 (instantiate_scev_r): Likewise.
18969 (instantiate_scev): Likewise.
18970 (resolve_mixers): Likewise.
18972 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
18975 * lra-constraints.c (best_small_class_operands_num): Remove.
18976 (process_alt_operands): Remove small_class_operands_num. Take
18977 small classes operands into losers and only if the operand is not
18978 matched. Modify debugging output.
18979 (curr_insn_transform): Remove best_small_class_operands_num.
18982 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
18984 * config/aarch64/aarch64-builtins.c
18985 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
18986 * config/aarch64/aarch64-simd-builtins.def
18987 (reduc_splus_): Add new modes.
18988 (reduc_uplus_): New.
18989 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
18990 (reduc_uplus_v4sf): Likewise.
18991 (reduc_splus_v4sf): Likewise.
18992 (aarch64_addv<mode>): Likewise.
18993 (reduc_uplus_<mode>): Likewise.
18994 (reduc_splus_<mode>): Likewise.
18995 (aarch64_addvv2di): Likewise.
18996 (reduc_uplus_v2di): Likewise.
18997 (reduc_splus_v2di): Likewise.
18998 (aarch64_addvv2si): Likewise.
18999 (reduc_uplus_v2si): Likewise.
19000 (reduc_splus_v2si): Likewise.
19001 (reduc_<sur>plus_<mode>): New.
19002 (reduc_<sur>plus_v2di): Likewise.
19003 (reduc_<sur>plus_v2si): Likewise.
19004 (reduc_<sur>plus_v4sf): Likewise.
19005 (aarch64_addpv4sf): Likewise.
19006 * config/aarch64/arm_neon.h
19007 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
19008 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
19009 add UNSPEC_SADDV, UNSPEC_UADDV.
19011 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
19013 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19015 * config/aarch64/arm_neon.h
19016 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
19018 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19020 * config/aarch64/aarch64-builtins
19021 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
19023 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19025 * config/aarch64/aarch64-simd-builtins.def
19026 (reduc_smax_): New.
19027 (reduc_smin_): Likewise.
19028 (reduc_umax_): Likewise.
19029 (reduc_umin_): Likewise.
19030 (reduc_smax_nan_): Likewise.
19031 (reduc_smin_nan_): Likewise.
19034 (smax): Update for V2SF, V4SF and V2DF modes.
19037 (smin_nan): Likewise.
19038 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
19039 (<su><maxmin><mode>3): ...This, refactor.
19040 (s<maxmin><mode>3): New.
19041 (<maxmin_uns><mode>3): Likewise.
19042 (reduc_<maxmin_uns>_<mode>): Refactor.
19043 (reduc_<maxmin_uns>_v4sf): Likewise.
19044 (reduc_<maxmin_uns>_v2si): Likewise.
19045 (aarch64_<fmaxmin><mode>: Remove.
19046 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
19048 (vmin<q>_f<32,64>): Likewise.
19049 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
19051 (su): Add mappings for smax, smin, umax, umin.
19053 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
19054 (FMAXMIN): Rename as...
19055 (FMAXMIN_UNS): ...This.
19057 (fmaxminv): Likewise.
19058 (fmaxmin): Likewise.
19060 (maxmin_uns_op): Likewise.
19062 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19064 * config/aarch64/arm_neon.h
19065 (vac<ge, gt><sd>_f<32, 64>): Rename to...
19066 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
19067 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
19069 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19071 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
19072 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
19074 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19076 * config/aarch64/aarch64-simd.md
19077 (vcond<mode>_internal): Handle special cases for constant masks.
19078 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
19079 (vcondu<mode><mode>): Likewise.
19080 (vcond<v_cmp_result><mode>): New.
19082 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19084 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
19085 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
19086 * config/aarch64/aarch64-simd-builtins.def
19087 (cmeq): Update to BUILTIN_VALLDI.
19092 * config/aarch64/arm_neon.h
19093 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
19094 to builtins or C as appropriate.
19096 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
19098 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
19100 (cmhi): Rename to...
19102 * config/aarch64/aarch64-simd.md
19103 (simd_mode): Add SF.
19104 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
19105 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
19106 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
19107 (cstore<mode>_neg): ...This.
19108 * config/aarch64/iterators.md
19110 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
19111 (COMPARISONS): New.
19112 (UCOMPARISONS): Likewise.
19113 (optab): Add missing comparisons.
19119 (VCMP_S): Likewise.
19120 (VCMP_U): Likewise.
19121 (V_cmp_result): Add DF, SF modes.
19122 (v_cmp_result): Likewise.
19124 (vmtype): Likewise.
19125 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
19127 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
19129 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
19130 define_insn to define_insn_and_split.
19131 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
19132 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
19133 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
19134 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
19135 (thumb2_negscc): Likewise.
19137 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
19139 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
19141 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
19143 * config/arm/thumb2.md: Remove trailing whitespaces.
19145 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19147 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
19148 Use gen_int_mode rather than GEN_INT.
19150 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
19152 * value-prof.c (stream_in_histogram_value): Remove the strayed
19155 2013-04-30 Richard Biener <rguenther@suse.de>
19157 PR middle-end/57122
19158 * cfghooks.c (split_edge): Properly check for the loop latch edge.
19160 2013-04-30 Richard Biener <rguenther@suse.de>
19162 PR middle-end/57107
19163 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
19165 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
19167 PR rtl-optimization/56957
19168 PR rtl-optimization/57105
19169 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
19170 variable. Use just INSN_UID for determining whether an insn
19171 should be only disconnected from the insn stream.
19172 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
19174 2013-04-30 Jakub Jelinek <jakub@redhat.com>
19176 PR tree-optimization/57104
19177 * tsan.c (instrument_expr): Don't instrument accesses to
19178 DECL_HARD_REGISTER VAR_DECLs.
19180 2013-04-30 Richard Biener <rguenther@suse.de>
19182 * function.h (loops_for_fn): New inline function.
19183 (set_loops_for_fn): Likewise.
19184 * cfgloop.h (place_new_loop): Add struct function parameter.
19185 (get_loop): Likewise.
19186 (get_loops): Likewise.
19187 (number_of_loops): Likewise.
19188 (fel_next): Adjust.
19189 (fel_init): Likewise.
19190 * cfg.c (get_loop_copy): Adjust.
19191 * cfgloop.c (flow_loops_dump): Likewise.
19192 (record_loop_exits): Likewise.
19193 (verify_loop_structure): Likewise.
19194 * cfgloopanal.c (mark_irreducible_loops): Likewise.
19195 (estimate_reg_pressure_cost): Likewise.
19196 (mark_loop_exit_edges): Likewise.
19197 * cfgloopmanip.c (place_new_loop): Likewise.
19198 (add_loop): Likewise.
19199 (duplicate_loop): Likewise.
19200 * graph.c (draw_cfg_nodes): Likewise.
19201 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
19202 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
19203 (extract_affine_chrec): Likewise.
19204 (build_scop_iteration_domain): Likewise.
19205 * graphite.c (graphite_initialize): Likewise.
19206 * ira-build.c (create_loop_tree_nodes): Likewise.
19207 (more_one_region_p): Likewise.
19208 (rebuild_regno_allocno_maps): Likewise.
19209 (mark_loops_for_removal): Likewise.
19210 (mark_all_loops_for_removal): Likewise.
19211 (remove_unnecessary_regions): Likewise.
19212 (ira_build): Likewise.
19213 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
19214 * loop-init.c (fix_loop_structure): Likewise.
19215 (gate_rtl_move_loop_invariants): Likewise.
19216 (gate_rtl_unswitch): Likewise.
19217 (gate_rtl_unroll_and_peel_loops): Likewise.
19218 (rtl_doloop): Likewise.
19219 * lto-streamer-in.c (input_cfg): Likewise.
19220 * lto-streamer-out.c (output_cfg): Likewise.
19221 * modulo-sched.c (sms_schedule): Likewise.
19222 * predict.c (tree_estimate_probability): Likewise.
19223 (tree_estimate_probability_driver): Likewise.
19224 (estimate_loops): Likewise.
19225 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
19226 (move_sese_region_to_fn): Likewise.
19227 (debug_loop_num): Likewise.
19228 * tree-chrec.c (chrec_evaluate): Likewise.
19229 (hide_evolution_in_other_loops_than_loop): Likewise.
19230 (chrec_component_in_loop_num): Likewise.
19231 (reset_evolution_in_loop): Likewise.
19232 (evolution_function_is_invariant_rec_p): Likewise.
19233 * tree-if-conv.c (main_tree_if_conversion): Likewise.
19234 * tree-inline.c (copy_loops): Likewise.
19235 (copy_cfg_body): Likewise.
19236 (tree_function_versioning): Likewise.
19237 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
19238 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
19240 (add_to_evolution_1): Likewise.
19241 (scev_const_prop): Likewise.
19242 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
19243 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
19244 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
19245 (tree_ssa_lim_initialize): Likewise.
19246 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
19247 (verify_loop_closed_ssa): Likewise.
19248 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
19249 (tree_ssa_loop_im): Likewise.
19250 (tree_ssa_loop_unswitch): Likewise.
19251 (tree_vectorize): Likewise.
19252 (check_data_deps): Likewise.
19253 (tree_ssa_loop_ivcanon): Likewise.
19254 (tree_ssa_loop_bounds): Likewise.
19255 (tree_complete_unroll): Likewise.
19256 (tree_complete_unroll_inner): Likewise.
19257 (tree_parallelize_loops): Likewise.
19258 (tree_ssa_loop_prefetch): Likewise.
19259 (tree_ssa_loop_ivopts): Likewise.
19260 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
19261 * tree-vectorizer.c (vectorize_loops): Likewise.
19263 2013-04-29 Mike Frysinger <vapier@gentoo.org>
19265 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
19266 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
19267 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
19268 with EABI_LINK_SPEC.
19270 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19273 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
19276 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
19279 * lra-constraints.c (process_alt_operands): Discourage a bit more
19280 using memory for pseudos. Print cost dump for alternatives.
19281 Modify cost values for conflicts with early clobbers.
19282 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
19284 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19287 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
19289 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19291 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
19292 from/to S register.
19293 (movdi_aarch64): Support LDR/STR from/to D register.
19295 2013-04-29 Ian Bolton <ian.bolton@arm.com>
19297 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
19298 or higher by default.
19300 2013-04-29 Richard Biener <rguenther@suse.de>
19302 PR middle-end/57075
19303 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
19304 even if not adding abnormal edges for calls that can make
19307 2013-04-29 Richard Biener <rguenther@suse.de>
19309 PR middle-end/57103
19310 * tree-cfg.c (move_stmt_op): Fix condition under which to update
19312 (move_stmt_r): Remove redundant checking.
19314 2013-04-29 Teresa Johnson <tejohnson@google.com>
19317 * basic-block.h (apply_scale): New function.
19318 (apply_probability): Use apply_scale.
19319 * gimple-streamer-in.c (input_bb): Ditto.
19320 * lto-streamer-in.c (input_cfg): Ditto.
19321 * lto-cgraph.c (merge_profile_summaries): Ditto.
19322 * tree-optimize.c (execute_fixup_cfg): Ditto.
19323 * tree-inline.c (copy_bb): Update comment to use apply_scale.
19324 (copy_edges_for_bb): Ditto.
19325 (copy_cfg_body): Ditto.
19327 2013-04-29 Tom de Vries <tom@codesourcery.com>
19329 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
19330 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
19331 (tail_merge_optimize): Handle current_loops == NULL.
19333 2013-04-26 Jeff Law <law@redhat.com>
19335 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
19336 (simplify_cond_using_ranges): Generalize code to simplify
19337 COND_EXPRs where one argument is a constant and the other
19338 is an SSA_NAME created by an integral type conversion.
19340 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19342 * config/arm/arm.md (store_minmaxsi): Use only when
19343 optimize_insn_for_size_p.
19345 2013-04-29 Christian Bruel <christian.bruel@st.com>
19348 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
19350 2013-04-29 Richard Biener <rguenther@suse.de>
19352 PR middle-end/57089
19353 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
19354 loop tree make sure to schedule a fixup for the child as well.
19355 (expand_omp_for_generic): Properly add loops.
19356 (expand_omp_for_static_nochunk): Likewise.
19357 (expand_omp_for_static_chunk): Likewise.
19358 (expand_omp_for): For the degenerate case fixup loops.
19359 (expand_omp_sections): Fix default bb placement in loops.
19360 (expand_omp_atomic_pipeline): Properly add loops.
19362 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19364 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
19366 2013-04-29 Tom de Vries <tom@codesourcery.com>
19368 * tree-ssa-tail-merge.c: Update header comment.
19370 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19372 * config/aarch64/arm_neon.h
19373 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
19374 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
19375 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
19376 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
19377 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
19378 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
19379 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
19380 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
19382 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19384 * config/aarch64/aarch64-simd.md
19385 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
19386 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
19387 fix_trunc, fixuns_trunc.
19388 (ftrunc<VDQF:mode>2): New.
19389 * config/aarch64/iterators.md (optab): Add fix, fixuns.
19390 (fix_trunc_optab): New.
19392 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19394 * config/aarch64/aarch64-builtins.c
19395 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
19396 iceilf, lround, iroundf.
19398 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
19401 * config/i386/i386.h (enum ix86_tune_indices)
19402 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
19403 New, split from X86_TUNE_INTER_UNIT_MOVES.
19404 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
19405 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
19406 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
19407 (TARGET_INTER_UNIT_MOVES): Remove.
19408 * config/i386/i386.c (initial_ix86_tune_features): Update.
19409 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
19410 (ix86_expand_convert_uns_didf_sse): Use
19411 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19412 (ix86_expand_vector_init_one_nonzero): Ditto.
19413 (ix86_expand_vector_init_interleave): Ditto.
19414 (inline_secondary_memory_needed): Return true for moves from SSE class
19415 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
19416 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
19417 * config/i386/constraints.md (Yi, Ym): Depend on
19418 TARGET_INTER_UNIT_MOVES_TO_VEC.
19419 (Yj, Yn): New constraints.
19420 * config/i386/i386.md (*movdi_internal): Change constraints of
19421 operand 1 from Yi to Yj and from Ym to Yn.
19422 (*movsi_internal): Ditto.
19423 (*movdf_internal): Ditto.
19424 (*movsf_internal): Ditto.
19425 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
19426 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
19427 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
19428 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
19429 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
19430 * config/i386/sse.md (movdi_to_sse): Ditto.
19431 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
19432 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
19433 TARGET_INTER_UNIT_MOVES.
19434 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
19435 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
19436 instead of TARGET_INTER_UNIT_MOVES.
19437 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
19438 operand 1 from Yi to Yj and from Ym to Yn.
19440 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19442 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
19443 (float_truncate_hi_): Likewise.
19444 (float_extend_lo_): Likewise.
19445 (float_truncate_lo_): Likewise.
19446 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
19447 (aarch64_float_extend_lo_v2df): Likewise.
19448 (vec_unpacks_hi_v4sf): Likewise.
19449 (aarch64_float_truncate_lo_v2sf): Likewise.
19450 (aarch64_float_truncate_hi_v4sf): Likewise.
19451 (vec_pack_trunc_v2df): Likewise.
19452 (vec_pack_trunc_df): Likewise.
19454 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19456 * config/aarch64/aarch64-builtins.c
19457 (aarch64_fold_builtin): Fold float conversions.
19458 * config/aarch64/aarch64-simd-builtins.def
19459 (floatv2si, floatv4si, floatv2di): New.
19460 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
19461 * config/aarch64/aarch64-simd.md
19462 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
19463 * config/aarch64/iterators.md (FLOATUORS): New.
19464 (optab): Add float, floatuns.
19465 (su_optab): Likewise.
19467 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19469 * config/aarch64/aarch64-builtins.c
19470 (aarch64_builtin_vectorized_function): Use new names for
19472 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
19473 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
19474 (fcvtzu): Split as...
19475 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
19476 (fcvtas): Split as...
19477 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
19478 (fcvtau): Split as...
19479 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
19480 (fcvtps): Split as...
19481 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
19482 (fcvtpu): Split as...
19483 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
19484 (fcvtms): Split as...
19485 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
19486 (fcvtmu): Split as...
19487 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
19488 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
19489 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
19490 (lfrintnusf, lfrintnudf): Likewise.
19491 * config/aarch64/aarch64-simd.md
19492 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
19494 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
19495 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
19496 (fcvt_pattern): Likewise.
19498 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19500 * config/aarch64/aarch64-simd.md
19501 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
19502 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
19504 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19506 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
19507 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
19508 (vrnd<a,m,n,p>_f32): Implement using builtins.
19509 (vrnd<i,x><q>_f<32, 64>): New.
19511 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
19513 * config/aarch64/aarch64-builtins.c
19514 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
19515 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
19516 (frintz): Rename to...
19518 (frintp): Rename to...
19520 (frintm): Rename to...
19522 (frinti): Rename to...
19523 (nearbyint): ...this.
19524 (frintx): Rename to...
19526 (frinta): Rename to...
19528 * config/aarch64/aarch64-simd.md
19529 (aarch64_frint<frint_suffix><mode>): Delete.
19530 (<frint_pattern><mode>2): Convert to insn.
19531 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
19532 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
19533 (frint_pattern): Likewise.
19534 (frint_suffix): Likewise.
19536 2013-04-29 Richard Biener <rguenther@suse.de>
19538 PR tree-optimization/57081
19539 * loop-init.c: Include tree-flow.h.
19540 (loop_optimizer_finalize): Free number of iteration estimates.
19541 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
19543 2013-04-29 Jakub Jelinek <jakub@redhat.com>
19545 PR tree-optimization/57083
19546 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
19547 non-singleton shift count range, zero extend low_bound for uns case.
19549 * config/i386/predicates.md (general_vector_operand): New predicate.
19550 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
19551 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
19552 if they aren't nonimmediate operands. If their original values
19553 satisfy const_vector_equal_evenodd_p, don't shift them.
19554 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
19555 predicates. For the SSE4.1 case force operands[{1,2}] into registers
19556 if not nonimmediate_operand.
19557 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
19558 instead of register_operand.
19559 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
19561 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
19563 * stor-layout.c (finalize_size_functions): Allocate a structure and
19564 reset cfun before dumping the functions.
19566 2013-04-27 Jakub Jelinek <jakub@redhat.com>
19568 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
19571 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
19572 use xop_pmacsdqh if uns_p.
19573 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
19574 the immediate rotate count.
19576 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
19578 * rtl.h (struct rtx_def): Add comment for field jump.
19579 (LRA_SUBREG_P): New macro.
19580 * recog.c (register_operand): Check LRA_SUBREG_P.
19581 * lra.c (lra): Add note at the end of RTL code. Align non-empty
19583 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
19584 (lra_final_code_change): Skip subreg change for operators.
19585 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
19586 if there are no operand changes.
19587 * lra-constraints.c (curr_insn_set): New.
19588 (match_reload): Set LRA_SUBREG_P.
19589 (emit_spill_move): Ditto.
19590 (check_and_process_move): Use curr_insn_set. Process only single
19591 set insns. Don't initialize sec_mem_p and change_p.
19592 (simplify_operand_subreg): Use LRA_SUBREG_P.
19593 (reg_in_class_p): New function.
19594 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
19595 of #ifdef. Add code to remove cycling.
19596 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
19597 non-null disp. Reload inner instead of disp when base and index
19598 are null. Try to put lo_sum into register.
19599 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
19600 (check_and_process_move): Move code for move cost check to
19601 simple_move_p. Remove equiv_substitution.
19602 (simple_move_p): New function.
19603 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
19604 curr_insn_set. Call check_and_process_move only for single set
19605 insns. Use the new function. Move call of check_and_process_move
19606 after operand equiv substitution and address process.
19608 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19611 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
19612 with nonlocal goto receivers or returns twice calls, ignore
19613 unininitialized values from abnormal edges to nl goto receiver
19614 or returns twice call.
19616 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19618 PR tree-optimization/57051
19619 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
19620 and VEC_RSHIFT_EXPR if shift count is a multiple of element
19623 2013-04-26 Richard Biener <rguenther@suse.de>
19625 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
19626 (expand_omp_taskreg): Likewise. Mark loops for fixup.
19627 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
19628 (fixup_loop_arrays_after_move): New function.
19629 (move_sese_region_to_fn): Properly outline the loop tree parts
19630 of the SESE region.
19632 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
19634 * config/i386/i386.md (type, unit): Fix long lines.
19636 2013-04-26 Richard Biener <rguenther@suse.de>
19638 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
19639 (lto-streamer-out.o): Likewise.
19640 * cfgloop.c (init_loops_structure): Export, add struct function
19641 argument and adjust.
19642 (flow_loops_find): Adjust.
19643 * cfgloop.h (enum loop_estimation): Add EST_LAST.
19644 (init_loops_structure): Declare.
19645 * lto-streamer-in.c: Include cfgloop.h.
19646 (input_cfg): Input the loop tree.
19647 * lto-streamer-out.c: Include cfgloop.h.
19648 (output_cfg): Output the loop tree.
19649 (output_struct_function_base): Do not drop PROP_loops.
19651 2013-03-26 Richard Biener <rguenther@suse.de>
19653 * tree-cfg.c (execute_build_cfg): Build the loop tree.
19654 (pass_build_cfg): Provide PROP_loops.
19655 (move_sese_region_to_fn): Remove loops that are outlined into fn
19657 * tree-inline.c: Include cfgloop.h.
19658 (initialize_cfun): Do not drop PROP_loops.
19659 (copy_loops): New function.
19660 (copy_cfg_body): Copy loop structure.
19661 (tree_function_versioning): Initialize destination loop tree.
19662 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
19663 (pass_parallelize_loops): Do IL verification.
19664 * loop-init.c (loop_optimizer_init): Fixup loops if required.
19665 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
19666 the CFG make sure we fixup loops as well.
19667 * tree-ssa-tail-merge.c: Include cfgloop.h.
19668 (replace_block_by): When merging loop latches mark loops for fixup.
19669 * lto-streamer-out.c (output_struct_function_base): Drop
19670 PROP_loops for now.
19671 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
19672 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
19673 * ipa-split.c: Include cfgloop.h.
19674 (split_function): Add the new return block to the loop tree root.
19675 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
19676 whether we have removed the forwarder block.
19677 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
19678 * cfgloop.h (place_new_loop): Declare.
19679 * cfgloopmanip.c (place_new_loop): Export.
19680 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
19681 (tree-switch-conversion.o): Likewise.
19682 (tree-complex.o): Likewise.
19683 (tree-inline.o): Likewise.
19684 (tree-ssa-tailmerge.o): Likewise.
19685 (ipa-split.o): Likewise.
19686 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
19687 (tree-ssa-copy.o): Likewise.
19688 * tree-switch-conversion.c: Include cfgloop.h
19689 (process_switch): If we emit a bit-test cascade, schedule loops
19691 * tree-complex.c: Include cfgloop.h.
19692 (expand_complex_div_wide): Properly add new basic-blocks to loops.
19693 * asan.c: Include cfgloop.h.
19694 (create_cond_insert_point): Properly add new basic-blocks to
19695 loops, schedule loop fixup.
19696 * cfgloop.c (verify_loop_structure): Check that looks are not
19698 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
19700 (expand_omp_for_generic): Likewise.
19701 (expand_omp_sections): Likewise.
19702 (expand_omp_atomic_pipeline): Schedule loops for fixup.
19703 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
19704 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
19705 is initialized, not when loops are present.
19706 * tree-parloops.c (parallelize_loops): Remove checking here.
19707 * passes.c (init_optimization_passes): Schedule a copy-propagation
19708 pass before complete unrolling of inner loops.
19710 2013-04-26 Jakub Jelinek <jakub@redhat.com>
19712 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
19713 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
19714 (colorize_init): Add argument to _WIN32 version.
19715 * toplev.c: Include diagnostic-color.h.
19716 (process_options): Default to -fdiagnostics-color=auto if
19717 GCC_COLORS env var is in the environment.
19718 * common.opt (fdiagnostics-color=): Add Var and Init.
19719 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
19720 env var is in the environment, the default is auto rather than never.
19722 * diagnostic.h (file_name_as_prefix): Add context argument.
19723 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
19724 the string as locus.
19725 * langhooks.c (lhd_print_error_function): Adjust caller.
19727 2013-04-25 Lawrence Crowl <crowl@google.com>
19729 * var-tracking.c (shared_hash_def::htab):
19730 Change type to hash_table. Update dependent calls and types.
19732 2013-04-25 Lawrence Crowl <crowl@google.com>
19734 * Makefile.in: Update as needed below.
19736 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
19737 Move declaration to after the type's method definitons.
19739 * attribs.c (htab_t scoped_attributes::attribute_hash):
19740 Change type to hash_table. Update dependent calls and types.
19742 * bitmap.c (htab_t bitmap_desc_hash):
19743 Change type to hash_table. Update dependent calls and types.
19745 * cselib.c (htab_t cselib_hash_table):
19746 Change type to hash_table. Update dependent calls and types.
19748 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
19749 (hash_string_slot_node): Move implementation into lto-streamer.h
19750 struct string_slot_hasher.
19751 (eq_string_slot_node): Likewise.
19753 * data-streamer-out.c: Update output_block::string_hash_table
19754 dependent calls and types.
19756 * dwarf2cfi.c (htab_t trace_index):
19757 Change type to hash_table. Update dependent calls and types.
19759 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
19760 Change type to hash_table. Update dependent calls and types.
19761 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
19762 (htab_t optimize_external_refs::map): Likewise.
19763 (htab_t output_comp_unit::extern_map): Likewise.
19764 (htab_t output_comdat_type_unit::extern_map): Likewise.
19765 (htab_t output_macinfo::macinfo_htab): Likewise.
19766 (htab_t optimize_location_lists::htab): Likewise.
19767 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
19769 * except.c (htab_t ehspec_hash_type):
19770 Change type to hash_table. Update dependent calls and types.
19771 (assign_filter_values::ttypes): Likewise.
19772 (assign_filter_values::ehspec): Likewise.
19773 (sjlj_assign_call_site_values::ar_hash): Likewise.
19774 (convert_to_eh_region_ranges::ar_hash): Likewise.
19776 * gcse.c (htab_t pre_ldst_table):
19777 Change type to hash_table. Update dependent calls and types.
19779 * ggc-common.c (htab_t saving_htab):
19780 Change type to hash_table. Update dependent calls and types.
19781 (htab_t loc_hash): Likewise.
19782 (htab_t ptr_hash): Likewise.
19783 (call_count): Rename ggc_call_count.
19784 (call_alloc): Rename ggc_call_alloc.
19785 (loc_descriptor): Rename make_loc_descriptor.
19786 (add_statistics): Rename ggc_add_statistics.
19788 * ggc-common.c (saving_htab):
19789 Change type to hash_table. Update dependent calls and types.
19791 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
19792 (push_gimplify_context): Likewise.
19793 (pop_gimplify_context): Likewise.
19794 (struct gimple_temp_hash_elt): Added.
19795 (struct gimplify_hasher): Likewise.
19796 (struct gimplify_ctx.temp_htab):
19797 Change type to hash_table. Update dependent calls and types.
19799 * gimple-fold.c: Include gimplify-ctx.h.
19801 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
19802 Change type to hash_table. Update dependent calls and types.
19803 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
19804 avoid potential global name collision.
19806 * gimplify.c: Include gimplify-ctx.h.
19807 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
19808 (htab_t gimplify_ctx::temp_htab):
19809 Update dependent calls and types for new type hash_table.
19810 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
19811 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
19813 * gimplify-ctx.h: New.
19814 (struct gimple_temp_hash_elt): Move from gimplify.c.
19815 (class gimplify_hasher): New.
19816 (struct gimplify_ctx): Move from gimple.h.
19817 (htab_t gimplify_ctx::temp_htab):
19818 Change type to hash_table. Update dependent calls and types.
19820 * graphite-clast-to-gimple.c: Include graphite-htab.h.
19821 (htab_t ivs_params::newivs_index):
19822 Change type to hash_table. Update dependent calls and types.
19823 (htab_t ivs_params::params_index): Likewise.
19824 (htab_t print_generated_program::params_index): Likewise.
19825 (htab_t gloog::newivs_index): Likewise.
19826 (htab_t gloog::params_index): Likewise.
19828 * graphite.c: Include graphite-htab.h.
19829 4htab_t graphite_transform_loops::bb_pbb_mapping):
19830 Change type to hash_table. Update dependent calls and types.
19832 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
19833 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
19834 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
19836 * graphite-dependences.c: Include graphite-htab.h.
19837 (loop_is_parallel_p): Change hash table type of parameter.
19839 * graphite-htab.h: New.
19840 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
19841 (extern find_pbb_via_hash): Move from graphite-poly.h.
19842 (extern loop_is_parallel_p): Move from graphite-poly.h.
19843 (extern get_loop_body_pbbs): Move from graphite-poly.h.
19845 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
19846 (extern loop_is_parallel_p): Move to graphite-htab.h.
19847 (extern get_loop_body_pbbs): Move to graphite-htab.h.
19849 * haifa-sched.c (htab_t delay_htab):
19850 Change type to hash_table. Update dependent calls and types.
19851 (htab_t delay_htab_i2): Likewise.
19853 * ira-color.c (htab_t allocno_hard_regs_htab):
19854 Change type to hash_table. Update dependent calls and types.
19856 * ira-costs.c (htab_t cost_classes_htab):
19857 Change type to hash_table. Update dependent calls and types.
19859 * loop-invariant.c (htab_t merge_identical_invariants::eq):
19860 Change type to hash_table. Update dependent calls and types.
19862 * loop-iv.c (htab_t bivs):
19863 Change type to hash_table. Update dependent calls and types.
19865 * loop-unroll.c (htab_t opt_info::insns_to_split):
19866 Change type to hash_table. Update dependent calls and types.
19867 (htab_t opt_info::insns_with_var_to_expand): Likewise.
19869 * lto-streamer.h (struct string_slot): Move from data-streamer.h
19870 (struct string_slot_hasher): New.
19871 (htab_t output_block::string_hash_table):
19872 Change type to hash_table. Update dependent calls and types.
19874 * lto-streamer-in.c (freeing_string_slot_hasher): New.
19875 (htab_t file_name_hash_table):
19876 Change type to hash_table. Update dependent calls and types.
19878 * lto-streamer-out.c: Update output_block::string_hash_table dependent
19881 * lto-streamer.c (htab_t tree_htab):
19882 Change type to hash_table. Update dependent calls and types.
19884 * omp-low.c: Include gimplify-ctx.h.
19886 * passes.c (htab_t name_to_pass_map):
19887 Change type to hash_table. Update dependent calls and types.
19888 (pass_traverse): Rename to passes_pass_traverse.
19890 * plugin.c (htab_t event_tab):
19891 Change type to hash_table. Update dependent calls and types.
19893 * postreload-gcse.c (htab_t expr_table):
19894 Change type to hash_table. Update dependent calls and types.
19895 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
19897 * sese.c (debug_rename_map_1): Make extern.
19898 (htab_t copy_bb_and_scalar_dependences::rename_map):
19899 Change type to hash_table. Update dependent calls and types.
19901 * sese.h (extern debug_rename_map): Move to .c file.
19903 * store-motion.c (htab_t store_motion_mems_table):
19904 Change type to hash_table. Update dependent calls and types.
19906 * trans-mem.c (htab_t tm_new_mem_hash):
19907 Change type to hash_table. Update dependent calls and types.
19909 * tree-browser.c (htab_t TB_up_ht):
19910 Change type to hash_table. Update dependent calls and types.
19912 * tree-cfg.c (htab_t discriminator_per_locus):
19913 Change type to hash_table. Update dependent calls and types.
19915 * tree-complex.c: Include tree-hasher.h
19916 (htab_t complex_variable_components):
19917 Change type to hash_table. Update dependent calls and types.
19919 * tree-eh.c (htab_t finally_tree):
19920 Change type to hash_table. Update dependent calls and types.
19922 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
19923 struct int_tree_hasher.
19924 (extern int_tree_map_eq): Likewise.
19925 (uid_decl_map_hash): Removed.
19926 (extern decl_tree_map_eq): Likewise.
19928 * tree-hasher.h: New.
19929 (struct int_tree_hasher): New.
19930 (typedef int_tree_htab_type): New.
19932 * tree-inline.c: Include gimplify-ctx.h.
19934 * tree-mudflap.c: Include gimplify-ctx.h.
19936 * tree-parloops.c: Include tree-hasher.h.
19937 (htab_t eliminate_local_variables_stmt::decl_address):
19938 Change type to hash_table. Update dependent calls and types.
19939 (htab_t separate_decls_in_region::decl_copies): Likewise.
19941 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
19942 Change type to hash_table. Update dependent calls and types.
19944 * tree-sra.c (candidates):
19945 Change type to hash_table. Update dependent calls and types.
19947 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
19949 (int_tree_map_hash): Likewise.
19951 * tree-ssa-dom.c (htab_t avail_exprs):
19952 Change type to hash_table. Update dependent calls and types.
19954 * tree-ssa-live.c (var_map_base_init::tree_to_index):
19955 Change type to hash_table. Update dependent calls and types.
19957 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
19958 Change type to hash_table. Update dependent calls and types.
19960 * tree-ssa-phiopt.c (seen_ssa_names):
19961 Change type to hash_table. Update dependent calls and types.
19963 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
19964 Change type to hash_table. Update dependent calls and types.
19966 * tree-ssa-uncprop.c (equiv):
19967 Change type to hash_table. Update dependent calls and types.
19969 2013-04-25 Jakub Jelinek <jakub@redhat.com>
19971 PR rtl-optimization/57003
19972 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
19973 call note_stores with kill_clobbered_value callback again after
19974 killing regs_invalidated_by_call.
19976 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19978 * config/aarch64/aarch64-simd.md
19979 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
19980 (aarch64_simd_bsl<mode>): Likewise.
19981 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
19983 2013-04-25 Marek Polacek <polacek@redhat.com>
19985 PR tree-optimization/57066
19986 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
19988 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19990 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
19992 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
19994 * config/aarch64/aarch64-builtins.c
19995 (aarch64_fold_builtin): New.
19996 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
19997 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
19998 * config/aarch64/aarch64-simd-builtins.def (abs): New.
19999 * config/aarch64/arm_neon.h
20000 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
20002 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20003 Tejas Belagod <tejas.belagod@arm.com>
20005 * config/aarch64/aarch64-builtins.c
20006 (aarch64_gimple_fold_builtin): New.
20007 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
20008 * config/aarch64/aarch64-simd-builtins.def (addv): New.
20009 * config/aarch64/aarch64-simd.md (addpv4sf): New.
20010 (addvv4sf): Update.
20011 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
20013 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20015 * config/aarch64/aarch64.md
20016 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
20018 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20020 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
20021 (*ngcsi_uxtw): New pattern.
20023 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20024 Julian Brown <julian@codesourcery.com>
20026 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
20027 (TB_DREG): Add T_V4HF.
20028 (v4hf_UP): New macro.
20029 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20030 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20031 Handle initialisation of V4HF. Adjust initialisation of reinterpret
20033 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
20034 (arm_vector_mode_supported_p): Handle V4HF.
20035 (arm_mangle_map): Handle V4HFmode.
20036 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
20037 * config/arm/arm_neon_builtins.def: Add entries for
20038 vcvtv4hfv4sf, vcvtv4sfv4hf.
20039 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
20040 (neon_vcvtv4hfv4sf): Likewise.
20041 * config/arm/neon-gen.ml: Handle half-precision floating point
20043 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
20044 * config/arm/arm_neon.h: Regenerate.
20045 * config/arm/neon.ml (type elts): Add F16.
20046 (type vectype): Add T_float16x4, T_floatHF.
20047 (type vecmode): Add V4HF.
20048 (type features): Add Requires_FP_bit feature.
20049 (elt_width): Handle F16.
20050 (elt_class): Likewise.
20051 (elt_of_class_width): Likewise.
20052 (mode_of_elt): Refactor.
20053 (type_for_elt): Handle F16, fix error messages.
20054 (vectype_size): Handle T_float16x4.
20055 (vcvt_sh): New function.
20056 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
20057 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
20058 (string_of_mode): Handle V4HF.
20059 * doc/arm-neon-intrinsics.texi: Regenerate.
20061 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
20063 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
20064 format specifier in 'X' case.
20066 2013-04-25 Alan Modra <amodra@gmail.com>
20069 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
20070 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
20071 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
20072 Repeat for many other rotate/shift and mask patterns using subregs.
20073 Name lshiftrt insns.
20074 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
20075 on WORDS_BIG_ENDIAN.
20077 2013-04-25 Alan Modra <amodra@gmail.com>
20079 * config.gcc: Support little-endian powerpc-linux targets.
20080 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
20081 (LINK_OS_LINUX_SPEC): Define.
20082 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
20083 Preserve MASK_LITTLE_ENDIAN.
20084 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
20085 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
20086 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
20087 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
20088 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
20089 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
20090 Correct fp word order for little-endian. Don't shift toc entries
20091 smaller than a word for little-endian.
20092 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
20093 (bswapdi2 splits): Correct low-part subreg for little-endian.
20094 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
20095 low/high where such is correct only for be.
20096 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
20097 little-endian for -mcall-aixdesc.
20099 2013-04-25 Alan Modra <amodra@gmail.com>
20101 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
20102 replace_equiv_address_nv.
20104 2013-04-25 Alan Modra <amodra@gmail.com>
20106 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
20108 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20111 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20112 * rtl.h (struct rtx_def): ...
20114 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20116 PR rtl-optimizations/57046
20117 * lra-constraints (split_reg): Set up lra_risky_transformations_p
20118 for multi-reg splits.
20120 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
20122 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
20124 2013-04-24 Sterling Augustine <saugustine@google.com>
20126 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
20127 (comp_dir_string, debug_str_dwo_section): New.
20128 (DEBUG_STR_DWO_SECTION): Rename to ...
20129 (DEBUG_DWO_STR_SECTION): ... this.
20130 (DEBUG_NORM_STR_SECTION): Delete.
20131 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
20132 (DEBUG_STR_DWO_SECTION_FLAGS): New.
20133 (find_AT_string): Move most logic to ...
20134 (find_AT_string_in_table): ... here. New.
20135 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
20136 add_skeleton_AT_string. Delete logic.
20137 (output_skeleton_debug_sections): Remove call to
20138 add_top_level_skeleton_die_attrs.
20139 (add_comp_dir_attribute): Move logic to comp_dir_string.
20140 (dwarf2out_init): Initialize debug_str_dwo_section.
20141 (output_indirect_string): Call find_string_form.
20142 (output_indirect_strings): Rewrite.
20143 (prune_unused_types): Empty skeleton_debug_str_hash.
20144 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
20145 (dwarf2out_finish): Call output_indirect_strings.
20147 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20149 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
20151 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
20153 * rtl.h (struct rtx_def): Add comment for field jump.
20154 (LRA_SUBREG_P): New macro.
20155 * recog.c (register_operand): Check LRA_SUBREG_P.
20156 * lra.c (lra): Add note at the end of RTL code. Align non-empty
20158 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
20159 (lra_final_code_change): Skip subreg change for operators.
20160 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
20161 if there are no operand changes.
20162 * lra-constraints.c (curr_insn_set): New.
20163 (match_reload): Set LRA_SUBREG_P.
20164 (emit_spill_move): Ditto.
20165 (check_and_process_move): Use curr_insn_set. Process only single
20166 set insns. Don't initialize sec_mem_p and change_p.
20167 (simplify_operand_subreg): Use LRA_SUBREG_P.
20168 (reg_in_class_p): New function.
20169 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
20170 of #ifdef. Add code to remove cycling.
20171 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
20172 non-null disp. Reload inner instead of disp when base and index
20173 are null. Try to put lo_sum into register.
20174 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
20175 (check_and_process_move): Move code for move cost check to
20176 simple_move_p. Remove equiv_substitution.
20177 (simple_move_p): New function.
20178 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
20179 curr_insn_set. Call check_and_process_move only for single set
20180 insns. Use the new function. Move call of check_and_process_move
20181 after operand equiv substitution and address process.
20183 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
20185 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
20186 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
20187 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
20189 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
20191 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
20193 2013-04-24 Marek Polacek <polacek@redhat.com>
20195 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
20196 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
20197 (select_loops_exit_conditions): Likewise.
20198 (number_of_iterations_for_all_loops): Likewise.
20199 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
20200 (scev_analysis): Likewise.
20202 2013-04-02 Catherine Moore <clm@codesourcery.com>
20203 Chao-ying Fu <fu@mips.com>
20205 * config/mips/micromips.md (jraddiusp): New pattern.
20206 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
20207 instruction if possible.
20209 2013-04-24 Alan Modra <amodra@gmail.com>
20211 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
20213 2013-04-24 Julian Brown <julian@codesourcery.com>
20214 Chung-Lin Tang <cltang@codesourcery.com>
20216 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
20217 dependency behavior in enumeration type DIE generation. Add TODO note
20218 to comments about future DW_FORM_sdata/udata re-work of related code.
20220 2013-04-23 Lawrence Crowl <crowl@google.com>
20222 * Makefile.in: Update as needed below.
20224 * hash-table.h (class hash_table):
20225 Correct many methods with parameter types compare_type to the correct
20226 value_type. (Correct code was unlikely to notice the change.)
20227 (hash_table::elements_with_deleted) New.
20228 (class hashtable::iterator): New.
20229 (hashtable::begin()): New.
20230 (hashtable::end()): New.
20231 (FOR_EACH_HASH_TABLE_ELEMENT): New.
20233 * statistics.c (statistics_hashes):
20234 Change type to hash_table. Update dependent calls and types.
20236 * tree-into-ssa.c (var_infos):
20237 Change type to hash_table. Update dependent calls and types.
20239 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
20240 Change type to hash_table. Update dependent calls and types.
20242 * tree-ssa-loop-im.c (struct mem_ref.refs):
20243 Change type to hash_table. Update dependent calls and types.
20245 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
20246 Change type to hash_table. Update dependent calls and types.
20248 * tree-ssa-sccvn.c (vn_tables_s::nary):
20249 Change type to hash_table. Update dependent calls and types.
20250 (vn_tables_s::phis): Likewise.
20251 (vn_tables_s::references): Likewise.
20253 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
20254 (vn_reference_eq): Update parameter and return types.
20256 * tree-ssa-structalias.c (pointer_equiv_class_table):
20257 Change type to hash_table. Update dependent calls and types.
20258 (location_equiv_class_table): Likewise.
20260 * tree-vect-data-refs.c: Consequential changes for making
20261 peeling a hash_table.
20263 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
20264 (destroy_loop_vec_info): Dependent hash_table update.
20266 * tree-vectorizer.h (peeling_htab):
20267 Change type to hash_table. Update dependent calls and types.
20269 2013-04-23 Shiva Chen <shiva0217@gmail.com>
20271 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
20272 to check the register content is equal or not.
20273 * lra-constraints.c (match_reload): Use lra_assign_reg_val
20274 to assign register content record.
20275 * lra-eliminations.c (update_reg_eliminate): Use
20276 lra_update_reg_val_offset to update register content offset.
20277 * lra-int.h (struct lra_reg): Add offset member.
20278 (lra_reg_val_equal_p): New static inline function.
20279 (lra_update_reg_val_offset): New static inline function.
20280 (lra_assign_reg_val): New static inline function.
20281 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
20282 to assign register content record.
20283 (initialize_lra_reg_info_element): Initial offset to zero.
20285 2013-04-23 Catherine Moore <clm@codesourcery.com>
20287 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
20288 operands. Record compression.
20290 2013-04-23 Xinliang David Li <davidxl@google.com>
20292 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
20294 2013-04-23 Richard Biener <rguenther@suse.de>
20296 PR middle-end/57036
20297 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
20298 parameter, only add abnormal goto edges from the copied body
20299 if the call could perform abnormal gotos.
20300 (copy_cfg_body): Adjust.
20302 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
20304 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
20306 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
20308 * coretypes.h (gimple_stmt_iterator): Add struct to make
20311 2013-04-23 Richard Biener <rguenther@suse.de>
20313 PR tree-optimization/57026
20314 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
20315 from SSA names occuring in abnormal PHI nodes.
20317 2013-04-22 Andi Kleen <ak@linux.intel.com>
20319 * lto/lto.c (print_lto_report_1): Fix LTO report names.
20321 2013-04-22 Andi Kleen <ak@linux.intel.com>
20323 * lto/lto.c (print_lto_report_1): Declare early.
20324 (read_cgraph_and_symbols): Call print_lto_report_1 early.
20326 2013-04-22 Andi Kleen <ak@linux.intel.com>
20328 * common.opt (-flto-report-wpa): Add.
20329 * doc/invoke.texi (-flto-report-wpa): Add.
20330 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
20333 2013-04-22 Xinliang David Li <davidxl@google.com>
20335 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
20336 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
20337 * Makefile.in: New dependency
20339 David Daney <ddaney.cavm@gmail.com>
20341 * configure.ac (gcc_cv_as_micromips_support): Use the
20342 --fatal-warnings option.
20343 * configure: Regenerate.
20345 2013-04-22 Marek Polacek <polacek@redhat.com>
20348 * tsan.c (instrument_expr): Don't instrument expression
20349 in case its size is zero.
20351 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
20355 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
20357 * config/alpha/alpha.c (TARGET_LRA_P): New define.
20359 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20361 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
20362 (gimple_stmt_iterator): New typedef.
20363 * gimple.h (gimple_stmt_iterator): Rename to...
20364 (gimple_stmt_iterator_d): ... This.
20365 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
20366 trees be valid for GIMPLE and GENERIC.
20367 (TARGET_GIMPLE_FOLD_BUILTIN): New.
20368 * gimple-fold.c (gimple_fold_call): Call target hook
20369 gimple_fold_builtin.
20370 * hooks.c (hook_bool_gsiptr_false): New.
20371 * hooks.h (hook_bool_gsiptr_false): New.
20372 * target.def (fold_stmt): New.
20373 * doc/tm.texi: Regenerate.
20375 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
20378 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
20379 a set sp if no stack realignment.
20381 2013-04-22 Nick Clifton <nickc@redhat.com>
20383 * config.gcc (tilegx-linux): Extend extra_objs rather than
20385 (tilepro-linux): Likewise.
20387 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20389 * config/aarch64/aarch64-builtins.c
20391 (CF0, CF1, CF2, CF3, CF4, CF10): New.
20392 (VAR<1-12>): Add MAP parameter.
20393 (BUILTIN_*): Likewise.
20394 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
20395 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
20396 (aarch64_ushl_n<mode>): Likewise.
20397 (aarch64_sshr_n<mode>): Likewise.
20398 (aarch64_ushr_n<mode>): Likewise.
20399 (aarch64_<maxmin><mode>): Likewise.
20400 (aarch64_sqrt<mode>): Likewise.
20401 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
20402 (vshr<q>_n_*): Likewise.
20404 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
20406 * config/aarch64/aarch64-builtins.c
20407 (aarch64_simd_builtin_type_mode): Handle SF types.
20409 (BUILTIN_GPF): Define.
20410 (aarch64_init_simd_builtins): Handle SF types.
20411 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
20412 (frecps): Likewise.
20413 (frecpx): Likewise.
20414 * config/aarch64/aarch64-simd.md
20415 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
20416 (aarch64_frecpe<mode>): New.
20417 (aarch64_frecps<mode>): Likewise.
20418 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
20419 (v8type): Add frecp<esx>.
20420 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
20421 (aarch64_frecps<mode>): Likewise.
20422 * config/aarch64/iterators.md (FRECP): New.
20423 (frecp_suffix): Likewise.
20424 * config/aarch64/arm_neon.h
20425 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
20427 2013-04-22 Christian Bruel <christian.bruel@st.com>
20430 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
20431 (REG_CLASS_NAMES): Idem.
20432 (REG_CLASS_CONTENTS): Idem.
20433 (REGCLASS_HAS_FP_REG): Idem.
20434 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
20435 (sh_conditional_register_usage): Idem.
20437 2013-04-21 Jeff Law <law@redhat.com>
20439 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
20440 (ssa_forward_propagate_and_combine): Use it.
20442 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20444 * lra.c: Update the flow chart diagram.
20446 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
20448 PR rtl-optimization/56847
20449 * lra-constraints.c (process_alt_operands): Discourage alternative
20450 with non-matche doffsettable memory constraint fro memory with
20453 2013-04-19 Richard Biener <rguenther@suse.de>
20455 PR tree-optimization/56982
20456 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
20458 * gimplify.c (gimplify_call_expr): Notice special calls.
20459 (gimplify_modify_expr): Likewise.
20460 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
20461 abnormal control flow receivers.
20462 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
20463 in the same way as cfun->has_nonlocal_labels.
20464 (gimple_purge_dead_abnormal_call_edges): Likewise.
20465 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
20466 receivers start a basic-block.
20468 2013-04-19 Richard Biener <rguenther@suse.de>
20470 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
20472 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
20473 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
20474 (SLP_TREE_LOAD_PERMUTATION): Add.
20475 (vect_transform_slp_perm_load): Adjust prototype.
20476 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
20477 (vect_free_slp_instance): Likewise.
20478 (vect_create_new_slp_node): Likewise.
20479 (vect_supported_slp_permutation_p): Remove.
20480 (vect_slp_rearrange_stmts): Adjust.
20481 (vect_supported_load_permutation_p): Likewise. Inline
20482 vect_supported_slp_permutation_p here.
20483 (vect_analyze_slp_instance): Compute load permutations per
20484 slp node instead of per instance.
20485 (vect_get_slp_defs): Adjust.
20486 (vect_transform_slp_perm_load): Likewise.
20487 (vect_schedule_slp_instance): Remove redundant code.
20488 (vect_schedule_slp): Remove hack for PR56270, add it ...
20489 * tree-vect-stmts.c (vectorizable_load): ... here, do not
20490 CSE loads for SLP. Adjust.
20492 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20494 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
20495 spelling in two comments.
20497 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
20500 * config/arm/arm.c (load_multiple_sequence): Require SP
20501 as base register for loads if SP is in the register list.
20503 2013-04-19 Martin Jambor <mjambor@suse.cz>
20505 PR tree-optimization/56718
20506 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
20507 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
20508 and made public. Adjusted all callers.
20509 (ipa_intraprocedural_devirtualization): New function.
20510 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
20511 (ipa_intraprocedural_devirtualization): Likewise.
20512 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
20514 2013-04-19 Richard Biener <rguenther@suse.de>
20516 PR tree-optimization/57000
20517 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
20519 2013-04-19 Terry Guo <terry.guo@arm.com>
20521 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
20523 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
20524 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
20525 (cortex_m4_fmacs): Use new reservations.
20526 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
20528 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
20530 PR rtl-optimization/56999
20531 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
20533 (lra_coalesce): Remove split_origin_bitmap and related code.
20534 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
20535 ranges if necessary.
20537 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
20539 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
20541 (ix86_expand_call): Remove clobbered_registers array and use
20542 x86_64_ms_sysv_extra_clobbered_registers instead.
20543 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
20545 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
20547 * config/i386/i386.md (*call_rex64_ms_sysv): Use
20548 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
20549 (*call_value_rex64_ms_sysv): Ditto.
20551 2013-04-18 Cary Coutant <ccoutant@google.com>
20553 * dwarf2out.c (output_pubnames): Check die_perennial_p of
20554 parent instead of die_mark.
20556 2013-04-18 Diego Novillo <dnovillo@google.com>
20558 * gimple.c (create_gimple_tmp): New.
20559 (get_expr_type): New.
20560 (build_assign): New.
20561 (build_type_cast): New.
20562 * gimple.h (enum ssa_mode): Define.
20563 (gimple_seq_set_location): New.
20564 * asan.c (build_check_stmt): Change some gimple_build_* calls
20565 to use build_assign and build_type_cast.
20567 2013-04-18 Richard Biener <rguenther@suse.de>
20569 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
20570 handle negative step. Remove redundant checks.
20571 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
20572 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
20573 for negative step and grouped loads fail to vectorize.
20575 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20577 * emit-rtl.c (reset_insn_used_flags): New function.
20578 (reset_all_used_flags): Use it.
20579 (verify_insn_sharing): New function.
20580 (verify_rtl_sharing): Fix verification for SEQUENCEs.
20582 2013-04-18 Jakub Jelinek <jakub@redhat.com>
20584 PR tree-optimization/56984
20585 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
20586 and (x >> M) >= N don't register any assertion if N << M is the
20589 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
20591 * lower-subreg.c (resolve_simple_move): If called self-recursive,
20592 do not delete_insn insns that have not yet been emitted, only
20593 unlink them with remove_insn.
20594 * df-scan.c (df_insn_delete): Revert r197492.
20596 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
20598 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
20599 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
20601 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20603 * config/arm/arm.md (movsicc_insn): Convert define_insn into
20604 define_insn_and_split.
20605 (and_scc,ior_scc,negscc): Likewise.
20606 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
20608 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
20610 * config/arm/arm.c (use_return_insn): Return 0 for targets that
20611 can benefit from using a sequence of LDRD instructions in epilogue
20612 instead of a single LDM instruction.
20614 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
20617 * doc/extend.texi: Fix typo.
20619 2013-04-17 Richard Biener <rguenther@suse.de>
20621 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
20622 (vect_build_slp_tree): ... here.
20623 (vect_build_slp_tree_1): Compute which stmts of the SLP group
20624 match. Remove special-casing of mismatched complex loads.
20625 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
20626 re-try the match with swapped commutative operands.
20627 (vect_supported_load_permutation_p): Remove special-casing of
20628 mismatched complex loads.
20629 (vect_analyze_slp_instance): Adjust.
20631 2013-04-17 Richard Biener <rguenther@suse.de>
20633 PR rtl-optimization/56921
20634 * cfgloop.h (struct loop): Add simple_loop_desc member.
20635 (struct niter_desc): Mark with GTY(()).
20636 (simple_loop_desc): Do not use aux field but simple_loop_desc.
20637 * loop-iv.c (get_simple_loop_desc): Likewise.
20638 (free_simple_loop_desc): Likewise.
20641 2013-04-16 Richard Biener <rguenther@suse.de>
20643 PR rtl-optimization/56921
20644 * loop-init.c (pass_rtl_move_loop_invariants): Add
20645 TODO_do_not_ggc_collect to todo_flags_finish.
20646 (pass_rtl_unswitch): Same.
20647 (pass_rtl_unroll_and_peel_loops): Same.
20648 (pass_rtl_doloop): Same.
20650 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
20652 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
20653 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
20654 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
20656 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
20657 * tree-streamer.c (record_common_node): Adjust reference in comment.
20659 2013-04-17 Terry Guo <terry.guo@arm.com>
20661 * config/arm/cortex-m4.md: Add a new bypass.
20663 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20665 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
20667 (*subs_<optab><mode>_multp2): New pattern.
20668 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20669 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
20671 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20673 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
20674 (*subs_mul_imm_<mode>): New pattern.
20676 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
20679 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
20680 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
20681 (vsx_movti_32bit): Change j->wa to O->wa.
20683 2013-04-16 Richard Biener <rguenther@suse.de>
20685 PR rtl-optimization/56921
20686 * loop-init.c (pass_rtl_move_loop_invariants): Add
20687 TODO_do_not_ggc_collect to todo_flags_finish.
20688 (pass_rtl_unswitch): Same.
20689 (pass_rtl_unroll_and_peel_loops): Same.
20690 (pass_rtl_doloop): Same.
20692 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
20694 * config/arm/arm.c (emit_multi_reg_push): New declaration
20695 for an existing function.
20696 (arm_emit_strd_push): New function.
20697 (arm_expand_prologue): Used here.
20698 (arm_emit_ldrd_pop): New function.
20699 (arm_expand_epilogue): Used here.
20700 (arm_get_frame_offsets): Update condition.
20701 (arm_emit_multi_reg_pop): Add a special case for load of a single
20702 register with writeback.
20704 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
20706 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
20709 2013-04-16 Richard Biener <rguenther@suse.de>
20711 PR tree-optimization/56756
20712 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
20713 (first_mem_ref_loc): New.
20714 (execute_sm): Place the load temporarily before a previous
20715 access instead of in the latch edge to ensure its SSA dependencies
20716 are defined at points dominating the load.
20718 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
20720 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
20721 correct fix by moving header and footer insn to the footer of
20722 the merged basic block. Clear BB_END of the merged-away block.
20724 PR middle-end/43631
20725 * emit-rtl.c (make_note_raw): New function.
20726 (link_insn_into_chain): New static inline function.
20727 (add_insn): Use it.
20728 (add_insn_before, add_insn_after): Factor insn chain linking code...
20729 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
20730 using link_insn_into_chain.
20731 (note_outside_basic_block_p): New helper function for emit_note_after
20732 and emit_note_before.
20733 (emit_note_after): Use nobb variant of add_insn_after if the note
20734 should not be contained in a basic block.
20735 (emit_note_before): Use nobb variant of add_insn_before if the note
20736 should not be contained in a basic block.
20737 (emit_note_copy): Use make_note_raw.
20738 (emit_note): Likewise.
20739 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
20740 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
20741 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
20742 the moved barrier the tail of the basic block it follows.
20743 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
20745 2013-04-15 Jakub Jelinek <jakub@redhat.com>
20747 PR tree-optimization/56962
20748 * gimple-ssa-strength-reduction.c (record_increment): Only set
20749 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
20750 either rhs1 or rhs2 is equal to c->base_expr.
20752 2013-04-15 Richard Biener <rguenther@suse.de>
20754 PR tree-optimization/56933
20755 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
20757 (GROUP_READ_WRITE_DEPENDENCE): Remove.
20758 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
20759 * tree-vect-data-refs.c (vect_analyze_group_access): Move
20760 dependence check ...
20761 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
20763 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
20764 GROUP_READ_WRITE_DEPENDENCE.
20766 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20768 * emit-rtl.c (reset_all_used_flags): New function.
20769 (verify_rtl_sharing): Call reset_all_used_flags before and after
20770 performing the checks.
20772 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20774 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
20775 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
20776 * config/arm/constraints.md (De): New constraint.
20777 * config/arm/neon.md (anddi3_neon): Delete.
20778 (neon_vand<mode>): Expand to standard anddi3 pattern.
20779 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
20780 Move earlier in the file.
20781 (neon_inv_logic_op2): Likewise.
20782 (arm_anddi_operand_neon): New predicate.
20784 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20786 * configure.ac (gcc_cv_ld_as_needed): Set
20787 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
20788 Use -z ignore, -z record on *-*-solaris2*.
20789 (HAVE_LD_AS_NEEDED): Update comment.
20790 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
20791 * configure: Regenerate.
20792 * config.in: Regenerate.
20793 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
20794 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
20795 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
20796 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
20797 equivalents. Fix markup.
20798 * doc/tm.texi: Regenerate.
20800 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
20802 * config/i386/i386.opt: New option mstack-protector-guard=.
20803 * config/i386/i386-opts.h: Add enum stack_protector_guard.
20804 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
20805 TARGET_SSP_TLS_GUARD.
20806 * config/i386/i386.c (ix86_option_override_internal): Set
20807 ix86_stack_protector_guard.
20808 * config/i386/i386.md (stack_protect_set): Enable for
20809 TARGET_SSP_TLS_GUARD only.
20810 (stack_protect_set_<mode>): Ditto.
20811 (stack_protect_test): Ditto.
20812 (stack_protect_test_<mode>): Ditto.
20813 * doc/invoke.texi (i386 Option): Document.
20815 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
20818 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
20819 (S_MODES): Set H_MODE bit.
20820 (SF_MODES): Set only S_MODE and SF_MODE bits.
20821 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
20822 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
20823 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
20824 <MODE_FLOAT>: Likewise.
20826 2013-04-15 Joey Ye <joey.ye@arm.com>
20828 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
20830 2013-04-15 Joey Ye <joey.ye@arm.com>
20832 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
20834 (thumb_far_jump_used_p): Count instruction size and set
20837 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
20839 * reorg.c (fill_simple_delay_slots): Reindent block of code.
20840 * resource.c (mark_target_live_regs): Reformat conditional block.
20842 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20844 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
20845 notes, they are emitted only just before final.
20846 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
20848 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
20850 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
20851 * cfgrtl.c (delete_insn): Call it here instead.
20852 * lra-spills.c (lra_final_code_change): Use delete_insn.
20853 * haifa-sched.c (sched_remove_insn): Likewise.
20854 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
20855 returning to the nop pool.
20856 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
20857 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
20859 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20861 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
20862 * doc/tm.texi: Regenerated.
20864 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
20866 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
20869 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
20871 * df-core.c (df_find_def): Compare register numbers.
20872 (df_find_use): Likewise.
20874 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
20877 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
20878 lra_in_progress for return.
20880 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20882 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
20883 define_insn into define_insn_and_split and emit movsicc patterns.
20885 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
20887 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
20889 2013-04-12 Richard Biener <rguenther@suse.de>
20891 * tree-pass.h (TODO_do_not_ggc_collect): New.
20892 * passes.c (execute_one_ipa_transform_pass): Honor
20893 TODO_do_not_ggc_collect.
20894 (execute_one_pass): Likewise.
20897 2013-04-10 Richard Biener <rguenther@suse.de>
20899 * passes.c (init_optimization_passes): Remove reload pass.
20900 * ira.c (do_reload): Merge into ...
20902 (rest_of_handle_reload): Remove.
20903 (pass_reload): Likewise.
20904 * config/i386/i386.c (ix86_option_override): Refer to ira instead
20905 of reload for vzeroupper pass placement.
20907 2013-04-12 Jakub Jelinek <jakub@redhat.com>
20909 PR tree-optimization/56918
20910 PR tree-optimization/56920
20911 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
20912 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
20913 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
20914 use wide_mul_with_sign method.
20916 2013-04-12 Richard Biener <rguenther@suse.de>
20918 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
20919 not be considered a gimple constant.
20921 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
20923 * fold-const.c (const_binop): Handle vector shifts by a scalar.
20924 (fold_binary_loc): Call const_binop also for mixed vector-scalar
20927 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
20928 Jakub Jelinek <jakub@redhat.com>
20930 * opts.c: Include diagnostic-color.h.
20931 (common_handle_option): Handle OPT_fdiagnostics_color_.
20932 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
20933 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
20934 (diagnostic-color.o): New.
20935 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
20936 (diagnostic_color_rule): New enum.
20937 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
20938 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
20939 the location string.
20940 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
20941 either NULL, or color kind.
20942 * diagnostic-color.c: New file.
20943 * diagnostic-color.h: New file.
20944 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
20946 * doc/invoke.texi (-fdiagnostics-color): Document.
20947 * pretty-print.h (pp_show_color): Define.
20948 (struct pretty_print_info): Add show_color field.
20949 * diagnostic.c: Include diagnostic-color.h.
20950 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
20951 macros. Colorize error:, warning: etc. strings and also the location
20953 (diagnostic_show_locus): Colorize the caret line.
20954 * pretty-print.c: Include diagnostic-color.h.
20955 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
20956 inside of %< %> quotes or quoted through q format modifier.
20958 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20960 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
20962 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20964 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
20965 code in CC_NZ mode.
20966 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
20969 2013-04-11 Marek Polacek <polacek@redhat.com>
20971 PR tree-optimization/48184
20972 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
20974 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
20976 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
20977 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
20978 (skip_simple_arithmetic): Tidy up.
20979 * tree.h (skip_simple_constant_arithmetic): Declare.
20981 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20983 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
20985 2013-04-11 Richard Biener <rguenther@suse.de>
20987 * tree-vect-loop.c (get_initial_def_for_induction): Properly
20988 generate vector constants.
20990 2013-04-11 Richard Biener <rguenther@suse.de>
20992 PR tree-optimization/56878
20993 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
20994 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
20996 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
20997 Prefer to align the DR with the most invariant base address.
20999 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
21001 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
21004 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
21006 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
21007 floating-point vector comparisons against 0.
21009 2013-04-11 Jakub Jelinek <jakub@redhat.com>
21011 PR tree-optimization/56899
21012 * fold-const.c (extract_muldiv_1): Apply distributive law
21013 only if TYPE_OVERFLOW_WRAPS (ctype).
21015 2013-04-11 Bin Cheng <bin.cheng@arm.com>
21018 * ira-costs.c (scan_one_insn): Check whether the source rtx of
21019 loading has side effect.
21021 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
21023 * config/sparc/sparc.c: Include tree-pass.h.
21024 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
21025 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
21026 head of file. Change return type. Split off gate function.
21027 (sparc_gate_work_around_errata): New function.
21028 (pass_work_around_errata): New pass definition.
21029 (insert_pass_work_around_errata) New pass insert definition to
21030 insert pass_work_around_errata just after delayed-branch scheduling.
21031 (sparc_option_override): Insert the pass.
21032 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
21034 2013-04-10 David S. Miller <davem@davemloft.net>
21036 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
21037 or -mcpu=hypersparc.
21039 * target.def (cstore_mode): New hook.
21040 * target.h: Include insn-codes.h
21041 * targhooks.c: Likewise.
21042 (default_cstore_mode): New function.
21043 * targhooks.h: Declare it.
21044 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
21045 * doc/tm.texi: Rebuild.
21046 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
21047 target hook, rather than inspecting the insn_data.
21048 * config/sparc/sparc.c (sparc_cstore_mode): New function.
21049 (TARGET_CSTORE_MODE): Redefine.
21050 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
21052 * config/sparc/predicates.md (cstore_result_operand): New special
21054 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
21055 Use it for operand 0.
21056 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
21057 (*snesi_special): Likewise.
21058 (*snesi_zero): Likewise.
21059 (*seqsi_zero): Likewise.
21060 (*sltu_insn): Likewise.
21061 (*sgeu_insn): Likewise.
21062 (*seqdi_special): Make operand 0 and comparison operation be of
21064 (*snedi_special): Likewise.
21065 (*snedi_special_vis3): Likewise.
21066 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
21067 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
21068 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
21069 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
21070 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
21071 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
21072 (*sltu_extend_sp64): Likewise.
21073 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
21074 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
21075 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
21076 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
21077 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
21079 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
21081 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
21082 (aarch64_start_file): Use the new function.
21084 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
21085 Jason Merrill <jason@redhat.com>
21087 * common.opt: Add -gdwarf.
21088 * opts.c (common_handle_option): Handle it.
21089 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
21091 2013-04-10 Richard Biener <rguenther@suse.de>
21093 * passes.c (execute_todo): Do not call ggc_collect conditional here.
21094 (execute_one_ipa_transform_pass): But unconditionally here.
21095 (execute_one_pass): And here.
21096 (init_optimization_passes): Remove reload pass.
21097 * tree-pass.h (TODO_ggc_collect): Remove.
21098 (pass_reload): Likewise.
21099 * ira.c (do_reload): Merge into ...
21101 (rest_of_handle_reload): Remove.
21102 (pass_reload): Likewise.
21103 * config/i386/i386.c (ix86_option_override): Refer to ira instead
21104 of reload for vzeroupper pass placement.
21105 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
21106 and todo_flags_finish of all passes.
21108 2013-04-10 Richard Biener <rguenther@suse.de>
21110 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
21111 first_const_oprnd field, rename first_def_type to first_op_type.
21112 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
21113 (vect_get_and_check_slp_defs): Always use the type of the
21114 operand. Allow mixed vect_external_def, vect_constant_def types.
21115 (vect_get_constant_vectors): Handle mixed vect_external_def,
21116 vect_constant_def types.
21118 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
21120 PR tree-optimization/55524
21121 * tree-ssa-math-opts.c
21122 (convert_mult_to_fma): Don't use an fms construct
21123 when we don't have an fms operation, but fnma, and it looks
21124 likely that we'll be able to use the latter.
21126 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
21128 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
21130 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
21131 inline fail caused by overwritable functions.
21133 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
21135 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
21136 unnecessary bits in the constant power of two case.
21138 2013-04-10 Richard Biener <rguenther@suse.de>
21140 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
21141 broken code swapping operands.
21142 (vect_build_slp_tree): Do not compute load permutations here.
21143 (vect_analyze_slp_instance): Compute load permutations here,
21144 after building the SLP tree.
21146 2013-04-09 Christian Bruel <christian.bruel@st.com>
21148 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
21149 of next/prev_real_insn.
21151 2013-04-09 Jan Hubicka <jh@suse.cz>
21153 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
21154 Drop aliased parameter.
21155 (function_and_variable_visibility): Do not handle alias pairs.
21156 * cgraph.c (varpool_externally_visible_p): Update prototype.
21157 * varpool.c (varpool_add_new_variable): Update.
21159 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21161 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
21163 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
21165 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
21167 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
21169 2013-04-09 Marek Polacek <polacek@redhat.com>
21171 PR tree-optimization/48762
21172 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
21174 2013-04-09 Richard Biener <rguenther@suse.de>
21176 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
21178 (vect_build_slp_tree): Likewise.
21179 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
21180 calculating the cost of a SLP instance.
21181 (vect_analyze_slp_instance): Use it from here, after building
21184 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21186 PR middle-end/56883
21187 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
21188 expand_omp_for_static_chunk): Use simple_p = true in
21189 force_gimple_operand_gsi calls when assigning to addressable decls.
21191 2013-04-09 Jeff Law <law@redhat.com>
21193 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
21194 when the boolean was created by converting a wider object which
21195 had a boolean range.
21197 2013-04-09 Richard Biener <rguenther@suse.de>
21199 * tree-vectorizer.h (slp_void_p): Remove.
21200 (slp_tree): Typedef before _slp_tree declaration.
21201 (struct _slp_tree): Use a vector of slp_tree as children.
21202 (vect_get_place_in_interleaving_chain): Remove.
21203 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
21205 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
21207 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
21208 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
21209 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
21210 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
21211 Use slp_node instead of slp_void_p and adjust.
21213 2013-04-09 Richard Biener <rguenther@suse.de>
21215 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
21216 work that is not necessary.
21218 2013-04-09 Jakub Jelinek <jakub@redhat.com>
21220 PR tree-optimization/56854
21221 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
21222 forward into clobber stmts if it would change MEM_REF lhs into
21225 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
21227 * tree.c (type_hash_lookup, type_hash_add): Make static.
21228 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
21230 2013-04-09 Richard Biener <rguenther@suse.de>
21232 * tree.h (unsave_expr_now): Remove.
21233 * tree-inline.c (mark_local_for_remap_r): Remove.
21234 (unsave_expr_1): Likewise.
21235 (unsave_r): Likewise.
21236 (unsave_expr_now): Likewise.
21237 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
21238 (propagate_tree_value): Likewise.
21240 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
21242 * doc/rtl.texi (sequence): Rewrite documentation to match the
21243 current use of SEQUENCE rtl objects.
21244 * rtl.def (SEQUENCE): Likewise.
21246 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
21247 Update documentation.
21248 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
21249 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
21251 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
21253 2013-04-08 Teresa Johnson <tejohnson@google.com>
21255 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
21256 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
21258 (estimate_edge_size_and_time): Add comment to suggest using rounding
21260 (estimate_node_size_and_time): Ditto.
21261 (remap_edge_change_prob): Use helper rounding divide methods.
21262 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
21263 (gimple_mod_pow2_value_transform): Ditto.
21264 (gimple_mod_subtract_transform): Ditto.
21265 (gimple_ic_transform): Ditto.
21266 (gimple_stringops_transform): Ditto.
21267 * stmt.c (conditional_probability): Ditto.
21268 (emit_case_dispatch_table): Ditto.
21269 * lto-cgraph.c (merge_profile_summaries): Ditto.
21270 * tree-optimize.c (execute_fixup_cfg): Ditto.
21271 * cfgcleanup.c (try_forward_edges): Ditto.
21272 * cfgloopmanip.c (scale_loop_profile): Ditto.
21274 (duplicate_loop_to_header_edge): Ditto.
21275 (lv_adjust_loop_entry_edge): Ditto.
21276 * tree-vect-loop.c (vect_transform_loop): Ditto.
21277 * profile.c (compute_branch_probabilities): Ditto.
21278 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
21279 * lto-streamer-in.c (input_cfg): Ditto.
21280 * gimple-streamer-in.c (input_bb): Ditto.
21281 * ipa-cp.c (update_profiling_info): Ditto.
21282 (update_specialized_profile): Ditto.
21283 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
21284 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
21286 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21287 (compute_trg_info): Ditto.
21288 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
21289 (purge_dead_edges): Ditto.
21290 * loop-unswitch.c (unswitch_loop): Ditto.
21291 * cgraphclones.c (cgraph_clone_edge): Ditto.
21292 (cgraph_clone_node): Ditto.
21293 * tree-inline.c (copy_bb): Ditto.
21294 (copy_edges_for_bb): Ditto.
21295 (initialize_cfun): Ditto.
21296 (copy_cfg_body): Ditto.
21297 (expand_call_inline): Ditto.
21299 2013-04-08 Kai Tietz <ktietz@redhat.com>
21301 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
21302 TARGET_CYGWIN64 by TARGET_64BIT.
21304 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21306 * config/epiphany/epiphany.md (GPR_1): New constant.
21307 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
21308 * config/epiphany/epiphany.c (gen_compare_reg):
21309 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
21310 is already in place.
21311 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
21312 Don't require being called during rtl expansion; If y operlaps r0,
21314 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
21315 (epiphany_expand_epilogue): Likewise.
21317 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
21318 Don't use CC_FPmode for ORDERED / UNORDERED.
21319 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
21321 * config/epiphany/constraints.md (CnL): New constraint.
21322 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
21323 * config/epiphany/predicates.md (add_operand): Allow 1024.
21325 * config/epiphany/epiphany.md (logical_op): New code iterator.
21326 (op_mnc): New code attribute.
21327 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
21328 (mov_f+1, mov_f+2): New peephole2 patterns.
21330 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
21331 (cstoresi4): Also allow re-use of zero result when doing a NE
21332 comparison to a non-zero operand.
21333 Use (clobber (scratch)) for first insn if the gpr output is not needed.
21335 * config/epiphany/epiphany.md (<insn_opname>v2si3):
21336 Use gen_addsi3_i / gen_subsi3_i.
21338 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21342 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
21343 contain anything but clobbers, at most one __builtin_stack_restore,
21344 optionally debug stmts and final resx, and if it has at least one
21345 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
21346 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
21347 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
21348 which isn't defaut definition, remove them.
21349 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
21350 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
21351 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
21352 with MEM_REF LHS with SSA_NAME address.
21354 2013-04-08 Jeff Law <law@redhat.com>
21356 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
21358 2013-04-08 Richard Biener <rguenther@suse.de>
21360 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
21362 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
21363 determined vector type.
21364 (vect_analyze_data_refs): Likewise.
21365 (vect_get_new_vect_var): Adjust.
21366 (vect_create_destination_var): Preserve SSA name versions.
21367 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
21368 not dump anything here.
21370 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
21372 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
21373 Add member lr_slot_known.
21374 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
21376 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
21377 Remove code that sets lr_slot_offset according to what a previous
21378 version of epiphany_emit_save_restore used to do.
21379 (epiphany_emit_save_restore): When doing an lr save or restore,
21380 set/verify lr_slot_known and lr_slot_offset.
21382 2013-04-08 Xinyu Qi <xyqi@marvell.com>
21385 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
21388 2013-04-08 Richard Biener <rguenther@suse.de>
21390 * alias.c (find_base_term): Fix thinko in previous change.
21392 2013-04-08 Jakub Jelinek <jakub@redhat.com>
21394 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
21395 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
21396 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
21397 if possible to compute val.
21398 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
21399 For QImode integers don't require anything about precision. Use
21400 const_with_all_bytes_same to find out if the constant doesn't have
21401 repeated bytes in it.
21403 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21405 * config/s390/s390.c (s390_expand_insv): Only accept insertions
21408 2013-04-08 Marek Polacek <polacek@redhat.com>
21410 PR rtl-optimization/48182
21411 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
21414 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21417 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
21418 nuses, make sure we have a label.
21420 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21423 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
21424 (rs6000_emit_swdiv_low_precision): Remove.
21425 (rs6000_emit_swdiv): Rewrite to handle between one and four
21426 iterations of Newton-Raphson generally; modify required number of
21427 iterations for some cases.
21428 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
21430 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
21432 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
21433 set-but-unused variable.
21435 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
21436 basic blocks of released function bodies garbage-collectable.
21438 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
21439 (struct rtl_opt_pass): Add TODO_df_finish.
21441 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
21443 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21445 * config/arm/constraints.md (q): New constraint.
21446 * config/arm/ldrdstrd.md: New file.
21447 * config/arm/arm.md (ldrdstrd.md) New include.
21448 (arm_movdi): Use "q" instead of "r" constraint
21449 for double-word memory access.
21450 (movdf_soft_insn): Likewise.
21451 * config/arm/vfp.md (movdi_vfp): Likewise.
21452 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
21453 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
21454 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
21455 (mem_ok_for_ldrd_strd): Likewise.
21456 (output_move_double): Update assertion.
21458 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21460 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
21462 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21464 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
21465 define_insn_and_split.
21466 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
21468 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21470 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
21471 define_insn_and_split.
21472 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
21473 (shiftsi3_compare): New pattern.
21474 (rrx): New pattern.
21475 * config/arm/unspecs.md (UNSPEC_RRX): New.
21477 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21479 * config/arm/arm.md (negdi_extendsidi): New pattern.
21480 (negdi_zero_extendsidi): Likewise.
21482 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21484 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
21485 define_insn_and_split.
21486 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
21487 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
21489 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21491 * config/arm/arm.md (arm_subdi3): Convert define_insn into
21492 define_insn_and_split.
21493 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
21494 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
21496 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21498 * config/arm/arm.md (subsi3_carryin): New pattern.
21499 (subsi3_carryin_const): Likewise.
21500 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
21501 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
21503 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21505 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
21507 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
21509 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
21510 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
21512 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21514 * config/arm/arm.c (arm_expand_builtin): Change fcode
21515 type to unsigned int.
21517 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21519 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
21521 2013-04-04 Ian Lance Taylor <iant@google.com>
21523 * doc/standards.texi (Standards): The Go frontend supports the Go 1
21526 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
21528 PR middle-end/56729
21529 * df-scan.c (df_insn_delete): Disable failing assert.
21531 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21533 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
21534 New function prototype.
21535 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
21536 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
21537 (arm_builtin_vectorized_function): New function.
21539 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21541 * config/arm/arm_neon_builtins.def: New file.
21542 * config/arm/arm.c (neon_builtin_data): Move contents to
21543 arm_neon_builtins.def.
21544 (enum arm_builtins): Include neon builtin definitions.
21545 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
21546 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
21548 2013-04-04 Marek Polacek <polacek@redhat.com>
21550 PR tree-optimization/48186
21551 * predict.c (maybe_hot_frequency_p): Return false if
21552 HOT_BB_FREQUENCY_FRACTION is 0.
21553 (cgraph_maybe_hot_edge_p): Likewise.
21555 2013-04-04 Richard Biener <rguenther@suse.de>
21557 PR tree-optimization/56826
21558 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
21561 2013-04-04 Richard Biener <rguenther@suse.de>
21563 PR tree-optimization/56213
21564 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
21565 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
21567 2013-04-04 Richard Biener <rguenther@suse.de>
21569 PR tree-optimization/56837
21570 * tree-loop-distribution.c (classify_partition): For non-zero
21571 values require that the value has the same precision as its
21572 mode to be useful as memset value.
21574 2013-04-03 Nick Clifton <nickc@redhat.com>
21576 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
21577 (fmssf4): Use fmsf.s on E3V5 architectures.
21578 (fnmasf4): Use fnmaf.s on E3V5 architectures.
21579 (fnmssf4): Use fnmsf.s on E3V5 architectures.
21581 2013-04-03 Jeff Law <law@redhat.com>
21583 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
21584 (lra-eliminations.o): Likewise.
21586 2013-04-03 Teresa Johnson <tejohnson@google.com>
21588 * gcov-io.c (compute_working_sets): Moved most of body of old
21589 compute_working_sets here from profile.c.
21590 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
21591 (gcov_working_set_t): Moved typedef here from basic-block.h
21592 (compute_working_set): Declare.
21593 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
21594 (get_working_sets): Renamed from compute_working_set,
21595 replace most of body with call to new compute_working_sets.
21596 (get_exec_counts): Replace call to compute_working_sets
21597 to get_working_sets.
21598 * profile.h (get_working_sets): Renamed from compute_working_set.
21599 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
21600 to get_working_sets.
21601 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
21602 * gcov-dump.c (dump_working_sets): New function.
21604 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
21606 * hwint.c (sext_hwi, zext_hwi): New functions.
21607 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
21608 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
21609 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
21610 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
21611 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
21612 (sext_hwi, zext_hwi): New functions.
21614 2013-04-03 Jeff Law <law@redhat.com>
21616 PR tree-optimization/56799
21617 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
21618 back test for widening conversion erroneously dropped in prior change.
21620 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21623 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
21624 instead of next_real_insn.
21626 2013-04-03 Marek Polacek <polacek@redhat.com>
21629 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
21631 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21634 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
21636 (thumb1_output_casesi): Likewise.
21637 (thumb2_output_casesi): Likewise.
21639 2013-04-03 Richard Biener <rguenther@suse.de>
21641 PR tree-optimization/56817
21642 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
21644 (tree_unroll_loops_completely_1): ... new function to manually
21645 walk the loop tree, properly defering outer loops of unrolled
21646 loops to later iterations.
21648 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21650 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
21651 (vectorizable_load): Likewise.
21652 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
21653 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
21655 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21657 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
21660 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
21662 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
21664 2013-04-03 Bin Cheng <bin.cheng@arm.com>
21666 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
21668 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21670 PR tree-optimization/56790
21671 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
21674 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
21676 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
21678 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
21679 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
21682 2013-04-03 Jakub Jelinek <jakub@redhat.com>
21685 * tree.h (force_folding_builtin_constant_p): New decl.
21686 * builtins.c (force_folding_builtin_constant_p): New variable.
21687 (fold_builtin_constant_p): Fold immediately also if
21688 force_folding_builtin_constant_p.
21690 2013-04-03 Richard Biener <rguenther@suse.de>
21692 PR tree-optimization/56812
21693 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
21694 DRs of the same interleaving chain are independent.
21696 2013-04-02 Jason Merrill <jason@redhat.com>
21698 * gdbinit.in (pbb): Use debug fn.
21700 2013-04-02 Lawrence Crowl <crowl@google.com>
21702 * sese.h (struct ivtype_map_elt_s): Remove unused.
21703 (extern debug_ivtype_map): Remove unused.
21704 (extern eq_ivtype_map_elts): Remove unused.
21705 * sese.c (debug_ivtype_map): Removed unused.
21706 (debug_ivtype_map_1): Removed unused.
21707 (debug_ivtype_elt): Remove unused.
21708 (eq_ivtype_map_elts): Remove unused.
21711 2013-04-02 Kai Tietz <ktietz@redhat.com>
21714 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
21715 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
21716 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
21718 (legitimize_pe_coff_symbol): Likewise.
21719 (is_imported_p): New helper-function.
21720 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
21721 for Windows x64 targets.
21722 (ix86_expand_prologue): Optimize for pe-coff targets.
21723 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
21724 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
21725 medium/large code-model.
21726 (legitimize_pic_address): Likewise.
21727 (legitimize_tls_address): Likewise.
21728 (ix86_expand_call): Likewise.
21729 (x86_output_mi_thunk): Likewise.
21730 (get_dllimport_decl): Add new beimport argument.
21731 (construct_plt_address): Don't assert for x64 pe-coff targets.
21732 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
21734 (SYMBOL_FLAG_STUBVAR): New macro.
21735 (SYMBOL_REF_STUBVAR_P): Likewise.
21736 * config/i386/winnt.c (stub_list): New structure.
21737 (stub_head): New local variable.
21738 (i386_pe_record_stub): New function.
21739 (i386_pe_file_end): Emit refptr-stubs.
21741 2013-04-02 Jakub Jelinek <jakub@redhat.com>
21743 PR rtl-optimization/56745
21744 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
21745 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
21748 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
21749 and both of them are MEM_REFs, just compare first argument for
21750 equality and attempt to deal even with differing offsets.
21753 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
21754 of gimple_clobber_p to be MEM_REF.
21755 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
21756 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
21757 after gimplification.
21758 * asan.c (get_mem_ref_of_assignment): Don't instrument
21759 gimple_clobber_p stmts.
21760 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
21761 gimple_clobber_p stmt if they have MEM_REF lhs and
21762 are dead because of another gimple_clobber_p stmt.
21763 * tree-ssa-live.c (clear_unused_block_pointer): Treat
21764 gimple_clobber_p stmts like debug stmts.
21765 (remove_unused_locals): Remove clobbers with MEM_REF lhs
21766 that refer to unused VAR_DECLs or uninitialized values.
21767 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
21768 gimple_clobber_p stmts if they refer to removed parameters.
21769 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
21772 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
21774 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
21775 using SWI48 mode attribute.
21777 2013-04-02 Wei Mi <wmi@google.com>
21779 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
21780 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
21781 *<rotate_insn><mode>3_mask in i386.md.
21783 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
21785 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
21787 2013-04-02 Richard Biener <rguenther@suse.de>
21789 PR tree-optimization/56778
21790 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
21791 Runtime alias tests are not supported for gather loads.
21792 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
21793 stmts referenced from SSA operands before updating SSA form.
21795 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
21796 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
21798 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
21799 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
21800 * config/arm/cortex-a53.md: New file.
21801 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
21802 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
21803 * config/arm/arm.c (arm_issue_rate): Likewise.
21804 * config/arm/arm-tune.md: Regenerate
21805 * config/arm/arm-tables.opt: Regenerate.
21806 * config/arm/arm-cores.def: Add cortex-a53.
21808 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
21810 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
21813 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21815 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
21816 scalar load/store operations using B/H registers.
21817 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
21819 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
21821 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
21823 * config/aarch64/aarch64.c
21824 (aarch64_simd_scalar_immediate_valid_for_move): New.
21825 * config/aarch64/aarch64-protos.h
21826 (aarch64_simd_scalar_immediate_valid_for_move): New.
21827 * config/aarch64/constraints.md (Dh, Dq): New.
21828 * config/aarch64/iterators.md (hq): New.
21830 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
21832 * reorg.c (get_branch_condition): Deal with conditional returns.
21833 (fill_simple_delay_slots): Remove dead code dealing with jumps.
21835 2013-04-01 Wei Mi <wmi@google.com>
21837 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
21838 Truncate operand 2 using %b asm operand modifier.
21839 (*<shift_insn><mode>3_mask): Ditto.
21840 (*<rotate_insn><mode>3_mask): Ditto.
21842 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
21844 PR middle-end/56798
21845 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
21847 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
21849 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
21851 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
21853 2013-03-30 Lawrence Crowl <crowl@google.com>
21855 * dse.c (clear_alias_sets): Remove never set.
21856 (disqualified_clear_alias_sets): Remove never set.
21857 (clear_alias_mode_pool): Remove never set.
21858 (dse_step0): Remove condition that is never true.
21859 (canon_address): Remove condition that is never true.
21860 (dse_step7): Remove condition that is never true.
21861 (rest_of_handle_dse): Remove condition that is never true.
21862 (rest_of_handle_dse::did_global): Remove never read from above.
21863 (dse_step2_spill): Remove never called from above.
21864 (dse_step5_spill): Remove never called from above.
21866 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
21868 * doc/md.texi (Standard Names) <casesi>: Update documentation for
21869 JUMP_TABLE_DATA changes.
21870 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
21871 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
21872 (Insns) <jump_table_data>: New entry.
21873 * doc/tm.texi: Regenerate.
21875 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
21877 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
21878 for table jump at the end of a basic block using tablejump_p.
21879 * targhooks.c (default_invalid_within_doloop): Likewise.
21880 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
21881 target hook implementation that is identical to the default hook.
21882 (rs6000_invalid_within_doloop): Remove.
21884 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
21885 unused variable from tablejump_p call.
21887 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
21888 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
21889 (INSN_DELETED_P): Likewise.
21890 (emit_jump_table_data): New prototype.
21891 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
21892 after 4th as unused.
21893 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
21894 * sched-vis.c (print_insn): Likewise.
21895 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
21896 insn for compatibility with back ends that use next_active_insn to
21897 identify jump table data.
21898 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
21899 (remove_insn): Likewise.
21900 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
21902 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
21903 (emit_jump_table_data): New function.
21905 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
21906 basic block, a JUMP_TABLE_DATA never is.
21907 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
21908 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
21909 off from code handling real insns.
21910 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
21911 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
21912 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
21913 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
21914 is not a NONDEBUG_INSN_P.
21915 * ira-costs.c (scan_one_insn): Likewise.
21916 * jump.c (mark_all_labels): Likewise.
21917 (mark_jump_label_1): Likewise.
21918 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
21919 * lra.c (get_insn_freq): Expect all insns reaching here to be in
21921 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
21922 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
21923 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
21924 JUMP_TABLE_DATA_P insns.
21925 (calculate_elim_costs_all_insns): Likewise.
21926 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
21927 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
21928 (delete_output_reload): Code style fixups.
21929 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
21930 insn flags on this non-insn.
21931 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
21932 as scheduling barriers, for pre-change compatibility.
21933 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
21934 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
21936 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
21937 redundant JUMP_TABLE_DATA_P test.
21938 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
21939 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
21940 (frv_for_each_packet): Likewise.
21941 * config/i386/i386.c (min_insn_size): Likewise.
21942 (ix86_avoid_jump_mispredicts): Likewise.
21943 * config/m32r/m32r.c (m32r_is_insn): Likewise.
21944 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
21945 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
21946 (mips16_insn_length): Robustify.
21947 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
21948 (mips16_split_long_branches): Likewise.
21949 * config/pa/pa.c (pa_combine_instructions): Likewise.
21950 * config/rs6000/rs6000.c (get_next_active_insn): Treat
21951 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
21952 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
21953 as contributing to pool range lengths.
21954 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
21955 Remove redundant JUMP_TABLE_DATA_P test.
21956 (sh_loop_align): Likewise.
21957 (split_branches): Likewise.
21958 (sh_insn_length_adjustment): Likewise.
21959 * config/spu/spu.c (get_branch_target): Likewise.
21961 2013-03-29 Jan Hubicka <jh@suse.cz>
21963 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
21964 gcov streaming; stream hot bb threshold to ltrans.
21965 * predict.c (get_hot_bb_threshold): Break out from ....
21966 (maybe_hot_count_p): ... here.
21967 (set_hot_bb_threshold): New function.
21968 * lto-section-in.c (lto_section_name): Add profile.
21969 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
21970 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
21971 and data-streamer.h
21972 (histogram_entry): New structure.
21973 (histogram, histogram_pool): New global vars.
21974 (histogram_hash): New structure.
21975 (histogram_hash::hash): New method.
21976 (histogram_hash::equal): Likewise.
21977 (account_time_size): New function.
21978 (cmp_counts): New function.
21979 (dump_histogram): New function.
21980 (ipa_profile_generate_summary): New function.
21981 (ipa_profile_write_summary): New function.
21982 (ipa_profile_read_summary): New function.
21983 (ipa_profile): Decide on threshold.
21984 (pass_ipa_profile): Add ipa_profile_write_summary and
21985 ipa_profile_read_summary.
21986 * Makefile.in (ipa.o): Update dependencies.
21987 * lto-streamer.h (LTO_section_ipa_profile): New section.
21989 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
21991 * tree.h (VAR_P): New.
21993 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
21996 * doc/invoke.texi ([-fwhole-program]): Fix typo.
21998 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
22000 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
22001 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
22002 (control_flow_insn_p): Likewise.
22003 * cfgrtl.c (duplicate_insn_chain): Likewise.
22004 * final.c (get_attr_length_1): Likewise.
22005 (shorten_branches): Likewise.
22006 (final_scan_insn): Likewise.
22007 * function.c (instantiate_virtual_regs): Likewise.
22008 * gcse.c (insert_insn_end_basic_block): Likewise.
22009 * ira-costs.c (scan_one_insn): Likewise.
22010 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
22011 * lra.c (check_rtl): Likewise.
22012 * reload1.c (elimination_costs_in_insn): Likewise.
22013 * reorg.c (follow_jumps): Likewise.
22015 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
22016 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
22017 (thumb_far_jump_used_p): Likewise.
22018 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
22019 (workaround_speculation): Likewise.
22020 (add_sched_insns_for_speculation): Likewise.
22021 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
22022 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
22023 (frv_for_each_packet): Likewise.
22024 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
22025 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
22026 (final_emit_insn_group_barriers): Likewise.
22027 * config/m32r/m32r.c (m32r_is_insn): Likewise.
22028 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
22029 (mips16_insn_length): Likewise.
22030 * config/pa/pa.c (pa_reorg): Likewise.
22031 (pa_combine_instructions): Likewise.
22032 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
22033 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
22034 (sh_reorg): Likewise.
22035 (split_branches): Likewise.
22036 * config/spu/spu.c (get_branch_target): Likewise.
22038 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
22041 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
22043 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
22044 Fix declaration name.
22046 2013-03-28 Lawrence Crowl <crowl@google.com>
22048 * graphds.h (struct graph.indicies): Remove unused.
22049 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
22050 (SCOP_ORIGINAL_PDDRS): Remove unused.
22051 * sese.h (extern insert_loop_close_phis): Removed unused.
22052 (extern insert_guard_phis): Removed unused.
22053 (extern ivtype_map_elt_info): Removed unused.
22054 (new_ivtype_map_elt): Removed unused.
22055 * sese.c (ivtype_map_elt_info): Removed unused.
22057 2013-03-28 Lawrence Crowl <crowl@google.com>
22059 * Makefile.in: Add several missing include dependences.
22061 (test-dump.o): New. This object is not added to any executable,
22062 but is present for ad-hoc testing.
22064 (debug (const bitmap_head_def &)): New.
22065 (debug (const bitmap_head_def *)): New.
22067 (extern debug (const bitmap_head_def &)): New.
22068 (extern debug (const bitmap_head_def *)): New.
22070 (debug (edge_def &)): New.
22071 (debug (edge_def *)): New.
22073 (debug (basic_block_def &)): New.
22074 (debug (basic_block_def *)): New.
22076 (dump_node (const_tree, int, FILE *)): Correct source file.
22078 (debug (die_struct &)): New.
22079 (debug (die_struct *)): New.
22081 (extern debug (die_struct &)): New.
22082 (extern debug (die_struct *)): New.
22083 * gimple-pretty-print.c
22084 (debug (gimple_statement_d &)): New.
22085 (debug (gimple_statement_d *)): New.
22086 * gimple-pretty-print.h
22087 (extern debug (gimple_statement_d &)): New.
22088 (extern debug (gimple_statement_d *)): New.
22090 (debug (ira_allocno_copy &)): New.
22091 (debug (ira_allocno_copy *)): New.
22092 (debug (ira_allocno &)): New.
22093 (debug (ira_allocno *)): New.
22095 (extern debug (ira_allocno_copy &)): New.
22096 (extern debug (ira_allocno_copy *)): New.
22097 (extern debug (ira_allocno &)): New.
22098 (extern debug (ira_allocno *)): New.
22100 (debug (live_range &)): New.
22101 (debug (live_range *)): New.
22103 (debug (lra_live_range &)): New.
22104 (debug (lra_live_range *)): New.
22106 (debug (lra_live_range &)): New.
22107 (debug (lra_live_range *)): New.
22109 (debug (omega_pb_d &)): New.
22110 (debug (omega_pb_d *)): New.
22112 (extern debug (omega_pb_d &)): New.
22113 (extern debug (omega_pb_d *)): New.
22115 (debug (const rtx_def &)): New.
22116 (debug (const rtx_def *)): New.
22118 (debug_tree (tree): Move within file.
22119 (debug_raw (const tree_node &)): New.
22120 (debug_raw (const tree_node *)): New.
22121 (dump_tree_via_hooks (const tree_node *, int)): New.
22122 (debug (const tree_node &)): New.
22123 (debug (const tree_node *)): New.
22124 (debug_verbose (const tree_node &)): New.
22125 (debug_verbose (const tree_node *)): New.
22126 (debug_head (const tree_node &)): New.
22127 (debug_head (const tree_node *)): New.
22128 (debug_body (const tree_node &)): New.
22129 (debug_body (const tree_node *)): New.
22130 (debug_vec_tree (tree): Move and reimplement in terms of dump.
22131 (debug (vec<tree, va_gc> &)): New.
22132 (debug (vec<tree, va_gc> *)): New.
22134 (extern debug (const rtx_def &)): New.
22135 (extern debug (const rtx_def *)): New.
22137 (debug_raw (simple_bitmap_def &)): New.
22138 (debug_raw (simple_bitmap_def *)): New.
22139 (debug (simple_bitmap_def &)): New.
22140 (debug (simple_bitmap_def *)): New.
22142 (extern debug (simple_bitmap_def &)): New.
22143 (extern debug (simple_bitmap_def *)): New.
22144 (extern debug_raw (simple_bitmap_def &)): New.
22145 (extern debug_raw (simple_bitmap_def *)): New.
22147 (debug (vinsn_def &)): New.
22148 (debug (vinsn_def *)): New.
22149 (debug_verbose (vinsn_def &)): New.
22150 (debug_verbose (vinsn_def *)): New.
22151 (debug (expr_def &)): New.
22152 (debug (expr_def *)): New.
22153 (debug_verbose (expr_def &)): New.
22154 (debug_verbose (expr_def *)): New.
22155 (debug (vec<rtx> &)): New.
22156 (debug (vec<rtx> *)): New.
22158 (extern debug (vinsn_def &)): New.
22159 (extern debug (vinsn_def *)): New.
22160 (extern debug_verbose (vinsn_def &)): New.
22161 (extern debug_verbose (vinsn_def *)): New.
22162 (extern debug (expr_def &)): New.
22163 (extern debug (expr_def *)): New.
22164 (extern debug_verbose (expr_def &)): New.
22165 (extern debug_verbose (expr_def *)): New.
22166 (extern debug (vec<rtx> &)): New.
22167 (extern debug (vec<rtx> *)): New.
22169 (_list_iter_cond_expr): Make inline instead of static.
22171 (debug (sreal &)): New.
22172 (debug (sreal *)): New.
22174 (extern debug (sreal &)): New.
22175 (extern debug (sreal *)): New.
22177 (extern debug_raw (const tree_node &)): New.
22178 (extern debug_raw (const tree_node *)): New.
22179 (extern debug (const tree_node &)): New.
22180 (extern debug (const tree_node *)): New.
22181 (extern debug_verbose (const tree_node &)): New.
22182 (extern debug_verbose (const tree_node *)): New.
22183 (extern debug_head (const tree_node &)): New.
22184 (extern debug_head (const tree_node *)): New.
22185 (extern debug_body (const tree_node &)): New.
22186 (extern debug_body (const tree_node *)): New.
22187 (extern debug (vec<tree, va_gc> &)): New.
22188 (extern debug (vec<tree, va_gc> *)): New.
22190 (debug (struct loop &)): New.
22191 (debug (struct loop *)): New.
22192 (debug_verbose (struct loop &)): New.
22193 (debug_verbose (struct loop *)): New.
22194 * tree-dump.c: Add header dependence.
22196 (extern debug (struct loop &)): New.
22197 (extern debug (struct loop *)): New.
22198 (extern debug_verbose (struct loop &)): New.
22199 (extern debug_verbose (struct loop *)): New.
22201 (debug (data_reference &)): New.
22202 (debug (data_reference *)): New.
22203 (debug (vec<data_reference_p> &)): New.
22204 (debug (vec<data_reference_p> *)): New.
22205 (debug (vec<ddr_p> &)): New.
22206 (debug (vec<ddr_p> *)): New.
22208 (extern debug (data_reference &)): New.
22209 (extern debug (data_reference *)): New.
22210 (extern debug (vec<data_reference_p> &)): New.
22211 (extern debug (vec<data_reference_p> *)): New.
22212 (extern debug (vec<ddr_p> &)): New.
22213 (extern debug (vec<ddr_p> *)): New.
22215 (debug (pt_solution &)): New.
22216 (debug (pt_solution *)): New.
22218 (extern debug (pt_solution &)): New.
22219 (extern debug (pt_solution *)): New.
22221 (debug (_var_map &)): New.
22222 (debug (_var_map *)): New.
22223 (debug (tree_live_info_d &)): New.
22224 (debug (tree_live_info_d *)): New.
22226 (extern debug (_var_map &)): New.
22227 (extern debug (_var_map *)): New.
22228 (extern debug (tree_live_info_d &)): New.
22229 (extern debug (tree_live_info_d *)): New.
22231 2013-03-28 Jan Hubicka <jh@suse.cz>
22233 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
22235 2013-03-28 Ian Bolton <ian.bolton@arm.com>
22237 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
22238 record only when desired or required.
22240 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
22242 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
22243 *vec_extractv2di_1_rex64. Use x64 isa attribute.
22245 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22247 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
22248 (*andsi3_compare0_uxtw): New pattern.
22249 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
22250 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
22252 2013-03-28 Jan Hubicka <jh@suse.cz>
22254 * data-streamer-in.c (streamer_read_gcov_count): New function.
22255 * gimple-streamer-out.c: Include value-prof.h.
22256 (output_gimple_stmt): Output histogram.
22257 (output_bb): Use streamer_write_gcov_count.
22258 * value-prof.c: Include data-streamer.h
22259 (dump_histogram_value): Add HIST_TYPE_MAX.
22260 (stream_out_histogram_value): New function.
22261 (stream_in_histogram_value): New function.
22262 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
22263 (stream_out_histogram_value, stream_in_histogram_value): Declare.
22264 * data-streamer-out.c (streamer_write_gcov_count): New function.
22265 (streamer_write_gcov_count_stream): New function.
22266 * lto-cgraph.c (lto_output_edge): Update counter streaming.
22267 (lto_output_node): Likewise.
22268 (input_node, input_edge): Likewise.
22269 * lto-streamer-out.c (output_cfg): Update streaming.
22270 * lto-streamer-in.c (input_cfg): Likewise.
22271 * data-streamer.h (streamer_write_gcov_count,
22272 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
22273 * gimple-streamer-in.c: Include value-prof.h
22274 (input_gimple_stmt): Input histograms.
22275 (input_bb): Update profile streaming.
22277 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22279 * genmodes.c (emit_max_int): New function.
22280 (emit_insn_modes_h): Added call to emit_max_function.
22281 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
22283 * machmode.def: Fixed comment.
22285 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
22287 * combine.c (try_combine): Removed useless assert.
22288 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
22290 2013-03-28 Marek Polacek <polacek@redhat.com>
22291 Richard Biener <rguenther@suse.de>
22293 PR tree-optimization/56695
22294 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
22295 build signed result of a vector comparison.
22296 * tree-cfg.c (verify_gimple_comparison): Check that a result
22297 of a vector comparison has signed type.
22299 2013-03-28 Richard Biener <rguenther@suse.de>
22301 PR tree-optimization/37021
22302 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
22303 do not restrict gaps between groups.
22304 * tree-vect-stmts.c (vectorizable_load): Properly account for
22305 a gap between groups.
22307 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
22309 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
22310 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
22313 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
22315 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
22316 * doc/extend.texi (Named Address Spaces): Ditto.
22317 (Variable Attributes): Ditto.
22319 2013-03-27 Kai Tietz <ktietz@redhat.com>
22321 * config.build: Add support for cygwin x64 target.
22322 * config.gcc: Likewise.
22323 * config.host: Likewise.
22324 * configure.ac: Likewise
22325 * configure: Regenerated.
22327 2013-03-27 Kai Tietz <ktietz@redhat.com>
22329 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
22330 * config/i386/t-cygwin-w64: New file.
22331 * config/i386/cygwin-w64.h: New file.
22332 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
22333 and add support for x64-cygwin target.
22334 (CPP_SPEC): Likewise.
22335 (CXX_WRAP_SPEC_LIST): Undefine before define.
22336 (LIBGCJ_SONAME): Use 15 as version.
22338 2013-03-27 Richard Biener <rguenther@suse.de>
22340 PR tree-optimization/56716
22341 * tree-ssa-structalias.c (perform_var_substitution): Adjust
22342 dumping for ref nodes.
22344 2013-03-27 Martin Jambor <mjambor@suse.cz>
22346 PR tree-optimization/55334
22347 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
22348 restricted pointers to arrays.
22350 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
22352 * Makefile.in (.SUFFIXES): Add .cc.
22353 (.c.o): Apply same recipe for implicit rule .cc.o.
22355 2013-03-27 Richard Biener <rguenther@suse.de>
22357 PR tree-optimization/37021
22358 * tree-vect-data-refs.c (vect_check_strided_load): Allow
22359 REALPART/IMAGPART_EXPRs around the supported refs.
22360 * tree-ssa-structalias.c (find_func_aliases): Assume that
22361 floating-point values are not used to transfer pointers.
22363 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
22365 * target.def (TARGET_HAS_IFUNC_P): New target hook.
22366 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
22367 * doc/tm.texi: Regenerate.
22368 * targhooks.h (default_has_ifunc_p): New.
22369 * targhooks.c (default_has_ifunc_p): Ditto.
22370 * config/linux-protos.h: New file.
22371 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
22372 hook for linux which disables support of indirect functions in android.
22373 * config/linux-android.c: New file.
22374 * config/t-linux-android.c: Ditto.
22375 * config.gcc: Added new object file linux-android.o.
22376 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
22377 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
22378 * varasm.c (do_assemble_alias): Likewise.
22379 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
22380 doesn't support indirect functions.
22381 * configure: Regenerate.
22383 2013-03-27 Bin Cheng <bin.cheng@arm.com>
22386 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
22387 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
22390 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22392 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
22394 2013-03-27 Terry Guo <terry.guo@arm.com>
22396 * config/arm/arm-cores.def: Added core cortex-r7.
22397 * config/arm/arm-tune.md: Regenerated.
22398 * config/arm/arm-tables.opt: Regenerated.
22399 * doc/invoke.texi: Added entry for core cortex-r7.
22401 2013-03-27 Walter Lee <walt@tilera.com>
22403 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
22404 double-decrement of next_scratch_regno.
22406 2013-03-27 Walter Lee <walt@tilera.com>
22408 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
22410 (insn_v1mulus): Ditto.
22411 (insn_v2muls): Ditto.
22413 2013-03-27 Walter Lee <walt@tilera.com>
22415 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
22416 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
22418 2013-03-27 Walter Lee <walt@tilera.com>
22420 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
22421 (*sibcall_value): Ditto.
22423 2013-03-27 Walter Lee <walt@tilera.com>
22425 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
22426 (insn_mnz_v8qi): ... this ...
22427 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
22429 (insn_v<n>mnz): Replaced by ...
22430 (insn_v1mnz): ... this ...
22431 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
22433 (insn_mz_<mode>): Replaced by ...
22434 (insn_mz_v8qi): ... this ...
22435 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
22437 (insn_v<n>mz): Replaced by ...
22438 (insn_v1mz): ... this ...
22439 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
22442 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
22444 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
22446 2013-03-26 Roland McGrath <mcgrathr@google.com>
22448 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
22449 than fprintf with a non-constant, non-format string.
22451 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22453 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
22454 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
22455 operand 0 predicate.
22456 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
22457 attribute. Use general_x64nomem_operand as operand 1 predicate.
22458 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
22459 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22460 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
22461 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
22462 (mov<mode>_insv_1): Remove expander. Merge insn with
22463 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
22464 Use general_x64nomem_operand as operand 1 predicate.
22465 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
22466 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
22467 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22468 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
22469 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
22470 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
22471 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
22472 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
22473 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
22474 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
22475 (general_x64nomem_operand): Ditto.
22477 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
22479 * config/rtems.opt: Add -pthread option.
22481 2013-03-26 Richard Biener <rguenther@suse.de>
22483 * alias.c (find_base_term): Avoid redundant and not used recursion.
22484 (base_alias_check): Get the initial base term from the caller.
22485 (true_dependence_1): Compute and pass base terms to base_alias_check.
22486 (write_dependence_p): Likewise.
22487 (may_alias_p): Likewise.
22489 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
22491 * config/aarch64/aarch64.c (aarch64_classify_address): Support
22492 PC-relative load in SI modes and above only.
22494 2013-03-26 Xinyu Qi <xyqi@marvell.com>
22496 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
22497 * config/arm/iwmmxt.md (WCGR0): Update.
22498 (WCGR1, WCGR2, WCGR3): Likewise.
22500 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
22502 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
22503 Use x64 and nox64 isa attributes.
22505 2013-03-26 Richard Biener <rguenther@suse.de>
22507 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
22508 alignment computations and rely on get_object_alignment_1
22509 for the !TYPE_P case.
22510 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
22512 2013-03-26 Walter Lee <walt@tilera.com>
22514 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
22515 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
22517 2013-03-25 Jeff Law <law@redhat.com>
22519 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
22520 check for INTEGRAL_TYPE_P that was missing due to checking in
22521 wrong version of prior patch.
22523 2013-03-25 Walter Lee <walt@tilera.com>
22525 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
22526 TILEGX_INSN_SHUFFLEBYTES1.
22527 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
22529 (tilegx_builtins): Ditto.
22530 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
22532 2013-03-25 Walter Lee <walt@tilera.com>
22534 * config/tilegx/tilegx.md (floatsisf2): New pattern.
22535 (floatunssisf2): New pattern.
22536 (floatsidf2): New pattern.
22537 (floatunssidf2): New pattern.
22539 2013-03-25 Walter Lee <walt@tilera.com>
22541 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
22542 tests for constraint J, K, N, P.
22544 2013-03-25 Walter Lee <walt@tilera.com>
22546 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
22547 Use indirect/pcrel encoding.
22548 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
22551 2013-03-25 Steve Ellcey <sellcey@mips.com>
22553 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
22554 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
22555 * config/mips/mips.c (mips_option_override): Set IMADD default.
22556 * config/mips/mips.h (PTF_AVOID_IMADD): New.
22557 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
22558 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
22559 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
22560 * doc/invoke.texi (-mimadd/-mno-imadd): New.
22562 2013-03-25 Jeff Law <law@redhat.com>
22564 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
22565 slightly to avoid creating and folding useless trees. Simplify
22566 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
22568 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
22570 * config/i386/i386.md (*zero_extendsidi2): Merge with
22571 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
22572 * config/i386/predicates.md (x86_64_zext_operand): Rename from
22573 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
22574 targets. Clarify comment.
22576 2013-03-25 Martin Jambor <mjambor@suse.cz>
22578 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
22579 pass-through jump functions differently.
22580 (ipa_read_jump_function): Likewise. Also use setter functions to set
22583 2013-03-25 Martin Jambor <mjambor@suse.cz>
22585 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
22586 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
22588 (ipa_get_indirect_edge_target): New function.
22589 (devirtualization_time_bonus): New parameter known_aggs, pass it to
22590 ipa_get_indirect_edge_target. Update all callers.
22591 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
22592 ipa_get_indirect_edge_target_1 instead of calling
22593 ipa_get_indirect_edge_target.
22594 (create_specialized_node): Pass aggvlas to
22595 ipcp_discover_new_direct_edges.
22597 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22599 * config/arm/arm.md (f_sels, f_seld): New types.
22600 (*cmov<mode>): New pattern.
22601 * config/arm/predicates.md (arm_vsel_comparison_operator): New
22604 2013-03-25 Kai Tietz <ktietz@redhat.com>
22606 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
22607 POSIX-printf for mingw-hosted builds.
22609 2013-03-25 Richard Biener <rguenther@suse.de>
22611 PR middle-end/56694
22612 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
22613 must-not-throw stmt location.
22615 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22617 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
22618 Emit load-acquire versions when acq is true.
22619 (arm_emit_store_exclusive): Add rel parameter.
22620 Emit store-release versions when rel is true.
22621 (arm_split_compare_and_swap): Use acquire-release instructions
22623 of barriers when appropriate.
22624 (arm_split_atomic_op): Likewise.
22625 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
22626 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
22627 (VUNSPEC_SLX): Likewise.
22628 (VUNSPEC_LDA): Likewise.
22629 (VUNSPEC_STL): Likewise.
22630 * config/arm/sync.md (atomic_load<mode>): New pattern.
22631 (atomic_store<mode>): Likewise.
22632 (arm_load_acquire_exclusive<mode>): Likewise.
22633 (arm_load_acquire_exclusivesi): Likewise.
22634 (arm_load_acquire_exclusivedi): Likewise.
22635 (arm_store_release_exclusive<mode>): Likewise.
22637 2013-03-25 Catherine Moore <clm@codesourcery.com>
22639 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
22640 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
22641 * config/mip/predicates.md (lwsp_swsp_operand,
22642 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
22643 sb16_operand, db4_operand, db7_operand, ib3_operand,
22644 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
22645 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
22646 andi16_operand): New predicates.
22647 * config/mips/mips.md (compression): New attribute.
22648 (enabled): New attribute.
22649 (length): Consider compression in computing length.
22650 (shift_compression): New code attribute.
22651 (*add<mode>3): New operands. Record compression.
22652 (sub<mode>3): Likewise.
22653 (one_cmpl<mode>2): Likewise.
22654 (*and<mode>3): Likewise.
22655 (*ior<mode>3): Likewise.
22656 (unnamed pattern for xor): Likewise.
22657 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
22658 (*<optab><mode>3): Likewise.
22659 (*mov<mode>_internal: Likewise.
22660 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
22661 (mips_unsigned_immediate_p): New.
22662 (umips_lwsp_swsp_address_p): New.
22663 (m16_based_address_p): New.
22664 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
22665 (mips_unsigned_immediate_p): New prototype.
22666 (lwsp_swsp_address_p): New prototype.
22667 (m16_based_address_p): New prototype.
22668 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
22669 (mips_signed_immediate_p): New function.
22670 (m16_based_address_p): New function.
22671 (lwsp_swsp_address_p): New function.
22672 (mips_print_operand_punctuation): Recognize short delay slot insns
22673 for microMIPS.add<mode>3"
22675 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22678 * config/arm/iterators.md (v_cmp_result): New mode attribute.
22679 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
22681 2013-03-25 Richard Biener <rguenther@suse.de>
22683 PR tree-optimization/56689
22684 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
22687 2013-03-25 Richard Biener <rguenther@suse.de>
22689 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
22691 (memory_references): Likewise.
22692 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
22693 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
22694 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
22695 (gather_mem_refs_in_loops): Fold into ...
22696 (analyze_memory_references): ... this. Move initialization
22697 to tree_ssa_lim_initialize.
22698 (fill_always_executed_in): Rename to ...
22699 (fill_always_executed_in_1): ... this.
22700 (fill_always_executed_in): Move contains_call computation to
22701 this new function from ...
22702 (tree_ssa_lim_initialize): ... here.
22703 (tree_ssa_lim): Call fill_always_executed_in.
22705 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22707 * postreload.c (reload_combine): Fix code detecting returns.
22709 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
22711 * function.c (emit_use_return_register_into_block): On cc0 targets,
22712 do not emit the sequence between cc0 setter and user.
22714 2013-03-25 Kai Tietz <ktietz@redhat.com>
22716 * config/i386/predicates.md (local_symbolic_operand): Interpret
22717 dll-imported symbols as none-local.
22719 2013-03-25 Richard Biener <rguenther@suse.de>
22721 * tree-ssa-loop-im.c (struct depend): Remove.
22722 (struct lim_aux_data): Make depends a vec of gimples.
22723 (free_lim_aux_data): Adjust.
22724 (add_dependency): Likewise.
22725 (set_level): Likewise.
22727 2013-03-25 Richard Biener <rguenther@suse.de>
22729 PR middle-end/56434
22730 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
22731 the pointer returned by calls with ECF_MALLOC set.
22733 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22735 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
22737 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
22739 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
22740 using MMXMODE mode iterator.
22741 (*move<mode>_internal): Merge with *movv2sf_internal and
22742 *movv2sf_internal_rex64 using MMXMODE mode iterator.
22744 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22746 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
22747 (record_last_mem_set_info): Likewise.
22749 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
22750 of XNEWVEC followed by memset.
22751 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
22753 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
22755 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
22756 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
22757 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
22758 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
22759 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
22760 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
22761 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
22762 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
22763 BARRIER_P instead of GET_CODE.
22765 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22767 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
22768 inaccuracy in the probing code.
22770 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
22771 (ctrapdi4): Likewise.
22773 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22775 * calls.c (expand_call): Add missing guard to code handling return
22776 of non-BLKmode structures in MSB.
22777 * function.c (expand_function_end): Likewise.
22779 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
22781 * combine.c (try_combine): Adjust comment. Do not add the set of
22782 insn #0 if the destination indirectly is set or dies in insn #2.
22783 Tidy up code to distribute a new note.
22785 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22787 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
22788 also for alternatives 16 and 17.
22790 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
22792 * config/i386/sse.md (*mov<mode>_internal): Merge with
22793 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
22794 Emit insn template depending on type attribute. Use
22795 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
22796 movd instead of movq mnemonic for interunit moves. Rewrite mode
22797 attribute calculation. Remove unit attribute calculation.
22798 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
22799 Set prefix_data16 attribute for DImode ssemov types.
22800 Use Ym instead of y for SSE-MMX conversion alternatives.
22801 Reorder operand constraints.
22803 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
22805 * df.h (df_insn_delete): Adjust prototype.
22806 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
22807 and let it decide whether mark the basic block dirty.
22808 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
22809 * df-scan.c (df_insn_info_delete): New helper function, split
22810 off from df_insn_delete.
22811 (df_scan_free_bb_info): Use it.
22812 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
22814 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
22815 that the insn is actually an insn and it has a non-NULL basic block.
22816 Do not mark basic block dirty if only deleting a DEBUG_INSN.
22818 2013-03-22 Richard Biener <rguenther@suse.de>
22820 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
22822 (mem_ref_alloc): Do not allocate them.
22823 (refs_independent_p): Do not query or maintain a cache.
22825 2013-03-22 Richard Biener <rguenther@suse.de>
22827 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
22828 (gather_mem_refs_in_loops): Do not compute it.
22829 (analyze_memory_references): Do not allocate it.
22830 (tree_ssa_lim_finalize): Do not free it.
22831 (for_all_locs_in_loop): Do not query all_refs_in_loop.
22833 2013-03-22 Richard Biener <rguenther@suse.de>
22835 * is-a.h (as_a): Use gcc_checking_assert.
22837 2013-03-22 Ian Bolton <ian.bolton@arm.com>
22839 * config/aarch64/aarch64.c (aarch64_print_operand): New
22840 format specifier for printing a constant in hex.
22841 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
22842 format specifier for printing second operand.
22844 2013-03-22 Richard Biener <rguenther@suse.de>
22846 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
22848 (gather_mem_refs_in_loops): Perform store accumulation here.
22849 (create_vop_ref_mapping_loop): Remove.
22850 (create_vop_ref_mapping): Likewise.
22851 (analyze_memory_references): Initialize refs_stored_in_loop.
22852 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
22853 (record_indep_loop): Remove.
22854 (record_dep_loop): New function.
22855 (ref_indep_loop_p_1): Adjust to only walk over references
22856 in the loop, not its subloops.
22857 (ref_indep_loop_p): Rename to ...
22858 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
22859 maintaining a more fine-grained cache.
22860 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
22861 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
22863 2013-03-22 Richard Biener <rguenther@suse.de>
22865 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
22866 (struct mem_ref): Make accesses_in_loop a vec of a vec of
22867 aggregate mem_ref_loc.
22868 (free_mem_ref_locs): Inline into ...
22869 (memref_free): ... this and adjust.
22870 (mem_ref_alloc): Adjust.
22871 (mem_ref_locs_alloc): Remove.
22872 (record_mem_ref_loc): Adjust.
22873 (get_all_locs_in_loop): Rewrite into ...
22874 (for_all_locs_in_loop): ... this iterator.
22875 (rewrite_mem_ref_loc): New functor.
22876 (rewrite_mem_refs): Use for_all_locs_in_loop.
22877 (sm_set_flag_if_changed): New functor.
22878 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
22879 (ref_always_accessed): New functor.
22880 (ref_always_accessed_p): Use for_all_locs_in_loop.
22882 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22884 * tree-pass.h (PROP_gimple_lvec): New.
22885 * passes.c (dump_properties): Handle PROP_gimple_lvec.
22886 (init_optimization_passes): Move pass_lower_vector.
22887 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
22889 (pass_lower_vector): Provide PROP_gimple_lvec.
22890 (pass_lower_vector_ssa): Likewise.
22891 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
22893 2013-03-21 Mark Wielaard <mjw@redhat.com>
22895 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
22897 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
22899 * config/i386/i386.md (*movdi_internal): Disparage slightly
22900 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
22901 conversion alternatives.
22903 2013-03-21 Jakub Jelinek <jakub@redhat.com>
22905 PR middle-end/48087
22906 * diagnostic.def (DK_WERROR): New kind.
22907 * diagnostic.h (werrorcount): Define.
22908 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
22909 promoted to DK_ERROR, increment DK_WERROR counter instead of
22911 * toplev.c (toplev_main): Call print_ignored_options even if
22912 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
22913 even if just werrorcount is non-zero.
22916 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
22918 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
22919 (string_cst_pool_decl): New function.
22920 (optimize_one_addr_into_implicit_ptr): New function.
22921 (resolve_addr_in_expr): Optimize DWARF location expression
22922 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
22923 which doesn't live in memory, but has DW_AT_location or
22924 DW_AT_const_value, or refers to a string literal, into
22925 DW_OP_GNU_implicit_pointer.
22926 (optimize_location_into_implicit_ptr): New function.
22927 (resolve_addr): If removing DW_AT_location of a variable because
22928 it was DW_OP_addr of address of the variable, but the variable doesn't
22929 live in memory, try to emit const value attribute for the initializer.
22931 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22933 * tree.h (VECTOR_TYPE_P): New macro.
22934 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
22935 TYPE_MODE): Use it.
22936 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
22937 VEC_COND_EXPR cannot be lvalues.
22938 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
22940 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
22942 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
22943 Restrict the transformation to equal modes.
22945 2013-03-21 Richard Biener <rguenther@suse.de>
22947 PR tree-optimization/39326
22948 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
22949 (MEM_ANALYZABLE): Adjust.
22950 (record_mem_ref_loc): Move bitmap ops ...
22951 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
22952 unanalyzable refs, do not record locations for it.
22953 (analyze_memory_references): Allocate ref zero as shared
22955 (refs_independent_p): Do not test for unanalyzed mems here.
22956 (ref_indep_loop_p_1): Special-case disambiguation against
22957 the unanalyzed ref.
22958 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
22960 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
22962 * config/arm/arm-protos.h (tune_params): Add
22963 prefer_neon_for_64bits field.
22964 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
22965 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
22966 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
22967 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
22968 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
22969 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
22970 (arm_option_override): Handle -mneon-for-64bits new option.
22971 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
22972 (prefer_neon_for_64bits): Declare new variable.
22973 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
22974 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
22975 (arch_enabled): Handle new arch types. Remove support for onlya8
22977 (one_cmpldi2): Use new arch names.
22978 (zero_extend<mode>di2, extend<mode>di2): Ditto.
22979 * config/arm/arm.opt (mneon-for-64bits): Add option.
22980 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
22981 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
22982 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
22984 * doc/invoke.texi (-mneon-for-64bits): Document.
22986 2013-03-21 Richard Biener <rguenther@suse.de>
22988 PR tree-optimization/39326
22989 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
22990 (sort_bbs_in_loop_postorder_cmp): New function.
22991 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
22993 2013-03-21 Richard Biener <rguenther@suse.de>
22995 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
22996 (vect_insert_into_interleaving_chain): Likewise.
22997 (vect_drs_dependent_in_basic_block): Inline ...
22998 (vect_slp_analyze_data_ref_dependence): ... here. New function,
23000 (vect_analyze_data_ref_dependence): ... here. Simplify.
23001 (vect_check_interleaving): Simplify.
23002 (vect_analyze_data_ref_dependences): Likewise. Split out ...
23003 (vect_slp_analyze_data_ref_dependences): ... this new function.
23004 (dr_group_sort_cmp): New function.
23005 (vect_analyze_data_ref_accesses): Compute data-reference groups
23006 here instead of in vect_analyze_data_ref_dependence. Use
23007 a more efficient algorithm.
23008 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
23009 vect_slp_analyze_data_ref_dependences. Call
23010 vect_analyze_data_ref_accesses earlier.
23011 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
23012 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
23013 (vect_slp_analyze_data_ref_dependences): New prototype.
23015 2013-03-21 Richard Biener <rguenther@suse.de>
23017 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
23018 ref is stored in the loop.
23019 (find_refs_for_sm): Walk only over all stores.
23020 (store_motion_loop): Allocate from lim_bitmap_obstack.
23021 (store_motion): Likewise.
23023 2013-03-21 Richard Biener <rguenther@suse.de>
23025 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
23026 Update virtual SSA form.
23028 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23030 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
23031 * configure: Regenerate.
23032 * config.in: Regenerate.
23033 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
23034 if !HAVE_LD_EH_FRAME_CIEV3.
23036 2013-03-21 Richard Biener <rguenther@suse.de>
23038 * tree-cfg.c (verify_expr_no_block): New function.
23039 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
23040 nor DECL_VALUE_EXPR have locations with associated blocks.
23041 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
23042 (clear_unused_block_pointer): Remove code dealing with
23043 blocks in DECL_DEBUG_EXPR locations.
23045 2013-03-21 Richard Biener <rguenther@suse.de>
23047 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
23048 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
23049 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
23050 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
23051 instead of DECL_DEBUG_EXPR_IS_FROM.
23052 * gimplify.c (gimplify_modify_expr): Likewise.
23053 * tree-cfg.c (verify_expr_location_1): Likewise.
23054 * tree-complex.c (create_one_component_var): Likewise.
23055 * tree-sra.c (create_access_replacement): Likewise.
23056 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
23057 (clear_unused_block_pointer): Likewise.
23058 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
23059 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
23060 * var-tracking.c (var_debug_decl): Likewise.
23061 (track_expr_p): Likewise.
23062 * tree-inline.c (add_local_variables): Likewise. Set
23063 DECL_HAS_DEBUG_EXPR_P after copying it.
23064 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
23065 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
23067 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
23070 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
23071 * configure: Regenerate.
23072 * config.in: Regenerate.
23073 * config/i386/i386.md (*movdf_internal): Use
23074 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
23075 movd instead of movq mnemonic for interunit moves.
23076 (*movdi_internal): Ditto.
23078 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23080 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
23081 (abd<mode>_3): New pattern.
23082 (aba<mode>_3): New pattern.
23083 (fabd<mode>_3): New pattern.
23085 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23087 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
23088 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
23089 occurrence of REGISTER_PREFIX as its empty string.
23091 2013-03-20 Jeff Law <law@redhat.com>
23093 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
23094 addititional equivalences for equality comparisons between an SSA_NAME
23095 and a constant where the SSA_NAME was set from a widening conversion.
23097 2013-03-20 Walter Lee <walt@tilera.com>
23099 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
23101 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23103 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
23104 depending on type attribute.
23105 (*movti_internal): Ditto.
23106 (*movtf_internal): Ditto.
23107 (*movxf_internal): Ditto.
23108 (*movdf_internal): Ditto.
23109 (*movsf_internal): Ditto.
23111 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23113 * config/i386/i386.md (*movti_internal): Set prefix attribute to
23114 maybe_vex for sselog1 and ssemov types.
23115 (*movdi_internal): Reorder operand constraints.
23116 (*movsi_internal): Ditto. Set prefix attribute to
23117 maybe_vex for sselog1 and ssemov types.
23118 (*movtf_internal): Set prefix attribute to maybe_vex
23119 for sselog1 and ssemov types.
23120 (*movdf_internal): Ditto. Set prefix_data16 attribute for
23121 DImode ssemov types. Reorder operand constraints.
23122 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
23123 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
23124 attribute for SImode ssemov types. Reorder operand constraints.
23126 2013-03-20 Martin Jambor <mjambor@suse.cz>
23128 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
23129 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
23131 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
23133 * config/rs6000/predicates.md (indexed_address, update_address_mem
23134 update_indexed_address_mem): New predicates.
23135 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
23136 attribute for load/store instructions.
23137 * config/rs6000/dfp.md (movsd_store): Likewise.
23138 (movsd_load): Likewise.
23139 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
23140 (unnamed HI->DI extend define_insn): Likewise.
23141 (unnamed SI->DI extend define_insn): Likewise.
23142 (unnamed QI->SI extend define_insn): Likewise.
23143 (unnamed QI->HI extend define_insn): Likewise.
23144 (unnamed HI->SI extend define_insn): Likewise.
23145 (unnamed HI->SI extend define_insn): Likewise.
23146 (extendsfdf2_fpr): Likewise.
23147 (movsi_internal1): Likewise.
23148 (movsi_internal1_single): Likewise.
23149 (movhi_internal): Likewise.
23150 (movqi_internal): Likewise.
23151 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
23152 attribute for load/store instructions.
23153 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
23155 (mov<mode>_softfloat): Likewise.
23156 (mov<mode>_hardfloat32): Likewise.
23157 (mov<mode>_hardfloat64): Likewise.
23158 (mov<mode>_softfloat64): Likewise.
23159 (movdi_internal32): Likewise.
23160 (movdi_internal64): Likewise.
23161 (probe_stack_<mode>): Likewise.
23163 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
23165 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
23166 floating point, and decimal floating point to reload iterator.
23168 * config/rs6000/constraints.md (wl constraint): New constraints to
23169 return FLOAT_REGS if certain options are used to reduce the number
23170 of separate patterns that exist in the file.
23171 (wx constraint): Likewise.
23172 (wz constraint): Likewise.
23174 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23175 -mdebug=reg, print wg, wl, wx, and wz constraints.
23176 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23177 Initialize the reload functions for 64-bit binary/decimal floating
23179 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23180 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23181 create the buffer on the stack to overcome not having a 32-bit
23183 (rs6000_emit_move): Likewise.
23184 (rs6000_secondary_memory_needed_rtx): Likewise.
23185 (rs6000_alloc_sdmode_stack_slot): Likewise.
23186 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23187 via xxlxor, just like DFmode 0.0.
23189 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
23190 define as 1 if we are running on a power7 or newer.
23191 (enum r6000_reg_class_enum): Add new constraints.
23193 * config/rs6000/dfp.md (movsd): Delete, combine with binary
23194 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23195 with other moves by using conditional constraits (wg). Use LFIWZX
23196 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
23197 (movsd splitter): Likewise.
23198 (movsd_hardfloat): Likewise.
23199 (movsd_softfloat): Likewise.
23201 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
23202 binary and decimal floating point moves.
23203 (fmove_ok): New attributes to combine binary and decimal floating
23204 point moves, and to combine power6x (mfpgpr) moves along normal
23206 (real_value_to_target): Likewise.
23207 (f32_lr): Likewise.
23208 (f32_lm): Likewise.
23209 (f32_li): Likewise.
23210 (f32_sr): Likewise.
23211 (f32_sm): Likewise.
23212 (f32_si): Likewise.
23213 (movsf): Combine binary and decimal floating point moves. Combine
23214 power6x (mfpgpr) moves with other moves by using conditional
23215 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
23216 (mov<mode> for SFmode/SDmode); Likewise.
23217 (SFmode/SDmode splitters): Likewise.
23218 (movsf_hardfloat): Likewise.
23219 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
23220 (movsf_softfloat): Likewise.
23221 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
23223 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
23224 wx and wz constraints.
23226 * config/rs6000/constraints.md (wg constraint): New constraint to
23227 return FLOAT_REGS if -mmfpgpr (power6x) was used.
23229 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
23232 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
23233 -mdebug=reg, print wg, wl, wx, and wz constraints.
23234 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
23235 Initialize the reload functions for 64-bit binary/decimal floating
23237 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
23238 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
23239 create the buffer on the stack to overcome not having a 32-bit
23241 (rs6000_emit_move): Likewise.
23242 (rs6000_secondary_memory_needed_rtx): Likewise.
23243 (rs6000_alloc_sdmode_stack_slot): Likewise.
23244 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
23245 via xxlxor, just like DFmode 0.0.
23247 * config/rs6000/dfp.md (movdd): Delete, combine with binary
23248 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
23249 with other moves by using conditional constraits (wg). Use LFIWZX
23250 and STFIWX for loading SDmode on power7.
23251 (movdd splitters): Likewise.
23252 (movdd_hardfloat32): Likewise.
23253 (movdd_softfloat32): Likewise.
23254 (movdd_hardfloat64_mfpgpr): Likewise.
23255 (movdd_hardfloat64): Likewise.
23256 (movdd_softfloat64): Likewise.
23258 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
23259 64-bit binary and decimal floating point moves.
23260 (FMOVE64X): Likewise.
23261 (movdf): Combine 64-bit binary and decimal floating point moves.
23262 Combine power6x (mfpgpr) moves with other moves by using
23263 conditional constraits (wg).
23264 (mov<mode> for DFmode/DDmode): Likewise.
23265 (DFmode/DDmode splitters): Likewise.
23266 (movdf_hardfloat32): Likewise.
23267 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
23268 (movdf_softfloat32): Likewise.
23269 (movdf_hardfloat64_mfpgpr): Likewise.
23270 (movdf_hardfloat64): Likewise.
23271 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
23272 (movdf_softfloat64): Likewise.
23273 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
23274 (reload_<mode>_load): Move to later in the file so they aren't in
23275 the middle of the floating point move insns.
23276 (reload_<mode>_store): Likewise.
23278 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
23281 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
23282 constraint if -mdebug=reg.
23283 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
23284 Enable using dd reload support if needed.
23286 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
23287 binary and decimal floating point moves in rs6000.md.
23288 (movtd_internal): Likewise.
23290 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
23291 decimal floating point moves.
23293 (movtf_internal): Likewise.
23294 (mov<mode>_internal, TDmode/TFmode): Likewise.
23295 (movtf_softfloat): Likewise.
23296 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
23298 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
23299 movdi_internal64, using wg constraint for move direct operations.
23300 (movdi_internal64): Likewise.
23302 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
23303 MODES_TIEABLE_P for selected modes. Print the numerical value of
23304 the various virtual registers. Use GPR/FPR first/last values,
23305 instead of hard coding the register numbers. Print which modes
23306 have reload functions registered.
23307 (rs6000_option_override_internal): If -mdebug=reg, trace the options
23308 settings before/after setting cpu, target and subtarget settings.
23309 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
23310 and for secondary reload failures in rs6000_secondary_reload_inner.
23311 (rs6000_secondary_reload_fail): Likewise.
23312 (rs6000_secondary_reload_inner): Likewise.
23314 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
23315 macros for first/last GPR and FPR registers.
23316 (LAST_GPR_REGNO): Likewise.
23317 (FIRST_FPR_REGNO): Likewise.
23318 (LAST_FPR_REGNO): Likewise.
23320 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
23321 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
23322 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
23323 (vcond<mode><mode>): Likewise.
23324 (vcondu<mode><mode>): Likewise.
23325 (vector_gtu<mode>): Likewise.
23326 (vector_gte<mode>): Likewise.
23327 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
23328 to prevent the compiler from converting DImode operations to TImode.
23329 (ior<mode>3): Likewise.
23330 (and<mode>3): Likewise.
23331 (one_cmpl<mode>2): Likewise.
23332 (nor<mode>3): Likewise.
23333 (andc<mode>3): Likewise.
23335 * config/rs6000/constraints.md (wt constraint): New constraint
23336 that returns VSX_REGS if TImode is allowed in VSX registers.
23338 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
23339 constant under VSX.
23341 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
23342 similar to TImode, but it is restricted to being in the GPRs.
23344 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
23345 TImode to occupy a single VSX register.
23347 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
23348 -mvsx-timode for power7/power8.
23349 (power7 cpu): Likewise.
23350 (power8 cpu): Likewise.
23352 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
23353 sure that TFmode/TDmode take up two registers if they are ever
23354 allowed in the upper VSX registers.
23355 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
23357 (rs6000_init_hard_regno_mode_ok): Likewise.
23358 (rs6000_debug_reg_global): Add debugging for PTImode and wt
23359 constraint. Print if LRA is turned on.
23360 (rs6000_option_override_internal): Give an error if -mvsx-timode
23361 and VSX is not enabled.
23362 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
23363 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
23364 to reg+offset addressing. Use PTImode when checking offset
23365 addresses for validity.
23366 (reg_offset_addressing_ok_p): Likewise.
23367 (rs6000_legitimate_offset_address_p): Likewise.
23368 (rs6000_legitimize_address): Likewise.
23369 (rs6000_legitimize_reload_address): Likewise.
23370 (rs6000_legitimate_address_p): Likewise.
23371 (rs6000_eliminate_indexed_memrefs): Likewise.
23372 (rs6000_emit_move): Likewise.
23373 (rs6000_secondary_reload): Likewise.
23374 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
23375 reloads to fpr registers to continue to use reg+offset addressing,
23376 but 64-bit reloads to altivec registers need reg+reg addressing.
23377 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
23378 it. Treat LO_SUM like a PLUS operation.
23379 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
23380 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
23381 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
23382 registers to share a register with a smaller sized type, since VSX
23383 puts scalars in the upper 64-bits.
23384 (print_operand): Add support for PTImode.
23385 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
23386 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
23387 registers, but don't have arithmetic support.
23388 (rs6000_memory_move_cost): Add test for VSX.
23389 (rs6000_opt_masks): Add -mvsx-timode.
23391 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
23394 (VSr): Use wt constraint for TImode.
23395 (VSv): Drop TImode support.
23396 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
23397 (vsx_movti_64bit): Likewise.
23398 (vsx_movti_32bit): Likewise.
23399 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
23400 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
23401 one '?' on the appropriate output constraint. Do not allow TImode
23402 logical operations on 32-bit systems.
23403 (vsx_ior<mode>3): Likewise.
23404 (vsx_xor<mode>3): Likewise.
23405 (vsx_one_cmpl<mode>2): Likewise.
23406 (vsx_nor<mode>3): Likewise.
23407 (vsx_andc<mode>3): Likewise.
23408 (vsx_concat_<mode>): Likewise.
23409 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
23411 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
23412 OPTION_MASK_VSX_TIMODE.
23413 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
23414 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
23416 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
23417 (TI2 iterator): New iterator for TImode, PTImode.
23418 (wd mode attribute): Add values for vector types.
23419 (movti_string): Replace TI move operations with operations for TImode
23420 and PTImode. Add support for TImode being allowed in VSX registers.
23421 (mov<mode>_string, TImode/PTImode): Likewise.
23422 (movti_ppc64): Likewise.
23423 (mov<mode>_ppc64, TImode/PTImode): Likewise.
23424 (TI mode splitters): Likewise.
23426 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
23429 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
23431 PR tree-optimization/56355
23432 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
23433 Also handle integers with undefined overflow.
23435 2013-03-20 Catherine Moore <clm@codesourcery.com>
23436 Maciej W. Rozycki <macro@codesourcery.com>
23437 Tom de Vries <tom@codesourcery.com>
23438 Nathan Sidwell <nathan@codesourcery.com>
23439 Iain Sandoe <iain@codesourcery.com>
23440 Nathan Froyd <froydnj@codesourcery.com>
23441 Chao-ying Fu <fu@mips.com>
23443 * doc/extend.texi: (micromips, nomicromips, nocompression):
23444 Document new function attributes.
23445 * doc/invoke.texi (minterlink-compressed, mmicromips,
23446 m14k, m14ke, m14kec): Document new options.
23447 (minterlink-mips16): Update documentation.
23448 * doc/md.texi (ZC, ZD): Document new constraints.
23449 * configure.ac (gcc_cv_as_micromips): Check if linker
23450 supports the .set micromips directive.
23451 * configure: Regenerate.
23452 * config.in: Regenerate.
23453 * config/mips/mips-tables.opt: Regenerate.
23454 * config/mips/micromips.md: New file.
23455 * constraints.md (ZC, ZD): New constraints.
23456 * config/mips/predicates.md (movep_src_register): New predicate.
23457 (movep_src_operand): New predicate.
23458 (non_volatile_mem_operand): New predicate.
23459 * config/mips/mips.md (multimem): New type.
23460 (length): Differentiate between 17-bit and 18-bit branch offsets.
23461 (MOVEP1, MOVEP2): New mode iterator.
23462 (mov_<load>l): Use ZC constraint.
23463 (mov_<load>r): Likewise.
23464 (mov_<store>l): Likewise.
23465 (mov_<store>r): Likewise.
23466 (*branch_equality<mode>_inverted): Add microMIPS support.
23467 (*branch_equality<mode>): Likewise.
23468 (*jump_absolute): Likewise.
23469 (indirect_jump_<mode>): Likewise.
23470 (tablejump_<mode>): Likewise.
23471 (<optab>_internal): Likewise.
23472 (sibcall_internal): Likewise.
23473 (sibcall_value_internal): Likewise.
23474 (prefetch): Use constraint ZD.
23475 * config/mips/mips.opt (minterlink-compressed): New option.
23476 (minterlink-mips16): Now an alias for minterlink-compressed.
23477 (mmicromips): New option.
23478 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23479 (compare_and_swap_12): Likewise.
23480 (sync_add<mode>): Likewise.
23481 (sync_<optab>_12): Likewise.
23482 (sync_old_<optab>_12): Likewise.
23483 (sync_new_<optab>_12): Likewise.
23484 (sync_nand_12): Likewise.
23485 (sync_old_nand_12): Likewise.
23486 (sync_new_nand_12): Likewise.
23487 (sync_sub<mode>): Likewise.
23488 (sync_old_add<mode>): Likewise.
23489 (sync_old_sub<mode>): Likewise.
23490 (sync_new_add<mode>): Likewise.
23491 (sync_new_sub<mode>): Likewise.
23492 (sync_<optab><mode>): Likewise.
23493 (sync_old_<optab><mode>): Likewise.
23494 (sync_new_<optab><mode>): Likewise.
23495 (sync_nand<mode>): Likewise.
23496 (sync_old_nand<mode>): Likewise.
23497 (sync_new_nand<mode>): Likewise.
23498 (sync_lock_test_and_set<mode>): Likewise.
23499 (test_and_set_12): Likewise.
23500 (atomic_compare_and_swap<mode>): Likewise.
23501 (atomic_exchange<mode>_llsc): Likewise.
23502 (atomic_fetch_add<mode>_llsc): Likewise.
23503 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23504 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23505 (umips_save_restore_pattern_p): Likewise.
23506 (umips_load_store_pair_p): Likewise.
23507 (umips_output_load_store_pair): Likewise.
23508 (umips_movep_target_p): Likewise.
23509 (umips_12bit_offset_address_p): Likewise.
23510 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23511 (mips_base_mips16): Rename this...
23512 (mips_base_compression_flags): ...to this. Update all uses.
23513 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23514 (mips_mips16_decl_p): Delete.
23515 (mips_nomips16_decl_p): Delete.
23516 (mips_get_compress_on_flags): New function.
23517 (mips_get_compress_off_flags): New function.
23518 (mips_get_compress_mode): New function.
23519 (mips_get_compress_on_name): New function.
23520 (mips_get_compress_off_name): New function.
23521 (mips_insert_attributes): Support multiple compression types.
23522 (mips_merge_decl_attributes): Likewise.
23523 (umips_12bit_offset_address_p): New function.
23524 (mips_start_function_definition): Emit .set micromips directive.
23525 (mips_call_may_need_jalx_p): New function.
23526 (mips_function_ok_for_sibcall): Add microMIPS support.
23527 (mips_print_operand_punctuation): Support short delay slots and
23529 (umips_swm_mask, umips_swm_encoding): New.
23530 (umips_build_save_restore): New function.
23531 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23532 (was_mips16_p): Remove.
23533 (old_compression_mode): New.
23534 (mips_set_compression_mode): New function.
23535 (mips_set_current_function): Add microMIPS support.
23536 (mips_option_override): Likewise.
23537 (umips_save_restore_pattern_p): New function.
23538 (umips_output_save_restore): New function.
23539 (umips_load_store_pair_p_1): New function.
23540 (umips_load_store_pair_p): New function.
23541 (umips_output_load_store_pair_1): New function.
23542 (umips_output_load_store_pair): New function.
23543 (umips_movep_target_p) New function.
23544 (mips_prepare_pch_save): Add microMIPS support.
23545 * config/mips/mips.h (TARGET_COMPRESSION): New.
23546 (TARGET_CPU_CPP_BUILTINS): Update macro
23547 to use new compression flags and to support microMIPS.
23548 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23549 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23550 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23551 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23552 (ASM_SPEC): Support mmicromips and mno-micromips.
23553 (M16STORE_REG_P): New macro.
23554 (MIPS_CALL): Support TARGET_MICROMIPS.
23555 (MICROMIPS_J): New macro.
23556 (mips_base_mips16): Rename this...
23557 (mips_base_compression_flags): ...to this.
23558 (UMIPS_12BIT_OFFSET_P): New macro.
23559 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23560 (MULTILIB_DIRNAMES): Likewise.
23561 2013-03-20 Richard Biener <rguenther@suse.de>
23563 PR tree-optimization/56661
23564 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
23565 the result does not have to be distinct.
23567 2013-03-20 Richard Biener <rguenther@suse.de>
23569 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
23572 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23573 Steven Bosscher <steven@gcc.gnu.org>
23575 PR rtl-optimization/56605
23576 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
23578 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
23581 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
23582 that require movd instead of movq.
23584 2013-03-20 Richard Biener <rguenther@suse.de>
23586 * tree-ssa-structalias.c (struct variable_info): Add pointer
23587 to the first field of an aggregate with sub-vars. Make
23588 this and the pointer to the next subfield its ID.
23589 (vi_next): New function.
23590 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
23591 storedanything_id, integer_id): Increment by one.
23592 (new_var_info, get_call_vi, lookup_call_clobber_vi,
23593 get_call_clobber_vi): Adjust.
23594 (solution_set_expand): Simplify and speedup.
23595 (solution_set_add): Inline into ...
23596 (set_union_with_increment): ... this. Adjust accordingly.
23597 (do_sd_constraint): Likewise.
23598 (do_ds_constraint): Likewise.
23599 (do_complex_constraint): Simplify.
23600 (build_pred_graph): Adjust.
23601 (solve_graph): Likewise. Simplify and speedup.
23602 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
23603 get_constraint_for_component_ref, get_constraint_for_1,
23604 first_vi_for_offset, first_or_preceding_vi_for_offset,
23605 create_function_info_for, create_variable_info_for_1,
23606 create_variable_info_for, intra_create_variable_infos): Adjust.
23607 (init_base_vars): Push NULL for ID zero.
23608 (compute_points_to_sets): Adjust.
23610 2013-03-20 Richard Biener <rguenther@suse.de>
23612 * cfgloop.c (verify_loop_structure): Streamline and avoid
23613 ICEing on corrupt loop tree.
23614 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
23617 2013-03-20 Richard Biener <rguenther@suse.de>
23619 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
23620 check whether an SSA update is needed.
23622 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
23624 * config/mips/constraints.md (T): Rename to...
23628 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
23629 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
23631 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23633 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
23634 (*subsi3_carryin_uxtw): Likewise.
23636 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23638 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
23639 (*rorsi3_insn_uxtw): Likewise.
23641 2013-03-19 Ian Bolton <ian.bolton@arm.com>
23643 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
23644 (*extrsi5_insn_uxtw): Likewise.
23646 2013-03-19 Richard Biener <rguenther@suse.de>
23648 PR tree-optimization/56273
23649 * passes.c (init_optimization_passes): Move second VRP after DOM.
23651 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
23653 * config/i386/i386.md (*movti_internal): Merge from
23654 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
23655 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
23656 nox64 isa attributes.
23658 2013-03-18 Richard Biener <rguenther@suse.de>
23660 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
23662 (merge_node_constraints): Likewise.
23663 (build_succ_graph): Likewise.
23664 (valid_graph_edge): Inline into single caller.
23665 (unify_nodes): Likewise. Use bitmap_set_bit return value
23667 (scc_visit): Fix formatting and variable use.
23668 (do_sd_constraint): Use gcc_checking_assert.
23669 (do_ds_constraint): Likewise.
23670 (do_complex_constraint): Likewise.
23671 (condense_visit): Likewise. Cleanup.
23672 (dump_pred_graph): New function.
23673 (perform_var_substitution): Dump the pred-graph before
23674 variable substitution.
23675 (find_equivalent_node): Use gcc_checking_assert.
23676 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
23678 2013-03-18 Richard Biener <rguenther@suse.de>
23680 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
23681 Remove cond_expr_stmt_list argument and do not gimplify the
23683 (vect_loop_versioning): Adjust.
23684 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
23685 Cleanup to use less temporaries.
23686 (vect_create_data_ref_ptr): Cleanup.
23688 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23690 PR tree-optimization/56635
23691 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
23692 require types_compatible_p types.
23694 2013-03-18 Nick Clifton <nickc@redhat.com>
23696 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
23697 spurious backslash.
23699 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
23700 Add missing line to comment describing function.
23702 2013-03-18 Richard Biener <rguenther@suse.de>
23704 PR tree-optimization/56210
23705 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
23706 Handle string / character search functions.
23707 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
23709 2013-03-18 Richard Biener <rguenther@suse.de>
23711 PR middle-end/56483
23712 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
23713 and implement properly.
23714 * gimple.h (gimple_cond_single_var_p): Remove.
23716 2013-03-18 Richard Biener <rguenther@suse.de>
23718 * tree-data-ref.h (find_data_references_in_loop): Declare.
23719 * tree-data-ref.c (get_references_in_stmt): Use a stack
23720 vector pre-allocated in the callers.
23721 (find_data_references_in_stmt): Adjust.
23722 (graphite_find_data_references_in_stmt): Likewise.
23723 (create_rdg_vertices): Likewise.
23724 (find_data_references_in_loop): Export.
23725 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
23726 Compute dependences here...
23727 (vect_analyze_data_refs): ...not here. When we encounter
23728 a non-vectorizable data reference in basic-block vectorization
23729 truncate the data reference vector. Do not bother to
23730 fixup data-dependence information for gather loads.
23731 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
23732 of data references, as reported.
23734 2013-03-18 Richard Biener <rguenther@suse.de>
23736 PR tree-optimization/3713
23737 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
23738 has_constants and expr.
23739 (stmt_has_constants): Properly valueize SSA names when deciding
23740 whether the stmt has constants.
23742 2013-03-18 Richard Biener <rguenther@suse.de>
23744 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
23745 whole function when there is nothing to do.
23746 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
23747 * tree-vectorizer.c (vectorize_loops): Update virtual and
23748 loop-closed SSA once.
23749 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
23751 2013-03-18 Richard Biener <rguenther@suse.de>
23753 PR middle-end/56113
23754 * domwalk.c (bb_postorder): New global static.
23755 (cmp_bb_postorder): New function.
23756 (walk_dominator_tree): Replace scheme imposing an order for
23757 visiting dominator sons by one sorting them at the time they
23758 are pushed on the stack.
23760 2013-03-18 Richard Biener <rguenther@suse.de>
23762 PR tree-optimization/39326
23763 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
23764 (struct mem_ref): Replace mem member with ao_ref typed member.
23765 (MEM_ANALYZABLE): Adjust.
23766 (memref_eq): Likewise.
23767 (mem_ref_alloc): Likewise.
23768 (gather_mem_refs_stmt): Likewise.
23769 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
23770 (execute_sm_if_changed_flag_set): Adjust.
23771 (execute_sm): Likewise.
23772 (ref_always_accessed_p): Likewise.
23773 (refs_independent_p): Likewise.
23774 (can_sm_ref_p): Likewise.
23776 2013-03-18 Jakub Jelinek <jakub@redhat.com>
23779 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
23780 return 1 even for !unsignedp.
23782 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23784 * config/i386/i386.md (isa): Add x64 and nox64.
23785 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
23786 (*pushtf): Enable *roF alternative for x64 isa only.
23787 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
23788 mode attribute of integer alternatives to DImode for TARGET_64BIT.
23789 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
23790 (*movtf_internal): Merge from *movtf_internal_rex64 and
23791 *movtf_internal_sse. Use x64 and nox64 isa attributes.
23792 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
23793 nox64 isa attributes.
23794 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
23795 nox64 isa attributes.
23796 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
23798 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23800 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23802 2013-03-17 Jakub Jelinek <jakub@redhat.com>
23805 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
23806 class names. Remove trailing comma after "ALL_REGS".
23808 2013-03-16 Jan Hubicka <jh@suse.cz>
23810 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
23811 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
23812 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
23813 of cgraph_get_create_node.
23814 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
23816 2013-03-16 Jason Merrill <jason@redhat.com>
23819 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
23820 with DW_AT_default_value.
23822 2013-03-16 Jakub Jelinek <jakub@redhat.com>
23824 * BASE-VER: Set to 4.9.0.
23826 2013-03-14 Andi Kleen <ak@linux.intel.com>
23829 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
23830 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
23831 Document _x* TSX intrinsics.
23833 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
23834 David Holsgrove <david.holsgrove@xilinx.com>
23836 * configure.ac: Add MicroBlaze TLS support detection.
23837 * configure: Regenerate.
23838 * config/microblaze/microblaze-protos.h
23839 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
23840 symbol_mentioned_p, label_mentioned_p): Add prototypes.
23841 * config/microblaze/microblaze.c (microblaze_address_type): Add
23842 ADDRESS_TLS and tls_reloc address types.
23843 (microblaze_address_info): Add tls_reloc.
23844 (TARGET_HAVE_TLS): Define.
23845 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
23846 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
23847 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
23848 load_tls_operand, microblaze_call_tls_get_addr,
23849 microblaze_legitimize_tls_address): New functions.
23850 (microblaze_classify_unspec): Handle UNSPEC_TLS.
23851 (get_base_reg): Use microblaze_tls_symbol_p.
23852 (microblaze_classify_address): Handle TLS.
23853 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
23854 label_mentioned_p and microblaze_tls_referenced_p.
23855 (microblaze_legitimize_address): Handle TLS.
23856 (microblaze_address_insns): Handle ADDRESS_TLS.
23857 (pic_address_needs_scratch): Handle TLS.
23858 (print_operand_address): Handle TLS.
23859 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
23860 (microblaze_expand_move): Handle TLS.
23861 (microblaze_legitimate_constant_p): Check
23862 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
23863 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
23864 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
23865 (PIC_OFFSET_TABLE_REGNUM): Set.
23866 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
23867 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
23868 (addsi3, movsi_internal2, movdf_internal): Update constraints
23869 * config/microblaze/predicates.md (arith_plus_operand): Define
23870 (move_operand): Redefine as move_src_operand,
23871 check microblaze_tls_referenced_p.
23873 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23875 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
23876 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
23878 2013-03-14 Ian Bolton <ian.bolton@arm.com>
23880 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
23883 2013-03-14 Jakub Jelinek <jakub@redhat.com>
23885 PR tree-optimization/53265
23886 * common.opt (Waggressive-loop-optimizations): New option.
23887 * tree-ssa-loop-niter.c: Include tree-pass.h.
23888 (do_warn_aggressive_loop_optimizations): New function.
23889 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
23890 if number_of_latch_executions returned constant.
23891 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
23892 early. If number_of_latch_executions returned constant, set
23893 nb_iterations_upper_bound back to it.
23894 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
23896 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
23897 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
23899 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
23900 (MULTILIB_OSDIRNAMES): Set.
23901 * genmultilib: If defaultosdirname doesn't start with :: , set
23902 defaultosdirname2 instead, clear it and emit two . multilib_raw
23903 entries instead of just one.
23905 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
23907 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23908 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
23909 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
23910 (SUBTARGET_OVERRIDE_OPTIONS): New.
23912 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
23915 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
23916 (musermode): Convert to Var(TARGET_USERMODE).
23917 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
23918 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
23919 * config/sh/sh.c (sh_option_override): Use
23920 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
23921 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
23923 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
23925 (udivsi3_i4_single, divsi3_i4_single): Use
23926 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
23928 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
23930 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
23933 2013-03-13 Richard Biener <rguenther@suse.de>
23935 PR tree-optimization/56608
23936 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
23937 calls when vectorizing basic-blocks.
23939 2013-03-13 Jakub Jelinek <jakub@redhat.com>
23942 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
23945 2013-03-12 Jakub Jelinek <jakub@redhat.com>
23947 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
23949 2013-03-11 Jan Hubicka <jh@suse.cz>
23952 * lto-streamer-out.c (output_symbol_p): Skip references from
23953 constructors of external variables.
23955 2013-03-11 Jan Hubicka <jh@suse.cz>
23957 PR middle-end/56571
23958 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
23960 * emit-rtl.c (verify_rtx_sharing): Likewise.
23961 (copy_insn_1): Likewise.
23962 * rtl.c (copy_rtx): Likewise.
23964 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
23967 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
23968 output_operand_lossage message.
23970 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
23973 * arm.c (shift_op): Validate RTL pattern on the fly.
23974 (arm_print_operand, case 'S'): Don't use shift_operator to validate
23977 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23980 * config/pa/pa.md (call_value): Check for calls to powf and direct to
23981 new call patterns that clobber %fr12.
23982 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
23983 split and postreload patterns.
23984 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
23985 registers %fr12 and %fr12R as call used.
23987 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
23989 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
23990 (canon_address, record_store, replace_read, check_mem_read_rtx,
23991 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
23992 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
23993 rest_of_handle_dse): Likewise.
23995 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
23997 PR middle-end/56524
23998 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
24000 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
24001 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
24002 (save_optabs_if_changed): Replace with...
24003 (init_tree_optimization_optabs): ...this.
24004 * optabs.c (save_optabs_if_changed): Rename to...
24005 (init_tree_optimization_optabs): ...this. Take the optimization node
24006 as argument. Do nothing if the base optabs are already correct.
24007 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
24008 to recompute optabs.
24009 * function.h (function): Remove optabs field.
24010 * function.c (invoke_set_current_function_hook): Call
24011 init_tree_optimization_optabs. Use the result to initialize
24014 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
24016 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
24017 if GTMA_HAS_NO_INSTRUMENTATION.
24018 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
24019 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
24020 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
24021 * gimple-pretty-print.c (dump_gimple_transaction): Handle
24022 GTMA_HAS_NO_INSTRUMENTATION.
24024 2013-03-08 Jakub Jelinek <jakub@redhat.com>
24026 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
24029 2013-03-08 Marek Polacek <polacek@redhat.com>
24030 Jakub Jelinek <jakub@redhat.com>
24032 PR tree-optimization/56478
24033 * predict.c (is_comparison_with_loop_invariant_p): Change the
24034 type of loop_step to tree.
24035 (predict_loops): Adjust.
24036 (predict_iv_comparison): Perform the computations on double_ints.
24038 2013-03-08 Richard Biener <rguenther@suse.de>
24040 PR tree-optimization/56570
24041 * tree-cfg.c (verify_expr_location_1): Verify locations for
24043 * tree-sra.c (create_access_replacement): Strip locations
24044 from DECL_DEBUG_EXPRs.
24046 2013-03-08 Richard Biener <rguenther@suse.de>
24048 * tree-inline.c (expand_call_inline): Do not associate
24049 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
24050 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
24052 2013-03-08 Richard Biener <rguenther@suse.de>
24054 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
24055 or block changes with -Og. Fix for location / block encoding
24056 changes and PHI arguments with locations.
24058 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
24060 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
24062 (struct output_info): Likewise.
24063 (register_overhead): Remove bad gcc_assert.
24064 (bitmap_find_bit): If there is only a single bitmap element, do not
24065 count a miss as a search.
24066 (print_statistics): Update for counter type changes.
24067 (dump_bitmap_statistics): Likewise. Print headers such that they
24068 are properly lined up with the printed counters.
24070 2013-03-07 Jakub Jelinek <jakub@redhat.com>
24072 PR tree-optimization/56559
24073 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
24074 check that it has only a single use.
24076 2013-03-07 Richard Biener <rguenther@suse.de>
24078 * doc/invoke.texi (fwhole-program): Discourage use in combination
24081 2013-03-06 Jakub Jelinek <jakub@redhat.com>
24083 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
24085 PR tree-optimization/56539
24086 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
24087 instead of GSI_CONTINUE_LINKING as last argument to
24088 force_gimple_operand_gsi. Adjust function comment.
24090 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
24093 PR middle-end/56548
24094 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
24095 promoted mode, convert the result back to the original mode.
24097 2013-03-06 Richard Biener <rguenther@suse.de>
24099 PR middle-end/56294
24100 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
24101 (insert_updated_phi_nodes_compare_uids): New function.
24102 (update_ssa): Sort symbols_to_rename after UID before
24103 traversing it to insert PHI nodes.
24105 2013-03-06 Richard Biener <rguenther@suse.de>
24107 PR middle-end/50494
24108 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
24109 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
24112 2013-02-13 Richard Biener <rguenther@suse.de>
24115 * varasm.c (output_constant_def_1): Get the decl representing
24116 the constant as argument.
24117 (output_constant_def): Wrap output_constant_def_1.
24118 (make_decl_rtl): Use output_constant_def_1 with the decl
24119 representing the constant.
24120 (build_constant_desc): Optionally re-use a decl already
24121 representing the constant.
24122 (tree_output_constant_def): Adjust.
24124 2013-03-06 Joey Ye <joey.ye@arm.com>
24127 * gcc.c (convert_white_space): New function.
24128 (main): Handles white space in function name.
24130 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
24133 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
24134 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
24135 to SH_DIV_CALL_TABLE for TARGET_SH2.
24136 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
24138 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
24139 call-table options.
24141 2013-03-05 Sterling Augustine <saugustine@google.com>
24142 Cary Coutant <ccoutant@google.com>
24145 * dwarf2out.c (resolve_addr): Don't call
24146 remove_loc_list_addr_table_entries a second time for the same
24149 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24152 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
24153 (avoid_complex_debug_insns): New function.
24154 (expand_debug_locations): Call it.
24156 PR rtl-optimization/56484
24157 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
24158 lifetimes of hard registers on small register class machines.
24160 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24162 * config/microblaze/microblaze-protos.h: Rename
24163 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
24164 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
24166 (microblaze_fast_interrupt_function_p): New function.
24167 (microblaze_is_interrupt_handler): Rename to
24168 microblaze_is_interrupt_variant and add fast_interrupt check.
24169 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
24170 (save_restore_insns): Likewise.
24171 (compute_frame_size): Likewise.
24172 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
24173 (microblaze_globalize_label): Likewise.
24174 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
24175 * config/microblaze/microblaze.md: Use wrapper
24176 microblaze_is_interrupt_variant.
24178 2013-03-05 Kai Tietz <ktietz@redhat.com>
24180 * sdbout.c (sdbout_one_type): Switch to current function's section
24181 supporting cold/hot.
24183 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
24185 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
24188 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24190 PR middle-end/56461
24191 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
24192 if VALGRIND_GET_VBITS is defined, temporarily make object
24193 memory all defined, and restore previous valgrind addressability
24194 and definability afterwards. Free this_object at the end.
24196 PR middle-end/56461
24197 * lra.c (lra): Call lra_clear_live_ranges if live_p,
24198 right before calling lra_create_live_ranges, also call it
24199 when clearing live_p. Only call lra_clear_live_ranges
24200 at the end if live_p.
24202 PR middle-end/56461
24203 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
24205 2013-03-05 Richard Biener <rguenther@suse.de>
24207 PR tree-optimization/56521
24208 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
24211 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
24214 * except.h (remove_unreachable_eh_regions): New prototype.
24215 * except.c (remove_eh_handler_splicer): New function, split out
24216 of remove_eh_handler.
24217 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
24218 warning about running it on many EH regions one at a time.
24219 (remove_unreachable_eh_regions_worker): New function, walk the
24220 EH tree in depth-first order and remove non-marked regions.
24221 (remove_unreachable_eh_regions): New function.
24222 * tree-eh.c (mark_reachable_handlers): New function, split out
24223 from remove_unreachable_handlers.
24224 (remove_unreachable_handlers): Use mark_reachable_handlers and
24225 remove_unreachable_eh_regions.
24226 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
24227 and remove_unreachable_eh_regions.
24229 2013-03-05 Richard Biener <rguenther@suse.de>
24231 PR middle-end/56525
24232 * loop-init.c (fix_loop_structure): Remove loops in two stages,
24233 not freeing them until the end.
24235 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24237 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
24239 2013-03-05 Richard Biener <rguenther@suse.de>
24241 PR tree-optimization/56270
24242 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
24243 of loads after scheduling an SLP instance.
24245 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24247 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
24249 (check_gcc_parallelize): Run guality.exp as a separate job from
24250 vect.exp with unsorted.exp and $(dg_target_exps) separately from
24251 struct-layout-1.exp with stackalign.exp.
24253 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
24255 PR middle-end/56461
24256 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
24257 load_index sbitmap even if some bit in it isn't set.
24259 PR middle-end/56461
24260 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
24261 (discover_iteration_bound_by_body_walk): Change queues to
24262 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
24263 spelling in comment. Call safe_push on queues[bound_index] directly.
24264 Release queues[queue_index] in every iteration unconditionally.
24265 Release bounds vector.
24267 PR middle-end/56461
24268 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
24269 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
24270 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
24273 2013-03-05 Richard Biener <rguenther@suse.de>
24276 * tree-inline.c (remap_blocks_to_null): New function.
24277 (expand_call_inline): When expanding a call stmt without
24278 an associated block inline remap all callee blocks to NULL.
24280 2013-03-05 Jakub Jelinek <jakub@redhat.com>
24282 PR rtl-optimization/56494
24283 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
24284 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
24285 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
24287 PR middle-end/56461
24288 * sel-sched-ir.c (free_sched_pools): Release
24289 succs_info_pool.stack[succs_info_pool.max_top] vectors too
24290 if succs_info_pool.max_top isn't -1.
24293 * opts.c (opts_obstack, opts_concat): Moved to...
24294 * opts-common.c (opts_obstack, opts_concat): ... here.
24296 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24298 PR middle-end/56461
24299 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
24301 2013-03-04 Martin Jambor <mjambor@suse.cz>
24303 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
24304 all appropriate places.
24306 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
24308 PR tree-optimization/56424
24309 * ipa-split.c (split_function): Do not set the RSO flag if result is
24310 not by reference and its type is a register type.
24312 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
24314 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
24315 (microblaze_legitimate_pic_operand): Likewise
24316 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
24317 new function microblaze_legitimate_pic_operand
24318 * config/microblaze/microblaze-protos.h
24319 (microblaze_legitimate_pic_operand): Declare.
24321 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24323 * config/microblaze/predicates.md (call_insn_simple_operand):
24324 New predicate for supported rtx code types.
24325 * config/microblaze/microblaze.md (call_internal1): Use
24326 call_insn_simple_operand predicate.
24328 2013-03-04 Jakub Jelinek <jakub@redhat.com>
24330 PR middle-end/56461
24331 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
24332 partitions.ordered_remove.
24334 PR middle-end/56461
24335 * tree-vect-stmts.c (vectorizable_conversion): Don't call
24336 vec_oprnds0.create (1) for modifier == NONE.
24338 PR middle-end/56461
24339 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
24340 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
24341 vec_oprnds1 right before pushing anything to it for
24344 PR middle-end/56461
24345 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
24346 set nbbs to 0 instead of having separate code path.
24347 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
24348 instead of false as last argument if returning NULL.
24350 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
24352 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
24353 the attribute is now called "target" instead of "option".
24354 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
24355 * doc/tm.texi.in (Target Attributes): Likewise document the correct
24356 attribute/pragma name for TARGET_OPTION_VALID_P and
24357 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
24358 * doc/tm.texi: Regenerated.
24360 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
24362 * config/microblaze/microblaze.c:
24363 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
24364 * config/microblaze/microblaze.h: Add -mxl-reorder to
24366 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
24367 instructions emitted if TARGET_REORDER.
24368 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
24369 or 0 for -m/-mno case, but initialises as 2 to detect default use case
24372 2013-03-01 Xinliang David Li <davidxl@google.com>
24374 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
24377 2013-03-01 Jakub Jelinek <jakub@redhat.com>
24379 PR middle-end/56461
24380 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
24381 vector even when returning true. Fix up function comment formatting.
24383 PR middle-end/56461
24384 * ira-build.c (ira_loop_nodes_count): New variable.
24385 (create_loop_tree_nodes): Initialize it.
24386 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
24388 PR middle-end/56461
24389 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
24390 method on dr_chain and result_chain.
24391 * tree-vect-stmts.c (vectorizable_store): Only call
24392 result_chain.create if j == 0.
24394 PR middle-end/56461
24395 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
24396 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
24397 before overwriting it.
24399 2013-03-01 Tobias Burnus <burnus@net-b.de>
24401 * doc/extended.texi (C Extensions): Change order in @menu
24403 (Other MIPS Built-in Functions): Move last MIPS entry before
24404 "picoChip Built-in Functions".
24405 (SH Built-in Functions): Move after RX Built-in Functions.
24406 * doc/gcc.texi (Introduction): Change order in @menu
24408 * doc/md.texi (Constraints): Ditto.
24409 * gty.texi (Type Information): Ditto.
24410 (User-provided marking routines for template types): Make
24412 * doc/invoke.texi (AArch64 Options): Move before
24413 "Adapteva Epiphany Options".
24415 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
24416 Jakub Jelinek <jakub@redhat.com>
24419 * asan.c (gate_asan): Lookup no_sanitize_address instead of
24420 no_address_safety_analysis attribute.
24421 * doc/extend.texi (no_address_safety_attribute): Rename to
24422 no_sanitize_address attribute, mention no_address_safety_analysis
24423 attribute as deprecated alias.
24425 2013-02-28 Jakub Jelinek <jakub@redhat.com>
24427 PR middle-end/56461
24428 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
24429 type to vec<vec<tree> > *.
24430 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
24431 to be vec<tree> instead of vec<tree> *, set vec_defs
24432 to vNULL and call vec_defs.create (number_of_vects), adjust other
24434 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
24435 vectorizable_condition): Adjust vect_get_slp_defs callers.
24437 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24439 * config/aarch64/aarch64.c
24440 (aarch64_float_const_representable): Remove unused variable.
24442 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24444 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
24446 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24448 * config/aarch64/aarch64-builtins.c
24449 (aarch64_init_simd_builtins): Make static.
24451 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
24453 * config/aarch64/aarch64.c
24454 (aarch64_simd_make_constant): Make static.
24456 2013-02-28 Martin Jambor <mjambor@suse.cz>
24458 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
24459 with no initialization to the RHS of debug statements.
24461 2013-02-28 Martin Jambor <mjambor@suse.cz>
24463 PR tree-optimization/56294
24464 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
24466 (get_access_replacement): Do not call create_access_replacement.
24467 Assert a replacement exists.
24468 (get_repl_default_def_ssa_name): Create the replacement declaration
24471 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24473 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
24474 final_end_function.
24476 2013-02-28 Marek Polacek <polacek@redhat.com>
24478 PR rtl-optimization/56466
24479 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
24480 if we're changing a loop.
24481 (peel_loops_completely): Likewise.
24483 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
24486 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
24488 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24491 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
24492 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
24493 INTX_FTYPE_FX, FX_FTYPE_INTX.
24494 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
24496 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24498 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
24499 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
24500 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
24501 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
24502 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
24503 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
24504 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
24505 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
24506 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
24507 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
24508 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
24509 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
24510 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
24511 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
24512 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
24513 (avrxmega6): Increase max flash segments from 5 to 6.
24514 * config/avr/t-multilib: Regenerate.
24515 * config/avr/avr-tables.opt: Regenerate.
24516 * doc/avr-mmcu.texi: Regenerate.
24518 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
24520 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
24521 (avr_device_to_arch): Rename to avr_device_to_ld.
24522 (avr_device_to_as): New prototype.
24523 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
24524 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
24525 * config/avr/driver-avr.c (avr_device_to_as): New.
24526 (avr_device_to_arch): Rename to avr_device_to_ld.
24528 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24530 PR middle-end/56461
24531 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
24532 method on dr_chain and result_chain.
24534 PR middle-end/56461
24535 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
24536 pointer_set_destroy on not_executed_last_iteration.
24538 PR middle-end/56461
24539 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
24541 PR middle-end/56461
24542 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
24543 FOR_EACH_DEFINED_FUNCTION when freeing state.
24545 PR middle-end/56461
24546 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
24548 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
24551 PR middle-end/56461
24552 * ipa-cp.c (decide_whether_version_node): Call vec_free on
24553 known_aggs[i].items and release known_aggs vector.
24555 PR middle-end/56461
24556 * ipa-reference.c (propagate): Free node_info even for alias nodes.
24558 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24560 * config/microblaze/microblaze.c (microblaze_emit_compare):
24561 Use xor for EQ/NE comparisions.
24562 * config/microblaze/microblaze.md (cstoresf4): Add constraints
24563 (cbranchsf4): Adjust operator to comparison_operator.
24565 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24567 PR middle-end/56461
24568 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
24570 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
24571 vec_safe_push, always update *slot.
24572 (redirect_edge_var_map_clear): Use vec_free.
24573 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
24574 (free_var_map_entry): Use vec_free.
24575 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
24576 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
24578 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
24580 PR middle-end/45472
24581 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
24582 when the may_trap_p bit of the exprs being merged differs.
24583 Reorder tests for speculativeness in the logical and operator.
24585 2013-02-27 Jakub Jelinek <jakub@redhat.com>
24587 * incpath.c (add_standard_paths): Use reconcat instead of concat
24588 where appropriate and avoid leaking memory.
24590 * opts.h: Include obstack.h.
24591 (opts_concat): New prototype.
24592 (opts_obstack): New declaration.
24593 * opts.c (opts_concat): New function.
24594 (opts_obstack): New variable.
24595 (init_options_struct): Call gcc_init_obstack on opts_obstack.
24596 (finish_options): Use opts_concat instead of concat
24597 and XOBNEWVEC instead of XNEWVEC.
24598 * opts-common.c (generate_canonical_option, decode_cmdline_option,
24599 generate_option): Likewise.
24600 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
24601 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
24604 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
24605 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
24607 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24609 PR middle-end/56461
24610 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
24612 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24614 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
24615 (arm_block_move_unaligned_straight): Likewise.
24616 (arm_adjust_block_mem): Likewise.
24618 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
24621 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
24622 temp, cond and label.
24623 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
24626 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
24627 * config/c6x/c6x.h (dbx_register_map): Update declaration.
24630 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
24631 of prologue/epilogue functions.
24634 * config/tilegx/tilegx.c (tilegx_expand_prologue):
24635 Remove unused variable cfa_offset.
24636 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
24639 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
24640 type promotion to unsigned.
24643 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
24644 for HOST_WIDE_INT of 32 bit / same size as int.
24645 (arm_block_move_unaligned_straight): Likewise.
24646 (arm_adjust_block_mem): Likewise.
24649 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
24652 2013-02-26 Marek Polacek <polacek@redhat.com>
24654 PR tree-optimization/56426
24655 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
24657 2013-02-26 Richard Biener <rguenther@suse.de>
24660 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
24661 unused variable loops.
24663 2013-02-26 Jakub Jelinek <jakub@redhat.com>
24665 PR tree-optimization/56448
24666 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
24667 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
24668 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
24669 later operands of the references, or even first operand for
24670 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
24672 PR tree-optimization/56443
24673 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
24674 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
24675 to type_for_mode langhook.
24677 2013-02-25 Matt Turner <mattst88@gmail.com>
24679 * doc/invoke.texi: Document r4700.
24681 2013-02-25 Richard Biener <rguenther@suse.de>
24683 PR tree-optimization/56175
24684 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
24686 (simplify_bitwise_binary): ... here. Also guard the conversion
24687 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
24689 2013-02-25 Catherine Moore <clm@codesourcery.com>
24692 2013-02-24 Catherine Moore <clm@codesourcery.com>
24693 Maciej W. Rozycki <macro@codesourcery.com>
24694 Tom de Vries <tom@codesourcery.com>
24695 Nathan Sidwell <nathan@codesourcery.com>
24696 Iain Sandoe <iain@codesourcery.com>
24697 Nathan Froyd <froydnj@codesourcery.com>
24698 Chao-ying Fu <fu@mips.com>
24700 * doc/extend.texi: (micromips, nomicromips, nocompression):
24701 Document new function attributes.
24702 * doc/invoke.texi (minterlink-compressed, mmicromips,
24703 m14k, m14ke, m14kec): Document new options.
24704 (minterlink-mips16): Update documentation.
24705 * doc/md.texi (ZC, ZD): Document new constraints.
24706 * configure.ac (gcc_cv_as_micromips): Check if linker
24707 supports the .set micromips directive.
24708 * configure: Regenerate.
24709 * config.in: Regenerate.
24710 * config/mips/mips-tables.opt: Regenerate.
24711 * config/mips/micromips.md: New file.
24712 * constraints.md (ZC, AD): New constraints.
24713 * config/mips/predicates.md (movep_src_register): New predicate.
24714 (movep_src_operand): New predicate.
24715 (non_volatile_mem_operand): New predicate.
24716 * config/mips/mips.md (multimem): New type.
24717 (length): Differentiate between 17-bit and 18-bit branch offsets.
24718 (MOVEP1, MOVEP2): New mode iterator.
24719 (mov_<load>l): Use ZC constraint.
24720 (mov_<load>r): Likewise.
24721 (mov_<store>l): Likewise.
24722 (mov_<store>r): Likewise.
24723 (*branch_equality<mode>_inverted): Add microMIPS support.
24724 (*branch_equality<mode>): Likewise.
24725 (*jump_absolute): Likewise.
24726 (indirect_jump_<mode>): Likewise.
24727 (tablejump_<mode>): Likewise.
24728 (<optab>_internal): Likewise.
24729 (sibcall_internal): Likewise.
24730 (sibcall_value_internal): Likewise.
24731 (prefetch): Use constraint ZD.
24732 * config/mips/mips.opt (minterlink-compressed): New option.
24733 (minterlink-mips16): Now an alias for minterlink-compressed.
24734 (mmicromips): New option.
24735 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24736 (compare_and_swap_12): Likewise.
24737 (sync_add<mode>): Likewise.
24738 (sync_<optab>_12): Likewise.
24739 (sync_old_<optab>_12): Likewise.
24740 (sync_new_<optab>_12): Likewise.
24741 (sync_nand_12): Likewise.
24742 (sync_old_nand_12): Likewise.
24743 (sync_new_nand_12): Likewise.
24744 (sync_sub<mode>): Likewise.
24745 (sync_old_add<mode>): Likewise.
24746 (sync_old_sub<mode>): Likewise.
24747 (sync_new_add<mode>): Likewise.
24748 (sync_new_sub<mode>): Likewise.
24749 (sync_<optab><mode>): Likewise.
24750 (sync_old_<optab><mode>): Likewise.
24751 (sync_new_<optab><mode>): Likewise.
24752 (sync_nand<mode>): Likewise.
24753 (sync_old_nand<mode>): Likewise.
24754 (sync_new_nand<mode>): Likewise.
24755 (sync_lock_test_and_set<mode>): Likewise.
24756 (test_and_set_12): Likewise.
24757 (atomic_compare_and_swap<mode>): Likewise.
24758 (atomic_exchange<mode>_llsc): Likewise.
24759 (atomic_fetch_add<mode>_llsc): Likewise.
24760 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24761 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24762 (umips_save_restore_pattern_p): Likewise.
24763 (umips_load_store_pair_p): Likewise.
24764 (umips_output_load_store_pair): Likewise.
24765 (umips_movep_target_p): Likewise.
24766 (umips_12bit_offset_address_p): Likewise.
24767 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24768 (mips_base_mips16): Rename this...
24769 (mips_base_compression_flags): ...to this. Update all uses.
24770 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24771 (mips_mips16_decl_p): Delete.
24772 (mips_nomips16_decl_p): Delete.
24773 (mips_get_compress_on_flags): New function.
24774 (mips_get_compress_off_flags): New function.
24775 (mips_get_compress_mode): New function.
24776 (mips_get_compress_on_name): New function.
24777 (mips_get_compress_off_name): New function.
24778 (mips_insert_attributes): Support multiple compression types.
24779 (mips_merge_decl_attributes): Likewise.
24780 (umips_12bit_offset_address_p): New function.
24781 (mips_start_function_definition): Emit .set micromips directive.
24782 (mips_call_may_need_jalx_p): New function.
24783 (mips_function_ok_for_sibcall): Add microMIPS support.
24784 (mips_print_operand_punctuation): Support short delay slots and
24786 (umips_swm_mask, umips_swm_encoding): New.
24787 (umips_build_save_restore): New function.
24788 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24789 (was_mips16_p): Remove.
24790 (old_compression_mode): New.
24791 (mips_set_compression_mode): New function.
24792 (mips_set_current_function): Add microMIPS support.
24793 (mips_option_override): Likewise.
24794 (umips_save_restore_pattern_p): New function.
24795 (umips_output_save_restore): New function.
24796 (umips_load_store_pair_p_1): New function.
24797 (umips_load_store_pair_p): New function.
24798 (umips_output_load_store_pair_1): New function.
24799 (umips_output_load_store_pair): New function.
24800 (umips_movep_target_p) New function.
24801 (mips_prepare_pch_save): Add microMIPS support.
24802 * config/mips/mips.h (TARGET_COMPRESSION): New.
24803 (TARGET_CPU_CPP_BUILTINS): Update macro
24804 to use new compression flags and to support microMIPS.
24805 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24806 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24807 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24808 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24809 (ASM_SPEC): Support mmicromips and mno-micromips.
24810 (M16STORE_REG_P): New macro.
24811 (MIPS_CALL): Support TARGET_MICROMIPS.
24812 (MICROMIPS_J): New macro.
24813 (mips_base_mips16): Rename this...
24814 (mips_base_compression_flags): ...to this.
24815 (UMIPS_12BIT_OFFSET_P): New macro.
24816 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24817 (MULTILIB_DIRNAMES): Likewise.
24819 2013-02-25 Tom de Vries <tom@codesourcery.com>
24821 PR rtl-optimization/56131
24822 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
24823 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
24824 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
24826 2013-02-25 Tobias Burnus <burnus@net-b.de>
24828 * doc/invoke.texi (-fsanitize=): Move from optimization
24829 to debugging options.
24831 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24833 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
24835 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
24836 Alexander Monakov <amonakov@ispras.ru>
24838 PR middle-end/56077
24839 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
24840 flush pending lists also on non-jumps. Adjust comment.
24842 2013-02-24 Catherine Moore <clm@codesourcery.com>
24843 Maciej W. Rozycki <macro@codesourcery.com>
24844 Tom de Vries <tom@codesourcery.com>
24845 Nathan Sidwell <nathan@codesourcery.com>
24846 Iain Sandoe <iain@codesourcery.com>
24847 Nathan Froyd <froydnj@codesourcery.com>
24848 Chao-ying Fu <fu@mips.com>
24850 * doc/extend.texi: (micromips, nomicromips, nocompression):
24851 Document new function attributes.
24852 * doc/invoke.texi (minterlink-compressed, mmicromips,
24853 m14k, m14ke, m14kec): Document new options.
24854 (minterlink-mips16): Update documentation.
24855 * doc/md.texi (ZC, ZD): Document new constraints.
24856 * configure.ac (gcc_cv_as_micromips): Check if linker
24857 supports the .set micromips directive.
24858 * configure: Regenerate.
24859 * config.in: Regenerate.
24860 * config/mips/mips-tables.opt: Regenerate.
24861 * config/mips/micromips.md: New file.
24862 * constraints.md (ZC, AD): New constraints.
24863 * config/mips/predicates.md (movep_src_register): New predicate.
24864 (movep_src_operand): New predicate.
24865 (non_volatile_mem_operand): New predicate.
24866 * config/mips/mips.md (multimem): New type.
24867 (length): Differentiate between 17-bit and 18-bit branch offsets.
24868 (MOVEP1, MOVEP2): New mode iterator.
24869 (mov_<load>l): Use ZC constraint.
24870 (mov_<load>r): Likewise.
24871 (mov_<store>l): Likewise.
24872 (mov_<store>r): Likewise.
24873 (*branch_equality<mode>_inverted): Add microMIPS support.
24874 (*branch_equality<mode>): Likewise.
24875 (*jump_absolute): Likewise.
24876 (indirect_jump_<mode>): Likewise.
24877 (tablejump_<mode>): Likewise.
24878 (<optab>_internal): Likewise.
24879 (sibcall_internal): Likewise.
24880 (sibcall_value_internal): Likewise.
24881 (prefetch): Use constraint ZD.
24882 * config/mips/mips.opt (minterlink-compressed): New option.
24883 (minterlink-mips16): Now an alias for minterlink-compressed.
24884 (mmicromips): New option.
24885 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
24886 (compare_and_swap_12): Likewise.
24887 (sync_add<mode>): Likewise.
24888 (sync_<optab>_12): Likewise.
24889 (sync_old_<optab>_12): Likewise.
24890 (sync_new_<optab>_12): Likewise.
24891 (sync_nand_12): Likewise.
24892 (sync_old_nand_12): Likewise.
24893 (sync_new_nand_12): Likewise.
24894 (sync_sub<mode>): Likewise.
24895 (sync_old_add<mode>): Likewise.
24896 (sync_old_sub<mode>): Likewise.
24897 (sync_new_add<mode>): Likewise.
24898 (sync_new_sub<mode>): Likewise.
24899 (sync_<optab><mode>): Likewise.
24900 (sync_old_<optab><mode>): Likewise.
24901 (sync_new_<optab><mode>): Likewise.
24902 (sync_nand<mode>): Likewise.
24903 (sync_old_nand<mode>): Likewise.
24904 (sync_new_nand<mode>): Likewise.
24905 (sync_lock_test_and_set<mode>): Likewise.
24906 (test_and_set_12): Likewise.
24907 (atomic_compare_and_swap<mode>): Likewise.
24908 (atomic_exchange<mode>_llsc): Likewise.
24909 (atomic_fetch_add<mode>_llsc): Likewise.
24910 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
24911 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
24912 (umips_save_restore_pattern_p): Likewise.
24913 (umips_load_store_pair_p): Likewise.
24914 (umips_output_load_store_pair): Likewise.
24915 (umips_movep_target_p): Likewise.
24916 (umips_12bit_offset_address_p): Likewise.
24917 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
24918 (mips_base_mips16): Rename this...
24919 (mips_base_compression_flags): ...to this. Update all uses.
24920 (mips_attribute_table): Add micromips, nomicromips and nocompression.
24921 (mips_mips16_decl_p): Delete.
24922 (mips_nomips16_decl_p): Delete.
24923 (mips_get_compress_on_flags): New function.
24924 (mips_get_compress_off_flags): New function.
24925 (mips_get_compress_mode): New function.
24926 (mips_get_compress_on_name): New function.
24927 (mips_get_compress_off_name): New function.
24928 (mips_insert_attributes): Support multiple compression types.
24929 (mips_merge_decl_attributes): Likewise.
24930 (umips_12bit_offset_address_p): New function.
24931 (mips_start_function_definition): Emit .set micromips directive.
24932 (mips_call_may_need_jalx_p): New function.
24933 (mips_function_ok_for_sibcall): Add microMIPS support.
24934 (mips_print_operand_punctuation): Support short delay slots and
24936 (umips_swm_mask, umips_swm_encoding): New.
24937 (umips_build_save_restore): New function.
24938 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
24939 (was_mips16_p): Remove.
24940 (old_compression_mode): New.
24941 (mips_set_compression_mode): New function.
24942 (mips_set_current_function): Add microMIPS support.
24943 (mips_option_override): Likewise.
24944 (umips_save_restore_pattern_p): New function.
24945 (umips_output_save_restore): New function.
24946 (umips_load_store_pair_p_1): New function.
24947 (umips_load_store_pair_p): New function.
24948 (umips_output_load_store_pair_1): New function.
24949 (umips_output_load_store_pair): New function.
24950 (umips_movep_target_p) New function.
24951 (mips_prepare_pch_save): Add microMIPS support.
24952 * config/mips/mips.h (TARGET_COMPRESSION): New.
24953 (TARGET_CPU_CPP_BUILTINS): Update macro
24954 to use new compression flags and to support microMIPS.
24955 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
24956 (MIPS_ARCH_FLOAT_SPEC): Likewise.
24957 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
24958 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
24959 (ASM_SPEC): Support mmicromips and mno-micromips.
24960 (M16STORE_REG_P): New macro.
24961 (MIPS_CALL): Support TARGET_MICROMIPS.
24962 (MICROMIPS_J): New macro.
24963 (mips_base_mips16): Rename this...
24964 (mips_base_compression_flags): ...to this.
24965 (UMIPS_12BIT_OFFSET_P): New macro.
24966 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
24967 (MULTILIB_DIRNAMES): Likewise.
24969 2013-02-24 Jakub Jelinek <jakub@redhat.com>
24972 * target-globals.c (save_target_globals): For init_reg_sets and
24973 target_reinit remporarily set this_fn_optabs to this_target_optabs.
24975 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
24977 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
24978 * config/aarch64/t-aarch64
24979 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
24981 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
24983 PR inline-asm/56148
24984 * lra-constraints.c (process_alt_operands): Reload operand
24985 conflicting with earlier clobber only if no more other conflicting
24988 2013-02-22 Jakub Jelinek <jakub@redhat.com>
24991 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
24992 if not linking a shared library.
24994 2013-02-22 Seth LaForge <sethml@google.com>
24996 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
24998 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
25000 * config/arm/arm.md (split for extendsidi): Update condition.
25001 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
25002 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
25003 (qhs_zextenddi_cstr): Likewise.
25005 2013-02-21 Jakub Jelinek <jakub@redhat.com>
25007 PR middle-end/56420
25008 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
25009 avoid signed wrapping.
25010 (expand_mult): Handle properly multiplication by
25011 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
25012 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
25013 in the compiler if coeff is HOST_WIDE_INT_MIN.
25014 (expand_divmod): Don't make ext_op1 static, change it's type to
25015 uhwi. Avoid undefined behavior in -INTVAL (op1).
25017 PR rtl-optimization/50339
25018 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
25020 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
25021 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
25022 into splitting_ashiftrt field.
25023 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
25025 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
25028 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
25030 PR middle-end/56108
25031 * trans-mem.c (execute_tm_mark): Do not expand transactions that
25032 are sure to go irrevocable.
25034 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
25036 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
25037 scalars are valid operands.
25039 2013-02-21 Martin Jambor <mjambor@suse.cz>
25041 PR tree-optimization/56310
25042 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
25043 only matching indices and non-negative final offsets.
25044 (intersect_aggregates_with_edge): Pass src_idx to
25045 agg_replacements_to_vector. Pass src_idx insstead of index to
25046 intersect_with_agg_replacements.
25048 2013-02-21 Martin Jambor <mjambor@suse.cz>
25050 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
25051 instead of hard-wired defaults.
25053 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
25055 * doc/invoke.texi (MIPS Options): Update documentation of the
25056 floating-point multiply-accumulate instruction restrictions.
25058 2013-02-21 Kostya Serebryany <kcc@google.com>
25060 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
25061 asan_shadow_offset on x86_64 linux.
25063 2013-02-21 Richard Biener <rguenther@suse.de>
25065 PR tree-optimization/56415
25067 2013-02-11 Richard Biener <rguenther@suse.de>
25069 PR tree-optimization/56273
25070 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25073 2013-02-21 Jakub Jelinek <jakub@redhat.com>
25076 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
25079 PR inline-asm/56405
25080 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
25081 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
25083 2013-02-20 Jan Hubicka <jh@suse.cz>
25085 PR tree-optimization/56265
25086 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
25087 when target is referenced for first time.
25089 2013-02-20 Richard Biener <rguenther@suse.de>
25091 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
25092 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
25093 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
25094 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
25095 not return anything.
25096 (rename_ssa_copies): Do not remove unused locals.
25097 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
25098 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
25099 * passes.c (execute_function_todo): Do not schedule unused locals
25100 removal if cleanup_tree_cfg did something.
25101 * tree-ssa-live.c (remove_unused_locals): Dump statistics
25102 about the number of removed locals.
25104 2013-02-20 Richard Biener <rguenther@suse.de>
25106 PR tree-optimization/56398
25107 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
25109 2013-02-20 Martin Jambor <mjambor@suse.cz>
25111 PR tree-optimization/55334
25112 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
25113 restricted pointers to arrays.
25115 2013-02-20 Richard Biener <rguenther@suse.de>
25116 Jakub Jelinek <jakub@redhat.com>
25118 PR tree-optimization/56396
25119 * tree-ssa-ccp.c (n_const_val): New static variable.
25120 (get_value): Return NULL for SSA names we don't have a lattice
25122 (ccp_initialize): Initialize n_const_val.
25123 * tree-ssa-copy.c (n_copy_of): New static variable.
25124 (init_copy_prop): Initialize n_copy_of.
25125 (get_value): Return NULL_TREE for SSA names we don't have a
25128 2013-02-20 Martin Jambor <mjambor@suse.cz>
25130 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
25132 2013-02-20 Richard Biener <rguenther@suse.de>
25134 * genpreds.c (write_lookup_constraint): Do not compare first
25135 letter of the constraint again.
25137 2013-02-20 Richard Biener <rguenther@suse.de>
25139 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
25141 (get_use_iv_cost): Terminate hashtable walk when coming across
25144 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
25146 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
25147 reassociation for avx2 targets.
25149 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25151 * config/microblaze/microblaze.c: microblaze_has_clz = 0
25152 Add version check for v8.10.a to enable microblaze_has_clz
25153 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
25154 version and TARGET_PATTERN_COMPARE check
25155 * config/microblaze/microblaze.md: New clzsi2 instruction
25157 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25159 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
25160 function before branching.
25162 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25164 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
25166 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
25168 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
25170 PR middle-end/55889
25171 * sel-sched.c: Include ira.h.
25172 (implicit_clobber_conflict_p): New function.
25173 (moveup_expr): Use it.
25174 * Makefile.in (sel-sched.o): Depend on ira.h.
25176 2013-02-19 Richard Biener <rguenther@suse.de>
25178 PR tree-optimization/56384
25179 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
25180 (vn_hash_type): Split out from ...
25181 (vn_hash_constant_with_type): ... here.
25182 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
25183 (vn_phi_eq): Compare types from vn_phi_s structure.
25184 (vn_phi_lookup): Populate vn_phi_s type.
25185 (vn_phi_insert): Likewise.
25187 2013-02-19 Jakub Jelinek <jakub@redhat.com>
25189 PR tree-optimization/56350
25190 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
25191 if haven't found reduction or nested cycle operand, rather than
25192 asserting we must find it.
25194 PR tree-optimization/56381
25195 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
25198 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
25199 Jakub Jelinek <jakub@redhat.com>
25202 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
25203 (swap_optab_enable): Same.
25204 (init_all_optabs): Use argument instead of global.
25205 * tree.h (struct tree_optimization_option): New field target_optabs.
25206 * expr.h (init_all_optabs): Add argument to prototype.
25207 (TREE_OPTIMIZATION_OPTABS): New.
25208 (save_optabs_if_changed): Protoize.
25209 * optabs.h: Declare this_fn_optabs.
25210 * optabs.c (save_optabs_if_changed): New.
25211 Declare this_fn_optabs.
25212 (init_optabs): Add argument to init_all_optabs() call.
25213 * function.c (invoke_set_current_function_hook): Handle per
25215 * function.h (struct function): New field optabs.
25216 * config/mips/mips.c (mips_set_mips16_mode): Handle when
25217 optimization_current_node has changed.
25218 * target-globals.h (save_target_globals_default_opts): Protoize.
25219 * target-globals.c (save_target_globals_default_opts): New.
25221 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25224 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
25225 registers %fr12 and %fr12R as call used.
25228 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
25229 and HImode, require all displacements to be an integer multiple of
25231 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
25232 only allow QImode and HImode when reload is in progress and strict is
25233 true. Likewise for symbolic addresses. Use base14_operand to check
25234 displacements in REG+BASE addresses.
25236 2013-02-18 Richard Biener <rguenther@suse.de>
25238 PR tree-optimization/56366
25239 * tree-vect-loop.c (get_initial_def_for_induction): Properly
25240 handle sign-conversion of outer-loop initial induction value.
25242 2013-02-18 Richard Biener <rguenther@suse.de>
25244 PR middle-end/56349
25245 * cfghooks.c (merge_blocks): If we merge a latch into another
25246 block adjust references to it.
25247 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
25248 (verify_loop_structure): Verify that a recorded latch is in fact
25251 2013-02-18 Richard Biener <rguenther@suse.de>
25253 PR tree-optimization/56321
25254 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
25255 order SSA name release and virtual operand unlinking.
25257 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25259 * config/microblaze/microblaze.md (save_stack_block): Define.
25260 (restore_stack_block): Likewise.
25262 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25264 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
25265 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
25266 * config/microblaze/microblaze.c (microblaze_option_override):
25267 Bail out early for PIC modes when target does not support PIC.
25269 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25271 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
25272 Replace with a microblaze version.
25273 (microblaze_trampoline_init): Adapt for microblaze.
25274 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
25277 2013-02-16 Jakub Jelinek <jakub@redhat.com>
25278 Dodji Seketeli <dodji@redhat.com>
25281 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
25282 (instrument_mem_region_access): Do not forget to always put
25283 instrumentation of the of 'base' and 'base + len' in a "if (len !=
25284 0) statement, even for cases where either 'base' or 'base + len'
25285 are not instrumented -- because they have been previously
25286 instrumented. Simplify the logic by putting all the statements
25287 instrument 'base + len' inside a sequence, and then insert that
25288 sequence right before the current insertion point. Then, to
25289 instrument 'base + len', just get an iterator on that statement.
25290 And do not forget to update the pointer to iterator the function
25291 received as argument.
25293 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
25295 PR rtl-optimization/56348
25296 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
25298 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
25300 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
25301 (clean_graph_dump_file): Pass base to start_graph_dump.
25303 2013-02-14 Richard Henderson <rth@redhat.com>
25306 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
25308 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
25310 * collect2-aix.h: Define F_LOADONLY.
25312 2013-02-14 Richard Biener <rguenther@suse.de>
25315 * varasm.c (output_constant_def_1): Get the decl representing
25316 the constant as argument.
25317 (output_constant_def): Wrap output_constant_def_1.
25318 (make_decl_rtl): Use output_constant_def_1 with the decl
25319 representing the constant.
25320 (build_constant_desc): Optionally re-use a decl already
25321 representing the constant.
25322 (tree_output_constant_def): Adjust.
25324 2013-02-14 Dodji Seketeli <dodji@redhat.com>
25327 * asan.c (instrument_builtin_call): Really put the length of the
25328 second source argument into src1_len.
25330 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25332 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
25333 argument. If it is false, don't create edge from then_bb to
25335 (insert_if_then_before_iter): Pass true to it.
25336 (build_check_stmt): Pass false to it.
25337 (transform_statements): Flush hash table only on extended basic
25338 block boundaries, rather than at the beginning of every bb.
25339 Don't flush hash table on nonfreeing_call_p calls.
25340 * tree-flow.h (nonfreeing_call_p): New prototype.
25341 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
25343 2013-02-13 David S. Miller <davem@davemloft.net>
25345 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
25347 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
25350 * ira.c (max_regno_before_ira): Move from ...
25352 (fix_reg_equiv_init): Use max_regno_before_ira instead of
25355 2013-02-13 Jakub Jelinek <jakub@redhat.com>
25357 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
25359 2013-02-13 Richard Biener <rguenther@suse.de>
25362 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
25363 globals in MEM_REFs.
25365 2013-02-13 Richard Biener <rguenther@suse.de>
25367 * loop-init.c (loop_optimizer_init): Clear loop state when
25368 re-initializing preserved loops.
25369 * loop-unswitch.c (unswitch_single_loop): Return whether
25370 we unswitched the loop. Do not verify loop state here.
25371 (unswitch_loops): When we unswitched a loop discover new loops.
25373 2013-02-13 Kostya Serebryany <kcc@google.com>
25375 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
25377 * sanitizer.def: Rename __asan_init to __asan_init_v1.
25379 2013-02-12 Dodji Seketeli <dodji@redhat.com>
25381 Avoid instrumenting duplicated memory access in the same basic block
25382 * Makefile.in (asan.o): Add new dependency on hash-table.h
25383 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
25384 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
25385 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
25386 (free_mem_ref_resources, has_mem_ref_been_instrumented)
25387 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
25388 (get_mem_ref_of_assignment): New functions.
25389 (get_mem_refs_of_builtin_call): Extract from
25390 instrument_builtin_call and tweak a little bit to make it fit with
25392 (instrument_builtin_call): Use the new
25393 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
25394 of is_gimple_builtin_call.
25395 (instrument_derefs, instrument_mem_region_access): Insert the
25396 instrumented memory reference into the hash table.
25397 (maybe_instrument_assignment): Renamed instrument_assignment into
25398 this, and change it to advance the iterator when instrumentation
25399 actually happened and return true in that case. This makes it
25400 homogeneous with maybe_instrument_assignment, and thus give a
25401 chance to callers to be more 'regular'.
25402 (transform_statements): Clear the memory reference hash table
25403 whenever we enter a new BB, when we cross a function call, or when
25404 we are done transforming statements. Use
25405 maybe_instrument_assignment instead of instrumentation. No more
25406 need to special case maybe_instrument_assignment and advance the
25407 iterator after calling it; it's now handled just like
25408 maybe_instrument_call. Update comment.
25410 2013-02-13 Richard Biener <rguenther@suse.de>
25412 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
25413 Fix loop discovery code.
25415 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
25417 PR inline-asm/56148
25418 * lra-constraints.c (process_alt_operands): Match early clobber
25419 operand with itself. Check conflicts with earlyclobber only if
25420 the operand is not reloaded. Prefer to reload conflicting operand
25421 if earlyclobber and matching operands are the same.
25423 2013-02-12 Richard Biener <rguenther@suse.de>
25426 * lto-streamer-out.c (write_symbol): Do not output symbols
25427 for hard register variables.
25429 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
25432 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
25433 (umulsidi3_insn, mulsidi3_insn): New insns.
25435 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
25437 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
25438 (struct tune_params): Add vec_costs field.
25439 * config/arm/arm.c (arm_builtin_vectorization_cost)
25440 (arm_add_stmt_cost): New functions.
25441 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
25442 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
25443 (arm_default_vec_cost): New struct of type cpu_vec_costs.
25444 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
25445 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
25446 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
25447 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
25449 2013-02-12 Richard Biener <rguenther@suse.de>
25452 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
25453 decls again if possible.
25455 2013-02-12 Richard Biener <rguenther@suse.de>
25457 PR middle-end/56288
25458 * tree-ssa.c (verify_ssa_name): Fix check, move
25459 SSA_NAME_IN_FREE_LIST check up.
25461 2013-02-12 Jakub Jelinek <jakub@redhat.com>
25462 Steven Bosscher <steven@gcc.gnu.org>
25464 PR rtl-optimization/56151
25465 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
25466 equal to op0 or op1, and last_insn pattern is CODE operation
25467 with MEM dest and one of the operands matches that MEM.
25469 2013-02-11 Sriraman Tallam <tmsriram@google.com>
25471 * doc/extend.texi: Document Function Multiversioning and "default"
25472 parameter string to target attribute.
25473 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
25474 target attribute parameter is "default".
25475 (ix86_compare_version_priority): Remove checks for target attribute.
25476 (ix86_mangle_function_version_assembler_name): Change error to sorry.
25477 Remove check for target attribute equal to NULL. Add assert.
25478 (ix86_generate_version_dispatcher_body): Change error to sorry.
25480 2013-02-11 Iain Sandoe <iain@codesourcery.com>
25481 Jack Howarth <howarth@bromo.med.uc.edu>
25482 Patrick Marlier <patrick.marlier@gmail.com>
25485 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
25486 define ENDFILE_SPEC as TM_DESTRUCTOR.
25487 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
25489 2013-02-11 Alexander Potapenko <glider@google.com>
25490 Jack Howarth <howarth@bromo.med.uc.edu>
25491 Jakub Jelinek <jakub@redhat.com>
25494 * config/darwin.c (cdtor_record): Rename ctor_record.
25495 (sort_cdtor_records): Rename sort_ctor_records.
25496 (finalize_dtors): New routine to sort destructors by
25497 priority before use in assemble_integer.
25498 (machopic_asm_out_destructor): Use finalize_dtors if needed.
25500 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
25502 PR rtl-optimization/56275
25503 * simplify-rtx.c (avoid_constant_pool_reference): Check that
25504 offset is non-negative and less than cmode size before
25505 calling simplify_subreg.
25507 2013-02-11 Richard Biener <rguenther@suse.de>
25509 PR tree-optimization/56264
25510 * cfgloop.h (fix_loop_structure): Adjust prototype.
25511 * loop-init.c (fix_loop_structure): Return the number of
25512 newly discovered loops.
25513 * tree-cfgcleanup.c (repair_loop_structures): When new loops
25514 are discovered, do a full loop-closed SSA rewrite.
25516 2013-02-11 Richard Biener <rguenther@suse.de>
25518 PR tree-optimization/56273
25519 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
25521 (check_array_ref): Fix missing newline in dumps.
25522 (search_for_addr_array): Likewise.
25524 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
25526 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
25528 2013-02-09 Jakub Jelinek <jakub@redhat.com>
25531 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
25533 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
25535 PR rtl-optimization/56246
25536 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25538 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
25539 constraints are satisfied.
25541 2013-02-08 Jeff Law <law@redhat.com>
25544 * emit-rtl.c (reg_is_parm_p): New function.
25545 * regs.h (reg_is_parm_p): New prototype.
25546 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
25547 callee-clobbered registers.
25549 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
25552 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
25553 If there is no implicit builtin declaration, just return NULL.
25555 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25557 * config/i386/sse.md (FMAMODEM): New mode iterator.
25558 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
25559 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
25561 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25563 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
25564 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
25565 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
25567 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25569 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
25570 (microblaze*-*-elf): Likewise.
25571 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
25573 * config/microblaze/microblaze-c.c: Add builtin defines for
25574 _LITTLE_ENDIAN and _BIG_ENDIAN.
25575 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
25576 add to TARGET_DEFAULT flags.
25577 Expand ASM_SPEC and LINK_SPEC.
25578 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
25579 * config/microblaze/microblaze.md: Update extendsidi2 and
25580 movdi_internal instructions to use low-order / high-order reg
25582 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
25583 options and inversemask / mask of LITTLE_ENDIAN.
25584 * config/microblaze/t-microblaze: Expand multilib options to
25585 include mlittle-endian (le) and update exceptions patterns.
25587 2013-02-08 Jakub Jelinek <jakub@redhat.com>
25589 PR rtl-optimization/56195
25590 * lra-constraints.c (get_reload_reg): Don't reuse regs
25591 if they have smaller mode than requested, if they have
25592 wider mode than requested, try to return a SUBREG.
25594 PR tree-optimization/56250
25595 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
25596 if type is unsigned and code isn't MULT_EXPR.
25598 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25600 PR tree-optimization/56064
25601 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
25602 bits according to mode.
25603 * fixed-value.h (fixed_from_double_int)
25604 (const_fixed_from_double_int): Adjust comments.
25606 2013-02-08 Richard Biener <rguenther@suse.de>
25609 * lto-streamer.h (struct data_in): Remove current_file, current_line
25610 and current_col members.
25611 * lto-streamer-out.c (lto_output_location): Stream changed bits
25612 en-block for efficiency.
25613 * lto-streamer-in.c (clear_line_info): Remove.
25614 (lto_input_location): Cache current file, line and column
25615 globally via local statics. Read changed bits en-block.
25616 (input_function): Do not call clear_line_info.
25617 (lto_read_body): Likewise.
25618 (lto_input_toplevel_asms): Likewise.
25620 2013-02-08 Michael Matz <matz@suse.de>
25622 PR tree-optimization/52448
25623 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
25624 (nt_call_phase): New static.
25625 (add_or_mark_expr): Only mark accesses with newer phase than any
25627 (nonfreeing_call_p): New.
25628 (nt_init_block): Update nt_call_phase, mark blocks as visited.
25629 (nt_fini_block): Keep blocks marked as visited.
25630 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
25632 2013-02-08 Richard Biener <rguenther@suse.de>
25634 * ira.c (ira): Free broken dominator information.
25636 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
25638 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
25640 2013-02-08 Marek Polacek <polacek@redhat.com>
25642 * cfgloop.c (verify_loop_structure): Add more checking of headers.
25644 2013-02-08 Richard Biener <rguenther@suse.de>
25646 PR middle-end/56181
25647 * cfgloop.h (flow_loops_find): Adjust.
25648 (bb_loop_header_p): Declare.
25649 * cfgloop.c (bb_loop_header_p): New function split out from ...
25650 (flow_loops_find): ... here. Adjust function signature,
25651 support incremental loop structure update.
25652 (verify_loop_structure): Cleanup. Verify a loop is a loop.
25653 * cfgloopmanip.c (fix_loop_structure): Move ...
25654 * loop-init.c (fix_loop_structure): ... here.
25655 (apply_loop_flags): Split out from ...
25656 (loop_optimizer_init): ... here.
25657 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
25658 in incremental mode, only remove dead loops here.
25660 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
25663 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
25664 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
25665 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
25666 (*round<mode>3.libgcc): New insns for fixed-modes.
25667 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
25668 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
25669 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
25670 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
25671 implementations. Define to __builtin_avr_absFX,
25672 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
25673 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
25674 __builtin_avr_countlsFX, respectively.
25675 * config/avr/avr-c.c (target.h): Include it.
25676 (enum avr_builtin_id): New enum.
25677 (avr_resolve_overloaded_builtin): New static function.
25678 (avr_register_target_pragmas): Use it to set
25679 targetm.resolve_overloaded_builtin.
25680 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
25681 tree nodes used by DEF_BUILTIN.
25682 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
25683 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
25684 <AVR_BUILTIN_xxBITS>: Same.
25686 2013-02-08 Richard Biener <rguenther@suse.de>
25688 * cfgloop.c (verify_loop_structure): Properly handle
25689 a loop exiting to another loop header.
25690 * ira-int.h (ira_loops): Remove.
25691 * ira.c (ira_loops): Remove.
25692 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
25693 (do_reload): Use loop_optimizer_finalize.
25694 * ira-build.c (create_loop_tree_nodes): Use get_loops and
25695 number_of_loops to access the loop tree.
25696 (more_one_region_p): Likewise.
25697 (finish_loop_tree_nodes): Likewise.
25698 (rebuild_regno_allocno_maps): Likewise.
25699 (mark_loops_for_removal): Likewise.
25700 (mark_all_loops_for_removal): Likewise.
25701 (remove_unnecessary_regions): Likewise.
25702 (ira_build): Likewise.
25703 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
25705 2013-02-08 Richard Biener <rguenther@suse.de>
25707 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
25708 * ipa-pure-const.c (analyze_function): Avoid calling
25709 mark_irreducible_loops twice.
25710 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
25712 2013-02-07 David S. Miller <davem@davemloft.net>
25714 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
25716 * var-tracking.c (vt_add_function_parameter): Test the presence of
25717 HAVE_window_save properly and do not remap argument registers when
25718 we have a leaf function.
25720 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
25723 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
25725 * config/i386/i386.c (ix86_print_operand): Ditto.
25727 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25729 * lra-constraints.c (process_alt_operands): Fix recently added comment.
25731 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
25733 PR rtl-optimization/56225
25734 * lra-constraints.c (process_alt_operands): Check that reload hard
25735 reg can hold value for strict_low_part.
25737 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25740 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
25741 dwarf2out_end_function.
25742 (in_first_function_p, maybe_at_text_label_p,
25743 first_loclabel_num_not_at_text_label): New variables.
25744 (dwarf2out_var_location): In the first function find out
25745 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
25746 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
25749 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
25751 PR rtl-optimization/56178
25752 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
25753 SUBREG of a register. Tidy up related block of code.
25754 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
25755 note if the source is a register or a SUBREG of a register.
25757 2013-02-07 Jakub Jelinek <jakub@redhat.com>
25760 * config/rs6000/rs6000.md (ptrm): New mode attr.
25761 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
25762 call_value_indirect_aix<pttrsize>,
25763 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
25766 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25768 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
25769 if -bnortl. Convert to strcmp and strncmp.
25771 2013-02-07 Alan Modra <amodra@gmail.com>
25774 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
25775 addresses won't wrap when offsetting.
25776 (rs6000_secondary_reload): Provide secondary reloads needed for
25777 wrapping LO_SUM addresses.
25779 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
25781 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
25782 MACH, just __MACH__.
25784 2013-02-06 Richard Biener <rguenther@suse.de>
25786 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
25787 instead of calling fix_loop_structure.
25789 2013-02-06 Jakub Jelinek <jakub@redhat.com>
25791 PR middle-end/56217
25792 * omp-low.c (use_pointer_for_field): Return false if
25793 lower_send_shared_vars doesn't generate any copy-out code.
25795 2013-02-06 Tom de Vries <tom@codesourcery.com>
25797 PR rtl-optimization/56131
25798 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
25799 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
25800 of the label is NULL. Add comment.
25802 2013-02-05 Jakub Jelinek <jakub@redhat.com>
25804 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
25807 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
25808 (STATIC_LIBTSAN_LIBS): Likewise.
25809 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
25810 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
25811 is defined, don't add anything else beyond that.
25812 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
25813 (LINK_COMMAND_SPEC): Use them.
25815 PR tree-optimization/56205
25816 * tree-stdarg.c (check_all_va_list_escapes): Return true if
25817 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
25818 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
25820 2013-02-05 Richard Biener <rguenther@suse.de>
25822 PR tree-optimization/53342
25823 PR tree-optimization/53185
25824 * tree-vectorizer.h (vect_check_strided_load): Remove.
25825 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
25826 not disallow peeling for vectorized strided loads.
25827 (vect_check_strided_load): Make static and simplify.
25828 (vect_analyze_data_refs): Adjust.
25829 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
25830 correctly when vectorizing strided loads.
25832 2013-02-05 Richard Biener <rguenther@suse.de>
25834 * doc/install.texi: Refer to ISL, not PPL.
25836 2013-02-05 Jan Hubicka <jh@suse.cz>
25838 PR tree-optimization/55789
25839 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
25841 2013-02-05 Jan Hubicka <jh@suse.cz>
25843 PR tree-optimization/55789
25844 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
25845 the dead call anyway.
25847 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
25850 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
25852 2013-02-04 Alexander Potapenko <glider@google.com>
25853 Jack Howarth <howarth@bromo.med.uc.edu>
25854 Jakub Jelinek <jakub@redhat.com>
25857 * config/darwin.c (sort_ctor_records): Stabilized qsort
25858 on constructor priority by using original position.
25859 (finalize_ctors): New routine to sort constructors by
25860 priority before use in assemble_integer.
25861 (machopic_asm_out_constructor): Use finalize_ctors if needed.
25863 2013-02-04 Jakub Jelinek <jakub@redhat.com>
25866 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
25867 about visibility on artificial decls.
25868 * config/sol2.c (solaris_assemble_visibility): Likewise.
25870 2013-02-04 Kai Tietz <ktietz@redhat.com>
25873 * config/i386/i386.c (function_value_ms_64): Add additional valtype
25874 argument and improve checking of return-argument types for 16-byte
25876 (ix86_function_value_1): Add additional valtype argument on call
25877 of function_value_64.
25878 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
25879 handling infunction_value_64 function.
25881 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
25883 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
25885 2013-02-04 Richard Biener <rguenther@suse.de>
25887 PR tree-optimization/56188
25888 * tree-ssa-structalias.c (label_visit): Consider case with
25889 initially non-empty points-to set.
25890 (perform_var_substitution): Dump node mapping and clean up.
25892 2013-02-04 Richard Guenther <rguenther@suse.de>
25895 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
25896 node prevail as last resort.
25897 (lto_symtab_merge_decls): Remove guard on LTRANS here.
25898 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
25900 2013-02-04 Richard Biener <rguenther@suse.de>
25902 PR tree-optimization/56113
25903 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
25905 (equiv_class_lookup_or_add): ... this.
25906 (label_visit): Adjust and fix error in previous patch.
25907 (perform_var_substitution): Adjust.
25909 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
25911 * config/sh/divtab.c: Fix formatting and comments throughout the file.
25912 * config/sh/sh4-300.md: Likewise.
25913 * config/sh/sh4a.md: Likewise.
25914 * config/sh/constraints.md: Likewise.
25915 * config/sh/sh.md: Likewise.
25916 * config/sh/netbsd-elf.h: Likewise.
25917 * config/sh/predicates.md: Likewise.
25918 * config/sh/sh-protos.h: Likewise.
25919 * config/sh/ushmedia.h: Likewise.
25920 * config/sh/linux.h: Likewise.
25921 * config/sh/sh.c: Likewise.
25922 * config/sh/superh.h: Likewise.
25923 * config/sh/elf.h: Likewise.
25924 * config/sh/sh4.md: Likewise.
25925 * config/sh/sh.h: Likewise.
25927 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25929 * config/pa/constraints.md: Adjust unused letters. Change "T"
25930 constraint to match_test floating_point_store_memory_operand().
25931 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
25932 (base14_operand): New.
25933 (floating_point_store_memory_operand): New.
25934 (integer_store_memory_operand): Revise to use base14_operand and
25935 reg_plus_base_memory_operand.
25936 (move_dest_operand): Allow symbolic_memory_operands.
25937 (symbolic_memory_operand): Check for LO_SOM.
25938 (symbolic_operand): Change default case to break.
25939 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
25940 CONST_DOUBLE values to be reloaded by putting them into memory when
25941 the destination is a floating point register.
25942 (movdf): Remove code to handle CONST_DOUBLE.
25944 (reload_indf_r1): New.
25945 (reload_insf_r1): New.
25946 Consistently use "Q" and "T" constraints with integer and floating
25947 point move instructions, respectively.
25948 (movdi): Remove FAIL.
25949 Change predicate for source operand unamed DImode move from
25950 general_operand to move_src_operand.
25951 (umulsidi3): Change predicate for destination operand to
25953 Likewise for similar unamed patterns.
25954 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
25955 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
25956 (hppa_legitimize_address): Simplify mask calculation.
25957 (pa_emit_move_sequence): Revised handling of secondary reloads from
25958 REG+D addresses for floating point loads and stores. Directly handle
25959 loading CONST0_RTX (mode) to a floating point register.
25960 (pa_secondary_reload): Handle reloading DF and SFmode constant values
25961 to floating point registers. Don't restrict secondary reloads to
25962 floating point registers to integer modes. Revise some comments and
25964 (TARGET_LEGITIMATE_ADDRESS_P): Define.
25965 (pa_legitimate_address_p): New.
25966 (pa_legitimize_reload_address): New.
25967 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
25968 (STRICT_REG_OK_FOR_BASE_P): New.
25969 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
25970 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
25972 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
25973 Andrew Dixie <andrewd@gentrack.com>
25975 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
25978 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
25980 * expmed.c (extract_bit_field_1): Pass the full width of the
25981 structure to get_best_reg_extraction_insn.
25983 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
25986 * configure.ac (use_cxa_atexit): Add AIX.
25987 * configure: Regenerate.
25989 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
25991 2013-02-01 Jakub Jelinek <jakub@redhat.com>
25994 * final.c (need_profile_function): New variable.
25995 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
25996 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
25997 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
25998 notes, targetm.asm_out.function_prologue doesn't emit anything,
25999 HAVE_prologue and profiler should be emitted before prologue,
26000 set need_profile_function instead of emitting it.
26001 (final_scan_insn): If need_profile_function, emit
26002 profile_function on the first NOTE_INSN_BASIC_BLOCK or
26003 NOTE_INSN_FUNCTION_BEG note.
26005 2013-02-01 Richard Henderson <rth@redhat.com>
26007 * config/rs6000/rs6000.md (smulditi3): New.
26010 * config/alpha/alpha.md (umulditi3): New.
26012 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
26014 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
26015 (ASM_OUTPUT_ALIGNED_LOCAL): New.
26017 2013-02-01 Richard Biener <rguenther@suse.de>
26019 PR tree-optimization/56113
26020 * tree-ssa-structalias.c (label_visit): Reduce work for
26021 single-predecessor nodes.
26023 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
26025 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
26026 range isn't testing for zero.
26028 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
26030 PR middle-end/56113
26031 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
26033 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
26034 Nick Clifton <nickc@redhat.com>
26036 * config/v850/constraints.md (Q): Define as a memory constraint.
26037 * config/v850/predicates.md (label_ref_operand): New predicate.
26038 (e3v5_shift_operand): New predicate.
26039 (ior_operator): New predicate.
26040 * config/v850/t-v850: Add e3v5 multilib.
26041 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
26042 (v850_gen_movdi): Prototype.
26043 * config/v850/v850.c: Add support for e3v5 architecture.
26044 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
26046 (construct_save_jarl): Add e3v5 long JARL support.
26047 (v850_adjust_insn_length): New function. Adjust length of call
26048 insns when using e3v5 instructions.
26049 (v850_gen_movdi): New function: Generate instructions to move a
26051 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
26052 (CPP_SPEC): Define __v850e3v5__ as appropriate.
26053 (TARGET_USE_FPU): Enable for e3v5.
26054 (CONST_OK_FOR_W): New macro.
26055 (ADJUST_INSN_LENGTH): Define.
26056 * config/v850/v850.md (UNSPEC_LOOP): Define.
26057 (attr cpu): Add v850e3v5.
26058 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
26059 (movdi): New pattern.
26060 (movdi_internal): New pattern.
26061 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
26062 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
26063 (cstoresf4): Likewise.
26064 (cstoredf4): Likewise.
26065 (insv): New pattern.
26066 (rotlso3_a): New pattern.
26067 (rotlsi3_b): New pattern
26068 (rotlsi3_v850e3v5): New pattern.
26069 (doloop_begin): New pattern.
26070 (fix_loop_counter): New pattern.
26071 (doloop_end): New pattern.
26072 (branch_normal): Add e3v5 long branch support.
26073 (branch_invert): Likewise.
26074 (branch_z_normal): Likewise.
26075 (branch_z_invert): Likewise.
26076 (branch_nz_normal): Likewise.
26077 (branch_nz_invert): Likewise.
26078 (call_internal_short): Add e3v5 register-indirect JARL support.
26079 (call_internal_long): Likewise.
26080 (call_value_internal_short): Likewise.
26081 (call_value_internal_long): Likewise.
26082 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
26083 (mloop): New option.
26084 * config.gcc: Add support for configuring v840e3v5 target.
26085 * doc/invoke.texi: Document new v850 specific command line options.
26087 2013-01-31 Paul Koning <ni1d@arrl.net>
26091 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
26092 children if parent is a class.
26093 (prune_unused_types_prune): Don't add DW_AT_declaration.
26095 2013-01-31 Richard Biener <rguenther@suse.de>
26097 PR tree-optimization/56157
26098 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
26099 match up operand with SLP child.
26101 2013-01-31 Jason Merrill <jason@redhat.com>
26104 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
26105 parameters the first time.
26106 (gen_scheduled_generic_parms_dies): Check completeness here.
26108 2013-01-31 Richard Biener <rguenther@suse.de>
26110 PR middle-end/53073
26111 * common.opt (faggressive-loop-optimizations): New flag,
26112 enabled by default.
26113 * doc/invoke.texi (faggressive-loop-optimizations): Document.
26114 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
26115 infer_loop_bounds_from_undefined by it.
26117 2013-01-31 Richard Biener <rguenther@suse.de>
26119 PR tree-optimization/56150
26120 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
26121 visit virtual operands.
26122 (find_uses_to_rename_bb): Likewise.
26124 2013-01-31 Richard Biener <rguenther@suse.de>
26126 PR tree-optimization/56150
26127 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
26128 mixed store non-store stmts.
26130 2013-01-30 Jakub Jelinek <jakub@redhat.com>
26133 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
26134 LIBASAN_EARLY_SPEC is defined.
26135 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
26136 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
26138 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
26141 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
26142 invalid args instead of ICEing on it.
26143 (ix86_valid_target_attribute_tree): Return error_mark_node if
26144 ix86_valid_target_attribute_inner_p failed.
26145 (ix86_valid_target_attribute_p): Return false only if
26146 ix86_valid_target_attribute_tree returned error_mark_node. Allow
26147 target("default") attribute.
26148 (sorted_attr_string): Change argument from const char * to tree,
26149 merge in all target attribute arguments rather than just one.
26150 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
26151 instead of free. Avoid using strcat.
26152 (ix86_mangle_function_version_assembler_name): Mangle
26153 target("default") as if no target attribute is present. Adjust
26154 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
26155 instead of xmalloc and XDELETEVEC instead of free.
26156 (ix86_function_versions): Don't return true if one of the decls
26157 doesn't have target attribute. If they don't and one of the decls
26158 is DECL_FUNCTION_VERSIONED, report an error. Adjust
26159 sorted_attr_string caller. Use XDELETEVEC instead of free.
26160 (ix86_supports_function_versions): Remove.
26161 (make_name): Fix up formatting.
26162 (make_dispatcher_decl): Remove resolver_name and its initialization.
26163 Avoid leaking memory.
26164 (is_function_default_version): Return true if there is
26165 target("default") attribute rather than no target attribute at all.
26166 (make_resolver_func): Avoid leaking memory.
26167 (ix86_generate_version_dispatcher_body): Likewise.
26168 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
26169 * target.def (supports_function_versions): Remove.
26170 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
26171 * doc/tm.texi: Regenerated.
26173 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
26175 PR rtl-optimization/56144
26176 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
26177 for values with side effects.
26179 2013-01-30 Richard Biener <rguenther@suse.de>
26181 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
26182 (sparseset_pop): Likewise.
26183 * cfganal.c (compute_idf): Likewise. Increase work-stack size
26184 to be able to use quick_push in the worker loop.
26186 2013-01-30 Marek Polacek <polacek@redhat.com>
26188 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
26190 2013-01-30 Richard Biener <rguenther@suse.de>
26193 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
26195 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
26197 PR tree-optimization/56064
26198 * fixed-value.c (fixed_from_double_int): New function.
26199 * fixed-value.h (fixed_from_double_int): New prototype.
26200 (const_fixed_from_double_int): New static inline function.
26201 * fold-const.c (native_interpret_fixed): New static function.
26202 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
26203 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
26204 (native_encode_fixed): New static function.
26205 (native_encode_expr) <FIXED_CST>: Use it.
26206 (native_interpret_int): Move double_int worker code to...
26207 * double-int.c (double_int::from_buffer): ...this new static method.
26208 * double-int.h (double_int::from_buffer): Prototype it.
26210 2013-01-30 Richard Biener <rguenther@suse.de>
26212 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
26213 New pointer-map and obstack.
26214 (init_alias_vars): Allocate pointer-map and obstack.
26215 (delete_points_to_sets): Free them.
26216 (find_what_var_points_to): Cache result.
26217 (find_what_p_points_to): Adjust for changed interface of
26218 find_what_var_points_to.
26219 (compute_points_to_sets): Likewise.
26220 (ipa_pta_execute): Likewise.
26222 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26224 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
26225 * configure: Regenerate.
26226 * config.in: Regenerate.
26227 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
26228 #nobits/#progbits if supported.
26230 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
26233 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
26234 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
26235 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
26237 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26239 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
26240 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
26242 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26244 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
26245 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
26247 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26249 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
26251 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
26252 * config/arm/cortex-a7.md: New bypasses using
26253 arm_mac_accumulator_is_result.
26255 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26257 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
26258 (cortex_a7_neon_mla): Likewise.
26259 (cortex_a7_fpfmad): New reservation.
26260 (cortex_a7_fpmacs): Use ffmas and update required units.
26261 (cortex_a7_fpmuld): Update required units and latency.
26262 (cortex_a7_fpmacd): Likewise.
26263 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
26264 (cortex_a7_neon). Likewise.
26265 (bypass) Update participating units.
26267 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
26269 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
26270 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
26272 * config/arm/vfp11.md (vfp_farith): Use ffmas.
26273 (vfp_fmul): Use ffmad.
26274 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
26275 (cortex_r4_fmacd): Use ffmad.
26276 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
26277 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
26278 (cortex_a9_fmacd): Use ffmad.
26279 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
26280 (cortex_a8_vfp_macd): Use ffmad.
26281 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
26282 (cortex_a5_fpmacd): Use ffmad.
26283 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
26284 (cortex_a15_vfp_macd): Use ffmad.
26285 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
26287 2013-01-29 Jason Merrill <jason@redhat.com>
26290 * varasm.c (default_assemble_visibility): Don't warn about
26291 visibility on artificial decls.
26293 2013-01-29 Richard Biener <rguenther@suse.de>
26295 PR tree-optimization/56113
26296 * tree-ssa-structalias.c (equiv_class_lookup): Also return
26298 (label_visit): Free duplicate bitmaps and record the leader instead.
26299 (perform_var_substitution): Adjust.
26301 2013-01-29 Richard Biener <rguenther@suse.de>
26303 PR tree-optimization/55270
26304 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
26305 the CFG, schedule loops for fixup.
26307 2013-01-29 Nick Clifton <nickc@redhat.com>
26309 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
26312 2013-01-28 Leif Ekblad <leif@rdos.net>
26314 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
26315 * config/i386/i386.h (TARGET_RDOS): New macro.
26316 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
26317 * config/i386/i386.c (ix86_option_override_internal): For 64bit
26318 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
26319 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
26320 DEFAULT_LARGE_SECTION_THRESHOLD.
26321 * config/i386/i386.md (R14_REG, R15_REG): New constants.
26322 * config/i386/rdos.h: New file.
26323 * config/i386/rdos64.h: New file.
26325 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
26328 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
26331 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26333 PR rtl-optimization/56117
26334 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
26335 call cselib_lookup_from_insn on the MEM before calling
26336 add_insn_mem_dependence.
26338 2013-01-28 Richard Biener <rguenther@suse.de>
26340 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
26341 to a stmt that didn't have one.
26342 (copy_phis_for_bb): Likewise for PHI arguments.
26343 (copy_debug_stmt): Likewise for debug stmts.
26345 2013-01-28 Richard Biener <rguenther@suse.de>
26347 PR tree-optimization/56034
26348 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
26349 (partition_builtin_p): Adjust.
26350 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
26351 it is the last partition.
26352 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
26353 up the vertex for the definition.
26354 (classify_partition): Classify whether a partition is a
26355 PKIND_REDUCTION, thus has uses outside of the loop.
26356 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
26357 Merge all PKIND_REDUCTION partitions into the last partition.
26358 (tree_loop_distribution): Seed partitions from reductions as well.
26360 2013-01-28 Jakub Jelinek <jakub@redhat.com>
26362 PR tree-optimization/56125
26363 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
26364 pow(x,c) into sqrt(x) * powi(x, n/2) or
26365 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
26366 optimizing for size.
26367 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
26368 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
26371 PR tree-optimization/56094
26372 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
26373 to UNKNOWN_LOCATION while gimplifying expr.
26375 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
26378 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
26379 operand 0 in movabs insn template for -masm=intel asm alternative.
26380 (*movabs<mode>_2): Ditto for operand 1.
26382 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
26385 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
26388 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26390 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
26393 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26395 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
26396 Undef to avoid warning.
26398 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
26400 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
26401 * configure: Regenerate.
26403 2013-01-25 Jakub Jelinek <jakub@redhat.com>
26405 PR tree-optimization/56098
26406 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
26407 for stmts with volatile ops.
26408 (cond_store_replacement): Don't optimize if assign has volatile ops.
26409 (cond_if_else_store_replacement_1): Don't optimize if either
26410 then_assign or else_assign have volatile ops.
26411 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
26414 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26416 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
26418 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26420 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
26421 missing ':' in asm example.
26423 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
26425 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
26426 entries into lane and laneq entries.
26427 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
26428 Remove AdvSIMD scalar modes.
26429 (aarch64_sq<r>dmulh_laneq<mode>): New.
26430 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
26432 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
26433 builtin implementations to relfect changes in RTL in aarch64-simd.md.
26434 * config/aarch64/iterators.md (VCOND): New.
26437 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
26440 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
26441 Add NULL LIBNAME argument to existing definitions.
26442 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
26443 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
26444 * config/avr/avr.c (DEF_BUILTIN): Same.
26445 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
26446 (avr_expand_builtin): Expand to a vanilla call if a libgcc
26447 implementation is available (DECL_ASSEMBLER_NAME is set).
26448 (avr_fold_absfx): New static function.
26449 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
26450 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
26451 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
26452 AVR_BUILTIN_ABSLLK.
26453 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
26454 (abshk, absk, abslk, absllk): Provide as static inline functions.
26456 2013-01-25 Marek Polacek <polacek@redhat.com>
26458 PR tree-optimization/56035
26459 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
26461 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
26463 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
26464 (*movtf_internal_rex64): Add (!o,C) alternative
26465 (*movxf_internal_rex64): Ditto.
26466 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
26468 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
26470 * doc/invoke.texi: fix typo.
26471 * doc/objc.texi: fix typo.
26473 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
26475 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
26476 for the first two alternatives.
26478 2013-01-24 Diego Novillo <dnovillo@google.com>
26480 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
26481 (ggc-zone.o): Remove.
26482 * configure.ac: Remove option --with-gc.
26483 * configure: Re-generate.
26484 * doc/install.texi: Remove documentation for --with-gc.
26485 * gengtype.c (write_enum_defn): Remove. Update all users.
26486 (write_Types_process_field): Remove generation of gt_e_* argument.
26487 (output_type_enum): Remove. Update all users.
26488 (write_enum_defn): Remove. Update all users.
26489 (enum alloc_zone): Remove. Update all users.
26490 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
26491 * ggc-common.c (ggc_splay_alloc): Remove first argument.
26492 Update all callers.
26493 (struct ptr_data): Remove field TYPE. Update all users.
26494 (gt_pch_note_object): Remove argument TYPE. Update all users.
26495 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
26497 * ggc-none.c (ggc_alloc_typed_stat): Remove.
26498 (struct alloc_zone): Remove.
26499 (ggc_internal_alloc_zone_stat): Remove.
26500 (ggc_internal_cleared_alloc_zone_stat): Remove.
26501 * ggc-page.c (ggc_alloc_typed_stat): Remove.
26502 (ggc_pch_count_object): Remove last argument. Update all users.
26503 (ggc_pch_alloc_object): Remove last argument. Update all users.
26504 (struct alloc_zone): Remove.
26505 * ggc-zone.c: Remove.
26506 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
26507 (struct alloc_zone): Remove.
26508 (ggc_alloc_typed_stat): Remove.
26509 (ggc_alloc_typed): Remove.
26510 (ggc_splay_alloc): Remove first argument.
26511 (rtl_zone): Remove. Update all users.
26512 (tree_zone): Remove. Update all users.
26513 (tree_id_zone): Remove. Update all users.
26514 (ggc_internal_zone_alloc_stat): Remove. Update all users.
26515 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
26516 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
26517 * tree-ssanames.c: Remove references to zone allocator in comments.
26519 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
26521 * config/avr/avr.c (avr_out_fract): Make register numbers that
26522 might be outside of source operand signed.
26524 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
26526 * config/i386/constraints.md (Yf): New constraint.
26527 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
26528 of f constraint to conditionaly disable x87 register preferences.
26529 (*movdf_internal): Ditto.
26530 (*movsf_internal): Ditto.
26532 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
26534 PR inline-asm/55934
26535 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
26536 that have operands with impossible constraints.
26537 Add a FIXME for a speed-up opportunity.
26538 * lra-constraints.c (process_alt_operands): Verify that a class
26539 selected from constraints on asms is valid for the operand mode.
26540 (curr_insn_transform): Remove incorrect comment.
26542 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
26544 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
26545 TOC operand is a valid symbol ref in the constant pool.
26547 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26549 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
26551 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
26554 * config/avr/stdfix.h: New file.
26555 * t-avr (stdfix-gcc.h): New rule to build it.
26556 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
26558 2013-01-23 Kostya Serebryany <kcc@google.com>
26560 * config/darwin.h: remove dependency on
26561 CoreFoundation (asan on Mac OS).
26563 2013-01-23 Jakub Jelinek <jakub@redhat.com>
26566 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
26567 instead of cmpdi_operand for first comparison operand.
26568 Don't assert that comparison operands aren't both constants.
26570 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26572 * doc/install.texi (Downloading the Source): Update references to
26573 downloading separate components.
26575 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
26577 * doc/extend.texi (__int128): Improve grammar.
26579 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
26582 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
26583 alternative to (o,r).
26584 (*movdi_internal_rex64): Remove (!o,n) alternative.
26585 (DImode immediate->memory splitter): Remove.
26586 (DImode immediate->memory peephole2): Remove.
26587 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
26588 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
26589 alternative to (!o,*r).
26590 (*movtf_internal_sse): New pattern.
26591 (*movxf_internal_rex64): New pattern.
26592 (*movxf_internal): Disable for TARGET_64BIT.
26593 (*movdf_internal_rex64): Remove (!o,F) alternative.
26595 2013-01-22 Jakub Jelinek <jakub@redhat.com>
26597 PR middle-end/56074
26598 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
26599 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
26600 * tree-vect-loop-manip.c (find_loop_location): Also ignore
26601 stmt locations where LOCATION_LOCUS of the stmt location is
26602 UNKNOWN_LOCATION or BUILTINS_LOCATION.
26605 * config/i386/i386.md (UNSPEC_STOS): New.
26606 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
26607 *strsetqi_1): Add UNSPEC_STOS.
26609 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
26612 * doc/invoke.texi: Remove left over -Wsynth example.
26614 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26616 PR tree-optimization/56051
26617 * fold-const.c (fold_binary_loc): Don't fold
26618 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
26619 a narrowing conversion, or widening conversion from signed
26622 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26624 PR rtl-optimization/56023
26625 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
26626 dependent on debug instruction.
26628 2013-01-21 Martin Jambor <mjambor@suse.cz>
26630 PR middle-end/56022
26631 * function.c (allocate_struct_function): Call
26632 invoke_set_current_function_hook earlier.
26634 2013-01-21 Jakub Jelinek <jakub@redhat.com>
26636 * reload1.c (init_reload): Only initialize reload_obstack
26637 during the first call.
26639 2013-01-21 Marek Polacek <polacek@redhat.com>
26641 * cfgloop.c (verify_loop_structure): Fix up grammar.
26643 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
26645 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
26646 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
26648 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26651 * config/arm/marvell-pj4.md: Update copyright year.
26652 Fix up use of alu to alu_reg and simple_alu_imm.
26654 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
26656 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
26658 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
26661 * lra-constraints.c (curr_insn_transform): Don't reuse original
26662 insn for secondary memory move when memory mode should be different.
26664 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26666 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
26667 atomic_storedi_1): New patterns.
26669 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
26671 btver2 pipeline descriptions.
26672 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
26674 * config/i386/i386.md (btver2_decode): New type attributes.
26675 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
26677 * config/i386/btver2.md: New file describing btver2 pipelines.
26679 2013-01-19 Andrew Pinski <apinski@cavium.com>
26681 PR tree-optimization/52631
26682 * tree-ssa-sccvn (visit_use): Before looking up the original
26683 statement, try looking up the simplified expression.
26685 2013-01-19 Anthony Green <green@moxielogic.com>
26687 * config/moxie/moxie.c (moxie_expand_prologue): Set
26688 current_function_static_stack_size.
26690 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26692 PR tree-optimization/56029
26693 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
26694 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
26696 2013-01-18 Sharad Singhai <singhai@google.com>
26698 PR tree-optimization/55995
26699 * dumpfile.c (dump_loc): Print location only if available.
26700 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
26702 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
26705 * lra-constraints.c (curr_insn_transform): Reuse original insn for
26706 secondary memory move.
26707 (inherit_reload_reg): Use rclass instead of cl for
26708 check_secondary_memory_needed_p.
26710 2013-01-18 Jakub Jelinek <jakub@redhat.com>
26712 PR middle-end/56015
26713 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
26714 the case where writing real complex part of target modifies op1.
26716 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
26718 * config/aarch64/aarch64-simd.md
26719 (aarch64_vcond_internal<mode>): Handle unordered cases.
26720 * config/aarch64/iterators.md (v_cmp_result): New.
26722 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
26723 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26725 * config/arm/marvell-pj4.md: New file.
26726 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
26727 * config/arm/arm.md (generic_sched): Add marvell_pj4.
26728 (generic_vfp): Likewise.
26729 * config/arm/arm-cores.def: Add marvell-pj4.
26730 * config/arm/arm-tune.md: Regenerate.
26731 * config/arm/arm-tables.opt: Regenerate.
26732 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
26733 * doc/invoke.texi: Document marvell-pj4.
26735 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
26737 * config/aarch64/arm_neon.h: Map scalar types to standard types.
26739 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26744 * var-tracking.c (negative_power_of_two_p): New.
26745 (global_get_addr_cache, local_get_addr_cache): New.
26746 (get_addr_from_global_cache, get_addr_from_local_cache): New.
26747 (vt_canonicalize_addr): Rewrite using the above. Adjust the
26749 (vt_stack_offset_p): Remove.
26750 (vt_canon_true_dep): Always canonicalize loc's address.
26751 (clobber_overlapping_mems): Make sure we have a MEM.
26752 (local_get_addr_clear_given_value): New.
26753 (val_reset): Clear local cached entries.
26754 (compute_bb_dataflow): Create and release the local cache.
26755 Disable duplicate MEMs clobbering.
26756 (emit_notes_in_bb): Clobber MEMs likewise.
26757 (vt_emit_notes): Create and release the local cache.
26758 (vt_initialize, vt_finalize): Create and release the global
26759 cache, respectively.
26760 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
26762 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26764 PR libmudflap/53359
26765 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
26766 not found in the symtab.
26768 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
26771 PR rtl-optimization/55547
26772 PR rtl-optimization/53827
26775 * alias.c (offset_overlap_p): New, factored out of...
26776 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
26777 the conservative special case for symbolic constants. Don't
26778 adjust zero sizes on alignment.
26780 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
26782 PR rtl-optimization/52573
26783 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
26784 REG_UNUSED for the same register.
26786 2013-01-17 Richard Biener <rguenther@suse.de>
26787 Marek Polacek <polacek@redhat.com>
26789 PR rtl-optimization/55833
26790 * loop-unswitch.c (unswitch_loops): Move loop verification...
26791 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
26792 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
26793 Set it to true when we're removing a loop from hierarchy tree in
26794 an irreducible region.
26795 (fix_bb_placements): Adjust caller.
26796 (fix_loop_placements): Likewise.
26798 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
26800 * config/avr/builtins.def (DEF_BUILTIN): Factor out
26801 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
26802 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
26803 Remove ID. Adjust comments.
26804 * config/avr/avr-c.c (avr_builtin_name): Remove.
26805 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
26806 * config/avr/avr.c (avr_tolower): New static function.
26807 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
26808 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
26809 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
26812 2013-01-17 Jan Hubicka <jh@suse.cz>
26814 PR tree-optimization/55273
26815 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
26817 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
26820 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
26821 store through atomic_store<mode>_1.
26822 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
26824 2013-01-17 Martin Jambor <mjambor@suse.cz>
26826 PR tree-optimizations/55264
26827 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
26828 for virtual methods.
26829 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
26830 virtual methods before inlining is over.
26831 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
26833 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
26836 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
26838 PR rtl-optimization/56005
26839 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
26840 pending reads for prefetch.
26842 2013-01-16 Ian Bolton <ian.bolton@arm.com>
26844 * config/aarch64/aarch64.md
26845 (*cstoresi_neg_uxtw): New pattern.
26846 (*cmovsi_insn_uxtw): New pattern.
26847 (*<optab>si3_uxtw): New pattern.
26848 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
26849 (*<optab>si3_insn_uxtw): New pattern.
26850 (*bswapsi2_uxtw): New pattern.
26852 2013-01-16 Richard Biener <rguenther@suse.de>
26854 * tree-inline.c (tree_function_versioning): Remove set but
26855 never used variable.
26857 2013-01-16 Richard Biener <rguenther@suse.de>
26859 PR tree-optimization/55964
26860 * tree-flow.h (rename_variables_in_loop): Remove.
26861 (rename_variables_in_bb): Likewise.
26862 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
26863 (copy_loop_before): Adjust and delete update-ssa status.
26864 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
26865 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
26866 (rename_variables_in_loop): Remove.
26867 (slpeel_update_phis_for_duplicate_loop): Likewise.
26868 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
26869 use available cfg machinery instead of duplicating it.
26870 Update PHI nodes and perform poor-mans SSA update here.
26871 (slpeel_tree_peel_loop_to_edge): Adjust.
26873 2013-01-16 Richard Biener <rguenther@suse.de>
26875 PR tree-optimization/54767
26876 PR tree-optimization/53465
26877 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
26878 (vrp_visit_phi_node): For PHI arguments coming via backedges
26879 drop all symbolical range information.
26880 (execute_vrp): Compute backedges.
26882 2013-01-16 Richard Biener <rguenther@suse.de>
26884 * doc/install.texi: Update CLooG and ISL requirements to
26887 2013-01-16 Christian Bruel <christian.bruel@st.com>
26890 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
26891 (broken_move): Handle UNSPECV_SP_SWITCH_B.
26892 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
26894 2013-01-16 DJ Delorie <dj@redhat.com>
26896 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
26897 (UNSPECV_SP_SWITCH_E): New.
26898 (sp_switch_1): Change to an unspec.
26899 (sp_switch_2): Change to an unspec. Don't use post-inc when we
26902 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
26904 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
26905 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
26906 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
26907 (expand_mem_thread_fence): Ditto.
26908 (expand_mem_signal_fence): Ditto.
26909 (expand_atomic_load): Ditto.
26910 (expand_atomic_store): Ditto.
26912 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
26914 PR rtl-optimization/55547
26915 PR rtl-optimization/53827
26918 * alias.c (memrefs_conflict_p): Set sizes to negative after
26921 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26924 * function.c (thread_prologue_and_epilogue_insns): Always
26925 add crtl->drap_reg to set_up_by_prologue.set, even if
26926 stack_realign_drap is false.
26928 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
26930 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
26931 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
26932 *call): Fix indention.
26934 2013-01-15 Tom de Vries <tom@codesourcery.com>
26937 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
26940 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
26942 PR rtl-optimization/55153
26943 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
26945 2013-01-15 Martin Jambor <mjambor@suse.cz>
26947 PR tree-optimization/55920
26948 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
26949 accesses as grp_to_be_debug_replaced.
26951 2013-01-15 Jakub Jelinek <jakub@redhat.com>
26953 PR tree-optimization/55920
26954 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
26955 there is non-useless type conversion needed from debug rhs to lhs,
26956 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
26958 2013-01-15 Joseph Myers <joseph@codesourcery.com>
26959 Mikael Pettersson <mikpe@it.uu.se>
26962 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
26964 (ASM_OUTPUT_CASE_LABEL): Remove.
26965 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
26966 * final.c (shorten_branches): Update alignment of labels before
26967 jump tables if CASE_VECTOR_SHORTEN_MODE.
26969 2013-01-15 Richard Biener <rguenther@suse.de>
26972 * system.h: Do not include gmp.h for building host tools.
26974 2013-01-15 Richard Biener <rguenther@suse.de>
26976 PR middle-end/55882
26977 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
26978 account for bitpos when computing alignment.
26980 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
26982 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
26983 (ix86_target_macros_internal): Likewise.
26985 * config/i386/i386.c (m_CORE2I7): Removed.
26986 (m_CORE_HASWELL): New macro.
26987 (m_CORE_ALL): Likewise.
26988 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
26989 (initial_ix86_arch_features): Likewise.
26990 (processor_target_table): Initializations for Core avx2.
26991 (cpu_names): New names "core-avx2".
26992 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
26993 PROCESSOR_CORE_HASWELL.
26994 (ix86_issue_rate): New case.
26995 (ia32_multipass_dfa_lookahead): Likewise.
26996 (ix86_sched_init_global): Likewise.
26998 * config/i386/i386.h (TARGET_HASWELL): New macro.
26999 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
27000 (processor_type): New PROCESSOR_HASWELL.
27002 2013-01-15 Jakub Jelinek <jakub@redhat.com>
27004 PR tree-optimization/55955
27005 * tree-vect-loop.c (vectorizable_reduction): Give up early on
27006 *SHIFT_EXPR and *ROTATE_EXPR codes.
27008 PR tree-optimization/48766
27009 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
27010 -ftrapv disable -fwrapv.
27012 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27015 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
27016 etc. to 1 and not to __flash.
27017 Use LL suffix for __INT24_MAX__ with -mint8.
27018 Use ULL suffix for __UINT24_MAX__ with -mint8.
27020 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27022 * config/avr/avr-arch.h
27023 (struct base_arch_s): Use typedef avr_arch_t instead.
27024 (struct arch_info_s): Use typedef avr_arch_info_t instead.
27025 (struct mcu_type_s): Use typedef avr_mcu_t instead.
27026 * config/avr/avr.c: Same.
27027 * config/avr/avr-devices.c: Same.
27028 * config/avr/driver-avr.c: Same.
27029 * config/avr/gen-avr-mmcu-texi.c: Same.
27030 * config/avr/avr-mcus.def: Adjust comment.
27032 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
27034 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
27035 * config/aarch64/iterators.md (VALLDI): New.
27037 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
27038 Andi Kleen <ak@linux.intel.com>
27041 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
27042 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
27045 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27047 * config/avr/avr-stdint.h: Remove trailing blanks.
27048 * config/avr/avr-log.h: Same.
27049 * config/avr/avr-arch.h: Same.
27050 * config/avr/avr-devices.c: Same.
27051 * config/avr/avr-dimode.md: Same.
27052 * config/avr/predicates.md: Same.
27053 * config/avr/avr-c.c: Same. And fix typo.
27055 * config/avr/avr-protos.h: Same. And:
27056 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
27057 (init_cumulative_args): Rename to avr_init_cumulative_args.
27058 (expand_prologue): Rename to avr_expand_prologue.
27059 (expand_epilogue): Rename to avr_expand_epilogue.
27060 (adjust_insn_length): Rename to avr_adjust_insn_length.
27061 (notice_update_cc): Rename to avr_notice_update_cc.
27062 (final_prescan_insn): Rename to avr_final_prescan_insn.
27063 * config/avr/avr.c: Same.
27064 * config/avr/avr.h: Same.
27065 * config/avr/avr.md: Remove trailing blanks.
27066 (prologue): Use avr_expand_prologue.
27067 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
27069 2013-01-14 Richard Biener <rguenther@suse.de>
27071 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
27072 verify_location, collect_subblocks): New functions.
27073 (verify_gimple_in_cfg): Verify that locations only reference
27074 BLOCKs in the functions BLOCK tree.
27076 2013-01-14 Richard Biener <rguenther@suse.de>
27078 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
27080 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
27082 (insert_out_of_ssa_copy_on_edge): Likewise.
27083 (rewrite_close_phi_out_of_ssa): Likewise.
27084 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
27086 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
27087 propagated constants.
27088 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
27091 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
27093 * config/avr/avr-modes.def: Add GPL copyright notice.
27095 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
27097 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
27098 MEMMODEL_MASK to determine memory model.
27099 (atomic_store<mode>): Ditto from operands[2].
27100 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
27102 2013-01-13 Jakub Jelinek <jakub@redhat.com>
27105 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
27106 (fold_gimple_assign): Don't call unshare_expr here.
27107 (fold_ctor_reference): Call unshare_expr.
27109 2013-01-13 Terry Guo <terry.guo@arm.com>
27111 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
27112 * doc/fragments.texi: Document MULTILIB_REUSE.
27113 * gcc.c (multilib_reuse): New internal spec.
27114 (set_multilib_dir): Also search multilib from multilib_reuse.
27115 * genmultilib (tmpmultilib3): Refactor code.
27116 (tmpmultilib4): Ditto.
27117 (multilib_reuse): New multilib argument.
27119 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
27121 * Makefile.in: Update copyright.
27123 2013-01-12 Tom de Vries <tom@codesourcery.com>
27125 PR middle-end/55890
27126 * calls.c (expand_call): Check if arg_nr is valid.
27128 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
27130 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
27131 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
27132 documentation. Add missing '__' in front of
27133 __builtin_ia32_packssdw256.
27135 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27138 * config/s390/s390.c (s390_preferred_reload_class): Do not return
27139 NO_REGS for larl operands.
27140 (s390_reload_larl_operand): Use s390_load_address instead of
27143 2013-01-11 Richard Biener <rguenther@suse.de>
27145 * tree-cfg.c (verify_node_sharing_1): Split out from ...
27146 (verify_node_sharing): ... here.
27147 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
27149 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
27151 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
27152 Substitute TREECHECKING.
27153 * configure: Regenerate.
27154 * Makefile.in (TREECHECKING): New.
27156 2013-01-11 Richard Guenther <rguenther@suse.de>
27158 PR tree-optimization/44061
27159 * tree-vrp.c (extract_range_basic): Compute zero as
27160 value-range for __builtin_constant_p of function parameters.
27162 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
27164 Update copyright years.
27166 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
27168 PR rtl-optimization/55672
27169 * lra-eliminations.c (mark_not_eliminable): Permit addition with
27170 const to be eliminable.
27172 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
27174 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
27175 * configure: Regenerate.
27177 2013-01-10 Richard Biener <rguenther@suse.de>
27179 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
27181 2013-01-10 Richard Biener <rguenther@suse.de>
27184 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
27185 locations for virtual PHI arguments.
27186 (rewrite_update_phi_arguments): Likewise.
27188 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
27190 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
27193 2013-01-10 Jakub Jelinek <jakub@redhat.com>
27195 PR tree-optimization/55921
27196 * tree-complex.c (expand_complex_asm): New function.
27197 (expand_complex_operations_1): Call it for GIMPLE_ASM.
27199 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27202 * config/s390/s390.c (s390_symref_operand_p)
27203 (s390_loadrelative_operand_p): Merge the two functions.
27204 (s390_check_qrst_address, print_operand_address): Add parameters
27205 to s390_loadrelative_operand_p invokation.
27206 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
27207 (s390_reload_larl_operand, s390_secondary_reload): Use
27208 s390_loadrelative_operand_p instead of s390_symref_operand_p.
27209 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
27211 2013-01-09 Mike Stump <mikestump@comcast.net>
27213 * dse.c (record_store): Remove unnecessary assert.
27215 2013-01-09 Jan Hubicka <jh@suse.cz>
27217 PR tree-optimization/55569
27218 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
27219 * cfgloop.h (scale_loop_profile): Likewise.
27221 2013-01-09 Jan Hubicka <jh@suse.cz>
27224 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
27226 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
27228 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
27230 PR middle-end/55114
27231 * expr.h (maybe_emit_group_store): Declare.
27232 * expr.c (maybe_emit_group_store): New function.
27233 * builtins.c (expand_builtin_int_roundingfn): Call it.
27234 (expand_builtin_int_roundingfn_2): Likewise.
27236 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
27238 PR rtl-optimization/55829
27239 * lra-constraints.c (match_reload): Add code for absent output.
27240 (curr_insn_transform): Add code for reloads of matched inputs
27243 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
27245 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
27246 attribute of movddup insn to DF.
27247 (*vec_interleave_lowv2df): Ditto.
27248 (vec_dupv2df): Ditto.
27250 2013-01-09 Jan Hubicka <jh@suse.cz>
27252 PR tree-optimiation/55875
27253 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
27254 EVERY_ITERATION parameter.
27255 (number_of_iterations_exit): Check if exit is executed every iteration.
27256 (idx_infer_loop_bounds): Similarly here.
27257 (n_of_executions_at_most): Simplify
27258 to only test for cases where statement is dominated by the
27259 particular bound; handle correctly the "postdominance" test.
27260 (scev_probably_wraps_p): Use max loop iterations info
27261 as a global bound first.
27263 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
27264 Nick Clifton <nickc@redhat.com>
27266 * config/v850/v850.md (cbranchsf4): New pattern.
27267 (cstoresf4): New pattern.
27268 (cbranchdf4): New pattern.
27269 (cstoredf4): New pattern.
27270 (movsicc): Disallow floating point comparisons.
27271 (cmpsf_le_insn): Fix order of operators.
27272 (cmpsf_lt_insn): Likewise.
27273 (cmpsf_eq_insn): Likewise.
27274 (cmpdf_le_insn): Likewise.
27275 (cmpdf_lt_insn): Likewise.
27276 (cmpdf_eq_insn): Likewise.
27277 (cmpsf_ge_insn): Use LE comparison.
27278 (cmpdf_ge_insn): Likewise.
27279 (cmpsf_gt_insn): Use LT comparison.
27280 (cmpdf_gt_insn): Likewise.
27281 (cmpsf_ne_insn): Delete pattern.
27282 (cmpdf_ne_insn): Delete pattern.
27283 * config/v850/v850.c (v850_gen_float_compare): Use
27284 gen_cmpdf_eq_insn for NE comparison.
27285 (v850_float_z_comparison_operator)
27286 (v850_float_nz_comparison_operator): Move from here ...
27287 * config/v850/predicates.md: ... to here. Move GT and GE
27288 comparisons into v850_float_z_comparison_operator.
27289 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
27291 (v850_float_nz_comparison_operator): Likewise.
27293 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27295 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
27296 with calls to gen_insvsi/gen_insvdi.
27298 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
27300 * config/i386/i386.c (initial_ix86_tune_features): Set up
27301 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
27303 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
27304 Jakub Jelinek <jakub@redhat.com>
27306 PR tree-optimization/48189
27307 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
27308 If nitercst is 0, don't predict the exit edge.
27310 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
27312 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
27313 in asm_fprintf with reg_names.
27314 (aarch64_print_operand_address): Likewise.
27315 (aarch64_return_addr): Likewise.
27316 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
27318 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27320 * config/pa/pa.h (VAL_U6_BITS_P): Define.
27321 (INT_U6_BITS): Likewise.
27322 * config/pa/predicates.md (uint6_operand): New predicate.
27323 (shift5_operand, shift6_operand): Likewise.
27324 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
27326 (lshrdi3): Use shift6_operand.
27327 (shrpsi4, shrpdi4): New insn patterns.
27328 (extzv): Delete expander.
27329 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
27330 predicates in unamed zero extract patterns. Tighten common constraint.
27331 (extv): Delete expander.
27332 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
27333 predicates in unamed sign extract patterns. Tighten common constraint.
27334 (insv): Delete expander.
27335 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
27336 predicates in unamed insert patterns. Tighten common constraint.
27337 Change uint32_operand predicate to uint6_operand predicate in unamed
27338 DImode pattern to insert constant values of type 1...1xxxx.
27340 2013-01-04 Jan Hubicka <jh@suse.cz>
27342 PR tree-optimization/55823
27343 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
27346 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27347 Uros Bizjak <ubizjak@gmail.com>
27349 PR rtl-optimization/55845
27350 * df-problems.c (can_move_insns_across): Stop scanning at
27351 volatile_insn_p source instruction or give up if
27352 across_from .. across_to range contains any volatile_insn_p
27355 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27357 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
27358 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
27360 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
27361 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
27363 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27366 * asan.c (asan_clear_shadow): New function.
27367 (asan_emit_stack_protection): Use it.
27369 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
27371 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
27372 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
27373 with tab instead of space.
27375 2013-01-08 Nick Clifton <nickc@redhat.com>
27377 * config/rl78/rl78.c (rl78_expand_prologue): Always select
27378 register bank 0 at the start of an interrupt handler.
27379 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
27382 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27384 * config/aarch64/aarch64-simd.md
27385 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
27386 (aarch64_simd_bsl): Likewise.
27387 (aarch64_vcond_internal<mode>): Likewise.
27388 (vcond<mode><mode>): Likewise.
27389 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
27390 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
27392 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
27394 * config/aarch64/aarch64-builtins.c
27395 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
27397 2013-01-08 Martin Jambor <mjambor@suse.cz>
27400 * tree-sra.c (analyze_access_subtree): Return true also after
27401 potentially creating a debug-only replacement.
27403 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27405 PR middle-end/55890
27406 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
27408 PR tree-optimization/54120
27409 * tree-vrp.c (range_fits_type_p): Don't allow
27410 src_precision < precision from signed vr to unsigned_p
27411 if vr->min or vr->max is negative.
27412 (simplify_float_conversion_using_ranges): Test can_float_p
27413 against CODE_FOR_nothing.
27415 2013-01-08 Jakub Jelinek <jakub@redhat.com>
27416 Richard Biener <rguenther@suse.de>
27418 PR middle-end/55851
27419 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
27420 types instead of just INTEGER_TYPE types.
27422 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
27424 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
27427 2013-01-07 Steve Ellcey <sellcey@mips.com>
27430 * config/mips/mips.opt: Change mad to mmad to match documentation.
27432 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27435 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
27436 .progmemx.data now.
27438 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27441 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
27442 (avr_addrspace_t): Add .section_name field.
27443 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
27445 (avr_addrspace): Same. Initialize .section_name. Remove last
27446 NULL entry. Put __memx into .progmemx.data.
27447 (progmem_section_prefix): Remove.
27448 (avr_asm_init_sections): No need to initialize progmem_section.
27449 (avr_asm_named_section): Use avr_addrspace[].section_name to get
27450 section name prefix.
27451 (avr_asm_select_section): Ditto. And use get_unnamed_section to
27452 retrieve the progmem section.
27453 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
27454 boundary to run over avr_addrspace[].
27455 (avr_register_target_pragmas): Ditto.
27457 2013-01-06 Jakub Jelinek <jakub@redhat.com>
27459 * varasm.c (output_constant_def_contents): For asan_protect_global
27460 protected strings, adjust DECL_ALIGN if needed, before testing for
27462 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
27463 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
27465 (output_object_block): For asan protected decls, emit asan padding
27466 after their contents.
27467 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
27468 (asan_finish_file): Test it here instead.
27470 2013-01-07 Nick Clifton <nickc@redhat.com>
27471 Matthias Klose <doko@debian.org>
27472 Doug Kwan <dougkwan@google.com>
27473 H.J. Lu <hongjiu.lu@intel.com>
27476 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
27478 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
27480 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
27482 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
27484 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
27486 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27489 * doc/install.texi (Cross-Compiler-Specific Options): Document
27492 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
27494 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
27495 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
27496 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
27497 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
27498 vqmovun_high_s64): Fix source operand number and update copyright.
27500 2013-01-07 Richard Biener <rguenther@suse.de>
27502 PR middle-end/55890
27503 * gimple.h (gimple_call_builtin_p): New overload.
27504 * gimple.c (validate_call): New function.
27505 (gimple_call_builtin_p): Likewise.
27506 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
27507 Use gimple_call_builtin_p.
27508 (find_func_clobbers): Likewise.
27509 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
27510 (strlen_optimize_stmt): Likewise.
27512 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27514 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
27515 (vld1q_dup_*): Likewise.
27516 (vld1_*): Likewise.
27517 (vld1q_*): Likewise.
27518 (vld1_lane_*): Likewise.
27519 (vld1q_lane_*): Likewise.
27521 2013-01-07 Richard Biener <rguenther@suse.de>
27523 * lto-streamer.h (LTO_minor_version): Bump to 2.
27525 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
27527 * config/aarch64/aarch64-protos.h
27528 (aarch64_const_double_zero_rtx_p): Rename to...
27529 (aarch64_float_const_zero_rtx_p): ...this.
27530 (aarch64_float_const_representable_p): New.
27531 (aarch64_output_simd_mov_immediate): Likewise.
27532 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
27533 move immediate case.
27534 * config/aarch64/aarch64.c
27535 (aarch64_const_double_zero_rtx_p): Rename to...
27536 (aarch64_float_const_zero_rtx_p): ...this.
27537 (aarch64_print_operand): Allow printing of new constants.
27538 (aarch64_valid_floating_const): New.
27539 (aarch64_legitimate_constant_p): Check for valid floating-point
27541 (aarch64_simd_valid_immediate): Likewise.
27542 (aarch64_vect_float_const_representable_p): New.
27543 (aarch64_float_const_representable_p): Likewise.
27544 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
27545 (aarch64_output_simd_mov_immediate): New.
27546 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
27547 (*movdf_aarch64): Likewise.
27548 * config/aarch64/constraints.md (Ufc): New.
27549 (Y): call aarch64_float_const_zero_rtx.
27550 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
27552 2013-01-07 Richard Biener <rguenther@suse.de>
27554 PR tree-optimization/55888
27555 PR tree-optimization/55862
27556 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
27557 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
27558 not if it is contained therein.
27560 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27562 * config/avr/t-avr: Typo.
27564 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
27567 * config/avr/t-avr: Don't automatically rebuild
27568 $(srcdir)/config/avr/t-multilib
27569 $(srcdir)/config/avr/avr-tables.opt
27570 $(srcdir)/doc/avr-mmcu.texi
27571 (avr-mcus): New phony target to build them on request.
27572 (s-avr-mlib, s-avr-mmcu-texi): Remove.
27573 * avr/avr-mcus.def: Adjust comments.
27575 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
27577 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
27579 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27581 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
27583 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
27585 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
27587 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
27589 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
27590 to generate profiling.
27591 * config/rs6000/aix64.h (LIB_SPEC): Same.
27593 2013-01-04 Andrew Pinski <apinski@cavium.com>
27595 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
27597 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
27599 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
27601 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
27603 (ix86_expand_move): Ditto.
27604 (ix86_zero_extend_to_Pmode): Ditto.
27605 (ix86_expand_call): Ditto.
27606 (ix86_expand_special_args_builtin): Ditto.
27607 (ix86_expand_builtin): Ditto.
27609 2013-01-04 Richard Biener <rguenther@suse.de>
27611 PR tree-optimization/55862
27612 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
27613 translating them through PHI nodes.
27615 2013-01-04 Martin Jambor <mjambor@suse.cz>
27617 PR tree-optimization/55755
27618 * tree-sra.c (sra_modify_assign): Do not check that an access has no
27619 children when trying to avoid producing a VIEW_CONVERT_EXPR.
27621 2013-01-04 Marek Polacek <polacek@redhat.com>
27623 PR middle-end/55859
27624 * opts.c (default_options_optimization): Clarify error message.
27626 2013-01-04 Richard Biener <rguenther@suse.de>
27628 PR middle-end/55863
27629 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
27632 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27635 * config/pa/pa.md (movsi): Revert previous change.
27636 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
27639 2013-01-03 Richard Henderson <rth@redhat.com>
27641 * config/i386/i386.c (ix86_expand_move): Always assign to op1
27642 after eliminating TLS symbols.
27644 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
27647 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
27648 * graphite-poly.c (debug_gmp_value): Likewise.
27650 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
27653 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
27654 selected code model, define __code_mode_small__, __code_model_medium__,
27655 __code_model_large__, __code_model_32__ or __code_model_kernel__.
27656 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
27657 xchg temporary register with %k. Declare temporary register as
27659 [__x86_64__]: For medium and large code models, preserve %rbx register.
27661 2013-01-03 Richard Biener <rguenther@suse.de>
27663 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
27664 (dump_subscript): Adjust.
27665 (finalize_ddr_dependent): Do not dump redundant info.
27666 (analyze_siv_subscript): Adjust.
27667 (subscript_dependence_tester): Likewise.
27668 (compute_affine_dependence): Likewise.
27670 2013-01-03 Richard Biener <rguenther@suse.de>
27673 2013-01-03 Richard Biener <rguenther@suse.de>
27675 PR tree-optimization/55857
27676 * tree-vect-stmts.c (vectorizable_load): Do not setup
27677 re-alignment for invariant loads.
27679 2013-01-02 Richard Biener <rguenther@suse.de>
27681 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27682 invariant load do not generate a vector load from the scalar location.
27684 2013-01-03 Richard Biener <rguenther@suse.de>
27686 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
27687 for not vectorizing.
27688 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
27689 not build INDIRECT_REFs, call get_name once only.
27690 (vect_create_data_ref_ptr): Likewise. Dump base object kind
27691 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
27693 2013-01-03 Richard Biener <rguenther@suse.de>
27695 PR tree-optimization/55857
27696 * tree-vect-stmts.c (vectorizable_load): Do not setup
27697 re-alignment for invariant loads.
27699 2013-01-03 Richard Biener <rguenther@suse.de>
27702 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
27703 prefer a built-in decl.
27705 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27707 * gcc.c (process_command): Update copyright notice dates.
27708 * gcov.c (print_version): Likewise.
27709 * gcov-dump.c (print_version): Likewise.
27711 PR rtl-optimization/55838
27712 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
27713 iv0.step, iv1.step and step.
27715 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27716 Marc Glisse <marc.glisse@inria.fr>
27718 PR tree-optimization/55832
27719 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
27720 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
27721 integer_{one,zero}_node.
27723 2013-01-03 Jakub Jelinek <jakub@redhat.com>
27726 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
27727 * var-tracking.c (reverse_op): Don't add reverse ops to
27728 VALUEs that have already
27729 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
27731 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
27733 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
27735 2013-01-02 Teresa Johnson <tejohnson@google.com>
27737 * dumpfile.c (dump_loc): Print filename with location.
27738 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
27739 new location_t parameter to emit complete unroll message with
27740 new dump framework.
27741 (canonicalize_loop_induction_variables): Compute loops location
27742 and pass to try_unroll_loop_completely.
27743 * loop-unroll.c (report_unroll_peel): New function.
27744 (peel_loops_completely): Use new dump format with location
27745 for main dumpfile message, and invoke report_unroll_peel on success.
27746 (decide_unrolling_and_peeling): Ditto.
27747 (decide_peel_once_rolling): Remove old dumpfile message subsumed
27748 by report_unroll_peel.
27749 (decide_peel_completely): Ditto.
27750 (decide_unroll_constant_iterations): Ditto.
27751 (decide_unroll_runtime_iterations): Ditto.
27752 (decide_peel_simple): Ditto.
27753 (decide_unroll_stupid): Ditto.
27754 * cfgloop.c (get_loop_location): New function.
27755 * cfgloop.h (get_loop_location): Declare.
27757 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27759 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
27762 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27764 PR middle-end/55198
27765 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
27766 BLKmode objects when EXPAND_MEMORY is specified.
27768 2013-01-02 Sriraman Tallam <tmsriram@google.com>
27770 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
27772 (fold_builtin_cpu): Do not share cpu model decls across statements.
27774 2013-01-02 Jason Merrill <jason@redhat.com>
27777 * tree.c (build_array_type_1): Revert earlier change.
27779 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
27781 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
27783 * config/aarch64/aarch64-tune.md: Re-generate.
27785 2013-01-02 Richard Biener <rguenther@suse.de>
27787 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
27788 invariant load do not generate a vector load from the scalar location.
27790 2013-01-02 Richard Biener <rguenther@suse.de>
27793 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
27794 * configure: Regenerate.
27796 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27798 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
27799 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
27800 (expand_builtin_int_roundingfn_2): Keep the original target around
27801 for the fallback case.
27803 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
27805 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
27806 to be clear for sign changes.
27808 2013-01-01 Jan Hubicka <jh@suse.cz>
27810 * ipa-inline-analysis.c: Fix formatting.
27812 2013-01-01 Jakub Jelinek <jakub@redhat.com>
27814 PR tree-optimization/55831
27815 * tree-vect-loop.c (get_initial_def_for_induction): Use
27816 gsi_after_labels instead of gsi_start_bb.
27818 Copyright (C) 2013 Free Software Foundation, Inc.
27820 Copying and distribution of this file, with or without modification,
27821 are permitted in any medium without royalty provided the copyright
27822 notice and this notice are preserved.