1 2013-11-13 Joseph Myers <joseph@codesourcery.com>
3 * doc/extend.texi (Statement Exprs, Typeof): Discuss __auto_type.
4 * ginclude/stdatomic.h (kill_dependency, atomic_store_explicit)
5 (atomic_load_explicit, atomic_exchange_explicit)
6 (atomic_compare_exchange_strong_explicit)
7 (atomic_compare_exchange_weak_explicit): Use __auto_type to
8 declare variable initialized with PTR argument.
10 2013-11-12 Jeff Law <law@redhat.com>
12 * tree-ssa-threadedge.c (thread_around_empty_blocks): New
13 argument backedge_seen_p. Set, use and pass it to children
15 (thread_through_normal_block): Similarly.
16 (thread_across_edge): Similarly.
18 * gimple-ssa-isolate-paths.c (check_loadstore): Mark discovered
19 memory references as volatile.
20 (insert_trap_and_remove_trailing_statements): Fix comment.
22 2013-11-12 Vladimir Makarov <vmakarov@redhat.com>
25 * ira-costs.c (record_operand_costs): Check operands number for
28 2013-11-12 Michael Meissner <meissner@linux.vnet.ibm.com>
31 * config/rs6000/rs6000.md (movdi_internal32): Eliminate
32 constraints that would allow DImode into the traditional Altivec
33 registers, but cause undesirable code generation when loading 0 as
35 (movdi_internal64): Likewise.
36 (cmp<mode>_fpr): Do not use %x for CR register output.
37 (extendsfdf2_fpr): Fix constraints when -mallow-upper-df and
38 -mallow-upper-sf debug switches are used.
40 2013-11-12 Andrew MacLeod <amacleod@redhat.com>
42 * gimple-expr.h (create_tmp_var_name, create_tmp_var_raw,
43 create_tmp_var, create_tmp_reg, mark_addressable, is_gimple_reg_rhs):
44 Relocate prototypes from gimple.h.
45 * gimplify.h: New File. Relocate some prototypes from gimple.h here.
46 (gimple_predicate, enum fallback, enum gimplify_status): Relocate
48 * gimple.h: Move some prototypes to gimplify.h.
49 (gimple_predicate, enum fallback, enum gimplify_status): Move to
51 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location):
52 Relocate from gimpify.c.
53 * gimple-expr.c (remove_suffix, tmp_var_id_num, create_tmp_var_name,
54 create_tmp_var_raw, create_tmp_var, create_tmp_reg, mark_addressable,
55 is_gimple_reg_rhs) Relocate from gimplify.c.
56 * gimplify.c (mark_addressable): Move to gimple-expr.c.
57 (gimple_seq_add_stmt_without_update): Move to gimple.c.
58 (remove_suffix, tmp_var_id_num, create_tmp_var_name, create_tmp_var_raw,
59 create_tmp_var, create_tmp_reg, is_gimple_reg_rhs): Move to
61 (should_carry_location_p): Move to gimple.c.
62 (gimple_do_not_emit_location_p, gimple_set_do_not_emit_location): Move
64 (annotate_one_with_location, annotate_all_with_location_after,
65 annotate_all_with_location): Move to gimple.c.
66 (compare_case_labels, sort_case_labels,
67 preprocess_case_label_vec_for_gimple): Move to gimple.c.
68 (rhs_predicate_for): Make static.
69 (gimplify_assign): Relocate from gimple.c.
70 * gimple.c (gimplify_assign): Move to gimplify.c.
71 (gimple_seq_add_stmt_without_update, should_carry_location_p,
72 annotate_one_with_location, annotate_all_with_location_after,
73 annotate_all_with_location, compare_case_labels, sort_case_labels,
74 preprocess_case_label_vec_for_gimple): Relocate from gimplify.c.
75 * tree.h (unshare_expr, unshare_expr_without_location,
76 mark_addressable): Move prototypes to gimplify.h.
77 * Makefile.in (GTFILES): gimple-expr.c now has the GTY tag for
79 * asan.c: Include gimplify.h rather than gimple.h.
80 * cfgloopmanip.c: Likewise.
81 * cgraphunit.c: Likewise.
82 * cilk-common.c: Likewise.
83 * dwarf2out.c: Dont include gimple.h.
84 * fold-const.c: Include gimplify.h rather than gimple.h.
85 * function.c: Likewise.
86 * gimple-fold.c: Likewise.
87 * gimple-ssa-strength-reduction.c: Likewise.
88 * graphite-clast-to-gimple.c: Likewise.
89 * graphite-sese-to-poly.c: Likewise.
90 * ipa-prop.c: Likewise.
91 * ipa-split.c: Likewise.
93 * langhooks.c: Dont include gimple.h.
94 * loop-init.c: Include gimplify.h rather than gimple.h.
95 * omp-low.c: Likewise.
97 * stor-layout.c: Likewise.
98 * targhooks.c: Likewise.
99 * trans-mem.c: Likewise.
100 * tree-affine.c: Likewise.
101 * tree-cfg.c: Likewise.
102 * tree-cfgcleanup.c: Likewise.
103 * tree-complex.c: Likewise.
104 * tree-if-conv.c: Likewise.
105 * tree-inline.c: Likewise.
106 * tree-iterator.c: Likewise.
107 * tree-loop-distribution.c: Likewise.
108 * tree-nested.c: Likewise.
109 * tree-parloops.c: Likewise.
110 * tree-predcom.c: Likewise.
111 * tree-profile.c: Likewise.
112 * tree-scalar-evolution.c: Likewise.
113 * tree-sra.c: Likewise.
114 * tree-ssa-address.c: Likewise.
115 * tree-ssa-ccp.c: Likewise.
116 * tree-ssa-dce.c: Likewise.
117 * tree-ssa-forwprop.c: Likewise.
118 * tree-ssa-ifcombine.c: Likewise.
119 * tree-ssa-loop-im.c: Likewise.
120 * tree-ssa-loop-ivopts.c: Likewise.
121 * tree-ssa-loop-manip.c: Likewise.
122 * tree-ssa-loop-niter.c: Likewise.
123 * tree-ssa-loop-prefetch.c: Likewise.
124 * tree-ssa-loop-unswitch.c: Likewise.
125 * tree-ssa-math-opts.c: Likewise.
126 * tree-ssa-phiopt.c: Likewise.
127 * tree-ssa-phiprop.c: Likewise.
128 * tree-ssa-pre.c: Likewise.
129 * tree-ssa-propagate.c: Likewise.
130 * tree-ssa-reassoc.c: Likewise.
131 * tree-ssa-sccvn.c: Likewise.
132 * tree-ssa-strlen.c: Likewise.
133 * tree-ssa.c: Likewise.
134 * tree-switch-conversio: Likewise.n.c
135 * tree-tailcall.c: Likewise.
136 * tree-vect-data-refs.c: Likewise.
137 * tree-vect-generic.c: Likewise.
138 * tree-vect-loop-manip.c: Likewise.
139 * tree-vect-loop.c: Likewise.
140 * tree-vect-patterns.c: Likewise.
141 * tree-vect-stmts.c: Likewise.
143 * value-prof.c: Likewise.
144 * config/aarch64/aarch64.c: Include gimplify.h instead of gimple.h.
145 * config/alpha/alpha.c: Likewise.
146 * config/darwin.c: Likewise.
147 * config/i386/i386.c: Likewise.
148 * config/ia64/ia64.c: Likewise.
149 * config/mep/mep.c: Likewise.
150 * config/mips/mips.c: Likewise.
151 * config/rs6000/rs6000.c: Likewise.
152 * config/s390/s390.c: Likewise.
153 * config/sh/sh.c: Likewise.
154 * config/sparc/sparc.c: Likewise.
155 * config/spu/spu.c: Likewise.
156 * config/stormy16/stormy16.c: Likewise.
157 * config/tilegx/tilegx.c: Likewise.
158 * config/tilepro/tilepro.c: Likewise.
159 * config/xtensa/xtensa.c: Likewise.
161 2013-11-12 Adam Butcher <adam@jessamine.co.uk>
163 * tree.c (grow_tree_vec_stat): New function ...
164 * tree.h (grow_tree_vec_stat) (grow_tree_vec): ... and its declaration
167 2013-11-12 Marek Polacek <polacek@redhat.com>
169 * final.c (update_alignments): Initialize label to NULL_RTX.
171 2013-11-12 Jeff Law <law@redhat.com>
173 * gimple-ssa-isolate-paths.c (check_loadstore): New function.
174 (insert_trap_and_remove_trailing_statements): New argument OP which
175 is the NULL pointer. Emit the trap after the load/store through
176 the NULL pointer. Simplify the RHS of a store through a NULL pointer
177 when trivial to do so.
178 (isolate_path): Corresponding changes.
179 (gimple_ssa_isolate_erroneous_path): Likewise.
181 2013-11-12 Teresa Johnson <tejohnson@google.com>
182 Jan Hubicka <jh@suse.cz>
184 * predict.c (drop_profile): New function.
185 (handle_missing_profiles): Ditto.
186 (counts_to_freqs): Don't overwrite estimated frequencies
187 when function has no profile counts.
188 * predict.h (handle_missing_profiles): Declare.
189 * tree-inline.c (freqs_to_counts): New function.
190 (copy_cfg_body): Invoke freqs_to_counts as needed.
191 * tree-profile.c (tree_profiling): Invoke handle_missing_profiles.
193 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
196 * config/i386/x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL):
198 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Set for m_HASWELL.
200 2013-11-12 H.J. Lu <hongjiu.lu@intel.com>
203 * config/i386/i386.c (ix86_option_override_internal): Check
204 X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL and
205 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL for
206 MASK_AVX256_SPLIT_UNALIGNED_LOAD and
207 MASK_AVX256_SPLIT_UNALIGNED_STORE.
209 * config/i386/x86-tune.def (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL):
210 Clear m_COREI7_AVX and update comments.
211 (X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): Likewise.
213 2013-11-12 Martin Jambor <mjambor@suse.cz>
215 PR rtl-optimization/10474
216 * ira.c (interesting_dest_for_shprep): New function.
217 (split_live_ranges_for_shrink_wrap): Likewise.
218 (find_moveable_pseudos): Move calculation of dominance info,
219 df_analysios and the final anlyses to...
220 (ira): ...here, call split_live_ranges_for_shrink_wrap.
222 2013-11-12 Bin Cheng <bin.cheng@arm.com>
224 * tree-ssa-loop-ivopts.c (force_expr_to_var_cost): Refactor the code.
225 Handle type conversion.
227 2013-11-11 Martin Liska <marxin.liska@gmail.com>
228 Jan Hubicka <jh@suse.cz>
230 * cgraph.c (dump_cgraph_node): Profile dump added.
231 * cgraph.h (struct cgraph_node): New time profile variable added.
232 * cgraphclones.c (cgraph_clone_node): Time profile is cloned.
233 * gcov-io.h (gcov_type): New profiler type introduced.
234 * ipa-profile.c (lto_output_node): Streaming for time profile added.
235 (input_node): Time profiler is read from LTO stream.
236 * predict.c (maybe_hot_count_p): Hot prediction changed.
237 * profile.c (instrument_values): New case for time profiler added.
238 (compute_value_histograms): Read of time profile.
239 * tree-pretty-print.c (dump_function_header): Time profiler is dumped.
240 * tree-profile.c (init_ic_make_global_vars): Time profiler function added.
241 (gimple_init_edge_profiler): TP function instrumentation.
242 (gimple_gen_time_profiler): New.
243 * value-prof.c (gimple_add_histogram_value): Support for time profiler
245 (dump_histogram_value): TP type added to dumps.
246 (visit_hist): More sensitive check that takes TP into account.
247 (gimple_find_values_to_profile): TP instrumentation.
248 * value-prof.h (hist_type): New histogram type added.
249 (struct histogram_value_t): Pointer to struct function added.
250 * libgcc/Makefile.in: New GCOV merge function for TP added.
251 * libgcov.c: function_counter variable introduced.
252 (_gcov_merge_time_profile): New.
253 (_gcov_time_profiler): New.
255 2013-11-11 Marc Glisse <marc.glisse@inria.fr>
256 Jeff Law <law@redhat.com>
258 * tree-ssa-alias.c (stmt_kills_ref_p_1): Use
259 ao_ref_init_from_ptr_and_size for builtins.
261 2013-11-11 Uros Bizjak <ubizjak@gmail.com>
262 H.J. Lu <hongjiu.lu@intel.com>
265 * config/i386/x86-tune.def
266 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
267 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES.
269 (TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Rename from
270 TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES. Update for renamed
271 X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES.
272 * config/i386/i386.c (ix86_expand_set_or_movmem): Use
273 TARGET_MISALIGNED_MOVE_STRING_PRO_EPILOGUES to calculate
274 misaligned_prologue_used. Check that
275 desired_aling <= epilogue_size_needed.
277 2013-11-11 Cong Hou <congh@google.com>
279 PR tree-optimization/59050
280 * tree-vect-data-refs.c (comp_dr_addr_with_seg_len_pair): Bug fix.
282 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
285 * expmed.c (emit_store_flag): Fail for const-const comparison.
287 2013-11-11 Tristan Gingold <gingold@adacore.com>
288 Eric Botcazou <ebotcazou@adacore.com>
290 * tree.h (CONSTRUCTOR_NO_CLEARING): Define.
291 * tree-core.h (CONSTRUCTOR_NO_CLEARING): Document it.
292 * tree.def (CONSTRUCTOR): Likewise.
293 * doc/generic.texi (CONSTRUCTOR): Likewise. Update description.
294 * gimplify.c (gimplify_init_constructor): Do not clear the object when
295 the constructor is incomplete and CONSTRUCTOR_NO_CLEARING is set.
297 2013-11-11 Basile Starynkevitch <basile@starynkevitch.net>
299 * toplev.c (toplev_main): Move PLUGIN_FINISH invocation before
302 2013-11-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
304 * config/arm/arm.c (arm_new_rtx_costs): Return after handling
307 2013-11-11 Joern Rennecke <joern.rennecke@embecosm.com>
309 * config/arc/arc.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
311 2013-11-08 Jeff Law <law@redhat.com>
313 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
314 threading paths first, then perform PHI node checks if applicable.
316 2013-11-10 Karlson2k <k2k@narod.ru>
317 Kai Tietz <ktietz@redhat.com>
320 * configure.ac: Adding for exported symbols check
321 and for rdynamic-check executable-extension.
322 * configure: Regenerated.
324 2013-11-10 Uros Bizjak <ubizjak@gmail.com>
326 * mode-switching.c (optimize_mode_switching): Mark block as
327 nontransparent, if last_mode at block exit is different from no_mode.
329 2013-11-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
331 * function.c (NAME__MAIN): Move to...
332 * cfgexpand.c (NAME__MAIN): ...here.
334 2013-11-09 Richard Sandiford <rdsandiford@googlemail.com>
336 * target.def (can_use_doloop_p): New hook.
337 * doc/tm.texi.in (TARGET_CAN_USE_DOLOOP_P): Add.
338 * doc/tm.texi: Regenerate.
339 * doc/md.texi (doloop_begin, doloop_end): Update documentation.
340 * hooks.h (hook_bool_dint_dint_uint_true): Declare.
341 * hooks.c (hook_bool_dint_dint_uint_true): New function.
342 * targhooks.h (can_use_doloop_if_innermost): Declare.
343 * targhooks.c (can_use_doloop_if_innermost): New function.
344 * target.h: Include double-int.h.
345 * loop-doloop.c (doloop_optimize): Call targetm.can_use_doloop_p.
346 Remove iteration count, maximum iteration count, loop depth and
347 enter-at-top inputs from doloop_begin and doloop_end.
348 * config/arc/arc.md (doloop_begin, doloop_end): Update for new
350 * config/arc/arc.c (arc_can_use_doloop_p): New function.
351 (TARGET_CAN_USE_DOLOOP_P): Define.
352 * config/arm/thumb2.md (doloop_end): Update for new interface.
353 * config/arm/arm.c (TARGET_CAN_USE_DOLOOP_P): Define.
354 * config/bfin/bfin.md (doloop_end): Update for new interface.
355 * config/bfin/bfin.c (bfin_can_use_doloop_p): New function.
356 (TARGET_CAN_USE_DOLOOP_P): Define.
357 * config/c6x/c6x.md (doloop_end): Update for new interface.
358 * config/ia64/ia64.md (doloop_end): Update for new interface.
359 * config/ia64/ia64.c (TARGET_CAN_USE_DOLOOP_P): Define.
360 * config/mep/mep.md (doloop_begin, doloop_end): Update for new
362 * config/mep/mep.c (mep_emit_doloop): Likewise.
363 (TARGET_CAN_USE_DOLOOP_P): Define.
364 * config/rs6000/rs6000.md (doloop_end): Update for new interface.
365 * config/rs6000/rs6000.c (TARGET_CAN_USE_DOLOOP_P): Define.
366 * config/s390/s390.md (doloop_end): Update for new interface.
367 * config/sh/sh.md (doloop_end): Likewise.
368 * config/spu/spu.md (doloop_end): Likewise.
369 * config/spu/spu.c (TARGET_CAN_USE_DOLOOP_P): Define.
370 * config/tilegx/tilegx.md (doloop_end): Update for new interface.
371 * config/tilegx/tilegx.c (TARGET_CAN_USE_DOLOOP_P): Define.
372 * config/tilepro/tilepro.md (doloop_end): Update for new interface.
373 * config/tilepro/tilepro.c (TARGET_CAN_USE_DOLOOP_P): Define.
374 * config/v850/v850.md (doloop_begin, doloop_end): Update for new
376 * config/v850/v850.c (TARGET_CAN_USE_DOLOOP_P): Define.
378 2013-11-08 H.J. Lu <hongjiu.lu@intel.com>
381 * doc/extend.texi: Move Cilk Plus Builtins node before Other
384 2013-11-08 Andrew MacLeod <amacleod@redhat.com>
385 Joseph Myers <joseph@codesourcery.com>
387 * ginclude/stdatomic.h: New file.
388 * Makefile.in (USER_H): Add stdatomic.h.
390 2013-11-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
392 * config/arm/arm.c (arm_new_rtx_costs): Break after handling
395 2013-11-08 Jeff Law <law@redhat.com>
397 * tree-ssa-threadupdate.h (delete_thread_path): Declare.
398 * tree-ssa-threadupdate.c (delete_thread_path): New function.
399 (ssa_redirect_edges, thread_block_1): Use it.
400 (thread_through_loop_header, mark_threaded_blocks): Likewise.
401 (thread_through_all_blocks, register_jump_thread): Likewise.
402 * tree-ssa-threadedge.c (thread_across_edge): Likewise.
404 2013-11-08 James Greenhalgh <james.greenhalgh@arm.com>
406 * config/arm/aarch-common.c
407 (search_term): New typedef.
408 (shift_rtx_costs): New array.
409 (arm_rtx_shift_left_p): New.
410 (arm_find_sub_rtx_with_search_term): Likewise.
411 (arm_find_sub_rtx_with_code): Likewise.
412 (arm_early_load_addr_dep): Add sanity checking.
413 (arm_no_early_alu_shift_dep): Likewise.
414 (arm_no_early_alu_shift_value_dep): Likewise.
415 (arm_no_early_mul_dep): Likewise.
416 (arm_no_early_store_addr_dep): Likewise.
418 2013-11-08 Richard Biener <rguenther@suse.de>
420 PR tree-optimization/59047
421 * tree-predcom.c (ref_at_iteration): Handle bitfield accesses properly.
423 2013-11-08 Ilya Enkovich <ilya.enkovich@intel.com>
425 * common.opt (fcheck-pointer-bounds): Move to ...
426 * c-family/c.opt: ... here.
427 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): Remove.
428 (LANG_HOOKS_INITIALIZER): Remove LANG_HOOKS_CHKP_SUPPORTED.
429 * langhooks.h (lang_hooks): Remove chkp_supported field.
430 * toplev.c (process_options): Remove chkp_supported check.
432 2013-11-08 Richard Biener <rguenther@suse.de>
434 PR tree-optimization/59038
435 PR tree-optimization/58955
436 * tree-loop-distribution.c (pg_add_dependence_edges): Revert
437 previous change. Handle known dependences correctly.
439 2013-11-08 Tom de Vries <tom@codesourcery.com>
441 * config/rs6000/t-xilinx: Remove duplicate contents.
443 2013-11-07 Andrew MacLeod <amacleod@redhat.com>
444 Joseph Myers <joseph@codesourcery.com>
446 * tree-core.h (enum cv_qualifier): Add TYPE_QUAL_ATOMIC.
447 (enum tree_index): Add TI_ATOMICQI_TYPE, TI_ATOMICHI_TYPE,
448 TI_ATOMICSI_TYPE, TI_ATOMICDI_TYPE and TI_ATOMICTI_TYPE.
449 (struct tree_base): Add atomic_flag field.
450 * tree.h (TYPE_ATOMIC): New accessor macro.
451 (TYPE_QUALS, TYPE_QUALS_NO_ADDR_SPACE): Add TYPE_QUAL_ATOMIC.
452 (TYPE_QUALS_NO_ADDR_SPACE_NO_ATOMIC): New macro.
453 (atomicQI_type_node, atomicHI_type_node, atomicSI_type_node)
454 (atomicDI_type_node, atomicTI_type_node): New macros for type nodes.
455 * tree.c (set_type_quals): Set TYPE_ATOMIC.
456 (find_atomic_core_type): New function.
457 (build_qualified_type): Adjust alignment for qualified types.
458 (build_atomic_base): New function
459 (build_common_tree_nodes): Build atomicQI_type_node,
460 atomicHI_type_node, atomicSI_type_node, atomicDI_type_node and
462 * print-tree.c (print_node): Print atomic qualifier.
463 * tree-pretty-print.c (dump_generic_node): Print atomic type attribute.
464 * target.def (atomic_assign_expand_fenv): New hook.
465 * doc/tm.texi.in (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New @hook.
466 * doc/tm.texi: Regenerate.
467 * targhooks.c (default_atomic_assign_expand_fenv): New function.
468 * targhooks.h (default_atomic_assign_expand_fenv): Declare.
469 * sync-builtins.def (__atomic_feraiseexcept): New built-in function.
470 * config/i386/i386-builtin-types.def (VOID_FTYPE_PUSHORT): New
472 * config/i386/i386.c (enum ix86_builtins): Add
473 IX86_BUILTIN_FNSTENV, IX86_BUILTIN_FLDENV, IX86_BUILTIN_FNSTSW and
475 (bdesc_special_args): Add __builtin_ia32_fnstenv,
476 __builtin_ia32_fldenv, __builtin_ia32_fnstsw and __builtin_ia32_fnclex.
477 (ix86_expand_builtin): Handle the new built-in functions.
478 (ix86_atomic_assign_expand_fenv): New function.
479 (TARGET_ATOMIC_ASSIGN_EXPAND_FENV): New macro.
480 * config/i386/i386.md (UNSPECV_FNSTENV, UNSPECV_FLDENV)
481 (UNSPECV_FNSTSW, UNSPECV_FNCLEX): New unspecs.
482 (fnstenv, fldenv, fnstsw, fnclex): New insns.
484 2013-11-07 Steve Ellcey <sellcey@mips.com>
486 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add fp64 directory.
487 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add -mfp64 flag.
488 (MULTILIB_DIRNAMES): Add fp64 directory.
489 (MULTILIB_EXCEPTIONS): Add new exclusions.
491 2013-11-07 Aldy Hernandez <aldyh@redhat.com>
493 * gimplify.c (gimple_regimplify_operands): Do not set
495 * graphite-sese-to-poly.c (remove_simple_copy_phi): Same.
496 (rewrite_close_phi_out_of_ssa): Same.
497 (rewrite_phi_out_of_ssa): Same.
498 (rewrite_degenerate_phi): Same.
499 (handle_scalar_deps_crossing_scop_limits): Same.
500 * tree-if-conv.c (predicate_scalar_phi): Same.
501 * tree-parloops.c (create_loads_for_reductions): Same.
502 (create_final_loads_for_reduction): Same.
503 (create_loads_and_stores_for_name): Same.
504 (transform_to_exit_first_loop): Same.
505 (create_parallel_loop): Same.
507 (move_computations_dom_walker::before_dom_children): Same.
508 * tree-ssa-loop-manip.c (rewrite_phi_with_iv): Same.
509 * tree-ssa-pre.c (eliminate_dom_walker::before_dom_children): Same.
510 * tree-ssa-propagate.c (substitute_and_fold): Same.
511 * tree-vect-loop.c (vect_finalize_reduction): Same.
512 * tree-vect-stmts.c (vectorizable_call): Same.
514 2013-11-07 Mike Stump <mikestump@comcast.net>
516 * config/pdp11/pdp11.c: Include dbxout.h.
517 * config/picochip/picochip.c: Likewise.
519 2013-11-07 Cong Hou <congh@google.com>
521 PR tree-optimization/56764
522 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
523 Combine alias checks if it is possible to amortize the runtime
524 overhead. Return the number of alias checks after merging.
525 * tree-vect-data-refs.c (vect_prune_runtime_alias_test_list):
526 Use the function vect_create_cond_for_alias_checks () to check
527 the number of alias checks.
529 2013-11-07 Jeff Law <law@redhat.com>
531 * varpool.c (ctor_for_folding): Fix typo in comment.
533 2013-11-07 Joern Rennecke <joern.rennecke@embecosm.com>
535 * config/arc/arc.c (arc_ifcvt): Use commutativity, e.g.:
536 reg_a := reg_b + reg_a ==> reg_a := reg_a + reg_b
538 2013-11-07 Jeff Law <law@redhat.com>
540 * doc/invoke.texi (-fisolate-erroneous-paths): Document.
542 * gimple-ssa-isolate-paths.c (gate_isolate_erroneous_paths):
543 No longer check if we have __builtin_trap, assume it's available.
545 2013-11-07 Diego Novillo <dnovillo@google.com>
547 * attribs.c (lookup_scoped_attribute_spec): Make static.
548 (get_attribute_namespace): Likewise.
549 * builtins.c (more_const_call_expr_args_p): Move from tree.h.
550 (validate_arglist): Move earlier in the file. Make static.
551 (expand_stack_restore): Move from stmt.c
552 (expand_stack_save): Move from stmt.c
553 (rewrite_call_expr_array): Move earlier in the file.
554 (rewrite_call_expr_valist): Likewise.
555 * cfgexpand.c: Include hard-reg-set.h before tree.h
558 (expand_asm_loc): Move from stmt.c.
559 (n_occurrences): Move from stmt.c.
560 (check_operand_nalternatives): Move from stmt.c.
561 (tree_conflicts_with_clobbers_p): Move from stmt.c.
562 (expand_asm_operands): Move from stmt.c
563 (expand_asm_stmt): Move from stmt.c
564 (expand_computed_goto): Move from stmt.c
565 (expand_goto): Move from stmt.c
566 (expand_null_return_1): Move from stmt.c
567 (expand_null_return): Move from stmt.c
568 (expand_value_return): Move from stmt.c
569 (expand_return): Move from stmt.c
570 (expand_main_function): Move from function.c
571 (stack_protect_prologue): Move from function.c
572 * cgraphclones.c (build_function_type_skip_args): Move from tree.c.
573 (build_function_decl_skip_args): Move from tree.c.
574 * explow.c (tree_expr_size): Move from tree.c.
575 * expr.c (addr_expr_of_non_mem_decl_p): Remove.
576 (fields_length): Move from tree.c.
577 * fold-const.c (size_low_cst): Move from tree.c.
578 (tree_expr_nonzero_warnv_p): Make static. Move earlier in the file.
579 (tree_expr_nonzero_p): Make static. Move earlier in the file.
580 (fold_build3_initializer_loc): Remove.
581 (tree_invalid_nonnegative_warnv_p): Make static.
582 * function.c (expand_main_function): Move to cfgexpand.c.
583 (stack_protect_prologue): Move to cfgexpand.c.
584 (set_insn_locations): Move earlier in the file.
585 * gimple-fold.c: Include langhooks.h.
586 (truth_type_for): Move from tree.c.
587 * print-tree.c (print_vec_tree): Remove.
588 * stmt.c (expand_computed_goto): Move to cfgexpand.c.
589 (expand_goto): Move to cfgexpand.c.
590 (n_occurrences): Move to cfgexpand.c.
591 (expand_asm_loc): Move to cfgexpand.c
592 (tree_conflicts_with_clobbers_p): Move to cfgexpand.c.
593 (expand_asm_operands): Move to cfgexpand.c.
594 (expand_asm_stmt): Move to cfgexpand.c.
595 (check_operand_nalternatives): Move to cfgexpand.c
596 (expand_null_return): Move to cfgexpand.c.
597 (expand_value_return): Move to cfgexpand.c.
598 (expand_null_return_1): Move to cfgexpand.c.
599 (expand_return): Move to cfgexpand.c.
600 (expand_stack_save): Move to builtins.c.
601 (expand_stack_restore): Move to builtins.c
602 * symtab.c: Include output.h.
603 (decl_assembler_name_hash): Move from tree.c.
604 (decl_assembler_name_equal): Move from tree.c.
605 * trans-mem.c (is_tm_safe_or_pure): Move from tree.h.
606 * tree-eh.c (in_array_bounds_p): Move from tree.c.
607 (range_in_array_bounds_p): Move from tree.c.
608 * tree-object-size.c (fini_object_sizes): Make static.
609 * tree-ssa-dom.c (iterative_hash_exprs_commutative): Move from tree.h.
610 * tree-vrp.c (ssa_name_nonnegative_p): Remove.
611 * tree.c (decl_assembler_name_equal): Move to symtab.c.
612 (tree_expr_size): Move to explow.c.
613 (decl_assembler_name_hash): Move to symtab.c.
615 (tree_expr_size): Move to explow.c.
616 (stabilize_reference_1): Move earlier in the file. Make static.
617 (omp_remove_redundant_declare_simd_attrs): Remove.
618 (simple_cst_list_equal): Move earlier in the file. Make static.
619 (size_low_cst): Move to fold-const.c.
620 (build_type_no_quals): Remove.
621 (build_function_type_skip_args): Move to cgraphclones.c.
622 (build_function_decl_skip_args): Move to cgraphclones.c.
623 (in_array_bounds_p): Move to tree-eh.c.
624 (range_in_array_bounds_p): Move to tree-eh.c.
625 (truth_type_for): Move to gimple-fold.c.
626 (list_equal_p): Remove.
627 * tree.h (decl_assembler_name_equal): Remove.
628 (decl_assembler_name_hash): Remove.
629 (truth_type_for): Remove.
630 (build_type_no_quals): Remove.
631 (build_function_decl_skip_args): Remove.
632 (in_array_bounds_p): Remove.
633 (range_in_array_bounds_p): Remove.
634 (size_low_cst): Remove.
635 (omp_remove_redundant_declare_simd_attrs): Remove.
636 (tree_expr_size): Remove.
637 (fields_length): Remove.
638 (stabilize_reference_1): Remove.
639 (expand_goto): Remove.
640 (expand_stack_save): Remove.
641 (expand_stack_restore): Remove.
642 (expand_return): Remove.
643 (fold_build3_initializer_loc): Remove.
644 (tree_expr_nonzero_p): Remove.
645 (tree_invalid_nonnegative_warnv_p): Remove.
646 (tree_expr_nonzero_warnv_p): Remove.
647 (fold_builtin_snprintf_chk): Remove.
648 (validate_arglist): Remove.
649 (iterative_hash_exprs_commutative): Move to tree-ssa-dom.c.
650 (simple_cst_list_equal): Remove.
652 (expand_main_function): Remove.
653 (stack_protect_prologue): Remove.
654 (print_vec_tree): Remove.
655 (lookup_scoped_attribute_spec): Remove.
656 (get_attribute_namespace): Remove.
657 (expand_computed_goto): Remove.
658 (expand_asm_stmt): Remove.
659 (list_equal_p): Remove.
660 (ssa_name_nonnegative_p): Remove.
661 (fini_object_sizes): Remove.
662 (addr_expr_of_non_mem_decl_p): Remove.
663 (is_tm_safe_or_pure): Move to trans-mem.c.
664 (more_const_call_expr_args_p): Remove.
665 (save_vtable_map_decl): Remove.
667 2013-11-07 Thomas Schwinge <thomas@codesourcery.com>
669 * doc/sourcebuild.texi (Top Level) <lto-plugin>: GNU ld can use
672 * config/arc/arc.h (LINK_COMMAND_SPEC): For -ftree-parallelize-loops=*,
673 link to libgomp and its dependencies.
674 * config/ia64/hpux.h (LIB_SPEC): Likewise.
675 * config/pa/pa-hpux11.h (LIB_SPEC): Likewise.
676 * config/pa/pa64-hpux.h (LIB_SPEC): Likewise.
677 * gcc.c (GOMP_SELF_SPECS): Update comment about libgomp's dependencies.
679 2013-11-07 Jakub Jelinek <jakub@redhat.com>
681 * tree-ssa-loop-niter.c: Include tree-ssanames.h.
682 (determine_value_range): Add loop argument. Use get_range_info to
684 (bound_difference): Adjust caller.
686 2013-11-07 Richard Biener <rguenther@suse.de>
687 Jakub Jelinek <jakub@redhat.com>
689 * tree-vrp.c (find_assert_locations): Pre-seed live bitmaps for loop
690 latches from header PHI arguments from the latch edge.
692 2013-11-07 Paolo Carlini <paolo.carlini@oracle.com>
695 * varasm.c (output_constant): Handle NULLPTR_TYPE.
697 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
699 * config/i386/i386.c (ix86_expand_set_or_movmem): Don't set
700 misaligned_prologue_used when it has been set.
702 2013-11-07 Yury Gribov <y.gribov@samsung.com>
703 Jakub Jelinek <jakub@redhat.com>
706 * asan.c (get_mem_refs_of_builtin_call): Allow
707 integer literals as addresses in instrumented builtins.
709 2013-11-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
711 * config/aarch64/aarch64.c (aarch64_legitimize_reload_address):
712 Explain why plus_constant is not used.
714 2013-11-07 Richard Biener <rguenther@suse.de>
716 * tree-ssa-ccp.c (canonicalize_float_value): Rename to ...
717 (canonicalize_value): ... this. Also handle stripping of
719 (get_value, set_lattice_value, get_value_for_expr): Adjust.
720 * gimple-fold.c (canonicalize_constructor_val): Strip TREE_OVERFLOW.
721 * tree-ssa-threadedge.c (set_ssa_name_value): Likewise.
723 2013-11-07 Richard Biener <rguenther@suse.de>
725 * tree-dfa.c (get_ref_base_and_extent): Fix casting.
727 2013-11-07 H.J. Lu <hongjiu.lu@intel.com>
730 * config/i386/i386.md (push peepholer/splitter): Use Pmode
731 with stack_pointer_rtx.
733 2013-11-07 Bin Cheng <bin.cheng@arm.com>
735 * tree-ssa-loop-ivopts.c (get_shiftadd_cost): Check equality
736 using operand_equal_p.
738 2013-11-07 Bin Cheng <bin.cheng@arm.com>
740 * tree-ssa-loop-ivopts.c (alloc_iv): Lower address expressions.
741 * tree-affine.c (get_inner_reference_aff): Return base.
742 * tree-affine.h (get_inner_reference_aff): Change prototype.
744 2013-11-06 Tobias Burnus <burnus@net-b.de>
746 * doc/invoke.texi (Wdate-time): Fix typo.
748 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
750 * config/sh/sh.md (addsf3, divsf3, divsf3_i, rsqrtsf2, cmpgtdf_t,
751 cmpeqdf_t, *ieee_ccmpeqdf_t, negdf2, sqrtdf2, absdf2): Use
752 fp_arith_reg_operand instead of arith_reg_operand.
754 2013-11-06 Oleg Endo <olegendo@gcc.gnu.org>
756 * config/sh/sh.md (adddi3): Remove empty constraints.
757 Remove can_create_pseudo_p and arith_reg_operand check.
758 (adddi3_compact, subdi3_compact, *negdi2): Remove constraints.
761 2013-11-06 Jeff Law <law@redhat.com>
762 Tom Tromey <tromey@redhat.com>
764 * gdbinit.in: Disable strict type checking.
766 2013-11-06 Vladimir Makarov <vmakarov@redhat.com>
768 * tree-pass.h (make_pass_live_range_shrinkage): New external.
769 * timevar.def (TV_LIVE_RANGE_SHRINKAGE): New.
770 * sched-rgn.c (gate_handle_live_range_shrinkage): New.
771 (rest_of_handle_live_range_shrinkage): Ditto
772 (class pass_live_range_shrinkage): Ditto.
773 (pass_data_live_range_shrinkage): Ditto.
774 (make_pass_live_range_shrinkage): Ditto.
775 * sched-int.h (initialize_live_range_shrinkage): New prototype.
776 (finish_live_range_shrinkage): Ditto.
777 * sched-deps.c (create_insn_reg_set): Make void return value.
778 * passes.def: Add pass_live_range_shrinkage.
779 * ira.c (update_equiv_regs): Don't move if flag_live_range_shrinkage.
780 * haifa-sched.c (live_range_shrinkage_p): New.
781 (initialize_live_range_shrinkage, finish_live_range_shrinkage):
783 (rank_for_schedule): Add code for pressure relief through live
785 (schedule_insn): Print more debug info.
786 (sched_init): Setup SCHED_PRESSURE_WEIGHTED for pressure relief
787 through live range shrinkage.
788 * doc/invoke.texi (-flive-range-shrinkage): New.
789 * common.opt (flive-range-shrinkage): New.
791 2013-11-06 Uros Bizjak <ubizjak@gmail.com>
794 * config/i386/i386.c (ix86_avx_u128_mode_needed): Require
795 AVX_U128_DIRTY mode for call_insn RTXes that use AVX256 registers.
796 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY mode for call_insn
797 RTXes that return in AVX256 register.
799 2013-11-06 Richard Biener <rguenther@suse.de>
801 PR tree-optimization/58653
802 * tree-predcom.c (ref_at_iteration): Rewrite to generate a MEM_REF.
803 (prepare_initializers_chain): Adjust.
805 2013-11-06 Andrew MacLeod <amacleod@redhat.com>
807 * gimple.h (block_in_transaction): Move to basic-block.h and rename.
808 (gimple_in_transaction): Use bb_in_transaction.
809 * basic-block.h (bb_in_transaction): Relocate here and rename.
810 * tree-ssa-loop-im.c (execute_sm): Use bb_in_transaction.
812 2013-11-06 Richard Biener <rguenther@suse.de>
814 * tree.c (drop_tree_overflow): New function.
815 * tree.h (drop_tree_overflow): Declare.
816 * gimplify.c (gimplify_expr): Drop TREE_OVERFLOW.
817 * tree-vrp.c (range_int_cst_singleton_p): Use
818 is_overflow_infinity instead of testing TREE_OVERFLOW.
819 (extract_range_from_assert): Likewise.
820 (zero_nonzero_bits_from_vr): Likewise.
821 (extract_range_basic): Likewise.
822 (register_new_assert_for): Use drop_tree_overflow.
823 (vrp_visit_phi_node): Likewise.
825 2013-11-06 Eric Botcazou <ebotcazou@adacore.com>
827 * config/i386/i386.c (ix86_expand_prologue): Optimize stack
828 checking for leaf functions without dynamic stack allocation.
829 * config/ia64/ia64.c (ia64_emit_probe_stack_range): Adjust.
830 (ia64_expand_prologue): Likewise.
831 * config/mips/mips.c (mips_expand_prologue): Likewise.
832 * config/rs6000/rs6000.c (rs6000_emit_prologue): Likewise.
833 * config/sparc/sparc.c (sparc_expand_prologue): Likewise.
834 (sparc_flat_expand_prologue): Likewise.
836 2013-11-06 James Greenhalgh <james.greenhalgh@arm.com>
838 * config/aarch64/arm_neon.h (__ST2_LANE_FUNC): Better model data size.
839 (__ST3_LANE_FUNC): Likewise.
840 (__ST4_LANE_FUNC): Likewise.
842 2013-11-06 Nick Clifton <nickc@redhat.com>
844 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define the
845 name returned by msp430_mcu_name.
846 (LIB_SPEC): If a -T option has not been specified then set a
847 default, mcu-specific, linker script.
848 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add more mcu names.
849 * config/msp430/msp430.c (msp430x_names): Likewise.
850 Alpha sort the names for ease of comparison.
851 (msp430_mcu_name): New function: Returns a string suitable for
852 use as a C preprocessor symbol based upon the name of the MCU
854 (msp430_option_override): Accept msp430x and msp430xv2 as generic
856 * config/msp430/msp430-protos.h (msp430_mcu_name): Prototype.
858 * gcc.c (do_spec_1): Do not insert a space after a %* substitution
859 unless it is the last part of a spec substring.
860 * doc/invoke.texi (Spec Files): Document space insertion
863 2013-11-06 Christian Bruel <christian.bruel@st.com>
865 * config/sh/sh-mem.cc (sh_expand_cmpnstr, sh_expand_cmpstr):
866 Factorize probabilities, Use adjust_address instead of
867 adjust_automodify_address when possible. Enable for optimize.
868 (sh_expand_strlen): New function.
869 * config/sh/sh-protos.h (sh_expand_strlen): Declare.
870 * config/sh/sh.md (strlensi): New pattern.
871 (UNSPEC_BUILTIN_STRLEN): Define.
873 2013-11-06 Jakub Jelinek <jakub@redhat.com>
876 * expr.c (get_bit_range): Handle *offset == NULL_TREE.
877 (expand_assignment): If *bitpos is negative, set *offset
878 and adjust *bitpos, so that it is not negative.
880 2013-11-06 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
882 * config/i386/bdver3.md : Added two additional decoder units
883 to support issue rate of 4 and remodeled vector unit.
884 * config/i386/i386.c (ix86_issue_rate): Issue rate for BD
885 architectures is set to 4.
886 * config/i386/i386.c (ia32_multipass_dfa_lookahead): DFA
887 lookahead is set to 4 for BD architectures.
889 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
891 * config/rs6000/rs6000.c (rs6000_option_override_internal):
892 Remove restriction against use of VSX instructions when generating
893 code for little endian mode.
895 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
897 * config/rs6000/altivec.md (mulv4si3): Ensure we generate vmulouh
898 for both big and little endian.
899 (mulv8hi3): Swap input operands for merge high and merge low
900 instructions for little endian.
902 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
904 * config/rs6000/altivec.md (vec_widen_umult_even_v16qi): Change
905 define_insn to define_expand that uses even patterns for big
906 endian and odd patterns for little endian.
907 (vec_widen_smult_even_v16qi): Likewise.
908 (vec_widen_umult_even_v8hi): Likewise.
909 (vec_widen_smult_even_v8hi): Likewise.
910 (vec_widen_umult_odd_v16qi): Likewise.
911 (vec_widen_smult_odd_v16qi): Likewise.
912 (vec_widen_umult_odd_v8hi): Likewise.
913 (vec_widen_smult_odd_v8hi): Likewise.
914 (altivec_vmuleub): New define_insn.
915 (altivec_vmuloub): Likewise.
916 (altivec_vmulesb): Likewise.
917 (altivec_vmulosb): Likewise.
918 (altivec_vmuleuh): Likewise.
919 (altivec_vmulouh): Likewise.
920 (altivec_vmulesh): Likewise.
921 (altivec_vmulosh): Likewise.
923 2013-11-05 Mike Stump <mikestump@comcast.net>
925 * Makefile.in (mostlyclean): Remove c-family objects.
927 2013-11-05 Ian Lance Taylor <iant@google.com>
929 * config/i386/sync.md (atomic_compare_and_swap<dwi>_doubleword):
930 If possible, add .cfi directives to record change to bx.
931 * config/i386/i386.c (ix86_emit_cfi): New function.
932 * config/i386/i386-protos.h (ix86_emit_cfi): Declare.
934 2013-11-05 Steven Bosscher <steven@gcc.gnu.org>
936 * rtlanal.c (tablejump_p): Expect a JUMP_TABLE_DATA to always follow
937 immediately after a label for a tablejump pattern.
939 * config/arm/arm.c (is_jump_table): Remove.
940 (create_fix_barrier): Use tablejump_p instead.
941 (arm_reorg): Likewise.
942 (thumb1_output_casesi): Expect JUMP_TABLE_DATA to always be NEXT_INSN.
943 (thumb2_output_casesi): Likewise.
944 * config/aarch64/aarch64.c (aarch64_output_casesi): Likewise.
945 * config/sh/sh.md (casesi_worker_1, casesi_worker_2,
946 casesi_shift_media, casesi_load_media): Likewise.
947 * config/iq2000/iq2000.md: Likewise (in anonymous define_insn).
948 * config/microblaze/microblaze.md: Likewise.
950 2013-11-05 Tobias Burnus <burnus@net-b.de>
952 * doc/invoke.texi (-Wdate-time): Document.
954 2013-11-05 Richard Sandiford <rdsandiford@googlemail.com>
956 * double-int.c (lshift_double, rshift_double): Remove
957 SHIFT_COUNT_TRUNCATED handling.
959 2013-11-05 Jeff Law <law@redhat.com>
961 * Makefile.in (OBJS): Add gimple-ssa-isolate-paths.o
962 * common.opt (-fisolate-erroneous-paths): Add option and documentation.
963 * gimple-ssa-isolate-paths.c: New file.
964 * gimple.c (check_loadstore): New function.
965 (infer_nonnull_range): Moved into gimple.c from tree-vrp.c
966 Verify OP is in the argument list and the argument corresponding
967 to OP is a pointer type. Use operand_equal_p rather than
968 pointer equality when testing if OP is on the nonnull list.
969 Use check_loadstore rather than count_ptr_derefs. Handle
970 GIMPLE_RETURN statements.
971 * tree-vrp.c (infer_nonnull_range): Remove.
972 * gimple.h (infer_nonnull_range): Declare.
973 * opts.c (default_options_table): Add OPT_fisolate_erroneous_paths.
974 * passes.def: Add pass_isolate_erroneous_paths.
975 * timevar.def (TV_ISOLATE_ERRONEOUS_PATHS): New timevar.
976 * tree-pass.h (make_pass_isolate_erroneous_paths): Declare.
977 * tree-ssa.c (struct count_ptr_d): Remove.
978 (count_ptr_derefs, count_uses_and_derefs): Remove.
979 * tree-ssa.h (count_uses_and_derefs): Remove.
981 2013-11-05 Jakub Jelinek <jakub@redhat.com>
983 PR rtl-optimization/58997
984 * loop-iv.c (iv_subreg): For IV_UNKNOWN_EXTEND, expect
985 get_iv_value to be in iv->mode rather than iv->extend_mode.
986 (iv_extend): Likewise. Otherwise, if iv->extend != extend,
987 use lowpart_subreg on get_iv_value before calling simplify_gen_unary.
988 * loop-unswitch.c (may_unswitch_on): Make sure op[i] is in the right
991 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
993 * gimple.h: Move some prototypes to gimple-expr.h and add to include
995 (extract_ops_from_tree, gimple_call_addr_fndecl, is_gimple_reg_type):
996 Move to gimple-expr.h.
997 * gimple-expr.h: New file. Relocate some prototypes from gimple.h.
998 (types_compatible_p, is_gimple_reg_type, is_gimple_variable,
999 is_gimple_id, virtual_operand_p, is_gimple_addressable,
1000 is_gimple_constant, extract_ops_from_tree, gimple_call_addr_fndecl):
1002 * gimple.c (extract_ops_from_tree_1, gimple_cond_get_ops_from_tree,
1003 gimple_set_body, gimple_body, gimple_has_body_p, is_gimple_lvalue,
1004 is_gimple_condexpr, is_gimple_addressable, is_gimple_constant,
1005 is_gimple_address, is_gimple_invariant_address,
1006 is_gimple_ip_invariant_address, is_gimple_min_invariant,
1007 is_gimple_ip_invariant, is_gimple_variable, is_gimple_id,
1008 virtual_operand_p, is_gimple_reg, is_gimple_val, is_gimple_asm_val,
1009 is_gimple_min_lval, is_gimple_call_addr, is_gimple_mem_ref_addr,
1010 gimple_decl_printable_name, useless_type_conversion_p,
1011 types_compatible_p, gimple_can_coalesce_p, copy_var_decl): Move to
1013 * gimple-expr.c: New File.
1014 (useless_type_conversion_p, gimple_set_body, gimple_body,
1015 gimple_has_body_p, gimple_decl_printable_name, copy_var_decl,
1016 gimple_can_coalesce_p, extract_ops_from_tree_1,
1017 gimple_cond_get_ops_from_tree, is_gimple_lvalue, is_gimple_condexpr,
1018 is_gimple_address, is_gimple_invariant_address,
1019 is_gimple_ip_invariant_address, is_gimple_min_invariant,
1020 is_gimple_ip_invariant, is_gimple_reg, is_gimple_val,
1021 is_gimple_asm_val, is_gimple_min_lval, is_gimple_call_addr,
1022 is_gimple_mem_ref_addr): Relocate here.
1023 * Makefile.in (OBJS): Add gimple-expr.o.
1025 2013-11-05 David Malcolm <dmalcolm@redhat.com>
1027 * gengtype-parse.c (struct_field_seq): Support empty structs.
1029 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1031 * config/i386/t-rtems (MULTILIB_MATCHES): Fix option typos.
1033 2013-11-05 Uros Bizjak <ubizjak@gmail.com>
1035 * config/i386/i386-c.c (ix86_target_macros): Define _SOFT_FLOAT
1037 * config/i386/rtemself.h (TARGET_OS_CPP_BUILTINS): Do not define
1039 (LONG_DOUBLE_TYPE_SIZE): New define.
1040 (LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Ditto.
1042 2013-11-05 Paolo Carlini <paolo.carlini@oracle.com>
1045 * doc/extend.texi [visibility ("visibility_type")]: Add example
1046 about visibility attribute on namespace declaration.
1048 2013-11-05 Richard Biener <rguenther@suse.de>
1051 * passes.def (all_passes): Start with pass_fixup_cfg again.
1053 2013-11-05 Richard Biener <rguenther@suse.de>
1055 PR tree-optimization/58955
1056 * tree-loop-distribution.c (pg_add_dependence_edges): Fix
1059 2013-11-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1061 * config/rs6000/vector.md (vec_pack_sfix_trunc_v2df): Adjust for
1063 (vec_pack_ufix_trunc_v2df): Likewise.
1065 2013-11-05 H.J. Lu <hongjiu.lu@intel.com>
1068 * doc/md.texi (@code{movmem@var{m}}): Specify Pmode as mode of
1069 pattern, instead of word_mode.
1071 * expr.c (emit_block_move_via_movmem): Don't use mode wider than
1073 (set_storage_via_setmem): Likewise.
1075 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
1077 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
1078 gimple.h and the rest of the condition in eliminate_build.
1079 (eliminate_build): Call new routine.
1080 * gimple.h (phi_ssa_name_p): Delete.
1082 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
1084 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
1085 case of no prefix and reserving zero slots, because when that's the
1086 case we'll never get here.
1087 * vec.h (va_heap::reserve): Don't try and handle
1088 vec_prefix::calculate_allocation returning zero because that should
1091 2013-11-05 Richard Biener <rguenther@suse.de>
1094 * tree-dfa.c (get_ref_base_and_extent): Merge common code
1095 in MEM_REF and TARGET_MEM_REF handling. Make sure to
1096 process trailing array detection before diving into the
1097 view-converted object (and possibly apply some extra offset).
1099 2013-11-05 Joseph Myers <joseph@codesourcery.com>
1101 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
1103 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
1105 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1107 PR tree-optimization/58958
1108 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
1109 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
1111 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
1113 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
1114 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
1116 2013-11-05 Jakub Jelinek <jakub@redhat.com>
1118 PR tree-optimization/58984
1119 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
1120 set *SIZE_P if non-NULL on success.
1121 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
1123 (ipcp_transform_function): Likewise. Punt if size of access
1124 is different from TYPE_SIZE on v->value's type.
1126 2013-11-05 Tobias Burnus <burnus@net-b.de>
1128 * doc/invoke.texi (-fopenmp-simd): Document new option.
1129 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
1130 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
1131 * tree.c (attribute_value_equal): Ditto.
1133 2013-11-04 Wei Mi <wmi@google.com>
1135 * sched-rgn.c (add_branch_dependences): Keep insns in
1136 a SCHED_GROUP at the end of BB to remain their location.
1138 2013-11-04 Wei Mi <wmi@google.com>
1140 * config/i386/i386.c (memory_address_length): Extract a part
1141 of code to rip_relative_addr_p.
1142 (rip_relative_addr_p): New Function.
1143 (ix86_macro_fusion_p): Ditto.
1144 (ix86_macro_fusion_pair_p): Ditto.
1145 * config/i386/i386.h: Add new tune features about macro-fusion.
1146 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1147 * doc/tm.texi: Generated.
1148 * doc/tm.texi.in: Ditto.
1149 * haifa-sched.c (try_group_insn): New Function.
1150 (group_insns_for_macro_fusion): Ditto.
1151 (sched_init): Call group_insns_for_macro_fusion.
1152 * target.def: Add two hooks: macro_fusion_p and
1153 macro_fusion_pair_p.
1155 2013-11-04 Kostya Serebryany <kcc@google.com>
1157 Update to match the changed asan API.
1158 * asan.c (asan_function_start): New function.
1159 (asan_emit_stack_protection): Update the string stored in the
1160 stack red zone to match new API. Store the PC of the current
1161 function in the red zone.
1162 (asan_global_struct): Update the __asan_global definition to match
1164 (asan_add_global): Ditto.
1165 * asan.h (asan_function_start): New prototype.
1166 * final.c (final_start_function): Call asan_function_start.
1167 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1
1170 2013-11-04 Wei Mi <wmi@google.com>
1172 * config/i386/i386-c.c (ix86_target_macros_internal): Separate
1173 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
1174 * config/i386/i386.c (ix86_option_override_internal): Ditto.
1175 (ix86_issue_rate): Ditto.
1176 (ix86_adjust_cost): Ditto.
1177 (ia32_multipass_dfa_lookahead): Ditto.
1178 (ix86_sched_init_global): Ditto.
1179 (get_builtin_code_for_version): Ditto.
1180 * config/i386/i386.h (enum target_cpu_default): Ditto.
1181 (enum processor_type): Ditto.
1182 * config/i386/x86-tune.def (DEF_TUNE): Ditto.
1184 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1186 PR rtl-optimization/58967
1187 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
1188 !lra_in_progress for mode sizes bigger word.
1190 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1192 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
1193 arguments to merge instruction for little endian.
1194 (vec_widen_umult_lo_v16qi): Likewise.
1195 (vec_widen_smult_hi_v16qi): Likewise.
1196 (vec_widen_smult_lo_v16qi): Likewise.
1197 (vec_widen_umult_hi_v8hi): Likewise.
1198 (vec_widen_umult_lo_v8hi): Likewise.
1199 (vec_widen_smult_hi_v8hi): Likewise.
1200 (vec_widen_smult_lo_v8hi): Likewise.
1202 2013-11-04 Ian Lance Taylor <iant@google.com>
1204 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
1205 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
1206 builtins that take pointers.
1207 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions if set.
1208 * lto-wrapper.c (merge_and_complain): Collect OPT_fnon_call_exceptions.
1209 (run_gcc): Pass -fnon-call-exceptions.
1211 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1213 * optabs.c (expand_vec_perm): Revert one incorrect line from
1216 PR tree-optimization/58978
1217 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
1218 use_stmt by single_imm_use directly. Only call single_imm_use
1221 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
1223 PR rtl-optimization/58968
1224 * lra-spills.c (return_regno_p): New function.
1225 (lra_final_code_change): Use it.
1227 2013-11-04 Joseph Myers <joseph@codesourcery.com>
1229 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document macros.
1230 * target.def (float_exceptions_rounding_supported_p): New hook.
1231 * targhooks.c (default_float_exceptions_rounding_supported_p): New
1233 * targhooks.h (default_float_exceptions_rounding_supported_p): Declare.
1234 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
1236 * doc/tm.texi: Regenerate.
1237 * config.gcc (powerpc*-*-linux*): Set extra_objs.
1238 * config/rs6000/rs6000-linux.c: New file.
1239 * config/rs6000/rs6000-protos.h
1240 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
1241 * config/rs6000/linux.h
1242 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
1243 * config/rs6000/linux64.h
1244 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
1245 * config/rs6000/t-linux (rs6000-linux.o): New rule.
1246 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
1248 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1250 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
1251 Replace the define_insn_and_split with a define_insn and two
1252 define_splits, with the split after reload re-permuting the source
1253 register to its original value.
1254 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
1255 (*vsx_le_perm_store_v8hi): Likewise.
1256 (*vsx_le_perm_store_v16qi): Likewise.
1258 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1260 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
1263 2013-11-04 Jakub Jelinek <jakub@redhat.com>
1265 PR tree-optimization/58946
1266 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
1267 bbs with bbinfo[idx].op != NULL before all blocks with
1268 bbinfo[idx].op == NULL.
1270 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
1272 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
1273 (avr_dump_double_int_hex): Likewise.
1274 (avr_log_vadump): Remove %D and %X handling.
1275 * config/avr/avr.c (avr_double_int_push_digit): Delete.
1276 (avr_map_op_t): Change map from double_int to unsigned int.
1277 (avr_map_op): Update accordingly.
1278 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
1279 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
1280 unsigned ints rather than double_ints.
1282 2013-11-03 Marek Polacek <polacek@redhat.com>
1284 Implement -fsanitize=vla-bound.
1285 * opts.c (common_handle_option): Handle vla-bound.
1286 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE): Define.
1287 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
1288 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
1290 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
1292 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
1295 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
1297 * config/i386/constraints.md (Ts, Tv): New address constrains.
1298 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
1299 constraint for address_no_seg_operand.
1300 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
1301 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
1302 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
1303 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
1304 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
1305 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
1306 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
1307 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
1308 constraint for vsib_address_operand.
1310 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
1312 * gcse.c (pre_delete): Remove references to regmove from comments.
1313 * recog.c: (validate_replace_rtx_1): Likewise.
1314 * config/rl78/rl78.c: Likewise.
1315 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
1316 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
1317 * common/config/mmix/mmix-common.c: Likewise.
1319 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
1321 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
1322 * gimplify.c (gimplify_compound_lval): Likewise.
1323 * graphite-clast-to-gimple.c (gloog): Likewise.
1324 * graphite-dependences.c (loop_is_parallel_p): Likewise.
1325 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
1326 (limit_scop); Likewise.
1327 (build_scops): Likewise.
1328 (dot_scop): Likewise.
1329 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
1330 (build_scop_drs): Likewise.
1331 (insert_stmts): Likewise.
1332 (insert_out_of_ssa_copy): Likewise.
1333 (remove_phi): Likewise.
1334 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
1335 * hw-doloop.c (discover_loop): Likewise.
1336 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
1337 * tree-dfa.c (dump_enumerated_decls): Likewise.
1338 * tree-if-conv.c (if_convertable_loop_p): Likewise.
1339 * tree-inline.c (tree_function_versioning): Likewise.
1340 * tree-loop-distribution.c (build_rdg): Likewise.
1341 (rdg_flag_vertex_and_dependent): Likewise.
1342 (distribute_loop): Likewise.
1343 * tree-parloops.c (loop_parallel_p): Likewise.
1344 (eliminate_local_variables): Likewise.
1345 (separate_decls_in_region): Likewise.
1346 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
1347 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
1348 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
1349 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
1350 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
1351 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
1352 (vectorizable_condition): Likewise.
1354 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
1356 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
1357 * configure: Regenerate.
1358 * config/i386/i386.md (*movdi_internal): Change
1359 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
1360 (*movdf_internal): Ditto.
1361 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
1362 * config/i386/sse.md (vec_concatv2di): Output interunit movq
1363 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
1365 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
1367 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
1370 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1372 Automated part of renaming of symtab_node_base to symtab_node.
1374 Patch autogenerated by rename_symtab.py from
1375 https://github.com/davidmalcolm/gcc-refactoring-scripts
1376 revision 58bb219cc090b2f4516a9297d868c245495ee622
1377 with ChangeLog entry fixed up by hand.
1379 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
1381 (cgraph_node_for_asm): Likewise.
1382 * cgraph.h (symtab_node_base): Likewise.
1383 (cgraph_node): Likewise.
1384 (varpool_node): Likewise.
1385 (is_a_helper <cgraph_node>::test): Likewise.
1386 (is_a_helper <varpool_node>::test): Likewise.
1387 (symtab_nodes): Likewise.
1388 (symtab_register_node): Likewise.
1389 (symtab_unregister_node): Likewise.
1390 (symtab_remove_node): Likewise.
1391 (symtab_get_node): Likewise.
1392 (symtab_node_for_asm): Likewise.
1393 (symtab_node_asm_name): Likewise.
1394 (symtab_node_name): Likewise.
1395 (symtab_insert_node_to_hashtable): Likewise.
1396 (symtab_add_to_same_comdat_group): Likewise.
1397 (symtab_dissolve_same_comdat_group_list): Likewise.
1398 (dump_symtab_node): Likewise.
1399 (debug_symtab_node): Likewise.
1400 (dump_symtab_base): Likewise.
1401 (verify_symtab_node): Likewise.
1402 (verify_symtab_base): Likewise.
1403 (symtab_used_from_object_file_p): Likewise.
1404 (symtab_alias_ultimate_target): Likewise.
1405 (symtab_resolve_alias): Likewise.
1406 (fixup_same_cpp_alias_visibility): Likewise.
1407 (symtab_for_node_and_aliases): Likewise.
1408 (symtab_nonoverwritable_alias): Likewise.
1409 (availability symtab_node_availability): Likewise.
1410 (symtab_semantically_equivalent_p): Likewise.
1411 (fixup_same_cpp_alias_visibility): Likewise.
1412 (symtab_prevail_in_asm_name_hash): Likewise.
1414 (varpool): Likewise.
1415 (varpool_first_variable): Likewise.
1416 (varpool_next_variable): Likewise.
1417 (varpool_first_static_initializer): Likewise.
1418 (varpool_next_static_initializer): Likewise.
1419 (varpool_first_defined_variable): Likewise.
1420 (varpool_next_defined_variable): Likewise.
1421 (cgraph_first_defined_function): Likewise.
1422 (cgraph_next_defined_function): Likewise.
1423 (cgraph_first_function): Likewise.
1424 (cgraph_next_function): Likewise.
1425 (cgraph_first_function_with_gimple_body): Likewise.
1426 (cgraph_next_function_with_gimple_body): Likewise.
1427 (symtab_alias_target): Likewise.
1428 (symtab_real_symbol_p): Likewise.
1429 (symtab_can_be_discarded): Likewise.
1430 * cgraphbuild.c (mark_address): Likewise.
1431 (mark_load): Likewise.
1432 (mark_store): Likewise.
1433 * cgraphunit.c (decide_is_symbol_needed): Likewise.
1435 (enqueue_node): Likewise.
1436 (referred_to_p): Likewise.
1437 (cgraph_process_same_body_aliases): Likewise.
1438 (analyze_functions): Likewise.
1439 (handle_alias_pairs): Likewise.
1440 (output_weakrefs): Likewise.
1441 (compile): Likewise.
1442 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1443 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1444 * ipa-prop.c (remove_described_reference): Likewise.
1445 (try_decrement_rdesc_refcount): Likewise.
1446 (ipa_edge_duplication_hook): Likewise.
1447 * ipa-ref.c (ipa_record_reference): Likewise.
1448 (ipa_maybe_record_reference): Likewise.
1449 (ipa_clone_ref): Likewise.
1450 (ipa_clone_references): Likewise.
1451 (ipa_clone_referring): Likewise.
1452 (ipa_find_reference): Likewise.
1453 (ipa_remove_stmt_references): Likewise.
1454 (ipa_clear_stmts_in_references): Likewise.
1455 * ipa-ref.h (symtab_node_base): Likewise.
1456 (ipa_ref): Likewise.
1457 (ipa_record_reference): Likewise.
1458 (ipa_maybe_record_reference): Likewise.
1459 (ipa_clone_references): Likewise.
1460 (ipa_clone_referring): Likewise.
1461 (ipa_clone_ref): Likewise.
1462 (ipa_find_reference): Likewise.
1463 (ipa_remove_stmt_references): Likewise.
1464 (ipa_clear_stmts_in_references): Likewise.
1465 * ipa-reference.c (ipa_reference_write_optimization_summary):
1467 * ipa.c (enqueue_node): Likewise.
1468 (process_references): Likewise.
1469 (walk_polymorphic_call_targets): Likewise.
1470 (symtab_remove_unreachable_nodes): Likewise.
1471 (address_taken_from_non_vtable_p): Likewise.
1472 (comdat_can_be_unshared_p_1): Likewise.
1473 (comdat_can_be_unshared_p): Likewise.
1474 (can_replace_by_local_alias): Likewise.
1475 (function_and_variable_visibility): Likewise.
1476 * is-a.h: Likewise (within example in comment).
1477 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
1478 (lto_symtab_encoder_encode): Likewise.
1479 (lto_symtab_encoder_delete_node): Likewise.
1480 (lto_symtab_encoder_in_partition_p): Likewise.
1481 (lto_set_symtab_encoder_in_partition): Likewise.
1482 (output_refs): Likewise.
1483 (compute_ltrans_boundary): Likewise.
1484 (output_symtab): Likewise.
1485 (input_node): Likewise.
1486 (input_ref): Likewise.
1487 (input_edge): Likewise.
1488 (input_cgraph_1): Likewise.
1489 (input_refs): Likewise.
1490 (output_cgraph_opt_summary): Likewise.
1491 (input_node_opt_summary): Likewise.
1492 (input_cgraph_opt_section): Likewise.
1493 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1495 * lto-streamer-out.c (lto_output): Likewise.
1496 (output_symbol_p): Likewise.
1497 (produce_symtab): Likewise.
1498 * lto-streamer.h (lto_encoder_entry): Likewise.
1499 (lto_free_function_in_decl_state_for_node): Likewise.
1500 (lto_symtab_encoder_encode): Likewise.
1501 (lto_symtab_encoder_delete_node): Likewise.
1502 (lto_symtab_encoder_in_partition_p): Likewise.
1503 (lto_set_symtab_encoder_in_partition): Likewise.
1504 (lto_symtab_encoder_lookup): Likewise.
1505 (lsei_node): Likewise.
1506 (lto_symtab_encoder_deref): Likewise.
1507 * symtab.c (symtab_hash): Likewise.
1508 (assembler_name_hash): Likewise.
1509 (symtab_nodes): Likewise.
1510 (hash_node): Likewise.
1511 (eq_node): Likewise.
1512 (hash_node_by_assembler_name): Likewise.
1513 (eq_assembler_name): Likewise.
1514 (insert_to_assembler_name_hash): Likewise.
1515 (unlink_from_assembler_name_hash): Likewise.
1516 (symtab_prevail_in_asm_name_hash): Likewise.
1517 (symtab_register_node): Likewise.
1518 (symtab_insert_node_to_hashtable): Likewise.
1519 (symtab_unregister_node): Likewise.
1520 (symtab_get_node): Likewise.
1521 (symtab_remove_node): Likewise.
1522 (symtab_initialize_asm_name_hash): Likewise.
1523 (symtab_node_for_asm): Likewise.
1524 (symtab_add_to_same_comdat_group): Likewise.
1525 (symtab_dissolve_same_comdat_group_list): Likewise.
1526 (symtab_node_asm_name): Likewise.
1527 (symtab_node_name): Likewise.
1528 (dump_symtab_base): Likewise.
1529 (dump_symtab_node): Likewise.
1530 (dump_symtab): Likewise.
1531 (debug_symtab_node): Likewise.
1532 (verify_symtab_base): Likewise.
1533 (verify_symtab_node): Likewise.
1534 (verify_symtab): Likewise.
1535 (symtab_used_from_object_file_p): Likewise.
1536 (symtab_node_availability): Likewise.
1537 (symtab_alias_ultimate_target): Likewise.
1538 (fixup_same_cpp_alias_visibility): Likewise.
1539 (symtab_resolve_alias): Likewise.
1540 (symtab_for_node_and_aliases): Likewise.
1541 (symtab_for_node_and_aliases): Likewise.
1542 (symtab_nonoverwritable_alias_1): Likewise.
1543 (symtab_nonoverwritable_alias): Likewise.
1544 (symtab_semantically_equivalent_p): Likewise.
1545 * value-prof.c (init_node_map): Likewise.
1546 * varasm.c (find_decl): Likewise.
1547 * varpool.c (varpool_node_for_asm): Likewise.
1548 (varpool_remove_unreferenced_decls): Likewise.
1550 2013-10-31 David Malcolm <dmalcolm@redhat.com>
1552 Manual part of renaming of symtab_node_base to symtab_node.
1554 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
1555 clashes with the preferred name for the base class.
1556 (const_symtab_node): Remove redundant typedef.
1558 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1560 * optabs.c (expand_vec_perm): Avoid vector mode punning
1561 SUBREGs in SET_DEST.
1562 * expmed.c (store_bit_field_1): Likewise.
1563 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
1564 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
1565 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
1566 * config/i386/i386.c (ix86_expand_vector_move_misalign,
1567 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
1568 ix86_expand_sse_unpack, ix86_expand_args_builtin,
1569 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
1570 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
1571 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
1572 expand_vec_perm_vpshufb2_vpermq,
1573 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
1574 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
1575 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
1576 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
1579 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
1581 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
1582 Use gcc_unreachable for unhandled modes. Do not check results of
1583 expand_simple_binop. If not expanded to target, move the result.
1585 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
1586 Shiva Chen <shiva0217@gmail.com>
1588 * config.gcc (nds32*-*-*): Add nds32 target.
1589 * config/nds32/nds32.c: New file.
1590 * config/nds32/nds32.h: New file.
1591 * config/nds32/nds32.md: New file.
1592 * config/nds32/constants.md: New file.
1593 * config/nds32/constraints.md: New file.
1594 * config/nds32/iterators.md: New file.
1595 * config/nds32/nds32-doubleword.md: New file.
1596 * config/nds32/nds32-intrinsic.md: New file.
1597 * config/nds32/nds32_intrinsic.h: New file.
1598 * config/nds32/nds32-modes.def: New file.
1599 * config/nds32/nds32-multiple.md: New file.
1600 * config/nds32/nds32.opt: New file.
1601 * config/nds32/nds32-opts.h: New file.
1602 * config/nds32/nds32-protos.h: New file.
1603 * config/nds32/nds32-peephole2.md: New file.
1604 * config/nds32/pipelines.md: New file.
1605 * config/nds32/predicates.md: New file.
1606 * config/nds32/t-mlibs: New file.
1607 * common/config/nds32: New directory and files.
1609 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
1610 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
1611 * doc/install.texi (Cross-Compiler-Specific Options): Document
1612 --with-nds32-lib for nds32 target.
1613 * doc/extend.texi (Function Attributes, Target Builtins): Document
1614 nds32 specific attributes.
1616 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
1618 * lra-constraints (process_alt_operands): Use the result
1619 elimination register for operand when matching constraints.
1621 2013-10-31 Jakub Jelinek <jakub@redhat.com>
1623 * tree-vrp.c (maybe_set_nonzero_bits): New function.
1624 (remove_range_assertions): Call it.
1626 * tree.c (tree_ctz): New function.
1627 * tree.h (tree_ctz): New prototype.
1628 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
1629 first argument from tree to const_tree.
1630 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
1631 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
1632 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
1634 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
1635 * tree-vect-loop.c (vect_analyze_loop_operations,
1636 vect_transform_loop): Don't force scalar loop for bound just because
1637 number of iterations is unknown, only do it if it is not known to be
1638 a multiple of vectorization_factor.
1639 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
1641 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
1642 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
1643 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
1644 (set_nonzero_bits, get_nonzero_bits): New prototypes.
1645 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
1646 a default def isn't partially constant.
1647 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
1648 is known to be partially zero.
1649 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
1650 to see if a default def isn't partially constant.
1651 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
1652 creation of a range, if VR_RANGE, try to improve nonzero_bits from
1654 (set_nonzero_bits, get_nonzero_bits): New functions.
1656 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
1657 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
1658 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
1659 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
1660 uses but in the condition and ASSERT_EXPR and the other successor of
1661 the predecessor bb is __builtin_unreachable (), set_range_info of the
1662 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
1664 2013-10-31 Martin Jambor <mjambor@suse.cz>
1666 PR rtl-optimization/58934
1668 2013-10-30 Martin Jambor <mjambor@suse.cz>
1669 PR rtl-optimization/10474
1670 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1672 (interesting_dest_for_shprep): New function.
1673 (split_live_ranges_for_shrink_wrap): Likewise.
1674 (ira): Calculate dominance info and df analysis. Call
1675 split_live_ranges_for_shrink_wrap.
1677 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
1678 Yury Gribov <y.gribov@samsung.com>
1681 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
1682 shadow pointer to avoid clobbering the main one.
1684 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
1686 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
1688 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1691 * ira-color.c (update_costs_from_copies): Add new parameter. Use
1692 it for calling update_costs_from_allocno.
1693 (assign_hard_reg): Call restore_costs_from_copies only for
1694 !retry_p. Pass new argument to update_costs_from_copies.
1695 (color_pass): Pass new argument to update_costs_from_copies.
1696 (ira_mark_allocation_change): Ditto.
1698 2013-10-30 Sharad Singhai <singhai@google.com>
1701 * opts.c (common_handle_option): Remove deprecated option
1702 -ftree-vectorizer-verbose.
1703 * doc/invoke.texi (Debugging Options): Ditto.
1704 * opts-global.c (handle_common_deferred_options): Ditto.
1705 (dump_remap_tree_vectorizer_verbose): Delete.
1706 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
1708 2013-10-30 DJ Delorie <dj@redhat.com>
1710 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
1711 that take no arguments.
1713 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
1716 * reload1.c (update_eliminables_and_spill): New function, broken
1718 (reload): Use it. Check for frame size change after frame size
1719 alignment, and call update_eliminables_and_spill first if continue-ing.
1721 2013-10-30 Cong Hou <congh@google.com>
1724 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
1725 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
1726 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
1728 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
1730 PR rtl-optimization/58369
1731 * reload1.c (compute_reload_subreg_offset): New function.
1732 (choose_reload_regs): Use it to pass endian-correct
1733 offset to subreg_regno_offset.
1735 2013-10-30 Tobias Burnus <burnus@net-b.de>
1738 * tree-cfg.c (replace_loop_annotate): Replace warning by
1741 2013-10-30 Jason Merrill <jason@redhat.com>
1743 * configure.ac (loose_warn): Add -Wno-format if
1744 --disable-build-format-warnings.
1746 2013-10-30 David Malcolm <dmalcolm@redhat.com>
1748 * cgraphunit.c (analyze_functions): Split symtab_node declarations
1749 onto multiple lines to make things easier for rename_symtab.py.
1751 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
1752 (symtab_semantically_equivalent_p): Likewise.
1754 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1757 * lra.c (check_rtl): Remove address check before LRA work.
1759 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
1761 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
1762 POINTER_PLUS_EXPR in the defining statement.
1764 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
1766 * regmove.c: Remove.
1767 * tree-pass.h (make_pass_regmove): Remove.
1768 * timevar.def (TV_REGMOVE): Remove.
1769 * passes.def (pass_regmove): Remove.
1770 * opts.c (default_options_table): Remove entry for regmove.
1771 * doc/passes.texi: Remove regmove pass description.
1772 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
1774 (-fdump-rtl-regmove): Ditto.
1775 * common.opt (foptimize-register-move, fregmove): Ignore.
1776 * Makefile.in (OBJS): Remove regmove.o.
1777 * regmove.c: Remove.
1778 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
1780 (struct ira_allocno) New member allocno_prefs.
1781 (ALLOCNO_PREFS): New macro.
1782 (ira_prefs, ira_prefs_num): New external vars.
1783 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
1784 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
1785 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
1787 (ira_add_allocno_copy_to_list): Remove prototype.
1788 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
1789 (ira_pref_iterator): New type.
1790 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
1791 (FOR_EACH_PREF): New macro.
1792 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
1793 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
1795 (ira_setup_alts): New function.
1796 (decrease_live_ranges_number): New function.
1797 (ira): Call the above function.
1798 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
1799 (ira_create_allocno): Initialize allocno prefs.
1800 (pref_pool, pref_vec): New static vars.
1801 (initiate_prefs, find_allocno_pref, ira_create_pref): New
1803 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
1804 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
1805 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
1806 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
1807 (ira_add_allocno_copy_to_list): Make static. Rename to
1808 add_allocno_copy_to_list.
1809 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
1810 swap_allocno_copy_ends_if_necessary.
1811 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
1812 ira_remove_allocno_prefs.
1813 (ira_flattening): Ditto.
1814 (ira_build): Call initiate_prefs, print_prefs.
1815 (ira_destroy): Call finish_prefs.
1816 * ira-color.c (struct update_cost_record): New.
1817 (struct allocno_color_data): Add new member update_cost_records.
1818 (update_cost_record_pool): New static var.
1819 (init_update_cost_records, get_update_cost_record): New functions.
1820 (free_update_cost_record_list, finish_update_cost_records): Ditto.
1821 (struct update_cost_queue_elem): Add member from.
1822 (initiate_cost_update): Call init_update_cost_records.
1823 (finish_cost_update): Call finish_update_cost_records.
1824 (queue_update_cost, get_next_update_cost): Add new param from.
1825 (Update_allocno_cost, update_costs_from_allocno): New functions.
1826 (update_costs_from_prefs): Ditto.
1827 (update_copy_costs): Rename to update_costs_from_copies.
1828 (restore_costs_from_copies): New function.
1829 (update_conflict_hard_regno_costs): Don't go back.
1830 (assign_hard_reg): Call restore_costs_from_copies. Add printing
1832 (pop_allocnos): Add priniting more debug info.
1833 (color_allocnos): Remove prefs for conflicting hard regs.
1834 Call update_costs_from_prefs.
1835 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
1836 (get_dup_num): Rename, modify, and move to ira.c
1837 (process_regs_for_copy): Add prefs.
1838 (add_insn_allocno_copies): Put src as first arg of
1839 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
1840 * ira-costs.c (record_reg_classes): Modify and move code into
1841 record_operands_costs.
1842 (find_costs_and_classes): Create prefs for the hard reg of small
1844 (process_bb_node_for_hard_reg_moves): Add prefs.
1846 2013-10-30 Richard Biener <rguenther@suse.de>
1849 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
1850 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
1852 (pre_and_rev_post_order_compute): ... which now wraps it.
1853 * graph.c (draw_cfg_nodes_no_loops): Use
1854 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
1857 2013-10-30 Christian Bruel <christian.bruel@st.com>
1859 * config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
1860 (sh_expand_cmpstr): Handle known align and schedule improvements.
1861 * config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
1862 * config/sh/sh.md (cmpstrnsi): New pattern.
1864 2013-10-30 Martin Jambor <mjambor@suse.cz>
1866 PR rtl-optimization/10474
1867 * ira.c (find_moveable_pseudos): Do not calculate dominance info
1869 (interesting_dest_for_shprep): New function.
1870 (split_live_ranges_for_shrink_wrap): Likewise.
1871 (ira): Calculate dominance info and df analysis. Call
1872 split_live_ranges_for_shrink_wrap.
1874 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
1877 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
1879 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1881 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
1882 * tree.h (POINTER_BOUNDS_P): New.
1883 (BOUNDED_TYPE_P): New.
1885 (pointer_bounds_type_node): New.
1886 * tree.c (build_common_tree_nodes): Initialize
1887 pointer_bounds_type_node.
1888 * gimple.h (gimple_call_get_nobnd_arg_index): New.
1889 (gimple_call_num_nobnd_args): New.
1890 (gimple_call_nobnd_arg): New.
1891 (gimple_return_retbnd): New.
1892 (gimple_return_set_retbnd): New
1893 * gimple.c (gimple_build_return): Increase number of ops
1894 for return statement.
1895 (gimple_call_get_nobnd_arg_index): New.
1896 * gimple-pretty-print.c (dump_gimple_return): Print second op.
1898 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1900 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
1901 with "chkp ctor" and "bnd_legacy" attributes.
1902 * gimplify.c (gimplify_init_constructor): Avoid infinite
1903 loop during gimplification of bounds initializer.
1905 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1907 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
1908 (handle_bnd_legacy): New.
1909 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
1910 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
1913 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
1915 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
1916 (BT_FN_PTR_CONST_PTR): New.
1917 (BT_FN_CONST_PTR_CONST_PTR): New.
1918 (BT_FN_PTR_CONST_PTR_SIZE): New.
1919 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
1920 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
1921 (BT_FN_VOID_CONST_PTR_SIZE): New.
1922 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
1923 * chkp-builtins.def: New.
1924 * builtins.def: include chkp-builtins.def.
1925 (DEF_CHKP_BUILTIN): New.
1926 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
1927 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
1928 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
1929 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
1930 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
1931 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
1932 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
1933 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
1934 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
1935 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
1936 * common.opt (fcheck-pointer-bounds): New.
1937 * toplev.c (process_options): Check Pointer Bounds Checker is
1939 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
1941 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
1943 * target.def (builtin_chkp_function): New.
1944 (chkp_bound_type): New.
1945 (chkp_bound_mode): New.
1946 (fn_abi_va_list_bounds_size): New.
1947 (load_bounds_for_arg): New.
1948 (store_bounds_for_arg): New.
1949 * targhooks.h (default_load_bounds_for_arg): New.
1950 (default_store_bounds_for_arg): New.
1951 (default_fn_abi_va_list_bounds_size): New.
1952 (default_chkp_bound_type): New.
1953 (default_chkp_bound_mode): New.
1954 (default_builtin_chkp_function): New.
1955 * targhooks.c (default_load_bounds_for_arg): New.
1956 (default_store_bounds_for_arg): New.
1957 (default_fn_abi_va_list_bounds_size): New.
1958 (default_chkp_bound_type): New.
1959 (default_chkp_bound_mode); New.
1960 (default_builtin_chkp_function): New.
1961 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
1962 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
1963 (TARGET_STORE_BOUNDS_FOR_ARG): New.
1964 (TARGET_BUILTIN_CHKP_FUNCTION): New.
1965 (TARGET_CHKP_BOUND_TYPE): New.
1966 (TARGET_CHKP_BOUND_MODE): New.
1967 * doc/tm.texi: Regenerated.
1968 * langhooks.h (lang_hooks): Add chkp_supported field.
1969 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
1970 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
1972 2013-10-29 Andrew Pinski <apinski@cavium.com>
1974 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
1975 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
1976 fails, combining the branches anyways.
1977 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
1978 increases the number of combinings.
1979 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
1981 2013-10-29 Mike Stump <mikestump@comcast.net>
1983 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
1984 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
1985 (make_vector_mode): Increase namebuf to 16.
1986 (emit_insn_modes_h): When processing BImode, don't
1987 also match partial int modes.
1988 (emit_class_narrowest_mode): Likewise.
1990 * config/bfin/bfin-modes.def: Add precision to PDI.
1991 * config/m32c/m32c-modes.def: Add precision to PSI.
1992 * config/msp430/msp430-modes.def: Add precision to PSI.
1993 * config/rs6000/rs6000-modes.def: Add precision to PTI.
1994 * config/sh/sh-modes.def: Add precision to PSI and PDI.
1996 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
1999 * config/sh/sh.md (*addc): Rename existing variations to ...
2000 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
2001 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
2002 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
2003 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
2005 2013-10-29 Teresa Johnson <tejohnson@google.com>
2008 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
2009 insanities when updating probabilities.
2011 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2013 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
2014 prettyprinter for cgraph_node to reflect the conversion of the
2015 symtable types to a C++ class hierarchy: it now *is* a
2016 symtab_node_base, rather than having one (named "symbol").
2018 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
2020 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
2021 __cilkrts_pop_frame. If matched, then return true for built-in
2023 (expand_builtin): Added BUILT_IN_CILK_DETACH and
2024 BUILT_IN_CILK_POP_FRAME case.
2025 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
2026 (lhs_cilk_detect_spawn): Likewise.
2027 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
2028 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
2029 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
2030 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
2031 (LANG_HOOKS_CILKPLUS): Likewise.
2032 * tree.h (CILK_SPAWN_FN): Likewise.
2033 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
2034 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
2035 (OBJS): Added cilk-common.o.
2036 (BUILTINS_DEF): Added cilk-builtins.def.
2037 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
2038 (lhd_cilk_detect_spawn): Likewise.
2039 * langhooks.h (lang_hooks_for_cilkplus): New struct.
2040 (struct lang_hooks): Added new field called "cilkplus."
2041 * cilk-common.c: New file.
2043 * cilk-builtins.def: Likewise.
2044 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
2045 "__cilk" macro and set it to 200.
2046 * function.h (struct function::cilk_frame_decl): New field.
2047 (struct function::is_cilk_function): Likewise.
2048 (struct function::calls_cilk_spawn): Likewise.
2049 * gimplify.c (gimplify_call_expr): Added a check if the function call
2050 being gimplified is a spawn detach point. If so, then add pop_frame
2051 and detach function calls.
2052 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
2053 for gimplifying _Cilk_spawn and _Cilk_sync statements.
2054 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
2055 function. If so, added a _Cilk_sync and gimplified it.
2056 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
2057 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
2058 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
2060 (can_inline_edge_p): Prevent inling of spawnee function.
2061 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
2062 for functions that use Cilk keywords.
2063 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
2064 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
2065 CILK_SYNC_STMT cases.
2066 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
2068 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
2069 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
2070 * passes.texi (Cilk Keywords): New section that describes the compiler
2071 code changes for handling Cilk Keywords.
2073 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2075 Patch autogenerated by refactor_symtab.py from
2076 https://github.com/davidmalcolm/gcc-refactoring-scripts
2077 revision 58bb219cc090b2f4516a9297d868c245495ee622
2079 * asan.c (asan_finish_file): Update for conversion of symtab types to
2080 a true class hierarchy.
2081 * cfgexpand.c (estimated_stack_frame_size): Likewise.
2082 * cgraph.c (cgraph_get_body): Likewise.
2083 (cgraph_get_create_real_symbol_node): Likewise.
2084 (verify_cgraph_node): Likewise.
2085 (verify_edge_corresponds_to_fndecl): Likewise.
2086 (verify_edge_count_and_frequency): Likewise.
2087 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
2088 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
2089 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
2090 (cgraph_node_cannot_return): Likewise.
2091 (cgraph_set_pure_flag_1): Likewise.
2092 (cgraph_set_const_flag_1): Likewise.
2093 (cgraph_set_nothrow_flag_1): Likewise.
2094 (cgraph_make_node_local_1): Likewise.
2095 (cgraph_for_node_and_aliases): Likewise.
2096 (cgraph_for_node_thunks_and_aliases): Likewise.
2097 (cgraph_node_can_be_local_p): Likewise.
2098 (cgraph_node_cannot_be_local_p_1): Likewise.
2099 (cgraph_function_body_availability): Likewise.
2100 (dump_cgraph_node): Likewise.
2101 (cgraph_rtl_info): Likewise.
2102 (cgraph_mark_address_taken_node): Likewise.
2103 (cgraph_remove_node): Likewise.
2104 (cgraph_release_function_body): Likewise.
2105 (cgraph_update_edges_for_call_stmt_node): Likewise.
2106 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
2107 (cgraph_make_edge_direct): Likewise.
2108 (cgraph_resolve_speculation): Likewise.
2109 (cgraph_speculative_call_info): Likewise.
2110 (cgraph_turn_edge_to_speculative): Likewise.
2111 (cgraph_create_edge_1): Likewise.
2112 (cgraph_set_call_stmt): Likewise.
2113 (cgraph_node_for_asm): Likewise.
2114 (cgraph_add_thunk): Likewise.
2115 (cgraph_same_body_alias): Likewise.
2116 (cgraph_create_function_alias): Likewise.
2117 (cgraph_create_node): Likewise.
2118 (cgraph_create_empty_node): Likewise.
2119 (record_function_versions): Likewise.
2120 (used_from_object_file_p): Likewise.
2121 * cgraph.h (symtab_can_be_discarded): Likewise.
2122 (symtab_real_symbol_p): Likewise.
2123 (cgraph_mark_force_output_node): Likewise.
2124 (cgraph_edge_recursive_p): Likewise.
2125 (symtab_alias_target): Likewise.
2126 (varpool_all_refs_explicit_p): Likewise.
2127 (varpool_can_remove_if_no_refs): Likewise.
2128 (cgraph_only_called_directly_or_aliased_p): Likewise.
2129 (cgraph_next_function_with_gimple_body): Likewise.
2130 (cgraph_first_function_with_gimple_body): Likewise.
2131 (cgraph_function_with_gimple_body_p): Likewise.
2132 (cgraph_next_function): Likewise.
2133 (cgraph_first_function): Likewise.
2134 (cgraph_next_defined_function): Likewise.
2135 (cgraph_first_defined_function): Likewise.
2136 (varpool_next_defined_variable): Likewise.
2137 (varpool_first_defined_variable): Likewise.
2138 (varpool_next_static_initializer): Likewise.
2139 (varpool_first_static_initializer): Likewise.
2140 (varpool_next_variable): Likewise.
2141 (varpool_first_variable): Likewise.
2142 (varpool_node_name): Likewise.
2143 (varpool): Likewise.
2145 (is_a_helper <varpool_node>::test): Likewise.
2146 (is_a_helper <cgraph_node>::test): Likewise.
2147 (varpool_variable_node): Likewise.
2148 (cgraph_function_or_thunk_node): Likewise.
2149 (varpool_alias_target): Likewise.
2150 (cgraph_alias_target): Likewise.
2151 (cgraph_node_name): Likewise.
2152 (varpool_node_asm_name): Likewise.
2153 (cgraph_node_asm_name): Likewise.
2154 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
2155 (cgraph_rebuild_references): Likewise.
2156 (rebuild_cgraph_edges): Likewise.
2157 (record_eh_tables): Likewise.
2158 (build_cgraph_edges): Likewise.
2159 (mark_store): Likewise.
2160 (mark_load): Likewise.
2161 (mark_address): Likewise.
2162 (record_type_list): Likewise.
2163 (record_reference): Likewise.
2164 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
2165 (cgraph_materialize_clone): Likewise.
2166 (cgraph_function_versioning): Likewise.
2167 (cgraph_copy_node_for_versioning): Likewise.
2168 (update_call_expr): Likewise.
2169 (cgraph_find_replacement_node): Likewise.
2170 (cgraph_create_virtual_clone): Likewise.
2171 (cgraph_clone_node): Likewise.
2172 * cgraphunit.c (compile): Likewise.
2173 (output_weakrefs): Likewise.
2174 (output_in_order): Likewise.
2175 (expand_function): Likewise.
2176 (assemble_thunks_and_aliases): Likewise.
2177 (expand_thunk): Likewise.
2178 (mark_functions_to_output): Likewise.
2179 (handle_alias_pairs): Likewise.
2180 (analyze_functions): Likewise.
2181 (walk_polymorphic_call_targets): Likewise.
2182 (varpool_finalize_decl): Likewise.
2183 (process_function_and_variable_attributes): Likewise.
2184 (cgraph_process_same_body_aliases): Likewise.
2185 (analyze_function): Likewise.
2186 (cgraph_add_new_function): Likewise.
2187 (cgraph_finalize_function): Likewise.
2188 (referred_to_p): Likewise.
2189 (cgraph_reset_node): Likewise.
2190 (cgraph_process_new_functions): Likewise.
2191 (enqueue_node): Likewise.
2192 (decide_is_symbol_needed): Likewise.
2193 * coverage.c (coverage_compute_profile_id): Likewise.
2194 * dbxout.c (dbxout_expand_expr): Likewise.
2195 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
2196 (reference_to_unused): Likewise.
2197 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
2198 * gimplify.c (unvisit_body): Likewise.
2199 (unshare_body): Likewise.
2200 * ipa-cp.c (ipcp_generate_summary): Likewise.
2201 (ipcp_decision_stage): Likewise.
2202 (identify_dead_nodes): Likewise.
2203 (decide_whether_version_node): Likewise.
2204 (decide_about_value): Likewise.
2205 (perhaps_add_new_callers): Likewise.
2206 (create_specialized_node): Likewise.
2207 (update_profiling_info): Likewise.
2208 (ipcp_propagate_stage): Likewise.
2209 (estimate_local_effects): Likewise.
2210 (good_cloning_opportunity_p): Likewise.
2211 (devirtualization_time_bonus): Likewise.
2212 (propagate_constants_accross_call): Likewise.
2213 (initialize_node_lattices): Likewise.
2214 (ipcp_cloning_candidate_p): Likewise.
2215 (determine_versionability): Likewise.
2216 (print_all_lattices): Likewise.
2217 (print_lattice): Likewise.
2218 (ipcp_discover_new_direct_edges): Likewise.
2219 * ipa-devirt.c (ipa_devirt): Likewise.
2220 (likely_target_p): Likewise.
2221 (update_type_inheritance_graph): Likewise.
2222 (possible_polymorphic_call_target_p): Likewise.
2223 (dump_possible_polymorphic_call_targets): Likewise.
2224 (devirt_variable_node_removal_hook): Likewise.
2225 (record_binfo): Likewise.
2226 (maybe_record_node): Likewise.
2227 (build_type_inheritance_graph): Likewise.
2228 * ipa-inline-analysis.c (inline_write_summary): Likewise.
2229 (inline_generate_summary): Likewise.
2230 (inline_analyze_function): Likewise.
2231 (do_estimate_growth): Likewise.
2232 (simple_edge_hints): Likewise.
2233 (estimate_node_size_and_time): Likewise.
2234 (estimate_edge_devirt_benefit): Likewise.
2235 (compute_inline_parameters): Likewise.
2236 (estimate_function_body_sizes): Likewise.
2237 (compute_bb_predicates): Likewise.
2238 (initialize_inline_failed): Likewise.
2239 (dump_inline_summary): Likewise.
2240 (dump_inline_edge_summary): Likewise.
2241 * ipa-inline-transform.c (inline_transform): Likewise.
2242 (preserve_function_body_p): Likewise.
2243 (save_inline_function_body): Likewise.
2244 (inline_call): Likewise.
2245 (clone_inlined_nodes): Likewise.
2246 (can_remove_node_now_p): Likewise.
2247 (can_remove_node_now_p_1): Likewise.
2248 * ipa-inline.c (early_inliner): Likewise.
2249 (early_inline_small_functions): Likewise.
2250 (inline_always_inline_functions): Likewise.
2251 (ipa_inline): Likewise.
2252 (flatten_function): Likewise.
2253 (inline_small_functions): Likewise.
2254 (speculation_useful_p): Likewise.
2255 (recursive_inlining): Likewise.
2256 (update_caller_keys): Likewise.
2257 (reset_edge_caches): Likewise.
2258 (update_edge_key): Likewise.
2259 (edge_badness): Likewise.
2260 (relative_time_benefit): Likewise.
2261 (want_inline_self_recursive_call_p): Likewise.
2262 (want_inline_small_function_p): Likewise.
2263 (want_early_inline_function_p): Likewise.
2264 (num_calls): Likewise.
2265 (can_early_inline_edge_p): Likewise.
2266 (can_inline_edge_p): Likewise.
2267 (report_inline_failed_reason): Likewise.
2268 * ipa-profile.c (ipa_profile): Likewise.
2269 (ipa_propagate_frequency): Likewise.
2270 (ipa_propagate_frequency_1): Likewise.
2271 (ipa_profile_generate_summary): Likewise.
2272 * ipa-prop.c (ipcp_transform_function): Likewise.
2273 (read_replacements_section): Likewise.
2274 (ipa_prop_read_section): Likewise.
2275 (ipa_modify_call_arguments): Likewise.
2276 (ipa_print_node_params): Likewise.
2277 (propagate_controlled_uses): Likewise.
2278 (update_indirect_edges_after_inlining): Likewise.
2279 (remove_described_reference): Likewise.
2280 (ipa_make_edge_direct_to_target): Likewise.
2281 (ipa_analyze_node): Likewise.
2282 (ipa_analyze_params_uses): Likewise.
2283 (ipa_compute_jump_functions): Likewise.
2284 (ipa_get_callee_param_type): Likewise.
2285 (ipa_print_node_jump_functions): Likewise.
2286 (ipa_initialize_node_params): Likewise.
2287 (ipa_populate_param_decls): Likewise.
2288 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
2289 (write_agg_replacement_chain): Likewise.
2290 (ipa_write_node_info): Likewise.
2291 (ipa_edge_duplication_hook): Likewise.
2292 (try_decrement_rdesc_refcount): Likewise.
2293 * ipa-pure-const.c (propagate_nothrow): Likewise.
2294 (propagate_pure_const): Likewise.
2295 (pure_const_read_summary): Likewise.
2296 (pure_const_write_summary): Likewise.
2297 (analyze_function): Likewise.
2298 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
2299 (ipa_ref_referring_ref_list): Likewise.
2300 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
2301 (ipa_remove_stmt_references): Likewise.
2302 (ipa_find_reference): Likewise.
2303 (ipa_dump_referring): Likewise.
2304 (ipa_dump_references): Likewise.
2305 (ipa_record_reference): Likewise.
2306 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
2307 (ipa_reference_write_optimization_summary): Likewise.
2308 (write_node_summary_p): Likewise.
2309 (propagate): Likewise.
2310 (read_write_all_from_decl): Likewise.
2311 (generate_summary): Likewise.
2312 (analyze_function): Likewise.
2313 (propagate_bits): Likewise.
2314 (ipa_reference_get_not_written_global): Likewise.
2315 (ipa_reference_get_not_read_global): Likewise.
2316 * ipa-split.c (execute_split_functions): Likewise.
2317 (split_function): Likewise.
2318 * ipa-utils.c (ipa_merge_profiles): Likewise.
2319 (dump_cgraph_node_set): Likewise.
2320 (ipa_reverse_postorder): Likewise.
2321 (ipa_edge_within_scc): Likewise.
2322 (ipa_get_nodes_in_cycle): Likewise.
2323 (ipa_free_postorder_info): Likewise.
2324 (ipa_reduced_postorder): Likewise.
2325 (searchc): Likewise.
2326 (recursive_call_p): Likewise.
2327 * ipa.c (ipa_cdtor_merge): Likewise.
2328 (record_cdtor_fn): Likewise.
2329 (function_and_variable_visibility): Likewise.
2330 (varpool_externally_visible_p): Likewise.
2331 (cgraph_externally_visible_p): Likewise.
2332 (comdat_can_be_unshared_p): Likewise.
2333 (comdat_can_be_unshared_p_1): Likewise.
2334 (address_taken_from_non_vtable_p): Likewise.
2335 (ipa_discover_readonly_nonaddressable_vars): Likewise.
2336 (symtab_remove_unreachable_nodes): Likewise.
2337 (walk_polymorphic_call_targets): Likewise.
2338 (process_references): Likewise.
2339 (enqueue_node): Likewise.
2340 (has_addr_references_p): Likewise.
2341 (cgraph_non_local_node_p_1): Likewise.
2342 * is-a.h (varpool_analyze_node): Likewise.
2343 * lto-cgraph.c (input_symtab): Likewise.
2344 (merge_profile_summaries): Likewise.
2345 (input_cgraph_1): Likewise.
2346 (input_edge): Likewise.
2347 (input_varpool_node): Likewise.
2348 (input_node): Likewise.
2349 (input_overwrite_node): Likewise.
2350 (compute_ltrans_boundary): Likewise.
2351 (output_refs): Likewise.
2352 (lto_output_varpool_node): Likewise.
2353 (lto_output_node): Likewise.
2354 (reachable_from_other_partition_p): Likewise.
2355 (referenced_from_other_partition_p): Likewise.
2356 (lto_output_edge): Likewise.
2357 (output_node_opt_summary): Likewise.
2358 (add_node_to): Likewise.
2359 (reachable_from_this_partition_p): Likewise.
2360 (lto_set_symtab_encoder_in_partition): Likewise.
2361 (lto_symtab_encoder_in_partition_p): Likewise.
2362 (lto_set_symtab_encoder_encode_initializer): Likewise.
2363 (lto_symtab_encoder_encode_initializer_p): Likewise.
2364 (lto_set_symtab_encoder_encode_body): Likewise.
2365 (lto_symtab_encoder_encode_body_p): Likewise.
2366 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
2368 * lto-streamer-in.c (lto_read_body): Likewise.
2369 (fixup_call_stmt_edges): Likewise.
2370 (fixup_call_stmt_edges_1): Likewise.
2371 * lto-streamer-out.c (produce_symtab): Likewise.
2372 (output_symbol_p): Likewise.
2373 (write_symbol): Likewise.
2374 (lto_output): Likewise.
2375 (copy_function): Likewise.
2376 (output_function): Likewise.
2377 * passes.c (function_called_by_processed_nodes_p): Likewise.
2378 (ipa_write_optimization_summaries): Likewise.
2379 (ipa_write_summaries): Likewise.
2380 (do_per_function_toporder): Likewise.
2381 (do_per_function): Likewise.
2382 (dump_passes): Likewise.
2383 * symtab.c (symtab_semantically_equivalent_p): Likewise.
2384 (symtab_nonoverwritable_alias): Likewise.
2385 (symtab_nonoverwritable_alias_1): Likewise.
2386 (symtab_for_node_and_aliases): Likewise.
2387 (symtab_resolve_alias): Likewise.
2388 (fixup_same_cpp_alias_visibility): Likewise.
2389 (symtab_alias_ultimate_target): Likewise.
2390 (symtab_used_from_object_file_p): Likewise.
2391 (verify_symtab_base): Likewise.
2392 (dump_symtab_base): Likewise.
2393 (symtab_node_name): Likewise.
2394 (symtab_node_asm_name): Likewise.
2395 (symtab_dissolve_same_comdat_group_list): Likewise.
2396 (symtab_add_to_same_comdat_group): Likewise.
2397 (symtab_unregister_node): Likewise.
2398 (symtab_insert_node_to_hashtable): Likewise.
2399 (symtab_register_node): Likewise.
2400 (unlink_from_assembler_name_hash): Likewise.
2401 (insert_to_assembler_name_hash): Likewise.
2402 (eq_assembler_name): Likewise.
2403 (hash_node_by_assembler_name): Likewise.
2404 (eq_node): Likewise.
2405 (hash_node): Likewise.
2406 * toplev.c (wrapup_global_declaration_2): Likewise.
2407 * trans-mem.c (ipa_tm_execute): Likewise.
2408 (ipa_tm_transform_clone): Likewise.
2409 (ipa_tm_transform_transaction): Likewise.
2410 (ipa_tm_transform_calls_redirect): Likewise.
2411 (ipa_tm_insert_gettmclone_call): Likewise.
2412 (ipa_tm_insert_irr_call): Likewise.
2413 (ipa_tm_create_version): Likewise.
2414 (ipa_tm_create_version_alias): Likewise.
2415 (ipa_tm_mark_forced_by_abi_node): Likewise.
2416 (ipa_tm_mark_force_output_node): Likewise.
2417 (ipa_tm_diagnose_tm_safe): Likewise.
2418 (ipa_tm_mayenterirr_function): Likewise.
2419 (ipa_tm_scan_irr_function): Likewise.
2420 (ipa_tm_note_irrevocable): Likewise.
2421 (ipa_tm_scan_calls_clone): Likewise.
2422 (get_cg_data): Likewise.
2423 * tree-eh.c (tree_could_trap_p): Likewise.
2424 * tree-emutls.c (ipa_lower_emutls): Likewise.
2425 (create_emultls_var): Likewise.
2426 (lower_emutls_function_body): Likewise.
2427 (gen_emutls_addr): Likewise.
2428 (emutls_decl): Likewise.
2429 (new_emutls_decl): Likewise.
2430 * tree-inline.c (tree_function_versioning): Likewise.
2431 (optimize_inline_calls): Likewise.
2432 (expand_call_inline): Likewise.
2433 (estimate_num_insns): Likewise.
2434 (copy_bb): Likewise.
2435 (delete_unreachable_blocks_update_callgraph): Likewise.
2436 * tree-nested.c (gimplify_all_functions): Likewise.
2437 (create_nesting_tree): Likewise.
2438 (check_for_nested_with_variably_modified): Likewise.
2439 * tree-pretty-print.c (dump_function_header): Likewise.
2440 * tree-profile.c (tree_profiling): Likewise.
2441 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
2442 (modify_function): Likewise.
2443 (convert_callers): Likewise.
2444 (convert_callers_for_node): Likewise.
2445 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
2446 (associate_varinfo_to_alias): Likewise.
2447 (create_variable_info_for): Likewise.
2448 (get_constraint_for_ssa_var): Likewise.
2449 * tree-vectorizer.c (increase_alignment): Likewise.
2450 * tree.c (find_decls_types_in_var): Likewise.
2451 (find_decls_types_in_node): Likewise.
2452 (free_lang_data_in_decl): Likewise.
2453 * value-prof.c (gimple_ic_transform): Likewise.
2454 (gimple_ic): Likewise.
2455 (check_ic_target): Likewise.
2456 (init_node_map): Likewise.
2457 * varasm.c (decl_binds_to_current_def_p): Likewise.
2458 (default_binds_local_p_1): Likewise.
2459 (dump_tm_clone_pairs): Likewise.
2460 (assemble_alias): Likewise.
2461 (find_decl): Likewise.
2462 (mark_decl_referenced): Likewise.
2463 * varpool.c (varpool_for_node_and_aliases): Likewise.
2464 (varpool_extra_name_alias): Likewise.
2465 (varpool_create_variable_alias): Likewise.
2466 (add_new_static_var): Likewise.
2467 (varpool_finalize_named_section_flags): Likewise.
2468 (varpool_remove_unreferenced_decls): Likewise.
2469 (enqueue_node): Likewise.
2470 (varpool_assemble_decl): Likewise.
2471 (assemble_aliases): Likewise.
2472 (varpool_analyze_node): Likewise.
2473 (cgraph_variable_initializer_availability): Likewise.
2474 (varpool_add_new_variable): Likewise.
2475 (ctor_for_folding): Likewise.
2476 (dump_varpool_node): Likewise.
2477 (varpool_remove_initializer): Likewise.
2478 (varpool_remove_node): Likewise.
2479 (varpool_node_for_decl): Likewise.
2480 (varpool_create_empty_node): Likewise.
2481 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
2482 (ix86_get_function_versions_dispatcher): Likewise.
2484 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2486 * cgraph.h (symtab_node_base): Convert to a class;
2487 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
2488 chain_next/prev from symtab_node_def.
2489 (cgraph_node): Inherit from symtab_node; add GTY option
2490 tag ("SYMTAB_FUNCTION").
2491 (varpool_node): Inherit from symtab_node; add GTY option
2492 tag ("SYMTAB_VARIABLE").
2493 (symtab_node_def): Remove.
2494 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
2495 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
2496 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
2497 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
2499 * ipa-ref.h (symtab_node_def): Drop.
2500 (symtab_node): Change underlying type from symtab_node_def to
2502 (const_symtab_node): Likwise.
2504 * is-a.h: Update examples in comment.
2506 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
2507 (assembler_name_hash): Likewise.
2509 2013-10-29 Martin Liska <marxin.liska@gmail.com>
2511 * doc/tree-ssa.texi (gimple_phi_result): Document.
2512 (gimple_phi_num_args, gimple_phi_arg): Likewise.
2513 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
2514 (PHI_RESULT, PHI_NUM_ARGS): Remove.
2515 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
2517 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2519 * expr.h: Revert change and include tree-core.h.
2520 * rtl.h: Revert change and don't include tree-core.h.
2522 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2524 * config/darwin.c: Include gimple.h.
2525 * config/i386/winnt.c: Likewise.
2527 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
2529 PR tree-optimization/19831
2530 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
2532 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
2534 * tree-outof-ssa.h: Remove include files.
2535 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
2537 * tree-ssa-coalesce.c: Likewise.
2538 * cfgexpand.c: Likewise.
2539 * tree-ssa-ter.c: Likewise.
2540 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
2541 * lto-streamer.h: Likewise.
2542 * cgraphbuild.c: Add gimple.h to include list.
2543 * data-streamer-in.c: Likewise.
2544 * ipa-cp.c: Likewise.
2545 * tree-streamer.c: Likewise.
2546 * lto-compress.c: Likewise.
2547 * ipa-reference.c: Likewise.
2548 * data-streamer-out.c: Likewise.
2549 * lto-cgraph.c: Likewise.
2550 * cgraphclones.c: Likewise.
2551 * ipa-utils.c: Likewise.
2552 * data-streamer.c: Likewise.
2553 * ipa-split.c: Likewise.
2554 * lto-section-in.c: Likewise.
2555 * tree-streamer-out.c: Likewise.
2556 * ipa-prop.c: Likewise.
2557 * tree-streamer-in.c: Likewise.
2558 * symtab.c: Likewise.
2559 * opts-global.c: Likewise.
2560 * lto-opts.c: Likewise.
2561 * lto-section-out.c: Likewise.
2562 * lto-streamer.c: Likewise.
2563 * rtl.h: Add tree-core.h to include list.
2564 * expr.h: Remove tree-core.h from include list.
2565 * gimple.h: Likewise.
2566 * ipa-utils.h: Likewise.
2567 * streamer-hooks.h: Likewise.
2568 * streamer-hooks.c: Include input.h.
2570 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2572 * config/arm/arm.c (cortexa7_extra_costs): New table.
2573 (arm_cortex_a7_tune): New.
2574 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
2576 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
2578 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
2580 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2582 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
2583 to use autogenerated markers for a class-hierarchy, every class
2584 must have a GTY marker.
2585 * gengtype.h (struct type): Add linked list of subclasses to
2586 the "s" member of the union.
2587 (add_subclass): New decl.
2588 * gengtype-state.c (read_state_struct_type): Set up subclass
2590 * gengtype.c (get_ultimate_base_class): New.
2591 (add_subclass): New.
2592 (new_structure): Set up subclass linked list.
2593 (set_gc_used_type): Propagate usage information to subclasses.
2594 (output_mangled_typename): Use get_ultimate_base_class.
2595 (walk_subclasses): Use the subclass linked list, avoiding an
2596 O(N^2) when writing out all types.
2597 (walk_type): Issue an error if the base class is missing a tag,
2598 rather than generating bogus C code. Add a gcc_unreachable
2599 default case, in case people omit tags from concrete subclasses,
2600 or get the values wrong.
2601 (write_func_for_structure): Issue an error for subclasses for
2602 which the base doesn't have a "desc", since otherwise the
2603 autogenerated routines for the base would silently fail to visit
2604 any subclass fields.
2605 (write_root): Use get_ultimate_base_class, tweaking constness of
2606 tp to match that function's signature.
2608 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2610 * doc/gty.texi (GTY Options): Add note about inheritance to
2611 description of desc and tag.
2612 (Inheritance and GTY): New.
2614 2013-10-29 David Malcolm <dmalcolm@redhat.com>
2616 * gengtype-parse.c (opts_have): Drop "static" so that
2617 we can use this from gengtype.c.
2618 * gengtype.c (set_gc_used_type): Mark any base class as used;
2619 update field traversal to visit inherited fields.
2620 (output_mangled_typename): Convert references to classes within
2621 an inheritance hierarchy to reference the ultimate base class,
2622 since only it will have gt_ functions.
2623 (get_string_option): New.
2624 (walk_subclasses): New.
2625 (walk_type): Treat GTY structs that have a "desc" as being the
2626 root of an inheritance hierarchy. Generate a switch on it
2627 within the marking function which walks all subclasses, adding
2628 cases for them via walk_subclasses. For subclasses, visit all
2629 fields of the type (including inherited ones).
2630 (write_func_for_structure): Don't write fns for subclasses, only
2631 for the ultimate base class within an inheritance hierarchy.
2632 Subclasses-marking will be handled by the base class marking functions.
2633 (write_types): Likewise.
2634 (write_local_func_for_structure): Likewise.
2635 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
2636 a "tag" option (and are thus concrete subclasses).
2637 (write_root): Use the marker function for the ultimate base class.
2638 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
2639 (opts_have): Add declaration.
2641 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
2643 * lra-spills.c (lra_final_code_change): Remove useless move insns
2644 originated from moves of pseudos.
2646 2013-10-28 Jeff Law <law@redhat.com>
2648 * lower-subreg.c (resolve_simple_move): Fix comment typo.
2650 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
2652 * df-scan.c (df_collection_rec): Adjust.
2653 (copy_defs): New constant.
2654 (copy_uses): Likewise.
2655 (copy_eq_uses): Likewise.
2656 (copy_mw): Likewise.
2657 (copy_all): Likewise.
2658 (df_insn_rescan): Adjust.
2659 (df_notes_rescan): Likewise.
2660 (df_swap_refs): Likewise.
2661 (df_sort_and_compress_refs): Likewise.
2662 (df_sort_and_compress_mws): Likewise.
2663 (df_install_refs): Likewise.
2664 (df_install_mws): Likewise.
2665 (df_refs_add_to_chains): Add flags parameter controlling which vectors
2667 (df_bb_refs_record): Adjust.
2668 (df_record_entry_block_defs): Likewise.
2669 (df_record_exit_block_defs): Likewise.
2670 (df_refs_verify): Likewise.
2671 (df_mws_verify): Likewise.
2672 (df_insn_refs_verify): Likewise.
2673 (df_bb_verify): Likewise.
2674 * ipa-pure-const.c (finish_state): Remove.
2675 (propagate): Adjust.
2676 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
2677 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
2678 var-tracking.c: Adjust.
2679 * vec.c (stack_vecs): Remove.
2680 (register_stack_vec): Likewise.
2681 (stack_vec_register_index): Likewise.
2682 (unregister_stack_vec): Likewise.
2683 * vec.h (struct va_stack): Remove.
2684 (struct vec<T, A, vl_ptr>): Specialize as
2685 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
2686 allocation strategy compatable with the vl_ptr layout.
2687 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
2688 specialization anyway.
2689 (class stack_vec): New class.
2690 (vec_stack_alloc): Remove.
2691 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
2693 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2694 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2695 Sergey Lega <sergey.s.lega@intel.com>
2696 Anna Tikhonova <anna.tikhonova@intel.com>
2697 Ilya Tocar <ilya.tocar@intel.com>
2698 Andrey Turetskiy <andrey.turetskiy@intel.com>
2699 Ilya Verbin <ilya.verbin@intel.com>
2700 Kirill Yukhin <kirill.yukhin@intel.com>
2701 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2703 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
2704 (*prefetch_avx512pf_<mode>): New.
2705 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
2706 (avx512f_maskcmp<mode>3): Ditto.
2707 (vashrv16si3): Ditto.
2709 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
2710 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
2711 Sergey Lega <sergey.s.lega@intel.com>
2712 Anna Tikhonova <anna.tikhonova@intel.com>
2713 Ilya Tocar <ilya.tocar@intel.com>
2714 Andrey Turetskiy <andrey.turetskiy@intel.com>
2715 Ilya Verbin <ilya.verbin@intel.com>
2716 Kirill Yukhin <kirill.yukhin@intel.com>
2717 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
2719 * config/i386/i386.md (any_truncate): New.
2720 (trunsuffix): Ditto.
2721 * config/i386/predicates.md (const_8_to_9_operand): New.
2722 (const_10_to_11_operand): Ditto.
2723 (const_12_to_13_operand): Ditto.
2724 (const_14_to_15_operand): Ditto.
2725 (const_16_to_19_operand): Ditto.
2726 (const_20_to_23_operand): Ditto.
2727 (const_24_to_27_operand): Ditto.
2728 (const_28_to_31_operand): Ditto.
2729 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
2730 (cvtusi2<ssescalarmodesuffix>32): New.
2731 (cvtusi2<ssescalarmodesuffix>64): Ditto.
2732 (ufloatv16siv16sf2): Ditto.
2733 (avx512f_fix_notruncv16sfv16si): Ditto.
2734 (avx512f_ufix_notruncv16sfv16si): Ditto.
2735 (avx512f_vcvtss2usi): Ditto.
2736 (avx512f_vcvtss2usiq): Ditto.
2737 (avx512f_vcvttss2usi): Ditto.
2738 (avx512f_vcvttss2usiq): Ditto.
2739 (avx512f_vcvtsd2usi): Ditto.
2740 (avx512f_vcvtsd2usiq): Ditto.
2741 (avx512f_vcvttsd2usi): Ditto.
2742 (avx512f_vcvttsd2usiq): Ditto.
2743 (ufloatv8siv8df): Ditto.
2744 (avx512f_cvtdq2pd512_2): Ditto.
2745 (avx512f_cvtpd2dq512): Ditto.
2746 (avx512f_ufix_notruncv8dfv8si): Ditto.
2747 (avx512f_cvtpd2ps512): Ditto.
2748 (vec_unpacks_lo_v16sf): Ditto.
2749 (vec_unpacks_hi_v16sf): Ditto.
2750 (vec_unpacks_float_hi_v16si): Ditto.
2751 (vec_unpacks_float_lo_v16si): Ditto.
2752 (avx512f_unpckhps512): Ditto.
2753 (avx512f_unpcklps512): Ditto.
2754 (avx512f_movshdup512): Ditto.
2755 (avx512f_movsldup512): Ditto.
2756 (vec_extract_lo_v32hi): Ditto.
2757 (vec_extract_hi_v32hi): Ditto.
2758 (vec_extract_lo_v64qi): Ditto.
2759 (vec_extract_hi_v64qi): Ditto.
2760 (avx512f_unpckhpd512): Ditto.
2761 (avx512f_movddup512): Ditto.
2762 (avx512f_unpcklpd512): Ditto.
2763 (*avx512f_unpcklpd512): Ditto.
2764 (avx512f_shufps512_1): Ditto.
2765 (avx512f_shufpd512_1): Ditto.
2766 (avx512f_interleave_highv8di): Ditto.
2767 (avx512f_interleave_lowv8di): Ditto.
2768 (PMOV_DST_MODE): Ditto.
2769 (pmov_src_mode): Ditto.
2770 (pmov_src_lower): Ditto.
2772 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
2773 (*avx512f_<code>v8div16qi2): Ditto.
2774 (*avx512f_<code>v8div16qi2_store): Ditto.
2775 (vec_widen_umult_even_v16si): Ditto.
2776 (*vec_widen_umult_even_v16si): Ditto.
2777 (vec_widen_smult_even_v16si): Ditto.
2778 (*vec_widen_smult_even_v16si): Ditto.
2779 (avx512f_interleave_highv16si): Ditto.
2780 (avx512f_interleave_lowv16si): Ditto.
2781 (avx512f_<code>v16qiv16si2): Ditto.
2782 (avx512f_<code>v16hiv16si2): Ditto.
2783 (avx512f_<code>v8qiv8di2): Ditto.
2784 (avx512f_<code>v8hiv8di2): Ditto.
2785 (avx512f_<code>v8siv8di2): Ditto.
2786 (avx512cd_maskb_vec_dupv8di): Ditto.
2787 (avx512cd_maskw_vec_dupv16si): Ditto.
2788 (avx512f_vcvtph2ps512): Ditto.
2789 (avx512f_vcvtps2ph512): Ditto.
2790 (VEC_EXTRACT_MODE): Extened with wider modes.
2791 (VEC_PERM_AVX2): Ditto.
2792 (VEC_PERM_CONST): Ditto.
2794 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
2796 * config/arc/arc.c (arc_ccfsm_post_advance):
2797 Add comment about TYPE_RETURN.
2799 2013-10-28 Bin Cheng <bin.cheng@arm.com>
2801 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
2802 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
2803 (strip_offset): Convert offset to unsigned number.
2805 2013-10-27 Tom de Vries <tom@codesourcery.com>
2807 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
2808 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
2810 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
2812 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
2813 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
2814 (TARGET_BIG_ENDIAN): New macro.
2815 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
2816 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
2817 * config/sh/sh.c: Likewise.
2818 * config/sh/sh.h: Likewise.
2820 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
2822 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
2823 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
2825 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2827 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
2829 (cmpstr_t, cmpstrsi): Fix formatting.
2831 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
2834 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
2835 addressing, do not use general_operand for memory operands.
2837 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
2840 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2841 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2843 2013-10-26 Jeff Law <law@redhat.com>
2845 * predict.c (PRED_MUDFLAP): Remove.
2846 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
2848 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
2849 (OBJS): Remove tree-nomudflap.o
2850 (GTFILES): Remove tree-mudflap.c
2851 * builtins.c (expand_builtin_alloc): Remove mudflap support.
2852 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2853 (mfwrap_spec, mflib_spec): Likewise.
2854 (cpp_unique_options, cc1_options, static_specs): Likewise.
2855 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
2856 * passes.def: Likewise.
2857 * toplev.c (compile_file, process_options): Likewise.
2858 * tree-inline.c (copy_tree_r): Likewise.
2859 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
2860 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
2861 (build_constant_desc, output_constant_def_contents): Likewise.
2862 (categorize_decl_for_section): Likewise.
2863 * tree-mudflap.c: Removed.
2864 * tree-mudflap.h: Removed.
2865 * tree-nomudflap.c: Removed.
2866 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
2867 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
2868 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
2869 * config/sol2.h (MFLIB_SPEC): Likewise.
2870 * doc/install.texi: Remove mudflap references.
2871 * doc/passes.texi: Similarly.
2872 * doc/sourcebuild.texi: Similarly.
2873 * doc/invoke.texi: Remove mudlfap related options.
2875 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2877 PR rtl-optimization/58759
2878 * lra-constraints.c (lra_constraints): Remove wrong condition to
2879 remove insn setting up an equivalent pseudo.
2881 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
2883 * config/rs6000/rs6000-protos.h
2884 (rs6000_secondary_memory_needed_mode): New prototype.
2885 * config/rs6000/rs6000.c: Include ira.h.
2886 (TARGET_LRA_P): Redefine.
2887 (rs6000_legitimate_offset_address_p): Call
2888 legitimate_constant_pool_address_p in strict mode for LRA.
2889 (rs6000_legitimate_address_p): Ditto.
2890 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
2891 (rs6000_emit_move): Add LRA version of code to generate load/store
2893 (rs6000_secondary_memory_needed_mode): New.
2894 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
2895 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
2896 constants, memory, and FP registers.
2897 (rs6000_lra_p): New.
2898 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
2899 * config/rs6000/rs6000.opt (mlra): New option.
2900 * lra-spills.c (lra_final_code_change): Remove useless move insns.
2902 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
2904 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
2905 has_single_use () and not do the conversion if has_single_use ()
2906 returns false for the multiplication result.
2908 2013-10-25 David Malcolm <dmalcolm@redhat.com>
2910 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
2911 to "NODE", since this works on a "tree", not an
2913 (CONSTANT_CLASS_P): Likewise.
2916 (INDIRECT_REF_P): Likewise.
2917 (REFERENCE_CLASS_P): Likewise.
2918 (COMPARISON_CLASS_P): Likewise.
2919 (UNARY_CLASS_P): Likewise.
2920 (BINARY_CLASS_P): Likewise.
2921 (STATEMENT_CLASS_P): Likewise.
2922 (VL_EXP_CLASS_P): Likewise.
2923 (EXPRESSION_CLASS_P): Likewise.
2924 (IS_TYPE_OR_DECL_P): Likewise.
2926 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
2928 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
2929 ADDR_EXPR in the defining statement.
2931 2013-10-25 Richard Biener <rguenther@suse.de>
2933 PR tree-optimization/58626
2934 * tree-loop-distribution.c (enum rdg_dep_type): Remove
2935 anti_dd, output_dd and input_dd.
2936 (struct rdg_edge): Remove level and relation members.
2937 (RDGE_LEVEL, RDGE_RELATION): Remove.
2938 (dot_rdg_1): Adjust.
2939 (create_rdg_edge_for_ddr): Remove.
2940 (create_rdg_edges_for_scalar): Adjust.
2941 (create_edge_for_control_dependence): Likewise.
2942 (create_rdg_edges): Split into ...
2943 (create_rdg_flow_edges): ... this
2944 (create_rdg_cd_edges): ... and this.
2946 (build_rdg): Likewise, do not compute data dependences or
2948 (pg_add_dependence_edges): New function.
2950 (distribute_loop): First apply all non-dependence based
2951 partition mergings. Then compute dependences between partitions
2952 and merge and order partitions according to them.
2954 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2956 PR rtl-optimization/58831
2957 * alias.c (init_alias_analysis): At the beginning of each iteration,
2958 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
2960 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
2962 * recog.c (search_ofs): New static variable moved from...
2963 (peep2_find_free_register): ...here.
2964 (peephole2_optimize): Initialize it.
2966 2013-10-25 Tobias Burnus <burnus@net-b.de>
2968 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
2970 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
2972 * config/i386/i386.h (TARGET_MPX): New define.
2973 (TARGET_MPX_P): Ditto.
2975 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2977 * config/i386/constraints.md (B): New.
2980 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
2981 * config/i386/i386-modes.def (BND32): New.
2983 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
2984 * config/i386/i386.c (isa_opts): Add mmpx.
2985 (regclass_map): Add bound registers.
2986 (dbx_register_map): Likewise.
2987 (dbx64_register_map): Likewise.
2988 (svr4_dbx_register_map): Likewise.
2990 (ix86_option_override_internal): Support MPX ISA.
2991 (ix86_conditional_register_usage): Support bound registers.
2992 (print_reg): Likewise.
2993 (ix86_code_end): Add MPX bnd prefix.
2994 (output_set_got): Likewise.
2995 (ix86_output_call_insn): Likewise.
2996 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
2997 (ix86_print_operand_punct_valid_p): Likewise.
2998 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
3000 (ix86_class_likely_spilled_p): Add bound regs support.
3001 (ix86_hard_regno_mode_ok): Likewise.
3002 (x86_order_regs_for_local_alloc): Likewise.
3003 (ix86_bnd_prefixed_insn_p): New.
3004 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
3005 (FIXED_REGISTERS): Add bound registers.
3006 (CALL_USED_REGISTERS): Likewise.
3007 (REG_ALLOC_ORDER): Likewise.
3008 (HARD_REGNO_NREGS): Likewise.
3010 (VALID_BND_REG_MODE): New.
3011 (FIRST_BND_REG): New.
3012 (LAST_BND_REG): New.
3013 (reg_class): Add BND_REGS.
3014 (REG_CLASS_NAMES): Likewise.
3015 (REG_CLASS_CONTENTS): Likewise.
3017 (ANY_BND_REG_P): New.
3019 (HI_REGISTER_NAMES): Add bound registers.
3020 * config/i386/i386.md (UNSPEC_BNDMK): New.
3021 (UNSPEC_BNDMK_ADDR): New.
3022 (UNSPEC_BNDSTX): New.
3023 (UNSPEC_BNDLDX): New.
3024 (UNSPEC_BNDLDX_ADDR): New.
3025 (UNSPEC_BNDCL): New.
3026 (UNSPEC_BNDCU): New.
3027 (UNSPEC_BNDCN): New.
3028 (UNSPEC_MPX_FENCE): New.
3031 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
3032 (length_immediate): Likewise.
3033 (prefix_0f): Likewise.
3035 (prefix_rep): Check for bnd prefix.
3036 (length_nobnd): New.
3037 (length): Use length_nobnd if specified.
3042 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
3045 (simple_return_internal): Likewise.
3046 (simple_return_pop_internal): Likewise.
3047 (*indirect_jump): Add MPX bnd prefix.
3048 (*tablejump_1): Likewise.
3049 (simple_return_internal_long): Likewise.
3050 (simple_return_indirect_internal): Likewise.
3054 (*mov<mode>_internal_mpx): New.
3055 (<mode>_<bndcheck>): New.
3056 (*<mode>_<bndcheck>): New.
3061 * config/i386/predicates.md (lea_address_operand): Rename to...
3062 (address_no_seg_operand): ... this.
3063 (address_mpx_no_base_operand): New.
3064 (address_mpx_no_index_operand): New.
3065 (bnd_mem_operator): New.
3066 * config/i386/i386.opt (mmpx): New.
3067 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
3068 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
3070 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
3072 * mode-classes.def (MODE_POINTER_BOUNDS): New.
3073 * tree.def (POINTER_BOUNDS_TYPE): New.
3074 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
3075 (POINTER_BOUNDS_MODE): New.
3076 (make_pointer_bounds_mode): New.
3077 * machmode.h (POINTER_BOUNDS_MODE_P): New.
3078 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
3079 (layout_type): Support POINTER_BOUNDS_TYPE.
3080 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
3081 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
3082 (type_contains_placeholder_1): Likewise.
3083 * tree.h (POINTER_BOUNDS_TYPE_P): New.
3084 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
3085 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
3087 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3089 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
3090 mode of address computation when calling memory_address_addr_space.
3092 2013-08-24 Richard Henderson <rth@twiddle.net>
3095 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
3096 instead of create_convert_operand_to.
3097 (maybe_emit_sync_lock_test_and_set): Likewise.
3098 (expand_atomic_compare_and_swap): Likewise.
3099 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
3101 2013-08-24 Sriraman Tallam <tmsriram@google.com>
3103 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
3104 Change param_is to use the struct and not the pointer to the struct.
3106 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
3108 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
3109 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3110 Move to targhooks.c.
3111 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
3112 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
3113 Relocate from gimplify.c.
3114 * targhooks.h: Add 2 prototypes.
3115 * tree.h: Delete 2 prototypes.
3117 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
3119 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
3120 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
3122 2013-10-24 Cong Hou <congh@google.com>
3124 * convert.c (convert_to_real): Guard those unsafe math function
3125 convertions with flag_unsafe_math_optimizations. Handle sqrt()
3128 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
3131 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
3133 (cgraph_create_edge): Use the new argument.
3134 (cgraph_create_indirect_edge): Likewise.
3136 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
3138 * config/arc/arc.c (arc_ccfsm_post_advance): Also handle
3140 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
3141 changes statep->state.
3143 2013-10-24 Tobias Burnus <burnus@net-b.de>
3146 * tree-cfg.c (replace_loop_annotate): New function.
3147 (execute_build_cfg): Call it.
3148 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
3149 * internal-fn.c (expand_ANNOTATE): New function.
3150 * internal-fn.def (ANNOTATE): Define as new internal function.
3151 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
3152 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
3153 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
3154 * doc/extend.texi (Pragmas): Document #pragma ivdep.
3155 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
3157 2013-10-17 Ian Bolton <ian.bolton@arm.com>
3158 Marcus Shawcroft <marcus.shawcroft@arm.com>
3160 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
3161 Special case reload SP+C into none GENERAL_REGS.
3163 2013-10-24 Michael Matz <matz@suse.de>
3165 * gengtype.c (is_file_equal): Check that files will be same length.
3167 2013-10-25 Christian Bruel <christian.bruel@st.com>
3169 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
3170 * config/sh/t-sh (sh-mem.o): New rule.
3171 * config/sh/sh-mem.cc (expand_block_move): Moved here.
3172 (sh_expand_cmpstr): New function.
3173 * config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
3174 * config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
3175 * config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
3176 (rotlhi3_8): Rename.
3178 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3180 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
3181 * Makefile.in (CXXDEPMODE): Assign and change users.
3182 (CCDEPMODE): Delete.
3183 * configure: Regenerate.
3185 2013-10-23 David Malcolm <dmalcolm@redhat.com>
3187 * gengtype-parse.c (require_without_advance): New.
3188 (type): For GTY-marked types that are not GTY((user)), parse any
3189 base classes, requiring them to be single-inheritance, and not
3190 be templates. For non-GTY-marked types and GTY((user)),
3191 continue to skip over any C++ inheritance specification.
3192 * gengtype-state.c (state_writer::write_state_struct_type):
3193 Write base class of type (if any).
3194 (read_state_struct_type): Read base class of type (if any).
3195 * gengtype.c (new_structure): Add a "base_class" parameter.
3196 (create_optional_field_): Update for new parameter to new_structure.
3197 (adjust_field_rtx_def): Likewise.
3198 (adjust_field_tree_exp): Likewise.
3199 * gengtype.h (struct type): Add "base_class" field to the s
3201 (new_structure): Add "base" parameter.
3203 2013-10-23 Sriraman Tallam <tmsriram@google.com>
3206 * config/i386/i386.c (ix86_option_override_internal):
3207 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
3208 (ix86_valid_target_attribute_tree):
3209 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
3210 Change TARGET_SSE to TARGET_SSE_P (opts->...)
3212 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3214 * tree-ssa-loop.h: Remove include files.
3215 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3216 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
3217 * graphite-clast-to-gimple.c: Likewise.
3218 * graphite-scop-detection.c: Likewise.
3219 * graphite-sese-to-poly.c: Likewise.
3220 * ipa-inline-analysis.c: Likewise.
3221 * ipa-pure-const.c: Likewise.
3222 * loop-init.c: Likewise.
3223 * passes.c: Likewise.
3224 * predict.c: Likewise.
3225 * tree-cfg.c: Likewise.
3226 * tree-cfgcleanup.c: Likewise.
3227 * tree-chrec.c: Likewise.
3228 * tree-data-ref.c: Likewise.
3229 * tree-loop-distribution.c: Likewise.
3230 * tree-parloops.c: Likewise.
3231 * tree-predcom.c: Likewise.
3232 * tree-scalar-evolution.c: Likewise.
3233 * tree-ssa-address.c: Likewise.
3234 * tree-ssa.c: Likewise.
3235 * tree-ssa-dce.c: Likewise.
3236 * tree-ssa-loop.c: Likewise.
3237 * tree-ssa-loop-im.c: Likewise.
3238 * tree-ssa-loop-ivcanon.c: Likewise.
3239 * tree-ssa-loop-ivopts.c: Likewise.
3240 * tree-ssa-loop-manip.c: Likewise.
3241 * tree-ssa-loop-niter.c: Likewise.
3242 * tree-ssa-loop-prefetch.c: Likewise.
3243 * tree-ssa-loop-unswitch.c: Likewise.
3244 * tree-ssa-reassoc.c: Likewise.
3245 * tree-vect-data-refs.c: Likewise.
3246 * tree-vect-loop.c: Likewise.
3247 * tree-vect-loop-manip.c: Likewise.
3248 * tree-vectorizer.c: Likewise.
3249 * tree-vect-stmts.c: Likewise.
3250 * tree-vrp.c: Likewise.
3252 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3254 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
3256 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3258 PR tree-optimization/58775
3259 PR tree-optimization/58791
3260 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
3261 (insert_stmt_after): Rewritten, don't move the stmt, but really
3263 (get_stmt_uid_with_default): Remove.
3264 (build_and_add_sum): Use insert_stmt_after and
3265 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
3266 (update_range_test): Set uid on stmts added by
3267 force_gimple_operand_gsi. Don't immediately modify statements
3268 in inter-bb optimization, just update oe->op values.
3269 (optimize_range_tests): Return bool whether any changed have been made.
3270 (update_ops): New function.
3271 (struct inter_bb_range_test_entry): New type.
3272 (maybe_optimize_range_tests): Perform statement changes here.
3273 (not_dominated_by, appears_later_in_bb, get_def_stmt,
3274 ensure_ops_are_available): Remove.
3275 (find_insert_point): Rewritten.
3276 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
3277 return LHS of the (new resp. old) stmt. Don't call
3278 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
3279 instead of last, move new stmt at the right place.
3280 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
3281 (negate_value): Likewise. Set uids.
3282 (break_up_subtract_bb): Initialize uids.
3283 (reassociate_bb): Adjust rewrite_expr_tree caller.
3284 (do_reassoc): Don't call renumber_gimple_stmt_uids.
3286 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
3289 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
3290 TARGET_32BIT final condition.
3291 (mulsi3_internal2 and splitter): Same.
3293 2013-10-23 Jeff Law <law@redhat.com>
3295 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
3296 through joiner blocks with abnormal outgoing edges.
3298 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
3299 Add parameter JOINERS, to allow/disallow threading through joiner
3301 (thread_block): New. Call thread_block_1.
3302 (mark_threaded_blocks): Remove code to filter out certain cases
3303 of threading through joiner blocks.
3304 (thread_through_all_blocks): Document how we can have a dangling
3305 edge AUX field and clear it.
3307 2013-10-23 Ian Lance Taylor <iant@google.com>
3309 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
3310 (C++ Dialect Options): Likewise.
3311 (Optimize Options): Likewise.
3313 2013-10-23 Tom de Vries <tom@codesourcery.com>
3315 PR tree-optimization/58805
3316 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
3318 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3320 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
3321 sequence based on get_range_info returned range.
3323 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
3325 * tree-ssa.h: Remove all #include's
3326 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
3327 * alias.c: Move required includes from tree-ssa.h.
3329 * builtins.c: Likewise.
3330 * calls.c: Likewise.
3331 * cfgexpand.c: Likewise.
3332 * cfghooks.c: Likewise.
3333 * cfgloop.c: Likewise.
3334 * cfgloopmanip.c: Likewise.
3335 * cgraph.c: Likewise.
3336 * cgraphbuild.c: Likewise.
3337 * cgraphclones.c: Likewise.
3338 * cgraphunit.c: Likewise.
3340 * except.c: Likewise.
3342 * final.c: Likewise.
3343 * fold-const.c: Likewise.
3344 * ggc-page.c: Likewise.
3345 * gimple-builder.c: Likewise.
3346 * gimple-fold.c: Likewise.
3347 * gimple-iterator.c: Likewise.
3348 * gimple-low.c: Likewise.
3349 * gimple-pretty-print.c: Likewise.
3350 * gimple-ssa-strength-reduction.c: Likewise.
3351 * gimple-streamer-in.c: Likewise.
3352 * gimple-streamer-out.c: Likewise.
3353 * gimplify.c: Likewise.
3354 * graphite-blocking.c: Likewise.
3355 * graphite-clast-to-gimple.c: Likewise.
3356 * graphite-dependences.c: Likewise.
3357 * graphite-interchange.c: Likewise.
3358 * graphite-optimize-isl.c: Likewise.
3359 * graphite-poly.c: Likewise.
3360 * graphite-scop-detection.c: Likewise.
3361 * graphite-sese-to-poly.c: Likewise.
3362 * graphite.c: Likewise.
3363 * ipa-cp.c: Likewise.
3364 * ipa-inline-analysis.c: Likewise.
3365 * ipa-inline-transform.c: Likewise.
3366 * ipa-inline.c: Likewise.
3367 * ipa-prop.c: Likewise.
3368 * ipa-pure-const.c: Likewise.
3369 * ipa-reference.c: Likewise.
3370 * ipa-split.c: Likewise.
3371 * ipa-utils.c: Likewise.
3372 * loop-init.c: Likewise.
3373 * lto-cgraph.c: Likewise.
3374 * lto-section-in.c: Likewise.
3375 * lto-section-out.c: Likewise.
3376 * lto-streamer-in.c: Likewise.
3377 * lto-streamer-out.c: Likewise.
3378 * lto-streamer.c: Likewise.
3379 * omp-low.c: Likewise.
3380 * passes.c: Likewise.
3381 * predict.c: Likewise.
3382 * print-tree.c: Likewise.
3383 * profile.c: Likewise.
3385 * targhooks.c: Likewise.
3386 * tracer.c: Likewise.
3387 * trans-mem.c: Likewise.
3388 * tree-call-cdce.c: Likewise.
3389 * tree-cfg.c: Likewise.
3390 * tree-cfgcleanup.c: Likewise.
3391 * tree-chrec.c: Likewise.
3392 * tree-complex.c: Likewise.
3393 * tree-data-ref.c: Likewise.
3394 * tree-dfa.c: Likewise.
3395 * tree-eh.c: Likewise.
3396 * tree-emutls.c: Likewise.
3397 * tree-if-conv.c: Likewise.
3398 * tree-inline.c: Likewise.
3399 * tree-into-ssa.c: Likewise.
3400 * tree-loop-distribution.c: Likewise.
3401 * tree-mudflap.c: Likewise.
3402 * tree-nested.c: Likewise.
3403 * tree-nrv.c: Likewise.
3404 * tree-object-size.c: Likewise.
3405 * tree-outof-ssa.c: Likewise.
3406 * tree-parloops.c: Likewise.
3407 * tree-phinodes.c: Likewise.
3408 * tree-predcom.c: Likewise.
3409 * tree-pretty-print.c: Likewise.
3410 * tree-profile.c: Likewise.
3411 * tree-scalar-evolution.c: Likewise.
3412 * tree-sra.c: Likewise.
3413 * tree-ssa-address.c: Likewise.
3414 * tree-ssa-alias.c: Likewise.
3415 * tree-ssa-ccp.c: Likewise.
3416 * tree-ssa-coalesce.c: Likewise.
3417 * tree-ssa-copy.c: Likewise.
3418 * tree-ssa-copyrename.c: Likewise.
3419 * tree-ssa-dce.c: Likewise.
3420 * tree-ssa-dom.c: Likewise.
3421 * tree-ssa-dse.c: Likewise.
3422 * tree-ssa-forwprop.c: Likewise.
3423 * tree-ssa-ifcombine.c: Likewise.
3424 * tree-ssa-live.c: Likewise.
3425 * tree-ssa-loop-ch.c: Likewise.
3426 * tree-ssa-loop-im.c: Likewise.
3427 * tree-ssa-loop-ivcanon.c: Likewise.
3428 * tree-ssa-loop-ivopts.c: Likewise.
3429 * tree-ssa-loop-manip.c: Likewise.
3430 * tree-ssa-loop-niter.c: Likewise.
3431 * tree-ssa-loop-prefetch.c: Likewise.
3432 * tree-ssa-loop-unswitch.c: Likewise.
3433 * tree-ssa-loop.c: Likewise.
3434 * tree-ssa-math-opts.c: Likewise.
3435 * tree-ssa-operands.c: Likewise.
3436 * tree-ssa-phiopt.c: Likewise.
3437 * tree-ssa-phiprop.c: Likewise.
3438 * tree-ssa-pre.c: Likewise.
3439 * tree-ssa-propagate.c: Likewise.
3440 * tree-ssa-reassoc.c: Likewise.
3441 * tree-ssa-sccvn.c: Likewise.
3442 * tree-ssa-sink.c: Likewise.
3443 * tree-ssa-strlen.c: Likewise.
3444 * tree-ssa-structalias.c: Likewise.
3445 * tree-ssa-tail-merge.c: Likewise.
3446 * tree-ssa-ter.c: Likewise.
3447 * tree-ssa-threadedge.c: Likewise.
3448 * tree-ssa-threadupdate.c: Likewise.
3449 * tree-ssa-uncprop.c: Likewise.
3450 * tree-ssa-uninit.c: Likewise.
3451 * tree-ssa.c: Likewise.
3452 * tree-ssanames.c: Likewise.
3453 * tree-stdarg.c: Likewise.
3454 * tree-streamer-in.c: Likewise.
3455 * tree-switch-conversion.c: Likewise.
3456 * tree-tailcall.c: Likewise.
3457 * tree-vect-data-refs.c: Likewise.
3458 * tree-vect-generic.c: Likewise.
3459 * tree-vect-loop-manip.c: Likewise.
3460 * tree-vect-loop.c: Likewise.
3461 * tree-vect-patterns.c: Likewise.
3462 * tree-vect-slp.c: Likewise.
3463 * tree-vect-stmts.c: Likewise.
3464 * tree-vectorizer.c: Likewise.
3465 * tree-vrp.c: Likewise.
3468 * value-prof.c: Likewise.
3469 * var-tracking.c: Likewise.
3470 * varpool.c: Likewise.
3471 * vtable-verify.c: Likewise.
3473 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3475 * config/tilegx/tilegx.c: Include "tree.h".
3477 2013-10-23 Jakub Jelinek <jakub@redhat.com>
3479 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
3480 the info, not after it.
3481 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
3482 dump_ssaname_info call.
3483 (pp_gimple_stmt_1): Adjust caller.
3484 (dump_phi_nodes): Likewise. Don't print "# " here.
3486 2013-10-22 Jan Hubicka <jh@suse.cz>
3488 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
3490 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
3491 * i386.c (expand_small_movmem_or_setmem): New function.
3492 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
3494 (alg_usable_p): Add support for value ranges; cleanup.
3495 (ix86_expand_set_or_movmem): Add support for misaligned moves.
3497 2013-10-22 Sterling Augustine <saugustine@google.com>
3499 * doc/invoke.texi: Document -ggnu-pubnames.
3500 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
3502 * dwarf2out.c: Include gdb/gdb-index.h.
3503 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
3504 debug_generate_pub_sections.
3505 (is_java, output_pubtables, output_pubname): New functions.
3506 (include_pubname_in_output): Handle debug_generate_pub_sections at
3508 (size_of_pubnames): Use new local space_for_flags based on
3509 debug_generate_pub_sections.
3510 (output_pubnames): Unify pubnames and pubtypes output logic.
3511 Genericize comments. Call output_pubname.
3512 (dwarf2out_finish): Move logic to output_pubnames and call it.
3514 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
3517 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
3518 Remove CCCmode handling.
3519 <case LTU>: Return 'c' suffix for CCCmode.
3520 <case GEU>: Return 'nc' suffix for CCCmode.
3521 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
3522 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
3523 (*sub<mode>3_cc_overflow): Ditto.
3524 (*subsi3_zext_cc_overflow): Ditto.
3526 2013-10-22 Steve Ellcey <sellcey@mips.com>
3528 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
3529 (AND (NOT OP1) (NOT OP2)).
3531 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
3533 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
3534 meaning of merge-high and merge-low masks for little endian; avoid
3535 use of vector-pack masks for little endian for mismatched modes.
3537 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3539 * config/tilepro/tilepro.c: Include "tree.h".
3541 2013-10-22 Andreas Schwab <schwab@suse.de>
3543 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
3546 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
3548 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
3549 [Ed Smith-Rowland]): New entries.
3550 ([Stephen M. Webb]): Update.
3552 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
3554 * tree-ssa-ter.h: Remove duplicate copy of file contents.
3556 2013-10-21 Marek Polacek <polacek@redhat.com>
3559 * fold-const.c (fold_range_test): Return 0 if the type is not
3562 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
3564 * system.h: Move hwint.h include further down.
3565 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
3566 gcc_checking_asserts.
3567 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
3569 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3571 Fix volatile issues in optimize_bit_field_compare.
3572 * fold-const.c (optimize_bit_field_compare): Bail out if
3573 lvolatilep or rvolatilep.
3575 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
3577 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
3578 and get_inner_reference returning different pmode for non-volatile
3579 bit-field members dependent on flag_strict_volatile_bitfields.
3580 * stor-layout.c (layout_decl): Remove special handling of
3581 flag_strict_volatile_bitfields.
3582 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
3583 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
3585 2013-10-21 Paulo Matos <pmatos@broadcom.com>
3587 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
3590 2013-10-21 Jeff Law <law@redhat.com>
3592 * tree-ssa-threadedge.c (thread_through_normal_block): New
3593 argument VISITED. Remove VISISTED as a local variable. When we
3594 have a threadable jump, verify the destination of the jump has not
3596 (thread_across_edge): Allocate VISITED bitmap once at function
3597 scope and use it throughout. Make sure to set appropriate bits in
3598 VISITED for E (start of jump thread path).
3599 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
3600 through a joiner if any edge on the path has a recorded jump thread.
3602 2013-10-21 Ian Lance Taylor <iant@google.com>
3604 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
3605 don't imply that attributes can solve all problems.
3606 (Directory Options): Fix typo.
3608 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3610 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
3611 extend and extend_add.
3613 2013-10-21 Richard Biener <rguenther@suse.de>
3615 PR tree-optimization/58794
3616 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
3617 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
3619 2013-10-21 Richard Biener <rguenther@suse.de>
3622 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
3623 to (T) X for sign-changing conversions (or no conversion).
3625 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
3627 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
3629 2013-10-20 Jan Hubicka <jh@suse.cz>
3631 * config/i386/i386-tune.def: Add comment; organize into categories
3633 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
3635 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
3636 argument. Update function comment.
3637 (expand_set_or_movmem_via_rep): New function combining
3638 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
3639 (expand_movmem_via_rep_mov): Remove.
3640 expand_setmem_via_rep_stos): Remove.
3641 (expand_movmem_epilogue): Update calls correspondingly.
3642 (expand_setmem_epilogue_via_loop): Likewise.
3644 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
3645 (expand_set_or_movmem_prologue): New function combining
3646 expand_movmem_prologue and expand_setmem_prologue.
3647 (expand_movmem_prologue): Remove.
3648 (expand_setmem_prologue): Remove.
3649 (expand_set_or_movmem_constant_prologue): New function combining
3650 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
3651 (expand_constant_movmem_prologue): Remove.
3652 (expand_constant_setmem_prologue): Remove.
3653 (promote_duplicated_reg): Allow vector-const0 value.
3654 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
3655 and ix86_expand_setmem.
3656 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
3657 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
3659 2013-10-21 Diego Novillo <dnovillo@google.com>
3661 * asan.c: Include tree.h
3662 * bb-reorder.c: Likewise.
3663 * cfgcleanup.c: Likewise.
3664 * cfgloopmanip.c: Likewise.
3665 * data-streamer-in.c: Likewise.
3666 * data-streamer-out.c: Likewise.
3667 * data-streamer.c: Likewise.
3668 * dwarf2cfi.c: Likewise.
3669 * graphite-blocking.c: Likewise.
3670 * graphite-clast-to-gimple.c: Likewise.
3671 * graphite-dependences.c: Likewise.
3672 * graphite-interchange.c: Likewise.
3673 * graphite-optimize-isl.c: Likewise.
3674 * graphite-poly.c: Likewise.
3675 * graphite-scop-detection.c: Likewise.
3676 * graphite-sese-to-poly.c: Likewise.
3677 * graphite.c: Likewise.
3678 * ipa-devirt.c: Likewise.
3679 * ipa-profile.c: Likewise.
3682 * loop-init.c: Likewise.
3683 * loop-unroll.c: Likewise.
3684 * lower-subreg.c: Likewise.
3685 * lto/lto-object.c: Likewise.
3686 * recog.c: Likewise.
3687 * reginfo.c: Likewise.
3688 * tree-loop-distribution.c: Likewise.
3689 * tree-parloops.c: Likewise.
3690 * tree-ssa-strlen.c: Likewise.
3691 * tree-streamer.c: Likewise.
3692 * value-prof.c: Likewise.
3693 * target-globals.c: Likewise.
3694 * expr.h: Include tree-core.h instead of tree.h.
3695 * gimple.h: Likewise.
3696 * ipa-prop.h: Likewise.
3697 * ipa-utils.h: Likewise.
3698 * lto-streamer.h: Likewise.
3699 * streamer-hooks.h: Likewise.
3700 * ipa-reference.h: Include cgraph.h instead of tree.h.
3701 * cgraph.h: Include basic-block.h instead of tree.h.
3702 * tree-streamer.h: Do not include tree.h.
3703 * genattrtab.c (write_header): Generate inclusion of
3705 * genautomata.c (main): Likewise.
3706 * genemit.c: Likewise.
3707 * genopinit.c: Likewise.
3708 * genoutput.c (output_prologue): Likewise.
3709 * genpeep.c: Likewise.
3711 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3713 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
3715 (vec_unpacku_hi_v8hi): Likewise.
3716 (vec_unpacku_lo_v16qi): Likewise.
3717 (vec_unpacku_lo_v8hi): Likewise.
3719 2013-10-20 Jan Hubicka <jh@suse.cz>
3721 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
3722 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
3723 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
3725 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
3727 * config/mips/mips.h (ISA_HAS_WSBH): Define.
3728 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
3730 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
3732 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
3735 * system.h: Undef m_slot.
3737 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
3739 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
3740 all elements for both endian flavors.
3742 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3745 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
3746 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
3747 and SI_REG for 64bit SYSV ABI targets.
3749 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
3751 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
3752 to multi_reg_return. Clarify that we are skipping USEs of multiple
3753 return registers. Use bool type where appropriate.
3755 2013-10-18 Jan Hubicka <jh@suse.cz>
3757 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
3759 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
3760 (X86_TUNE_PUSH_MEMORY): Likewise.
3761 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
3762 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
3763 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
3765 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
3766 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
3768 (ix86_option_override_internal): Update to use tune features instead
3771 2013-10-18 Cong Hou <congh@google.com>
3773 PR tree-optimization/58508
3774 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
3775 statement that contains data refs with zero-step.
3777 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3779 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
3780 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
3781 * gimple-low.c (gimple_check_call_arg,
3782 gimple_check_call_matching_types): Move to cgraph.c.
3783 * gimple-low.h: Remove prototype.
3784 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
3785 Relocate from gimple-low.c.
3786 * cgraph.h: Add prototype. Don't include basic-block.h.
3787 * gimplify.c: Add gimple-low to include list.
3788 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
3789 * tree-eh.c: Add gimple-low to include list.
3790 * tree-nested.c: Likewise.
3791 * cfgexpand.c: Add tree-ssa-address.h to include list.
3793 * gimple-fold.c: Likewise.
3794 * gimple-ssa-strength-reduction.c: Likewise.
3795 * trans-mem.c: Likewise.
3796 * tree-mudflap.c: Likewise.
3797 * tree-ssa-loop-ivopts.c: Likewise.
3798 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
3799 (degenerate_phi_result): Move to tree-phinodes.c.
3800 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
3801 * tree-ssa-threadedge.c: Likewise.
3802 * tree-vrp.c: Likewise.
3803 * tree-phinodes.c (degenerate_phi_result): Relocate here.
3804 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
3805 * tree-phinodes.h (degenerate_phi_result): Add prototype.
3806 * tree-ssa-copy.c: Include tree-ssa-dom.h.
3807 * tree-ssa-forwprop.c: Likewise.
3808 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
3809 pass_data_cleanup_cfg_post_optimizing,
3810 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
3811 * tree-optimize.c: Delete File.
3812 * graphite.c: Include tree-cfgcleanup.h.
3813 * passes.c: Likewise.
3814 * tree-cfg.c: Likewise.
3815 * tree-profile.c: Likewise.
3816 * tree-ssa-dse.c: Likewise.
3817 * tree-ssa-loop-ivcanon.c: Likewise.
3818 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
3819 * tree-outof-ssa.c: Include sbitmap.h.
3820 * tree-ssa-live.c: Likewise.
3821 * tree-ssa-propagate.c: Likewise.
3822 * tree-ssa-structalias.c: Likewise.
3823 * tree-stdarg.c: Likewise.
3824 * Makefile.in (OBJS): Delete tree-optimize.o.
3825 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
3826 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
3827 * varasm.c: Include basic-block.h.
3828 * cfgloop.h: Include function.h instead of basic-block.h
3829 (bb_loop_depth): Move to cfgloop.c.
3830 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
3832 2013-10-18 Teresa Johnson <tejohnson@google.com>
3834 * predict.c (probably_never_executed): Compare frequency-based
3835 count to number of training runs.
3836 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
3838 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3840 * config/arm/arm.c (cortexa9_extra_costs): New table.
3841 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
3843 2013-10-18 Jeff Law <law@redhat.com>
3845 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
3847 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
3848 (register_jump_thread): Add "registered_jump_thread" debug
3850 * dbgcnt.def (registered_jump_thread): New debug counter.
3852 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3854 * config/rs6000/rs6000.c: Include cgraph.h.
3856 2013-10-18 Teresa Johnson <tejohnson@google.com>
3858 * tree-ssa-tail-merge.c (replace_block_by): Update edge
3859 weights during merging.
3861 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
3863 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
3864 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
3865 * tree-ssa-operands.h: Remove prototype.
3866 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
3867 * gimple.c (virtual_operand_p): Relocate from gimple.c.
3868 * gimple.h: Add prototype.
3869 * gimple-ssa.h: Include tree-ssa-operands.h.
3870 * tree-dump.c: Add tree-cfg.h to include list.
3871 * tree-ssa-alias.c: Add ipa-reference.h to include list.
3872 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
3873 * config/i386/i386.c: Don't include tree-flow.h.
3874 * config/rs6000/rs6000.c: Likewise.
3876 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3878 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
3880 2013-10-18 Richard Biener <rguenther@suse.de>
3882 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
3883 or TYPE_UNSIGNED of integral types.
3884 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
3885 NULL_TREE for zero-precision integral types.
3887 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
3889 * config/aarch64/arm_neon.h
3890 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
3892 2013-10-17 Sriraman Tallam <tmsriram@google.com>
3895 * opth-gen.awk: Define target_flags_explicit.
3897 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3899 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
3900 fields to the reg_addr array that describes the valid addressing
3901 mode for any register, general purpose registers, floating point
3902 registers, and Altivec registers.
3903 (FIRST_RELOAD_REG_CLASS): Likewise.
3904 (LAST_RELOAD_REG_CLASS): Likewise.
3905 (struct reload_reg_map_type): Likewise.
3906 (reload_reg_map_type): Likewise.
3907 (RELOAD_REG_VALID): Likewise.
3908 (RELOAD_REG_MULTIPLE): Likewise.
3909 (RELOAD_REG_INDEXED): Likewise.
3910 (RELOAD_REG_OFFSET): Likewise.
3911 (RELOAD_REG_PRE_INCDEC): Likewise.
3912 (RELOAD_REG_PRE_MODIFY): Likewise.
3913 (reg_addr): Likewise.
3914 (mode_supports_pre_incdec_p): New helper functions to say whether
3915 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
3916 (mode_supports_pre_modify_p): Likewise.
3917 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
3918 print the valid address mode bits for each mode.
3919 (rs6000_debug_print_mode): Likewise.
3920 (rs6000_debug_reg_global): Likewise.
3921 (rs6000_setup_reg_addr_masks): New function to set up the address
3922 mask bits for each type.
3923 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
3924 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
3925 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
3926 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
3927 PRE_MODIFY are supported.
3928 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
3929 registers, instead of {src,dest}_av_p.
3930 (rs6000_print_options_internal): Tweak the debug output slightly.
3932 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3934 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
3937 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
3939 * tree-flow.h (struct omp_region): Move to omp-low.c.
3940 Remove omp_ prototypes and variables.
3941 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
3942 (copy_var_decl): Relocate prototype from tree-flow.h.
3943 * gimple.c (copy_var_decl): Relocate from omp-low.c.
3944 * tree.h: Move prototype to omp-low.h.
3945 * omp-low.h: New File. Relocate prototypes here.
3946 * omp-low.c (struct omp_region): Make local here.
3947 (root_omp_region): Make static.
3948 (copy_var_decl) Move to gimple.c.
3949 (new_omp_region): Make static.
3950 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
3951 * tree-cfg.c: Include omp-low.h.
3952 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
3953 * gimplify.c: Include omp-low.h.
3954 * tree-parloops.c: Likewise.
3956 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
3958 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
3959 operands are in memory, prefer to force non-matched operand 1 to
3962 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
3965 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
3966 restrict TImode addresses to single indirect registers if both
3967 -mquad-memory and -mvsx-timode are used.
3968 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
3969 we should emit load/store quad. Remove using %y for quad memory
3972 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
3973 constraints to allow load/store quad on machines where TImode is
3974 not allowed in VSX registers. Use 'n' instead of 'F' constraint
3975 for TImode to load integer constants.
3977 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3979 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
3981 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
3983 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
3984 handling of STACK_REG.
3986 2013-10-17 Richard Biener <rguenther@suse.de>
3988 PR tree-optimization/58143
3989 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
3991 (rewrite_to_defined_overflow): Likewise.
3992 (move_computations_dom_walker::before_dom): Rewrite stmts
3993 with undefined signed overflow that are not always executed
3994 into unsigned arithmetic.
3996 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
3999 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
4000 explicit isa flag to be an options variable, instead of using
4001 global_options_set. Remove define from rs6000.h.
4002 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
4004 * config/rs6000/rs6000.c (rs6000_option_override_internal):
4005 Initialize rs6000_isa_flags_explicit.
4006 (rs6000_function_specific_save): Add gcc_options* parameter, so
4007 that the powerpc builds after the 2013-10-15 changes.
4008 (rs6000_function_specific_restore): Likewise.
4010 2013-10-16 DJ Delorie <dj@redhat.com>
4012 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
4013 op is a REG before checking REGNO.
4014 (rl78_alloc_physical_registers): Verify pattern is a SET before
4017 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4019 * config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
4021 (vec_unpacks_lo_v4sf): Likewise.
4022 (vec_unpacks_float_hi_v4si): Likewise.
4023 (vec_unpacks_float_lo_v4si): Likewise.
4024 (vec_unpacku_float_hi_v4si): Likewise.
4025 (vec_unpacku_float_lo_v4si): Likewise.
4027 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4029 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
4030 (vsx_concat_v2sf): Likewise.
4032 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
4034 * config/aarch64/aarch64.md
4035 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
4037 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
4039 PR tree-optimization/58697
4040 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
4041 estimated_loop_iterations_int.
4042 (max_stmt_executions_int): Call get_max_loop_iterations_int.
4043 (get_max_loop_iterations_int): New. HWINT version of
4044 get_max_loop_iterations.
4045 * cfgloop.h: Add prototypes.
4046 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
4047 * loop-unroll.c (decide_peel_once_rolling): Call
4048 get_estimated_loop_iterations_int.
4049 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
4050 * tree-ssa-loop-niter.h: Tweak prototypes.
4052 2013-10-16 David Malcolm <dmalcolm@redhat.com>
4054 * gengtype-parse.c (struct_field_seq): Ignore access-control
4055 keywords ("public:" etc).
4057 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
4059 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
4060 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
4062 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
4064 * config/arm/arm.opt (mlra): New option.
4065 * config/arm/arm.c (arm_lra_p): New function.
4066 (TARGET_LRA_P): Define.
4068 2013-10-16 Paulo Matos <pmatos@broadcom.com>
4070 * tree-core.h (tree_code_name): Remove.
4071 * tree.h (get_tree_code_name): New prototype.
4072 * tree.c (tree_code_name): Make static.
4073 (get_tree_code_name): New function.
4074 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
4075 tree_class_check_failed, tree_range_check_failed,
4076 tree_not_class_check_failed, omp_clause_check_failed,
4077 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
4078 wrapper get_tree_code_name instead of calling tree_code_name directly.
4079 * tree-vrp.c (dump_asserts_for): Likewise.
4080 * tree-dump.c (dequeue_and_dump): Likewise.
4081 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
4082 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
4083 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
4084 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
4085 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
4086 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
4087 dump_gimple_omp_for): Likewise.
4088 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
4089 * tree-ssa-pre.c (print_pre_expr): Likewise.
4090 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
4091 * print-tree.c (print_node_brief, print_node): Likewise.
4092 * gimple.c (gimple_check_failed): Likewise.
4093 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
4094 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
4095 * config/mep/mep.c (mep_validate_vliw): Likewise.
4096 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
4097 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
4099 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
4101 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
4104 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
4106 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
4107 (MULTILIB_MATCHES): Add multilib for -march=v8.
4109 2013-10-15 Sriraman Tallam <tmsriram@google.com>
4112 * optc-save-gen.awk: Add extra parameter to the save and restore
4114 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
4115 * tree.c (build_optimization_node): New parameter. Add extra parameter
4116 to call to cl_optimization_save.
4117 (build_target_option_node): New parameter. Add extra parameter
4118 to call to cl_target_option_save.
4119 * tree.h (build_optimization_node): New parameter.
4120 (build_target_option_node): New parameter.
4121 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
4122 build_optimization_node and build_target_option_node.
4123 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
4124 (handle_pragma_push_options): Ditto.
4125 * toplev.c (process_options): Ditto.
4126 * opts.c (init_options_struct): Check for opts_set non-null.
4127 * target.def (target_option.save): New parameter.
4128 (target_option.restore): New parameter.
4129 * tm.texi: Generate.
4130 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
4131 (ix86_pragma_target_parse): Ditto.
4132 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
4134 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
4135 to build_optimization_node and build_target_option_node.
4136 (rs6000_valid_attribute_p): Ditto.
4137 (rs6000_pragma_target_parse): Ditto.
4138 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
4140 * config/i386/i386.h:
4141 TARGET_64BIT_P: New Macro
4142 TARGET_MMX_P: New Macro.
4143 TARGET_3DNOW_P: New Macro.
4144 TARGET_3DNOW_A_P: New Macro.
4145 TARGET_SSE_P: New Macro.
4146 TARGET_SSE2_P: New Macro.
4147 TARGET_SSE3_P: New Macro.
4148 TARGET_SSSE3_P: New Macro.
4149 TARGET_SSE4_1_P: New Macro.
4150 TARGET_SSE4_2_P: New Macro.
4151 TARGET_AVX_P: New Macro.
4152 TARGET_AVX2_P: New Macro.
4153 TARGET_AVX512F_P: New Macro.
4154 TARGET_AVX512PF_P: New Macro.
4155 TARGET_AVX512ER_P: New Macro.
4156 TARGET_AVX512CD_P: New Macro.
4157 TARGET_FMA_P: New Macro.
4158 TARGET_SSE4A_P: New Macro.
4159 TARGET_FMA4_P: New Macro.
4160 TARGET_XOP_P: New Macro.
4161 TARGET_LWP_P: New Macro.
4162 TARGET_ABM_P: New Macro.
4163 TARGET_BMI_P: New Macro.
4164 TARGET_BMI2_P: New Macro.
4165 TARGET_LZCNT_P: New Macro.
4166 TARGET_TBM_P: New Macro.
4167 TARGET_POPCNT_P: New Macro.
4168 TARGET_SAHF_P: New Macro.
4169 TARGET_MOVBE_P: New Macro.
4170 TARGET_CRC32_P: New Macro.
4171 TARGET_AES_P: New Macro.
4172 TARGET_PCLMUL_P: New Macro.
4173 TARGET_CMPXCHG16B_P: New Macro.
4174 TARGET_FSGSBASE_P: New Macro.
4175 TARGET_RDRND_P: New Macro.
4176 TARGET_F16C_P: New Macro.
4177 TARGET_RTM_P: New Macro.
4178 TARGET_HLE_P: New Macro.
4179 TARGET_RDSEED_P: New Macro.
4180 TARGET_PRFCHW_P: New Macro.
4181 TARGET_ADX_P: New Macro.
4182 TARGET_FXSR_P: New Macro.
4183 TARGET_XSAVE_P: New Macro.
4184 TARGET_XSAVEOPT_P: New Macro.
4185 TARGET_LP64_P: New Macro.
4186 TARGET_X32_P: New Macro.
4187 TARGET_FPMATH_DEFAULT_P: New Macro.
4188 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
4189 * config/i386/i386.c (ix86_option_override_internal): New parameters.
4191 Change ix86_tune_string to access opts->x_ix86_tune_string.
4192 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
4193 Change ix86_arch_string to access opts->x_ix86_arch_string.
4194 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
4195 Change ix86_pmode to access opts->x_ix86_pmode.
4196 Change ix86_abi to access opts->x_ix86_abi.
4197 Change ix86_cmodel to access opts->x_ix86_cmodel.
4198 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
4199 Change ix86_isa_flags_explicit to access
4200 opts->x_ix86_isa_flags_explicit.
4201 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
4202 Change ix86_regparm to access opts->x_ix86_regparm.
4203 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
4204 Change ix86_preferred_stack_boundary_arg to access
4205 opts->x_ix86_preferred_stack_boundary_arg.
4206 Change ix86_force_align_arg_pointer to access
4207 opts->x_ix86_force_align_arg_pointer.
4208 Change ix86_incoming_stack_boundar_arg to access
4209 opts->x_ix86_incoming_stack_boundar_arg.
4210 Change ix86_fpmath to access opts->x_ix86_fpmath.
4211 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
4212 Change ix86_recip_name to access opts->x_ix86_recip_name.
4213 Change ix86_stack_protector_guard to access
4214 opts->x_ix86_stack_protector_guard.
4215 Change ix86_tune_memcpy_strategy to access
4216 opts->x_ix86_tune_memcpy_strategy.
4217 Change ix86_tune_memset_strategy to access
4218 opts->x_ix86_tune_memset_strategy.
4219 Change global_options to access opts.
4220 Change global_options_set to access opts_set.
4221 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
4222 Change TARGET_MMX to TARGET_MMX_P (opts->...).
4223 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
4224 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
4225 Change TARGET_SSE to TARGET_SSE_P (opts->...).
4226 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
4227 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
4228 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
4229 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
4230 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
4231 Change TARGET_AVX to TARGET_AVX_P (opts->...).
4232 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
4233 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
4234 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
4235 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
4236 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
4237 Change TARGET_FMA to TARGET_FMA_P (opts->...).
4238 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
4239 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
4240 Change TARGET_XOP to TARGET_XOP_P (opts->...).
4241 Change TARGET_LWP to TARGET_LWP_P (opts->...).
4242 Change TARGET_ABM to TARGET_ABM_P (opts->...).
4243 Change TARGET_BMI to TARGET_BMI_P (opts->...).
4244 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
4245 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
4246 Change TARGET_TBM to TARGET_TBM_P (opts->...).
4247 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
4248 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
4249 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
4250 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
4251 Change TARGET_AES to TARGET_AES_P (opts->...).
4252 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
4253 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
4254 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
4255 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
4256 Change TARGET_F16C to TARGET_F16C_P (opts->...).
4257 Change TARGET_RTM to TARGET_RTM_P (opts->...).
4258 Change TARGET_HLE to TARGET_HLE_P (opts->...).
4259 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
4260 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
4261 Change TARGET_ADX to TARGET_ADX_P (opts->...).
4262 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
4263 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
4264 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
4265 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
4266 Change TARGET_X32 to TARGET_X32_P (opts->...).
4267 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
4268 Change TARGET_FLOAT_RETURNS_IN_80387 to
4269 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
4270 (ix86_function_specific_save): New parameter. Use opts-> fields
4271 to replace global fields.
4272 (ix86_function_specific_restore): Ditto.
4273 (ix86_valid_target_attribute_inner_p): New parameters.
4275 Fix call to ix86_handle_option and set_option.
4276 (ix86_valid_target_attribute_tree): New parameters.
4277 Change global_options to access opts.
4278 Change global_options_set to access opts_set.
4279 Fix call to ix86_valid_target_attribute_inner_p.
4280 Change ix86_tune_string to access opts->x_ix86_tune_string.
4281 Change ix86_arch_string to access opts->x_ix86_arch_string.
4282 Change ix86_fpmath to access opts->x_ix86_fpmath
4283 Fix call to ix86_option_override_internal.
4284 Fix call to ix86_add_new_builtins.
4285 Fix calls to build_optimization_node and build_target_option_node.
4286 (ix86_valid_target_attribute_p): Remove access to global_options.
4287 Use new gcc_options structure func_options.
4288 Fix call to ix86_valid_target_attribute_tree.
4289 Fix call to build_optimization_node.
4290 (get_builtin_code_for_version): Fix call to
4291 ix86_valid_target_attribute_tree.
4293 2013-10-15 David Malcolm <dmalcolm@redhat.com>
4295 * Makefile.in (PICFLAG): New.
4296 (enable_host_shared): New.
4297 (INTERNAL_CFLAGS): Use PICFLAG.
4298 (LIBIBERTY): Use pic build of libiberty.a if configured with
4299 --enable-host-shared.
4300 * configure.ac: Add --enable-host-shared, setting up new
4302 * configure: Regenerate.
4303 * doc/install.texi (--enable-shared): Add note contrasting it with ...
4304 (--enable-host-shared): New option.
4306 2013-10-15 Richard Biener <rguenther@suse.de>
4308 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
4309 for built-in functions.
4311 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
4313 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
4314 (optimize_range_tests_1): New function,
4315 extracted from optimize_range_tests.
4316 (optimize_range_tests_xor): Similarly.
4317 (optimize_range_tests_diff): New function.
4318 (optimize_range_tests): Use optimize_range_tests_1.
4320 2013-10-15 Cong Hou <congh@google.com>
4322 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
4323 requirement of the reduction pattern so that one operand of the
4324 reduction operation can come from outside of the loop.
4326 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4328 * config/arm/neon-schedgen.ml: Remove.
4329 * config/arm/cortex-a9-neon.md: Remove comment regarding
4332 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4334 * config/arm/types: Remove old neon types.
4336 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4338 * config/arm/cortex-a7.md
4339 (cortex_a7_neon_type): New.
4340 (cortex_a7_neon_mul): Update for new types.
4341 (cortex_a7_neon_mla): Likewise.
4342 (cortex_a7_neon): Likewise.
4344 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4346 * config/arm/cortex-a15-neon.md
4347 (cortex_a15_neon_type): New,
4349 (cortex_a15_neon_int_1): Remove.
4350 (cortex_a15_neon_int_2): Likewise.
4351 (cortex_a15_neon_int_3): Likewise.
4352 (cortex_a15_neon_int_4): Likewise.
4353 (cortex_a15_neon_int_5): Likewise.
4354 (cortex_a15_neon_vqneg_vqabs): Likewise.
4355 (cortex_a15_neon_vmov): Likewise.
4356 (cortex_a15_neon_vaba): Likewise.
4357 (cortex_a15_neon_vaba_qqq): Likewise.
4358 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4359 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4360 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4362 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4363 (cortex_a15_neon_mla_qqq_8_16): Likewise.
4364 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
4365 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4366 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4367 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
4368 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4369 (cortex_a15_neon_shift_1): Likewise.
4370 (cortex_a15_neon_shift_2): Likewise.
4371 (cortex_a15_neon_shift_3): Likewise.
4372 (cortex_a15_neon_vshl_ddd): Likewise.
4373 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4374 (cortex_a15_neon_vsra_vrsra): Likewise.
4375 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4376 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4377 (cortex_a15_neon_bp_3cycle): Likewise.
4378 (cortex_a15_neon_ldm_2): Likewise.
4379 (cortex_a15_neon_stm_2): Likewise.
4380 (cortex_a15_neon_mcr): Likewise.
4381 (cortex_a15_neon_mrc): Likewise.
4382 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
4383 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
4384 (cortex_a15_neon_fp_vmul_ddd): Likewise.
4385 (cortex_a15_neon_fp_vmul_qqd): Likewise.
4386 (cortex_a15_neon_fp_vmla_ddd): Likewise.
4387 (cortex_a15_neon_fp_vmla_qqq): Likewise.
4388 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
4389 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
4390 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4391 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4392 (cortex_a15_neon_bp_simple): Likewise.
4393 (cortex_a15_neon_bp_2cycle): Likewise.
4394 (cortex_a15_neon_bp_3cycle): Likewise.
4395 (cortex_a15_neon_vld1_1_2_regs): Likewise.
4396 (cortex_a15_neon_vld1_3_4_regs): Likewise.
4397 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4398 (cortex_a15_neon_vld2_4_regs): Likewise.
4399 (cortex_a15_neon_vld3_vld4): Likewise.
4400 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4401 (cortex_a15_neon_vst1_3_4_regs): Likewise.
4402 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
4403 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
4404 (cortex_a15_neon_vst3_vst4): Rename to...
4405 (cortex_a15_neon_vst4): This, update for new attributes.
4406 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
4407 (cortex_a15_neon_vld3_vld4_lane): Likewise.
4408 (cortex_a15_neon_vst1_vst2_lane): Likewise.
4409 (cortex_a15_neon_vst3_vst4_lane): Likewise.
4410 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
4411 (cortex_a15_neon_ldm_2): Likewise.
4412 (cortex_a15_neon_stm_2): Likewise.
4413 (cortex_a15_neon_mcr): Likewise.
4414 (cortex_a15_neon_mcr_2_mcrr): Likewise.
4415 (cortex_a15_neon_mrc): Likewise.
4416 (cortex_a15_neon_mrrc): Likewise.
4418 (cortex_a15_neon_abd): New.
4419 (cortex_a15_neon_abd_q): Likewise.
4420 (cortex_a15_neon_aba): Likewise.
4421 (cortex_a15_neon_aba_q): Likewise.
4422 (cortex_a15_neon_acc): Likewise.
4423 (cortex_a15_neon_acc_q): Likewise.
4424 (cortex_a15_neon_arith_basic): Likewise.
4425 (cortex_a15_neon_arith_complex): Likewise.
4426 (cortex_a15_neon_multiply): Likewise.
4427 (cortex_a15_neon_multiply_q): Likewise.
4428 (cortex_a15_neon_mla): Likewise.
4429 (cortex_a15_neon_mla_q): Likewise.
4430 (cortex_a15_neon_sat_mla_long): Likewise.
4431 (cortex_a15_neon_shift_acc): Likewise.
4432 (cortex_a15_neon_shift_imm_basic): Likewise.
4433 (cortex_a15_neon_shift_imm_complex): Likewise.
4434 (cortex_a15_neon_shift_reg_basic): Likewise.
4435 (cortex_a15_neon_shift_reg_basic_q): Likewise.
4436 (cortex_a15_neon_shift_reg_complex): Likewise.
4437 (cortex_a15_neon_shift_reg_complex_q): Likewise.
4438 (cortex_a15_neon_fp_negabs): Likewise
4439 (cortex_a15_neon_fp_arith): Likewise
4440 (cortex_a15_neon_fp_arith_q): Likewise
4441 (cortex_a15_neon_fp_cvt_int): Likewise
4442 (cortex_a15_neon_fp_cvt_int_q): Likewise
4443 (cortex_a15_neon_fp_cvt_16): Likewise
4444 (cortex_a15_neon_fp_mul): Likewise
4445 (cortex_a15_neon_fp_mul_q): Likewise
4446 (cortex_a15_neon_fp_mla): Likewise
4447 (cortex_a15_neon_fp_mla_q): Likewise
4448 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
4449 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
4450 (cortex_a15_neon_bitops): Likewise.
4451 (cortex_a15_neon_bitops_q): Likewise.
4452 (cortex_a15_neon_from_gp): Likewise.
4453 (cortex_a15_neon_from_gp_q): Likewise.
4454 (cortex_a15_neon_tbl3_tbl4): Likewise.
4455 (cortex_a15_neon_zip_q): Likewise.
4456 (cortex_a15_neon_to_gp): Likewise.
4457 (cortex_a15_neon_load_a): Likewise.
4458 (cortex_a15_neon_load_b): Likewise.
4459 (cortex_a15_neon_load_c): Likewise.
4460 (cortex_a15_neon_load_d): Likewise.
4461 (cortex_a15_neon_load_e): Likewise.
4462 (cortex_a15_neon_load_f): Likewise.
4463 (cortex_a15_neon_store_a): Likewise.
4464 (cortex_a15_neon_store_b): Likewise.
4465 (cortex_a15_neon_store_c): Likewise.
4466 (cortex_a15_neon_store_d): Likewise.
4467 (cortex_a15_neon_store_e): Likewise.
4468 (cortex_a15_neon_store_f): Likewise.
4469 (cortex_a15_neon_store_g): Likewise.
4470 (cortex_a15_neon_store_h): Likewise.
4471 (cortex_a15_vfp_to_from_gp): Likewise.
4473 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4475 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
4477 (cortex_a9_neon_vshl_ddd): Remove.
4478 (cortex_a9_neon_vst3_vst4): Likewise.
4479 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4481 (cortex_a9_neon_bit_ops_q): New.
4483 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
4484 (cortex_a9_neon_int_2): Likewise.
4485 (cortex_a9_neon_int_3): Likewise.
4486 (cortex_a9_neon_int_4): Likewise.
4487 (cortex_a9_neon_int_5): Likewise.
4488 (cortex_a9_neon_vqneg_vqabs): Likewise.
4489 (cortex_a9_neon_vmov): Likewise.
4490 (cortex_a9_neon_vaba): Likewise.
4491 (cortex_a9_neon_vaba_qqq): Likewise.
4492 (cortex_a9_neon_shift_1): Likewise.
4493 (cortex_a9_neon_shift_2): Likewise.
4494 (cortex_a9_neon_shift_3): Likewise.
4495 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4496 (cortex_a9_neon_vsra_vrsra): Likewise.
4497 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4498 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4499 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4501 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4502 (cortex_a9_neon_mla_qqq_8_16): Likewise.
4503 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4505 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4506 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4507 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
4508 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4509 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
4510 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
4511 (cortex_a9_neon_fp_vsum): Likewise.
4512 (cortex_a9_neon_fp_vmul_ddd): Likewise.
4513 (cortex_a9_neon_fp_vmul_qqd): Likewise.
4514 (cortex_a9_neon_fp_vmla_ddd): Likewise.
4515 (cortex_a9_neon_fp_vmla_qqq): Likewise.
4516 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
4517 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
4518 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4519 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4520 (cortex_a9_neon_bp_simple): Likewise.
4521 (cortex_a9_neon_bp_2cycle): Likewise.
4522 (cortex_a9_neon_bp_3cycle): Likewise.
4523 (cortex_a9_neon_ldr): Likewise.
4524 (cortex_a9_neon_str): Likewise.
4525 (cortex_a9_neon_vld1_1_2_regs): Likewise.
4526 (cortex_a9_neon_vld1_3_4_regs): Likewise.
4527 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4528 (cortex_a9_neon_vld2_4_regs): Likewise.
4529 (cortex_a9_neon_vld3_vld4): Likewise.
4530 (cortex_a9_neon_vld1_vld2_lane): Likewise.
4531 (cortex_a9_neon_vld3_vld4_lane): Likewise.
4532 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
4533 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4534 (cortex_a9_neon_vst1_3_4_regs): Likewise.
4535 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
4536 (cortex_a9_neon_vst1_vst2_lane): Likewise.
4537 (cortex_a9_neon_vst3_vst4_lane): Likewise.
4538 (cortex_a9_neon_mcr): Likewise.
4539 (cortex_a9_neon_mcr_2_mcrr): Likewise.
4540 (cortex_a9_neon_mrc): Likewise.
4541 (cortex_a9_neon_mrrc): Likewise.
4543 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4545 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
4547 (cortex_a8_neon_vshl_ddd): Remove.
4548 (cortex_a8_neon_vst3_vst4): Likewise.
4549 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
4551 (cortex_a8_neon_bit_ops_q): New.
4553 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
4554 (cortex_a8_neon_int_2): Likewise..
4555 (cortex_a8_neon_int_3): Likewise.
4556 (cortex_a8_neon_int_5): Likewise.
4557 (cortex_a8_neon_vqneg_vqabs): Likewise.
4558 (cortex_a8_neon_int_4): Likewise.
4559 (cortex_a8_neon_vaba): Likewise.
4560 (cortex_a8_neon_vaba_qqq): Likewise.
4561 (cortex_a8_neon_shift_1): Likewise.
4562 (cortex_a8_neon_shift_2): Likewise.
4563 (cortex_a8_neon_shift_3): Likewise.
4564 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
4565 (cortex_a8_neon_vsra_vrsra): Likewise.
4566 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4567 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
4568 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
4570 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
4571 (cortex_a8_neon_mla_qqq_8_16): Likewise.
4572 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
4574 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
4575 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
4576 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
4577 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
4578 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
4579 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
4580 (cortex_a8_neon_fp_vsum): Likewise.
4581 (cortex_a8_neon_fp_vmul_ddd): Likewise.
4582 (cortex_a8_neon_fp_vmul_qqd): Likewise.
4583 (cortex_a8_neon_fp_vmla_ddd): Likewise.
4584 (cortex_a8_neon_fp_vmla_qqq): Likewise.
4585 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
4586 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
4587 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
4588 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
4589 (cortex_a8_neon_bp_simple): Likewise.
4590 (cortex_a8_neon_bp_2cycle): Likewise.
4591 (cortex_a8_neon_bp_3cycle): Likewise.
4592 (cortex_a8_neon_ldr): Likewise.
4593 (cortex_a8_neon_str): Likewise.
4594 (cortex_a8_neon_vld1_1_2_regs): Likewise.
4595 (cortex_a8_neon_vld1_3_4_regs): Likewise.
4596 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
4597 (cortex_a8_neon_vld2_4_regs): Likewise.
4598 (cortex_a8_neon_vld3_vld4): Likewise.
4599 (cortex_a8_neon_vld1_vld2_lane): Likewise.
4600 (cortex_a8_neon_vld3_vld4_lane): Likewise.
4601 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
4602 (cortex_a8_neon_vst1_3_4_regs): Likewise.
4603 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
4604 (cortex_a8_neon_vst1_vst2_lane): Likewise.
4605 (cortex_a8_neon_vst3_vst4_lane): Likewise.
4606 (cortex_a8_neon_mcr): Likewise.
4607 (cortex_a8_neon_mcr_2_mcrr): Likewise.
4608 (cortex_a8_neon_mrc): Likewise.
4609 (cortex_a8_neon_mrrc): Likewise.
4611 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4613 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
4615 * config/aarch64/aarch64-simd.md (neon_type): Remove.
4616 (aarch64_simd_dup<mode>): Add "type" attribute.
4617 (aarch64_dup_lane<mode>): Likewise.
4618 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
4619 (*aarch64_simd_mov<mode>): Likewise.
4620 (aarch64_simd_mov_from_<mode>low): Likewise.
4621 (aarch64_simd_mov_from_<mode>high): Likewise.
4622 (orn<mode>3): Likewise.
4623 (bic<mode>3): Likewise.
4624 (add<mode>3): Likewise.
4625 (sub<mode>3): Likewise.
4626 (mul<mode>3): Likewise.
4627 (*aarch64_mul3_elt<mode>): Likewise.
4628 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4629 (*aarch64_mul3_elt_to_128df): Likewise.
4630 (*aarch64_mul3_elt_to_64v2df): Likewise.
4631 (neg<mode>2): Likewise.
4632 (abs<mode>2): Likewise.
4633 (abd<mode>_3): Likewise.
4634 (aba<mode>_3): Likewise.
4635 (fabd<mode>_3): Likewise.
4636 (*fabd_scalar<mode>3): Likewise.
4637 (and<mode>3): Likewise.
4638 (ior<mode>3): Likewise.
4639 (xor<mode>3): Likewise.
4640 (one_cmpl<mode>2): Likewise.
4641 (aarch64_simd_vec_set<mode>): Likewise.
4642 (aarch64_simd_lshr<mode>): Likewise.
4643 (aarch64_simd_ashr<mode>): Likewise.
4644 (aarch64_simd_imm_shl<mode>): Likewise.
4645 (aarch64_simd_reg_sshl<mode): Likewise.
4646 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
4647 (aarch64_simd_reg_shl<mode>_signed): Likewise.
4648 (aarch64_simd_vec_setv2di): Likewise.
4649 (aarch64_simd_vec_set<mode>): Likewise.
4650 (aarch64_mla<mode>): Likewise.
4651 (*aarch64_mla_elt<mode>): Likewise.
4652 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4653 (aarch64_mls<mode>): Likewise.
4654 (*aarch64_mls_elt<mode>): Likewise.
4655 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4656 (<su><maxmin><mode>3): Likewise.
4657 (move_lo_quad_<mode>): Likewise.
4658 (aarch64_simd_move_hi_quad_<mode>): Likewise.
4659 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
4660 (vec_pack_trunc_<mode>): Likewise.
4661 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
4662 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
4663 (*aarch64_<su>mlal_lo<mode>): Likewise.
4664 (*aarch64_<su>mlal_hi<mode>): Likewise.
4665 (*aarch64_<su>mlsl_lo<mode>): Likewise.
4666 (*aarch64_<su>mlsl_hi<mode>): Likewise.
4667 (*aarch64_<su>mlal<mode>): Likewise.
4668 (*aarch64_<su>mlsl<mode>): Likewise.
4669 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
4670 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
4671 (add<mode>3): Likewise.
4672 (sub<mode>3): Likewise.
4673 (mul<mode>3): Likewise.
4674 (div<mode>3): Likewise.
4675 (neg<mode>2): Likewise.
4676 (abs<mode>2): Likewise.
4677 (fma<mode>4): Likewise.
4678 (*aarch64_fma4_elt<mode>): Likewise.
4679 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4680 (*aarch64_fma4_elt_to_128df): Likewise.
4681 (*aarch64_fma4_elt_to_64v2df): Likewise.
4682 (fnma<mode>4): Likewise.
4683 (*aarch64_fnma4_elt<mode>): Likewise.
4684 (*aarch64_fnma4_elt_<vswap_width_name><mode>
4685 (*aarch64_fnma4_elt_to_128df): Likewise.
4686 (*aarch64_fnma4_elt_to_64v2df): Likewise.
4687 (<frint_pattern><mode>2): Likewise.
4688 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
4689 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
4690 (vec_unpacks_lo_v4sf): Likewise.
4691 (aarch64_float_extend_lo_v2df): Likewise.
4692 (vec_unpacks_hi_v4sf): Likewise.
4693 (aarch64_float_truncate_lo_v2sf): Likewise.
4694 (aarch64_float_truncate_hi_v4sf): Likewise.
4695 (aarch64_vmls<mode>): Likewise.
4696 (<su><maxmin><mode>3): Likewise.
4697 (<maxmin_uns><mode>3): Likewise.
4698 (reduc_<sur>plus_<mode>): Likewise.
4699 (reduc_<sur>plus_v2di): Likewise.
4700 (reduc_<sur>plus_v2si): Likewise.
4701 (reduc_<sur>plus_<mode>): Likewise.
4702 (aarch64_addpv4sf): Likewise.
4703 (clz<mode>2): Likewise.
4704 (reduc_<maxmin_uns>_<mode>): Likewise.
4705 (reduc_<maxmin_uns>_v2di): Likewise.
4706 (reduc_<maxmin_uns>_v2si): Likewise.
4707 (reduc_<maxmin_uns>_<mode>): Likewise.
4708 (reduc_<maxmin_uns>_v4sf): Likewise.
4709 (aarch64_simd_bsl<mode>_internal): Likewise.
4710 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
4711 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
4712 (aarch64_get_lane<mode>): Likewise.
4713 (*aarch64_combinez<mode>): Likewise.
4714 (aarch64_combine<mode>): Likewise.
4715 (aarch64_simd_combine<mode>): Likewise.
4716 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
4717 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
4718 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
4719 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
4720 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
4721 (aarch64_<sur>h<addsub><mode>): Likewise.
4722 (aarch64_<sur><addsub>hn<mode>): Likewise.
4723 (aarch64_<sur><addsub>hn2<mode>): Likewise.
4724 (aarch64_pmul<mode>): Likewise.
4725 (aarch64_<su_optab><optab><mode>): Likewise.
4726 (aarch64_<sur>qadd<mode>): Likewise.
4727 (aarch64_sqmovun<mode>): Likewise.
4728 (aarch64_<sur>qmovn<mode>): Likewise.
4729 (aarch64_s<optab><mode>): Likewise.
4730 (aarch64_sq<r>dmulh<mode>): Likewise.
4731 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4732 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
4733 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
4734 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
4735 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4736 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
4737 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
4738 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
4739 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
4740 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
4741 (aarch64_sqdmull<mode>): Likewise.
4742 (aarch64_sqdmull_lane<mode>_internal): Likewise.
4743 (aarch64_sqdmull_n<mode>): Likewise.
4744 (aarch64_sqdmull2<mode>_internal): Likewise.
4745 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
4746 (aarch64_sqdmull2_n<mode>_internal): Likewise.
4747 (aarch64_<sur>shl<mode>): Likewise.
4748 (aarch64_<sur>q<r>shl<mode>
4749 (aarch64_<sur>shll_n<mode>): Likewise.
4750 (aarch64_<sur>shll2_n<mode>): Likewise.
4751 (aarch64_<sur>shr_n<mode>): Likewise.
4752 (aarch64_<sur>sra_n<mode>): Likewise.
4753 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
4754 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
4755 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
4756 (aarch64_cm<optab><mode>): Likewise.
4757 (aarch64_cm<optab>di): Likewise.
4758 (aarch64_cm<optab><mode>): Likewise.
4759 (aarch64_cm<optab>di): Likewise.
4760 (aarch64_cmtst<mode>): Likewise.
4761 (aarch64_cmtstdi): Likewise.
4762 (aarch64_cm<optab><mode>): Likewise.
4763 (*aarch64_fac<optab><mode>): Likewise.
4764 (aarch64_addp<mode>): Likewise.
4765 (aarch64_addpdi): Likewise.
4766 (sqrt<mode>2): Likewise.
4767 (vec_load_lanesoi<mode>): Likewise.
4768 (vec_store_lanesoi<mode>): Likewise.
4769 (vec_load_lanesci<mode>): Likewise.
4770 (vec_store_lanesci<mode>): Likewise.
4771 (vec_load_lanesxi<mode>): Likewise.
4772 (vec_store_lanesxi<mode>): Likewise.
4773 (*aarch64_mov<mode>): Likewise.
4774 (aarch64_ld2<mode>_dreg): Likewise.
4775 (aarch64_ld2<mode>_dreg): Likewise.
4776 (aarch64_ld3<mode>_dreg): Likewise.
4777 (aarch64_ld3<mode>_dreg): Likewise.
4778 (aarch64_ld4<mode>_dreg): Likewise.
4779 (aarch64_ld4<mode>_dreg): Likewise.
4780 (aarch64_tbl1<mode>): Likewise.
4781 (aarch64_tbl2v16qi): Likewise.
4782 (aarch64_combinev16qi): Likewise.
4783 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
4784 (aarch64_st2<mode>_dreg): Likewise.
4785 (aarch64_st2<mode>_dreg): Likewise.
4786 (aarch64_st3<mode>_dreg): Likewise.
4787 (aarch64_st3<mode>_dreg): Likewise.
4788 (aarch64_st4<mode>_dreg): Likewise.
4789 (aarch64_st4<mode>_dreg): Likewise.
4790 (*aarch64_simd_ld1r<mode>): Likewise.
4791 (aarch64_frecpe<mode>): Likewise.
4792 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
4793 (aarch64_frecps<mode>): Likewise.
4795 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
4797 * config/arm/iterators.md (V_elem_ch): New.
4799 (VQH_type): Likewise.
4800 * config/arm/arm.md (is_neon_type): New.
4801 (conds): Use is_neon_type.
4802 (anddi3_insn): Update type attribute.
4803 (xordi3_insn): Likewise.
4804 (one_cmpldi2): Likewise.
4805 * config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
4806 * config/arm/neon.md (neon_mov): Update type attribute.
4807 (*movmisalign<mode>_neon_store): Likewise.
4808 (*movmisalign<mode>_neon_load): Likewise.
4809 (vec_set<mode>_internal): Likewise.
4810 (vec_set<mode>_internal): Likewise.
4811 (vec_setv2di_internal): Likewise.
4812 (vec_extract<mode>): Likewise.
4813 (vec_extract<mode>): Likewise.
4814 (vec_extractv2di): Likewise.
4815 (*add<mode>3_neon): Likewise.
4816 (adddi3_neon): Likewise.
4817 (*sub<mode>3_neon): Likewise.
4818 (subdi3_neon): Likewise.
4819 (fma<VCVTF:mode>4): Likewise.
4820 (fma<VCVTF:mode>4_intrinsic): Likewise.
4821 (*fmsub<VCVTF:mode>4): Likewise.
4822 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
4823 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
4824 (ior<mode>3): Likewise.
4825 (and<mode>3): Likewise.
4826 (orn<mode>3_neon): Likewise.
4827 (orndi3_neon): Likewise.
4828 (bic<mode>3_neon): Likewise.
4829 (bicdi3_neon): Likewise.
4830 (xor<mode>3): Likewise.
4831 (one_cmpl<mode>2): Likewise.
4832 (abs<mode>2): Likewise.
4833 (neg<mode>2): Likewise.
4834 (negdi2_neon): Likewise.
4835 (*umin<mode>3_neon): Likewise.
4836 (*umax<mode>3_neon): Likewise.
4837 (*smin<mode>3_neon): Likewise.
4838 (*smax<mode>3_neon): Likewise.
4839 (vashl<mode>3): Likewise.
4840 (vashr<mode>3_imm): Likewise.
4841 (vlshr<mode>3_imm): Likewise.
4842 (ashl<mode>3_signed): Likewise.
4843 (ashl<mode>3_unsigned): Likewise.
4844 (neon_load_count): Likewise.
4845 (ashldi3_neon_noclobber): Likewise.
4846 (ashldi3_neon): Likewise.
4847 (signed_shift_di3_neon): Likewise.
4848 (unsigned_shift_di3_neon): Likewise.
4849 (ashrdi3_neon_imm_noclobber): Likewise.
4850 (lshrdi3_neon_imm_noclobber): Likewise.
4851 (<shift>di3_neon): Likewise.
4852 (widen_ssum<mode>3): Likewise.
4853 (widen_usum<mode>3): Likewise.
4854 (quad_halves_<code>v4si): Likewise.
4855 (quad_halves_<code>v4sf): Likewise.
4856 (quad_halves_<code>v8hi): Likewise.
4857 (quad_halves_<code>v16qi): Likewise.
4858 (reduc_splus_v2di): Likewise.
4859 (neon_vpadd_internal<mode>): Likewise.
4860 (neon_vpsmin<mode>): Likewise.
4861 (neon_vpsmax<mode>): Likewise.
4862 (neon_vpumin<mode>): Likewise.
4863 (neon_vpumax<mode>): Likewise.
4864 (*ss_add<mode>_neon): Likewise.
4865 (*us_add<mode>_neon): Likewise.
4866 (*ss_sub<mode>_neon): Likewise.
4867 (*us_sub<mode>_neon): Likewise.
4868 (neon_vadd<mode>_unspec): Likewise.
4869 (neon_vaddl<mode>): Likewise.
4870 (neon_vaddw<mode>): Likewise.
4871 (neon_vhadd<mode>): Likewise.
4872 (neon_vqadd<mode>): Likewise.
4873 (neon_vaddhn<mode>): Likewise.
4874 (neon_vmul<mode>): Likewise.
4875 (neon_vfms<VCVTF:mode>): Likewise.
4876 (neon_vmlal<mode>): Likewise.
4877 (neon_vmls<mode>): Likewise.
4878 (neon_vmlsl<mode>): Likewise.
4879 (neon_vqdmulh<mode>): Likewise.
4880 (neon_vqdmlal<mode>): Likewise.
4881 (neon_vqdmlsl<mode>): Likewise.
4882 (neon_vmull<mode>): Likewise.
4883 (neon_vqdmull<mode>): Likewise.
4884 (neon_vsub<mode>_unspec): Likewise.
4885 (neon_vsubl<mode>): Likewise.
4886 (neon_vsubw<mode>): Likewise.
4887 (neon_vqsub<mode>): Likewise.
4888 (neon_vhsub<mode>): Likewise.
4889 (neon_vsubhn<mode>): Likewise.
4890 (neon_vceq<mode>): Likewise.
4891 (neon_vcge<mode>): Likewise.
4892 (neon_vcgeu<mode>): Likewise.
4893 (neon_vcgt<mode>): Likewise.
4894 (neon_vcgtu<mode>): Likewise.
4895 (neon_vcle<mode>): Likewise.
4896 (neon_vclt<mode>): Likewise.
4897 (neon_vcage<mode>): Likewise.
4898 (neon_vcagt<mode>): Likewise.
4899 (neon_vtst<mode>): Likewise.
4900 (neon_vabd<mode>): Likewise.
4901 (neon_vabdl<mode>): Likewise.
4902 (neon_vaba<mode>): Likewise.
4903 (neon_vabal<mode>): Likewise.
4904 (neon_vmax<mode>): Likewise.
4905 (neon_vmin<mode>): Likewise.
4906 (neon_vpaddl<mode>): Likewise.
4907 (neon_vpadal<mode>): Likewise.
4908 (neon_vpmax<mode>): Likewise.
4909 (neon_vpmin<mode>): Likewise.
4910 (neon_vrecps<mode>): Likewise.
4911 (neon_vrsqrts<mode>): Likewise.
4912 (neon_vqabs<mode>): Likewise.
4913 (neon_vqneg<mode>): Likewise.
4914 (neon_vcls<mode>): Likewise.
4915 (clz<mode>2): Likewise.
4916 (popcount<mode>2): Likewise.
4917 (neon_vrecpe<mode>): Likewise.
4918 (neon_vrsqrte<mode>): Likewise.
4919 (neon_vget_lane<mode>_sext_internal): Likewise.
4920 (neon_vget_lane<mode>_zext_internal): Likewise.
4921 (neon_vdup_n<mode>): Likewise.
4922 (neon_vdup_n<mode>): Likewise.
4923 (neon_vdup_nv2di): Likewise.
4924 (neon_vdup_lane<mode>_interal): Likewise.
4925 (*neon_vswp<mode>): Likewise.
4926 (neon_vcombine<mode>): Likewise.
4927 (float<mode><V_cvtto>2): Likewise.
4928 (floatuns<mode><V_cvtto>2): Likewise.
4929 (fix_trunc<mode><V_cvtto>2): Likewise.
4930 (fixuns_trunc<mode><V_cvtto>2
4931 (neon_vcvt<mode>): Likewise.
4932 (neon_vcvt<mode>): Likewise.
4933 (neon_vcvtv4sfv4hf): Likewise.
4934 (neon_vcvtv4hfv4sf): Likewise.
4935 (neon_vcvt_n<mode>): Likewise.
4936 (neon_vcvt_n<mode>): Likewise.
4937 (neon_vmovn<mode>): Likewise.
4938 (neon_vqmovn<mode>): Likewise.
4939 (neon_vqmovun<mode>): Likewise.
4940 (neon_vmovl<mode>): Likewise.
4941 (neon_vmul_lane<mode>): Likewise.
4942 (neon_vmul_lane<mode>): Likewise.
4943 (neon_vmull_lane<mode>): Likewise.
4944 (neon_vqdmull_lane<mode>): Likewise.
4945 (neon_vqdmulh_lane<mode>): Likewise.
4946 (neon_vqdmulh_lane<mode>): Likewise.
4947 (neon_vmla_lane<mode>): Likewise.
4948 (neon_vmla_lane<mode>): Likewise.
4949 (neon_vmlal_lane<mode>): Likewise.
4950 (neon_vqdmlal_lane<mode>): Likewise.
4951 (neon_vmls_lane<mode>): Likewise.
4952 (neon_vmls_lane<mode>): Likewise.
4953 (neon_vmlsl_lane<mode>): Likewise.
4954 (neon_vqdmlsl_lane<mode>): Likewise.
4955 (neon_vext<mode>): Likewise.
4956 (neon_vrev64<mode>): Likewise.
4957 (neon_vrev32<mode>): Likewise.
4958 (neon_vrev16<mode>): Likewise.
4959 (neon_vbsl<mode>_internal): Likewise.
4960 (neon_vshl<mode>): Likewise.
4961 (neon_vqshl<mode>): Likewise.
4962 (neon_vshr_n<mode>): Likewise.
4963 (neon_vshrn_n<mode>): Likewise.
4964 (neon_vqshrn_n<mode>): Likewise.
4965 (neon_vqshrun_n<mode>): Likewise.
4966 (neon_vshl_n<mode>): Likewise.
4967 (neon_vqshl_n<mode>): Likewise.
4968 (neon_vqshlu_n<mode>): Likewise.
4969 (neon_vshll_n<mode>): Likewise.
4970 (neon_vsra_n<mode>): Likewise.
4971 (neon_vsri_n<mode>): Likewise.
4972 (neon_vsli_n<mode>): Likewise.
4973 (neon_vtbl1v8qi): Likewise.
4974 (neon_vtbl2v8qi): Likewise.
4975 (neon_vtbl3v8qi): Likewise.
4976 (neon_vtbl4v8qi): Likewise.
4977 (neon_vtbl1v16qi): Likewise.
4978 (neon_vtbl2v16qi): Likewise.
4979 (neon_vcombinev16qi): Likewise.
4980 (neon_vtbx1v8qi): Likewise.
4981 (neon_vtbx2v8qi): Likewise.
4982 (neon_vtbx3v8qi): Likewise.
4983 (neon_vtbx4v8qi): Likewise.
4984 (*neon_vtrn<mode>_insn): Likewise.
4985 (*neon_vzip<mode>_insn): Likewise.
4986 (*neon_vuzp<mode>_insn): Likewise.
4987 (neon_vld1<mode>): Likewise.
4988 (neon_vld1_lane<mode>): Likewise.
4989 (neon_vld1_lane<mode>): Likewise.
4990 (neon_vld1_dup<mode>): Likewise.
4991 (neon_vld1_dup<mode>): Likewise.
4992 (neon_vld1_dupv2di): Likewise.
4993 (neon_vst1<mode>): Likewise.
4994 (neon_vst1_lane<mode>): Likewise.
4995 (neon_vst1_lane<mode>): Likewise.
4996 (neon_vld2<mode>): Likewise.
4997 (neon_vld2<mode>): Likewise.
4998 (neon_vld2_lane<mode>): Likewise.
4999 (neon_vld2_lane<mode>): Likewise.
5000 (neon_vld2_dup<mode>): Likewise.
5001 (neon_vst2<mode>): Likewise.
5002 (neon_vst2<mode>): Likewise.
5003 (neon_vst2_lane<mode>): Likewise.
5004 (neon_vst2_lane<mode>): Likewise.
5005 (neon_vld3<mode>): Likewise.
5006 (neon_vld3qa<mode>): Likewise.
5007 (neon_vld3qb<mode>): Likewise.
5008 (neon_vld3_lane<mode>): Likewise.
5009 (neon_vld3_lane<mode>): Likewise.
5010 (neon_vld3_dup<mode>): Likewise.
5011 (neon_vst3<mode>): Likewise.
5012 (neon_vst3qa<mode>): Likewise.
5013 (neon_vst3qb<mode>): Likewise.
5014 (neon_vst3_lane<mode>): Likewise.
5015 (neon_vst3_lane<mode>): Likewise.
5016 (neon_vld4<mode>): Likewise.
5017 (neon_vld4qa<mode>): Likewise.
5018 (neon_vld4qb<mode>): Likewise.
5019 (neon_vld4_lane<mode>): Likewise.
5020 (neon_vld4_lane<mode>): Likewise.
5021 (neon_vld4_dup<mode>): Likewise.
5022 (neon_vst4<mode>): Likewise.
5023 (neon_vst4qa<mode>): Likewise.
5024 (neon_vst4qb<mode>): Likewise.
5025 (neon_vst4_lane<mode>): Likewise.
5026 (neon_vst4_lane<mode>): Likewise.
5027 (neon_vec_unpack<US>_lo_<mode>): Likewise.
5028 (neon_vec_unpack<US>_hi_<mode>): Likewise.
5029 (neon_vec_<US>mult_lo_<mode>): Likewise.
5030 (neon_vec_<US>mult_hi_<mode>): Likewise.
5031 (neon_vec_<US>shiftl_<mode>): Likewise.
5032 (neon_unpack<US>_<mode>): Likewise.
5033 (neon_vec_<US>mult_<mode>): Likewise.
5034 (vec_pack_trunc_<mode>): Likewise.
5035 (neon_vec_pack_trunc_<mode>): Likewise.
5036 (neon_vabd<mode>_2): Likewise.
5037 (neon_vabd<mode>_3): Likewise.
5039 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5041 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
5042 (load_pair): Update type attribute.
5043 (store_pair): Update type attribute.
5044 * config/aarch64/iterators.md (q): New.
5046 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
5048 * config/arm/types.md: Add new types for Neon insns.
5050 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5051 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5052 Sergey Lega <sergey.s.lega@intel.com>
5053 Anna Tikhonova <anna.tikhonova@intel.com>
5054 Ilya Tocar <ilya.tocar@intel.com>
5055 Andrey Turetskiy <andrey.turetskiy@intel.com>
5056 Ilya Verbin <ilya.verbin@intel.com>
5057 Kirill Yukhin <kirill.yukhin@intel.com>
5058 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5060 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
5061 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
5062 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
5064 (srcp14<mode>): Ditto.
5065 (rsqrt14<mode>): Ditto.
5066 (rsqrt14<mode>): Ditto.
5067 (avx512f_vmscalef<mode>): Ditto.
5068 (avx512f_scalef<mode>): Ditto.
5069 (avx512f_getexp<mode>): Ditto.
5070 (avx512f_sgetexp<mode>): Ditto.
5071 (avx512f_fixupimm<mode>): Ditto.
5072 (avx512f_sfixupimm<mode>): Ditto.
5073 (avx512f_rndscale<mode>): Ditto.
5074 (*avx512er_exp2<mode>): Ditto.
5075 (*avx512er_rcp28<mode>): Ditto.
5076 (avx512er_rsqrt28<mode>): Ditto.
5077 (avx512f_getmant<mode>): Ditto.
5078 (avx512f_getmant<mode>): Ditto.
5079 (avx512f_rndscale<mode>): Fix formatting.
5081 2013-10-15 Martin Jambor <mjambor@suse.cz>
5083 * ipa-utils.h (ipa_edge_within_scc): Declare.
5084 * ipa-cp.c (edge_within_scc): Moved...
5085 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
5087 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5088 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5089 Sergey Lega <sergey.s.lega@intel.com>
5090 Anna Tikhonova <anna.tikhonova@intel.com>
5091 Ilya Tocar <ilya.tocar@intel.com>
5092 Andrey Turetskiy <andrey.turetskiy@intel.com>
5093 Ilya Verbin <ilya.verbin@intel.com>
5094 Kirill Yukhin <kirill.yukhin@intel.com>
5095 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5097 * config/i386/predicates.md (const_8_to_15_operand): New.
5098 (const_16_to_31_operand): Ditto.
5099 * config/i386/sse.md (V8FI): New.
5101 (reduc_splus_v8df): Ditto.
5102 (reduc_splus_v16sf): Ditto.
5103 (avx512f_vextract<shuffletype>32x4_1): Ditto.
5104 (vec_extract_hi_<mode>): Ditto.
5105 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
5106 (vec_set_lo_<mode>): Ditto.
5107 (vec_set_hi_<mode>): Ditto.
5108 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
5109 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
5110 (avx512f_pshufd_1): Ditto.
5111 (avx512f_broadcast<mode>): Ditto.
5112 (avx512f_broadcast<mode>): Ditto.
5113 (define_split): Split vec_extract_lo into move.
5114 (ssequartermode): Ditto.
5115 (ssedoublemode): Extened with wider modes.
5116 (vec_extract_lo_<mode>): Ditto.
5118 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5119 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5120 Sergey Lega <sergey.s.lega@intel.com>
5121 Anna Tikhonova <anna.tikhonova@intel.com>
5122 Ilya Tocar <ilya.tocar@intel.com>
5123 Andrey Turetskiy <andrey.turetskiy@intel.com>
5124 Ilya Verbin <ilya.verbin@intel.com>
5125 Kirill Yukhin <kirill.yukhin@intel.com>
5126 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5128 * config/i386/predicates.md (register_or_constm1_operand): New.
5129 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
5130 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
5131 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
5132 UNSPEC_SCATTER_PREFETCH
5134 (avx512f_ucmp<mode>3): Ditto.
5135 (avx512f_vternlog<mode>): Ditto.
5136 (avx512f_align<mode>): Ditto.
5137 (<shift_insn><mode>3): Ditto.
5138 (avx512f_<rotate>v<mode>): Ditto.
5139 (avx512f_<rotate><mode>): Ditto.
5140 (avx512f_eq<mode>3): Ditto.
5141 (avx512f_eq<mode>3_1): Ditto.
5142 (avx512f_gt<mode>3): Ditto.
5143 (avx512f_testm<mode>3): Ditto.
5144 (avx512f_testnm<mode>3): Ditto.
5145 (avx512pf_gatherpf<mode>): Ditto.
5146 (*avx512pf_gatherpf<mode>_mask): Ditto.
5147 (*avx512pf_gatherpf<mode>): Ditto.
5148 (avx512pf_scatterpf<mode>): Ditto.
5149 (*avx512pf_scatterpf<mode>_mask): Ditto.
5150 (*avx512pf_scatterpf<mode>): Ditto.
5151 (avx512f_vec_dup_gpr<mode>): Ditto.
5152 (clz<mode>2): Ditto.
5153 (conflict<mode>): Ditto.
5154 (REDUC_SMINMAX_MODE): Extened with wider modes.
5155 (reduc_<code>_<mode>): Ditto.
5156 (vlshr<mode>3): Ditto.
5157 (vashl<mode>3): Ditto.
5159 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
5160 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5161 Sergey Lega <sergey.s.lega@intel.com>
5162 Anna Tikhonova <anna.tikhonova@intel.com>
5163 Ilya Tocar <ilya.tocar@intel.com>
5164 Andrey Turetskiy <andrey.turetskiy@intel.com>
5165 Ilya Verbin <ilya.verbin@intel.com>
5166 Kirill Yukhin <kirill.yukhin@intel.com>
5167 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5169 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
5172 (avx512fmaskmode): Ditto.
5173 (bcstscalarsuff): Ditto.
5174 (avx512f_blendm<mode>): Ditto.
5175 (cmp_imm_predicate): Ditto.
5176 (avx512f_cmp<mode>3): Ditto.
5177 (avx512f_vec_dup<mode>): Ditto.
5178 (avx512f_vec_dup_mem<mode>): Ditto.
5179 (avx512f_vpermi2var<mode>3): Ditto.
5180 (avx512f_vpermt2var<mode>3): Ditto.
5181 (vec_init<mode>): Ditto.
5182 (avx512f_gathersi<mode>): Ditto.
5183 (*avx512f_gathersi<mode>): Ditto.
5184 (*avx512f_gathersi<mode>_2): Ditto.
5185 (avx512f_gatherdi<mode>): Ditto.
5186 (*avx512f_gatherdi<mode>): Ditto.
5187 (*avx512f_gatherdi<mode>_2): Ditto.
5188 (avx512f_scattersi<mode>): Ditto.
5189 (*avx512f_scattersi<mode>): Ditto.
5190 (avx512f_scatterdi<mode>): Ditto.
5191 (*avx512f_scatterdi<mode>): Ditto.
5192 (sseintprefix): Extened with wider modes.
5193 (VEC_GATHER_IDXSI): Ditto.
5194 (VEC_GATHER_IDXDI): Ditto.
5195 (VEC_GATHER_SRCDI): Ditto.
5197 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
5198 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
5200 * config/arm/t-aprofile: New file.
5201 * config.gcc: Handle --with-multilib-list option.
5203 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
5205 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
5208 2013-10-15 Richard Biener <rguenther@suse.de>
5210 * tree-loop-distribution.c (build_empty_rdg): Inline into
5212 (rdg_flag_vertex): Inline into single user.
5213 (rdg_flag_vertex_and_dependent): Likewise.
5214 (build_rdg_partition_for_vertex): Remove processed bitmap.
5215 (rdg_build_partitions): Simplify.
5217 2013-10-15 Richard Biener <rguenther@suse.de>
5219 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
5220 Restructure forwarding through conversions and copies to
5221 avoid performing copy-propagation the wrong way. Adjust
5222 recursion invocations.
5223 (forward_propagate_addr_expr): Add argument stating if we
5224 are recursing from a single-use.
5225 (ssa_forward_propagate_and_combine): Adjust.
5227 2013-10-14 David Malcolm <dmalcolm@redhat.com>
5229 * dumpfile.h (gcc::dump_manager): New class, to hold state
5230 relating to dumpfile management.
5231 (get_dump_file_name): Remove in favor of method of dump_manager.
5232 (dump_initialized_p): Likewise.
5233 (dump_start): Likewise.
5234 (dump_finish): Likewise.
5235 (dump_switch_p): Likewise.
5236 (dump_register): Likewise.
5237 (get_dump_file_info): Likewise.
5238 * context.c (gcc::context::context): Construct the dump_manager
5240 * context.h (gcc::context::get_dumps): New.
5241 (gcc::context::m_dumps): New.
5242 * coverage.c (coverage_init): Port to dump_manager API.
5243 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
5244 (extra_dump_files_in_use): Likewise.
5245 (extra_dump_files_alloced): Likewise.
5246 (gcc::dump_manager::dump_manager): New.
5247 (dump_register): Convert to...
5248 (gcc::dump_manager::dump_register): ...method, replacing
5249 function-static next_dump with m_next_dump field.
5250 (get_dump_file_info): Convert to...
5251 (gcc::dump_manager::get_dump_file_info): ...method.
5252 (get_dump_file_name): Convert to...
5253 (gcc::dump_manager::get_dump_file_name): ...method.
5254 (dump_start): Convert to...
5255 (gcc::dump_manager::dump_start): ...method.
5256 (dump_finish): Convert to...
5257 (gcc::dump_manager::dump_finish): ...method.
5258 (dump_begin): Replace body with...
5259 (gcc::dump_manager::dump_begin): ...new method.
5260 (dump_phase_enabled_p): Convert to...
5261 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5262 (dump_phase_enabled_p): Convert to...
5263 (gcc::dump_manager::dump_phase_enabled_p): ...method.
5264 (dump_initialized_p): Convert to...
5265 (gcc::dump_manager::dump_initialized_p): ...method.
5266 (dump_flag_name): Replace body with...
5267 (gcc::dump_manager::dump_flag_name): ...new method.
5268 (dump_enable_all): Convert to...
5269 (gcc::dump_manager::dump_enable_all): ...new method.
5270 (opt_info_enable_passes): Convert to...
5271 (gcc::dump_manager::opt_info_enable_passes): ...new method.
5272 (dump_switch_p_1): Convert to...
5273 (gcc::dump_manager::dump_switch_p_1): ...new method.
5274 (dump_switch_p): Convert to...
5275 (gcc::dump_manager::dump_switch_p): ...new method.
5276 (opt_info_switch_p): Port to dump_manager API.
5277 (enable_rtl_dump_file): Likewise.
5278 * opts-global.c (handle_common_deferred_options): Port to new
5280 * passes.c (pass_manager::finish_optimization_passes): Likewise.
5281 (pass_manager::register_one_dump_file): Likewise.
5282 (pass_manager::register_pass): Likewise.
5283 (pass_init_dump_file): Likewise.
5284 (pass_fini_dump_file): Likewise.
5285 * statistics.c (statistics_early_init): Likewise.
5287 2013-10-14 Richard Biener <rguenther@suse.de>
5289 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
5290 iterative_hash_canonical_type, gimple_canonical_type_hash,
5291 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
5292 gimple_register_canonical_type, print_gimple_types_stats,
5293 free_gimple_type_tables): Move to lto/lto.c
5294 (gt-gimple.h): Do not include.
5295 * gimple.h (gimple_register_canonical_type,
5296 print_gimple_types_stats, free_gimple_type_tables): Remove.
5297 * Makefile.in (GTFILES): Remove gimple.c.
5299 2013-10-14 Travis Snoozy <quandary@remstate.com>
5302 * config/msp430/msp430.c (msp430_option_override): Correct thinko
5303 scanning for msp430x targets.
5305 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
5308 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
5309 (registers_ok_for_ldd_peep): Move around.
5310 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
5311 * config/sparc/sparc.md (widening peepholes): Use it.
5313 2013-10-14 Richard Biener <rguenther@suse.de>
5317 * gimple.c (iterative_hash_canonical_type): Make sure to
5318 record the hash into the correct hashtable slot.
5320 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
5322 PR rtl-optimization/58662
5323 * combine.c (try_combine): Take into account death nodes on I2 when
5324 splitting a PARALLEL of two independent SETs. Fix dump message.
5326 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
5329 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
5330 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
5331 * config/sh/t-sh (sh_treg_combine.o): New entry.
5332 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
5333 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
5334 (register_sh_passes): New function. Register sh_treg_combine pass.
5335 (sh_option_override): Invoke it.
5336 (sh_canonicalize_comparison): Handle op0_preserve_value.
5337 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
5338 opportunities. Canonicalize branch condition.
5339 (nott): Allow only if pseudos can be created for non-SH2A.
5341 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
5344 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
5345 (ix86_expand_movmem): Replace copy_addr_to_reg with
5346 ix86_copy_addr_to_reg.
5347 (ix86_expand_setmem): Likewise.
5349 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
5351 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
5352 provided by ix86_fp_compare_mode instead of CCFPUmode.
5354 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
5356 * config/aarch64/arm_neon.h
5357 (vtbx<1,3>_<psu>8): Fix register constriants.
5359 2013-10-11 Jeff Law <law@redhat.com>
5361 PR tree-optimization/58640
5362 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
5363 threading paths that cross over two loop entry points.
5365 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5367 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
5368 handle vector float as well.
5369 (*vsx_le_perm_load_v4si): Likewise.
5370 (*vsx_le_perm_store_v2di): Likewise.
5371 (*vsx_le_perm_store_v4si): Likewise.
5373 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5375 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
5376 directly to circumvent subtract from splat{31} workaround.
5377 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
5379 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
5380 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
5381 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
5382 (altivec_vperm_<mode>): Convert to define_insn_and_split to
5383 separate big and little endian logic.
5384 (*altivec_vperm_<mode>_internal): New define_insn.
5385 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
5386 separate big and little endian logic.
5387 (*altivec_vperm_<mode>_uns_internal): New define_insn.
5388 (vec_permv16qi): Add little endian logic.
5390 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
5392 * doc/extend.texi (returns_nonnull): Remove arguments.
5394 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5395 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5396 Sergey Lega <sergey.s.lega@intel.com>
5397 Anna Tikhonova <anna.tikhonova@intel.com>
5398 Ilya Tocar <ilya.tocar@intel.com>
5399 Andrey Turetskiy <andrey.turetskiy@intel.com>
5400 Ilya Verbin <ilya.verbin@intel.com>
5401 Kirill Yukhin <kirill.yukhin@intel.com>
5402 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5404 * config/i386/sse.md (VI48F_256_512): New.
5405 (avx2_permvar<mode>): Change to ...
5406 (<avx2_avx512f>_permvar<mode>): This.
5408 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5409 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5410 Sergey Lega <sergey.s.lega@intel.com>
5411 Anna Tikhonova <anna.tikhonova@intel.com>
5412 Ilya Tocar <ilya.tocar@intel.com>
5413 Andrey Turetskiy <andrey.turetskiy@intel.com>
5414 Ilya Verbin <ilya.verbin@intel.com>
5415 Kirill Yukhin <kirill.yukhin@intel.com>
5416 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5418 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
5419 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
5420 * config/i386/sse.md (VI4_AVX): New.
5421 (sf2simodelower): Ditto.
5422 (sse2_cvtps2dq): Change to ...
5423 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
5425 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5426 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5427 Sergey Lega <sergey.s.lega@intel.com>
5428 Anna Tikhonova <anna.tikhonova@intel.com>
5429 Ilya Tocar <ilya.tocar@intel.com>
5430 Andrey Turetskiy <andrey.turetskiy@intel.com>
5431 Ilya Verbin <ilya.verbin@intel.com>
5432 Kirill Yukhin <kirill.yukhin@intel.com>
5433 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5435 * config/i386/sse.md (V_512): New.
5437 (vcond<V_512:mode><VF_512:mode>): Ditto.
5438 (vcond<V_512:mode><VI_512:mode>): Ditto.
5439 (vcondu<V_512:mode><VI_512:mode>): Ditto.
5441 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5442 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5443 Sergey Lega <sergey.s.lega@intel.com>
5444 Anna Tikhonova <anna.tikhonova@intel.com>
5445 Ilya Tocar <ilya.tocar@intel.com>
5446 Andrey Turetskiy <andrey.turetskiy@intel.com>
5447 Ilya Verbin <ilya.verbin@intel.com>
5448 Kirill Yukhin <kirill.yukhin@intel.com>
5449 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5451 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
5452 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
5454 (fma<mode>4): Removed condition.
5455 (fms<mode>4): Ditto.
5456 (fnma<mode>4): Ditto.
5457 (fnms<mode>4): Ditto.
5458 (fma4i_fmadd_<mode>): Ditto.
5459 (*fma_fmadd_<mode>): Ditto.
5460 (*fma_fmsub_<mode>): Ditto.
5461 (*fma_fnmadd_<mode>): Ditto.
5462 (*fma_fnmsub_<mode>): Ditto.
5463 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
5464 (*fma_fmaddsub_<mode>): Ditto.
5465 (*fma_fmsubadd_<mode>): Ditto.
5466 (*fmai_fmadd_<mode>): Ditto.
5467 (*fmai_fmsub_<mode>): Ditto.
5468 (*fmai_fnmadd_<mode>): Ditto.
5469 (*fmai_fnmsub_<mode>): Ditto.
5471 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5472 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5473 Sergey Lega <sergey.s.lega@intel.com>
5474 Anna Tikhonova <anna.tikhonova@intel.com>
5475 Ilya Tocar <ilya.tocar@intel.com>
5476 Andrey Turetskiy <andrey.turetskiy@intel.com>
5477 Ilya Verbin <ilya.verbin@intel.com>
5478 Kirill Yukhin <kirill.yukhin@intel.com>
5479 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5481 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
5482 (VI124_256): Changed to ...
5483 (VI124_256_48_512): This.
5484 (ssepackmode): Extended with wider modes.
5485 (<code><mode>3): Changed iterator.
5486 (*avx2_<code><mode>3): Ditto.
5487 (vec_pack_trunc_<mode>): Ditto.
5489 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5490 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5491 Sergey Lega <sergey.s.lega@intel.com>
5492 Anna Tikhonova <anna.tikhonova@intel.com>
5493 Ilya Tocar <ilya.tocar@intel.com>
5494 Andrey Turetskiy <andrey.turetskiy@intel.com>
5495 Ilya Verbin <ilya.verbin@intel.com>
5496 Kirill Yukhin <kirill.yukhin@intel.com>
5497 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5499 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
5500 (VI8F_256_512): Ditto.
5501 (abs<mode>2): Changed iterator.
5502 (avx2_perm<mode>): Changed to ...
5503 (<avx2_avx512f>_perm<mode>): This.
5504 (avx2_perm<mode>_1): Changed to ...
5505 (<avx2_avx512f>_perm<mode>_1): This.
5507 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5508 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5509 Sergey Lega <sergey.s.lega@intel.com>
5510 Anna Tikhonova <anna.tikhonova@intel.com>
5511 Ilya Tocar <ilya.tocar@intel.com>
5512 Andrey Turetskiy <andrey.turetskiy@intel.com>
5513 Ilya Verbin <ilya.verbin@intel.com>
5514 Kirill Yukhin <kirill.yukhin@intel.com>
5515 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5517 * config/i386/sse.md (VI48_AVX512F): New.
5518 (VI48_AVX2): Changed to ...
5519 (VI48_AVX2_48_AVX512F): This.
5520 (avx2_ashrv<mode>): Changed to ...
5521 (<avx2_avx512f>_ashrv<mode>): This.
5522 (avx2_<shift_insn>v<mode>): Changed to ...
5523 (<avx2_avx512f>_<shift_insn>v<mode>): This.
5525 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5526 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5527 Sergey Lega <sergey.s.lega@intel.com>
5528 Anna Tikhonova <anna.tikhonova@intel.com>
5529 Ilya Tocar <ilya.tocar@intel.com>
5530 Andrey Turetskiy <andrey.turetskiy@intel.com>
5531 Ilya Verbin <ilya.verbin@intel.com>
5532 Kirill Yukhin <kirill.yukhin@intel.com>
5533 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5535 * config/i386/sse.md (VI4_AVX512F): New.
5536 (VI8_AVX2_AVX512F): Ditto.
5537 (mul<mode>3): Extended with wider modes.
5538 (*<sse4_1_avx2>_mul<mode>3): Ditto.
5539 (mul<mode>3): Ditto.
5540 (vec_widen_<s>mult_odd_<mode>): Ditto.
5542 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5543 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5544 Sergey Lega <sergey.s.lega@intel.com>
5545 Anna Tikhonova <anna.tikhonova@intel.com>
5546 Ilya Tocar <ilya.tocar@intel.com>
5547 Andrey Turetskiy <andrey.turetskiy@intel.com>
5548 Ilya Verbin <ilya.verbin@intel.com>
5549 Kirill Yukhin <kirill.yukhin@intel.com>
5550 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5552 * config/i386/sse.md (VI2_AVX512F): New.
5553 (VI124_AVX512F): Ditto.
5554 (sseunpackmode): Extended with wider modes.
5555 (sseunpackfltmode): Ditto.
5556 (vec_unpacks_float_hi_<mode>): Ditto.
5557 (vec_unpacks_float_lo_<mode>): Ditto.
5558 (vec_unpacku_float_hi_<mode>): Ditto.
5559 (vec_unpacku_float_lo_<mode>): Ditto.
5560 (vec_unpacks_lo_<mode>): Ditto.
5561 (vec_unpacks_hi_<mode>): Ditto.
5562 (vec_unpacku_lo_<mode>): Ditto.
5563 (vec_unpacku_hi_<mode>): Ditto.
5565 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5566 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5567 Sergey Lega <sergey.s.lega@intel.com>
5568 Anna Tikhonova <anna.tikhonova@intel.com>
5569 Ilya Tocar <ilya.tocar@intel.com>
5570 Andrey Turetskiy <andrey.turetskiy@intel.com>
5571 Ilya Verbin <ilya.verbin@intel.com>
5572 Kirill Yukhin <kirill.yukhin@intel.com>
5573 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5575 * config/i386/i386.md (multdiv): New.
5576 (multdiv_mnemonic): Ditto.
5577 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
5578 (<sse>_vm<multdiv_mnemonic><mode>3): This.
5579 (<sse>_vmdiv<mode>3): Removed.
5581 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5582 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5583 Sergey Lega <sergey.s.lega@intel.com>
5584 Anna Tikhonova <anna.tikhonova@intel.com>
5585 Ilya Tocar <ilya.tocar@intel.com>
5586 Andrey Turetskiy <andrey.turetskiy@intel.com>
5587 Ilya Verbin <ilya.verbin@intel.com>
5588 Kirill Yukhin <kirill.yukhin@intel.com>
5589 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5591 * config/i386/sse.md (V): Extended with wider modes.
5593 (ssehalfvecmode): Ditto.
5595 (ssepackfltmode): Ditto.
5596 (avx_vec_concat<mode>): Ditto.
5597 (V_256_512): New iterator.
5598 (VF2_512_256): Ditto.
5599 (si2dfmode): New attribute.
5600 (si2dfmodelower): Ditto.
5602 (concat_tg_mode): Ditto.
5603 (floatv4siv4df2): Changed to ...
5604 (float<si2dfmodelower><mode>2): This.
5605 (avx_cvtps2pd256): Changed to ...
5606 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
5607 (vec_pack_trunc_v4df): Changed to ...
5608 (vec_pack_trunc_<mode>): This.
5609 (avx_vpermil<mode>): Changed to ...
5610 (<sse2_avx_avx512f>_vpermil<mode>): This.
5611 (<fixsuffix>fix_truncv8dfv8si2): New.
5612 (vec_pack_sfix_trunc_v8df): Ditto.
5613 (avx512f_rndscale<mode>): Ditto.
5614 (avx512f_roundpd512): Ditto.
5615 (vec_pack_ufix_trunc_<mode>): Updated iterator.
5617 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5618 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5619 Sergey Lega <sergey.s.lega@intel.com>
5620 Anna Tikhonova <anna.tikhonova@intel.com>
5621 Ilya Tocar <ilya.tocar@intel.com>
5622 Andrey Turetskiy <andrey.turetskiy@intel.com>
5623 Ilya Verbin <ilya.verbin@intel.com>
5624 Kirill Yukhin <kirill.yukhin@intel.com>
5625 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5627 * config/i386/i386.md (any_fix): New iterator.
5628 (fixsuffix): New attribute.
5629 * config/i386/sse.md (VF1): Extened with wider modes.
5633 (sseintvecmodelower): Ditto.
5634 (ssescalarmode): Ditto.
5635 (ssescalarnum): Ditto.
5637 (ssexmmmode): Ditto.
5638 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
5639 (<sse>_rcp<mode>2): Change iterator.
5640 (rsqrt<mode>2): Ditto.
5641 (<sse>_rsqrt<mode>2): Ditto.
5642 (avx2_vec_dup<mode>): Ditto.
5643 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
5644 (round<mode>2_sfix): Ditto.
5645 (avx2_pbroadcast<mode>): Ditto.
5646 (*andnot<mode>3): Handle XI mode.
5647 (*<code><mode>3): Ditto.
5648 (AVXTOSSEMODE): Removed.
5649 (avx_vpermil<mode>): Changed to ...
5650 (<sse2_avx_avx512f>_vpermil<mode>): This.
5652 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5653 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5654 Sergey Lega <sergey.s.lega@intel.com>
5655 Anna Tikhonova <anna.tikhonova@intel.com>
5656 Ilya Tocar <ilya.tocar@intel.com>
5657 Andrey Turetskiy <andrey.turetskiy@intel.com>
5658 Ilya Verbin <ilya.verbin@intel.com>
5659 Kirill Yukhin <kirill.yukhin@intel.com>
5660 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5662 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
5663 (<sse>_comi): Ditto.
5664 (<sse>_ucomi): Ditto.
5665 (sse_cvtss2siq_2): Ditto.
5666 (sse2_cvtsd2si): Ditto.
5667 (sse2_cvtsd2siq): Ditto.
5668 (sse2_cvttsd2si): Ditto.
5669 (sse2_cvttsd2siq): Ditto.
5670 (<shift_insn><mode>3): Ditto.
5671 (sse2_cvtsi2sdq): Update constraint and prefix.
5672 (sse_cvtsi2ss): Update prefix.
5673 (sse_cvtsi2ssq): Ditto.
5675 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5677 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
5678 ignore internal calls.
5680 2013-10-11 Richard Biener <rguenther@suse.de>
5682 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
5683 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
5684 TREE_OVERFLOW is set.
5686 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
5688 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
5690 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
5691 * gimple.h: Likewise.
5693 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
5695 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
5698 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
5700 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
5701 * configure: Regenerate.
5703 2013-10-11 Jakub Jelinek <jakub@redhat.com>
5705 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
5706 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
5707 formatting fixes, use pp_colon instead of pp_character (..., ':'),
5708 similarly pp_right_paren.
5709 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
5710 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
5711 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
5712 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
5714 (omp_declare_simd_clauses_equal,
5715 omp_remove_redundant_declare_simd_attrs): New functions.
5716 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
5717 (walk_tree_1): Handle new OpenMP 4.0 clauses.
5718 * tree.h (OMP_LOOP_CHECK): Define.
5719 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
5720 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
5721 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
5722 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
5723 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
5724 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
5725 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
5726 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
5727 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
5728 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
5729 OMP_CLAUSE_SIMDLEN_EXPR): Define.
5730 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
5731 (omp_remove_redundant_declare_simd_attrs): New prototype.
5732 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
5733 GIMPLE_OMP_TEAMS): New codes.
5734 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
5735 * omp-low.c (struct omp_context): Add cancel_label and cancellable
5737 (target_nesting_level): New variable.
5738 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
5739 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
5740 for collapse > 1 static schedule unless ordered.
5741 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
5742 (determine_parallel_type): Adjust get_ws_args_for caller.
5743 (install_var_field): Handle mask & 4 for double indirection.
5744 (scan_sharing_clauses): Ignore shared clause on teams construct.
5745 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
5746 (create_omp_child_function): If inside target or declare target
5747 constructs, set "omp declare target" attribute on the child function.
5748 (find_combined_for): New function.
5749 (scan_omp_parallel): Handle combined loops.
5750 (scan_omp_target, scan_omp_teams): New functions.
5751 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
5752 restrictions and set ctx->cancellable for cancellable constructs.
5753 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
5754 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
5755 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
5756 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
5757 (omp_clause_aligned_alignment): New function.
5758 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
5759 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
5760 on teams constructs. Handle user defined reductions and new
5762 (lower_reduction_clauses): Don't set placeholder to address of ref
5763 if it has already the right type.
5764 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
5765 (expand_parallel_call): Use the new non-_start suffixed builtins,
5766 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
5767 and GOMP_parallel_end after the call.
5768 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
5769 (expand_omp_for_init_counts): Handle combined loops.
5770 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
5772 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
5773 end of cancellable loops.
5774 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
5775 Likewise. Handle collapse > 1 loops.
5776 (expand_omp_simd): Handle combined loops.
5777 (expand_omp_for): Add inner_stmt argument, adjust callers of
5778 expand_omp_for* functions, use expand_omp_for_static*chunk even
5779 for collapse > 1 unless ordered.
5780 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
5781 of cancellable sections.
5782 (expand_omp_single): Remove need_barrier variable, just rely on
5783 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
5784 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
5785 (expand_omp_atomic_load, expand_omp_atomic_store,
5786 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
5787 (expand_omp_target): New function.
5788 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
5789 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
5790 (build_omp_regions_1): Immediately close region for
5791 GF_OMP_TARGET_KIND_UPDATE.
5792 (maybe_add_implicit_barrier_cancel): New function.
5793 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
5795 (lower_omp_single): Likewise. Add clobber after the barrier.
5796 (lower_omp_taskgroup): New function.
5797 (lower_omp_for): Handle combined loops. Adjust
5798 lower_rec_input_clauses caller. Handle cancellation.
5799 (lower_depend_clauses): New function.
5800 (lower_omp_taskreg): Lower depend clauses. Adjust
5801 lower_rec_input_clauses caller. Add clobber after the call. Handle
5803 (lower_omp_target, lower_omp_teams): New functions.
5804 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
5805 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
5806 and GOMP_cancellation_point calls.
5807 (lower_omp): Fold stmts inside of target region.
5808 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
5809 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5810 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
5811 (BT_FN_VOID_OMPFN_PTR_UINT,
5812 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
5813 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
5814 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
5815 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
5816 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
5817 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
5818 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
5819 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
5820 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
5821 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
5822 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
5823 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
5824 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
5825 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
5826 BUILT_IN_GOMP_PARALLEL_END.
5827 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
5828 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
5829 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
5830 GF_OMP_FOR_KIND_DISTRIBUTE.
5831 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
5832 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
5833 (dump_gimple_omp_return): Print lhs if it has any.
5834 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
5835 gimple_omp_atomic_seq_cst_p.
5836 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
5837 and GIMPLE_OMP_TEAMS.
5838 * langhooks.c (lhd_omp_mappable_type): New function.
5839 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
5840 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
5842 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
5843 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
5844 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
5845 (struct gimplify_omp_ctx): Add combined_loop field.
5846 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
5847 on stmts inside of target region.
5848 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
5849 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
5850 ORT_TARGET and ORT_TARGET_DATA.
5851 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
5853 (omp_notice_threadprivate_variable): Complain about threadprivate
5854 variables in target region.
5855 (omp_notice_variable): Complain about vars with non-mappable type
5856 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
5857 (omp_check_private): Ignore ORT_TARGET* regions.
5858 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
5859 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
5860 (find_combined_omp_for): New function.
5861 (gimplify_omp_for): Handle gimplification of combined loops.
5862 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
5864 (gimplify_omp_target_update): New function.
5865 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
5866 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
5867 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
5868 (gimplify_body): If fndecl has "omp declare target" attribute, add
5869 implicit ORT_TARGET context around it.
5870 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
5871 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
5872 * tree-nested.c (convert_nonlocal_reference_stmt,
5873 convert_local_reference_stmt, convert_gimple_call): Handle
5874 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5875 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
5876 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
5877 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
5878 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
5879 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
5880 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
5881 BUILT_IN_GOMP_LOOP_END_CANCEL,
5882 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
5883 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
5884 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
5885 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
5886 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
5887 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
5888 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
5889 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
5890 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
5891 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
5892 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
5893 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
5894 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
5895 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
5896 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
5897 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5898 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
5899 gimple_build_omp_teams): New functions.
5900 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
5901 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
5902 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
5903 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5904 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
5905 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
5906 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
5907 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
5908 GF_OMP_ATOMIC_SEQ_CST): New.
5909 (gimple_build_omp_taskgroup, gimple_build_omp_target,
5910 gimple_build_omp_teams): New prototypes.
5911 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
5912 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5913 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
5914 GIMPLE_OMP_SINGLE as end of range.
5915 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
5916 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
5917 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
5918 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
5919 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
5920 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
5921 gimple_omp_target_kind, gimple_omp_target_set_kind,
5922 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
5923 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
5924 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
5925 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
5926 gimple_omp_teams_set_clauses): New inlines.
5927 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
5928 and GIMPLE_OMP_TASKGROUP.
5929 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
5930 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
5931 enum omp_clause_proc_bind_kind): New.
5932 (union omp_clause_subcode): Add depend_kind, map_kind and
5933 proc_bind_kind fields.
5934 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
5935 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
5936 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
5937 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
5938 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
5940 2013-10-10 Teresa Johnson <tejohnson@google.com>
5942 * predict.c (tree_estimate_probability): Add new parameter
5943 for estimate_bb_frequencies.
5944 (estimate_bb_frequencies): Add new parameter to force estimation.
5945 (rebuild_frequencies): When max frequency in function is small,
5946 recompute counts from frequencies.
5947 * predict.h (estimate_bb_frequencies): New parameter.
5949 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5951 * ipa-inline.c (ipa_inline): Fix leak of "order" when
5952 optimizations are disabled.
5954 2013-10-10 David Malcolm <dmalcolm@redhat.com>
5956 * coverage.c (coverage_finish): Fix leak of da_file_name.
5958 2013-10-10 Jan Hubicka <jh@suse.cz>
5960 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
5961 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
5964 2013-10-10 Jakub Jelinek <jakub@redhat.com>
5967 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
5968 if any labels are in FALLTHRU_BB, use a special label emitted
5969 immediately after the asm goto insn rather than label_rtx
5971 (expand_asm_stmt): Adjust caller.
5972 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
5973 edge if the last insn in predecessor is a jump with single successor,
5974 but it isn't simplejump_p.
5976 2013-10-10 Richard Biener <rguenther@suse.de>
5978 PR tree-optimization/58656
5979 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
5981 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5983 * gimplify.c: Include expr.h and tm_p.h for targets with special
5984 va-arg padding requirements.
5986 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
5988 * tree-flow.h: Move some prototypes to gimple.h.
5989 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
5990 * gimple.h: Relocate some prototypes from tree-flow.h
5991 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
5993 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
5994 (build_va_arg_indirect_ref): Relocate and make static.
5995 (std_gimplify_va_arg_expr): Relocate here.
5996 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
5997 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
5999 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6001 * doc/md.texi: Document the mnemonic attribute.
6003 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6006 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
6007 (set_attr_alternative x ...) when searching for user defined
6010 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
6012 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
6014 2013-10-09 Easwaran Raman <eraman@google.com>
6016 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
6017 * cfgexpand.c (defer_stack_allocation): ...use here
6018 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
6020 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
6022 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
6023 (operand_equal_for_value_replacement): New function, extracted from
6024 value_replacement and enhanced to catch more cases.
6025 (value_replacement): Use operand_equal_for_value_replacement.
6027 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6029 * loop-doloop.c (doloop_modify, doloop_optimize): Use
6030 get_max_loop_iterations.
6032 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6034 * config/arm/aarch-common.c (arm_early_load_addr_dep):
6035 Place comment above function.
6037 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6039 * tree-flow.h: Remove all remaining prototypes, enums and structs that
6040 are not related to tree-cfg.c.
6041 * tree-ssa-address.h: New file. Relocate prototypes.
6042 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
6043 (addr_for_mem_ref): New. Combine call to get_address_description and
6044 return addr_for_mem_ref.
6045 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
6046 * tree-ssa-live.h: Adjust prototypes.
6047 * passes.c: Include tree-ssa-live.h.
6048 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
6049 * graphite.c (graphite_transform_loops): Make static.
6050 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6051 make_pass_graphite, pass_data_graphite_transforms,
6052 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
6053 * ipa-pure-const.c (warn_function_noreturn): Make static.
6054 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6055 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6056 Relocate from tree-cfg.c
6057 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
6059 (execute_warn_function_noreturn, gate_warn_function_noreturn,
6060 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
6061 Move to ipa-pure-const.c.
6062 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
6063 Relocate from tree-optimize.c.
6064 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
6065 make_pass_fixup_cfg): Move to tree-cfg.c.
6066 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
6067 Relocate some prototypes.
6068 * tree-data-ref.h (tree_check_data_deps) Add prototype.
6069 * tree-dump.c (dump_function_to_file): Remove prototype.
6070 Add tree-flow.h to the include file.
6071 * tree-dump.h: Remove prototype.
6072 * tree-parloops.h: New File. Add prototypes.
6073 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
6074 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
6075 from tree-ssa-loop.c.
6076 * tree-predcom.c (run_tree_predictive_commoning,
6077 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
6078 Relocate here from tree-ssa-loop.c.
6079 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
6080 ssa_name_values.release ().
6081 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
6082 (ssa_name_values): Relocate from tree-flow.h.
6083 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
6084 * tree-ssa-loop.c (run_tree_predictive_commoning,
6085 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
6086 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
6087 make_pass_graphite, pass_data_graphite_transforms,
6088 make_pass_graphite_transforms, gate_tree_parallelize_loops,
6089 tree_parallelize_loops, pass_data_parallelize_loops,
6090 make_pass_parallelize_loops): Move to other files.
6091 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
6093 * tree.h: Remove prototypes from tree-address.c.
6095 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6097 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
6098 (struct int_tree_map): Move to tree-hasher.h
6099 (SCALE, LABEL, PERCENT): Move to gimple.h
6100 * tree-flow-inline.h: Delete. Move functions to other files.
6101 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
6102 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
6103 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
6104 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
6105 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
6106 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
6107 * tree-hasher.h: Don't include tree-flow.h.
6108 (struct int_tree_map): Relocate from tree-flow.h.
6109 * tree-sra.c (contains_view_convert_expr_p): Relocate from
6110 tree-flow-inline.h and make static.
6111 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
6113 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
6114 tree-flow-inline.h and make static.
6115 * tree.h (is_global_var, may_be_aliased): Relocate from
6117 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
6118 * value-prof.c: No longer include tree-flow-inline.h.
6119 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
6121 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
6123 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
6124 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
6125 (enum move_pos): Move to tree-ssa-loop-im.h
6126 * cfgloop.h: Move some prototypes.
6127 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
6128 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
6129 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
6130 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
6131 (loop_containing_stmt): Relocate from tree-flow-inline.h.
6132 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
6133 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
6134 (enum move_pos): Relocate here.
6135 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
6137 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
6138 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
6139 make_pass_lim): Relocate here from tree-ssa-loop.c.
6140 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
6142 (loop_edge_to_cancel, unloop_loops): Make static.
6143 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6144 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
6145 (tree_complete_unroll, gate_tree_complete_unroll,
6146 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
6147 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6148 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
6149 * tree-ssa-loop-ivopts.c: Remove local prototypes.
6150 (stmt_invariant_in_loop_p): Remove unused function.
6151 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
6152 * tree-ssa-loop-manip.h: New file. Add prototypes.
6153 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
6154 (gcov_type_to_double_int): Move to cfgloop.h.
6155 (double_int_cmp, bound_index,
6156 estimate_numbers_of_iterations_loop): Make static.
6157 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
6158 (max_loop_iterations): Factor out get_max_loop_iterations.
6159 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
6161 * tree-ssa-loop-niter.h: New file. Add prototypes.
6162 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
6163 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
6164 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
6165 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
6166 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
6167 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
6168 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
6169 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
6170 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
6171 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
6172 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
6173 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
6174 pass_data_complete_unroll, make_pass_complete_unroll,
6175 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
6176 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
6177 tree-ssa-loop-ivcanon.c.
6178 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
6179 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
6180 tree-ssa-loop-prefetch.c.
6181 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
6183 (get_lsm_tmp_name): Relocate and add suffix parameter.
6184 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
6185 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
6186 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
6187 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
6188 (get_estimated_loop_iterations): Factor out accessor from
6189 estimated_loop_iterations in tree-ssa-loop-niter.c.
6190 (get_max_loop_iterations): Factor out accessor from
6191 _max_loop_iterations in tree-ssa-niter.c.
6192 * loop-unroll.c (decide_unroll_constant_iterations,
6193 decide_unroll_runtime_iterations, decide_peel_simple,
6194 decide_unroll_stupid): Use new get_* accessors.
6196 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
6198 PR tree-optimization/20318
6199 * doc/extend.texi (returns_nonnull): New function attribute.
6200 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
6202 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
6203 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
6205 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
6208 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
6209 false if both components are bitfields.
6211 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6213 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
6214 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
6215 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
6216 * config/aarch64/aarch64.h
6217 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
6218 * config/aarch64/aarch64-simd-builtins.def
6219 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
6221 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6223 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
6225 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6227 * config/aarch64/arm_neon.h (vdiv_f64): Added.
6229 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
6231 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
6232 (vneg_f64): New intrinsic.
6233 (vneg_s8): Asm replaced with C.
6234 (vneg_s16): Likewise.
6235 (vneg_s32): Likewise.
6236 (vneg_s64): New intrinsic.
6237 (vnegq_f32): Asm replaced with C.
6238 (vnegq_f64): Likewise.
6239 (vnegq_s8): Likewise.
6240 (vnegq_s16): Likewise.
6241 (vnegq_s32): Likewise.
6242 (vnegq_s64): Likewise.
6244 2013-10-09 Renlin Li <Renlin.Li@arm.com>
6246 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
6248 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6250 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
6251 packed stack special handling.
6252 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
6253 back to fixed stack slots for FPRs saved due to stdarg.
6255 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6257 * config/s390/s390.c (s390_frame_info): Restructure function.
6259 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6261 * config/s390/s390.c (struct s390_frame_layout): New field
6263 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
6264 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
6265 regs_ever_clobbered to char*.
6266 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
6267 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6268 (s390_register_info): Enable FPR save slots. Move/Copy some
6269 functionality into ...
6270 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
6271 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
6273 (s390_frame_info): Do gpr slot allocation here now. stdarg does
6274 not imply a stack frame.
6275 (s390_init_frame_layout): Remove variable clobbered_regs.
6276 (s390_update_register_info): Remove function.
6277 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
6278 cannot be used for register renaming.
6279 (s390_hard_regno_scratch_ok): New function.
6280 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
6281 (s390_initial_elimination_offset): Change offset calculation of
6282 the return address pointer.
6283 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6284 (restore_gprs): Set frame related flag.
6285 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6286 (s390_emit_prologue): Call s390_register_info instead of
6287 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
6288 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
6289 (s390_optimize_prologue): Call s390_optimize_register_info.
6290 Try to remove also FPR slot save/restore INSNs. Remove frame
6291 related flags from restore INSNs.
6293 2013-10-08 DJ Delorie <dj@redhat.com>
6295 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
6298 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
6299 avoid conflict with the MI use of %c.
6300 * config/rl78/rl78-real.md: change %c to %C throughout.
6301 * config/rl78/rl78-virt.md: Likewise.
6303 2013-10-08 Jan Hubicka <jh@suse.cz>
6305 * config/i386/i386.c (ix86_option_override_internal): Switch
6306 to SSE math for -ffast-math when target ISA supports SSE2.
6308 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
6310 * tree-flow.h: Remove some prototypes.
6311 * tree.h: Remove some protypes, add a couple.
6312 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
6313 using_eh_for_cleanups_p): Add interface routines for front ends.
6314 * tree-eh.h: New file. Add protoptyes.
6315 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
6316 (add_stmt_to_eh_lp_fn): Make static.
6317 (lower_try_finally): Use new using_eh_for_cleanups_p.
6318 * emit-rtl.c: Include tree-eh.h.
6319 * gimple.h: Include tree-eh.h.
6321 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
6323 PR tree-optimization/58480
6324 * tree-vrp.c (infer_nonnull_range): New function.
6325 (infer_value_range): Call infer_nonnull_range.
6327 2013-10-08 Dehao Chen <dehao@google.com>
6329 PR tree-optimization/58619
6330 * tree-inline.c (copy_phis_for_bb): Combine location data
6333 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
6336 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
6337 RTX_FRAME_RELATED_P on INSN.
6339 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6341 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
6342 (altivec_expand_vec_perm_const): Call it.
6344 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6346 * config/rs6000/vector.md (mov<mode>): Emit permuted move
6347 sequences for LE VSX loads and stores at expand time.
6348 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
6350 * config/rs6000/rs6000.c (rs6000_const_vec): New.
6351 (rs6000_gen_le_vsx_permute): New.
6352 (rs6000_gen_le_vsx_load): New.
6353 (rs6000_gen_le_vsx_store): New.
6354 (rs6000_gen_le_vsx_move): New.
6355 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
6356 (*vsx_le_perm_load_v4si): New.
6357 (*vsx_le_perm_load_v8hi): New.
6358 (*vsx_le_perm_load_v16qi): New.
6359 (*vsx_le_perm_store_v2di): New.
6360 (*vsx_le_perm_store_v4si): New.
6361 (*vsx_le_perm_store_v8hi): New.
6362 (*vsx_le_perm_store_v16qi): New.
6363 (*vsx_xxpermdi2_le_<mode>): New.
6364 (*vsx_xxpermdi4_le_<mode>): New.
6365 (*vsx_xxpermdi8_le_V8HI): New.
6366 (*vsx_xxpermdi16_le_V16QI): New.
6367 (*vsx_lxvd2x2_le_<mode>): New.
6368 (*vsx_lxvd2x4_le_<mode>): New.
6369 (*vsx_lxvd2x8_le_V8HI): New.
6370 (*vsx_lxvd2x16_le_V16QI): New.
6371 (*vsx_stxvd2x2_le_<mode>): New.
6372 (*vsx_stxvd2x4_le_<mode>): New.
6373 (*vsx_stxvd2x8_le_V8HI): New.
6374 (*vsx_stxvd2x16_le_V16QI): New.
6376 2013-10-07 Renlin Li <Renlin.Li@arm.com>
6378 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
6380 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6382 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
6383 loop to work also for 31bit ABI.
6384 Save the stack pointer for frame_size > 0.
6386 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6388 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
6389 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
6390 constraint letters from expanders.
6391 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
6392 retry count to general_operand.
6393 ("tabort"): Give operand 0 a mode.
6394 ("tabort_1"): Add mode and constraint letter for operand 0.
6395 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
6397 2013-10-04 Jeff Law <law@redhat.com>
6399 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
6401 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
6403 (thread_across_edge): Here. Call it.
6405 2013-10-04 Cary Coutant <ccoutant@google.com>
6407 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
6408 discarding a location list expression (or a piece of one).
6410 2013-10-03 Jan Hubicka <jh@suse.cz>
6412 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
6413 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
6414 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6416 2013-10-03 Jan Hubicka <jh@suse.cz>
6418 * config/i386/i386.c (ix86_option_override_internal): Do not enable
6419 accumulate-outgoing-args when producing unwind info.
6421 2013-10-03 Wei Mi <wmi@google.com>
6423 * lra-constraints.c (insert_move_for_subreg): New function
6424 extracted from simplify_operand_subreg.
6425 (simplify_operand_subreg): Add reload for paradoxical subreg.
6427 2013-10-03 Rong Xu <xur@google.com>
6429 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
6430 the candidate of builtin_expect such that we should fix the
6431 size/time estimation.
6432 (estimate_function_body_sizes): Do the acutally size/time fix-up
6435 2013-10-03 Rong Xu <xur@google.com>
6437 * predict.c (tree_predict_by_opcode): Get the probability
6438 for builtin_expect from param builtin_expect_probability.
6439 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
6440 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
6441 * doc/invoke.texi: Add documentation for builtin-expect-probability.
6443 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
6446 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
6447 * calls.c (alloca_call_p): Use get_callee_fndecl.
6448 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
6449 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
6451 (vrp_visit_stmt): Remove duplicated code.
6453 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
6455 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
6456 ceildf2, btruncdf2, instead of vsx_* name.
6458 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
6459 iterators to only do V2DF and V4SF here. Move the DF code to
6460 rs6000.md where it is combined with SF mode. Replace <VSv> with
6461 just 'v' since only vector operations are handled with these insns
6462 after moving the DF support to rs6000.md.
6463 (vsx_sub<mode>3): Likewise.
6464 (vsx_mul<mode>3): Likewise.
6465 (vsx_div<mode>3): Likewise.
6466 (vsx_fre<mode>2): Likewise.
6467 (vsx_neg<mode>2): Likewise.
6468 (vsx_abs<mode>2): Likewise.
6469 (vsx_nabs<mode>2): Likewise.
6470 (vsx_smax<mode>3): Likewise.
6471 (vsx_smin<mode>3): Likewise.
6472 (vsx_sqrt<mode>2): Likewise.
6473 (vsx_rsqrte<mode>2): Likewise.
6474 (vsx_fms<mode>4): Likewise.
6475 (vsx_nfma<mode>4): Likewise.
6476 (vsx_copysign<mode>3): Likewise.
6477 (vsx_btrunc<mode>2): Likewise.
6478 (vsx_floor<mode>2): Likewise.
6479 (vsx_ceil<mode>2): Likewise.
6480 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
6481 (vsx_sminsf3): Likewise.
6482 (vsx_fmadf4): Likewise.
6483 (vsx_fmsdf4): Likewise.
6484 (vsx_nfmadf4): Likewise.
6485 (vsx_nfmsdf4): Likewise.
6486 (vsx_cmpdf_internal1): Likewise.
6488 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
6489 simpler to select whether a target has SPE or traditional floating
6490 point support in iterators.
6491 (TARGET_DF_SPE): Likewise.
6492 (TARGET_SF_FPR): Likewise.
6493 (TARGET_DF_FPR): Likewise.
6494 (TARGET_SF_INSN): Macros to say whether floating point support
6495 exists for a given operation for expanders.
6496 (TARGET_DF_INSN): Likewise.
6498 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
6499 combining of SF/DF mode operations, using both traditional and VSX
6507 (abs<mode>2): Combine SF/DF modes using traditional floating point
6508 instructions. Add support for using the upper DF registers with
6509 VSX support, and SF registers with power8-vector support. Update
6510 expanders for operations supported by both the SPE and traditional
6511 floating point units.
6512 (abs<mode>2_fpr): Likewise.
6513 (nabs<mode>2): Likewise.
6514 (nabs<mode>2_fpr): Likewise.
6515 (neg<mode>2): Likewise.
6516 (neg<mode>2_fpr): Likewise.
6517 (add<mode>3): Likewise.
6518 (add<mode>3_fpr): Likewise.
6519 (sub<mode>3): Likewise.
6520 (sub<mode>3_fpr): Likewise.
6521 (mul<mode>3): Likewise.
6522 (mul<mode>3_fpr): Likewise.
6523 (div<mode>3): Likewise.
6524 (div<mode>3_fpr): Likewise.
6525 (sqrt<mode>3): Likewise.
6526 (sqrt<mode>3_fpr): Likewise.
6527 (fre<Fs>): Likewise.
6528 (rsqrt<mode>2): Likewise.
6529 (cmp<mode>_fpr): Likewise.
6530 (smax<mode>3): Likewise.
6531 (smin<mode>3): Likewise.
6532 (smax<mode>3_vsx): Likewise.
6533 (smin<mode>3_vsx): Likewise.
6534 (negsf2): Delete SF operations that are merged with DF.
6541 (fmasf4_fpr): Likewise.
6542 (fmssf4_fpr): Likewise.
6543 (nfmasf4_fpr): Likewise.
6544 (nfmssf4_fpr): Likewise.
6545 (sqrtsf2): Likewise.
6546 (rsqrtsf_internal1): Likewise.
6547 (smaxsf3): Likewise.
6548 (sminsf3): Likewise.
6549 (cmpsf_internal1): Likewise.
6550 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
6551 (negdf2): Delete DF operations that are merged with SF.
6553 (nabsdf2): Likewise.
6559 (rsqrtdf_internal1): Likewise.
6560 (fmadf4_fpr): Likewise.
6561 (fmsdf4_fpr): Likewise.
6562 (nfmadf4_fpr): Likewise.
6563 (nfmsdf4_fpr): Likewise.
6564 (sqrtdf2): Likewise.
6565 (smaxdf3): Likewise.
6566 (smindf3): Likewise.
6567 (cmpdf_internal1): Likewise.
6568 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
6569 (btrunc<mode>2): Delete separate expander, and combine with the
6570 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
6571 (btrunc<mode>2_fpr): Likewise.
6572 (ceil<mode>2): Likewise.
6573 (ceil<mode>2_fpr): Likewise.
6574 (floor<mode>2): Likewise.
6575 (floor<mode>2_fpr): Likewise.
6576 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
6577 Add support for using the upper registers with VSX and
6578 power8-vector. Move insns to be closer to the define_expands. On
6579 VSX systems, prefer the traditional form of FMA over the VSX
6580 version, since the traditional form allows the target not to
6581 overlap with the inputs.
6582 (fms<mode>4_fpr): Likewise.
6583 (nfma<mode>4_fpr): Likewise.
6584 (nfms<mode>4_fpr): Likewise.
6586 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6587 Richard Earnshaw <richard.earnshaw@arm.com>
6589 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
6590 (struct mult_cost_table): Likewise.
6591 (struct mem_cost_table): Likewise.
6592 (struct fp_cost_table): Likewise.
6593 (struct vector_cost_table): Likewise.
6594 (cpu_cost_table): Likewise.
6595 * config/arm/arm.opt (mold-rts-costs): New option.
6596 (mnew-generic-costs): Likewise.
6597 * config/arm/arm.c (generic_extra_costs): New table.
6598 (cortexa15_extra_costs): Likewise.
6599 (arm_slowmul_tune): Use NULL as new costs.
6600 (arm_fastmul_tune): Likewise.
6601 (arm_strongarm_tune): Likewise.
6602 (arm_xscale_tune): Likewise.
6603 (arm_9e_tune): Likewise.
6604 (arm_v6t2_tune): Likewise.
6605 (arm_cortex_a5_tune): Likewise.
6606 (arm_cortex_a9_tune): Likewise.
6607 (arm_v6m_tune): Likewise.
6608 (arm_fa726te_tune): Likewise.
6609 (arm_cortex_a15_tune): Use cortex15_extra_costs.
6610 (arm_cortex_tune): Use generict_extra_costs.
6611 (shifter_op_p): New function.
6612 (arm_unspec_cost): Likewise.
6613 (LIBCALL_COST): Define.
6614 (arm_new_rtx_costs): New function.
6615 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
6616 table is available. Use old costs otherwise unless mnew-generic-costs
6618 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
6619 (cpu_cost_table): Declare.
6621 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
6624 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
6625 (*subs_mul_imm_<mode>)
6626 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
6627 (*sub_<shift>_<mode>)
6628 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
6629 Remove k constraint.
6631 2013-10-03 Ian Bolton <ian.bolton@arm.com>
6633 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
6635 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
6637 2013-10-02 Teresa Johnson <tejohnson@google.com>
6639 * predict.c (probably_never_executed): New function.
6640 (probably_never_executed_bb_p): Invoke probably_never_executed.
6641 (probably_never_executed_edge_p): Ditto.
6642 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
6643 Treat profile insanities conservatively.
6645 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
6647 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
6649 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
6651 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
6652 use it. Use smaller increase for scratch. Don't increase reject
6653 for early clobber scratch.
6654 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
6655 setting eliminated regs except setting fp from hfp.
6656 (lra_eliminate): Check lra_insn_recog_data on NULL.
6658 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
6661 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
6662 setting -mvsx-timode by default until the underlying problem is fixed.
6663 (RS6000_CPU, power7 defaults): Likewise.
6665 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6667 * config/x-linux (host-linux.o): Remove header dependencies.
6668 Use $(COMPILE) and $(POSTCOMPILE).
6669 * config/t-linux-android (linux-android.o): Ditto.
6671 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6673 * Makefile.in (expmed.o-warn): Remove.
6675 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6677 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
6678 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
6680 2013-10-02 Teresa Johnson <tejohnson@google.com>
6682 * dojump.c (do_jump_1): Divide probability between
6683 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
6685 2013-10-02 Tom Tromey <tromey@redhat.com>
6687 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
6689 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6691 * tree-flow.h: Remove some prototypes.
6692 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
6693 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
6694 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
6695 mark_virtual_phi_result_for_renaming): Relocate here.
6696 * tree-into-ssa.h: Add prototypes.
6697 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
6698 single_pred_before_succ_order.
6699 (blocks_in_phiopt_order): Rename and move to cfganal.c.
6700 (nonfreeing_call_p) Move to gimple.c.
6701 * cfganal.c (single_pred_before_succ_order): Move and renamed from
6703 * basic-block.h (single_pred_before_succ_order): Add prototype.
6704 * gimple.c (nonfreeing_call_p): Relocate here.
6705 * gimple.h: Add prototype.
6706 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
6707 * tree-ssa-dom.h: New file. Relocate prototypes here.
6708 * tree-ssa.h: Include tree-ssa-dom.h.
6710 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
6712 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
6713 Use $(COMPILE) and $(POSTCOMPILE).
6715 * config/alpha/x-alpha (driver-alpha.o): Ditto.
6717 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6719 * tree-flow.h: Remove some prototypes.
6720 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
6721 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6722 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
6723 propagate_tree_value*): Move from here to...
6724 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
6725 propagate_tree_value*): Relocate here.
6726 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
6727 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
6728 * gimple-fold.c: Remove gimple-fold.h from include list.
6729 * tree-vrp.c: Remove gimple-fold.h from include list.
6730 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
6731 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
6732 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
6733 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
6734 * sese.c: Add tree-ssa-propagate.h to include list.
6736 2013-10-02 Richard Biener <rguenther@suse.de>
6738 * tree-loop-distribution.c: Include tree-vectorizer.h for
6740 (enum partition_kind): Remove PKIND_REDUCTION.
6741 (struct partition_s): Remove has_writes member, add reduction_p member.
6742 (partition_alloc): Adjust.
6743 (partition_builtin_p): Likewise.
6744 (partition_has_writes): Remove.
6745 (partition_reduction_p): New function.
6746 (partition_merge_into): Likewise.
6747 (generate_code_for_partition): Commonize builtin partition
6749 (rdg_cannot_recompute_vertex_p): Remove.
6750 (already_processed_vertex_p): Likewise.
6751 (rdg_flag_vertex): Do not set has_writes.
6752 (classify_partition): Adjust.
6753 (rdg_build_partitions): Do not set has_writes, treat all
6754 partitions as useful.
6755 (distribute_loop): Record number of library calls generated. Adjust.
6756 (tree_loop_distribution): Report number of loops and library
6757 calls generated as opt-info.
6759 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
6761 * tree-flow.h: Include new .h files. Move prototypes.
6762 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
6763 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
6764 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
6765 * tree-pretty-print.h: Add prototypes from tree-flow.h.
6766 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
6767 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
6768 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
6769 * tree.h (get_ref_base_and_extent): Move prototype out.
6770 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
6772 * gimple-low.h: New File. Add prototypes from tree-flow.h.
6773 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
6774 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
6775 * tree-scalar-evolution.c: Include tree.h.
6776 * sese.c: Include tree.h.
6777 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
6778 * dwarf2out.c: Include tree-dfa.h.
6779 * tree-chrec.c: Include tree.h.
6780 * tree-data-ref.c: Include tree.h.
6782 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
6784 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
6787 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6789 * config/t-sol2 (sol2-c.o): Remove header dependencies.
6790 Use $(COMPILE) and $(POSTCOMPILE).
6791 (sol2-cxx.o): Likewise.
6792 (sol2-stubs.o): Likewise.
6794 * config/x-solaris (host-solaris.o): Likewise.
6796 * config/sparc/t-sparc (sparc.o): Remove.
6797 (sparc-c.o): Remove header dependencies.
6798 Use $(COMPILE) and $(POSTCOMPILE).
6799 * config/sparc/x-sparc: Likewise.
6801 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
6803 * config/arc/arc-opts.h: Add 2013 to Copyright years.
6804 * config/arc/arc700.md: Likewise.
6805 * config/arc/arc-modes.def: Likewise.
6806 * config/arc/arc-simd.h: Likewise.
6807 * config/arc/t-arc-uClibc: Likewise.
6808 * config/arc/t-arc-newlib: Likewise.
6810 2013-10-02 Renlin Li <renlin.li@arm.com>
6812 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
6814 (aarch64_expand_epilogue): Likewise.
6816 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
6817 Yufeng Zhang <yufeng.zhang@arm.com>
6819 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
6821 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
6822 'base_in' represent a conversion and legal_cast_p_1 holds; set
6823 'base_in' with the returned value from get_unwidened.
6825 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
6827 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
6828 plus_constant is not used.
6830 2013-10-01 Wei Mi <wmi@google.com>
6832 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
6833 * config/i386/i386.md: Add define_peephole2 to
6834 break partial reg stall for cvtss2sd/cvtsd2ss.
6836 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6838 * config/arc/arc.c (pass_arc_ifcvt::clone):
6839 Update for ctxt_ -> m_ctxt change.
6841 2013-10-01 Jeff Law <law@redhat.com>
6843 * tree-ssa-threadupdate.c (struct redirection_data): Delete
6844 outgoing_edge and intermediate_edge fields. Instead store the path.
6845 (redirection_data::hash): Hash on the last edge's destination index.
6846 (redirection_data::equal): Check the entire thread path.
6847 (lookup_redirectio_data): Corresponding changes.
6848 (create_edge_and_update_destination_phis): Likewise.
6849 (thread_single_edge): Likewise.
6851 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6852 Diego Novillo <dnovillo@google.com>
6854 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
6855 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
6856 (vld32wh_insn, vld32wl_insn): Delete commented-out old
6857 versions of these patterns.
6859 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
6860 (__builtin_arc_aligned): Likewise.
6862 * config/arc/arc.md: Expand adc_0 comment stating the intended
6863 purpose and why it isn't ready.
6864 Replace commented out call_value_via_label_mixed with a
6865 plain comment about bl_s.
6867 * config/arc/arc.c (stdio.h): Don't include directly.
6868 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
6869 Remove if (1) condition.
6870 (arc_encode_section_info): Fix comment.
6872 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6874 * config/arc/arc.c (arc_conditional_register_usage):
6875 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
6876 Also set reg_alloc_order for DMA config regs.
6878 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6879 Jeremy Bennett <jeremy.bennett@embecosm.com>
6881 * doc/install.texi (--with-cpu): Mention ARC.
6882 (arc-*-elf32): New paragraph.
6883 (arc-linux-uclibc): Likewise.
6884 * doc/md.texi (Machine Constraints): Add ARC part.
6885 * doc/invoke.texi: (menu): Add ARC Options.
6886 (Machine Dependent Options) <ARC Options>: Add synopsis.
6887 (node ARC Options): Add.
6888 * doc/extend.texi (long_call / short_call attribute): Add ARC.
6889 (ARC Built-in Functions): New section defining
6890 generic ARC built-in functions.
6891 (ARC SIMD Built-in Functions): New section defining SIMD specific
6893 (Declaring Attributes of Functions): Extended
6894 description of short_call and long_call attributes for ARC and
6895 added index entries.
6897 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
6898 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
6899 Joern Rennecke <joern.rennecke@embecosm.com>
6900 Muhammad Khurram Riaz <khurram.riaz@arc.com>
6901 Brendan Kehoe <brendan@zen.org>
6902 Michael Eager <eager@eagercon.com>
6903 Simon Cook <simon.cook@embecosm.com>
6904 Jeremy Bennett <jeremy.bennett@embecosm.com>
6906 * config/arc, common/config/arc: New directories.
6908 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
6909 Brendan Kehoe <brendan@zen.org>
6910 Simon Cook <simon.cook@embecosm.com>
6912 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
6914 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6916 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
6917 * tree-ssa-coalesce.h: New. Move prototype to here.
6918 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
6919 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
6920 (gimple_can_coalesce_p): Move to...
6921 * gimple.c (gimple_can_coalesce_p): Here.
6923 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
6925 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
6926 (dump_decl_set): Move to gimple.c.
6927 * gimple.h: Don't include tree-ssa-operands.h.
6928 (dump_decl_set): Add prototype.
6929 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6930 Move to gimple-ssa.h.
6931 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
6932 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
6933 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
6934 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
6935 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
6937 (dump_decl_set): Relocate here.
6938 * gimple-ssa.h: New file.
6939 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
6940 Relocate from gimple.h.
6941 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
6942 * tree-ssa-operands.c (swap_ssa_operands): Rename from
6943 swap_tree_operands and remove non-ssa path.
6944 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
6945 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
6947 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
6948 vect_is_simple_reduction_1): Use swap_ssa_operands.
6949 * tree-flow.h: Move various prototypes to tree-phinodes.h.
6950 (enum need_phi_state): Move to tree-into-ssa.c.
6951 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6952 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
6953 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
6954 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
6955 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6956 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6957 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6958 num_imm_uses): Move to ssa-iterators.h.
6959 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
6960 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
6962 (op_iter_done, op_iter_next_def, op_iter_next_tree,
6963 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
6964 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
6965 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
6966 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
6967 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
6968 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
6969 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
6970 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
6971 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
6972 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
6973 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
6974 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
6975 (set_phi_nodes): Move to tree-phinodes.h.
6976 * tree-ssa-operands.h (enum ssa_op_iter_type,
6977 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
6978 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
6979 (dump_decl_set): Remove prototype.
6980 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
6981 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
6982 (set_phi_nodes): Relocate from tree-flow-inline.h.
6983 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
6985 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
6986 include list. Temporarily add gimple.h to include list.
6987 * ssa-iterators.h: New file.
6988 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
6989 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
6990 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
6991 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
6992 Relocate from tree-ssa-operands.h.
6993 (delink_imm_use, link_imm_use_to_list, link_imm_use,
6994 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
6995 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
6996 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
6997 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
6998 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
6999 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
7000 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
7001 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
7002 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
7003 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
7004 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
7005 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
7006 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
7007 Relocate from tree-flow-inline.h.
7008 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
7010 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
7013 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
7014 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
7015 Insert '\t' to output template.
7016 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
7017 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
7018 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
7019 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
7021 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
7023 * doc/install.texi (Host/target specific installation notes for GCC):
7024 Put @anchor before @heading.
7025 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
7026 Use @email for email addresses.
7028 2013-10-01 Jeff Law <law@redhat.com>
7030 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
7031 a vec. Only delete the path if we create one without successfully
7032 registering a jump thread.
7033 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
7035 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
7036 (paths): New vector of jump threading paths.
7037 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
7038 (THREAD_PATH): New accessor macro for the entire thread path.
7039 (lookup_redirection_data): Get intermediate and final outgoing edge
7040 from the thread path.
7041 (create_edge_and_update_destination_phis): Copy the threading path.
7042 (ssa_fix_duplicate_block_edges): Get edges and block types from the
7043 jump threading path.
7044 (ssa_redirect_edges): Get edges and block types from the jump threading
7045 path. Free the path vector.
7046 (thread_block): Get edges from the jump threading path. Look at the
7047 entire path to see if we thread to a loop exit. If we cancel a jump
7048 thread request, then free the path vector.
7049 (thread_single_edge): Get edges and block types from the jump threading
7050 path. Free the path vector.
7051 (thread_through_loop_header): Get edges and block types from the jump
7052 threading path. Free the path vector.
7053 (mark_threaded_blocks): Iterate over the vector of paths and store
7054 the path on the appropriate edge. Get edges and block types from the
7055 jump threading path.
7056 (mark_threaded_blocks): Get edges and block types from the jump
7057 threading path. Free the path vector.
7058 (thread_through_all_blocks): Use the vector of paths rather than
7059 a vector of 3-edge sets.
7060 (register_jump_thread): Accept pointer to a path vector rather
7061 than the path vector itself. Store the path vector for later use.
7064 2013-10-01 Jakub Jelinek <jakub@redhat.com>
7065 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7068 * config/s390/s390.c (s390_split_branches): Modify check for table
7070 (s390_chunkify_start): Rearrange table jump insn check in order to
7071 deal with compare and branch insns correctly.
7073 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
7077 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
7078 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
7079 define_insn_and_split.
7080 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
7081 (shiftsi3_compare): New pattern.
7083 * config/arm/unspecs.md (UNSPEC_RRX): New.
7085 2013-10-01 Alan Modra <amodra@gmail.com>
7087 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
7088 casing inout operands.
7090 2013-10-01 Richard Biener <rguenther@suse.de>
7092 PR tree-optimization/58553
7093 * tree-loop-distribution.c (struct partition_s): Add niter member.
7094 (classify_partition): Populate niter member for the partition
7095 and properly identify whether the relevant store happens before
7096 or after the loop exit.
7097 (generate_memset_builtin): Use niter member from the partition.
7098 (generate_memcpy_builtin): Likewise.
7100 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7102 * vec.h (vec_prefix, vec): Prefix member names with "m_".
7103 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
7105 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
7107 * basic-block.h (edge_list): Prefix member names with "m_".
7108 * context.h (context): Likewise.
7109 * domwalk.h (dom_walker): Likewise.
7110 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
7111 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
7112 * hash-table.h (hash_table): Likewise.
7113 * machmode.h (bit_field_mode_iterator): Likewise.
7114 * pass_manager.h (pass_list): Likewise.
7115 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
7116 * tree-pass.h (pass_data): Likewise.
7117 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
7118 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
7119 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
7120 * asan.c (pass_data_asan): Update accordingly.
7121 * cfganal.c (control_dependences::find_control_dependence): Likewise.
7122 (control_dependences::control_dependences): Likewise.
7123 (control_dependences::~control_dependences): Likewise.
7124 (control_dependences::~control_dependences): Likewise.
7125 (control_dependences::get_edges_dependent_on): Likewise.
7126 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
7127 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
7128 * context.c (gcc::context::context): Likewise.
7129 * cprop.c (pass_rtl_cprop::clone): Likewise.
7130 * domwalk.c (dom_walker::walk): Likewise.
7131 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
7132 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
7133 * mode-switching.c (pass_mode_switching::clone): Likewise.
7134 * passes.c (opt_pass::opt_pass): Likewise.
7135 (pass_manager::pass_manager): Likewise.
7136 * predict.c (pass_strip_predict_hints::clone): Likewise.
7137 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
7138 (pass_split_all_insns::clone): Likewise.
7139 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
7141 (bit_field_mode_iterator::next_mode): Likewise.
7142 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
7143 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
7144 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
7145 * tree-complex.c (pass_lower_complex::clone): Likewise.
7146 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
7147 * tree-object-size.c (pass_object_sizes::clone): Likewise.
7148 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
7149 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
7150 (pass_fold_builtins::clone): Likewise.
7151 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
7152 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
7153 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
7154 (pass_cd_dce::clone): Likewise.
7155 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
7156 (pass_phi_only_cprop::clone): Likewise.
7157 * tree-ssa-dse.c (pass_dse::clone): Likewise.
7158 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
7159 * tree-ssa-loop.c (pass_lim::clone): Likewise.
7160 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
7161 * tree-ssa-pre.c (pass_fre::clone): Likewise.
7162 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
7163 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
7164 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
7165 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
7166 * tree-vrp.c (pass_vrp::clone): Likewise.
7167 * tsan.c (pass_tsan::clone): Likewise.
7169 2013-09-30 Jakub Jelinek <jakub@redhat.com>
7172 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
7173 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
7176 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
7177 optimization, punt if sign_bit_p looked through any zero extension.
7179 2013-09-30 Teresa Johnson <tejohnson@google.com>
7181 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
7182 Update redirected out edge count in joiner case.
7183 (ssa_redirect_edges): Common the joiner and non-joiner cases
7184 so that joiner case gets profile updates.
7186 2013-09-30 Richard Biener <rguenther@suse.de>
7188 PR tree-optimization/58554
7189 * tree-loop-distribution.c (classify_partition): Require
7190 unconditionally executed stores for memcpy and memset recognition.
7191 (tree_loop_distribution): Calculate dominance info.
7193 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
7195 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
7196 (NO_PROFILE_COUNTERS): Likewise.
7197 (PROFILE_HOOK): Likewise.
7198 (FUNCTION_PROFILER): Likewise.
7199 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
7201 2013-09-30 Iain Sandoe <iain@codesourcery.com>
7203 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
7204 calls and defines in TARGET_MACHO conditional.
7205 (load_macho_picbase_di): Likewise.
7206 (reload_macho_picbase): Likewise.
7207 (reload_macho_picbase_si): Likewise.
7208 (reload_macho_picbase_di): Likewise.
7209 (nonlocal_goto_receiver): Likewise.
7211 2013-09-30 Nick Clifton <nickc@redhat.com>
7213 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
7214 that use the MSP430X ISA.
7215 (msp430_option_override): Scan -mmcu command line option for any
7216 MCU name that supports the MSP430X ISA.
7217 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
7218 -mmcu options which enable the MSP430X ISA.
7220 2013-09-30 Richard Biener <rguenther@suse.de>
7223 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
7224 before looking for setjmp-like calls.
7226 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7229 * config/darwin-protos.h (machopic_get_function_picbase): New.
7230 * config/darwin.c (machopic_get_function_picbase): New.
7231 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
7232 label for a new func. (load_macho_picbase_di): Likewise.
7233 (reload_macho_picbase): New expand.
7234 (reload_macho_picbase_si): New insn.
7235 (reload_macho_picbase_di): New insn.
7236 (nonlocal_goto_receiver): New define and split.
7237 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
7238 (unspecv enum): Add UNSPECV_NLGR.
7240 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7242 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
7243 that altivec registers are correctly sized on Darwin.
7245 2013-09-29 Iain Sandoe <iain@codesourcery.com>
7247 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
7248 darwin-driver.o): Use COMPILE and POSTCOMPILE.
7249 * config/x-darwin (host-darwin.o): Likewise.
7250 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
7251 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
7252 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
7254 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
7256 * doc/invoke.texi: Fix usage of @tie{} command.
7258 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
7260 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
7262 2013-09-28 Jan Hubicka <jh@suse.cz>
7264 * config/i386/x86-tune.def: Add documentation for each of the options;
7267 2013-09-28 Jan Hubicka <jh@suse.cz>
7269 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
7271 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
7272 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
7273 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
7274 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
7275 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
7277 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
7279 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
7280 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
7281 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
7282 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
7283 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
7284 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
7285 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
7286 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
7287 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
7288 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
7289 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
7290 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
7291 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
7292 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
7293 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
7294 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
7295 gimple.h, godump.c, graphite-clast-to-gimple.c,
7296 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
7297 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
7298 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
7299 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
7300 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
7301 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
7302 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
7303 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
7304 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
7305 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
7306 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
7307 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
7308 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
7309 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
7310 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
7311 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
7312 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
7313 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
7314 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
7315 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
7316 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
7317 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
7318 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
7319 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
7320 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
7321 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
7322 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
7323 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
7324 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
7325 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
7326 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
7327 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
7328 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
7329 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
7330 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
7331 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
7332 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
7333 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
7334 whitespace before "(".
7336 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
7338 * expr.h (extract_bit_field): Remove packedp parameter.
7339 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
7340 from forward declaration.
7341 (store_split_bit_field): Remove packedp arg from calls to
7342 extract_fixed_bit_field.
7343 (extract_bit_field_1): Remove packedp parameter and packedp
7344 argument from recursive calls and calls to extract_fixed_bit_field.
7345 (extract_bit_field): Remove packedp parameter and corresponding
7346 arg to extract_bit_field_1.
7347 (extract_fixed_bit_field): Remove packedp parameter. Remove code
7349 (extract_split_bit_field): Remove packedp arg from call to
7350 extract_fixed_bit_field.
7351 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
7352 (copy_blkmode_from_reg): Likewise.
7353 (copy_blkmode_to_reg): Likewise.
7354 (read_complex_part): Likewise.
7355 (store_field): Likewise.
7356 (expand_expr_real_1): Likewise.
7357 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
7358 to extract_bit_field.
7359 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
7360 call to extract_bit_field.
7361 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
7362 call to extract_bit_field.
7363 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
7364 and special packedp behavior from -fstrict-volatile-bitfields
7367 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7369 * lra-eliminations.c (init_elim_table): Guard value_p.
7371 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
7373 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
7374 DFmode, DImode, and SFmode in the upper VSX registers based on the
7375 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
7376 if -mpower8-vector. Combine -mvsx-timode handling with the rest
7377 of the VSX register handling.
7379 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
7381 (zero_extendsidi2_lfiwzx): Add support for loading into the
7382 Altivec registers with -mpower8-vector. Use wu/wv constraints to
7383 only do VSX memory options on Altivec registers.
7384 (extendsidi2_lfiwax): Likewise.
7385 (extendsfdf2_fpr): Likewise.
7386 (mov<mode>_hardfloat, SF/SD modes): Likewise.
7387 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
7388 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
7389 (movdi_internal64): Likewise.
7391 2013-09-27 Xinliang David Li <davidxl@google.com>
7393 * opts.c (finish_options): Adjust parameters
7394 according to vect cost model.
7395 (common_handle_option): Set dynamic vect cost
7397 targhooks.c (default_add_stmt_cost): Compute stmt cost
7399 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
7400 Use helper function.
7401 * tree-vectorizer.h (unlimited_cost_model): New function.
7402 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
7403 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
7405 (vect_enhance_data_refs_alignment): Ditto.
7406 * flag-types.h: New enum.
7407 * common/config/i386/i386-common.c (ix86_option_init_struct):
7408 No need to initialize vect_cost_model flag.
7409 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
7412 2013-09-27 Diego Novillo <dnovillo@google.com>
7414 * gimple.h (enum ssa_mode): Remove.
7416 2013-09-27 Paulo Matos <pmatos@broadcom.com>
7418 * cfgloop.h (number_of_loops): Fix typo in check for null.
7420 2013-09-27 Jakub Jelinek <jakub@redhat.com>
7423 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
7424 are children of outermost saved_cfun's loop, and set it up to
7425 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
7426 if loop != loop0 and SESE region contains bbs that belong to loop0.
7428 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7430 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
7431 (binary_scale_code_p, get_base_term, get_index_term): New functions.
7432 (set_address_segment, set_address_base, set_address_index)
7433 (set_address_disp): Accept the argument unconditionally.
7434 (baseness): Remove must_be_base_p and must_be_index_p checks.
7435 (decompose_normal_address): Classify as much as possible in the
7438 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
7440 * cse.c (count_reg_usage): Handle INT_LIST.
7441 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
7442 * reginfo.c (reg_scan_mark_refs): Likewise.
7443 * reload1.c (eliminate_regs_1): Likewise.
7445 2013-09-27 Iain Sandoe <iain@codesourcery.com>
7448 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
7451 2013-09-27 Richard Biener <rguenther@suse.de>
7453 PR tree-optimization/58459
7454 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
7455 restriction not propagating into loops.
7457 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
7459 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
7460 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
7461 * doc/tree-ssa.texi (Walking use-def chains): Delete.
7463 2013-09-26 Richard Biener <rguenther@suse.de>
7465 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
7467 2013-09-26 Richard Biener <rguenther@suse.de>
7469 * alias.h (component_uses_parent_alias_set): Rename to ...
7470 (component_uses_parent_alias_set_from): ... this.
7471 * alias.c (component_uses_parent_alias_set): Rename to ...
7472 (component_uses_parent_alias_set_from): ... this and return
7474 (reference_alias_ptr_type_1): Use the result from
7475 component_uses_parent_alias_set_from instead of stripping
7476 components one at a time.
7477 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
7479 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7481 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
7482 Move prototypes to...
7483 * tree-ssa-ter.h: New File. Move prototypes here.
7484 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
7485 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
7486 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7487 from is_replaceable_p.
7488 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
7489 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
7490 (process_replaceable): Use ter_is_replaceable_p.
7491 (find_replaceable_in_bb): Use ter_is_replaceable_p.
7492 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
7493 newly refactored ssa_is_replaceable_p.
7494 * cfgexpand.c: Include tree-outof-ssa.h.
7495 * ssaexpand.h: Delete.
7497 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
7499 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
7500 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
7502 (create_gimple_tmp): Delete.
7503 (get_expr_type, build_assign, build_type_cast): Move to...
7504 * gimple-builder.c: New File.
7505 (get_expr_type): Relocate from gimple.c.
7506 (build_assign, build_type_cast): Change to only create ssanames.
7507 * gimple.h: Move prototypes to...
7508 * gimple-builder.h: New File. Here.
7509 * tree-ssa.h: And here.
7510 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
7511 count_uses_and_derefs): Relocate from gimple.c.
7512 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
7513 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
7514 * tree-ssa-math-opts (execute_cse_reciprocals): Use
7515 gimple_replace_ssa_lhs.
7516 * asan.c: Include gimple-builder.h.
7517 * Makefile.in: Add gimple-builder.o.
7519 2013-09-26 Richard Biener <rguenther@suse.de>
7521 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
7522 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
7523 (loe_visit_block): Use gcc_checking_assert.
7524 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
7525 gimple_assign_ssa_name_copy_p.
7526 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
7529 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
7531 * config/rs6000/t-rs6000 (rs6000.o): Remove.
7532 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
7534 2013-09-26 Richard Biener <rguenther@suse.de>
7536 PR tree-optimization/58539
7537 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
7538 the fact that debug statements are not taking part in loop-closed
7541 2013-09-26 Nick Clifton <nickc@redhat.com>
7543 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
7544 time warning message.
7545 (msp430_print_operand_raw): Delete unused letter parameter.
7546 (TARGET_PRINT_OPERAND_ADDRESS): Define.
7547 (msp430_print_operand_address): New function.
7548 (msp430_print_operand): Move address printing code from here to
7550 * config/msp430/msp430.md (movsipsi2): Add comment in generated
7552 (zero_extendpsisi2): Likewise.
7553 (extendpsisi2): New pattern.
7554 (andneghi3): New pattern.
7556 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
7558 * config/aarch64/aarch64.opt (mlra): New option.
7559 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
7560 (TARGET_LRA_P): Define.
7562 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
7564 * expr.c (expand_assignment): Remove obsolete comment.
7566 2013-09-25 Jeff Law <law@redhat.com>
7568 * tree-flow.h (thread_through_all_blocks): Prototype moved into
7569 tree-ssa-threadupdate.h.
7570 (register_jump_thread): Similarly.
7571 * tree-ssa-threadupdate.h: New header file.
7572 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
7573 * tree-vrp.c: Likewise.
7574 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
7575 (thread_around_empty_blocks): Change type of path vector argument to
7576 an edge,type pair from just an edge. Initialize both elements when
7577 appending to a jump threading path. Tweak references to elements
7579 (thread_across_edge): Similarly. Release memory for the elements
7581 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
7582 (dump_jump_thread_path): New function broken out from
7583 register_jump_thread.
7584 (register_jump_thread): Use dump_jump_thread_path. Change type of
7585 path vector entries. Search the path for NULL edges and dump
7586 the path if one is found. Tweak the conversion of path to 3-edge
7587 form to use the block copy type information embedded in the path.
7589 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7591 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
7593 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
7594 Vladimir Makarov <vmakarov@redhat.com>
7596 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
7597 from the least significant bit.
7598 (strip_address_mutations): Add bitfield operations handling.
7599 (must_be_index_p): Add shifting and rotate operations handling.
7600 (set_address_base): Use must_be_base_p predicate.
7601 (set_address_index): Use must_be_index_p predicate.
7603 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
7604 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7605 Sergey Lega <sergey.s.lega@intel.com>
7606 Anna Tikhonova <anna.tikhonova@intel.com>
7607 Ilya Tocar <ilya.tocar@intel.com>
7608 Andrey Turetskiy <andrey.turetskiy@intel.com>
7609 Ilya Verbin <ilya.verbin@intel.com>
7610 Kirill Yukhin <kirill.yukhin@intel.com>
7611 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7613 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
7615 (ix86_expand_vector_move_misalign): Support new unaligned load and
7616 stores and use new names.
7617 (CODE_FOR_sse2_storedqu): Rename to ...
7618 (CODE_FOR_sse2_storedquv16qi): ... this.
7619 (CODE_FOR_sse2_loaddqu): Rename to ...
7620 (CODE_FOR_sse2_loaddquv16qi): ... this.
7621 (CODE_FOR_avx_loaddqu256): Rename to ...
7622 (CODE_FOR_avx_loaddquv32qi): ... this.
7623 (CODE_FOR_avx_storedqu256): Rename to ...
7624 (CODE_FOR_avx_storedquv32qi): ... this.
7625 * config/i386/i386.md (fpint_logic): New.
7626 * config/i386/sse.md (VMOVE): Extend for AVX512.
7630 (VI_UNALIGNED_LOADSTORE): Ditto.
7631 (sse2_avx_avx512f): Ditto.
7632 (sse2_avx2): Extend for AVX512.
7633 (sse4_1_avx2): Ditto.
7634 (avx2_avx512f): New.
7635 (sse): Extend for AVX512.
7638 (avxsizesuffix): Ditto.
7639 (sseintvecmode): Ditto.
7641 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
7642 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
7643 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7644 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
7645 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
7646 (<sse2_avx_avx512f>_storedqu<mode): ... this.
7647 (<sse>_movnt<mode>): Replace constraint "x" with "v".
7648 (STORENT_MODE): Extend for AVX512.
7649 (*absneg<mode>2): Replace constraint "x" with "v".
7650 (*mul<mode>3): Ditto.
7651 (*ieee_smin<mode>3): Ditto.
7652 (*ieee_smax<mode>3): Ditto.
7653 (avx_cmp<mode>3): Replace VF with VF_128_256.
7654 (*<sse>_maskcmp<mode>3_comm): Ditto.
7655 (<sse>_maskcmp<mode>3): Ditto.
7656 (<sse>_andnot<mode>3): Extend for AVX512.
7657 (<code><mode>3, anylogic): Replace VF with VF_128_256.
7658 (<code><mode>3, fpint_logic): New.
7659 (*<code><mode>3): Extend for AVX512.
7660 (avx512flogicsuff): New.
7661 (avx512f_<logic><mode>): Ditto.
7662 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
7664 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
7665 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
7666 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
7667 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
7668 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
7669 (xop_vpermil2<mode>3): Ditto.
7670 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
7671 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
7672 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
7673 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
7675 2013-09-25 Tom Tromey <tromey@redhat.com>
7677 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
7678 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
7679 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
7680 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
7681 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
7682 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
7683 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
7684 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
7685 (GRAPHITE_HTAB_H): Remove.
7687 2013-09-25 Tom Tromey <tromey@redhat.com>
7689 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
7691 2013-09-25 Tom Tromey <tromey@redhat.com>
7693 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
7695 2013-09-25 Tom Tromey <tromey@redhat.com>
7697 * config/i386/t-i386 (i386.o): Remove.
7698 (i386-c.o): Use COMPILE and POSTCOMPILE.
7700 2013-09-25 Tom Tromey <tromey@redhat.com>
7702 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
7704 2013-09-25 Tom Tromey <tromey@redhat.com>
7706 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
7707 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
7708 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
7709 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
7710 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
7711 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
7712 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
7713 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
7714 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
7715 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
7716 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
7717 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
7718 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
7719 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
7720 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
7721 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
7722 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
7723 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
7724 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
7725 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
7726 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
7727 (tree-ssa-pre.o, tree-ssa-sccvn.o)
7728 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
7729 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
7730 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
7731 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
7732 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
7733 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
7734 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
7735 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
7736 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
7737 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
7738 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
7739 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
7740 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
7741 (graphite-clast-to-gimple.o, graphite-dependences.o)
7742 (graphite-interchange.o, graphite-poly.o)
7743 (graphite-scop-detection.o, graphite-sese-to-poly.o)
7744 (graphite-optimize-isl.o, tree-vect-loop.o)
7745 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
7746 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
7747 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
7748 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
7749 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
7750 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
7751 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
7752 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
7753 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
7754 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
7755 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
7756 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
7757 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
7758 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
7759 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
7760 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
7761 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
7762 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
7763 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
7764 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
7765 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
7766 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
7767 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
7768 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
7769 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
7770 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
7771 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
7772 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
7773 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
7774 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
7775 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
7776 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
7777 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
7778 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
7779 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
7780 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
7781 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
7782 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
7783 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
7784 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
7785 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
7786 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
7787 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
7788 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
7789 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
7790 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
7791 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
7792 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
7793 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
7794 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
7795 (gcov-dump.o): Remove.
7796 (default-c.o): Use COMPILE and POSTCOMPILE.
7797 (CFLAGS-gcc.o): New variable.
7798 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
7800 2013-09-25 Tom Tromey <tromey@redhat.com>
7802 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
7803 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
7804 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
7805 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
7806 (c-family/c-pragma.o, c-family/c-pretty-print.o)
7807 (c-family/c-semantics.o, c-family/c-ada-spec.o)
7808 (c-family/array-notation-common.o, c-family/stub-objc.o)
7809 (c-family/c-ubsan.o): Remove.
7811 2013-09-25 Tom Tromey <tromey@redhat.com>
7813 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
7815 2013-09-25 Tom Tromey <tromey@redhat.com>
7817 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
7818 to add -DENABLE_SHARED_LIBGCC.
7819 (gcc.o): Don't use subshell.
7821 2013-09-25 Tom Tromey <tromey@redhat.com>
7823 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
7824 * configure.ac: Don't invoke AM_PROG_CC_C_O.
7825 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
7826 * configure, config.in: Rebuild.
7828 2013-09-25 Tom Tromey <tromey@redhat.com>
7830 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
7831 (COMPILE, POSTCOMPILE): New variables.
7832 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
7833 (DEPFILES): New variable.
7834 Include ".Po" files.
7835 * configure.ac: Add checks for dependency checking.
7836 * configure, aclocal.m4: Regenerate.
7838 2013-09-25 Tom Tromey <tromey@redhat.com>
7840 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7841 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
7843 2013-09-25 Tom Tromey <tromey@redhat.com>
7845 * Makefile.in (generated_files): Add options.h,
7846 target-hooks-def.h, insn-opinit.h,
7847 common/common-target-hooks-def.h, pass-instances.def,
7848 c-family/c-target-hooks-def.h.
7850 2013-09-25 Jeff Law <law@redhat.com>
7852 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
7853 than foo[foo.length () - 1] to access last member in a vec.
7854 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
7856 2013-09-25 Richard Biener <rguenther@suse.de>
7859 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
7861 2013-09-25 Jan Hubicka <jh@suse.cz>
7863 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
7866 2013-09-25 Marek Polacek <polacek@redhat.com>
7869 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
7870 when determining the type name.
7872 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
7874 * config/sh/sh.md: Fix formatting.
7876 2013-09-24 Xinliang David Li <davidxl@google.com>
7878 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
7879 max peel iterations parameter.
7880 * param.def: New parameter.
7881 * doc/invoke.texi: Document New parameter.
7883 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
7885 * gimple-pretty-print.c: Various whitespace tweaks.
7886 * tree-core.h: Likewise.
7887 * tree-pretty-print.c: Likewise.
7888 * tree-ssa-alias.c: Likewise.
7889 * tree-ssa-copy.c: Likewise.
7890 * tree-ssanames.c: Likewise.
7891 * tree-ssanames.h: Likewise.
7892 * tree-vrp.c: Likewise.
7894 2013-09-24 Alan Modra <amodra@gmail.com>
7898 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
7899 for output operands that disallow regs. Don't use EXPAND_WRITE on
7902 2013-09-24 Richard Biener <rguenther@suse.de>
7905 * tree.c (reference_alias_ptr_type): Move ...
7906 * alias.c (reference_alias_ptr_type): ... here and implement
7907 in terms of the new reference_alias_ptr_type_1.
7908 (ref_all_alias_ptr_type_p): New helper.
7909 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
7910 use ref_all_alias_ptr_type_p.
7911 (get_deref_alias_set): Add flag_strict_aliasing check here.
7912 (reference_alias_ptr_type_1): New function, split out from ...
7913 (get_alias_set): ... here.
7914 (alias_ptr_types_compatible_p): New function.
7915 * alias.h (reference_alias_ptr_type): Declare.
7916 (alias_ptr_types_compatible_p): Likewise.
7917 * tree.h (reference_alias_ptr_type): Remove.
7918 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
7919 to compare MEM_REF alias types.
7921 2013-09-24 Richard Biener <rguenther@suse.de>
7923 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
7925 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
7927 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
7928 reload helper function arrays into a single array reg_addr.
7929 (reload_fpr_gpr): Likewise.
7930 (reload_gpr_vsx): Likewise.
7931 (reload_vsx_gpr): Likewise.
7932 (struct rs6000_reg_addr): Likewise.
7933 (reg_addr): Likewise.
7934 (rs6000_debug_reg_global): Change rs6000_vector_reload,
7935 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
7936 (rs6000_init_hard_regno_mode_ok): Likewise.
7937 (rs6000_secondary_reload_direct_move): Likewise.
7938 (rs6000_secondary_reload): Likewise.
7940 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
7941 constraints: wu, ww, and wy. Repurpose wv constraint added during
7942 power8 changes. Put wg constraint in alphabetical order.
7944 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
7945 for future work to add ISA 2.07 VSX single precision support.
7946 (-mvsx-scalar-double): Change default from -1 to 1, update
7947 documentation comment.
7948 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
7949 (-mupper-regs-df): New debug switch to control whether DF values
7950 can go in the traditional Altivec registers.
7951 (-mupper-regs-sf): New debug switch to control whether SF values
7952 can go in the traditional Altivec registers.
7954 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
7956 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
7957 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
7958 Add new constraints, wu/ww/wy. Repurpose wv constraint.
7959 (rs6000_debug_legitimate_address_p): Print if we are running
7960 before, during, or after reload.
7961 (rs6000_secondary_reload): Add a comment.
7962 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
7964 * config/rs6000/constraints.md (wa constraint): Sort w<x>
7965 constraints. Update documentation string.
7966 (wd constraint): Likewise.
7967 (wf constraint): Likewise.
7968 (wg constraint): Likewise.
7969 (wn constraint): Likewise.
7970 (ws constraint): Likewise.
7971 (wt constraint): Likewise.
7972 (wx constraint): Likewise.
7973 (wz constraint): Likewise.
7974 (wu constraint): New constraint for ISA 2.07 SFmode scalar
7976 (ww constraint): Likewise.
7977 (wy constraint): Likewise.
7978 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
7979 in the previous submissions.
7980 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
7982 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
7984 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
7985 (REG_BR_PROB): Say that the probability is stored in an int_list.
7986 * reg-notes.def: Update commentary to mention INT_LIST.
7987 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
7988 (INT_LIST): New rtx.
7989 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
7990 * rtlanal.c (int_reg_note_p): New function.
7991 (alloc_reg_note): Assert that the note does not have an int argument.
7992 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
7993 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
7994 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
7995 rather than an INSN_LIST. Handle INT_LIST.
7996 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
7997 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
7998 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
7999 Manipulate them as ints rather than rtxes.
8000 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
8001 * regmove.c (copy_src_to_dest): Likewise.
8002 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
8004 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
8005 into the cases that need it.
8006 * config/arm/arm.c (arm_unwind_emit): Likewise.
8008 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
8009 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
8010 * loop-doloop.c (add_test, doloop_modify): Likewise.
8011 * loop-unswitch.c (compare_and_jump_seq): Likewise.
8012 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
8013 * predict.c (combine_predictions_for_insn): Likewise.
8014 * print-rtl.c (print_rtx): Handle INT_LIST.
8015 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
8016 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
8017 * config/arm/arm.c (emit_unlikely_jump): Likewise.
8018 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
8019 (ix86_split_fp_branch, predict_jump): Likewise.
8020 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
8021 * config/sh/sh.c (expand_cbranchsi4): Likewise.
8022 * config/spu/spu.c (ea_load_store_inline): Likewise.
8024 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
8025 value of a REG_BR_PROB note.
8026 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
8027 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
8028 * emit-rtl.c (try_split): Likewise.
8029 * predict.c (br_prob_note_reliable_p): Likewise.
8030 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
8031 * reorg.c (mostly_true_jump): Likewise.
8032 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
8033 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
8034 * config/i386/i386.c (ix86_print_operand): Likewise.
8035 * config/ia64/ia64.c (ia64_print_operand): Likewise.
8036 * config/mmix/mmix.c (mmix_print_operand): Likewise.
8037 * config/rs6000/rs6000.c (output_cbranch): Likewise.
8038 * config/s390/s390.c (s390_expand_tbegin): Likewise.
8039 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
8040 * config/sparc/sparc.c (output_cbranch): Likewise.
8041 * config/spu/spu.c (get_branch_target): Likewise.
8042 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
8043 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
8045 2013-09-23 Jan Hubicka <jh@suse.cz>
8047 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
8049 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
8050 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
8052 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
8053 (gimple_fold_call): Dump inconsistent devirtualizations; add
8054 sanity check for type based devirtualizations.
8055 * ipa-prop.c: Include ipa-utils.h
8056 (ipa_intraprocedural_devirtualization): Add sanity check.
8057 (try_make_edge_direct_virtual_call): Likewise.
8059 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
8061 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
8062 assignment statements.
8064 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
8066 * gimple-pretty-print.c (dump_ssaname_info): New function.
8067 (dump_gimple_phi): Call it.
8068 (pp_gimple_stmt_1): Likewise.
8069 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
8070 (range_info_def): Declare.
8071 * tree-pretty-print.c (pp_double_int): New function.
8072 (dump_generic_node): Call it.
8073 * tree-pretty-print.h (pp_double_int): Declare.
8074 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
8075 * tree-ssanames.h (range_info_def): New structure.
8076 (value_range_type): Move definition here.
8077 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
8079 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
8081 (set_range_info): New function.
8082 (get_range_info): Likewise.
8083 (duplicate_ssa_name_range_info): Likewise.
8084 (duplicate_ssa_name_fn): Check pointer type and call
8085 duplicate_ssa_name_range_info.
8086 * tree-ssa-copy.c (fini_copy_prop): Likewise.
8087 * tree-vrp.c (value_range_type): Remove definition, now in
8089 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
8090 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
8091 (SSA_NAME_RANGE_INFO): New macro.
8093 2013-09-23 Richard Biener <rguenther@suse.de>
8095 PR tree-optimization/58464
8096 * tree-ssa-pre.c (phi_trans_lookup): Remove.
8097 (phi_trans_add): Change to add conditionally on being not present.
8098 (phi_translate_1): Remove recursion detection here.
8099 (phi_translate): Pre-seed the cache with NULL to catch
8100 recursion here in a more generic way.
8101 (bitmap_find_leader): Adjust comment.
8102 (get_representative_for): Dump value-numbers.
8103 (create_expression_by_pieces): Likewise.
8104 (insert_into_preds_of_block): Likewise.
8106 2013-09-23 Christian Bruel <christian.bruel@st.com>
8109 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
8110 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
8112 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
8115 2013-09-20 Renlin Li <renlin.li@arm.com>
8117 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
8119 (aarch64_expand_epilogue): Likewise.
8120 (aarch64_legitimize_reload_address): Likewise.
8122 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
8124 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
8125 NULL_TREE before pushing them onto the vector. Likewise for labels.
8127 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
8129 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
8130 comparison operators when -fno-trapping-math is in effect.
8131 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
8132 comparison operators in TFmode and assert that unsupported operators
8134 (ia64_print_operand): Likewise.
8136 2013-09-21 Jan Hubicka <jh@suse.cz>
8138 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
8139 (sse_typeless_stores): Enable for core
8140 (sse_load0_by_pxor): Likewise.
8141 (four_jump_limit): Disable for core.
8142 (pad_returns): Likewise.
8143 (avoid_vector_decode): Likewise.
8144 (fuse_cmp_and_branch): Enable for cores.
8145 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
8147 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8150 * config/pa/pa.c (pa_option_override): Disable auto increment and
8151 decrement instructions until reload is completed.
8153 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
8154 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
8155 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
8157 2013-09-20 DJ Delorie <dj@redhat.com>
8158 Nick Clifton <nickc@redhat.com>
8160 * config/rl78/rl78.c: Various whitespace and comment tweaks.
8161 (need_to_save): Save bank 0 on interrupts.
8162 (characterize_address): Strip far address wrappers.
8163 (rl78_as_legitimate_address): Likewise.
8164 (transcode_memory_rtx): Likewise.
8165 (rl78_peep_movhi_p): Disable this peephole after devirt.
8166 (rl78_propogate_register_origins): Forget all origins when a
8168 * config/rl78/rl78-virt.md: Various whitespace tweaks.
8169 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
8171 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
8172 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
8173 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
8175 * config/rl78/constraints.md: Various whitespace and paren tweaks.
8177 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
8179 * config/pa/pa.md: In "scc" insn patterns, change output template to
8180 handle const0_rtx in reg_or_0_operand operands.
8182 2013-09-20 Martin Husemann <martin@NetBSD.org>
8185 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
8186 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
8188 2013-09-20 Richard Biener <rguenther@suse.de>
8191 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
8192 remembering SSA name version and block index.
8193 (new_scev_info_str): Adjust.
8194 (hash_scev_info): Likewise. Also hash the block index.
8195 (eq_scev_info): Adjust.
8196 (find_var_scev_info): Likewise.
8197 (struct instantiate_cache_entry): Remove.
8198 (struct instantiate_cache_type): Use a htab to map name, block
8200 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8201 (get_instantiated_value_entry): Likewise.
8202 (hash_idx_scev_info, eq_idx_scev_info): New functions.
8203 (instantiate_scev_name): Adjust.
8205 2013-09-20 Jeff Law <law@redhat.com>
8207 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
8209 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
8211 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
8212 Call aarch64_simd_expand_args to update op[argc].
8214 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8216 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
8219 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
8221 * gengtype.c (file_rules): Added rule for *.cc files.
8222 (get_output_file_with_visibility): Give fatal message when no
8225 2013-09-20 Renlin Li <renlin.li@arm.com>
8227 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
8228 (aarch64_expand_epilogue): Likewise.
8229 (aarch64_legitimize_reload_address): Likewise.
8231 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
8234 * expr.c (expand_assignment): Remove misalignp code path.
8236 2013-09-20 Marek Polacek <polacek@redhat.com>
8239 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
8240 TYPE_PRECISION. Add asserts.
8242 2013-09-20 Richard Biener <rguenther@suse.de>
8244 PR tree-optimization/58453
8245 * tree-loop-distribution.c (distribute_loop): Apply the cost
8246 model for -ftree-loop-distribute-patterns, too.
8248 2013-09-20 Richard Biener <rguenther@suse.de>
8251 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
8252 make type comparison less strict.
8254 2013-09-20 Alan Modra <amodra@gmail.com>
8256 * configure: Regenerate.
8257 * aclocal.m4: Regenerate.
8259 2013-09-20 Marek Polacek <polacek@redhat.com>
8262 * doc/extend.texi: Document that attribute used is meant to be used
8263 on variables with static storage duration.
8265 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8267 PR tree-optimization/58472
8268 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
8269 simd_lane_access set inv_p = false.
8270 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
8271 the simduid magic VAR_DECL.
8273 2013-09-19 Jan Hubicka <jh@suse.cz>
8275 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
8277 2013-09-17 Jeff Law <law@redhat.com>
8279 * tree-ssa-dom.c (record_temporary_equivalences): New function
8280 split out of dom_opt_dom_walker::after_dom_children.
8281 (dom_opt_dom_walker::thread_across_edge): Move common code
8282 in here from dom_opt_dom_walker::after_dom_children.
8283 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
8285 2013-09-19 Jan Hubicka <jh@suse.cz>
8287 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
8288 (TARGET_GENERIC): Use PROCESOR_GENERIC
8289 (enum processor_type): Unify generic32 and 64.
8290 * i386.md (cpu): Likewise.
8291 * x86-tune.def (use_leave): Enable for generic32.
8292 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
8293 * athlon.md: Change generic64 to generic in all occurences.
8294 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
8295 (ix86_target_macros_internal): Likewise.
8296 * driver-i386.c (host_detect_local_cpu): Likewise.
8297 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
8299 (generic_memcpy, generic_memset, generic_cost): This one.
8300 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
8301 (m_GENERIC32, m_GENERIC64): Remove.
8302 (m_GENERIC): Turn into one flag.
8303 (processor_target): Unify generic tunnings.
8304 (ix86_option_override_internal): Replace generic32/64 by generic.
8305 (ix86_issue_rate): Likewise.
8306 (ix86_adjust_cost): Likewise.
8308 2013-09-19 Jan Hubicka <jh@suse.cz>
8310 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
8311 of speculative flag.
8313 2013-09-19 Jakub Jelinek <jakub@redhat.com>
8315 * omp-low.c (expand_omp_sections): Always pass len - 1 to
8316 GOMP_sections_start, even if !exit_reachable.
8318 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8320 * lra-constraints.c (need_for_all_save_p): Use macro
8321 HARD_REGNO_CALL_PART_CLOBBERED.
8322 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
8323 set up pseudo conflict hard regs.
8325 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
8328 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
8331 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
8333 PR rtl-optimization/58438
8334 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
8335 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
8336 from previous subpasses.
8338 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
8340 * arm.c (arm_get_frame_offsets): Validate architecture supports
8341 LDRD/STRD before accepting the tuning preference.
8342 (arm_expand_prologue): Likewise.
8343 (arm_expand_epilogue): Likewise.
8345 2013-09-18 Richard Biener <rguenther@suse.de>
8347 PR tree-optimization/58417
8348 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
8349 have chrecs with symbols defined in the loop as operands.
8350 (chrec_fold_multiply): Likewise.
8351 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
8352 parameters before folding binary operations.
8353 (struct instantiate_cache_entry_hasher): Remove.
8354 (struct instantiate_cache_type): Use a pointer-map.
8355 (instantiate_cache_type::instantiate_cache_type): New function.
8356 (instantiate_cache_type::get): Likewise.
8357 (instantiate_cache_type::set): Likewise.
8358 (instantiate_cache_type::~instantiate_cache_type): Adjust.
8359 (get_instantiated_value_entry): Likewise.
8360 (global_cache): New global.
8361 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
8362 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
8363 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
8364 (instantiate_scev_name): Adjust.
8365 (instantiate_scev): Construct global instead of local cache.
8366 (resolve_mixers): Likewise.
8368 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
8369 Paolo Carlini <paolo.carlini@oracle.com>
8372 * doc/implement-cxx.texi: Fix references to the C++ standards.
8374 2013-09-18 Jakub Jelinek <jakub@redhat.com>
8376 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
8377 * tree-vect-data-refs.c (vect_analyze_data_refs): For
8378 simd_lane_access drs, update also DR_ALIGNED_TO.
8380 2013-09-18 Marek Polacek <polacek@redhat.com>
8383 * doc/extend.texi: Document no_sanitize_undefined attribute.
8384 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
8385 no_sanitize_undefined attribute.
8387 2013-09-18 Nick Clifton <nickc@redhat.com>
8389 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
8390 (ASM_DECLARE_FUNCTION_NAME): Define.
8392 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
8394 * compare-elim.c (find_comparison_dom_walker): New class
8395 (find_comparisons_in_bb): Rename to
8396 find_comparison_dom_walker::before_dom_children
8397 (find_comparisons): Adjust
8398 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
8400 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
8401 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
8402 (init_walk_dominator_tree): Remove declaration.
8403 (fini_walk_dominator_tree): Remove declaration.
8404 * fwprop.c (single_def_use_dom_walker): New class
8405 (single_def_use_enter_block): Convert to
8406 single_def_use_dom_walker::before_dom_children.
8407 (single_def_use_leave_block): Convert to
8408 single_def_use_dom_walker::after_dom_children.
8409 (build_single_def_use_links): Adjust.
8410 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
8412 (find_candidates_in_block): Convert to
8413 find_candidates_dom_walker::before_dom_children.
8414 (execute_strength_reduction): Adjust.
8415 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
8416 (sese_dom_walker): New class.
8417 (sese_dom_walker::sese_dom_walker): New constructor.
8418 (sese_dom_walker::~sese_dom_walker): New destructor.
8419 (build_sese_conditions_before): Convert to
8420 sese_dom_walker::before_dom_children.
8421 (build_sese_conditions_after): Convert to
8422 sese_dom_walker::after_dom_children.
8423 (build_poly_scop): Adjust
8424 * tree-into-ssa.c (rewrite_dom_walker): New class
8425 (rewrite_enter_block): Convert to
8426 rewrite_dom_walker::before_dom_children.
8427 (rewrite_leave_block): Convert to
8428 rewrite_dom_walker::after_dom_children.
8429 (rewrite_update_dom_walker): New class.
8430 (rewrite_update_enter_block): Convert to
8431 rewrite_update_dom_walker::before_dom_children.
8432 (rewrite_update_leave_block): Convert to
8433 rewrite_update_dom_walker::after_dom_children.
8434 (rewrite_blocks, rewrite_into_ssa): Adjust.
8435 (mark_def_dom_walker): New class.
8436 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
8437 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
8438 (mark_def_sites_blocks): Convert to
8439 mark_def_dom_walker::before_dom_children.
8440 (mark_def_site_blocks): Remove.
8441 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
8442 (tree_ssa_dominator_optimize): Adjust.
8443 (dom_thread_across_edge): Convert to method
8444 dom_opt_dom_walker::thread_across_edge.
8445 (dom_opt_enter_block): Convert to member function
8446 dom_opt_dom_walker::before_dom_children.
8447 (dom_opt_leave_block): Convert to member function
8448 dom_opt_dom_walker::after_dom_children.
8449 * tree-ssa-dse.c (dse_dom_walker): New class.
8450 (dse_enter_block): Convert to member function
8451 dse_dom_walker::before_dom_children.
8452 (tree_ssa_dse): Adjust.
8453 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
8454 (determine_invariantness_stmt): Convert to method
8455 invariantness_dom_walker::before_dom_children.
8456 (determine_invariantness): Remove
8457 (move_computations_dom_walker): New class.
8458 (move_computations_stmt): Convert to method
8459 move_computations_dom_walker::before_dom_children.
8460 (move_computations, tree_ssa_lim): Adjust.
8461 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
8462 (nt_init_block): Convert to method
8463 notrappping_dom_walker::before_dom_children.
8464 (nt_fini_block): Convert to method
8465 method nontrapping_dom_walker::after_dom_children.
8466 (get_non_trapping): Adjust.
8467 * tree-ssa-pre.c (eliminate_dom_walker): New class.
8468 (eliminate_bb): Convert to method
8469 eliminate_dom_walker::before_dom_children.
8470 (eliminate_leave_block): Convert to method
8471 eliminate_dom_walker::after_dom_children.
8472 (eliminate): Adjust.
8473 * tree-ssa-strlen.c (strlen_dom_walker): New class.
8474 (strlen_enter_block): Convert to method
8475 strlen_dom_walker::before_dom_children.
8476 (strlen_leave_block): Convert to method
8477 method strlen_dom_walker::after_dom_children.
8478 (tree_ssa_strlen): Adjust.
8479 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
8480 (tree_ssa_uncprop): Adjust.
8481 (uncprop_leave_block): Convert to method
8482 uncprop_dom_walker::after_dom_children.
8483 (uncprop_leave_block): Convert to method
8484 uncprop_dom_walker::before_dom_children.
8486 2013-09-18 Bin Cheng <bin.cheng@arm.com>
8488 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
8489 branch in same basic block. Check both src and dest of the move insn.
8491 2013-09-17 Nick Clifton <nickc@redhat.com>
8493 * config/rl78/rl78-real.md (bf): New pattern.
8495 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
8496 (rl78_print_operand): Do not put a # before a %B.
8497 * config/rl78/rl78.opt: Tweak doc strings.
8499 2013-09-17 DJ Delorie <dj@redhat.com>
8501 * config/rl78/constraints.md (Wcv): Allow up to $r31.
8502 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
8503 (rl78_option_override): Likewise, if -mallregs.
8504 (is_virtual_register): Likewise.
8505 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
8506 (REGNO_OK_FOR_BASE_P): Likewise.
8507 * config/rl78/rl78.opt (-mallregs): New.
8509 2013-09-17 Nick Clifton <nickc@redhat.com>
8511 * config/rl78/rl78.c (need_to_save): Change return type to bool.
8512 For interrupt functions: save all call clobbered registers if the
8513 interrupt handler is not a leaf function.
8514 (rl78_expand_prologue): Always recompute the frame information.
8515 For interrupt functions: only select bank 0 if one of the bank 0
8516 registers is going to be psuhed.
8518 2013-09-17 DJ Delorie <dj@redhat.com>
8520 * config/rl78/constraints.md: For each W* constraint, rename to C*
8521 and create a W* constraint that checks for an optional ES: prefix
8523 * config/rl78/rl78.md (UNS_ES_ADDR): New.
8524 (es_addr): New. Used to wrap far addresses.
8525 * config/rl78/rl78-protos.h (rl78_es_addr): New.
8526 (rl78_es_base): New.
8527 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
8528 wrapped far addresses.
8529 (rl78_print_operand_1): Unwrap far addresses before processing.
8530 (rl78_lo16): Wrap far addresses in unspecs.
8531 (rl78_es_addr): New.
8532 (rl78_es_base): New.
8533 (insn_ok_now): Check for not-yet-wrapped far addresses.
8534 (transcode_memory_rtx): Properly re-wrap far addresses.
8536 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
8538 * config/sparc/t-rtems: Add leon3 multilibs.
8540 2013-09-17 Cong Hou <congh@google.com>
8542 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
8543 when checking the dot production pattern. The type of rhs operand
8544 of multiply is now checked correctly.
8546 2013-09-17 Jeff Law <law@redhat.com>
8548 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8549 edge implied equivalences into successor phis.
8550 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
8551 here from tree-ssa-threadedge.c.
8552 (mark_threaded_blocks): When threading through a joiner, if both
8553 successors of the joiner's clone reach the same block, verify the
8554 PHI arguments are equal. If not, cancel the jump threading request.
8555 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
8556 tree-ssa-threadupdate.c
8557 (thread_across_edge): Don't check PHI argument equality when
8558 threading through joiner block here.
8560 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
8562 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
8563 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
8564 (warn_uninit, warn_uninitialized_vars,
8565 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8566 Move to tree-ssa-uninit.c.
8567 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
8568 (has_undefined_value_p): New. Pass dependant parts of
8569 ssa_undefined_value_p.
8570 (uninit_undefined_value_p): Use has_undefined_value_p.
8571 (warn_uninit, warn_uninitialized_vars,
8572 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
8573 Move from tree-ssa.c.
8574 * tree-ssa.h: Adjust prototypes.
8576 2013-09-17 Jan Hubicka <jh@suse.cz>
8579 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
8580 * ipa-inline.c (can_inline_edge_p): Do not downgrade
8581 FUNCTION_NOT_OPTIMIZED.
8582 * ipa-inline-analysis.c (compute_inline_parameters): Function
8583 not optimized is not inlinable unless it is alwaysinline.
8584 (inline_analyze_function): Force calls in not optimized
8585 function not inlinable.
8587 2013-09-17 Jan Hubicka <jh@suse.cz>
8590 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
8592 * ipa.c (function_and_variable_visibility): Likewise.
8593 * ipa-profile.c (ipa_profile): Likewise.
8595 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
8598 * cgraph.c (cgraph_function_body_availability): Check for ifunc
8599 attribute, and don't inline the resolver in this case.
8601 2013-09-17 Teresa Johnson <tejohnson@google.com>
8603 * coverage.c (get_coverage_counts): Add missing newline.
8605 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8607 PR tree-optimization/58088
8608 * fold-const.c (mask_with_trailing_zeros): New function.
8609 (fold_binary_loc): Make sure we don't recurse infinitely
8610 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
8611 Use mask_with_trailing_zeros where appropriate.
8613 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
8615 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
8616 of 'prev' var to get better distance estimation.
8618 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
8620 * tree-inline.h (struct copy_body_data): Add transform_parameter.
8621 * tree-inline.c (is_parameter_of): New predicate.
8622 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
8623 a parameter has been remapped.
8624 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
8625 (optimize_inline_calls): Initialize transform_parameter.
8626 (copy_gimple_seq_and_replace_locals): Likewise.
8627 (tree_function_versioning): Likewise.
8628 (maybe_inline_call_in_expr): Likewise.
8630 2013-09-17 Nick Clifton <nickc@redhat.com>
8632 * config/msp430/msp430-protos.h: Add prototypes for new functions.
8633 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
8635 (is_attr_func): New function.
8636 (msp430_is_interrupt_func): New function.
8637 (is_naked_func): New function.
8638 (is_reentrant_func): New function.
8639 (is_critical_func): New function.
8640 (msp430_start_function): Add annotations for function attributes.
8641 (msp430_attr): New function.
8642 (msp430_attribute_table): New.
8643 (msp430_function_section): New function.
8644 (TARGET_ASM_FUNCTION_SECTION): Define.
8645 (msp430_builtin): New enum.
8646 (msp430_init_builtins): New function.
8647 (msp430_builtin_devl): New function.
8648 (msp430_expand_builtin): New function.
8649 (TARGET_INIT_BUILTINS): Define.
8650 (TARGET_EXPAND_BUILTINS): Define.
8651 (TARGET_BUILTIN_DECL): Define.
8652 (msp430_expand_prologue): Add support for naked, interrupt,
8653 critical and reentrant functions.
8654 (msp430_expand_epilogue): Likewise.
8655 (msp430_print_operand): Handle 'O' character.
8656 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
8658 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
8659 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
8660 (pushm): Use a 'n' rather than an 'i' constraint.
8661 (msp_return): Add generation of the interrupt return instruction.
8662 (disable_interrupts): New pattern.
8663 (enable_interrupts): New pattern.
8664 (push_intr_state): New pattern.
8665 (pop_intr_state): New pattern.
8666 (bic_SR): New pattern.
8667 (bis_SR): New pattern.
8668 * doc/extend.texi: Document MSP430 function attributes and builtin
8671 2013-09-17 Richard Biener <rguenther@suse.de>
8673 PR tree-optimization/58432
8674 * tree-loop-distribution.c (tree_loop_distribution): Also
8675 scan PHIs for outside loop uses and seed a partition from them.
8677 2013-09-17 Bin Cheng <bin.cheng@arm.com>
8679 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
8680 (restructure_reference): Call backtrace_base_for_ref.
8682 2013-09-17 Alan Modra <amodra@gmail.com>
8685 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
8688 2013-09-16 DJ Delorie <dj@redhat.com>
8690 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
8691 vregs location for RL78/G10.
8692 (rl78_expand_prologue): Avoid SEL on G10.
8693 (rl78_expand_epilogue): Likewise.
8694 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
8695 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
8696 __RL78_G10__ when appropriate.
8697 (ASM_SPEC): Pass -mg10 along to the assembler.
8698 * config/rl78/rl78.md (sel_rb): Disable for G10.
8699 * config/rl78/rl78.opt: Add -mg10 option.
8700 * config/rl78/t-rl78: Add -mg10 multilib.
8702 2013-09-16 Xinliang David Li <davidxl@google.com>
8704 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
8705 * omp-low.c (omp_max_vf): Ditto.
8706 (expand_omp_simd): Ditto.
8707 * tree-vectorizer.c (vectorize_loops): Ditto.
8708 (gate_vect_slp): Ditto.
8709 (gate_increase_alignment): Ditto.
8710 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
8711 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
8712 (gate_tree_vectorize): Name change.
8713 (tree_vectorize): Ditto.
8714 (pass_vectorize::gate): Call new function.
8715 (pass_vectorize::execute): Ditto.
8716 * opts.c: O3 default setting change.
8717 (finish_options): Check new flag.
8718 * doc/invoke.texi: Document new flags.
8719 * common.opt: New flags.
8721 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
8723 * doc/tm.texi.in (Cond Exec Macros): Remove node.
8724 (Condition Code): Don't reference it.
8725 * doc/tm.texi: Regenerate.
8727 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
8730 * lra-constraints.c (undo_optional_reloads): Consider all optional
8731 reload even if it did not get a hard reg.
8733 2013-09-16 Teresa Johnson <tejohnson@google.com>
8735 * dumpfile.c (dump_loc): Remove newline emission.
8736 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
8737 emission to dump_printf_loc calls where missing.
8738 (vect_mark_for_runtime_alias_test): Ditto.
8739 (vect_analyze_data_ref_dependence): Ditto.
8740 (vect_analyze_data_ref_dependences): Ditto.
8741 (vect_slp_analyze_data_ref_dependence): Ditto.
8742 (vect_slp_analyze_data_ref_dependences): Ditto.
8743 (vect_compute_data_ref_alignment): Ditto.
8744 (vect_update_misalignment_for_peel): Ditto.
8745 (vect_verify_datarefs_alignment): Ditto.
8746 (vector_alignment_reachable_p): Ditto.
8747 (vect_get_data_access_cost): Ditto.
8748 (vect_enhance_data_refs_alignment): Ditto.
8749 (vect_find_same_alignment_drs): Ditto.
8750 (vect_analyze_data_refs_alignment): Ditto.
8751 (vect_analyze_group_access): Ditto.
8752 (vect_analyze_data_ref_access): Ditto.
8753 (vect_analyze_data_ref_accesses): Ditto.
8754 (vect_prune_runtime_alias_test_list): Ditto.
8755 (vect_analyze_data_refs): Ditto.
8756 (vect_create_addr_base_for_vector_ref): Ditto.
8757 (vect_create_data_ref_ptr): Ditto.
8758 (vect_grouped_store_supported): Ditto.
8759 (vect_grouped_load_supported): Ditto.
8760 * value-prof.c (check_counter): Ditto.
8761 (check_ic_target): Ditto.
8762 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
8763 (vect_recog_widen_mult_pattern): Ditto.
8764 (vect_recog_widen_sum_pattern): Ditto.
8765 (vect_recog_over_widening_pattern): Ditto.
8766 (vect_recog_widen_shift_pattern): Ditto.
8767 (vect_recog_rotate_pattern): Ditto.
8768 (vect_recog_vector_vector_shift_pattern): Ditto.
8769 (vect_recog_divmod_pattern): Ditto.
8770 (vect_recog_mixed_size_cond_pattern): Ditto.
8771 (vect_recog_bool_pattern): Ditto.
8772 (vect_pattern_recog_1): Ditto.
8773 (vect_pattern_recog): Ditto.
8774 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
8775 (vect_is_simple_iv_evolution): Ditto.
8776 (vect_analyze_scalar_cycles_1): Ditto.
8777 (vect_get_loop_niters): Ditto.
8778 (vect_analyze_loop_1): Ditto.
8779 (vect_analyze_loop_form): Ditto.
8780 (vect_analyze_loop_operations): Ditto.
8781 (vect_analyze_loop_2): Ditto.
8782 (vect_analyze_loop): Ditto.
8783 (report_vect_op): Ditto.
8784 (vect_is_slp_reduction): Ditto.
8785 (vect_is_simple_reduction_1): Ditto.
8786 (vect_get_known_peeling_cost): Ditto.
8787 (vect_estimate_min_profitable_iters): Ditto.
8788 (vect_model_reduction_cost): Ditto.
8789 (vect_model_induction_cost): Ditto.
8790 (get_initial_def_for_induction): Ditto.
8791 (vect_create_epilog_for_reduction): Ditto.
8792 (vectorizable_reduction): Ditto.
8793 (vectorizable_induction): Ditto.
8794 (vectorizable_live_operation): Ditto.
8795 (vect_loop_kill_debug_uses): Ditto.
8796 (vect_transform_loop): Ditto.
8797 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
8798 (vect_stmt_relevant_p): Ditto.
8799 (process_use): Ditto.
8800 (vect_mark_stmts_to_be_vectorized): Ditto.
8801 (vect_model_simple_cost): Ditto.
8802 (vect_model_promotion_demotion_cost): Ditto.
8803 (vect_model_store_cost): Ditto.
8804 (vect_get_store_cost): Ditto.
8805 (vect_model_load_cost): Ditto.
8806 (vect_get_load_cost): Ditto.
8807 (vect_init_vector_1): Ditto.
8808 (vect_get_vec_def_for_operand): Ditto.
8809 (vect_finish_stmt_generation): Ditto.
8810 (vectorizable_call): Ditto.
8811 (vectorizable_conversion): Ditto.
8812 (vectorizable_assignment): Ditto.
8813 (vectorizable_shift): Ditto.
8814 (vectorizable_operation): Ditto.
8815 (vectorizable_store): Ditto.
8816 (vectorizable_load): Ditto.
8817 (vectorizable_condition): Ditto.
8818 (vect_analyze_stmt): Ditto.
8819 (vect_transform_stmt): Ditto.
8820 (vect_is_simple_use): Ditto.
8821 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
8822 (vect_can_advance_ivs_p): Ditto.
8823 (vect_update_ivs_after_vectorizer): Ditto.
8824 (vect_do_peeling_for_loop_bound): Ditto.
8825 (vect_gen_niters_for_prolog_loop): Ditto.
8826 (vect_update_inits_of_drs): Ditto.
8827 (vect_create_cond_for_alias_checks): Ditto.
8828 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
8829 (vect_build_slp_tree_1): Ditto.
8830 (vect_supported_load_permutation_p): Ditto.
8831 (vect_analyze_slp_instance): Ditto.
8832 (vect_analyze_slp): Ditto.
8833 (vect_make_slp_decision): Ditto.
8834 (vect_detect_hybrid_slp): Ditto.
8835 (vect_bb_vectorization_profitable_p): Ditto.
8836 (vect_slp_analyze_bb_1): Ditto.
8837 (vect_update_slp_costs_according_to_vf): Ditto.
8838 (vect_get_mask_element): Ditto.
8839 (vect_transform_slp_perm_load): Ditto.
8840 (vect_schedule_slp_instance): Ditto.
8841 (vect_schedule_slp): Ditto.
8842 (vect_slp_transform_bb): Ditto.
8843 * profile.c (read_profile_edge_counts): Ditto.
8844 (compute_branch_probabilities): Ditto.
8845 * coverage.c (get_coverage_counts): Ditto.
8847 2013-09-16 Diego Novillo <dnovillo@google.com>
8849 * tree-core.h: Add missing comment lines from refactoring of tree.h.
8851 2013-09-16 Jan Hubicka <jh@suse.cz>
8853 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
8854 abstract functions; for static functions check the presence of body.
8856 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8858 * config/aarch64/aarch64-simd-builtins.def (fma): New.
8859 * config/aarch64/aarch64-simd.md
8860 (aarch64_mla_elt<mode>): New.
8861 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
8862 (aarch64_mls_elt<mode>): Likewise.
8863 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
8864 (aarch64_fma4_elt<mode>): Likewise.
8865 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
8866 (aarch64_fma4_elt_to_128v2df): Likewise.
8867 (aarch64_fma4_elt_to_64df): Likewise.
8868 (fnma<mode>4): Likewise.
8869 (aarch64_fnma4_elt<mode>): Likewise.
8870 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
8871 (aarch64_fnma4_elt_to_128v2df): Likewise.
8872 (aarch64_fnma4_elt_to_64df): Likewise.
8873 * config/aarch64/iterators.md (VDQSF): New.
8874 * config/aarch64/arm_neon.h
8875 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
8876 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
8878 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8880 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
8881 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
8882 (aarch64_mul3_elt_to_128df): Likewise.
8883 (aarch64_mul3_elt_to_64v2df): Likewise.
8884 * config/aarch64/iterators.md (VEL): Also handle DFmode.
8886 (VMUL_CHANGE_NLANES) Likewise.
8889 * config/aarch64/arm_neon.h
8890 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
8892 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
8894 * config/aarch64/arm_neon.h
8895 (vcvtx_high_f32_f64): Fix parameters.
8897 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8898 Uros Bizjak <ubizjak@gmail.com>
8900 * config/alpha.c: Include tree-ssanames.h.
8902 2013-09-16 Richard Biener <rguenther@suse.de>
8904 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
8905 (dot_rdg_1): Handle control_dd.
8906 (create_edge_for_control_dependence): New function.
8907 (create_rdg_edges): Add control dependences if asked for.
8908 (build_rdg): Likewise.
8909 (generate_loops_for_partition): If there are not necessary
8910 control stmts remove all their dependencies.
8911 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
8912 (distribute_loop): Pass on control dependences.
8913 (tree_loop_distribution): Compute control dependences and remove
8914 restriction on number of loop nodes.
8916 2013-09-16 Jakub Jelinek <jakub@redhat.com>
8918 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
8921 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
8923 * cse.c (try_const_anchors): Punt on CC modes.
8925 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8927 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
8929 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
8932 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
8935 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8938 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
8940 (darwin_objc1_section): Likewise.
8941 (darwin_file_end): Emit Image Info section when required.
8943 2013-09-14 Jan Hubicka <jh@suse.cz>
8945 * tree-into-ssa.c (gate_into_ssa): New.
8946 (pass_data_build_ssa): Use it.
8947 * cgraph.h (expand_thunk): Update prototype.
8948 * cgraphunit.c (analyze_function): Expand thunks early.
8949 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
8950 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
8951 set lowered flag; do not add new function.
8952 (assemble_thunks_and_aliases): Update.
8953 * tree-ssa.c (gate_init_datastructures): New gate.
8954 (pass_data_init_datastructures): Use it.
8956 2013-09-14 Iain Sandoe <iain@codesourcery.com>
8959 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
8960 xmm register set described in the psABI.
8962 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
8964 * dwarf2out.c (should_emit_struct_debug): Add check
8965 for type_decl variable is not NULL.
8967 2013-09-13 Jacek Caban <jacek@codeweavers.com>
8969 * config.gcc: Use new winnt-c.c target hooks
8970 * config/t-winnt: New file
8971 * config/winnt-c.c: New file
8972 * doc/tm.texi.in: Document new hook
8973 * doc/tm.texi: Regenerated
8975 2013-09-13 Jan Hubicka <jh@suse.cz>
8978 * ipa-inline.c (check_callers): New function.
8979 (check_caller_edge): Remove.
8980 (want_inline_function_to_all_callers_p): Also permit alises that are
8982 (inline_to_all_callers): Terminate the walk when devirtualization
8983 introduce new calls.
8985 2013-09-13 Jan Hubicka <jh@suse.cz>
8987 * ipa-inline-analysis.c (struct growth_data): Add node.
8988 (do_estimate_growth_1): Fix detection of recursion.
8990 2013-09-13 Jakub Jelinek <jakub@redhat.com>
8992 PR tree-optimization/58392
8993 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
8994 to avoid shadowing of outer loop variable. If
8995 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
8996 replace_by_duplicate_decl simduid of loops that have it set and
8997 set dest_cfun->has_simduid_loops and/or
8998 dest_cfun->has_force_vect_loops.
8999 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
9000 instead of maybe_lookup_decl.
9001 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
9002 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
9003 call. Copy over force_vect and copy and remap simduid. Set
9004 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
9005 (copy_cfg_body): Remove blocks_to_copy argument. Use
9006 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
9007 caller. Don't set cfun->has_simduid_loops and/or
9008 cfun->has_force_vect_loops here.
9009 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
9011 (expand_call_inline, tree_function_versioning): Adjust copy_body
9014 2013-09-13 Martin Jambor <mjambor@suse.cz>
9017 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
9018 the assert if the edge was a speculative one.
9020 2013-09-13 Richard Biener <rguenther@suse.de>
9022 * tree-data-ref.h (known_dependences_p): Move here ...
9023 * tree-loop-distribution.c (known_dependences_p): ... from here.
9024 (dump_rdg_component, debug_rdg_component): Remove.
9026 (generate_loops_for_partition): Use gimple_uid instead of
9027 relying on matching stmt visit order.
9028 (rdg_build_partitions): Take starting stmt vector.
9029 (ldist_gen): Merge into ...
9030 (distribute_loop): ... this function. Do not compute starting
9032 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
9034 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9036 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
9037 Set type attribute correctly. Set predicable_short_it attribute.
9038 (cmpsi_shiftsi): Remove %? from output template.
9040 2013-09-13 Richard Biener <rguenther@suse.de>
9042 * tree-loop-distribution.c (struct rdg_component,
9043 rdg_defs_used_in_other_loops_p, free_rdg_components,
9044 rdg_build_components): Remove.
9045 (stmts_from_loop): Do not record virtual PHIs.
9046 (generate_loops_for_partition): Skip virtual PHIs.
9047 (build_rdg_partition_for_component): Rename to ...
9048 (build_rdg_partition_for_vertex): ... this and adjust.
9049 (rdg_build_partitions): Take a vector of starting vertices
9050 instead of components. Remove unnecessary leftover handling.
9051 (ldist_gen): Do not build components or record other stores.
9052 (distribute_loop): Do not distribute loops containing stmts
9055 2013-09-13 Christian Bruel <christian.bruel@st.com>
9058 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
9060 2013-09-13 Kai Tietz <ktietz@redhat.com>
9062 * config.gcc: Separate cases for mingw and cygwin targets,
9063 and add 64-bit cygwin target case.
9065 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
9066 dll-export inline-functions.
9067 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
9069 2013-09-13 Jeff Law <law@redhat.com>
9073 2013-09-06 Jeff Law <law@redhat.com>
9075 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9076 edge implied equivalences into successor phis.
9078 2013-09-12 DJ Delorie <dj@redhat.com>
9080 * config/rl78/rl78-virt.md: Change from | to \; for asm line
9083 2013-09-12 Brooks Moses <bmoses@google.com>
9086 * Makefile.in: Do not install driver binaries in $(target)/bin.
9088 2013-09-12 DJ Delorie <dj@redhat.com>
9090 * config/rl78/rl78.opt (mrelax): New.
9091 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
9092 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
9094 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
9096 (rl78_expand_epilogue): Likewise.
9098 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
9101 * lra-eliminations.c (remove_reg_equal_offset_note): New.
9102 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
9103 pointer elimination with using remove_reg_equal_offset_note.
9105 2013-09-12 DJ Delorie <dj@redhat.com>
9107 * config/msp430/: New port.
9108 * config.gcc (msp430): Added.
9109 * doc/invoke.texi: Document MSP430 options.
9110 * doc/install.texi: Document msp430-elf
9111 * doc/md.texi: Document msp430-elf
9112 * doc/contrib.texi: Document msp430-elf
9114 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
9117 2013-09-12 Martin Jambor <mjambor@suse.cz>
9120 * ipa-prop.c (remove_described_reference): Give up if the edge in the
9121 reference descriptor is NULL.
9122 (ipa_edge_removal_hook): If owning a reference descriptor, set its
9125 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9127 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
9128 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
9129 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
9130 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
9131 make_temp_ssa_name): move to tree-ssanames.h
9132 * tree-ssa-alias.h: Move prototype.
9133 * tree-ssa.h: Include tree-ssanames.h.
9134 * tree-ssanames.c (FREE_SSANAMES): Move to here.
9135 * tree-ssanames.h: New. Move items from tree-flow*.h
9136 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
9138 2013-09-12 Richard Biener <rguenther@suse.de>
9140 PR tree-optimization/58404
9141 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
9142 propagate non-invariant addresses into dereferences wrapped
9143 in component references.
9145 2013-09-12 Richard Biener <rguenther@suse.de>
9147 PR tree-optimization/58402
9148 * passes.def: Move pass_late_warn_uninitialized later.
9150 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
9152 * tree-ssa.h: New. Move content from tree-flow.h and
9154 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
9155 Move prototypes belonging to tree-ssa.c.
9156 * tree-flow-inline.h (redirect_edge_var_map_def,
9157 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
9159 * gimple.h: Adjust prototypes.
9160 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
9162 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
9163 * tree.h: Move prototype to tree-ssa.h.
9164 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
9165 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
9166 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
9167 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
9168 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
9169 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
9170 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
9171 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
9172 graphite-blocking.c, graphite-clast-to-gimple.c,
9173 graphite-dependences.c, graphite-interchange.c,
9174 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
9175 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
9176 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
9177 ipa-reference.c, ipa-split.c, ipa-utils.c,
9178 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
9179 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
9180 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
9181 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
9182 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
9183 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
9184 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
9185 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
9186 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
9187 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
9188 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
9189 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
9190 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
9191 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
9192 value-prof.c, var-tracking.c,
9193 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
9195 2013-09-12 Richard Biener <rguenther@suse.de>
9197 PR tree-optimization/58396
9198 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
9199 (build_rdg): Take a loop-nest parameter, fix memleaks.
9200 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
9202 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
9204 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
9207 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
9209 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
9210 The FMA instruction names should have a 'v' prefix.
9212 2013-09-12 Richard Biener <rguenther@suse.de>
9214 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
9215 (dot_rdg): Use popen instead of system in optional code.
9216 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
9217 (already_processed_vertex_p): Adjust.
9218 (has_anti_or_output_dependence, predecessor_has_mem_write,
9219 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
9220 rdg_flag_uses): Remove.
9221 (rdg_flag_vertex): Simplify.
9222 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
9224 (build_rdg_partition_for_component): Process the first vertex
9225 of a component only.
9226 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
9228 2013-09-12 Alan Modra <amodra@gmail.com>
9230 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
9232 2013-09-11 DJ Delorie <dj@redhat.com>
9233 Nick Clifton <nickc@redhat.com>
9235 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
9236 (rl78_stack_based_mem): New.
9237 * config/rl78/constraints.md (Iv08): New.
9246 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
9248 (movsi): Change from expand to insn-and-split.
9249 (ashrsi3): Clobber AX.
9253 * config/rl78/rl78.md (CC_REG): Fix.
9254 (addsi3): Allow memory and immediate operands.
9255 (addsi3_internal): Split into...
9256 (addsi3_internal_virt): ...new, and ...
9257 (addsi3_internal_real): ...new.
9259 (subsi3_internal_virt): New.
9260 (subsi3_internal_real): New.
9261 (mulsi3): Add memory operand.
9262 (mulsi3_rl78): Likewise.
9263 (mulsi3_g13): Likewise.
9264 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
9265 (cbranchqi4_real): Add more constraint options.
9266 (cbranchhi4_real): Expand pattern.
9267 (cbranchhi4_real_signed): New.
9268 (cbranchhi4_real_inverted): New.
9269 (cbranchsi4_real_lt): New.
9270 (cbranchsi4_real_ge): New.
9271 (cbranchsi4_real_signed): New.
9272 (cbranchsi4_real): New.
9274 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
9276 (lshrsi3_virt): Likewise.
9277 (ashlsi3_virt): Likewise.
9278 (cbranchqi4_virt_signed): New.
9279 (cbranchhi4_virt_signed): New.
9280 (cbranchsi4_virt): New.
9281 * config/rl78/rl78.c: Whitespace fixes throughout.
9282 (move_elim_pass): New.
9283 (pass_data_rl78_move_elim): New.
9284 (pass_rl78_move_elim): New.
9285 (make_pass_rl78_move_elim): New.
9286 (rl78_devirt_info): Run devirt earlier.
9287 (rl78_move_elim_info): New.
9288 (rl78_asm_file_start): Register it.
9289 (rl78_split_movsi): New.
9290 (rl78_as_legitimate_address): Allow virtual base registers when
9292 (rl78_addr_space_convert): Remove spurious debug stuff.
9293 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
9294 (rl78_print_operand): More cases for not printing '#'.
9295 (rl78_expand_compare): Remove most of the logic.
9296 (content_memory): New.
9297 (clear_content_memory): New.
9298 (get_content_index): New.
9299 (get_content_name): New.
9300 (display_content_memory): New.
9301 (update_content): New.
9302 (record_content): New.
9303 (already_contains): New.
9304 (insn_ok_now): Re-recog insns with virtual registers.
9305 (add_postponed_content_update): New.
9306 (process_postponed_content_update): New.
9307 (gen_and_emit_move): New.
9308 (transcode_memory_rtx): Record new location content.
9309 Use gen_and_emit_move.
9310 (force_into_acc): New.
9311 (move_to_acc): Use gen_and_emit_move.
9312 (move_from_acc): Likewise.
9313 (move_acc_to_reg): Likewise.
9314 (move_to_x): Likewise.
9315 (move_to_hl): Likewise.
9316 (move_to_de): Likewise.
9317 (rl78_alloc_physical_registers_op1): Record location content.
9318 (has_constraint): New.
9319 (rl78_alloc_physical_registers_op2): Record location content.
9321 (rl78_alloc_physical_registers_ro1): Likewise.
9322 (rl78_alloc_physical_registers_cmp): Likewise.
9323 (rl78_alloc_physical_registers_umul): Likewise.
9324 (rl78_alloc_address_registers_macax): New.
9325 (rl78_alloc_physical_registers): Initialize and set location
9326 content memory as needed.
9327 (rl78_reorg): Make sure split2 is called.
9328 (rl78_rtx_costs): New.
9330 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
9332 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
9333 for (not (neg ...)) and (neg (not ...)) cases.
9335 2013-09-11 Richard Biener <rguenther@suse.de>
9338 * passes.def: Split critical edges before late uninit warning passes.
9339 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
9341 2013-09-11 Jakub Jelinek <jakub@redhat.com>
9343 PR tree-optimization/58385
9344 * fold-const.c (build_range_check): If both low and high are NULL,
9345 use omit_one_operand_loc to preserve exp side-effects.
9347 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9349 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
9351 2013-09-11 Richard Biener <rguenther@suse.de>
9353 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
9354 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
9355 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
9356 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
9357 stmts_from_loop, known_dependences_p, build_empty_rdg,
9358 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
9359 * tree-loop-distribution.c: ... here.
9360 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
9361 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
9362 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
9363 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
9364 * tree-loop-distribution.c: ... here.
9365 * tree-loop-distribution.c: Include gimple-pretty-print.h.
9366 (struct partition_s): Add loops member.
9367 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
9368 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
9369 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
9371 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9372 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
9373 Sergey Lega <sergey.s.lega@intel.com>
9374 Anna Tikhonova <anna.tikhonova@intel.com>
9375 Ilya Tocar <ilya.tocar@intel.com>
9376 Andrey Turetskiy <andrey.turetskiy@intel.com>
9377 Ilya Verbin <ilya.verbin@intel.com>
9378 Kirill Yukhin <kirill.yukhin@intel.com>
9379 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
9381 * config/i386/constraints.md (k): New.
9383 * config/i386/i386.c (const regclass_map): Add new mask registers.
9384 (dbx_register_map): Ditto.
9385 (dbx64_register_map): Ditto.
9386 (svr4_dbx_register_map): Ditto.
9387 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
9389 (ix86_preferred_reload_class): Disable constants for mask registers.
9390 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
9391 (ix86_hard_regno_mode_ok): Support new mask registers.
9392 (x86_order_regs_for_local_alloc): Ditto.
9393 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
9394 (FIXED_REGISTERS): Add new mask registers.
9395 (CALL_USED_REGISTERS): Ditto.
9396 (REG_ALLOC_ORDER): Ditto.
9397 (VALID_MASK_REG_MODE): New.
9398 (FIRST_MASK_REG): Ditto.
9399 (LAST_MASK_REG): Ditto.
9400 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
9401 (MAYBE_MASK_CLASS_P): New.
9402 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
9403 (REG_CLASS_CONTENTS): Ditto.
9404 (MASK_REGNO_P): New.
9405 (ANY_MASK_REG_P): Ditto.
9406 (HI_REGISTER_NAMES): Add new mask registers.
9407 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
9408 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
9410 (attribute "type"): Add mskmov, msklog.
9411 (attribute "length_immediate"): Support them.
9412 (attribute "memory"): Ditto.
9413 (attribute "prefix_0f"): Ditto.
9414 (*movhi_internal): Support new mask registers.
9415 (*movqi_internal): Ditto.
9416 (define_split): Split out clobber pattern is a logic
9417 insn on mask registers.
9418 (*k<logic><mode>): New.
9419 (*andhi_1): Extend to support mask regs.
9420 (*andqi_1): Extend to support mask regs.
9422 (define_split): Split and-not to and and not if operands
9424 (*<code><mode>_1): Separate HI mode to new pattern...
9425 (*<code>hi_1): This.
9426 (*<code>qi_1): Extend to support mask regs.
9428 (kortestzhi): Ditto.
9429 (kortestchi): Ditto.
9431 (*one_cmpl<mode>2_1): Remove HImode and handle it...
9432 (*one_cmplhi2_1): ...Here, now with mask registers support.
9433 (*one_cmplqi2_1): Support new mask registers.
9434 (HI/QImode arithmetics splitter): Don't split if mask registers
9436 (HI/QImode not splitter): Ditto.
9437 * config/i386/predicated.md (mask_reg_operand): New.
9438 (general_reg_operand): Ditto.
9440 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
9442 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
9443 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
9445 2013-09-10 Jeff Law <law@redhat.com>
9447 PR tree-optimization/58380
9448 * tree-ssa-threadupdate.c (thread_block): Recognize another case
9449 of threading through a buried loop header.
9451 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
9452 return value for single successor case.
9454 2013-09-10 Jan Hubicka <jh@suse.cz>
9456 * ipa-devirt.c (ipa_devirt): Enable with LTO.
9458 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
9461 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
9462 support conditional execution.
9463 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
9465 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
9467 * lra.c (lra): Clear lra_optional_reload_pseudos before every
9469 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
9470 Check destination too to check move insn.
9471 (undo_optional_reloads): Add check that the original peudo did not
9472 changed its allocation and the optional reload was inherited on last
9473 inheritance pass. Break loop after deciding to keep optional reload.
9474 (lra_undo_inheritance): Add check that inherited pseudo still in
9477 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
9479 * config/aarch64/aarch64.md (generic_sched): New.
9480 * config/aarch64/aarch64-generic.md (load): Make conditional
9481 on generic_sched attribute.
9482 (nonload): Likewise.
9484 2013-09-10 Jan Hubicka <jh@suse.cz>
9486 * lto-cgraph.c: Include ipa-utils.h.
9487 (compute_ltrans_boundary): Also add possible targets into the boundary.
9489 2013-09-10 Jan Hubicka <jh@suse.cz>
9491 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
9492 VAR_DECL of vtable rather than full expression.
9494 2013-09-10 Jan Hubicka <jh@suse.cz>
9495 Paolo Carlini <paolo.carlini@oracle.com>
9497 * cgraphunit.c (analyze_functions): Save input_location, set it
9498 to UNKNOWN_LOCATION and restore it at the end.
9500 2013-09-10 Martin Jambor <mjambor@suse.cz>
9502 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
9503 represented by a thunk.
9505 2013-09-10 Jeff Law <law@redhat.com>
9507 PR tree-optimization/58343
9508 * tree-ssa-threadupdate.c (thread_block): Identify and disable
9509 jump threading requests through loop headers buried in the middle
9510 of a jump threading path.
9512 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
9513 in return value/type.
9515 2013-09-10 Jakub Jelinek <jakub@redhat.com>
9517 PR rtl-optimization/58365
9518 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
9519 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
9522 2013-09-10 Richard Biener <rguenther@suse.de>
9524 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
9525 * tree-data-ref.c (create_rdg_vertices): Collect all data
9526 references, signal failure to the caller, use data-ref API.
9527 (build_rdg): Compute data references only once. Maintain lifetime
9528 of data references and data dependences from within RDG.
9529 (free_rdg): Free dependence relations.
9530 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
9531 inventing extra dependences.
9532 (distribute_loop): Update for RDG API changes.
9534 2013-09-10 Kai Tietz <ktietz@redhat.com>
9536 * doc/invoke.texi (fms-extensions): Document changed
9537 behavior for ms-abi targets.
9538 * config/i386/i386.c (ix86_option_override_internal):
9539 Set default value of option -fms-extension for ms-abi targets.
9541 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
9543 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
9545 2013-09-10 Alan Modra <amodra@gmail.com>
9548 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
9550 2013-09-10 Alan Modra <amodra@gmail.com>
9552 * config/rs6000/predicates.md (add_cint_operand): New.
9553 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
9554 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
9555 using add_cint_operand.
9556 (largetoc_high_plus_aix): Likewise.
9558 2013-09-09 Jakub Jelinek <jakub@redhat.com>
9560 PR tree-optimization/58364
9561 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
9562 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
9563 the current range can't be an unconditional true or false.
9565 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
9567 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
9569 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
9571 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
9573 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
9576 * doc/invoke.texi (-Wdelete-incomplete): Document it.
9578 2013-09-09 Ian Bolton <ian.bolton@arm.com>
9580 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
9581 NO_REGS for immediate that can't be moved directly into FP_REGS.
9583 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
9585 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
9586 comparison with negated operand.
9587 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
9590 2013-09-09 Richard Biener <rguenther@suse.de>
9593 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
9594 of a subloop record all its block as affecting loop-closed SSA form.
9596 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9598 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
9599 of an rtx/bitpos pair.
9600 (store_fixed_bit_field): Update accordingly.
9602 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9604 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
9606 * builtins.c (expand_errno_check): Likewise.
9607 * dwarf2cfi.c (init_return_column_size): Likewise.
9608 * except.c (sjlj_mark_call_sites): Likewise.
9609 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
9610 * lra-constraints.c (emit_inc): Likewise.
9611 * ree.c (combine_set_extension): Likewise.
9612 * regmove.c (fixup_match_2): Likewise.
9613 * reload1.c (inc_for_reload): Likewise.
9615 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9617 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
9618 (force_to_mode, simplify_shift_const_1, simplify_comparison):
9619 Use gen_int_mode with the mode of the associated simplify_* call.
9620 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
9621 * expmed.c (expand_shift_1): Likewise.
9622 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9623 * loop-iv.c (iv_number_of_iterations): Likewise.
9624 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9625 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
9627 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9629 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
9630 of the associated expand_* call.
9631 (asan_emit_stack_protection): Likewise.
9632 * builtins.c (round_trampoline_addr): Likewise.
9633 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
9634 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
9635 (emit_store_flag): Likewise.
9636 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
9637 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
9639 * function.c (instantiate_virtual_regs_in_insn): Likewise.
9640 * ifcvt.c (noce_try_store_flag_constants): Likewise.
9641 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
9642 * modulo-sched.c (generate_prolog_epilog): Likewise.
9643 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
9644 (expand_ctz, expand_ffs, expand_unop): Likewise.
9646 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9648 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
9649 of the associated gen_rtx_* call.
9650 * caller-save.c (init_caller_save): Likewise.
9651 * combine.c (find_split_point, make_extraction): Likewise.
9652 (make_compound_operation): Likewise.
9653 * dwarf2out.c (mem_loc_descriptor): Likewise.
9654 * explow.c (plus_constant, probe_stack_range): Likewise.
9655 * expmed.c (expand_mult_const): Likewise.
9656 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
9657 * reload1.c (init_reload): Likewise.
9658 * valtrack.c (cleanup_auto_inc_dec): Likewise.
9659 * var-tracking.c (adjust_mems): Likewise.
9660 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
9661 rather than gen_rtx_fmt_ee.
9663 2013-09-09 Jan Hubicka <jh@suse.cz>
9666 * value-prof.c (gimple_ic): Copy also abnormal edges.
9668 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
9670 * asan.c (asan_shadow_cst): Use gen_int_mode.
9672 2013-09-08 Jan Hubicka <jh@suse.cz>
9674 * ipa-profile.c: Add toplevel comment.
9675 (ipa_propagate_frequency_1): Be more conservative when profile is read.
9676 (contains_hot_call_p): New function.
9677 (ipa_propagate_frequency): Set frequencies based on counts when
9679 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
9680 profile; do not tamper with profile after inlining if it is read.
9682 2013-09-08 Jan Hubicka <jh@suse.cz>
9684 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
9687 2013-09-08 Jan Hubicka <jh@suse.cz>
9689 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
9692 2013-09-08 Jeff Law <law@redhat.com>
9695 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
9698 2013-09-08 Andi Kleen <ak@linux.intel.com>
9700 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
9702 2013-09-08 Jan Hubicka <jh@suse.cz>
9704 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
9706 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
9708 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
9709 for non-debug insns.
9710 * lra.c (new_insn_reg): Take the containing insn as a parameter.
9711 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
9712 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
9715 2013-09-08 Jan Hubicka <jh@suse.cz>
9717 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
9718 targets and devirtualize to BUILT_IN_UNREACHABLE.
9719 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
9720 * ipa.c (walk_polymorphic_call_targets): New function.
9721 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
9722 functions; use the new timevar.
9723 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
9724 was removed from the program.
9725 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
9726 not consider it in the walk when its vtable is dead.
9727 (possible_polymorphic_call_targets_1): Pass anonymous flag to
9729 (devirt_variable_node_removal_hook): New function.
9730 (possible_polymorphic_call_targets): Also register
9731 devirt_variable_node_removal_hook.
9732 (ipa_devirt): Do not do non-speculative devirtualization.
9733 (gate_ipa_devirt): One execute if devirtualizing speculatively.
9735 2013-09-08 Jan Hubicka <jh@suse.cz>
9737 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
9738 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
9739 varpool_remove_variable_insertion_hook): Declare.
9740 * varpool.c (varpool_node_hook_list): New structure.
9741 (first_varpool_node_removal_hook,
9742 first_varpool_variable_insertion_hook): New variables.
9743 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
9744 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
9745 varpool_remove_variable_insertion_hook,
9746 varpool_call_variable_insertion_hooks): New functions.
9747 (varpool_remove_node): Use it.
9749 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
9752 * diagnostic.c (diagnostic_build_prefix): When s.file is
9753 "<built-in>" don't output line and column numbers.
9755 2013-09-06 Jan Hubicka <jh@suse.cz>
9757 * cgraphunit.c (expand_thunk): Get body before touching arguments.
9758 * lto-streamer-out.c: Stream thunks, too.
9759 * lto-streamer-in.c (input_function): Pop cfun here
9760 (lto_read_body): Instead of here.
9762 2013-09-06 Caroline Tice <cmtice@google.com>
9764 * doc/install.texi: Add documentation for the --enable-vtable-verify
9765 and the --disable-libvtv configure options.
9767 2013-09-06 Jeff Law <law@redhat.com>
9769 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
9770 edge implied equivalences into successor phis.
9772 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
9774 * resource.c (mark_referenced_resources): Handle COND_EXEC.
9776 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
9778 * resource.c (mark_target_live_regs): Compute resources taking
9779 into account if a call is predicated or not.
9781 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9783 * toplev.c (output_stack_usage): Be prepared for suffixes created by
9784 the compiler in the function names.
9786 2013-09-06 Jan Hubicka <jh@suse.cz>
9789 * ipa-inline.c (has_caller_p): New function.
9790 (want_inline_function_to_all_callers_p): Use it.
9791 (sum_callers, inline_to_all_callers): Break out from ...
9792 (ipa_inline): ... here.
9794 2013-09-06 Jan Hubicka <jh@suse.cz>
9796 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
9797 only when AVX is enabled.
9799 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9801 * config/aarch64/aarch64.md
9802 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
9803 is fpsimd_<load/store>2.
9804 (load_pair<mode>): Likewise.
9805 (store_pair<mode>): Likewise.
9807 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9809 * config/arm/types.md (type): Add "mrs" type.
9810 * config/aarch64/aarch64.md
9811 (aarch64_load_tp_hard): Make type "mrs".
9813 (load_tp_hard): Make type "mrs".
9814 * config/arm/cortex-a15.md: Update with new attributes.
9815 * config/arm/cortex-a5.md: Update with new attributes.
9816 * config/arm/cortex-a53.md: Update with new attributes.
9817 * config/arm/cortex-a7.md: Update with new attributes.
9818 * config/arm/cortex-a8.md: Update with new attributes.
9819 * config/arm/cortex-a9.md: Update with new attributes.
9820 * config/arm/cortex-m4.md: Update with new attributes.
9821 * config/arm/cortex-r4.md: Update with new attributes.
9822 * config/arm/fa526.md: Update with new attributes.
9823 * config/arm/fa606te.md: Update with new attributes.
9824 * config/arm/fa626te.md: Update with new attributes.
9825 * config/arm/fa726te.md: Update with new attributes.
9827 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9829 * config/aarch64/aarch64.md
9830 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
9831 (*movtf_aarch64): Likewise.
9833 (thumb1_movdi_insn): Use "multiple" for type where more than one
9834 instruction is used for a move.
9835 (*arm32_movhf): Likewise.
9836 (*thumb_movdf_insn): Likewise.
9838 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9840 * config/arm/types.md (type): Rename fcpys to fmov.
9842 (*arm_movsi_vfp): Rename type fcpys as fmov.
9843 (*thumb2_movsi_vfp): Likewise
9844 (*movhf_vfp_neon): Likewise
9845 (*movhf_vfp): Likewise
9846 (*movsf_vfp): Likewise
9847 (*thumb2_movsf_vfp): Likewise
9848 (*movsfcc_vfp): Likewise
9849 (*thumb2_movsfcc_vfp): Likewise
9850 * config/aarch64/aarch64-simd.md
9851 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
9852 * config/aarch64/aarch64.md
9853 (*movsi_aarch64): Replace type mov_reg with fmovs.
9854 (*movdi_aarch64): Likewise
9855 (*movsf_aarch64): Likewise
9856 (*movdf_aarch64): Likewise
9858 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
9859 * config/arm/iwmmxt.md
9860 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
9861 * config/arm/arm1020e.md: Update with new attributes.
9862 * config/arm/cortex-a15-neon.md: Update with new attributes.
9863 * config/arm/cortex-a5.md: Update with new attributes.
9864 * config/arm/cortex-a53.md: Update with new attributes.
9865 * config/arm/cortex-a7.md: Update with new attributes.
9866 * config/arm/cortex-a8-neon.md: Update with new attributes.
9867 * config/arm/cortex-a9.md: Update with new attributes.
9868 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9869 * config/arm/cortex-r4f.md: Update with new attributes.
9870 * config/arm/marvell-pj4.md: Update with new attributes.
9871 * config/arm/vfp11.md: Update with new attributes.
9873 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9875 * config/aarch64/aarch64.md
9876 (*madd<mode>): Fix type attribute.
9877 (*maddsi_uxtw): Likewise.
9878 (*msub<mode>): Likewise.
9879 (*msubsi_uxtw): Likewise.
9880 (<su_optab>maddsidi4): Likewise.
9881 (<su_optab>msubsidi4): Likewise.
9883 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9885 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
9886 * config/arm/arm.md (core_cycles): Remove fdiv.
9887 * config/arm/vfp.md:
9888 (*sqrtsf2_vfp): Update for attribute changes.
9889 (*sqrtdf2_vfp): Likewise.
9890 * config/aarch64/aarch64.md:
9891 (sqrt<mode>2): Update for attribute changes.
9892 * config/arm/arm1020e.md: Update with new attributes.
9893 * config/arm/cortex-a15-neon.md: Update with new attributes.
9894 * config/arm/cortex-a5.md: Update with new attributes.
9895 * config/arm/cortex-a53.md: Update with new attributes.
9896 * config/arm/cortex-a7.md: Update with new attributes.
9897 * config/arm/cortex-a8-neon.md: Update with new attributes.
9898 * config/arm/cortex-a9.md: Update with new attributes.
9899 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9900 * config/arm/cortex-r4f.md: Update with new attributes.
9901 * config/arm/marvell-pj4.md: Update with new attributes.
9902 * config/arm/vfp11.md: Update with new attributes.
9904 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9906 * config/arm/types.md
9907 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
9908 * config/aarch64/aarch64.md
9909 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
9911 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
9912 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
9913 (float<GPI:mode><GPF:mode>2): Likewise.
9915 (*truncsisf2_vfp): Update with new attributes.
9916 (*truncsidf2_vfp): Likewise.
9917 (fixuns_truncsfsi2): Likewise.
9918 (fixuns_truncdfsi2): Likewise.
9919 (*floatsisf2_vfp): Likewise.
9920 (*floatsidf2_vfp): Likewise.
9921 (floatunssisf2): Likewise.
9922 (floatunssidf2): Likewise.
9923 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
9924 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
9925 * config/arm/arm1020e.md: Update with new attributes.
9926 * config/arm/cortex-a15-neon.md: Update with new attributes.
9927 * config/arm/cortex-a5.md: Update with new attributes.
9928 * config/arm/cortex-a53.md: Update with new attributes.
9929 * config/arm/cortex-a7.md: Update with new attributes.
9930 * config/arm/cortex-a8-neon.md: Update with new attributes.
9931 * config/arm/cortex-a9.md: Update with new attributes.
9932 * config/arm/cortex-m4-fpu.md: Update with new attributes.
9933 * config/arm/cortex-r4f.md: Update with new attributes.
9934 * config/arm/marvell-pj4.md: Update with new attributes.
9935 * config/arm/vfp11.md: Update with new attributes.
9937 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9939 * config/aarch64/arm_neon.h
9940 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
9941 (vqtbx<1,2,3,4><q>_s8): Likewise.
9943 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9945 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
9946 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
9947 (add<mode>3): Add type attribute.
9948 (add<mode>3): Likewise.
9949 (usadd<mode>3): Likewise.
9950 (ssadd<mode>3): Likewise.
9951 (sub<mode>3): Likewise.
9952 (sub<mode>3): Likewise.
9953 (ussub<mode>3): Likewise.
9954 (sssub<mode>3): Likewise.
9955 (ssmulsa3): Likewise.
9956 (usmulusa3): Likewise.
9957 (arm_usatsihi): Likewise.
9959 (*movdi_vfp): Add types for all instructions.
9960 (*movdi_vfp_cortexa8): Likewise.
9961 (*movhf_vfp_neon): Likewise.
9962 (*movhf_vfp): Likewise.
9963 (*movdf_vfp): Likewise.
9964 (*thumb2_movdf_vfp): Likewise.
9965 (*thumb2_movdfcc_vfp): Likewise.
9966 * config/arm/arm.md: Add type attribute to all insn patterns.
9967 (*thumb1_adddi3): Add type attribute.
9968 (*arm_adddi3): Likewise.
9969 (*adddi_sesidi_di): Likewise.
9970 (*adddi_zesidi_di): Likewise.
9971 (*thumb1_addsi3): Likewise.
9972 (addsi3_compare0): Likewise.
9973 (*addsi3_compare0_scratch): Likewise.
9974 (*compare_negsi_si): Likewise.
9975 (cmpsi2_addneg): Likewise.
9976 (*addsi3_carryin_<optab>): Likewise.
9977 (*addsi3_carryin_alt2_<optab>): Likewise.
9978 (*addsi3_carryin_clobercc_<optab>): Likewise.
9979 (*subsi3_carryin): Likewise.
9980 (*subsi3_carryin_const): Likewise.
9981 (*subsi3_carryin_compare): Likewise.
9982 (*subsi3_carryin_compare_const): Likewise.
9983 (*arm_subdi3): Likewise.
9984 (*thumb_subdi3): Likewise.
9985 (*subdi_di_zesidi): Likewise.
9986 (*subdi_di_sesidi): Likewise.
9987 (*subdi_zesidi_di): Likewise.
9988 (*subdi_sesidi_di): Likewise.
9989 (*subdi_zesidi_ze): Likewise.
9990 (thumb1_subsi3_insn): Likewise.
9991 (*arm_subsi3_insn): Likewise.
9992 (*anddi3_insn): Likewise.
9993 (*anddi_zesidi_di): Likewise.
9994 (*anddi_sesdi_di): Likewise.
9995 (*ne_zeroextracts): Likewise.
9996 (*ne_zeroextracts): Likewise.
9997 (*ite_ne_zeroextr): Likewise.
9998 (*ite_ne_zeroextr): Likewise.
9999 (*anddi_notdi_di): Likewise.
10000 (*anddi_notzesidi): Likewise.
10001 (*anddi_notsesidi): Likewise.
10002 (andsi_notsi_si): Likewise.
10003 (thumb1_bicsi3): Likewise.
10004 (*iordi3_insn): Likewise.
10005 (*iordi_zesidi_di): Likewise.
10006 (*iordi_sesidi_di): Likewise.
10007 (*thumb1_iorsi3_insn): Likewise.
10008 (*xordi3_insn): Likewise.
10009 (*xordi_zesidi_di): Likewise.
10010 (*xordi_sesidi_di): Likewise.
10011 (*arm_xorsi3): Likewise.
10012 (*andsi_iorsi3_no): Likewise.
10013 (*smax_0): Likewise.
10014 (*smax_m1): Likewise.
10015 (*arm_smax_insn): Likewise.
10016 (*smin_0): Likewise.
10017 (*arm_smin_insn): Likewise.
10018 (*arm_umaxsi3): Likewise.
10019 (*arm_uminsi3): Likewise.
10020 (*minmax_arithsi): Likewise.
10021 (*minmax_arithsi_): Likewise.
10022 (*satsi_<SAT:code>): Likewise.
10023 (arm_ashldi3_1bit): Likewise.
10024 (arm_ashrdi3_1bit): Likewise.
10025 (arm_lshrdi3_1bit): Likewise.
10026 (*arm_negdi2): Likewise.
10027 (*thumb1_negdi2): Likewise.
10028 (*arm_negsi2): Likewise.
10029 (*thumb1_negsi2): Likewise.
10030 (*negdi_extendsid): Likewise.
10031 (*negdi_zero_extend): Likewise.
10032 (*arm_abssi2): Likewise.
10033 (*thumb1_abssi2): Likewise.
10034 (*arm_neg_abssi2): Likewise.
10035 (*thumb1_neg_abss): Likewise.
10036 (one_cmpldi2): Likewise.
10037 (extend<mode>di2): Likewise.
10038 (*compareqi_eq0): Likewise.
10039 (*arm_extendhisi2addsi): Likewise.
10040 (*arm_movdi): Likewise.
10041 (*thumb1_movdi_insn): Likewise.
10042 (*arm_movt): Likewise.
10043 (*thumb1_movsi_insn): Likewise.
10044 (pic_add_dot_plus_four): Likewise.
10045 (pic_add_dot_plus_eight): Likewise.
10046 (tls_load_dot_plus_eight): Likewise.
10047 (*thumb1_movhi_insn): Likewise.
10048 (*thumb1_movsf_insn): Likewise.
10049 (*movdf_soft_insn): Likewise.
10050 (*thumb_movdf_insn): Likewise.
10051 (cbranchsi4_insn): Likewise.
10052 (cbranchsi4_scratch): Likewise.
10053 (*negated_cbranchsi4): Likewise.
10054 (*tbit_cbranch): Likewise.
10055 (*tlobits_cbranch): Likewise.
10056 (*tstsi3_cbranch): Likewise.
10057 (*cbranchne_decr1): Likewise.
10058 (*addsi3_cbranch): Likewise.
10059 (*addsi3_cbranch_scratch): Likewise.
10060 (*arm_cmpdi_insn): Likewise.
10061 (*arm_cmpdi_unsig): Likewise.
10062 (*arm_cmpdi_zero): Likewise.
10063 (*thumb_cmpdi_zero): Likewise.
10064 (*deleted_compare): Likewise.
10065 (*mov_scc): Likewise.
10066 (*mov_negscc): Likewise.
10067 (*mov_notscc): Likewise.
10068 (*cstoresi_eq0_thumb1_insn): Likewise.
10069 (cstoresi_nltu_thumb1): Likewise.
10070 (cstoresi_ltu_thu): Likewise.
10071 (thumb1_addsi3_addgeu): Likewise.
10072 (*arm_jump): Likewise.
10073 (*thumb_jump): Likewise.
10074 (*check_arch2): Likewise.
10075 (arm_casesi_internal): Likewise.
10076 (thumb1_casesi_dispatch): Likewise.
10077 (*arm_indirect_jump): Likewise.
10078 (*thumb1_indirect_jump): Likewise.
10080 (*and_scc): Likewise.
10081 (*ior_scc): Likewise.
10082 (*compare_scc): Likewise.
10083 (*cond_move): Likewise.
10084 (*cond_arith): Likewise.
10085 (*cond_sub): Likewise.
10086 (*cmp_ite0): Likewise.
10087 (*cmp_ite1): Likewise.
10088 (*cmp_and): Likewise.
10089 (*cmp_ior): Likewise.
10090 (*ior_scc_scc): Likewise.
10091 (*ior_scc_scc_cmp): Likewise.
10092 (*and_scc_scc): Likewise.
10093 (*and_scc_scc_cmp): Likewise.
10094 (*and_scc_scc_nod): Likewise.
10095 (*negscc): Likewise.
10096 (movcond_addsi): Likewise.
10097 (movcond): Likewise.
10098 (*ifcompare_plus_move): Likewise.
10099 (*if_plus_move): Likewise.
10100 (*ifcompare_move_plus): Likewise.
10101 (*if_move_plus): Likewise.
10102 (*ifcompare_arith_arith): Likewise.
10103 (*if_arith_arith): Likewise.
10104 (*ifcompare_arith_move): Likewise.
10105 (*if_arith_move): Likewise.
10106 (*ifcompare_move_arith): Likewise.
10107 (*if_move_arith): Likewise.
10108 (*ifcompare_move_not): Likewise.
10109 (*if_move_not): Likewise.
10110 (*ifcompare_not_move): Likewise.
10111 (*if_not_move): Likewise.
10112 (*ifcompare_shift_move): Likewise.
10113 (*if_shift_move): Likewise.
10114 (*ifcompare_move_shift): Likewise.
10115 (*if_move_shift): Likewise.
10116 (*ifcompare_shift_shift): Likewise.
10117 (*ifcompare_not_arith): Likewise.
10118 (*ifcompare_arith_not): Likewise.
10119 (*if_arith_not): Likewise.
10120 (*ifcompare_neg_move): Likewise.
10121 (*if_neg_move): Likewise.
10122 (*ifcompare_move_neg): Likewise.
10123 (*if_move_neg): Likewise.
10124 (prologue_thumb1_interwork): Likewise.
10125 (*cond_move_not): Likewise.
10126 (*sign_extract_onebit): Likewise.
10127 (*not_signextract_onebit): Likewise.
10128 (stack_tie): Likewise.
10129 (align_4): Likewise.
10130 (align_8): Likewise.
10131 (consttable_end): Likewise.
10132 (consttable_1): Likewise.
10133 (consttable_2): Likewise.
10134 (consttable_4): Likewise.
10135 (consttable_8): Likewise.
10136 (consttable_16): Likewise.
10137 (*thumb1_tablejump): Likewise.
10138 (prefetch): Likewise.
10139 (force_register_use): Likewise.
10140 (thumb_eh_return): Likewise.
10141 (load_tp_hard): Likewise.
10142 (load_tp_soft): Likewise.
10143 (tlscall): Likewise.
10144 (*arm_movtas_ze): Likewise.
10145 (*arm_rev): Likewise.
10146 (*arm_revsh): Likewise.
10147 (*arm_rev16): Likewise.
10148 * config/arm/thumb2.md
10149 (*thumb2_smaxsi3): Likewise.
10150 (*thumb2_sminsi3): Likewise.
10151 (*thumb32_umaxsi3): Likewise.
10152 (*thumb2_uminsi3): Likewise.
10153 (*thumb2_negdi2): Likewise.
10154 (*thumb2_abssi2): Likewise.
10155 (*thumb2_neg_abss): Likewise.
10156 (*thumb2_movsi_insn): Likewise.
10157 (tls_load_dot_plus_four): Likewise.
10158 (*thumb2_movhi_insn): Likewise.
10159 (*thumb2_mov_scc): Likewise.
10160 (*thumb2_mov_negs): Likewise.
10161 (*thumb2_mov_negs): Likewise.
10162 (*thumb2_mov_nots): Likewise.
10163 (*thumb2_mov_nots): Likewise.
10164 (*thumb2_movsicc_): Likewise.
10165 (*thumb2_movsfcc_soft_insn): Likewise.
10166 (*thumb2_indirect_jump): Likewise.
10167 (*thumb2_and_scc): Likewise.
10168 (*thumb2_ior_scc): Likewise.
10169 (*thumb2_ior_scc_strict_it): Likewise.
10170 (*thumb2_cond_move): Likewise.
10171 (*thumb2_cond_arith): Likewise.
10172 (*thumb2_cond_ari): Likewise.
10173 (*thumb2_cond_sub): Likewise.
10174 (*thumb2_negscc): Likewise.
10175 (*thumb2_movcond): Likewise.
10176 (thumb2_casesi_internal): Likewise.
10177 (thumb2_casesi_internal_pic): Likewise.
10178 (*thumb2_alusi3_short): Likewise.
10179 (*thumb2_mov<mode>_shortim): Likewise.
10180 (*thumb2_addsi_short): Likewise.
10181 (*thumb2_subsi_short): Likewise.
10182 (thumb2_addsi3_compare0): Likewise.
10183 (*thumb2_cbz): Likewise.
10184 (*thumb2_cbnz): Likewise.
10185 (*thumb2_one_cmplsi2_short): Likewise.
10186 (*thumb2_negsi2_short): Likewise.
10187 (*orsi_notsi_si): Likewise.
10188 * config/arm/arm1020e.md: Update with new attributes.
10189 * config/arm/arm1026ejs.md: Update with new attributes.
10190 * config/arm/arm1136jfs.md: Update with new attributes.
10191 * config/arm/arm926ejs.md: Update with new attributes.
10192 * config/arm/cortex-a15.md: Update with new attributes.
10193 * config/arm/cortex-a5.md: Update with new attributes.
10194 * config/arm/cortex-a53.md: Update with new attributes.
10195 * config/arm/cortex-a7.md: Update with new attributes.
10196 * config/arm/cortex-a8.md: Update with new attributes.
10197 * config/arm/cortex-a9.md: Update with new attributes.
10198 * config/arm/cortex-m4.md: Update with new attributes.
10199 * config/arm/cortex-r4.md: Update with new attributes.
10200 * config/arm/fa526.md: Update with new attributes.
10201 * config/arm/fa606te.md: Update with new attributes.
10202 * config/arm/fa626te.md: Update with new attributes.
10203 * config/arm/fa726te.md: Update with new attributes.
10205 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
10207 * config/aarch64/aarch64-simd.md
10208 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
10209 <vwx> iterator to ensure correct register choice.
10210 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
10211 (aarch64_sqdmull_n<mode>): Likewise.
10212 (aarch64_sqdmull2_n<mode>_internal): Likewise.
10213 * config/aarch64/arm_neon.h
10214 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
10215 (vml<as><q>_n_<su>16): Likewise.
10216 (vml<as>l_high_lane<q>_<su>16): Likewise.
10217 (vml<as>l_high_n_<su>16): Likewise.
10218 (vml<as>l_lane<q>_<su>16): Likewise.
10219 (vml<as>l_n_<su>16): Likewise.
10220 (vmul<q>_lane<q>_<su>16): Likewise.
10221 (vmul<q>_n_<su>16): Likewise.
10222 (vmull_lane<q>_<su>16): Likewise.
10223 (vmull_n_<su>16): Likewise.
10224 (vmull_high_lane<q>_<su>16): Likewise.
10225 (vmull_high_n_<su>16): Likewise.
10226 (vqrdmulh<q>_n_s16): Likewise.
10228 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
10230 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
10231 have the correct lane parameter.
10233 2013-09-06 Richard Biener <rguenther@suse.de>
10235 * cfganal.c (control_dependences::~control_dependences):
10236 Properly free all of the vector.
10238 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
10241 * config/i386/i386.c (ix86_conditional_register_usage):
10242 Proper initialize extended SSE registers.
10244 2013-09-06 Jan Hubicka <jh@suse.cz>
10246 PR tree-optimization/58311
10247 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
10249 2013-09-06 Jan Hubicka <jh@suse.cz>
10251 * Makefile.in (tree-sra.o): Update dependencies.
10252 * tree-sra.c: Include ipa-utils.h
10253 (scan_function): Use recursive_call_p.
10254 (has_caller_p): New function.
10255 (cgraph_for_node_and_aliases): Count also callers of aliases.
10257 2013-09-06 Jan Hubicka <jh@suse.cz>
10259 PR middle-end/58094
10260 * cgraph.h (symtab_semantically_equivalent_p): Declare.
10261 * tree-tailcall.c: Include ipa-utils.h.
10262 (find_tail_calls): Use it.
10263 * ipa-pure-const.c (check_call): Likewise.
10264 * ipa-utils.c (recursive_call_p): New function.
10265 * ipa-utils.h (recursive_call_p): Dclare.
10266 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
10267 (symtab_semantically_equivalent_p): New function.
10268 * Makefile.in (tree-tailcall.o): Update dependencies.
10270 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
10272 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
10273 non-inlinable part.
10275 2013-09-06 Richard Biener <rguenther@suse.de>
10277 * lto-streamer.h (lto_global_var_decls): Remove.
10278 * Makefile.in (OBJS): Remove lto-symtab.o.
10279 (lto-symtab.o): Remove.
10280 (GTFILES): Remove lto-symtab.c
10281 * lto-symtab.c: Move to lto/
10283 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10285 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
10286 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
10287 (UNSPEC_FPINT_RINT): New constant definitions.
10288 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
10289 definition with 2 attributes.
10290 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
10291 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
10294 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10296 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
10297 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
10298 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
10299 attribute to "z196_alone".
10300 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
10303 2013-09-06 Richard Biener <rguenther@suse.de>
10305 * basic-block.h (class control_dependences): New.
10306 * tree-ssa-dce.c (control_dependence_map): Remove.
10308 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
10309 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
10310 find_pdom, find_control_dependence, find_all_control_dependences):
10312 (mark_control_dependent_edges_necessary,
10313 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
10314 tree_dce_done, perform_tree_ssa_dce): Adjust.
10315 * cfganal.c (set_control_dependence_map_bit,
10316 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
10317 find_all_control_dependences): Move from tree-ssa-dce.c and
10318 implement as methods of control_dependences class.
10319 (control_dependences::control_dependences): New.
10320 (control_dependences::~control_dependences): Likewise.
10321 (control_dependences::get_edges_dependent_on): Likewise.
10322 (control_dependences::get_edge): Likewise.
10324 2013-09-04 Jan Hubicka <jh@suse.cz>
10326 * tree.c (types_same_for_odr): Drop overactive check.
10327 * ipa-devirt.c (hash_type_name): Likewise.
10329 2013-09-04 Jan Hubicka <jh@suse.cz>
10331 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
10332 (analyze_functions): ... here.
10334 2013-09-04 Jan Hubicka <jh@suse.cz>
10336 PR middle-end/58201
10337 * cgraphunit.c (analyze_functions): Clear AUX fields
10338 after processing; initialize assembler name has.
10340 2013-09-05 Jeff Law <law@redhat.com>
10342 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
10343 from thread_around_empty_block. Record threading path into PATH.
10344 Recurse if threading through the initial block is successful.
10345 (thread_across_edge): Corresponding changes to slightly simplify.
10347 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10349 * config/aarch64/aarch64.md
10350 (type): Remove frecpe, frecps, frecpx.
10351 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
10352 fix to be a TARGET_SIMD instruction.
10353 (aarch64_frecps): Remove.
10354 * config/aarch64/aarch64-simd.md
10355 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
10356 (aarch64_frecps<mode>): Handle all float/vector of float modes.
10358 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10359 Sofiane Naci <sofiane.naci@arm.com>
10361 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
10362 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
10363 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
10364 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
10365 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
10366 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
10367 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
10368 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
10369 into "clz, "rbit". Rename "shift" to "shift_imm".
10370 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
10371 changes. Update for attribute changes all occurrences of arlo_* and
10373 * config/arm/arm-fixed.md: Update for attribute changes
10374 all occurrences of arlo_* types.
10375 * config/arm/thumb2.md: Update for attribute changes all occurrences
10377 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
10378 (cortexa7_older_only): Likewise.
10379 (cortexa7_younger): Likewise.
10380 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
10381 (1020alu_shift_op): Likewise.
10382 (1020alu_shift_reg_op): Likewise.
10383 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
10384 (alu_shift_op): Likewise.
10385 (alu_shift_reg_op): Likewise.
10386 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
10387 (11_alu_shift_op): Likewise.
10388 (11_alu_shift_reg_op): Likewise.
10389 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
10390 (9_alu_shift_reg_op): Likewise.
10391 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
10393 (cortex_a15_alu_shift): Likewise.
10394 (cortex_a15_alu_shift_reg): Likewise.
10395 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
10397 (cortex_a5_alu_shift): Likewise.
10398 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
10400 (cortex_a53_alu_shift): Likewise.
10401 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
10403 (cortex_a7_alu_reg): Likewise.
10404 (cortex_a7_alu_shift): Likewise.
10405 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
10407 (cortex_a8_alu_shift): Likewise.
10408 (cortex_a8_alu_shift_reg): Likewise.
10409 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
10410 (cortex_a9_dp_shift): Likewise.
10411 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
10413 * config/arm/cortex-r4.md
10414 (cortex_r4_alu): Update for attribute changes.
10415 (cortex_r4_mov): Likewise.
10416 (cortex_r4_alu_shift_reg): Likewise.
10417 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
10418 (526_alu_shift_op): Likewise.
10419 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
10420 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
10421 (626te_alu_shift_op): Likewise.
10422 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
10423 (726te_alu_shift_op): Likewise.
10424 (726te_alu_shift_reg_op): Likewise.
10425 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
10426 (mp626_alu_shift_op): Likewise.
10427 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
10428 (pj4_alu_conds): Likewise.
10429 (pj4_shift): Likewise.
10430 (pj4_shift_conds): Likewise.
10431 (pj4_alu_shift): Likewise.
10432 (pj4_alu_shift_conds): Likewise.
10433 * config/aarch64/aarch64.md: Update for attribute change
10434 all occurrences of arlo_* and shift* types.
10436 2013-09-05 Mike Stump <mikestump@comcast.net>
10438 * tree.h: Move documentation for tree_function_decl to tree-core.h
10439 with the declaration.
10441 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
10444 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
10445 looking for widest mode.
10447 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
10449 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
10451 2013-09-05 Richard Biener <rguenther@suse.de>
10453 PR tree-optimization/58137
10454 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
10455 Do not create vectors of pointers.
10456 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
10457 types for the components of the vector initializer.
10458 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
10459 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
10461 2013-09-05 Martin Jambor <mjambor@suse.cz>
10463 * ipa-prop.c (remove_described_reference): Accept missing references,
10464 return false if that hppens, otherwise return true.
10465 (cgraph_node_for_jfunc): New function.
10466 (try_decrement_rdesc_refcount): Likewise.
10467 (try_make_edge_direct_simple_call): Use them.
10468 (ipa_edge_removal_hook): Remove references from rdescs.
10469 (ipa_edge_duplication_hook): Clone rdescs and their references
10470 when the new edge has the same caller as the old one.
10471 * cgraph.c (cgraph_resolve_speculation): Remove speculative
10472 reference before removing any edges.
10474 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
10476 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
10478 * thumb2.md (thumb2_storewb_parisi): New pattern.
10480 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
10482 * config/aarch64/aarch64-option-extensions.def: Add
10483 AARCH64_OPT_EXTENSION of 'crc'.
10484 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
10485 (AARCH64_ISA_CRC): Ditto.
10486 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
10487 description of the CRC extension.
10489 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
10491 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
10492 * config/rs6000/linux.h: Ditto.
10493 * alpha/linux.h: Ditto.
10494 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
10495 no_c99_libc_has_function.
10496 * config/c6x/uclinux-elf.h: Ditto.
10497 * config/lm32/uclinux-elf.h: Ditto.
10498 * config/m68k/uclinux.h: Ditto.
10499 * config/moxie/uclinux.h: Ditto.
10500 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
10501 (crisv32-*-linux*, cris-*-linux*): Ditto.
10502 * config/bfin/bfin.c: Include "tm_p.h".
10504 2013-09-05 Richard Biener <rguenther@suse.de>
10506 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
10507 check for a definition without a basic-block.
10509 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
10510 Sofiane Naci <sofiane.naci@arm.com>
10512 * config/aarch64/aarch64.md
10513 (*movti_aarch64): Rename r_2_f and f_2_r.
10514 (*movsf_aarch64): Likewise.
10515 (*movdf_aarch64): Likewise.
10516 (*movtf_aarch64): Likewise.
10517 (aarch64_movdi_<mode>low): Likewise.
10518 (aarch64_movdi_<mode>high): Likewise.
10519 (aarch64_mov<mode>high_di): Likewise.
10520 (aarch64_mov<mode>low_di): Likewise.
10521 (aarch64_movtilow_tilow): Likewise.
10522 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
10523 values to config/arm/types.md
10524 (attribute "conds"): Update for attribute change.
10525 (anddi3_insn): Likewise.
10526 (iordi3_insn): Likewise.
10527 (xordi3_insn): Likewise.
10528 (one_cmpldi2): Likewise.
10529 * config/arm/types.md (type): Add Neon types.
10530 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
10531 use "type" attribute.
10532 (movmisalign<mode>_neon_store): Likewise.
10533 (movmisalign<mode>_neon_load): Likewise.
10534 (vec_set<mode>_internal): Likewise.
10535 (vec_setv2di_internal): Likewise.
10536 (vec_extract<mode>): Likewise.
10537 (vec_extractv2di): Likewise.
10538 (add<mode>3_neon): Likewise.
10539 (adddi3_neon): Likewise.
10540 (sub<mode>3_neon): Likewise.
10541 (subdi3_neon): Likewise.
10542 (mul<mode>3_neon): Likewise.
10543 (mul<mode>3add<mode>_neon): Likewise.
10544 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
10545 (fma<VCVTF:mode>4)): Likewise.
10546 (fma<VCVTF:mode>4_intrinsic): Likewise.
10547 (fmsub<VCVTF:mode>4)): Likewise.
10548 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
10549 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
10550 (ior<mode>3): Likewise.
10551 (and<mode>3): Likewise.
10552 (anddi3_neon): Likewise.
10553 (orn<mode>3_neon): Likewise.
10554 (orndi3_neon): Likewise.
10555 (bic<mode>3_neon): Likewise.
10556 (bicdi3_neon): Likewise.
10557 (xor<mode>3): Likewise.
10558 (one_cmpl<mode>2): Likewise.
10559 (abs<mode>2): Likewise.
10560 (neg<mode>2): Likewise.
10561 (umin<mode>3_neon): Likewise.
10562 (umax<mode>3_neon): Likewise.
10563 (smin<mode>3_neon): Likewise.
10564 (smax<mode>3_neon): Likewise.
10565 (vashl<mode>3): Likewise.
10566 (vashr<mode>3_imm): Likewise.
10567 (vlshr<mode>3_imm): Likewise.
10568 (ashl<mode>3_signed): Likewise.
10569 (ashl<mode>3_unsigned): Likewise.
10570 (neon_load_count): Likewise.
10571 (ashldi3_neon_noclobber): Likewise.
10572 (signed_shift_di3_neon): Likewise.
10573 (unsigned_shift_di3_neon): Likewise.
10574 (ashrdi3_neon_imm_noclobber): Likewise.
10575 (lshrdi3_neon_imm_noclobber): Likewise.
10576 (widen_ssum<mode>3): Likewise.
10577 (widen_usum<mode>3): Likewise.
10578 (quad_halves_<code>v4si): Likewise.
10579 (quad_halves_<code>v4sf): Likewise.
10580 (quad_halves_<code>v8hi): Likewise.
10581 (quad_halves_<code>v16qi): Likewise.
10582 (reduc_splus_v2di): Likewise.
10583 (neon_vpadd_internal<mode>): Likewise.
10584 (neon_vpsmin<mode>): Likewise.
10585 (neon_vpsmax<mode>): Likewise.
10586 (neon_vpumin<mode>): Likewise.
10587 (neon_vpumax<mode>): Likewise.
10588 (ss_add<mode>_neon): Likewise.
10589 (us_add<mode>_neon): Likewise.
10590 (ss_sub<mode>_neon): Likewise.
10591 (us_sub<mode>_neon): Likewise.
10592 (neon_vadd<mode>_unspec): Likewise.
10593 (neon_vaddl<mode>): Likewise.
10594 (neon_vaddw<mode>): Likewise.
10595 (neon_vhadd<mode>): Likewise.
10596 (neon_vqadd<mode>): Likewise.
10597 (neon_vaddhn<mode>): Likewise.
10598 (neon_vmul<mode>): Likewise.
10599 (neon_vmla<mode>): Likewise.
10600 (neon_vmlal<mode>): Likewise.
10601 (neon_vmls<mode>): Likewise.
10602 (neon_vmlsl<mode>): Likewise.
10603 (neon_vqdmulh<mode>): Likewise.
10604 (neon_vqdmlal<mode>): Likewise.
10605 (neon_vqdmlsl<mode>): Likewise.
10606 (neon_vmull<mode>): Likewise.
10607 (neon_vqdmull<mode>): Likewise.
10608 (neon_vsub<mode>_unspec): Likewise.
10609 (neon_vsubl<mode>): Likewise.
10610 (neon_vsubw<mode>): Likewise.
10611 (neon_vqsub<mode>): Likewise.
10612 (neon_vhsub<mode>): Likewise.
10613 (neon_vsubhn<mode>): Likewise.
10614 (neon_vceq<mode>): Likewise.
10615 (neon_vcge<mode>): Likewise.
10616 (neon_vcgeu<mode>): Likewise.
10617 (neon_vcgt<mode>): Likewise.
10618 (neon_vcgtu<mode>): Likewise.
10619 (neon_vcle<mode>): Likewise.
10620 (neon_vclt<mode>): Likewise.
10621 (neon_vcage<mode>): Likewise.
10622 (neon_vcagt<mode>): Likewise.
10623 (neon_vtst<mode>): Likewise.
10624 (neon_vabd<mode>): Likewise.
10625 (neon_vabdl<mode>): Likewise.
10626 (neon_vaba<mode>): Likewise.
10627 (neon_vabal<mode>): Likewise.
10628 (neon_vmax<mode>): Likewise.
10629 (neon_vmin<mode>): Likewise.
10630 (neon_vpaddl<mode>): Likewise.
10631 (neon_vpadal<mode>): Likewise.
10632 (neon_vpmax<mode>): Likewise.
10633 (neon_vpmin<mode>): Likewise.
10634 (neon_vrecps<mode>): Likewise.
10635 (neon_vrsqrts<mode>): Likewise.
10636 (neon_vqabs<mode>): Likewise.
10637 (neon_vqneg<mode>): Likewise.
10638 (neon_vcls<mode>): Likewise.
10639 (clz<mode>2): Likewise.
10640 (popcount<mode>2): Likewise.
10641 (neon_vrecpe): Likewise.
10642 (neon_vrsqrte): Likewise.
10643 (neon_vget_lane<mode>_sext_internal): Likewise.
10644 (neon_vget_lane<mode>_zext_internal): Likewise.
10645 (neon_vdup_n<mode>): Likewise.
10646 (neon_vdup_nv2di): Likewise.
10647 (neon_vdpu_lane<mode>_internal): Likewise.
10648 (neon_vswp<mode>): Likewise.
10649 (float<mode><V_cvtto>2): Likewise.
10650 (floatuns<mode><V_cvtto>2): Likewise.
10651 (fix_trunc<mode><V_cvtto>)2): Likewise
10652 (fixuns_trunc<mode><V_cvtto)2): Likewise.
10653 (neon_vcvt<mode>): Likewise.
10654 (neon_vcvtv4sfv4hf): Likewise.
10655 (neon_vcvtv4hfv4sf): Likewise.
10656 (neon_vcvt_n<mode>): Likewise.
10657 (neon_vmovn<mode>): Likewise.
10658 (neon_vqmovn<mode>): Likewise.
10659 (neon_vqmovun<mode>): Likewise.
10660 (neon_vmovl<mode>): Likewise.
10661 (neon_vmul_lane<mode>): Likewise.
10662 (neon_vmull_lane<mode>): Likewise.
10663 (neon_vqdmull_lane<mode>): Likewise.
10664 (neon_vqdmulh_lane<mode>): Likewise.
10665 (neon_vmla_lane<mode>): Likewise.
10666 (neon_vmlal_lane<mode>): Likewise.
10667 (neon_vqdmlal_lane<mode>): Likewise.
10668 (neon_vmls_lane<mode>): Likewise.
10669 (neon_vmlsl_lane<mode>): Likewise.
10670 (neon_vqdmlsl_lane<mode>): Likewise.
10671 (neon_vext<mode>): Likewise.
10672 (neon_vrev64<mode>): Likewise.
10673 (neon_vrev32<mode>): Likewise.
10674 (neon_vrev16<mode>): Likewise.
10675 (neon_vbsl<mode>_internal): Likewise.
10676 (neon_vshl<mode>): Likewise.
10677 (neon_vqshl<mode>): Likewise.
10678 (neon_vshr_n<mode>): Likewise.
10679 (neon_vshrn_n<mode>): Likewise.
10680 (neon_vqshrn_n<mode>): Likewise.
10681 (neon_vqshrun_n<mode>): Likewise.
10682 (neon_vshl_n<mode>): Likewise.
10683 (neon_vqshl_n<mode>): Likewise.
10684 (neon_vqshlu_n<mode>): Likewise.
10685 (neon_vshll_n<mode>): Likewise.
10686 (neon_vsra_n<mode>): Likewise.
10687 (neon_vsri_n<mode>): Likewise.
10688 (neon_vsli_n<mode>): Likewise.
10689 (neon_vtbl1v8qi): Likewise.
10690 (neon_vtbl2v8qi): Likewise.
10691 (neon_vtbl3v8qi): Likewise.
10692 (neon_vtbl4v8qi): Likewise.
10693 (neon_vtbx1v8qi): Likewise.
10694 (neon_vtbx2v8qi): Likewise.
10695 (neon_vtbx3v8qi): Likewise.
10696 (neon_vtbx4v8qi): Likewise.
10697 (neon_vtrn<mode>_internal): Likewise.
10698 (neon_vzip<mode>_internal): Likewise.
10699 (neon_vuzp<mode>_internal): Likewise.
10700 (neon_vld1<mode>): Likewise.
10701 (neon_vld1_lane<mode>): Likewise.
10702 (neon_vld1_dup<mode>): Likewise.
10703 (neon_vld1_dupv2di): Likewise.
10704 (neon_vst1<mode>): Likewise.
10705 (neon_vst1_lane<mode>): Likewise.
10706 (neon_vld2<mode>): Likewise.
10707 (neon_vld2_lane<mode>): Likewise.
10708 (neon_vld2_dup<mode>): Likewise.
10709 (neon_vst2<mode>): Likewise.
10710 (neon_vst2_lane<mode>): Likewise.
10711 (neon_vld3<mode>): Likewise.
10712 (neon_vld3qa<mode>): Likewise.
10713 (neon_vld3qb<mode>): Likewise.
10714 (neon_vld3_lane<mode>): Likewise.
10715 (neon_vld3_dup<mode>): Likewise.
10716 (neon_vst3<mode>): Likewise.
10717 (neon_vst3qa<mode>): Likewise.
10718 (neon_vst3qb<mode>): Likewise.
10719 (neon_vst3_lane<mode>): Likewise.
10720 (neon_vld4<mode>): Likewise.
10721 (neon_vld4qa<mode>): Likewise.
10722 (neon_vld4qb<mode>): Likewise.
10723 (neon_vld4_lane<mode>): Likewise.
10724 (neon_vld4_dup<mode>): Likewise.
10725 (neon_vst4<mode>): Likewise.
10726 (neon_vst4qa<mode>): Likewise.
10727 (neon_vst4qb<mode>): Likewise.
10728 (neon_vst4_lane<mode>): Likewise.
10729 (neon_vec_unpack<US>_lo_<mode>): Likewise.
10730 (neon_vec_unpack<US>_hi_<mode>): Likewise.
10731 (neon_vec_<US>mult_lo_<mode>): Likewise.
10732 (neon_vec_<US>mult_hi_<mode>): Likewise.
10733 (neon_vec_<US>shiftl_<mode>): Likewise.
10734 (neon_unpack<US>_<mode>): Likewise.
10735 (neon_vec_<US>mult_<mode>): Likewise.
10736 (vec_pack_trunc_<mode>): Likewise.
10737 (neon_vec_pack_trunk_<mode>): Likewise.
10738 (neon_vabd<mode>_2): Likewise.
10739 (neon_vabd<mode>_3): Likewise.
10740 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
10741 (thumb2_movsi_vfp): Likewise.
10742 (movdi_vfp): Likewise.
10743 (movdi_vfp_cortexa8): Likewise.
10744 (movhf_vfp_neon): Likewise.
10745 (movhf_vfp): Likewiwse.
10746 (movsf_vfp): Likewiwse.
10747 (thumb2_movsf_vfp): Likewiwse.
10748 (movdf_vfp): Likewise.
10749 (thumb2_movdf_vfp): Likewise.
10750 (movsfcc_vfp): Likewise.
10751 (thumb2_movsfcc_vfp): Likewise.
10752 (movdfcc_vfp): Likewise.
10753 (thumb2_movdfcc_vfp): Likewise.
10754 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
10755 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
10756 (v10_v2c): Likewise.
10757 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
10759 (cortex_a15_neon_int_2): Likewise.
10760 (cortex_a15_neon_int_3): Likewise.
10761 (cortex_a15_neon_int_4): Likewise.
10762 (cortex_a15_neon_int_5): Likewise.
10763 (cortex_a15_neon_vqneg_vqabs): Likewise.
10764 (cortex_a15_neon_vmov): Likewise.
10765 (cortex_a15_neon_vaba): Likewise.
10766 (cortex_a15_neon_vaba_qqq): Likewise.
10767 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10768 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10769 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
10770 scalar_64_32_long_scalar): Likewise.
10771 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10772 (cortex_a15_neon_mla_qqq_8_16): Likewise.
10773 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10774 lotype_qdd_64_32_long): Likewise.
10775 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10776 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10777 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
10778 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10779 (cortex_a15_neon_shift_1): Likewise.
10780 (cortex_a15_neon_shift_2): Likewise.
10781 (cortex_a15_neon_shift_3): Likewise.
10782 (cortex_a15_neon_vshl_ddd): Likewise.
10783 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10784 (cortex_a15_neon_vsra_vrsra): Likewise.
10785 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
10786 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
10787 (cortex_a15_neon_fp_vmul_ddd): Likewise.
10788 (cortex_a15_neon_fp_vmul_qqd): Likewise.
10789 (cortex_a15_neon_fp_vmla_ddd): Likewise.
10790 (cortex_a15_neon_fp_vmla_qqq): Likewise.
10791 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
10792 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
10793 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10794 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10795 (cortex_a15_neon_bp_simple): Likewise.
10796 (cortex_a15_neon_bp_2cycle): Likewise.
10797 (cortex_a15_neon_bp_3cycle): Likewise.
10798 (cortex_a15_neon_vld1_1_2_regs): Likewise.
10799 (cortex_a15_neon_vld1_3_4_regs): Likewise.
10800 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10801 (cortex_a15_neon_vld2_4_regs): Likewise.
10802 (cortex_a15_neon_vld3_vld4): Likewise.
10803 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10804 (cortex_a15_neon_vst1_3_4_regs): Likewise.
10805 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
10806 (cortex_a15_neon_vst3_vst4): Likewise.
10807 (cortex_a15_neon_vld1_vld2_lane): Likewise.
10808 (cortex_a15_neon_vld3_vld4_lane" 10
10809 (cortex_a15_neon_vst1_vst2_lane): Likewise.
10810 (cortex_a15_neon_vst3_vst4_lane): Likewise.
10811 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
10812 (cortex_a15_neon_ldm_2): Likewise.0
10813 (cortex_a15_neon_stm_2): Likewise.
10814 (cortex_a15_neon_mcr): Likewise.
10815 (cortex_a15_neon_mcr_2_mcrr): Likewise.
10816 (cortex_a15_neon_mrc): Likewise.
10817 (cortex_a15_neon_mrrc): Likewise.
10818 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
10820 (cortex_a15_alu_shift): Likewise.
10821 (cortex_a15_alu_shift_reg): Likewise.
10822 (cortex_a15_mult32): Likewise.
10823 (cortex_a15_mult64): Likewise.
10824 (cortex_a15_block): Likewise.
10825 (cortex_a15_branch): Likewise.
10826 (cortex_a15_load1): Likewise.
10827 (cortex_a15_load3): Likewise.
10828 (cortex_a15_store1): Likewise.
10829 (cortex_a15_store3): Likewise.
10830 (cortex_a15_call): Likewise.
10831 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
10832 (cortex_a5_f2r): Likewise.
10833 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
10835 (cortex_a53_f2r): Likewise.
10836 * config/arm/cortex-a7.md
10837 (cortex_a7_branch): Update for attribute change.
10838 (cortex_a7_call): Likewise.
10839 (cortex_a7_alu_imm): Likewise.
10840 (cortex_a7_alu_reg): Likewise.
10841 (cortex_a7_alu_shift): Likewise.
10842 (cortex_a7_mul): Likewise.
10843 (cortex_a7_load1): Likewise.
10844 (cortex_a7_store1): Likewise.
10845 (cortex_a7_load2): Likewise.
10846 (cortex_a7_store2): Likewise.
10847 (cortex_a7_load3): Likewise.
10848 (cortex_a7_store3): Likewise.
10849 (cortex_a7_load4): Likewise.
10850 (cortex_a7_store4): Likewise.
10851 (cortex_a7_fpalu): Likewise.
10852 (cortex_a7_fconst): Likewise.
10853 (cortex_a7_fpmuls): Likewise.
10854 (cortex_a7_neon_mul): Likewise.
10855 (cortex_a7_fpmacs): Likewise.
10856 (cortex_a7_neon_mla: Likewise.
10857 (cortex_a7_fpmuld: Likewise.
10858 (cortex_a7_fpmacd: Likewise.
10859 (cortex_a7_fpfmad: Likewise.
10860 (cortex_a7_fdivs: Likewise.
10861 (cortex_a7_fdivd: Likewise.
10862 (cortex_a7_r2f: Likewise.
10863 (cortex_a7_f2r: Likewise.
10864 (cortex_a7_f_flags: Likewise.
10865 (cortex_a7_f_loads: Likewise.
10866 (cortex_a7_f_loadd: Likewise.
10867 (cortex_a7_f_stores: Likewise.
10868 (cortex_a7_f_stored: Likewise.
10869 (cortex_a7_neon): Likewise.
10870 * config/arm/cortex-a8-neon.md
10871 (cortex_a8_neon_mrc): Update for attribute change.
10872 (cortex_a8_neon_mrrc): Likewise.
10873 (cortex_a8_neon_int_1): Likewise.
10874 (cortex_a8_neon_int_2): Likewise.
10875 (cortex_a8_neon_int_3): Likewise.
10876 (cortex_a8_neon_int_4): Likewise.
10877 (cortex_a8_neon_int_5): Likewise.
10878 (cortex_a8_neon_vqneg_vqabs): Likewise.
10879 (cortex_a8_neon_vmov): Likewise.
10880 (cortex_a8_neon_vaba): Likewise.
10881 (cortex_a8_neon_vaba_qqq): Likewise.
10882 (cortex_a8_neon_vsma): Likewise.
10883 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10884 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10885 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10886 long_scalar): Likewise.
10887 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10888 (cortex_a8_neon_mla_qqq_8_16): Likewise.
10889 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10890 long_scalar_qdd_64_32_long): Likewise.
10891 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10892 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10893 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
10894 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10895 (cortex_a8_neon_shift_1): Likewise.
10896 (cortex_a8_neon_shift_2): Likewise.
10897 (cortex_a8_neon_shift_3): Likewise.
10898 (cortex_a8_neon_vshl_ddd): Likewise.
10899 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10900 (cortex_a8_neon_vsra_vrsra): Likewise.
10901 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
10902 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
10903 (cortex_a8_neon_fp_vsum): Likewise.
10904 (cortex_a8_neon_fp_vmul_ddd): Likewise.
10905 (cortex_a8_neon_fp_vmul_qqd): Likewise.
10906 (cortex_a8_neon_fp_vmla_ddd): Likewise.
10907 (cortex_a8_neon_fp_vmla_qqq): Likewise.
10908 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
10909 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
10910 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10911 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10912 (cortex_a8_neon_bp_simple): Likewise.
10913 (cortex_a8_neon_bp_2cycle): Likewise.
10914 (cortex_a8_neon_bp_3cycle): Likewise.
10915 (cortex_a8_neon_ldr): Likewise.
10916 (cortex_a8_neon_str): Likewise.
10917 (cortex_a8_neon_vld1_1_2_regs): Likewise.
10918 (cortex_a8_neon_vld1_3_4_regs): Likewise.
10919 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10920 (cortex_a8_neon_vld2_4_regs): Likewise.
10921 (cortex_a8_neon_vld3_vld4): Likewise.
10922 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10923 (cortex_a8_neon_vst1_3_4_regs): Likewise.
10924 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
10925 (cortex_a8_neon_vst3_vst4): Likewise.
10926 (cortex_a8_neon_vld1_vld2_lane): Likewise.
10927 (cortex_a8_neon_vld3_vld4_lane): Likewise.
10928 (cortex_a8_neon_vst1_vst2_lane): Likewise.
10929 (cortex_a8_neon_vst3_vst4_lane): Likewise.
10930 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
10931 (cortex_a8_neon_mcr): Likewise.
10932 (cortex_a8_neon_mcr_2_mcrr): Likewise.
10933 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
10934 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
10936 (ca9_neon_mrrc): Likewise.
10937 (cortex_a9_neon_int_1): Likewise.
10938 (cortex_a9_neon_int_2): Likewise.
10939 (cortex_a9_neon_int_3): Likewise.
10940 (cortex_a9_neon_int_4): Likewise.
10941 (cortex_a9_neon_int_5): Likewise.
10942 (cortex_a9_neon_vqneg_vqabs): Likewise.
10943 (cortex_a9_neon_vmov): Likewise.
10944 (cortex_a9_neon_vaba): Likewise.
10945 (cortex_a9_neon_vaba_qqq): Likewise.
10946 (cortex_a9_neon_vsma): Likewise.
10947 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10948 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
10949 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
10950 long_scalar): Likewise.
10951 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
10952 (cortex_a9_neon_mla_qqq_8_16): Likewise.
10953 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
10954 long_scalar_qdd_64_32_long): Likewise.
10955 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
10956 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
10957 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
10958 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
10959 (cortex_a9_neon_shift_1): Likewise.
10960 (cortex_a9_neon_shift_2): Likewise.
10961 (cortex_a9_neon_shift_3): Likewise.
10962 (cortex_a9_neon_vshl_ddd): Likewise.
10963 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
10964 (cortex_a9_neon_vsra_vrsra): Likewise.
10965 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
10966 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
10967 (cortex_a9_neon_fp_vsum): Likewise.
10968 (cortex_a9_neon_fp_vmul_ddd): Likewise.
10969 (cortex_a9_neon_fp_vmul_qqd): Likewise.
10970 (cortex_a9_neon_fp_vmla_ddd): Likewise.
10971 (cortex_a9_neon_fp_vmla_qqq): Likewise.
10972 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
10973 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
10974 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
10975 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
10976 (cortex_a9_neon_bp_simple): Likewise.
10977 (cortex_a9_neon_bp_2cycle): Likewise.
10978 (cortex_a9_neon_bp_3cycle): Likewise.
10979 (cortex_a9_neon_ldr): Likewise.
10980 (cortex_a9_neon_str): Likewise.
10981 (cortex_a9_neon_vld1_1_2_regs): Likewise.
10982 (cortex_a9_neon_vld1_3_4_regs): Likewise.
10983 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
10984 (cortex_a9_neon_vld2_4_regs): Likewise.
10985 (cortex_a9_neon_vld3_vld4): Likewise.
10986 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
10987 (cortex_a9_neon_vst1_3_4_regs): Likewise.
10988 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
10989 (cortex_a9_neon_vst3_vst4): Likewise.
10990 (cortex_a9_neon_vld1_vld2_lane): Likewise.
10991 (cortex_a9_neon_vld3_vld4_lane): Likewise.
10992 (cortex_a9_neon_vst1_vst2_lane): Likewise.
10993 (cortex_a9_neon_vst3_vst4_lane): Likewise.
10994 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
10995 (cortex_a9_neon_mcr): Likewise.
10996 (cortex_a9_neon_mcr_2_mcrr): Likewise.
10997 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
10998 (cortex_a9_fps): Likewise.
10999 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
11001 (cortex_m4_fmuls): Likewise.
11002 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
11004 (cortex_r4_mrc): Likewise.
11005 * config/arm/iterators.md: Update comment referring to neon_type.
11006 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
11007 (iwmmxt_movsi_insn): Likewise.
11008 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
11010 (pj4_core_to_vfp): Likewise.
11011 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
11013 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
11014 (vfp_fstore): Likewise.
11015 * doc/md.texi: Change references to neon_type to refer to type.
11017 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11019 * tree.h (DECL_BUILT_IN): Fix typo in comment.
11021 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
11023 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
11024 lglobl if not weak.
11026 2013-09-04 Easwaran Raman <eraman@google.com>
11028 PR middle-end/57370
11029 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
11030 (build_and_add_sum): Use it.
11031 (appears_later_in_bb): Simplify code.
11033 2013-09-04 Teresa Johnson <tejohnson@google.com>
11035 * dumpfile.c (dump_finish): Don't close stderr/stdout.
11037 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
11039 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
11041 2013-09-04 Jan Hubicka <jh@suse.cz>
11043 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
11044 * ipa-devirt.c: Include diganostic.h
11045 (odr_type_d): Add types and types_set.
11046 (hash_type_name): Work for types with vtables during LTO.
11047 (odr_hasher::remove): Fix comment; destroy types_set.
11048 (add_type_duplicate): New function,
11049 (get_odr_type): Use it.
11050 (dump_type_inheritance_graph): Dump type duplicates.
11051 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
11053 * tree.c (types_same_for_odr): Give exact answers on types with
11056 2013-09-04 Dodji Seketeli <dodji@redhat.com>
11058 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
11059 explaining their differences.
11061 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
11063 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
11065 2013-09-03 Jeff Law <law@redhat.com>
11067 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
11068 when not threading through a joiner block. Pass joiner/no joiner
11069 state to register_jump_thread.
11070 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
11071 state from argument rather than implying on path length.
11072 Dump the entire jump thread path into debugging dump.
11073 * tree-flow.h (register_jump_thread): Update prototype.
11075 2013-08-29 Xinliang David Li <davidxl@google.com>
11077 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11078 Remove a trivial gcc_assert.
11080 2013-08-29 Xinliang David Li <davidxl@google.com>
11082 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
11083 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
11084 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
11085 Delay base decl alignment adjustment.
11086 * tree-vectorizer.c (vect_destroy_datarefs): New function.
11087 * tree-vectorizer.h: New data structure.
11088 (set_dr_misalignment): New function.
11089 (dr_misalignment): Ditto.
11090 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
11091 (vectorizable_load): Ditto.
11092 (ensure_base_align): New function.
11093 (vectorize_loops): Add dbg_cnt support.
11094 (execute_vect_slp): Ditto.
11095 * dbgcnt.def: New debug counter.
11096 * Makefile: New dependency.
11098 2013-09-03 Meador Inge <meadori@codesourcery.com>
11102 2013-08-30 Meador Inge <meadori@codesourcery.com>
11104 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11106 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
11108 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
11109 function descriptor.
11111 2013-09-03 Richard Biener <rguenther@suse.de>
11113 * tree-affine.c (add_elt_to_tree): Fix association issue,
11114 avoid useless converts and make sure to always return a
11115 properly typed result.
11117 2013-09-03 Richard Biener <rguenther@suse.de>
11119 PR middle-end/57656
11120 * fold-const.c (negate_expr_p): Fix division case.
11121 (negate_expr): Likewise.
11123 2013-09-03 Richard Biener <rguenther@suse.de>
11126 * tree-streamer-out.c: Include tm.h.
11127 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
11129 2013-09-03 Jan Hubicka <jh@suse.cz>
11131 * tree-profile.c (tree_profiling): Cleanup CFG when done.
11133 2013-09-03 Alan Modra <amodra@gmail.com>
11135 * config.gcc (powerpc*-*-linux*): Add support for little-endian
11136 multilibs to big-endian target and vice versa.
11137 * config/rs6000/t-linux64: Use := assignment on all vars.
11138 (MULTILIB_EXTRA_OPTS): Remove fPIC.
11139 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
11140 * config/rs6000/t-linux64le: New file.
11141 * config/rs6000/t-linux64bele: New file.
11142 * config/rs6000/t-linux64lebe: New file.
11144 2013-09-02 Jan Hubicka <jh@suse.cz>
11146 * ipa-inline-transform.c (inline_transform): Do not
11147 optimize_inline_calls when not optimizing.
11149 2013-09-02 Jan Hubicka <jh@suse.cz>
11151 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
11152 duplicated nodes for assembler names.
11153 * symtab.c (symtab_unregister_node): Do not attempt to unlink
11154 hard registers from assembler name hash.
11156 2013-09-02 Jan Hubicka <jh@suse.cz>
11158 * ipa-split.c (execute_split_functions): Split externally visible
11159 functions called once.
11161 2013-09-02 Martin Jambor <mjambor@suse.cz>
11164 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
11165 linked list. When finding the correct duplicate, also consider also
11166 the caller in additon to its inlined_to node.
11168 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
11170 * config/aarch64/aarch64-simd-builtins.def
11171 (dup_lane_scalar): Remove.
11172 * config/aarch64/aarch64-simd.md
11173 (aarch64_simd_dup): Add 'w->w' alternative.
11174 (aarch64_dup_lane<mode>): Allow for VALL.
11175 (aarch64_dup_lane_scalar<mode>): Remove.
11176 (aarch64_dup_lane_<vswap_width_name><mode>): New.
11177 (aarch64_get_lane_signed<mode>): Add w->w altenative.
11178 (aarch64_get_lane_unsigned<mode>): Likewise.
11179 (aarch64_get_lane<mode>): Likewise.
11180 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
11181 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
11182 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
11183 (VCON): Change container of V2SF.
11184 (vswap_width_name): Likewise.
11185 * config/aarch64/arm_neon.h
11186 (__aarch64_vdup_lane_any): New.
11187 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11188 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
11189 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
11191 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
11193 PR middle-end/56382
11194 * expr.c (emit_move_complex): Do not move complex FP values as parts if
11195 the source or the destination is a single hard register.
11197 2013-09-02 Richard Biener <rguenther@suse.de>
11199 PR middle-end/57511
11200 * tree-scalar-evolution.c (instantiate_scev_name): Allow
11203 2013-09-02 Richard Biener <rguenther@suse.de>
11205 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
11206 arithmetic to sizetype.
11208 2013-09-02 Bin Cheng <bin.cheng@arm.com>
11210 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
11211 Find auto-increment use both before and after candidate.
11213 2013-09-02 Marek Polacek <polacek@redhat.com>
11215 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
11217 2013-09-01 Jan Hubicka <jh@suse.cz>
11219 * Makefile.in: Add ipa-profile.o
11220 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
11221 * cgraph.c (struct cgraph_propagate_frequency_data,
11222 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
11223 ipa-profile.c; replace cgraph_ by ipa_ prefix.
11224 * cgraph.h (cgraph_propagate_frequency): Remove.
11225 * ipa-inline-analysis.c: Include ipa-utils.h;
11226 drop duplicated cfgloop.h.
11227 (inline_update_callee_summaries): Update.
11228 * ipa-profile.c: New file.
11229 * ipa-utils.h (ipa_propagate_frequency): Declare.
11230 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
11231 data-streamer.h, value-prof.h.
11232 (symtab_remove_unreachable_nodes): Update profile.
11233 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
11234 account_time_size, cmp_counts, dump_histogram,
11235 ipa_profile_generate_summary, ipa_profile_write_summary,
11236 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
11237 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
11238 Move to ipa-profile.c.
11240 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
11242 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
11244 2013-09-01 Jan Hubicka <jh@suse.cz>
11246 * common.opt (fdevirtualize-speculatively): New function.
11247 * invoke.texi (fdevirtualize-speculatively): Document.
11248 * ipa-devirt.c: Include ipa-inline.h
11249 (likely_target_p): New function.
11250 (ipa_devirt): New function.
11251 (gate_ipa_devirt): New function.
11252 (pass_data_ipa_devirt): New static var.
11253 (pass_ipa_devirt): Likewise.
11254 (make_pass_ipa_devirt): New function.
11255 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
11256 (common_handle_option): Disable devirtualization when
11257 value range profiling is available.
11258 * passes.def (pass_ipa_devirt): Add.
11259 * timever.def (TV_IPA_DEVIRT): New timevar.
11260 * tree-pass.h (make_pass_ipa_devirt):
11262 2013-09-01 Iain Sandoe <iain@codesourcery.com>
11264 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
11265 include sanitize(undefined).
11267 2013-08-31 Diego Novillo <dnovillo@google.com>
11269 * Makefile.in (TREE_CORE_H): Define.
11271 (GTFILES): Add tree-core.h.
11272 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
11274 * tree-core.h: New file.
11275 Move all data structures, enum, typedefs, global
11276 declarations and constants from ...
11277 * tree.h: ... here.
11279 2013-08-31 Jan Hubicka <jh@suse.cz>
11281 * bulitins.c (expand_builtin): Do not early exit for gcov
11282 instrumented functions.
11284 2013-08-31 Marek Polacek <polacek@redhat.com>
11286 * ubsan.c: Include tm_p.h.
11288 2013-08-31 Jan Hubicka <jh@suse.cz>
11290 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
11293 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
11294 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
11295 * tree-cfg.c (verify_gimple_label): ... here.
11296 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
11297 (ipa_merge_profiles): New function.
11298 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
11299 (lto_input_function_body): Likewise.
11300 * ipa-utils.h (ipa_merge_profiles): Declare.
11301 * lto-streamer.h (lto_input_function_body): Update prototype.
11302 (emit_label_in_global_context_p): Remove.
11303 * lto-symtab.c: Include ipa-utils.h
11304 (lto_cgraph_replace_node): Use ipa_merge_profiles.
11306 2013-08-31 Jan Hubicka <jh@suse.cz>
11308 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
11310 2013-08-31 Jan Hubicka <jh@suse.cz>
11312 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
11314 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
11316 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
11317 "cmp" RTX before signed_comparison_operator check to account
11318 for "code" changes.
11320 2013-08-30 Jan Hubicka <jh@suse.cz>
11322 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
11323 (detect_type_change_1): Rename to ...
11324 (detect_type_change): ... this one; early return on non-polymorphic
11326 (detect_type_change_ssa): Add comp_type parameter; update
11327 use of detect_type_change.
11328 (compute_complex_assign_jump_func): Add param_type parameter;
11329 update use of detect_type_change_ssa.
11330 (compute_complex_ancestor_jump_func): Likewise.
11331 (ipa_get_callee_param_type): New function.
11332 (ipa_compute_jump_functions_for_edge): Compute parameter type;
11333 update calls to the jump function computation functions.
11335 2013-08-30 Teresa Johnson <tejohnson@google.com>
11336 Steven Bosscher <steven@gcc.gnu.org>
11338 * cfgrtl.c (fixup_new_cold_bb): New routine.
11339 (commit_edge_insertions): Invoke fixup_partitions.
11340 (find_partition_fixes): New routine.
11341 (fixup_partitions): Ditto.
11342 (verify_hot_cold_block_grouping): Update comments.
11343 (rtl_verify_edges): Invoke find_partition_fixes.
11344 (rtl_verify_bb_pointers): Update comments.
11345 (rtl_verify_bb_layout): Ditto.
11346 * basic-block.h (probably_never_executed_edge_p): Declare.
11347 (fixup_partitions): Ditto.
11348 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
11349 * bb-reorder.c (sanitize_hot_paths): New function.
11350 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
11351 sanitize_hot_paths.
11352 * predict.c (probably_never_executed_edge_p): New routine.
11353 * cfg.c (check_bb_profile): Add partition insanity warnings.
11355 2013-08-30 Meador Inge <meadori@codesourcery.com>
11357 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
11359 2013-08-30 Marek Polacek <polacek@redhat.com>
11361 * Makefile.in (ubsan.o): Add.
11362 (c-family/c-ubsan.o): Add.
11363 (builtins.o): Add ubsan.h dependency.
11364 * ubsan.h: New file.
11365 * ubsan.c: New file.
11366 * common.opt: Add -fsanitize=undefined option.
11367 (flag_sanitize): Add variable.
11368 (fsanitize=): Add option. Add Driver.
11369 (fsanitize=thread): Remove option.
11370 (fsanitize=address): Likewise.
11371 (static-libubsan): New option.
11372 * doc/invoke.texi: Document the new flag and -static-libubsan.
11373 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
11374 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
11375 * builtin-attrs.def (ATTR_COLD): Define.
11376 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11377 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
11378 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
11379 * flag-types.h (sanitize_code): New enum.
11380 * opts.c (common_handle_option): Parse command line arguments
11381 of -fsanitize=. Add -fsanitize=unreachable option.
11382 * varasm.c (get_variable_section): Adjust.
11383 (assemble_noswitch_variable): Likewise.
11384 (assemble_variable): Likewise.
11385 (output_constant_def_contents): Likewise.
11386 (categorize_decl_for_section): Likewise.
11387 (place_block_symbol): Likewise.
11388 (output_object_block): Likewise.
11389 * builtins.def: Likewise.
11390 * toplev.c (compile_file): Likewise.
11391 (process_options): Likewise.
11392 * cppbuiltin.c: Likewise.
11393 * tsan.c (tsan_pass): Likewise.
11394 (tsan_gate): Likewise.
11395 (tsan_gate_O0): Likewise.
11396 * cfgexpand.c (partition_stack_vars): Likewise.
11397 (expand_stack_vars): Likewise.
11398 (defer_stack_allocation): Likewise.
11399 (expand_used_vars): Likewise.
11400 * cfgcleanup.c (old_insns_match_p): Likewise.
11401 * asan.c (asan_finish_file): Likewise.
11402 (asan_instrument): Likewise.
11403 (gate_asan): Likewise.
11404 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
11405 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
11406 (asan_global_struct): Use pointer_sized_int_node instead
11407 calling build_nonstandard_integer_type.
11408 (initialize_sanitizer_builtins): Likewise.
11409 (asan_finish_file): Likewise.
11410 * gcc.c: Document %{%:function(args):X}.
11411 (static_spec_functions): Add sanitize.
11412 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
11413 store funcval != NULL there.
11414 (do_spec_1): Adjust handle_spec_function caller.
11415 (handle_braces): Allow %:function(args) as condition.
11416 (sanitize_spec_function): New function.
11417 (ADD_STATIC_LIBUBSAN_LIBS): Define.
11418 (LIBUBSAN_SPEC): Likewise.
11419 (LIBUBSAN_EARLY_SPEC): Likewise.
11420 (SANITIZER_SPEC): Handle libubsan.
11421 (SANITIZER_EARLY_SPEC): Likewise.
11422 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
11423 instead of fsanitize=address.
11424 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
11425 instead of fsanitize=address*.
11426 * builtins.c: Include ubsan.h.
11427 (fold_builtin_0): Instrument __builtin_unreachable.
11428 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
11429 instead of flag_asan.
11430 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
11431 (pointer_sized_int_node): Define.
11432 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
11434 2013-08-30 Mike Stump <mikestump@comcast.net>
11436 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
11439 2013-08-29 Jan Hubicka <jh@suse.cz>
11441 * cgraph.c (cgraph_function_body_availability): Handle weakref
11443 * passes.def: Remove pass_fixup_cfg.
11444 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
11445 track when we need to remove functions.
11446 (gate_ipa_inline): Execute inlining always; add comment why.
11447 (pass_data_ipa_inline): Remove TODO_remove_functions.
11448 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
11449 do not produce summaries.
11450 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
11451 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
11452 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
11453 constant pool and vtable.
11455 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
11457 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
11458 New arm_neon.h's internal macros to specify 64-bit constants.
11459 Avoid using stdint.h's macros.
11461 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
11463 * recog.c (verify_changes): Verify that changes[i].old is non-zero
11464 before applying REG_P.
11466 2013-08-30 Jakub Jelinek <jakub@redhat.com>
11468 PR tree-optimization/58277
11469 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
11470 after seeing too many stmts with vdef in between dombb and current
11471 bb, invalidate everything.
11473 2013-08-30 Richard Biener <rguenther@suse.de>
11475 * fold-const.c (fold_single_bit_test): Fix overflow test.
11477 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
11479 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
11480 and which can live in a register, always retrieve the value on entry.
11481 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
11482 passed by invisible reference specially.
11483 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
11484 (vt_add_function_parameter): Correctly deal with a parameter passed by
11485 invisible reference.
11487 2013-08-30 Jan Hubicka <jh@suse.cz>
11489 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
11491 2013-08-30 Richard Biener <rguenther@suse.de>
11493 PR tree-optimization/58228
11494 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
11495 allow invariant loads in nested loop vectorization.
11497 2013-08-30 Richard Biener <rguenther@suse.de>
11499 PR tree-optimization/58223
11500 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
11501 (has_anti_or_output_dependence): ... this and adjust to also
11502 look for output dependences.
11503 (mark_nodes_having_upstream_mem_writes): Adjust.
11504 (rdg_flag_uses): Likewise.
11506 2013-08-30 Richard Biener <rguenther@suse.de>
11508 PR tree-optimization/58010
11509 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
11510 assert that we have a loop-closed PHI.
11512 2013-08-29 Jan Hubicka <jh@suse.cz>
11514 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
11515 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
11516 * lto-section-in.c (lto_free_function_in_decl_state): New function.
11517 (lto_free_function_in_decl_state_for_node): New function.
11519 2013-08-29 Xinliang David Li <davidxl@google.com>
11521 * loop-unroll.c (report_unroll_peel): Minor message change.
11522 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
11523 Emit alignment peeling message with default -fopt-info.
11524 (vect_loop_versioning): Emit loop version info message.
11525 * tree-vectorizer.c (vectorize_loops): Minor message change.
11526 (execute_vect_slp): Ditto.
11528 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
11530 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
11531 of the clone from the DECL_NAME of the original function.
11533 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
11535 * passes.c (register_pass): Add overload.
11536 * tree-pass.h (register_pass): Forward declare it. Add comment.
11538 2013-08-29 Jan Hubicka <jh@suse.cz>
11540 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
11541 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
11542 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
11543 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11544 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
11545 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
11546 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
11547 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
11549 2013-08-29 Teresa Johnson <tejohnson@google.com>
11551 * dumpfile.c (dump_loc): Output column number.
11552 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
11553 * doc/invoke.texi: Document optall -fopt-info flag.
11554 * profile.c (read_profile_edge_counts): Use new dump framework.
11555 (compute_branch_probabilities): Ditto.
11556 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
11557 when pass not in any opt group.
11558 * pass_manager.h (pass_manager::get_pass_profile): New method.
11559 * value-prof.c (check_counter): Use new dump framework.
11560 (check_ic_target): Ditto.
11561 * coverage.c (get_coverage_counts): Ditto.
11562 (coverage_init): Setup new dump framework.
11564 2013-08-29 Richard Biener <rguenther@suse.de>
11566 PR tree-optimization/58246
11567 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
11568 handle the dominance check inside a basic-block.
11570 2013-08-29 Richard Biener <rguenther@suse.de>
11572 PR middle-end/57287
11573 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
11574 of default defs that appear in abnormal PHI nodes.
11576 2013-08-29 Richard Biener <rguenther@suse.de>
11578 PR tree-optimization/57685
11579 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
11580 single-use operands to avoid exponential complexity.
11582 2013-08-28 Dehao Chen <dehao@google.com>
11584 * ipa-inline.c (edge_badness): Fix integer underflow.
11586 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
11588 * gtm-builtins.def (_ITM_free): Declare leaf.
11590 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11593 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
11594 (*tls_local_dynamic_base_64_largepic): Likewise.
11595 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
11596 Remove predicate from call operand.
11597 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
11598 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
11600 2013-08-28 Jeff Law <law@redhat.com>
11602 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
11603 checks for the number of predecessors and successors allowed.
11604 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
11605 which require copying a joiner block if there is a request which
11606 is a subpath that requires no joiner block copying.
11608 2013-08-28 Jan Hubicka <jh@suse.cz>
11610 * lto-streamer-out.c (DFS_write_tree_body): Drop
11611 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
11612 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
11613 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
11614 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
11615 Do not read DECL_ERROR_ISSUED.
11616 (unpack_ts_decl_with_vis_value_fields): Do not read
11618 (lto_input_ts_binfo_tree_pointers): Do not read
11619 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
11620 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
11621 write DECL_ERROR_ISSUED..
11622 (pack_ts_decl_with_vis_value_fields): Do not write
11624 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
11625 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
11626 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
11627 * tree.h (tree_decl_common): Update comment.
11628 (DECL_ERROR_ISSUED): Remove.
11630 2013-08-28 Jakub Jelinek <jakub@redhat.com>
11632 PR middle-end/58257
11633 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
11635 2013-08-28 Jan Hubicka <jh@suse.cz>
11637 * builtins.def (free): Declare leaf.
11639 2013-08-27 David Malcolm <dmalcolm@redhat.com>
11641 * gdbhooks.py: New.
11642 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
11643 * configure: Regenerate.
11645 2013-08-27 Martin Jambor <mjambor@suse.cz>
11647 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
11648 (ipa_ancestor_jf_data): Likewise.
11649 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
11650 (ipa_get_jf_pass_through_type_preserved): New function.
11651 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
11652 (ipa_get_jf_ancestor_type_preserved): New function.
11653 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
11654 (ipa_get_jf_ancestor_result): Likewise.
11655 (propagate_vals_accross_pass_through): Use
11656 ipa_get_jf_pass_through_result to do all the value mappings.
11657 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
11658 type_preserved flag.
11659 (ipa_set_jf_cst_copy): New function.
11660 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
11661 (ipa_set_jf_arith_pass_through): Likewise.
11662 (ipa_set_ancestor_jf): Likewise.
11663 (compute_complex_assign_jump_func): Set type_preserved instead of
11665 (ipa_compute_jump_functions_for_edge): Likewise.
11666 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
11667 (update_jump_functions_after_inlining): Update type_preserved.
11668 Explicitely create jump functions when combining one with pass_through.
11669 (ipa_write_jump_function): Stream the type_preserved flags.
11670 (ipa_read_jump_function): Likewise.
11672 2013-08-27 Jakub Jelinek <jakub@redhat.com>
11673 Aldy Hernandez <aldyh@redhat.com>
11675 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
11676 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
11677 * function.h (struct function): Add has_force_vect_loops and
11679 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
11680 * gimple.c (gimple_build_omp_critical): Add KIND argument and
11682 * gimple.def: Update CLAUSES comments.
11683 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
11684 (gimple_build_omp_for): Add argument to prototype.
11685 (gimple_omp_for_kind): New.
11686 (gimple_omp_for_set_kind): New.
11687 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
11688 GOVD_DATA_SHARE_CLASS.
11689 (enum omp_region_type): Add ORT_SIMD.
11690 (gimple_add_tmp_var): Handle ORT_SIMD.
11691 (gimplify_var_or_parm_decl): Same.
11692 (is_gimple_stmt): Same.
11693 (omp_firstprivatize_variable): Same.
11694 (omp_add_variable): Only use splay_tree_insert if lookup failed.
11695 (omp_notice_variable): Handle ORT_SIMD.
11696 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
11697 (omp_check_private): Handle ORT_SIMD.
11698 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11699 OMP_CLAUSE_SAFELEN.
11700 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
11701 Handle OMP_CLAUSE_LASTPRIVATE.
11702 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
11703 OMP_CLAUSE_SAFELEN.
11704 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
11705 (gimplify_expr): Handle OMP_SIMD.
11706 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
11707 (expand_GOMP_SIMD_VF): New.
11708 (expand_GOMP_SIMD_LAST_LANE): New.
11709 * internal-fn.def (GOMP_SIMD_LANE): New.
11710 (GOMP_SIMD_VF): New.
11711 (GOMP_SIMD_LAST_LANE): New.
11712 * omp-low.c: Include target.h.
11713 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
11714 OMP_CLAUSE_SAFELEN.
11715 (check_omp_nesting_restrictions): Same.
11717 (lower_rec_simd_input_clauses): New.
11718 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
11720 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
11721 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
11722 (expand_omp_build_assign): New.
11723 (expand_omp_for_init_counts): New.
11724 (expand_omp_for_init_vars): New.
11725 (extract_omp_for_update_vars): New.
11726 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
11727 and rewrite accordingly.
11728 (expand_omp_simd): New.
11729 (expand_omp_for): Use expand_omp_simd.
11730 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
11731 (lower_omp_for): Do not lower the body.
11732 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
11733 in their own loops.
11734 * tree-flow.h (find_omp_clause): Remove prototype.
11735 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
11736 forcing vectorization of the loop, or if flag_tree_vectorize.
11737 (gate_tree_if_conversion): Similarly.
11738 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
11739 gimple_build_omp_for.
11740 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
11741 * tree-parloops (create_parallel_loop): Pass kind argument to
11742 gimple_build_omp_for.
11743 * tree-pretty-print.c (dump_omp_clause): Add cases for
11744 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
11745 OMP_CLAUSE__SIMDUID_.
11746 (dump_generic_node): Handle OMP_SIMD.
11747 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
11748 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
11749 unroll OMP_SIMD loops here.
11750 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
11751 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
11753 (vect_analyze_data_refs): Handle simd loops.
11754 * tree-vect-loop.c (vectorizable_live_operation): Handle
11756 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
11757 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
11758 (vectorizable_load): Same.
11759 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
11760 (struct simduid_to_vf): New.
11761 (simduid_to_vf::hash): New.
11762 (simduid_to-vf::equal): New.
11763 (struct simd_array_to_simduid): New.
11764 (simd_array_to_simduid::hash): New.
11765 (simd_array_to_simduid::equal): New.
11766 (adjust_simduid_builtins): New.
11767 (struct note_simd_array_uses_struct): New.
11768 (note_simd_array_uses_cb): New.
11769 (note_simd_array_uses): New.
11770 (vectorize_loops): Handle simd hints and adjust simd builtins
11772 * tree-vectorizer.h (struct _stmt_vec_info): Add
11773 simd_lane_access_p field.
11774 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
11775 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
11776 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
11777 (omp_clause_code_name): Same.
11778 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
11779 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
11780 * tree.def (OMP_SIMD): New entry.
11781 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
11782 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
11783 (OMP_CLAUSE_DECL): Adjust range for new clauses.
11784 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
11785 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
11786 (OMP_CLAUSE_LINEAR_STEP): New.
11787 (OMP_CLAUSE_SAFELEN_EXPR): New.
11788 (OMP_CLAUSE__SIMDUID__DECL): New.
11789 (find_omp_clause): New prototype.
11791 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
11793 * config/i386/driver-i386.c (host_detect_local_cpu): Update
11794 Haswell processor detection.
11796 2013-08-27 Christian Widmer <shadow@umbrox.de>
11799 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
11800 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
11801 AVX2 capable processors.
11803 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
11805 * config/aarch64/arm_neon.h: Replace all inline asm implementations
11806 of vget_low_* with implementations in terms of other intrinsics.
11808 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
11810 PR middle-end/57219
11811 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
11812 values to -1, 0 and 1.
11814 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
11816 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
11817 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
11818 (<optab><mode>3_insn): Remove.
11819 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
11820 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
11821 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
11822 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
11823 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
11824 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
11825 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
11826 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
11827 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
11828 (ror<mode>3_insn): Likewise.
11829 * config/aarch64/predicates.md (aarch64_simd_register): New.
11831 2013-08-27 Richard Biener <rguenther@suse.de>
11833 PR tree-optimization/57521
11834 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
11835 one edge is non-critical.
11836 (find_phi_replacement_condition): Make sure to use a non-critical
11837 edge. Cleanup and remove old bug workarounds.
11838 (bb_postdominates_preds): Remove.
11839 (if_convertible_loop_p_1): Do not compute post-dominators.
11840 (combine_blocks): Do not free post-dominators.
11841 (main_tree_if_conversion): Likewise.
11842 (pass_data_if_conversion): Add TODO_verify_ssa.
11844 2013-08-27 DJ Delorie <dj@redhat.com>
11846 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
11848 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
11850 * function.c (assign_parm_find_data_types): Set passed_mode and
11851 nominal_mode to the TYPE_MODE of nominal_type for the built
11852 pointer type in case of the struct-pass-by-reference.
11854 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11856 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
11857 (UINT16_TYPE): Change default to "unsigned int".
11859 * config/avr/avr.opt (mfract-convert-truncate): New option.
11860 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
11861 is set, round negative fractional integers according to n1169
11862 when converting to integer types.
11864 2013-08-26 Jan Hubicka <jh@suse.cz>
11866 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
11867 methods can not be called indirectly when their address is not taken.
11869 2013-08-26 Jan Hubicka <jh@suse.cz>
11871 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
11874 2013-08-26 Jan Hubicka <jh@suse.cz>
11876 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
11879 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11881 * reload.c (find_valid_class): Allow classes that do not include
11882 FIRST_PSEUDO_REGISTER - 1.
11884 2013-08-26 Jan Hubicka <jh@suse.cz>
11886 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
11887 fix edge count/frequency when speculation failed; fix type check
11888 for the direct call.
11890 2013-08-26 Jan Hubicka <jh@suse.cz>
11892 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
11894 2013-08-26 Jan Hubicka <jh@suse.cz>
11896 * ipa-inline-transform.c (inline_transform): Be ready for basic block
11897 to be changed by edge redirection.
11899 2013-08-26 Jan Hubicka <jh@suse.cz>
11901 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
11902 formating; add sanity check.
11903 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
11904 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
11905 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
11906 (resolve_noninline_speculation): Update callee keys, too.
11908 2013-08-26 Jan Hubicka <jh@suse.cz>
11910 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
11911 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
11913 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
11915 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
11918 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
11920 * config/i386/i386.c (ix86_debug_options): Remove prototype.
11921 (x86_64_elf_select_section): Ditto.
11922 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
11924 (ix86_pass_by_reference): Ditto.
11925 (output_set_got): Ditto.
11926 (ix86_unary_operator_ok): Ditto.
11927 (ix86_expand_builtin): Ditto.
11929 2013-08-23 Jan Hubicka <jh@suse.cz>
11931 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
11933 2013-08-23 Jan Hubicka <jh@suse.cz>
11935 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
11936 (tree_decl_with_vis): Add FINAL field.
11938 2013-08-23 Jeff Law <law@redhat.com>
11940 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
11941 the debugging dump when the expression is fully redundant.
11943 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
11945 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
11946 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
11947 * pretty-print.c (pp_formatted_text_data): Likewise.
11948 (pp_write_text_to_stream): Likewise.
11949 (pp_write_text_as_dot_label_to_stream): Likewise.
11950 (pp_append_r): Likewise.
11951 (pp_format): Likewise.
11952 (pp_flush): Likewise.
11953 (pp_clear_output_area): Likewise.
11954 (pp_append_text): Likewise.
11955 (pp_formatted_text): Likewise.
11956 (pp_remaining_character_count_for_line): Likewise.
11957 (pp_newline): Likewise.
11958 (pp_character): Likewise.
11959 (output_buffer::~output_buffer): Define.
11960 (pretty_printer::~pretty_printer): Destruct output buffer.
11961 * pretty-print.h (output_buffer::~output_buffer): Declare.
11962 (pretty_printer::~pretty_printer): Declare virtual.
11964 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
11967 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
11968 HOST_WIDE_INT_M1U): New macros.
11969 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
11970 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
11971 unsigned -1 for lshift.
11972 * cse.c (cse_insn): Likewise.
11973 * double-int.c (rshift_double, lshift_double): Likewise.
11974 * builtins.c (fold_builtin_bitop): Likewise.
11975 * combine.c (force_to_mode): Likewise.
11976 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
11977 * simplify-rtx.c (simplify_const_unary_operation,
11978 simplify_const_binary_operation): Likewise.
11979 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
11980 check_va_list_escapes): Likewise.
11981 * rtlanal.c (nonzero_bits1): Likewise.
11982 * expmed.c (expand_smod_pow2): Likewise.
11983 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
11985 2013-08-23 Jan Hubicka <jh@suse.cz>
11987 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
11988 as having address taken.
11990 2013-08-23 Jan Hubicka <jh@suse.cz>
11992 * ipa-utils.h (method_class_type): Declare.
11993 * ipa-devirt.c (method_class_type): Export.
11995 * cgraphunit.c (analyze_functions): Do basic devirtualization;
11996 do not walk base classes of anonymous types.
11998 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
12000 PR rtl-optimization/58220
12001 PR regression/58221
12002 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
12003 handle SEQUENCE insns properly.
12005 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
12007 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
12009 (pp_newline_and_indent): Likewise.
12010 (pp_separate_with): Likewise.
12011 * pretty-print.c (pp_newline_and_flush): Define.
12012 (pp_newline_and_indent): Likewise.
12013 (pp_separate_with): Likewise.
12015 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12018 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
12019 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
12021 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
12023 * config/i386/predicates.md (ext_sse_reg_operand): New.
12024 * config/i386/i386.md (*movti_internal): Use
12025 predicate to determine if EVEX is needed.
12026 (*movsi_internal): Ditto.
12027 (*movdf_internal): Ditto.
12028 (*movsf_internal): Ditto.
12029 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
12031 2013-08-23 Jakub Jelinek <jakub@redhat.com>
12033 PR tree-optimization/58209
12034 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
12035 (find_tail_calls): Give up for pointer result types if m is non-NULL.
12036 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
12037 emit POINTER_PLUS_EXPR.
12038 (create_tailcall_accumulator): For pointer result type accumulate in
12041 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
12043 * configure.ac: Add backslashes missing from the last change.
12044 * configure: Regenerate.
12046 2013-08-22 Jan Hubicka <jh@susue.cz>
12048 * ipa.c (function_and_variable_visibility): First remember function
12049 was global and then make it local.
12051 2013-08-22 Julian Brown <julian@codesourcery.com>
12053 * configure.ac: Add aarch64 to list of arches which use "nop" in
12055 * configure: Regenerate.
12057 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12059 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
12060 gnu_libc_has_function.
12061 * config/s390/tpf.h: Likewise.
12063 2013-08-22 Jan Hubicka <jh@susue.cz>
12065 * timevar.c (validate_phases): Add cast.
12067 2013-08-22 Jan Hubicka <jh@susue.cz>
12069 * timevar.c (validate_phases): Use size_t for memory.
12070 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
12072 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
12074 * pretty-print.h (output_buffer::output_buffer): Declare.
12075 (pretty_printer::pretty_printer): Likewise.
12076 (pp_construct): Remove.
12077 * pretty-print.c (output_buffer::output_buffer): Define.
12078 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
12079 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
12080 (print_gimple_expr): Likewise.
12081 (print_gimple_seq): Likewise.
12082 (gimple_dump_bb): Likewise.
12083 * sched-vis.c (dump_value_slim): Likewise.
12084 (dump_insn_slim): Likewise.
12085 (dump_rtl_slim): Likewise.
12086 (str_pattern_slim): Likewise.
12087 * tree-mudflap.c (mf_varname_tree): Likewise.
12088 * graph.c (print_graph_cfg): Likewise.
12089 (start_graph_dump): Likewise.
12090 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
12092 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
12093 pretty printer initialization.
12094 * coretypes.h (diagnostic_context): Remove superflous type alias
12096 (pretty_printer): Likewise. Declare directly as a class.
12097 (pretty_print_info): Remove declaration as class.
12098 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
12099 and pp_clear_output_area.
12100 (asan_add_global): Likewise.
12102 2013-08-22 Jan Hubicka <jh@suse.cz>
12104 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
12105 * ipa-utils.h (update_type_inheritance_graph): Declare.
12106 (possible_polymorphic_call_target_p): Declare.
12107 (possible_polymorphic_call_target_p): New.
12108 * ipa-devirt.c: Update toplevel comments.
12109 (cached_polymorphic_call_targets): Move up.
12110 (odr_type_d): Move ID down.
12111 (polymorphic_type_binfo_p): Update comment.
12112 (odr_hasher::remove): Likewise;
12113 (get_odr_type): Set anonymous_namespace.
12114 (dump_odr_type): Dump it.
12115 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
12116 (maybe_record_node): Record node in cached_polymorphic_call_targets.
12117 (record_binfo): Add comment.
12118 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
12120 (devirt_node_removal_hook): Do not iCE when cache is freed.
12121 (possible_polymorphic_call_target_p): New predicate.
12122 (update_type_inheritance_graph): New function.
12124 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
12125 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
12126 Sergey Lega <sergey.s.lega@intel.com>
12127 Anna Tikhonova <anna.tikhonova@intel.com>
12128 Ilya Tocar <ilya.tocar@intel.com>
12129 Andrey Turetskiy <andrey.turetskiy@intel.com>
12130 Ilya Verbin <ilya.verbin@intel.com>
12131 Kirill Yukhin <kirill.yukhin@intel.com>
12132 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
12134 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
12135 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
12136 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
12137 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
12138 (OPTION_MASK_ISA_AVX2_UNSET): Update.
12139 (OPTION_MASK_ISA_AVX512F_UNSET): New.
12140 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
12141 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
12142 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
12143 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
12144 OPT_mavx512pf, OPT_mavx512er cases.
12145 * config/i386/constraints.md (v): New constraint.
12146 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
12147 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
12148 (bit_AVX512CD): New.
12149 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
12150 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
12151 * config/i386/i386-c.c (ix86_target_macros_internal):
12152 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
12154 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
12155 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
12156 * config/i386/i386.c (regclass_map, dbx_register_map)
12157 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
12158 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
12159 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
12160 -mavx512pf options.
12161 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
12162 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
12163 -mavx512pf options. Fix formatting.
12164 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
12165 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
12166 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
12167 -mavx512cd, -mavx512pf options.
12168 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
12169 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
12170 (ix86_preferred_output_reload_class): Replace SSE_REGS with
12172 (ix86_hard_regno_mode_ok): Support 512-bit registers.
12173 (ix86_set_reg_reg_cost): Ditto.
12174 (x86_order_regs_for_local_alloc): Ditto.
12175 (MAX_VECT_LEN): Extend to 64-byte.
12176 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
12177 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
12178 (TARGET_AVX512ER, TARGET_AVX512CD): New.
12179 (BIGGEST_ALIGNMENT): Extend to 512-bits.
12180 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
12181 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
12182 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
12183 (SSE_REG_MODE_P): Support new modes.
12184 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
12185 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
12186 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
12187 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
12188 (REG_CLASS_CONTENTS): Add new registers.
12189 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
12190 (EXT_REX_SSE_REGNO_P): New.
12191 (HI_REGISTER_NAMES): Add new registers.
12192 * config/i386/i386.md: Define constants for new registers.
12193 (mode): Add new 512-bit modes.
12194 (prefix): Support evex prefix.
12195 (isa): Support avx512f, noavx512f, fma_avx512f.
12196 (ssemodesuffix): Add new 512-bit modes.
12198 (*movxi_internal_avx512f): Ditto.
12199 (*movdi_internal): Replace constraint "x" with the new constraint "v".
12201 (*movsi_internal): Likewise.
12202 (*movdf_internal): Likewise.
12203 (*movsf_internal): Likewise.
12204 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
12205 (<code><mode>3): Likewise.
12206 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
12208 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
12209 with the new constraint "v".
12210 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
12212 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
12213 with the new constraint "v".
12214 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
12215 (<sse2>_storedqu<avxsizesuffix>): Likewise.
12216 (*<plusminus_insn><mode>3): Likewise.
12217 (<sse>_vm<plusminus_insn><mode>3): Likewise.
12218 (*mul<mode>3): Likewise.
12219 (<sse>_vmmul<mode>3): Likewise.
12220 (<sse>_div<mode>3): Likewise.
12221 (<sse>_vmdiv<mode>3): Likewise.
12222 (<sse>_sqrt<mode>2): Likewise.
12223 (<sse>_vmsqrt<mode>2): Likewise.
12224 (*<code><mode>3_finite): Likewise.
12225 (*<code><mode>3) <smaxmin>: Likewise.
12226 (<sse>_vm<code><mode>3): Likewise.
12227 (*<code><mode>3) <any_logic>: Likewise.
12228 (*fma_fmadd_<mode>): Likewise.
12229 (*fma_fmsub_<mode>): Likewise.
12230 (*fma_fnmadd_<mode>): Likewise.
12231 (*fma_fnmsub_<mode>): Likewise.
12232 (*fma_fmaddsub_<mode>): Likewise.
12233 (*fma_fmsubadd_<mode>): Likewise.
12234 (*fmai_fmadd_<mode>): Likewise.
12235 (*fmai_fmsub_<mode>): Likewise.
12236 (*fmai_fnmadd_<mode>): Likewise.
12237 (*fmai_fnmsub_<mode>): Likewise.
12238 (sse_cvtsi2ss): Likewise.
12239 (sse_cvtsi2ssq): Likewise.
12240 (sse_cvtss2si): Likewise.
12241 (sse_cvtss2si_2): Likewise.
12242 (sse_cvtss2siq): Likewise.
12243 (sse_cvtss2siq_2): Likewise.
12244 (sse_cvttss2si): Likewise.
12245 (sse_cvtss2siq_2): Likewise.
12246 (float<sseintvecmodelower><mode>2): Likewise.
12247 (sse2_cvtsd2si_2): Likewise.
12248 (sse2_cvtsd2siq_2): Likewise.
12249 (*<plusminus_insn><mode>3): Likewise.
12250 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
12251 (*<sse4_1_avx2>_mul<mode>3): Likewise.
12252 (ashr<mode>3): Likewise.
12253 (<shift_insn><mode>3): Likewise.
12254 (avx2_<code><mode>3): Likewise.
12255 (*avx2_<code><mode>3): Likewise.
12256 (*andnot<mode>3): Likewise.
12257 (*<code><mode>3) <any_logic>: Likewise.
12258 (abs<mode>2): Likewise.
12259 (avx2_permvar<mode>): Likewise.
12260 (avx2_perm<mode>_1): Likewise.
12261 (*avx_vpermilp<mode>): Likewise.
12262 (avx_vpermilvar<mode>3): Likewise.
12263 (avx2_ashrv<mode>): Likewise.
12264 (avx2_<shift_insn>v<mode>): Likewise.
12265 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
12267 * doc/rtl.texi: Document XImode.
12269 2013-08-21 Jeff Law <law@redhat.com>
12271 * tree-flow.h (register_jump_thread): Pass vector of edges
12272 instead of each important edge.
12273 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
12274 thread path into a vector and pass that to register_jump_thread.
12275 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
12276 passed in edge vector to the current 3-edge form.
12279 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12281 * dce.c (fini_dce): Call df_analyze again just in case
12282 delete_unmarked_insns removed anything.
12284 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
12286 * reload.h (struct reg_equivs): Rename to ..
12287 (struct reg_equivs_s): .. this.
12289 2013-08-20 Martin Liska <marxin.liska@gmail.com>
12291 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
12293 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
12295 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
12297 2013-08-21 Jeff Law <law@redhat.com>
12299 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
12300 simplify assignments too. If the RHS collapses to a singleton
12301 range, then return the value for the range.
12303 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
12305 * config/i386/sse.md (V16): Rename to...
12307 (mov<mode>): Update iterator name.
12308 (*mov<mode>_internal): Ditto.
12309 (push<mode>1): Ditto.
12310 (movmisalign<mode>): Ditto.
12312 2013-08-20 Jan Hubicka <jh@suse.cz>
12315 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
12316 entry for direct edges.
12317 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
12319 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12321 Revert my last two changes, r201865 and r201864:
12324 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12326 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12327 instances can own GC refs.
12329 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12330 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12331 (gcc::context::gt_pch_nx): Likewise.
12332 (gcc::context::gt_pch_nx): Likewise.
12333 * ggc.h (gt_ggc_mx <T>): New.
12334 (gt_pch_nx_with_op <T>): New.
12335 (gt_pch_nx <T>): New.
12336 * passes.c (opt_pass::gt_ggc_mx): New.
12337 (opt_pass::gt_pch_nx): New.
12338 (opt_pass::gt_pch_nx_with_op): New.
12339 (pass_manager::gt_ggc_mx): New.
12340 (pass_manager::gt_pch_nx): New.
12341 (pass_manager::gt_pch_nx_with_op): New.
12342 (pass_manager::operator new): Use
12343 ggc_internal_cleared_alloc_stat rather than xcalloc.
12344 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12345 (pass_manager::gt_ggc_mx): New.
12346 (pass_manager::gt_pch_nx): New.
12347 (pass_manager::gt_pch_nx_with_op): New.
12348 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12349 (opt_pass::operator new): New.
12350 (opt_pass::gt_ggc_mx): New.
12351 (opt_pass::gt_pch_nx): New.
12352 (opt_pass::gt_pch_nx_with_op): New.
12355 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12357 * Makefile.in (GTFILES): Add context.h.
12358 * context.c (gcc::context::operator new): New.
12359 (gcc::context::gt_ggc_mx): New.
12360 (gcc::context::gt_pch_nx): New.
12361 (gcc::context::gt_pch_nx): New.
12362 * context.h (gcc::context): Add GTY((user)) marking.
12363 (gcc::context::operator new): New.
12364 (gcc::context::gt_ggc_mx): New.
12365 (gcc::context::gt_pch_nx): New.
12366 (gcc::context::gt_pch_nx): New.
12367 (g): Add GTY marking.
12368 (gt_ggc_mx (gcc::context *)): New.
12369 (gt_pch_nx (gcc::context *)): New.
12370 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12371 void *cookie)): New.
12372 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12374 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
12376 * dce.c (fini_dce): Call df_analyze again just in case
12377 delete_unmarked_insns removed anything.
12379 2013-08-20 Teresa Johnson <tejohnson@google.com>
12381 PR rtl-optimizations/57451
12382 * final.c (reemit_insn_block_notes): Prevent lexical blocks
12383 from crossing split section boundaries.
12385 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
12387 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
12388 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
12389 with MULTLIB_DEFAULTS.
12391 2013-08-20 Nick Clifton <nickc@redhat.com>
12393 * target.def (narrow_volatile_bitfield): Note that the default
12394 value is false, not !TARGET_STRICT_ALIGN.
12395 * doc/tm.texi: Regenerate.
12397 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
12399 Fix LIB_SPEC for systems without libpthread.
12401 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
12402 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
12404 * config/i386/linux-common.h: Likewise.
12405 * config/mips/linux-common.h: Likewise.
12407 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
12409 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
12412 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12414 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
12415 instances can own GC refs.
12417 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
12418 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
12419 (gcc::context::gt_pch_nx): Likewise.
12420 (gcc::context::gt_pch_nx): Likewise.
12421 * ggc.h (gt_ggc_mx <T>): New.
12422 (gt_pch_nx_with_op <T>): New.
12423 (gt_pch_nx <T>): New.
12424 * passes.c (opt_pass::gt_ggc_mx): New.
12425 (opt_pass::gt_pch_nx): New.
12426 (opt_pass::gt_pch_nx_with_op): New.
12427 (pass_manager::gt_ggc_mx): New.
12428 (pass_manager::gt_pch_nx): New.
12429 (pass_manager::gt_pch_nx_with_op): New.
12430 (pass_manager::operator new): Use
12431 ggc_internal_cleared_alloc_stat rather than xcalloc.
12432 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
12433 (pass_manager::gt_ggc_mx): New.
12434 (pass_manager::gt_pch_nx): New.
12435 (pass_manager::gt_pch_nx_with_op): New.
12436 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
12437 (opt_pass::operator new): New.
12438 (opt_pass::gt_ggc_mx): New.
12439 (opt_pass::gt_pch_nx): New.
12440 (opt_pass::gt_pch_nx_with_op): New.
12442 2013-08-20 David Malcolm <dmalcolm@redhat.com>
12444 * Makefile.in (GTFILES): Add context.h.
12445 * context.c (gcc::context::operator new): New.
12446 (gcc::context::gt_ggc_mx): New.
12447 (gcc::context::gt_pch_nx): New.
12448 (gcc::context::gt_pch_nx): New.
12449 * context.h (gcc::context): Add GTY((user)) marking.
12450 (gcc::context::operator new): New.
12451 (gcc::context::gt_ggc_mx): New.
12452 (gcc::context::gt_pch_nx): New.
12453 (gcc::context::gt_pch_nx): New.
12454 (g): Add GTY marking.
12455 (gt_ggc_mx (gcc::context *)): New.
12456 (gt_pch_nx (gcc::context *)): New.
12457 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
12458 void *cookie)): New.
12459 * gengtype.c (open_base_files) <ifiles>: Add context.h.
12461 2013-08-20 Alan Modra <amodra@gmail.com>
12464 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
12465 (rs6000_emit_epilogue): Likewise.
12467 2013-08-19 Dehao Chen <dehao@google.com>
12469 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
12471 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
12472 Jakub Jelinek <jakub@redhat.com>
12474 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
12475 (BUILT_IN_FABSD64): Likewise.
12476 (BUILT_IN_FABSD128): Likewise.
12477 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
12478 (fold_builtin_1): Likewise.
12479 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
12480 destination and source operands.
12481 (*abstd2_fpr): Likewise.
12482 (*nabstd2_fpr): Likewise.
12484 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
12486 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
12489 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
12491 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
12494 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
12496 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
12497 * builtins.c (default_libc_has_function): New.
12498 (gnu_libc_has_function): Ditto.
12499 (no_c99_libc_has_function): Ditto.
12500 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
12501 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
12502 (fold_builtin_sincos): Likewise.
12503 (fold_builtin_cexp): Likewise.
12504 * builtins.def (DEF_C94_BUILTIN): Likewise.
12505 (DEF_C99_BUILTIN): Likewise.
12506 (DEF_C99_C90RES_BUILTIN): Likewise.
12507 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
12508 definitions to using this define.
12509 * config/darwin-protos.h (darwin_libc_has_function): New.
12510 * config/darwin.c (darwin_libc_has_function): Ditto.
12511 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
12512 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
12513 * config/darwin.h: Ditto.
12514 * config/elfos.h: Ditto.
12515 * config/freebsd.h: Ditto.
12516 * config/i386/cygming.h: Ditto.
12517 * config/i386/djgpp.h: Ditto.
12518 * config/i386/i386-interix.h: Ditto.
12519 * config/microblaze/microblaze.h: Ditto.
12520 * config/mmix/mmix.h: Ditto.
12521 * config/gnu-user.h: Ditto.
12522 * config/ia64/hpux.h: Ditto.
12523 * config/pa/pa-hpux.h: Ditto.
12524 * config/pdp11/pdp11.h: Ditto.
12525 * config/picochip/picochip.h: Ditto.
12526 * config/linux.h: Ditto.
12527 * config/netbsd.h: Ditto.
12528 * config/openbsd.h: Ditto.
12529 * config/rs6000/aix43.h: Ditto.
12530 * config/rs6000/aix51.h: Ditto.
12531 * config/rs6000/aix52.h: Ditto.
12532 * config/rs6000/aix53.h: Ditto.
12533 * config/rs6000/aix61.h: Ditto.
12534 * config/rs6000/darwin.h: Ditto.
12535 * config/rs6000/linux.h: Ditto.
12536 * config/rs6000/linux64.h: Ditto.
12537 * config/s390/tpf.h: Ditto.
12538 * config/sol2-10.h: Ditto.
12539 * config/sol2.h: Ditto.
12540 * config/vms/vms.h: Ditto.
12541 * config/vxworks.h: Ditto.
12542 * config/linux-android.c (linux_android_libc_has_function):
12543 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
12544 * config/linux-protos.h (linux_android_libc_has_function):
12546 * config/i386/i386.c (ix86_libc_has_function): New.
12547 * config/i386/i386-protos.h
12548 (ix86_libc_has_function): New declaration.
12549 * config/i386/i386.md
12550 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
12551 ("isinf<mode>2): Likewise.
12552 * convert.c (convert_to_integer): Using new target hook
12553 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
12554 TARGET_C99_FUNCTIONS.
12555 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
12556 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
12557 * coretypes.h (function_class): New enum for different
12558 classes of functions.
12559 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
12560 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
12561 (TARGET_HAS_SINCOS): Likewise.
12562 (TARGET_LIBC_HAS_FUNCTION): New.
12563 * doc/tm.texi: Regenerated.
12564 * targhooks.h (default_libc_has_function): New declaration.
12565 (no_c99_libc_has_function): Ditto.
12566 (gnu_libc_has_function): Ditto.
12567 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
12568 and TARGET_HAS_SINCOS.
12570 2013-08-18 Jan Hubicka <jh@suse.cz>
12572 * Makeifle-in (ipa-devirt.o): New.
12573 (GTFILES): Add ipa-utils.h and ipa-devirt.c
12574 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
12575 (analyze_functions): Look into possible targets of polymorphic call.
12576 * dumpfile.c (dump_files): Add type-inheritance dump.
12577 * dumpfile.h (TDI_inheritance): New.
12578 * ipa-devirt.c: New file.
12579 * ipa-utils.h (odr_type_d): Forward declare.
12580 (odr_type): New type.
12581 (build_type_inheritance_graph): Declare.
12582 (possible_polymorphic_call_targets): Declare and introduce inline
12583 variant when only edge is pased.
12584 (dump_possible_polymorphic_call_targets): Likewise.
12585 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
12586 * tree.c (type_in_anonymous_namespace_p): Break out from ...
12587 (types_same_for_odr): ... here.
12588 * tree.h (type_in_anonymous_namespace_p): Declare.
12590 2013-08-18 Jakub Jelinek <jakub@redhat.com>
12592 PR tree-optimization/58006
12593 * tree-parloops.c (take_address_of): Don't ICE if get_name
12595 (eliminate_local_variables_stmt): Remove clobber stmts.
12597 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
12599 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
12600 error message is issued for an alias to undefined symbol.
12602 2013-08-18 Jan Hubicka <jh@suse.cz>
12604 * cgraph.c (cgraph_create_indirect_edge): Discover
12605 polymorphic calls and record basic info into indirect_info.
12606 * gimple-fold.c (gimple_fold_call): When doing BINFO based
12607 devirtualization, ignore objc function calls.
12608 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
12609 call with no parm index info.
12610 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
12611 * tree.c (virtual_method_call_p): New function.
12612 * tree.h (virtual_method_call_p): Declare.
12614 2013-08-16 Jan Hubicka <jh@suse.cz>
12616 PR middle-end/58179
12617 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
12619 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
12621 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
12624 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12626 * gengtype.c (type_for_name): Add special-case support for
12627 locating types within the "gcc::" namespace.
12628 (open_base_files): Emit a "using namespace gcc" directive.
12630 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
12633 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
12634 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
12636 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
12637 array instead of each individual operand as a separate argument.
12638 (emit_fusion_gpr_load): Likewise.
12639 (expand_fusion_gpr_load): Add new function declaration.
12641 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
12642 signature to have the operands passed as an array, instead of as
12643 separate arguments. Allow ZERO_EXTEND to be in the memory
12644 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
12645 depend on the register live/dead flags when peepholes are run.
12646 (expand_fusion_gpr_load): New function to be called from the
12647 peephole2 pass, to change the register that addis sets to be the
12649 (emit_fusion_gpr_load): Change the calling signature to have the
12650 operands passed as an array, instead of as separate arguments.
12651 Allow ZERO_EXTEND to be in the memory address, and also
12652 SIGN_EXTEND if -mpower8-fusion-sign.
12654 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
12655 unspec enumeration.
12656 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
12657 adjust the register addis loads up in the peephole2 pass. Do not
12658 depend on the register live/dead state when the peephole pass is done.
12660 2013-08-16 David Malcolm <dmalcolm@redhat.com>
12662 * gengtype.c (create_user_defined_type): Ensure that the kind
12663 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
12664 declaration is seen before the GTY((user)) marking.
12666 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
12669 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
12671 2013-08-16 Jan Hubicka <jh@suse.cz>
12673 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
12674 arugment expected_type.
12675 (gimple_fold_call): Use it.
12676 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
12677 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
12678 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
12679 (try_make_edge_direct_virtual_call): Likewise.
12680 * tree.c (obj_type_ref_class): New.
12681 * tree.h (obj_type_ref_class): Use it.
12683 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
12685 * sched-vis.c (rtl_slim_pp_initialized): Remove.
12686 (rtl_slim_pp): Likewise.
12687 (init_rtl_slim_pretty_print): Likewise.
12688 (dump_value_slim): Don't call it. Use local pretty printer.
12689 (dump_insn_slim): Likewise.
12690 (dump_rtl_slim): Likewise.
12691 (str_pattern_slim): Likewise.
12692 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
12695 2013-08-16 Jakub Jelinek <jakub@redhat.com>
12697 PR tree-optimization/58164
12698 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
12699 walk gimple_goto_dest of GIMPLE_GOTO.
12701 PR tree-optimization/58165
12702 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
12703 bi_call must be the last stmt in a bb, don't split_block, instead
12704 use fallthru edge from it and give up if there is none.
12705 Release conds vector when returning early.
12707 2013-08-14 Xinliang David Li <davidxl@google.com>
12709 * config/i386/i386.c (ix86_option_override_internal):
12710 Remove unused variable and field.
12712 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12715 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
12716 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
12717 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
12718 and Linux, correct BLKmode alignment when 128-bit alignment is
12719 required and compatibility flag is not set.
12720 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
12721 for zero-size arguments when compatibility flag is not set.
12723 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12725 PR tree-optimization/58145
12726 * tree-sra.c (build_ref_for_offset): If prev_base has
12727 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
12729 2013-08-14 Xinliang David Li <davidxl@google.com>
12731 * config/i386/i386.c (ix86_option_override_internal):
12732 Fix uninitialized variable error.
12734 2013-08-14 Xinliang David Li <davidxl@google.com>
12736 * config/i386/i386.opt: Define two new options.
12737 * config/i386/x86-tune.def: Add arch selector field in macros.
12738 * config/i386/i386.h: Adjust macro definition.
12739 * config/i386/i386.c (ix86_option_override_internal):
12741 (parse_mtune_ctrl_str): New function.
12742 (set_ix86_tune_features): New function.
12743 (ix86_function_specific_restore): Call the new helper function.
12745 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
12747 PR rtl-optimization/57662
12748 * sel-sched.c (code_motion_process_successors): When the current insn
12749 is removed after the recursive traversal, break from the loop.
12750 Add comments and debug printouts.
12752 2013-08-14 Jakub Jelinek <jakub@redhat.com>
12753 Alexandre Oliva <aoliva@redhat.com>
12756 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
12757 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
12758 there also UNSPEC_PLTOFF.
12760 2013-08-14 Marek Polacek <polacek@redhat.com>
12762 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
12763 NUM_CONDITIONS bit positions.
12765 2013-08-13 Cary Coutant <ccoutant@google.com>
12767 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
12768 (attr_checksum): Hash vector contents instead of pointer.
12769 (attr_checksum_ordered): Likewise.
12771 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
12773 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
12774 when Pmode != word_mode. Add length_address attribute.
12775 (sse3_monitor_<mode>): Merge from sse3_monitor and
12776 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
12777 Pmode != word_mode. Update insn length attribute.
12778 * config/i386/i386.c (ix86_option_override_internal): Update
12779 ix86_gen_monitor selection for merged sse3_monitor insn.
12781 2013-08-13 Julian Brown <julian@codesourcery.com>
12783 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
12784 perform invalid legitimization on greater-than-word-size modes for
12785 TARGET_E500_DOUBLE.
12787 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
12789 * ira.c (setup_class_translate_array): Use aclass instead of cl
12790 for classes not fully covered by allocno classes.
12792 2013-08-13 Jakub Jelinek <jakub@redhat.com>
12794 PR tree-optimization/57661
12795 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
12796 * tree-inline.c (tree_function_versioning): Initialize it.
12797 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
12798 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
12799 that is not being copied.
12802 * asan.c (instrument_strlen_call): Fix typo in comment.
12803 Use char * type even for the lhs of POINTER_PLUS_EXPR.
12805 2013-08-13 Steve Ellcey <sellcey@mips.com>
12807 * config/mips/mips.md (prefetch): Use lw instead of ld on
12808 loongson in 32bit mode.
12810 2013-08-13 Nick Clifton <nickc@redhat.com>
12812 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
12814 2013-08-13 Jan Hubicka <jh@suse.cz>
12816 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
12817 introduced edge; fix typo in sanity check.
12818 (cgraph_resolve_speculation): Export; improve diagnostic.
12819 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
12820 speculation at type mismatch.
12821 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
12822 (cgraph_resolve_speculation): Declare.
12823 (symtab_can_be_discarded): New function.
12824 * value-prof.c (gimple_ic_transform): Remove actual transform code.
12825 * ipa-inline-transform.c (speculation_removed): New global var.
12826 (clone_inlined_nodes): See if speculation can be removed.
12827 (inline_call): If speculations was removed, we growths may not match.
12828 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
12829 (speculation_useful_p): New function.
12830 (resolve_noninline_speculation): New function.
12831 (inline_small_functions): Resolve useless speculations.
12832 * ipa-inline.h (speculation_useful_p): Declare
12833 * ipa.c (can_replace_by_local_alias): Simplify.
12834 (ipa_profile): Produce speculative calls in non-lto, too;
12835 add simple cost model; produce local aliases.
12837 2013-08-13 David Malcolm <dmalcolm@redhat.com>
12839 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
12842 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
12844 * config/i386/i386.c (ix86_function_versions): Use error + inform.
12846 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
12848 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
12849 iterator instead of X87MODEF.
12851 2013-08-12 Perez Read <netfirewall@gmail.com>
12854 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
12855 operand 0 for intel asm alternative.
12856 (*movabs<mode>_2): Ditto for operand 1.
12858 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
12860 * config/aarch64/arm_none.h
12861 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
12863 2013-08-12 Nick Clifton <nickc@redhat.com>
12865 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
12867 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
12869 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
12870 expand for QI/HImode operand to produce more effictive code for
12871 unsigned char(short) --> float(double) conversion.
12873 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
12875 * doc/invoke.texi: Mention that -ftls-model does not force the final
12878 2013-08-12 Marek Polacek <polacek@redhat.com>
12879 Marc Glisse <marc.glisse@inria.fr>
12881 PR tree-optimization/57980
12882 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
12883 when creating -1 constant.
12885 2013-08-10 Jan Hubicka <jh@suse.cz>
12887 Workaround binutils PR14342.
12888 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
12889 (gimple_init_edge_profiler): Likewise.
12890 (gimple_gen_ic_func_profiler): Likewise.
12892 2013-08-09 Jan Hubicka <jh@suse.cz>
12894 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
12896 2013-08-09 Xinliang David Li <davidxl@google.com>
12898 * config/i386/stringop.def: New file.
12899 * config/i386/stringop.opt: New file.
12900 * config/i386/i386-opts.h: Include stringopt.def.
12901 * config/i386/i386.opt: Include stringopt.opt.
12902 * config/i386/i386.c (ix86_option_override_internal):
12903 Override default size based stringop inline strategies with options.
12904 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
12907 2013-08-09 Jan Hubicka <jh@suse.cz>
12909 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
12911 2013-08-09 Jan Hubicka <jh@suse.cz>
12913 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
12915 (dump_cgraph_node): Dump profile-id.
12916 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
12917 and common_target_probability.
12918 * lto-cgraph.c (lto_output_edge): Stream common targets.
12919 (lto_output_node): Stream profile ids.
12920 (input_node): Stream profile ids.
12921 (input_edge): Stream common targets.
12922 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
12923 * ipa.c: Include value-prof.h
12924 (ipa_profile_generate_summary): Turn indirect call statement histograms
12925 into common targets.
12926 (ipa_profile): Turn common targets into speculative edges.
12928 2013-08-09 Jan Hubicka <jh@suse.cz>
12930 * cgraph.h (cgraph_node): Add profile_id.
12931 * value-prof.c (cgraph_node_map): Turn into pointer_map.
12932 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
12933 (del_node_map): Update.
12934 (find_func_by_funcdef_no): Replace by ...
12935 (find_func_by_profile_id): ... this one.
12936 (gimple_ic_transform): Do not remove useful histograms when
12937 speculation is not done; dump info when indirect call removal
12939 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
12940 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
12941 (__gcov_indirect_call_profiler_v2): .. this one.
12942 * profile.h (init_node_map): Update.
12943 * coverage.c (coverage_compute_profile_id): New function.
12944 * coverage.h (coverage_compute_profile_id): Declare.
12945 * tree-profile.c (init_ic_make_global_vars): Make
12946 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
12947 (gimple_init_edge_profiler): Update prototype of
12948 __gcov_indirect_call_profiler.
12949 (gimple_gen_ic_func_profiler): Simplify.
12950 (tree_profiling): Use init_node_map
12952 2013-08-09 Jan Hubicka <jh@suse.cz>
12954 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
12955 non-speculative refs.
12956 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
12957 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
12958 (cgraph_set_call_stmt): Likewise.
12959 (cgraph_create_edge_1): Fix release checking compilatoin;
12960 clear lto_stmt_uid.
12961 (cgraph_free_edge): Free indirect info.
12962 (cgraph_turn_edge_to_speculative): New function.
12963 (cgraph_speculative_call_info): New function.
12964 (cgraph_make_edge_direct): Return direct edge; handle speculation.
12965 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
12966 (dump_cgraph_node): Dump speculation.
12967 (verify_edge_count_and_frequency): Accept speculative edges.
12968 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
12969 (verify_cgraph_node): Handle speculation.
12970 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
12971 (cgraph_set_call_stmt): Update prototype.
12972 (cgraph_make_edge_direct): Update prototype.
12973 (cgraph_speculative_call_info): Declare.
12974 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
12975 to change; update call of ipa_find_references.
12976 * ipa-ref.c (ipa_record_reference): Fix return value; clear
12977 lto_stmt_uid and speculative flags.
12978 (ipa_dump_references): Dump speculation.
12979 (ipa_clone_references): Clone speculative flag.
12980 (ipa_clone_referring): Likewise.
12981 (ipa_clone_ref): New function.
12982 (ipa_find_reference): Look into lto_stmt_uids
12983 (ipa_clear_stmts_in_references): Do not clear speculative calls.
12984 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
12985 (ipa_find_reference): Update declaration.
12986 (ipa_clone_ref): Declare.
12987 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
12988 stream speculative flag.
12989 (lto_output_ref): Stream statements uids and speculation.
12990 (input_ref): Likewise.
12991 (input_edge): Stream speuclation.
12992 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
12993 (cgraph_set_call_stmt_including_clones): Handle speculation.
12994 * ipa-inline.c (heap_edge_removal_hook): New function.
12995 (inline_small_functions): Register it.
12996 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
12997 also initialize refs.
12998 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
13000 (try_make_edge_direct_simple_call): Likewise.
13001 (try_make_edge_direct_simple_call): Likewise.
13002 (update_indirect_edges_after_inlining): Likewise.
13003 (remove_described_reference): Look proper lto_stmt_uid.
13004 (propagate_controlled_uses): Likewise.
13005 (propagate_controlled_uses): Liekwise.
13006 * tree-inline.c (copy_bb): Copy speculative edges.
13007 (redirect_all_calls): New function.
13008 (copy_cfg_body): Do redirection after loop info is updated.
13009 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
13011 2013-08-09 Jan Hubicka <jh@suse.cz>
13013 * lto-streamer-out.c (output_function): Renumber PHIs.
13014 * lto-streamer-in.c (input_function): Likewise.
13016 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
13018 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
13019 (get_lane_unsigned): Likewise.
13020 (dup_lane_scalar): Likewise.
13021 (get_lane): enable for VALL.
13022 * config/aarch64/aarch64-simd.md
13023 (aarch64_dup_lane_scalar<mode>): Remove.
13024 (aarch64_get_lane_signed<mode>): Likewise.
13025 (aarch64_get_lane_unsigned<mode>): Likewise.
13026 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
13027 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
13028 (aarch64_get_lane<mode>): Enable for all vector modes.
13029 (aarch64_get_lanedi): Remove misleading constraints.
13030 * config/aarch64/arm_neon.h
13031 (__aarch64_vget_lane_any): Define.
13032 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
13033 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
13034 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
13035 * config/aarch64/iterators.md (VDQQH): New.
13036 (VDQQHS): Likewise.
13037 (vwcore): Likewise.
13039 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
13041 * configure.ac: Add GAS check for LEON instructions on SPARC.
13042 * configure: Regenerate.
13043 * config.in: Likewise.
13044 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
13046 * config/sparc/sparc.opt (LEON, LEON3): New masks.
13047 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
13049 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
13050 (AS_LEON_FLAG): New macro.
13051 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
13052 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
13053 Deal with LEON and LEON3 for the memory model.
13054 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
13055 (atomic_compare_and_swap<mode>_1): Likewise.
13056 (*atomic_compare_and_swap<mode>_1): Likewise.
13058 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
13060 * config/arm/neon.md (vcond): Fix floating-point vector
13061 comparisons against 0.
13063 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
13065 * lra-constraints.c (emit_spill_move): Remove assert.
13066 (process_alt_operands): Add more debugging
13067 output. Increase reject for spilling into memory. Decrease
13068 reject for reloading scratch.
13069 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
13071 2013-08-08 Steve Ellcey <sellcey@mips.com>
13073 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
13074 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
13075 micromips incompatible. Add nan2008.
13076 (MULTILIB_DIRNAMES): Add nan2008.
13077 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13078 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
13079 and micromips incompatible. Add nan2008.
13080 (MULTILIB_DIRNAMES): Add nan2008.
13081 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
13083 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
13085 PR rtl-optimization/58079
13086 * combine.c (combine_simplify_rtx): Avoid using SUBST if
13087 simplify_comparison has widened a comparison with an integer.
13089 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13091 * config/arm/neon.md (movmisalign<mode>): Disable when we
13092 don't allow unaligned accesses.
13093 (*movmisalign<mode>_neon_store): Likewise.
13094 (*movmisalign<mode>_neon_load): Likewise.
13095 (*movmisalign<mode>_neon_store): Likewise.
13096 (*movmisalign<mode>_neon_load): Likewise.
13098 2013-08-08 Jan Hubicka <jh@suse.cz>
13100 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
13101 (make_pass_rebuild_cgraph_edges): Also clear references.
13102 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
13103 * ipa-inline-transform.c (inline_transform): Remove all references
13105 * cgraphunit.c (expand_function): Remove all references after
13107 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
13108 (ipa_find_reference): Rewrite to iterator.
13109 (remove_stmt_references): Likewise.
13110 (ipa_clear_stmts_in_references): New function.
13111 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
13112 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
13114 * ipa-split.c (split_function): Remove references in split function.
13116 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13119 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
13120 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
13122 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
13125 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
13126 suggested mode for the assignment isn't compatible with the
13127 registers required.
13129 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
13132 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
13134 2013-08-07 Xinliang David Li <davidxl@google.com>
13136 * config/i386/i386.opt: New option -mtune-ctrl=.
13137 * config/i386/x86-tune.def: New file.
13138 * config/i386/i386.h: include x86-tune.def.
13139 * config/i386/i386.c (ix86_option_override_internal):
13140 Parsing -mtune-ctrl= option and set tune features.
13142 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
13145 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
13146 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
13149 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13151 * rtl.h (update_alignments): Declare.
13152 * final.c (grow_label_align): New function extracted from...
13153 (shorten_branches): ...here. Call it.
13154 (update_alignments): New function.
13155 * reorg.c (sibling_labels): New variable.
13156 (get_label_before): Add SIBLING parameter. If it is non-zero, push
13157 the new label along with it onto the sibling_labels vector.
13158 (fill_simple_delay_slots): Adjust call to get_label_before.
13159 (fill_slots_from_thread): Likewise.
13160 (relax_delay_slots): Likewise.
13161 (make_return_insns): Likewise.
13162 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
13164 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
13166 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
13167 document its semantics.
13168 (diagnostic_report_diagnostic): Adjust accordingly.
13170 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13172 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
13173 (sparc_option_override): ...and port to new C++ pass API.
13174 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
13176 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
13178 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
13180 2013-08-06 Caroline Tice <cmtice@google.com>
13182 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
13183 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
13184 * tree-pass.h: Add pass_vtable_verify.
13185 * varasm.c (assemble_variable): Add code to properly set the comdat
13186 section and name for the .vtable_map_vars section.
13187 (assemble_vtyv_preinit_initializer): New function.
13188 (default_sectin_type_flags): Make sure .vtable_map_vars section has
13190 * output.h: Add function decl for assemble_vtv_preinit_initializer.
13191 * vtable-verify.c: New file.
13192 * vtable-verify.h: New file.
13193 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
13194 initialiation levels.
13195 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
13196 * passes.def: Insert pass_vtable_verify.
13197 * aclocal.m4: Reorder includes.
13198 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
13199 -fvtv-counts options.
13200 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
13201 as appropriate, if -fvtable-verify=... is used.
13202 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
13203 -fvtable-verify=... is used.
13204 * Makefile.in (OBJS): Add vtable-verify.o to list.
13205 (vtable-verify.o): Add new build rule.
13206 (GTFILES): Add vtable-verify.c to list.
13207 * common.opt (fvtable-verify=): New flag.
13208 (vtv_priority): Values for fvtable-verify= flag.
13209 (fvtv-counts): New flag.
13210 (fvtv-debug): New flag.
13211 * tree.h (save_vtable_map_decl): New extern function decl.
13213 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13215 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
13216 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
13217 (pass_data_rl78_devirt): ...new pass_data instance and...
13218 (make_pass_rl78_devirt): ...new function.
13219 (rl78_asm_file_start): Port pass registration to new C++ API.
13221 2013-08-07 David Malcolm <dmalcolm@redhat.com>
13223 * coretypes.h (rtl_opt_pass): Add.
13224 (gcc::context): Add.
13225 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
13226 (epiphany_init): Port to new C++ pass API.
13227 (epiphany_optimize_mode_switching): Likewise.
13228 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
13229 (pass_manager::get_pass_mode_switching): New.
13230 (pass_manager::get_pass_peephole2): New.
13231 * mode-switching.c (pass_mode_switching): Add clone method.
13232 * recog.c (pass_peephole2): Add clone method.
13233 (pass_split_all_insns): Add clone method.
13235 2013-08-06 David Malcolm <dmalcolm@redhat.com>
13237 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
13238 (mips_option_override): ...here, porting to new C++ API for passes.
13240 2013-08-06 Jan Hubicka <jh@suse.cz>
13242 * cgraph.c (cgraph_get_body): New function based on lto.c
13244 * cgraph.h (cgraph_get_body): Declare.
13245 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
13247 * cgraphunit.c (expand_function): Get body prior expanding.
13248 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
13249 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
13251 * passes.c (do_per_function_toporder): Get body.
13252 * tree-inline.c (expand_call_inline): Get body prior inlining it.
13253 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
13255 2013-08-06 Martin Jambor <mjambor@suse.cz>
13258 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
13259 is not re-finalized. Rename second parameter to no_collect.
13261 2013-08-06 Martin Jambor <mjambor@suse.cz>
13263 PR middle-end/58041
13264 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
13265 MEM_REF has proper alignment information.
13267 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
13270 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
13272 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
13273 rtx (*) (rtx, ...) with insn_gen_fn.
13274 * genoutput.c (output_insn_data): Cast gen_? function pointers to
13275 insn_gen_fn::stored_funcptr. Add initializer braces.
13277 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13279 Rewrite how instances of passes are cloned to remove assumptions
13280 about their sizes (thus allowing pass subclasses to have
13281 additional data fields, albeit non-GC-managed ones at this point).
13283 * passes.c (make_pass_instance): Now that passes have clone
13284 methods, rewrite this function to eliminate XNEW and memcpy
13285 calls that used hardcoded sizes. Since this function no longer
13286 creates pass instances, rename it to...
13287 (add_pass_instance): ...this. Document the old way that passes were
13288 numbered and flagged, and rework this function to continue using it.
13289 (next_pass_1): Add an initial_pass argument for use by
13291 (position_pass): When adding multiple instances of a pass, use
13292 the pass's clone method, rather than relying on the XNEW/memcpy
13293 within the former make_pass_instance (now add_pass_instance).
13294 (pass_manager::pass_manager): When invoking next_pass_1, also supply
13295 the initial instance of the current pass within the pass manager.
13297 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13299 This is the automated part of the conversion of passes from C
13300 structs to C++ classes.
13302 Patch autogenerated by refactor_passes.py from
13303 https://github.com/davidmalcolm/gcc-refactoring-scripts
13304 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
13306 * asan.c (pass_asan): Convert from a global struct to a subclass of
13307 gimple_opt_pass along with...
13308 (pass_data_asan): ...new pass_data instance and...
13309 (make_pass_asan): ...new function.
13310 (pass_asan_O0): Convert from a global struct to a subclass of
13311 gimple_opt_pass along with...
13312 (pass_data_asan_O0): ...new pass_data instance and...
13313 (make_pass_asan_O0): ...new function.
13314 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
13315 subclass of rtl_opt_pass along with...
13316 (pass_data_inc_dec): ...new pass_data instance and...
13317 (make_pass_inc_dec): ...new function.
13318 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
13319 a subclass of rtl_opt_pass along with...
13320 (pass_data_reorder_blocks): ...new pass_data instance and...
13321 (make_pass_reorder_blocks): ...new function.
13322 (pass_duplicate_computed_gotos): Convert from a global struct to a
13323 subclass of rtl_opt_pass along with...
13324 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
13325 (make_pass_duplicate_computed_gotos): ...new function.
13326 (pass_partition_blocks): Convert from a global struct to a subclass of
13327 rtl_opt_pass along with...
13328 (pass_data_partition_blocks): ...new pass_data instance and...
13329 (make_pass_partition_blocks): ...new function.
13330 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
13331 struct to a subclass of rtl_opt_pass along with...
13332 (pass_data_branch_target_load_optimize1): ...new pass_data instance
13334 (make_pass_branch_target_load_optimize1): ...new function.
13335 (pass_branch_target_load_optimize2): Convert from a global struct to a
13336 subclass of rtl_opt_pass along with...
13337 (pass_data_branch_target_load_optimize2): ...new pass_data instance
13339 (make_pass_branch_target_load_optimize2): ...new function.
13340 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
13341 of rtl_opt_pass along with...
13342 (pass_data_jump): ...new pass_data instance and...
13343 (make_pass_jump): ...new function.
13344 (pass_jump2): Convert from a global struct to a subclass of
13345 rtl_opt_pass along with...
13346 (pass_data_jump2): ...new pass_data instance and...
13347 (make_pass_jump2): ...new function.
13348 * cfgexpand.c (pass_expand): Convert from a global struct to a
13349 subclass of rtl_opt_pass along with...
13350 (pass_data_expand): ...new pass_data instance and...
13351 (make_pass_expand): ...new function.
13352 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
13353 of rtl_opt_pass along with...
13354 (pass_data_free_cfg): ...new pass_data instance and...
13355 (make_pass_free_cfg): ...new function.
13356 (pass_into_cfg_layout_mode): Convert from a global struct to a
13357 subclass of rtl_opt_pass along with...
13358 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
13359 (make_pass_into_cfg_layout_mode): ...new function.
13360 (pass_outof_cfg_layout_mode): Convert from a global struct to a
13361 subclass of rtl_opt_pass along with...
13362 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
13363 (make_pass_outof_cfg_layout_mode): ...new function.
13364 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
13365 struct to a subclass of gimple_opt_pass along with...
13366 (pass_data_build_cgraph_edges): ...new pass_data instance and...
13367 (make_pass_build_cgraph_edges): ...new function.
13368 (pass_rebuild_cgraph_edges): Convert from a global struct to a
13369 subclass of gimple_opt_pass along with...
13370 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
13371 (make_pass_rebuild_cgraph_edges): ...new function.
13372 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
13373 subclass of gimple_opt_pass along with...
13374 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
13376 (make_pass_remove_cgraph_callee_edges): ...new function.
13377 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
13378 struct to a subclass of rtl_opt_pass along with...
13379 (pass_data_stack_adjustments): ...new pass_data instance and...
13380 (make_pass_stack_adjustments): ...new function.
13381 * combine.c (pass_combine): Convert from a global struct to a subclass
13382 of rtl_opt_pass along with...
13383 (pass_data_combine): ...new pass_data instance and...
13384 (make_pass_combine): ...new function.
13385 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
13386 global struct to a subclass of rtl_opt_pass along with...
13387 (pass_data_compare_elim_after_reload): ...new pass_data instance
13389 (make_pass_compare_elim_after_reload): ...new function.
13390 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
13391 of rtl_opt_pass along with...
13392 (pass_data_rtl_cprop): ...new pass_data instance and...
13393 (make_pass_rtl_cprop): ...new function.
13394 * cse.c (pass_cse): Convert from a global struct to a subclass of
13395 rtl_opt_pass along with...
13396 (pass_data_cse): ...new pass_data instance and...
13397 (make_pass_cse): ...new function.
13398 (pass_cse2): Convert from a global struct to a subclass of
13399 rtl_opt_pass along with...
13400 (pass_data_cse2): ...new pass_data instance and...
13401 (make_pass_cse2): ...new function.
13402 (pass_cse_after_global_opts): Convert from a global struct to a
13403 subclass of rtl_opt_pass along with...
13404 (pass_data_cse_after_global_opts): ...new pass_data instance and...
13405 (make_pass_cse_after_global_opts): ...new function.
13406 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
13407 of rtl_opt_pass along with...
13408 (pass_data_ud_rtl_dce): ...new pass_data instance and...
13409 (make_pass_ud_rtl_dce): ...new function.
13410 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
13411 rtl_opt_pass along with...
13412 (pass_data_fast_rtl_dce): ...new pass_data instance and...
13413 (make_pass_fast_rtl_dce): ...new function.
13414 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
13415 a subclass of rtl_opt_pass along with...
13416 (pass_data_df_initialize_opt): ...new pass_data instance and...
13417 (make_pass_df_initialize_opt): ...new function.
13418 (pass_df_initialize_no_opt): Convert from a global struct to a
13419 subclass of rtl_opt_pass along with...
13420 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
13421 (make_pass_df_initialize_no_opt): ...new function.
13422 (pass_df_finish): Convert from a global struct to a subclass of
13423 rtl_opt_pass along with...
13424 (pass_data_df_finish): ...new pass_data instance and...
13425 (make_pass_df_finish): ...new function.
13426 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
13427 rtl_opt_pass along with...
13428 (pass_data_rtl_dse1): ...new pass_data instance and...
13429 (make_pass_rtl_dse1): ...new function.
13430 (pass_rtl_dse2): Convert from a global struct to a subclass of
13431 rtl_opt_pass along with...
13432 (pass_data_rtl_dse2): ...new pass_data instance and...
13433 (make_pass_rtl_dse2): ...new function.
13434 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
13435 subclass of rtl_opt_pass along with...
13436 (pass_data_dwarf2_frame): ...new pass_data instance and...
13437 (make_pass_dwarf2_frame): ...new function.
13438 * except.c (pass_set_nothrow_function_flags): Convert from a global
13439 struct to a subclass of rtl_opt_pass along with...
13440 (pass_data_set_nothrow_function_flags): ...new pass_data instance
13442 (make_pass_set_nothrow_function_flags): ...new function.
13443 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
13444 subclass of rtl_opt_pass along with...
13445 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
13447 (make_pass_convert_to_eh_region_ranges): ...new function.
13448 * final.c (pass_compute_alignments): Convert from a global struct to a
13449 subclass of rtl_opt_pass along with...
13450 (pass_data_compute_alignments): ...new pass_data instance and...
13451 (make_pass_compute_alignments): ...new function.
13452 (pass_final): Convert from a global struct to a subclass of
13453 rtl_opt_pass along with...
13454 (pass_data_final): ...new pass_data instance and...
13455 (make_pass_final): ...new function.
13456 (pass_shorten_branches): Convert from a global struct to a subclass of
13457 rtl_opt_pass along with...
13458 (pass_data_shorten_branches): ...new pass_data instance and...
13459 (make_pass_shorten_branches): ...new function.
13460 (pass_clean_state): Convert from a global struct to a subclass of
13461 rtl_opt_pass along with...
13462 (pass_data_clean_state): ...new pass_data instance and...
13463 (make_pass_clean_state): ...new function.
13464 * function.c (pass_instantiate_virtual_regs): Convert from a global
13465 struct to a subclass of rtl_opt_pass along with...
13466 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
13467 (make_pass_instantiate_virtual_regs): ...new function.
13468 (pass_leaf_regs): Convert from a global struct to a subclass of
13469 rtl_opt_pass along with...
13470 (pass_data_leaf_regs): ...new pass_data instance and...
13471 (make_pass_leaf_regs): ...new function.
13472 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
13473 subclass of rtl_opt_pass along with...
13474 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
13476 (make_pass_thread_prologue_and_epilogue): ...new function.
13477 (pass_match_asm_constraints): Convert from a global struct to a
13478 subclass of rtl_opt_pass along with...
13479 (pass_data_match_asm_constraints): ...new pass_data instance and...
13480 (make_pass_match_asm_constraints): ...new function.
13481 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
13482 subclass of rtl_opt_pass along with...
13483 (pass_data_rtl_fwprop): ...new pass_data instance and...
13484 (make_pass_rtl_fwprop): ...new function.
13485 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
13486 rtl_opt_pass along with...
13487 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
13488 (make_pass_rtl_fwprop_addr): ...new function.
13489 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
13490 rtl_opt_pass along with...
13491 (pass_data_rtl_pre): ...new pass_data instance and...
13492 (make_pass_rtl_pre): ...new function.
13493 (pass_rtl_hoist): Convert from a global struct to a subclass of
13494 rtl_opt_pass along with...
13495 (pass_data_rtl_hoist): ...new pass_data instance and...
13496 (make_pass_rtl_hoist): ...new function.
13497 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
13498 subclass of gimple_opt_pass along with...
13499 (pass_data_lower_cf): ...new pass_data instance and...
13500 (make_pass_lower_cf): ...new function.
13501 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
13502 from a global struct to a subclass of gimple_opt_pass along with...
13503 (pass_data_strength_reduction): ...new pass_data instance and...
13504 (make_pass_strength_reduction): ...new function.
13505 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
13506 of rtl_opt_pass along with...
13507 (pass_data_rtl_ifcvt): ...new pass_data instance and...
13508 (make_pass_rtl_ifcvt): ...new function.
13509 (pass_if_after_combine): Convert from a global struct to a subclass of
13510 rtl_opt_pass along with...
13511 (pass_data_if_after_combine): ...new pass_data instance and...
13512 (make_pass_if_after_combine): ...new function.
13513 (pass_if_after_reload): Convert from a global struct to a subclass of
13514 rtl_opt_pass along with...
13515 (pass_data_if_after_reload): ...new pass_data instance and...
13516 (make_pass_if_after_reload): ...new function.
13517 * init-regs.c (pass_initialize_regs): Convert from a global struct to
13518 a subclass of rtl_opt_pass along with...
13519 (pass_data_initialize_regs): ...new pass_data instance and...
13520 (make_pass_initialize_regs): ...new function.
13521 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
13522 of ipa_opt_pass_d along with...
13523 (pass_data_ipa_cp): ...new pass_data instance and...
13524 (make_pass_ipa_cp): ...new function.
13525 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
13526 global struct to a subclass of gimple_opt_pass along with...
13527 (pass_data_inline_parameters): ...new pass_data instance and...
13528 (make_pass_inline_parameters): ...new function.
13529 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
13530 subclass of gimple_opt_pass along with...
13531 (pass_data_early_inline): ...new pass_data instance and...
13532 (make_pass_early_inline): ...new function.
13533 (pass_ipa_inline): Convert from a global struct to a subclass of
13534 ipa_opt_pass_d along with...
13535 (pass_data_ipa_inline): ...new pass_data instance and...
13536 (make_pass_ipa_inline): ...new function.
13537 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
13538 struct to a subclass of gimple_opt_pass along with...
13539 (pass_data_local_pure_const): ...new pass_data instance and...
13540 (make_pass_local_pure_const): ...new function.
13541 (pass_ipa_pure_const): Convert from a global struct to a subclass of
13542 ipa_opt_pass_d along with...
13543 (pass_data_ipa_pure_const): ...new pass_data instance and...
13544 (make_pass_ipa_pure_const): ...new function.
13545 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
13546 to a subclass of ipa_opt_pass_d along with...
13547 (pass_data_ipa_reference): ...new pass_data instance and...
13548 (make_pass_ipa_reference): ...new function.
13549 * ipa-split.c (pass_split_functions): Convert from a global struct to
13550 a subclass of gimple_opt_pass along with...
13551 (pass_data_split_functions): ...new pass_data instance and...
13552 (make_pass_split_functions): ...new function.
13553 (pass_feedback_split_functions): Convert from a global struct to a
13554 subclass of gimple_opt_pass along with...
13555 (pass_data_feedback_split_functions): ...new pass_data instance and...
13556 (make_pass_feedback_split_functions): ...new function.
13557 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
13558 global struct to a subclass of simple_ipa_opt_pass along with...
13559 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
13561 (make_pass_ipa_function_and_variable_visibility): ...new function.
13562 (pass_ipa_free_inline_summary): Convert from a global struct to a
13563 subclass of simple_ipa_opt_pass along with...
13564 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
13565 (make_pass_ipa_free_inline_summary): ...new function.
13566 (pass_ipa_whole_program_visibility): Convert from a global struct to a
13567 subclass of ipa_opt_pass_d along with...
13568 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
13570 (make_pass_ipa_whole_program_visibility): ...new function.
13571 (pass_ipa_profile): Convert from a global struct to a subclass of
13572 ipa_opt_pass_d along with...
13573 (pass_data_ipa_profile): ...new pass_data instance and...
13574 (make_pass_ipa_profile): ...new function.
13575 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
13576 ipa_opt_pass_d along with...
13577 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
13578 (make_pass_ipa_cdtor_merge): ...new function.
13579 * ira.c (pass_ira): Convert from a global struct to a subclass of
13580 rtl_opt_pass along with...
13581 (pass_data_ira): ...new pass_data instance and...
13582 (make_pass_ira): ...new function.
13583 (pass_reload): Convert from a global struct to a subclass of
13584 rtl_opt_pass along with...
13585 (pass_data_reload): ...new pass_data instance and...
13586 (make_pass_reload): ...new function.
13587 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
13588 subclass of rtl_opt_pass along with...
13589 (pass_data_cleanup_barriers): ...new pass_data instance and...
13590 (make_pass_cleanup_barriers): ...new function.
13591 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
13592 of rtl_opt_pass along with...
13593 (pass_data_loop2): ...new pass_data instance and...
13594 (make_pass_loop2): ...new function.
13595 (pass_rtl_loop_init): Convert from a global struct to a subclass of
13596 rtl_opt_pass along with...
13597 (pass_data_rtl_loop_init): ...new pass_data instance and...
13598 (make_pass_rtl_loop_init): ...new function.
13599 (pass_rtl_loop_done): Convert from a global struct to a subclass of
13600 rtl_opt_pass along with...
13601 (pass_data_rtl_loop_done): ...new pass_data instance and...
13602 (make_pass_rtl_loop_done): ...new function.
13603 (pass_rtl_move_loop_invariants): Convert from a global struct to a
13604 subclass of rtl_opt_pass along with...
13605 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
13606 (make_pass_rtl_move_loop_invariants): ...new function.
13607 (pass_rtl_unswitch): Convert from a global struct to a subclass of
13608 rtl_opt_pass along with...
13609 (pass_data_rtl_unswitch): ...new pass_data instance and...
13610 (make_pass_rtl_unswitch): ...new function.
13611 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
13612 subclass of rtl_opt_pass along with...
13613 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
13615 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13616 (pass_rtl_doloop): Convert from a global struct to a subclass of
13617 rtl_opt_pass along with...
13618 (pass_data_rtl_doloop): ...new pass_data instance and...
13619 (make_pass_rtl_doloop): ...new function.
13620 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
13621 a subclass of rtl_opt_pass along with...
13622 (pass_data_lower_subreg): ...new pass_data instance and...
13623 (make_pass_lower_subreg): ...new function.
13624 (pass_lower_subreg2): Convert from a global struct to a subclass of
13625 rtl_opt_pass along with...
13626 (pass_data_lower_subreg2): ...new pass_data instance and...
13627 (make_pass_lower_subreg2): ...new function.
13628 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
13629 struct to a subclass of ipa_opt_pass_d along with...
13630 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
13631 (make_pass_ipa_lto_gimple_out): ...new function.
13632 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
13633 of ipa_opt_pass_d along with...
13634 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
13635 (make_pass_ipa_lto_finish_out): ...new function.
13636 * mode-switching.c (pass_mode_switching): Convert from a global struct
13637 to a subclass of rtl_opt_pass along with...
13638 (pass_data_mode_switching): ...new pass_data instance and...
13639 (make_pass_mode_switching): ...new function.
13640 * modulo-sched.c (pass_sms): Convert from a global struct to a
13641 subclass of rtl_opt_pass along with...
13642 (pass_data_sms): ...new pass_data instance and...
13643 (make_pass_sms): ...new function.
13644 * omp-low.c (pass_expand_omp): Convert from a global struct to a
13645 subclass of gimple_opt_pass along with...
13646 (pass_data_expand_omp): ...new pass_data instance and...
13647 (make_pass_expand_omp): ...new function.
13648 (pass_lower_omp): Convert from a global struct to a subclass of
13649 gimple_opt_pass along with...
13650 (pass_data_lower_omp): ...new pass_data instance and...
13651 (make_pass_lower_omp): ...new function.
13652 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
13653 of gimple_opt_pass along with...
13654 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
13655 (make_pass_diagnose_omp_blocks): ...new function.
13656 * passes.c (pass_early_local_passes): Convert from a global struct to
13657 a subclass of simple_ipa_opt_pass along with...
13658 (pass_data_early_local_passes): ...new pass_data instance and...
13659 (make_pass_early_local_passes): ...new function.
13660 (pass_all_early_optimizations): Convert from a global struct to a
13661 subclass of gimple_opt_pass along with...
13662 (pass_data_all_early_optimizations): ...new pass_data instance and...
13663 (make_pass_all_early_optimizations): ...new function.
13664 (pass_all_optimizations): Convert from a global struct to a subclass
13665 of gimple_opt_pass along with...
13666 (pass_data_all_optimizations): ...new pass_data instance and...
13667 (make_pass_all_optimizations): ...new function.
13668 (pass_all_optimizations_g): Convert from a global struct to a subclass
13669 of gimple_opt_pass along with...
13670 (pass_data_all_optimizations_g): ...new pass_data instance and...
13671 (make_pass_all_optimizations_g): ...new function.
13672 (pass_rest_of_compilation): Convert from a global struct to a subclass
13673 of rtl_opt_pass along with...
13674 (pass_data_rest_of_compilation): ...new pass_data instance and...
13675 (make_pass_rest_of_compilation): ...new function.
13676 (pass_postreload): Convert from a global struct to a subclass of
13677 rtl_opt_pass along with...
13678 (pass_data_postreload): ...new pass_data instance and...
13679 (make_pass_postreload): ...new function.
13680 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
13681 subclass of rtl_opt_pass along with...
13682 (pass_data_gcse2): ...new pass_data instance and...
13683 (make_pass_gcse2): ...new function.
13684 * postreload.c (pass_postreload_cse): Convert from a global struct to
13685 a subclass of rtl_opt_pass along with...
13686 (pass_data_postreload_cse): ...new pass_data instance and...
13687 (make_pass_postreload_cse): ...new function.
13688 * predict.c (pass_profile): Convert from a global struct to a subclass
13689 of gimple_opt_pass along with...
13690 (pass_data_profile): ...new pass_data instance and...
13691 (make_pass_profile): ...new function.
13692 (pass_strip_predict_hints): Convert from a global struct to a subclass
13693 of gimple_opt_pass along with...
13694 (pass_data_strip_predict_hints): ...new pass_data instance and...
13695 (make_pass_strip_predict_hints): ...new function.
13696 * recog.c (pass_peephole2): Convert from a global struct to a subclass
13697 of rtl_opt_pass along with...
13698 (pass_data_peephole2): ...new pass_data instance and...
13699 (make_pass_peephole2): ...new function.
13700 (pass_split_all_insns): Convert from a global struct to a subclass of
13701 rtl_opt_pass along with...
13702 (pass_data_split_all_insns): ...new pass_data instance and...
13703 (make_pass_split_all_insns): ...new function.
13704 (pass_split_after_reload): Convert from a global struct to a subclass
13705 of rtl_opt_pass along with...
13706 (pass_data_split_after_reload): ...new pass_data instance and...
13707 (make_pass_split_after_reload): ...new function.
13708 (pass_split_before_regstack): Convert from a global struct to a
13709 subclass of rtl_opt_pass along with...
13710 (pass_data_split_before_regstack): ...new pass_data instance and...
13711 (make_pass_split_before_regstack): ...new function.
13712 (pass_split_before_sched2): Convert from a global struct to a subclass
13713 of rtl_opt_pass along with...
13714 (pass_data_split_before_sched2): ...new pass_data instance and...
13715 (make_pass_split_before_sched2): ...new function.
13716 (pass_split_for_shorten_branches): Convert from a global struct to a
13717 subclass of rtl_opt_pass along with...
13718 (pass_data_split_for_shorten_branches): ...new pass_data instance
13720 (make_pass_split_for_shorten_branches): ...new function.
13721 * ree.c (pass_ree): Convert from a global struct to a subclass of
13722 rtl_opt_pass along with...
13723 (pass_data_ree): ...new pass_data instance and...
13724 (make_pass_ree): ...new function.
13725 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
13726 subclass of rtl_opt_pass along with...
13727 (pass_data_stack_regs): ...new pass_data instance and...
13728 (make_pass_stack_regs): ...new function.
13729 (pass_stack_regs_run): Convert from a global struct to a subclass of
13730 rtl_opt_pass along with...
13731 (pass_data_stack_regs_run): ...new pass_data instance and...
13732 (make_pass_stack_regs_run): ...new function.
13733 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
13734 subclass of rtl_opt_pass along with...
13735 (pass_data_cprop_hardreg): ...new pass_data instance and...
13736 (make_pass_cprop_hardreg): ...new function.
13737 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
13738 subclass of rtl_opt_pass along with...
13739 (pass_data_reginfo_init): ...new pass_data instance and...
13740 (make_pass_reginfo_init): ...new function.
13741 * regmove.c (pass_regmove): Convert from a global struct to a subclass
13742 of rtl_opt_pass along with...
13743 (pass_data_regmove): ...new pass_data instance and...
13744 (make_pass_regmove): ...new function.
13745 * regrename.c (pass_regrename): Convert from a global struct to a
13746 subclass of rtl_opt_pass along with...
13747 (pass_data_regrename): ...new pass_data instance and...
13748 (make_pass_regrename): ...new function.
13749 * reorg.c (pass_delay_slots): Convert from a global struct to a
13750 subclass of rtl_opt_pass along with...
13751 (pass_data_delay_slots): ...new pass_data instance and...
13752 (make_pass_delay_slots): ...new function.
13753 (pass_machine_reorg): Convert from a global struct to a subclass of
13754 rtl_opt_pass along with...
13755 (pass_data_machine_reorg): ...new pass_data instance and...
13756 (make_pass_machine_reorg): ...new function.
13757 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
13758 of rtl_opt_pass along with...
13759 (pass_data_sched): ...new pass_data instance and...
13760 (make_pass_sched): ...new function.
13761 (pass_sched2): Convert from a global struct to a subclass of
13762 rtl_opt_pass along with...
13763 (pass_data_sched2): ...new pass_data instance and...
13764 (make_pass_sched2): ...new function.
13765 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
13766 to a subclass of rtl_opt_pass along with...
13767 (pass_data_stack_ptr_mod): ...new pass_data instance and...
13768 (make_pass_stack_ptr_mod): ...new function.
13769 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
13770 to a subclass of rtl_opt_pass along with...
13771 (pass_data_rtl_store_motion): ...new pass_data instance and...
13772 (make_pass_rtl_store_motion): ...new function.
13773 * tracer.c (pass_tracer): Convert from a global struct to a subclass
13774 of gimple_opt_pass along with...
13775 (pass_data_tracer): ...new pass_data instance and...
13776 (make_pass_tracer): ...new function.
13777 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
13778 to a subclass of gimple_opt_pass along with...
13779 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
13780 (make_pass_diagnose_tm_blocks): ...new function.
13781 (pass_lower_tm): Convert from a global struct to a subclass of
13782 gimple_opt_pass along with...
13783 (pass_data_lower_tm): ...new pass_data instance and...
13784 (make_pass_lower_tm): ...new function.
13785 (pass_tm_init): Convert from a global struct to a subclass of
13786 gimple_opt_pass along with...
13787 (pass_data_tm_init): ...new pass_data instance and...
13788 (make_pass_tm_init): ...new function.
13789 (pass_tm_mark): Convert from a global struct to a subclass of
13790 gimple_opt_pass along with...
13791 (pass_data_tm_mark): ...new pass_data instance and...
13792 (make_pass_tm_mark): ...new function.
13793 (pass_tm_edges): Convert from a global struct to a subclass of
13794 gimple_opt_pass along with...
13795 (pass_data_tm_edges): ...new pass_data instance and...
13796 (make_pass_tm_edges): ...new function.
13797 (pass_tm_memopt): Convert from a global struct to a subclass of
13798 gimple_opt_pass along with...
13799 (pass_data_tm_memopt): ...new pass_data instance and...
13800 (make_pass_tm_memopt): ...new function.
13801 (pass_ipa_tm): Convert from a global struct to a subclass of
13802 simple_ipa_opt_pass along with...
13803 (pass_data_ipa_tm): ...new pass_data instance and...
13804 (make_pass_ipa_tm): ...new function.
13805 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
13806 subclass of gimple_opt_pass along with...
13807 (pass_data_call_cdce): ...new pass_data instance and...
13808 (make_pass_call_cdce): ...new function.
13809 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
13810 subclass of gimple_opt_pass along with...
13811 (pass_data_build_cfg): ...new pass_data instance and...
13812 (make_pass_build_cfg): ...new function.
13813 (pass_split_crit_edges): Convert from a global struct to a subclass of
13814 gimple_opt_pass along with...
13815 (pass_data_split_crit_edges): ...new pass_data instance and...
13816 (make_pass_split_crit_edges): ...new function.
13817 (pass_warn_function_return): Convert from a global struct to a
13818 subclass of gimple_opt_pass along with...
13819 (pass_data_warn_function_return): ...new pass_data instance and...
13820 (make_pass_warn_function_return): ...new function.
13821 (pass_warn_function_noreturn): Convert from a global struct to a
13822 subclass of gimple_opt_pass along with...
13823 (pass_data_warn_function_noreturn): ...new pass_data instance and...
13824 (make_pass_warn_function_noreturn): ...new function.
13825 (pass_warn_unused_result): Convert from a global struct to a subclass
13826 of gimple_opt_pass along with...
13827 (pass_data_warn_unused_result): ...new pass_data instance and...
13828 (make_pass_warn_unused_result): ...new function.
13829 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
13830 a subclass of gimple_opt_pass along with...
13831 (pass_data_merge_phi): ...new pass_data instance and...
13832 (make_pass_merge_phi): ...new function.
13833 * tree-complex.c (pass_lower_complex): Convert from a global struct to
13834 a subclass of gimple_opt_pass along with...
13835 (pass_data_lower_complex): ...new pass_data instance and...
13836 (make_pass_lower_complex): ...new function.
13837 (pass_lower_complex_O0): Convert from a global struct to a subclass of
13838 gimple_opt_pass along with...
13839 (pass_data_lower_complex_O0): ...new pass_data instance and...
13840 (make_pass_lower_complex_O0): ...new function.
13841 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
13842 subclass of gimple_opt_pass along with...
13843 (pass_data_lower_eh): ...new pass_data instance and...
13844 (make_pass_lower_eh): ...new function.
13845 (pass_refactor_eh): Convert from a global struct to a subclass of
13846 gimple_opt_pass along with...
13847 (pass_data_refactor_eh): ...new pass_data instance and...
13848 (make_pass_refactor_eh): ...new function.
13849 (pass_lower_resx): Convert from a global struct to a subclass of
13850 gimple_opt_pass along with...
13851 (pass_data_lower_resx): ...new pass_data instance and...
13852 (make_pass_lower_resx): ...new function.
13853 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
13854 of gimple_opt_pass along with...
13855 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
13856 (make_pass_lower_eh_dispatch): ...new function.
13857 (pass_cleanup_eh): Convert from a global struct to a subclass of
13858 gimple_opt_pass along with...
13859 (pass_data_cleanup_eh): ...new pass_data instance and...
13860 (make_pass_cleanup_eh): ...new function.
13861 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
13862 to a subclass of simple_ipa_opt_pass along with...
13863 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
13864 (make_pass_ipa_lower_emutls): ...new function.
13865 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
13866 a subclass of gimple_opt_pass along with...
13867 (pass_data_if_conversion): ...new pass_data instance and...
13868 (make_pass_if_conversion): ...new function.
13869 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
13870 subclass of gimple_opt_pass along with...
13871 (pass_data_build_ssa): ...new pass_data instance and...
13872 (make_pass_build_ssa): ...new function.
13873 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
13874 global struct to a subclass of gimple_opt_pass along with...
13875 (pass_data_loop_distribution): ...new pass_data instance and...
13876 (make_pass_loop_distribution): ...new function.
13877 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
13878 subclass of gimple_opt_pass along with...
13879 (pass_data_mudflap_1): ...new pass_data instance and...
13880 (make_pass_mudflap_1): ...new function.
13881 (pass_mudflap_2): Convert from a global struct to a subclass of
13882 gimple_opt_pass along with...
13883 (pass_data_mudflap_2): ...new pass_data instance and...
13884 (make_pass_mudflap_2): ...new function.
13885 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
13886 subclass of gimple_opt_pass along with...
13887 (pass_data_mudflap_1): ...new pass_data instance and...
13888 (make_pass_mudflap_1): ...new function.
13889 (pass_mudflap_2): Convert from a global struct to a subclass of
13890 gimple_opt_pass along with...
13891 (pass_data_mudflap_2): ...new pass_data instance and...
13892 (make_pass_mudflap_2): ...new function.
13893 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
13894 gimple_opt_pass along with...
13895 (pass_data_nrv): ...new pass_data instance and...
13896 (make_pass_nrv): ...new function.
13897 (pass_return_slot): Convert from a global struct to a subclass of
13898 gimple_opt_pass along with...
13899 (pass_data_return_slot): ...new pass_data instance and...
13900 (make_pass_return_slot): ...new function.
13901 * tree-object-size.c (pass_object_sizes): Convert from a global struct
13902 to a subclass of gimple_opt_pass along with...
13903 (pass_data_object_sizes): ...new pass_data instance and...
13904 (make_pass_object_sizes): ...new function.
13905 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
13906 global struct to a subclass of gimple_opt_pass along with...
13907 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
13909 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13910 (pass_fixup_cfg): Convert from a global struct to a subclass of
13911 gimple_opt_pass along with...
13912 (pass_data_fixup_cfg): ...new pass_data instance and...
13913 (make_pass_fixup_cfg): ...new function.
13914 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
13915 (make_pass_mudflap_1): ...new function.
13916 (pass_mudflap_2): Replace declaration with that of...
13917 (make_pass_mudflap_2): ...new function.
13918 (pass_asan): Replace declaration with that of...
13919 (make_pass_asan): ...new function.
13920 (pass_asan_O0): Replace declaration with that of...
13921 (make_pass_asan_O0): ...new function.
13922 (pass_tsan): Replace declaration with that of...
13923 (make_pass_tsan): ...new function.
13924 (pass_tsan_O0): Replace declaration with that of...
13925 (make_pass_tsan_O0): ...new function.
13926 (pass_lower_cf): Replace declaration with that of...
13927 (make_pass_lower_cf): ...new function.
13928 (pass_refactor_eh): Replace declaration with that of...
13929 (make_pass_refactor_eh): ...new function.
13930 (pass_lower_eh): Replace declaration with that of...
13931 (make_pass_lower_eh): ...new function.
13932 (pass_lower_eh_dispatch): Replace declaration with that of...
13933 (make_pass_lower_eh_dispatch): ...new function.
13934 (pass_lower_resx): Replace declaration with that of...
13935 (make_pass_lower_resx): ...new function.
13936 (pass_build_cfg): Replace declaration with that of...
13937 (make_pass_build_cfg): ...new function.
13938 (pass_early_tree_profile): Replace declaration with that of...
13939 (make_pass_early_tree_profile): ...new function.
13940 (pass_cleanup_eh): Replace declaration with that of...
13941 (make_pass_cleanup_eh): ...new function.
13942 (pass_sra): Replace declaration with that of...
13943 (make_pass_sra): ...new function.
13944 (pass_sra_early): Replace declaration with that of...
13945 (make_pass_sra_early): ...new function.
13946 (pass_early_ipa_sra): Replace declaration with that of...
13947 (make_pass_early_ipa_sra): ...new function.
13948 (pass_tail_recursion): Replace declaration with that of...
13949 (make_pass_tail_recursion): ...new function.
13950 (pass_tail_calls): Replace declaration with that of...
13951 (make_pass_tail_calls): ...new function.
13952 (pass_tree_loop): Replace declaration with that of...
13953 (make_pass_tree_loop): ...new function.
13954 (pass_tree_loop_init): Replace declaration with that of...
13955 (make_pass_tree_loop_init): ...new function.
13956 (pass_lim): Replace declaration with that of...
13957 (make_pass_lim): ...new function.
13958 (pass_tree_unswitch): Replace declaration with that of...
13959 (make_pass_tree_unswitch): ...new function.
13960 (pass_predcom): Replace declaration with that of...
13961 (make_pass_predcom): ...new function.
13962 (pass_iv_canon): Replace declaration with that of...
13963 (make_pass_iv_canon): ...new function.
13964 (pass_scev_cprop): Replace declaration with that of...
13965 (make_pass_scev_cprop): ...new function.
13966 (pass_empty_loop): Replace declaration with that of...
13967 (make_pass_empty_loop): ...new function.
13968 (pass_record_bounds): Replace declaration with that of...
13969 (make_pass_record_bounds): ...new function.
13970 (pass_graphite): Replace declaration with that of...
13971 (make_pass_graphite): ...new function.
13972 (pass_graphite_transforms): Replace declaration with that of...
13973 (make_pass_graphite_transforms): ...new function.
13974 (pass_if_conversion): Replace declaration with that of...
13975 (make_pass_if_conversion): ...new function.
13976 (pass_loop_distribution): Replace declaration with that of...
13977 (make_pass_loop_distribution): ...new function.
13978 (pass_vectorize): Replace declaration with that of...
13979 (make_pass_vectorize): ...new function.
13980 (pass_slp_vectorize): Replace declaration with that of...
13981 (make_pass_slp_vectorize): ...new function.
13982 (pass_complete_unroll): Replace declaration with that of...
13983 (make_pass_complete_unroll): ...new function.
13984 (pass_complete_unrolli): Replace declaration with that of...
13985 (make_pass_complete_unrolli): ...new function.
13986 (pass_parallelize_loops): Replace declaration with that of...
13987 (make_pass_parallelize_loops): ...new function.
13988 (pass_loop_prefetch): Replace declaration with that of...
13989 (make_pass_loop_prefetch): ...new function.
13990 (pass_iv_optimize): Replace declaration with that of...
13991 (make_pass_iv_optimize): ...new function.
13992 (pass_tree_loop_done): Replace declaration with that of...
13993 (make_pass_tree_loop_done): ...new function.
13994 (pass_ch): Replace declaration with that of...
13995 (make_pass_ch): ...new function.
13996 (pass_ccp): Replace declaration with that of...
13997 (make_pass_ccp): ...new function.
13998 (pass_phi_only_cprop): Replace declaration with that of...
13999 (make_pass_phi_only_cprop): ...new function.
14000 (pass_build_ssa): Replace declaration with that of...
14001 (make_pass_build_ssa): ...new function.
14002 (pass_build_alias): Replace declaration with that of...
14003 (make_pass_build_alias): ...new function.
14004 (pass_build_ealias): Replace declaration with that of...
14005 (make_pass_build_ealias): ...new function.
14006 (pass_dominator): Replace declaration with that of...
14007 (make_pass_dominator): ...new function.
14008 (pass_dce): Replace declaration with that of...
14009 (make_pass_dce): ...new function.
14010 (pass_dce_loop): Replace declaration with that of...
14011 (make_pass_dce_loop): ...new function.
14012 (pass_cd_dce): Replace declaration with that of...
14013 (make_pass_cd_dce): ...new function.
14014 (pass_call_cdce): Replace declaration with that of...
14015 (make_pass_call_cdce): ...new function.
14016 (pass_merge_phi): Replace declaration with that of...
14017 (make_pass_merge_phi): ...new function.
14018 (pass_split_crit_edges): Replace declaration with that of...
14019 (make_pass_split_crit_edges): ...new function.
14020 (pass_pre): Replace declaration with that of...
14021 (make_pass_pre): ...new function.
14022 (pass_profile): Replace declaration with that of...
14023 (make_pass_profile): ...new function.
14024 (pass_strip_predict_hints): Replace declaration with that of...
14025 (make_pass_strip_predict_hints): ...new function.
14026 (pass_lower_complex_O0): Replace declaration with that of...
14027 (make_pass_lower_complex_O0): ...new function.
14028 (pass_lower_complex): Replace declaration with that of...
14029 (make_pass_lower_complex): ...new function.
14030 (pass_lower_vector): Replace declaration with that of...
14031 (make_pass_lower_vector): ...new function.
14032 (pass_lower_vector_ssa): Replace declaration with that of...
14033 (make_pass_lower_vector_ssa): ...new function.
14034 (pass_lower_omp): Replace declaration with that of...
14035 (make_pass_lower_omp): ...new function.
14036 (pass_diagnose_omp_blocks): Replace declaration with that of...
14037 (make_pass_diagnose_omp_blocks): ...new function.
14038 (pass_expand_omp): Replace declaration with that of...
14039 (make_pass_expand_omp): ...new function.
14040 (pass_expand_omp_ssa): Replace declaration with that of...
14041 (make_pass_expand_omp_ssa): ...new function.
14042 (pass_object_sizes): Replace declaration with that of...
14043 (make_pass_object_sizes): ...new function.
14044 (pass_strlen): Replace declaration with that of...
14045 (make_pass_strlen): ...new function.
14046 (pass_fold_builtins): Replace declaration with that of...
14047 (make_pass_fold_builtins): ...new function.
14048 (pass_stdarg): Replace declaration with that of...
14049 (make_pass_stdarg): ...new function.
14050 (pass_early_warn_uninitialized): Replace declaration with that of...
14051 (make_pass_early_warn_uninitialized): ...new function.
14052 (pass_late_warn_uninitialized): Replace declaration with that of...
14053 (make_pass_late_warn_uninitialized): ...new function.
14054 (pass_cse_reciprocals): Replace declaration with that of...
14055 (make_pass_cse_reciprocals): ...new function.
14056 (pass_cse_sincos): Replace declaration with that of...
14057 (make_pass_cse_sincos): ...new function.
14058 (pass_optimize_bswap): Replace declaration with that of...
14059 (make_pass_optimize_bswap): ...new function.
14060 (pass_optimize_widening_mul): Replace declaration with that of...
14061 (make_pass_optimize_widening_mul): ...new function.
14062 (pass_warn_function_return): Replace declaration with that of...
14063 (make_pass_warn_function_return): ...new function.
14064 (pass_warn_function_noreturn): Replace declaration with that of...
14065 (make_pass_warn_function_noreturn): ...new function.
14066 (pass_cselim): Replace declaration with that of...
14067 (make_pass_cselim): ...new function.
14068 (pass_phiopt): Replace declaration with that of...
14069 (make_pass_phiopt): ...new function.
14070 (pass_forwprop): Replace declaration with that of...
14071 (make_pass_forwprop): ...new function.
14072 (pass_phiprop): Replace declaration with that of...
14073 (make_pass_phiprop): ...new function.
14074 (pass_tree_ifcombine): Replace declaration with that of...
14075 (make_pass_tree_ifcombine): ...new function.
14076 (pass_dse): Replace declaration with that of...
14077 (make_pass_dse): ...new function.
14078 (pass_nrv): Replace declaration with that of...
14079 (make_pass_nrv): ...new function.
14080 (pass_rename_ssa_copies): Replace declaration with that of...
14081 (make_pass_rename_ssa_copies): ...new function.
14082 (pass_sink_code): Replace declaration with that of...
14083 (make_pass_sink_code): ...new function.
14084 (pass_fre): Replace declaration with that of...
14085 (make_pass_fre): ...new function.
14086 (pass_check_data_deps): Replace declaration with that of...
14087 (make_pass_check_data_deps): ...new function.
14088 (pass_copy_prop): Replace declaration with that of...
14089 (make_pass_copy_prop): ...new function.
14090 (pass_vrp): Replace declaration with that of...
14091 (make_pass_vrp): ...new function.
14092 (pass_uncprop): Replace declaration with that of...
14093 (make_pass_uncprop): ...new function.
14094 (pass_return_slot): Replace declaration with that of...
14095 (make_pass_return_slot): ...new function.
14096 (pass_reassoc): Replace declaration with that of...
14097 (make_pass_reassoc): ...new function.
14098 (pass_rebuild_cgraph_edges): Replace declaration with that of...
14099 (make_pass_rebuild_cgraph_edges): ...new function.
14100 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
14101 (make_pass_remove_cgraph_callee_edges): ...new function.
14102 (pass_build_cgraph_edges): Replace declaration with that of...
14103 (make_pass_build_cgraph_edges): ...new function.
14104 (pass_local_pure_const): Replace declaration with that of...
14105 (make_pass_local_pure_const): ...new function.
14106 (pass_tracer): Replace declaration with that of...
14107 (make_pass_tracer): ...new function.
14108 (pass_warn_unused_result): Replace declaration with that of...
14109 (make_pass_warn_unused_result): ...new function.
14110 (pass_diagnose_tm_blocks): Replace declaration with that of...
14111 (make_pass_diagnose_tm_blocks): ...new function.
14112 (pass_lower_tm): Replace declaration with that of...
14113 (make_pass_lower_tm): ...new function.
14114 (pass_tm_init): Replace declaration with that of...
14115 (make_pass_tm_init): ...new function.
14116 (pass_tm_mark): Replace declaration with that of...
14117 (make_pass_tm_mark): ...new function.
14118 (pass_tm_memopt): Replace declaration with that of...
14119 (make_pass_tm_memopt): ...new function.
14120 (pass_tm_edges): Replace declaration with that of...
14121 (make_pass_tm_edges): ...new function.
14122 (pass_split_functions): Replace declaration with that of...
14123 (make_pass_split_functions): ...new function.
14124 (pass_feedback_split_functions): Replace declaration with that of...
14125 (make_pass_feedback_split_functions): ...new function.
14126 (pass_strength_reduction): Replace declaration with that of...
14127 (make_pass_strength_reduction): ...new function.
14128 (pass_ipa_lower_emutls): Replace declaration with that of...
14129 (make_pass_ipa_lower_emutls): ...new function.
14130 (pass_ipa_function_and_variable_visibility): Replace declaration with
14132 (make_pass_ipa_function_and_variable_visibility): ...new function.
14133 (pass_ipa_tree_profile): Replace declaration with that of...
14134 (make_pass_ipa_tree_profile): ...new function.
14135 (pass_early_local_passes): Replace declaration with that of...
14136 (make_pass_early_local_passes): ...new function.
14137 (pass_ipa_whole_program_visibility): Replace declaration with that
14139 (make_pass_ipa_whole_program_visibility): ...new function.
14140 (pass_ipa_lto_gimple_out): Replace declaration with that of...
14141 (make_pass_ipa_lto_gimple_out): ...new function.
14142 (pass_ipa_increase_alignment): Replace declaration with that of...
14143 (make_pass_ipa_increase_alignment): ...new function.
14144 (pass_ipa_inline): Replace declaration with that of...
14145 (make_pass_ipa_inline): ...new function.
14146 (pass_ipa_free_lang_data): Replace declaration with that of...
14147 (make_pass_ipa_free_lang_data): ...new function.
14148 (pass_ipa_free_inline_summary): Replace declaration with that of...
14149 (make_pass_ipa_free_inline_summary): ...new function.
14150 (pass_ipa_cp): Replace declaration with that of...
14151 (make_pass_ipa_cp): ...new function.
14152 (pass_ipa_reference): Replace declaration with that of...
14153 (make_pass_ipa_reference): ...new function.
14154 (pass_ipa_pure_const): Replace declaration with that of...
14155 (make_pass_ipa_pure_const): ...new function.
14156 (pass_ipa_pta): Replace declaration with that of...
14157 (make_pass_ipa_pta): ...new function.
14158 (pass_ipa_lto_finish_out): Replace declaration with that of...
14159 (make_pass_ipa_lto_finish_out): ...new function.
14160 (pass_ipa_tm): Replace declaration with that of...
14161 (make_pass_ipa_tm): ...new function.
14162 (pass_ipa_profile): Replace declaration with that of...
14163 (make_pass_ipa_profile): ...new function.
14164 (pass_ipa_cdtor_merge): Replace declaration with that of...
14165 (make_pass_ipa_cdtor_merge): ...new function.
14166 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
14168 (make_pass_cleanup_cfg_post_optimizing): ...new function.
14169 (pass_init_datastructures): Replace declaration with that of...
14170 (make_pass_init_datastructures): ...new function.
14171 (pass_fixup_cfg): Replace declaration with that of...
14172 (make_pass_fixup_cfg): ...new function.
14173 (pass_expand): Replace declaration with that of...
14174 (make_pass_expand): ...new function.
14175 (pass_instantiate_virtual_regs): Replace declaration with that of...
14176 (make_pass_instantiate_virtual_regs): ...new function.
14177 (pass_rtl_fwprop): Replace declaration with that of...
14178 (make_pass_rtl_fwprop): ...new function.
14179 (pass_rtl_fwprop_addr): Replace declaration with that of...
14180 (make_pass_rtl_fwprop_addr): ...new function.
14181 (pass_jump): Replace declaration with that of...
14182 (make_pass_jump): ...new function.
14183 (pass_jump2): Replace declaration with that of...
14184 (make_pass_jump2): ...new function.
14185 (pass_lower_subreg): Replace declaration with that of...
14186 (make_pass_lower_subreg): ...new function.
14187 (pass_cse): Replace declaration with that of...
14188 (make_pass_cse): ...new function.
14189 (pass_fast_rtl_dce): Replace declaration with that of...
14190 (make_pass_fast_rtl_dce): ...new function.
14191 (pass_ud_rtl_dce): Replace declaration with that of...
14192 (make_pass_ud_rtl_dce): ...new function.
14193 (pass_rtl_dce): Replace declaration with that of...
14194 (make_pass_rtl_dce): ...new function.
14195 (pass_rtl_dse1): Replace declaration with that of...
14196 (make_pass_rtl_dse1): ...new function.
14197 (pass_rtl_dse2): Replace declaration with that of...
14198 (make_pass_rtl_dse2): ...new function.
14199 (pass_rtl_dse3): Replace declaration with that of...
14200 (make_pass_rtl_dse3): ...new function.
14201 (pass_rtl_cprop): Replace declaration with that of...
14202 (make_pass_rtl_cprop): ...new function.
14203 (pass_rtl_pre): Replace declaration with that of...
14204 (make_pass_rtl_pre): ...new function.
14205 (pass_rtl_hoist): Replace declaration with that of...
14206 (make_pass_rtl_hoist): ...new function.
14207 (pass_rtl_store_motion): Replace declaration with that of...
14208 (make_pass_rtl_store_motion): ...new function.
14209 (pass_cse_after_global_opts): Replace declaration with that of...
14210 (make_pass_cse_after_global_opts): ...new function.
14211 (pass_rtl_ifcvt): Replace declaration with that of...
14212 (make_pass_rtl_ifcvt): ...new function.
14213 (pass_into_cfg_layout_mode): Replace declaration with that of...
14214 (make_pass_into_cfg_layout_mode): ...new function.
14215 (pass_outof_cfg_layout_mode): Replace declaration with that of...
14216 (make_pass_outof_cfg_layout_mode): ...new function.
14217 (pass_loop2): Replace declaration with that of...
14218 (make_pass_loop2): ...new function.
14219 (pass_rtl_loop_init): Replace declaration with that of...
14220 (make_pass_rtl_loop_init): ...new function.
14221 (pass_rtl_move_loop_invariants): Replace declaration with that of...
14222 (make_pass_rtl_move_loop_invariants): ...new function.
14223 (pass_rtl_unswitch): Replace declaration with that of...
14224 (make_pass_rtl_unswitch): ...new function.
14225 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
14226 (make_pass_rtl_unroll_and_peel_loops): ...new function.
14227 (pass_rtl_doloop): Replace declaration with that of...
14228 (make_pass_rtl_doloop): ...new function.
14229 (pass_rtl_loop_done): Replace declaration with that of...
14230 (make_pass_rtl_loop_done): ...new function.
14231 (pass_web): Replace declaration with that of...
14232 (make_pass_web): ...new function.
14233 (pass_cse2): Replace declaration with that of...
14234 (make_pass_cse2): ...new function.
14235 (pass_df_initialize_opt): Replace declaration with that of...
14236 (make_pass_df_initialize_opt): ...new function.
14237 (pass_df_initialize_no_opt): Replace declaration with that of...
14238 (make_pass_df_initialize_no_opt): ...new function.
14239 (pass_reginfo_init): Replace declaration with that of...
14240 (make_pass_reginfo_init): ...new function.
14241 (pass_inc_dec): Replace declaration with that of...
14242 (make_pass_inc_dec): ...new function.
14243 (pass_stack_ptr_mod): Replace declaration with that of...
14244 (make_pass_stack_ptr_mod): ...new function.
14245 (pass_initialize_regs): Replace declaration with that of...
14246 (make_pass_initialize_regs): ...new function.
14247 (pass_combine): Replace declaration with that of...
14248 (make_pass_combine): ...new function.
14249 (pass_if_after_combine): Replace declaration with that of...
14250 (make_pass_if_after_combine): ...new function.
14251 (pass_ree): Replace declaration with that of...
14252 (make_pass_ree): ...new function.
14253 (pass_partition_blocks): Replace declaration with that of...
14254 (make_pass_partition_blocks): ...new function.
14255 (pass_match_asm_constraints): Replace declaration with that of...
14256 (make_pass_match_asm_constraints): ...new function.
14257 (pass_regmove): Replace declaration with that of...
14258 (make_pass_regmove): ...new function.
14259 (pass_split_all_insns): Replace declaration with that of...
14260 (make_pass_split_all_insns): ...new function.
14261 (pass_fast_rtl_byte_dce): Replace declaration with that of...
14262 (make_pass_fast_rtl_byte_dce): ...new function.
14263 (pass_lower_subreg2): Replace declaration with that of...
14264 (make_pass_lower_subreg2): ...new function.
14265 (pass_mode_switching): Replace declaration with that of...
14266 (make_pass_mode_switching): ...new function.
14267 (pass_sms): Replace declaration with that of...
14268 (make_pass_sms): ...new function.
14269 (pass_sched): Replace declaration with that of...
14270 (make_pass_sched): ...new function.
14271 (pass_ira): Replace declaration with that of...
14272 (make_pass_ira): ...new function.
14273 (pass_reload): Replace declaration with that of...
14274 (make_pass_reload): ...new function.
14275 (pass_clean_state): Replace declaration with that of...
14276 (make_pass_clean_state): ...new function.
14277 (pass_branch_prob): Replace declaration with that of...
14278 (make_pass_branch_prob): ...new function.
14279 (pass_value_profile_transformations): Replace declaration with that
14281 (make_pass_value_profile_transformations): ...new function.
14282 (pass_postreload_cse): Replace declaration with that of...
14283 (make_pass_postreload_cse): ...new function.
14284 (pass_gcse2): Replace declaration with that of...
14285 (make_pass_gcse2): ...new function.
14286 (pass_split_after_reload): Replace declaration with that of...
14287 (make_pass_split_after_reload): ...new function.
14288 (pass_branch_target_load_optimize1): Replace declaration with that
14290 (make_pass_branch_target_load_optimize1): ...new function.
14291 (pass_thread_prologue_and_epilogue): Replace declaration with that
14293 (make_pass_thread_prologue_and_epilogue): ...new function.
14294 (pass_stack_adjustments): Replace declaration with that of...
14295 (make_pass_stack_adjustments): ...new function.
14296 (pass_peephole2): Replace declaration with that of...
14297 (make_pass_peephole2): ...new function.
14298 (pass_if_after_reload): Replace declaration with that of...
14299 (make_pass_if_after_reload): ...new function.
14300 (pass_regrename): Replace declaration with that of...
14301 (make_pass_regrename): ...new function.
14302 (pass_cprop_hardreg): Replace declaration with that of...
14303 (make_pass_cprop_hardreg): ...new function.
14304 (pass_reorder_blocks): Replace declaration with that of...
14305 (make_pass_reorder_blocks): ...new function.
14306 (pass_branch_target_load_optimize2): Replace declaration with that
14308 (make_pass_branch_target_load_optimize2): ...new function.
14309 (pass_leaf_regs): Replace declaration with that of...
14310 (make_pass_leaf_regs): ...new function.
14311 (pass_split_before_sched2): Replace declaration with that of...
14312 (make_pass_split_before_sched2): ...new function.
14313 (pass_compare_elim_after_reload): Replace declaration with that of...
14314 (make_pass_compare_elim_after_reload): ...new function.
14315 (pass_sched2): Replace declaration with that of...
14316 (make_pass_sched2): ...new function.
14317 (pass_stack_regs): Replace declaration with that of...
14318 (make_pass_stack_regs): ...new function.
14319 (pass_stack_regs_run): Replace declaration with that of...
14320 (make_pass_stack_regs_run): ...new function.
14321 (pass_df_finish): Replace declaration with that of...
14322 (make_pass_df_finish): ...new function.
14323 (pass_compute_alignments): Replace declaration with that of...
14324 (make_pass_compute_alignments): ...new function.
14325 (pass_duplicate_computed_gotos): Replace declaration with that of...
14326 (make_pass_duplicate_computed_gotos): ...new function.
14327 (pass_variable_tracking): Replace declaration with that of...
14328 (make_pass_variable_tracking): ...new function.
14329 (pass_free_cfg): Replace declaration with that of...
14330 (make_pass_free_cfg): ...new function.
14331 (pass_machine_reorg): Replace declaration with that of...
14332 (make_pass_machine_reorg): ...new function.
14333 (pass_cleanup_barriers): Replace declaration with that of...
14334 (make_pass_cleanup_barriers): ...new function.
14335 (pass_delay_slots): Replace declaration with that of...
14336 (make_pass_delay_slots): ...new function.
14337 (pass_split_for_shorten_branches): Replace declaration with that of...
14338 (make_pass_split_for_shorten_branches): ...new function.
14339 (pass_split_before_regstack): Replace declaration with that of...
14340 (make_pass_split_before_regstack): ...new function.
14341 (pass_convert_to_eh_region_ranges): Replace declaration with that
14343 (make_pass_convert_to_eh_region_ranges): ...new function.
14344 (pass_shorten_branches): Replace declaration with that of...
14345 (make_pass_shorten_branches): ...new function.
14346 (pass_set_nothrow_function_flags): Replace declaration with that of...
14347 (make_pass_set_nothrow_function_flags): ...new function.
14348 (pass_dwarf2_frame): Replace declaration with that of...
14349 (make_pass_dwarf2_frame): ...new function.
14350 (pass_final): Replace declaration with that of...
14351 (make_pass_final): ...new function.
14352 (pass_rtl_seqabstr): Replace declaration with that of...
14353 (make_pass_rtl_seqabstr): ...new function.
14354 (pass_release_ssa_names): Replace declaration with that of...
14355 (make_pass_release_ssa_names): ...new function.
14356 (pass_early_inline): Replace declaration with that of...
14357 (make_pass_early_inline): ...new function.
14358 (pass_inline_parameters): Replace declaration with that of...
14359 (make_pass_inline_parameters): ...new function.
14360 (pass_update_address_taken): Replace declaration with that of...
14361 (make_pass_update_address_taken): ...new function.
14362 (pass_convert_switch): Replace declaration with that of...
14363 (make_pass_convert_switch): ...new function.
14364 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
14365 to a subclass of simple_ipa_opt_pass along with...
14366 (pass_data_ipa_tree_profile): ...new pass_data instance and...
14367 (make_pass_ipa_tree_profile): ...new function.
14368 * tree-sra.c (pass_sra_early): Convert from a global struct to a
14369 subclass of gimple_opt_pass along with...
14370 (pass_data_sra_early): ...new pass_data instance and...
14371 (make_pass_sra_early): ...new function.
14372 (pass_sra): Convert from a global struct to a subclass of
14373 gimple_opt_pass along with...
14374 (pass_data_sra): ...new pass_data instance and...
14375 (make_pass_sra): ...new function.
14376 (pass_early_ipa_sra): Convert from a global struct to a subclass of
14377 gimple_opt_pass along with...
14378 (pass_data_early_ipa_sra): ...new pass_data instance and...
14379 (make_pass_early_ipa_sra): ...new function.
14380 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
14381 subclass of gimple_opt_pass along with...
14382 (pass_data_ccp): ...new pass_data instance and...
14383 (make_pass_ccp): ...new function.
14384 (pass_fold_builtins): Convert from a global struct to a subclass of
14385 gimple_opt_pass along with...
14386 (pass_data_fold_builtins): ...new pass_data instance and...
14387 (make_pass_fold_builtins): ...new function.
14388 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
14389 subclass of gimple_opt_pass along with...
14390 (pass_data_copy_prop): ...new pass_data instance and...
14391 (make_pass_copy_prop): ...new function.
14392 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
14393 global struct to a subclass of gimple_opt_pass along with...
14394 (pass_data_rename_ssa_copies): ...new pass_data instance and...
14395 (make_pass_rename_ssa_copies): ...new function.
14396 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
14397 subclass of gimple_opt_pass along with...
14398 (pass_data_dce): ...new pass_data instance and...
14399 (make_pass_dce): ...new function.
14400 (pass_dce_loop): Convert from a global struct to a subclass of
14401 gimple_opt_pass along with...
14402 (pass_data_dce_loop): ...new pass_data instance and...
14403 (make_pass_dce_loop): ...new function.
14404 (pass_cd_dce): Convert from a global struct to a subclass of
14405 gimple_opt_pass along with...
14406 (pass_data_cd_dce): ...new pass_data instance and...
14407 (make_pass_cd_dce): ...new function.
14408 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
14409 subclass of gimple_opt_pass along with...
14410 (pass_data_dominator): ...new pass_data instance and...
14411 (make_pass_dominator): ...new function.
14412 (pass_phi_only_cprop): Convert from a global struct to a subclass of
14413 gimple_opt_pass along with...
14414 (pass_data_phi_only_cprop): ...new pass_data instance and...
14415 (make_pass_phi_only_cprop): ...new function.
14416 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
14417 subclass of gimple_opt_pass along with...
14418 (pass_data_dse): ...new pass_data instance and...
14419 (make_pass_dse): ...new function.
14420 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
14421 a subclass of gimple_opt_pass along with...
14422 (pass_data_forwprop): ...new pass_data instance and...
14423 (make_pass_forwprop): ...new function.
14424 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
14425 struct to a subclass of gimple_opt_pass along with...
14426 (pass_data_tree_ifcombine): ...new pass_data instance and...
14427 (make_pass_tree_ifcombine): ...new function.
14428 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
14429 subclass of gimple_opt_pass along with...
14430 (pass_data_ch): ...new pass_data instance and...
14431 (make_pass_ch): ...new function.
14432 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
14433 subclass of gimple_opt_pass along with...
14434 (pass_data_tree_loop): ...new pass_data instance and...
14435 (make_pass_tree_loop): ...new function.
14436 (pass_tree_loop_init): Convert from a global struct to a subclass of
14437 gimple_opt_pass along with...
14438 (pass_data_tree_loop_init): ...new pass_data instance and...
14439 (make_pass_tree_loop_init): ...new function.
14440 (pass_lim): Convert from a global struct to a subclass of
14441 gimple_opt_pass along with...
14442 (pass_data_lim): ...new pass_data instance and...
14443 (make_pass_lim): ...new function.
14444 (pass_tree_unswitch): Convert from a global struct to a subclass of
14445 gimple_opt_pass along with...
14446 (pass_data_tree_unswitch): ...new pass_data instance and...
14447 (make_pass_tree_unswitch): ...new function.
14448 (pass_predcom): Convert from a global struct to a subclass of
14449 gimple_opt_pass along with...
14450 (pass_data_predcom): ...new pass_data instance and...
14451 (make_pass_predcom): ...new function.
14452 (pass_vectorize): Convert from a global struct to a subclass of
14453 gimple_opt_pass along with...
14454 (pass_data_vectorize): ...new pass_data instance and...
14455 (make_pass_vectorize): ...new function.
14456 (pass_graphite): Convert from a global struct to a subclass of
14457 gimple_opt_pass along with...
14458 (pass_data_graphite): ...new pass_data instance and...
14459 (make_pass_graphite): ...new function.
14460 (pass_graphite_transforms): Convert from a global struct to a subclass
14461 of gimple_opt_pass along with...
14462 (pass_data_graphite_transforms): ...new pass_data instance and...
14463 (make_pass_graphite_transforms): ...new function.
14464 (pass_check_data_deps): Convert from a global struct to a subclass of
14465 gimple_opt_pass along with...
14466 (pass_data_check_data_deps): ...new pass_data instance and...
14467 (make_pass_check_data_deps): ...new function.
14468 (pass_iv_canon): Convert from a global struct to a subclass of
14469 gimple_opt_pass along with...
14470 (pass_data_iv_canon): ...new pass_data instance and...
14471 (make_pass_iv_canon): ...new function.
14472 (pass_scev_cprop): Convert from a global struct to a subclass of
14473 gimple_opt_pass along with...
14474 (pass_data_scev_cprop): ...new pass_data instance and...
14475 (make_pass_scev_cprop): ...new function.
14476 (pass_record_bounds): Convert from a global struct to a subclass of
14477 gimple_opt_pass along with...
14478 (pass_data_record_bounds): ...new pass_data instance and...
14479 (make_pass_record_bounds): ...new function.
14480 (pass_complete_unroll): Convert from a global struct to a subclass of
14481 gimple_opt_pass along with...
14482 (pass_data_complete_unroll): ...new pass_data instance and...
14483 (make_pass_complete_unroll): ...new function.
14484 (pass_complete_unrolli): Convert from a global struct to a subclass of
14485 gimple_opt_pass along with...
14486 (pass_data_complete_unrolli): ...new pass_data instance and...
14487 (make_pass_complete_unrolli): ...new function.
14488 (pass_parallelize_loops): Convert from a global struct to a subclass
14489 of gimple_opt_pass along with...
14490 (pass_data_parallelize_loops): ...new pass_data instance and...
14491 (make_pass_parallelize_loops): ...new function.
14492 (pass_loop_prefetch): Convert from a global struct to a subclass of
14493 gimple_opt_pass along with...
14494 (pass_data_loop_prefetch): ...new pass_data instance and...
14495 (make_pass_loop_prefetch): ...new function.
14496 (pass_iv_optimize): Convert from a global struct to a subclass of
14497 gimple_opt_pass along with...
14498 (pass_data_iv_optimize): ...new pass_data instance and...
14499 (make_pass_iv_optimize): ...new function.
14500 (pass_tree_loop_done): Convert from a global struct to a subclass of
14501 gimple_opt_pass along with...
14502 (pass_data_tree_loop_done): ...new pass_data instance and...
14503 (make_pass_tree_loop_done): ...new function.
14504 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
14505 struct to a subclass of gimple_opt_pass along with...
14506 (pass_data_cse_reciprocals): ...new pass_data instance and...
14507 (make_pass_cse_reciprocals): ...new function.
14508 (pass_cse_sincos): Convert from a global struct to a subclass of
14509 gimple_opt_pass along with...
14510 (pass_data_cse_sincos): ...new pass_data instance and...
14511 (make_pass_cse_sincos): ...new function.
14512 (pass_optimize_bswap): Convert from a global struct to a subclass of
14513 gimple_opt_pass along with...
14514 (pass_data_optimize_bswap): ...new pass_data instance and...
14515 (make_pass_optimize_bswap): ...new function.
14516 (pass_optimize_widening_mul): Convert from a global struct to a
14517 subclass of gimple_opt_pass along with...
14518 (pass_data_optimize_widening_mul): ...new pass_data instance and...
14519 (make_pass_optimize_widening_mul): ...new function.
14520 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
14521 subclass of gimple_opt_pass along with...
14522 (pass_data_phiopt): ...new pass_data instance and...
14523 (make_pass_phiopt): ...new function.
14524 (pass_cselim): Convert from a global struct to a subclass of
14525 gimple_opt_pass along with...
14526 (pass_data_cselim): ...new pass_data instance and...
14527 (make_pass_cselim): ...new function.
14528 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
14529 subclass of gimple_opt_pass along with...
14530 (pass_data_phiprop): ...new pass_data instance and...
14531 (make_pass_phiprop): ...new function.
14532 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
14533 subclass of gimple_opt_pass along with...
14534 (pass_data_pre): ...new pass_data instance and...
14535 (make_pass_pre): ...new function.
14536 (pass_fre): Convert from a global struct to a subclass of
14537 gimple_opt_pass along with...
14538 (pass_data_fre): ...new pass_data instance and...
14539 (make_pass_fre): ...new function.
14540 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
14541 subclass of gimple_opt_pass along with...
14542 (pass_data_reassoc): ...new pass_data instance and...
14543 (make_pass_reassoc): ...new function.
14544 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
14545 subclass of gimple_opt_pass along with...
14546 (pass_data_sink_code): ...new pass_data instance and...
14547 (make_pass_sink_code): ...new function.
14548 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
14549 subclass of gimple_opt_pass along with...
14550 (pass_data_strlen): ...new pass_data instance and...
14551 (make_pass_strlen): ...new function.
14552 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
14553 struct to a subclass of gimple_opt_pass along with...
14554 (pass_data_build_alias): ...new pass_data instance and...
14555 (make_pass_build_alias): ...new function.
14556 (pass_build_ealias): Convert from a global struct to a subclass of
14557 gimple_opt_pass along with...
14558 (pass_data_build_ealias): ...new pass_data instance and...
14559 (make_pass_build_ealias): ...new function.
14560 (pass_ipa_pta): Convert from a global struct to a subclass of
14561 simple_ipa_opt_pass along with...
14562 (pass_data_ipa_pta): ...new pass_data instance and...
14563 (make_pass_ipa_pta): ...new function.
14564 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
14565 subclass of gimple_opt_pass along with...
14566 (pass_data_uncprop): ...new pass_data instance and...
14567 (make_pass_uncprop): ...new function.
14568 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
14569 global struct to a subclass of gimple_opt_pass along with...
14570 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
14571 (make_pass_late_warn_uninitialized): ...new function.
14572 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
14573 to a subclass of gimple_opt_pass along with...
14574 (pass_data_init_datastructures): ...new pass_data instance and...
14575 (make_pass_init_datastructures): ...new function.
14576 (pass_early_warn_uninitialized): Convert from a global struct to a
14577 subclass of gimple_opt_pass along with...
14578 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
14579 (make_pass_early_warn_uninitialized): ...new function.
14580 (pass_update_address_taken): Convert from a global struct to a
14581 subclass of gimple_opt_pass along with...
14582 (pass_data_update_address_taken): ...new pass_data instance and...
14583 (make_pass_update_address_taken): ...new function.
14584 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
14585 struct to a subclass of gimple_opt_pass along with...
14586 (pass_data_release_ssa_names): ...new pass_data instance and...
14587 (make_pass_release_ssa_names): ...new function.
14588 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
14589 subclass of gimple_opt_pass along with...
14590 (pass_data_stdarg): ...new pass_data instance and...
14591 (make_pass_stdarg): ...new function.
14592 * tree-switch-conversion.c (pass_convert_switch): Convert from a
14593 global struct to a subclass of gimple_opt_pass along with...
14594 (pass_data_convert_switch): ...new pass_data instance and...
14595 (make_pass_convert_switch): ...new function.
14596 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
14597 to a subclass of gimple_opt_pass along with...
14598 (pass_data_tail_recursion): ...new pass_data instance and...
14599 (make_pass_tail_recursion): ...new function.
14600 (pass_tail_calls): Convert from a global struct to a subclass of
14601 gimple_opt_pass along with...
14602 (pass_data_tail_calls): ...new pass_data instance and...
14603 (make_pass_tail_calls): ...new function.
14604 * tree-vect-generic.c (pass_lower_vector): Convert from a global
14605 struct to a subclass of gimple_opt_pass along with...
14606 (pass_data_lower_vector): ...new pass_data instance and...
14607 (make_pass_lower_vector): ...new function.
14608 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
14609 gimple_opt_pass along with...
14610 (pass_data_lower_vector_ssa): ...new pass_data instance and...
14611 (make_pass_lower_vector_ssa): ...new function.
14612 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
14613 to a subclass of gimple_opt_pass along with...
14614 (pass_data_slp_vectorize): ...new pass_data instance and...
14615 (make_pass_slp_vectorize): ...new function.
14616 (pass_ipa_increase_alignment): Convert from a global struct to a
14617 subclass of simple_ipa_opt_pass along with...
14618 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
14619 (make_pass_ipa_increase_alignment): ...new function.
14620 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
14621 gimple_opt_pass along with...
14622 (pass_data_vrp): ...new pass_data instance and...
14623 (make_pass_vrp): ...new function.
14624 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
14625 subclass of simple_ipa_opt_pass along with...
14626 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
14627 (make_pass_ipa_free_lang_data): ...new function.
14628 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
14629 gimple_opt_pass along with...
14630 (pass_data_tsan): ...new pass_data instance and...
14631 (make_pass_tsan): ...new function.
14632 (pass_tsan_O0): Convert from a global struct to a subclass of
14633 gimple_opt_pass along with...
14634 (pass_data_tsan_O0): ...new pass_data instance and...
14635 (make_pass_tsan_O0): ...new function.
14636 * var-tracking.c (pass_variable_tracking): Convert from a global
14637 struct to a subclass of rtl_opt_pass along with...
14638 (pass_data_variable_tracking): ...new pass_data instance and...
14639 (make_pass_variable_tracking): ...new function.
14640 * web.c (pass_web): Convert from a global struct to a subclass of
14641 rtl_opt_pass along with...
14642 (pass_data_web): ...new pass_data instance and...
14643 (make_pass_web): ...new function.
14644 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
14645 declaration with that of...
14646 (make_pass_mode_switch_use): ...new function.
14647 (pass_resolve_sw_modes): Replace declaration with that of...
14648 (make_pass_resolve_sw_modes): ...new function.
14649 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
14650 from a global struct to a subclass of rtl_opt_pass along with...
14651 (pass_data_mode_switch_use): ...new pass_data instance and...
14652 (make_pass_mode_switch_use): ...new function.
14653 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
14654 from a global struct to a subclass of rtl_opt_pass along with...
14655 (pass_data_resolve_sw_modes): ...new pass_data instance and...
14656 (make_pass_resolve_sw_modes): ...new function.
14657 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
14658 struct to a subclass of rtl_opt_pass along with...
14659 (pass_data_insert_vzeroupper): ...new pass_data instance and...
14660 (make_pass_insert_vzeroupper): ...new function.
14661 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
14662 global struct to a subclass of rtl_opt_pass along with...
14663 (pass_data_work_around_errata): ...new pass_data instance and...
14664 (make_pass_work_around_errata): ...new function.
14665 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
14666 struct to a subclass of rtl_opt_pass along with...
14667 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
14668 (make_pass_mips_machine_reorg2): ...new function.
14670 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14672 * passes.c (pass_manager::operator new): New.
14674 2013-08-05 David Malcolm <dmalcolm@redhat.com>
14676 Handwritten part of conversion of passes to C++ classes.
14678 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
14679 (toplev.o): Add dep on PASS_MANAGER_H.
14680 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
14681 of early local pases to reflect this moving from a global to a
14682 member of gcc::pass_manager.
14683 (cgraph_add_new_function): Likewise.
14684 * lto-cgraph.c (lto_output_node): Update for conversion of
14685 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
14686 * passes.c (opt_pass::clone): New.
14687 (opt_pass::gate): New.
14688 (opt_pass::execute): New.
14689 (opt_pass::opt_pass): New.
14690 (pass_manager::execute_early_local_passes): New.
14691 (pass_manager::execute_pass_mode_switching): new.
14692 (finish_optimization_passes): Convert to...
14693 (pass_manager::finish_optimization_passes): ...this.
14694 (finish_optimization_passes): Update for conversion of passes to
14696 (register_dump_files_1): Use has_gate since we cannot portably
14697 check a vtable entry against NULL.
14698 (dump_one_pass): Likewise.
14699 (ipa_write_summaries_2): Likewise.
14700 (ipa_write_optimization_summaries_1): Likewise.
14701 (ipa_read_summaries_1): Likewise.
14702 (ipa_read_optimization_summaries_1): Likewise.
14703 (execute_ipa_stmt_fixups): Likewise.
14704 (pass_manager::pass_manager): Rewrite pass-creation, invoking
14705 pass-creation functions rather than wiring up globals, and
14706 storing the results in fields of pass_manager generated using
14707 pass-instances.def.
14708 (pass_manager::dump_profile_report): Update for conversion of
14709 passes to C++ classes.
14710 (pass_manager::execute_ipa_summary_passes): Likewise.
14711 (execute_one_ipa_transform_pass): Likewise.
14712 (execute_one_pass): Use has_gate and has_execute since we cannot
14713 portably check a vtable entry against NULL.
14714 * pass_manager.h (pass_manager::finish_optimization_passes): New.
14715 (pass_manager): Use pass-instances.def to add fields for the
14716 various pass instances.
14717 * toplev.c (finalize): Update for move of
14718 finish_optimization_passes to a method of gcc::pass_manager.
14719 * toplev.h (finish_optimization_passes): Move to method of class
14721 * tree-pass.h (struct pass_data): New.
14722 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
14723 (opt_pass::gate): Convert to virtual function.
14724 (opt_pass::~opt_pass): New.
14725 (opt_pass::clone): New.
14726 (opt_pass::execute): Convert to virtual function.
14727 (opt_pass::opt_pass): New.
14728 (opt_pass::ctxt_): new.
14729 (gimple_opt_pass): Convert to subclass of opt_pass.
14730 (gimple_opt_pass::gimple_opt_pass): New.
14731 (rtl_opt_pass): Convert to subclass of opt_pass.
14732 (rtl_opt_pass::rtl_opt_pass): New.
14733 (ipa_opt_pass_d): Convert to subclass of opt_pass.
14734 (ipa_opt_pass_d::ipa_opt_pass_d): New.
14735 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
14736 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
14737 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
14738 invocation of pass_mode_switching to reflect this moving from a
14739 global to a member of gcc::pass_manager.
14740 (ix86_option_override): Rework how pass_insert_vzeroupper is
14741 added to the pass_manager to reflect autogenerated changes.
14742 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
14744 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
14746 PR rtl-optimization/57708
14747 * recog.c (peep2_find_free_register): Validate all regs in a
14750 2013-08-05 Jan Hubicka <jh@suse.cz>
14753 * cgraph.c (verify_cgraph_node): Accept local flags from other
14755 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
14756 (function_and_variable_visibility): Likewise.
14757 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
14759 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14761 * graph.c (init_graph_slim_pretty_print): Remove.
14762 (print_graph_cfg): Do not call it. Use local pretty printer.
14763 (start_graph_dump): Likewise.
14765 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14767 * gimple-pretty-print.c (buffer): Remove.
14768 (initialized): Likewise.
14769 (maybe_init_pretty_print): Likewise.
14770 (print_gimple_stmt): Do not call it. Use non-static local
14771 pretty_printer variable.
14772 (print_gimple_expr): Likewise.
14773 (print_gimple_seq): Likewise.
14774 (gimple_dump_bb): Likewise.
14776 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
14778 * asan.c (asan_pp): Remove.
14779 (asan_pp_initialized): Likewise.
14780 (asan_pp_initialize): Likewise.
14781 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
14782 (asan_emit_stack_protection): Tidy. Use local pretty printer.
14783 (asan_add_global): Likewise.
14785 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
14787 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
14788 * diagnostic.h (diagnostic_flush_buffer): Adjust.
14789 * pretty-print.c (pp_formatted_text_data): Likewise.
14790 (pp_indent): Rename from pp_base_indent.
14791 (pp_format): Rename from pp_base_format.
14792 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
14793 (pp_format_verbatim): Rename from pp_base_format_verbatim.
14794 (pp_flush): Rename from pp_base_flush.
14795 (pp_set_line_maximum_length): Rename from
14796 pp_base_set_line_maximum_length.
14797 (pp_clear_output_area): Rename from pp_base_clear_output_area.
14798 (pp_set_prefix): Rename from pp_base_set_prefix.
14799 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
14800 (pp_emit_prefix): Rename from pp_base_emit_prefix.
14801 (pp_append_text): Rename from pp_base_append_text.
14802 (pp_formatted_text): Rename from pp_base_formatted_text.
14803 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
14804 (pp_remaining_character_count_for_line): Rename from
14805 pp_base_remaining_character_count_for_line.
14806 (pp_newline): Rename from pp_base_newline.
14807 (pp_character): Rename from pp_base_character.
14808 (pp_string): Rename from pp_base_string.
14809 (pp_maybe_space): Rename from pp_base_maybe_space.
14810 * asan.c (asan_pp_string): Adjust.
14811 (asan_emit_stack_protection): Likewise.
14812 (asan_add_global): Likewise.
14813 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
14814 * tree-mudflap.c (mf_varname_tree): Likewise.
14815 * tree-pretty-print.c (pp_tree_identifier): Rename from
14816 pp_base_tree_identifier.
14817 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
14818 Declare as function.
14820 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14822 * pretty-print.h (pp_bar_bar): New.
14823 (pp_ampersand_ampersand): Likewise.
14824 (pp_less_equal): Likewise.
14825 (pp_greater_equal): Likewise.
14826 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
14827 printer functions instead of pp_string or operators and punctuators.
14828 (dump_gimple_call): Likewise.
14829 (dump_gimple_omp_for): Likewise.
14830 (dump_gimple_transaction): Likewise.
14831 (dump_gimple_phi): Likewise.
14832 (pp_gimple_stmt_1): Likewise.
14833 * sched-vis.c (print_insn): Likewise.
14834 * tree-mudflap.c (mf_varname_tree): Likewise.
14835 * tree-pretty-print.c (dump_block_node): Likewise.
14836 (dump_generic_node): Likewise.
14838 2013-08-02 Jan Hubicka <jh@suse.cz>
14840 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
14842 * lto-streamer-out.c (tree_is_indexable): Results decls and
14843 parm decls are not indexable.
14844 (DFS_write_tree_body): Do not follow args and results.
14845 (hash_tree): Likewise.
14846 (output_functions): Rearrange so struct function is needed
14847 only when real body is output; be able to also ouptut abstract
14848 functions; output DECL_ARGUMENTS and DECL_RESULT.
14849 (lto_output): When not in WPA, ale store abstract functions.
14850 (write_symbol): Do not care about RESULT_DECL.
14851 (output_symbol_p): Handle correctly sbtract decls.
14852 * lto-streamer-in.c (input_function): Rearrange so struct
14853 function can be NULL at entry; allow streaming of
14854 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
14855 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
14856 sanity check during LTO.
14857 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
14858 RESULT_DECl and DECL_ARGUMENTS.
14859 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
14862 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
14864 * pretty-print.h (pp_underscore): New.
14866 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
14867 printer functions instead of pp_character.
14868 (dump_binary_rhs): Likewise.
14869 (dump_ternary_rhs): Likewise.
14870 (dump_gimple_call_args): Likewise.
14871 (pp_points_to_solution): Likewise.
14872 (dump_gimple_call): Likewise.
14873 (dump_gimple_switch): Likewise.
14874 (dump_gimple_cond): Likewise.
14875 (dump_gimple_bind): Likewise.
14876 (dump_gimple_try): Likewise.
14877 (dump_gimple_omp_for): Likewise.
14878 (dump_gimple_omp_continue): Likewise.
14879 (dump_gimple_omp_single): Likewise.
14880 (dump_gimple_omp_sections): Likewise.
14881 (dump_gimple_omp_block): Likewise.
14882 (dump_gimple_omp_critical): Likewise.
14883 (dump_gimple_transaction): Likewise.
14884 (dump_gimple_asm): Likewise.
14885 (dump_gimple_phi): Likewise.
14886 (dump_gimple_omp_parallel): Likewise.
14887 (dump_gimple_omp_task): Likewise.
14888 (dump_gimple_omp_atomic_load): Likewise.
14889 (dump_gimple_omp_atomic_store): Likewise.
14890 (dump_gimple_mem_ops): Likewise.
14891 (pp_gimple_stmt_1): Likewise.
14892 (pp_cfg_jump): Likewise.
14893 (dump_implicit_edges): Likewise.
14894 (gimple_dump_bb_for_graph): Likewise.
14895 * graph.c (draw_cfg_node): Likewise.
14896 * langhooks.c (lhd_print_error_function): Likewise.
14897 * sched-vis.c (print_exp): Likewise.
14898 (print_value): Likewise.
14899 (print_pattern): Likewise.
14900 (print_insn): Likewise.
14901 (rtl_dump_bb_for_graph): Likewise.
14902 * tree-pretty-print.c (dump_function_declaration): Likewise.
14903 (dump_array_domain): Likewise.
14904 (dump_omp_clause): Likewise.
14905 (dump_location): Likewise.
14906 (dump_generic_node): Likewise.
14907 (print_struct_decl): Likewise.
14908 * diagnostic.c (diagnostic_show_locus): Use pp_space.
14910 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14912 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
14913 candidate table when replacing a candidate statement.
14914 (replace_rhs_if_not_dup): Likewise.
14915 (replace_one_candidate): Likewise.
14917 2013-08-02 Jan Hubicka <jh@suse.cz>
14918 Martin Liska <marxin.liska@gmail.com>
14920 * cgraphunit.c (add_new_function): Fix logic when adding from
14922 (assemble_thunk): Rename to ...
14923 (expand_thunk); .. this one; export; get it working with
14924 general functions; make produced gimple valid.
14925 * cgraph.h (expand_thunk): Declare.
14927 2013-08-02 Jan Hubicka <jh@suse.cz>
14929 * ipa-cp.c (gather_context_independent_values): Use
14930 ipa_get_param_move_cost.
14931 (get_replacement_map): Remove PARAM; move parameter folding
14933 (create_specialized_node): Update.
14934 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
14935 assert that we have gimple body; update move_cost.
14936 (count_formal_params): Assert that we have gimple body.
14937 (ipa_dump_param): New function.
14938 (ipa_alloc_node_params): Break out from ...
14939 (ipa_initialize_node_params): ... here.
14940 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
14941 (ipa_write_node_info): Stream move costs.
14942 (ipa_read_node_info): Read move costs.
14943 (ipa_update_after_lto_read): Do not recompute node params.
14944 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
14945 (ipa_get_param): Check we are not in WPA.
14946 (ipa_get_param_move_cost): New.
14947 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
14948 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
14949 parm numbers to be present.
14951 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14953 PR rtl-optimization/58048
14954 * lra-constraints.c (process_alt_operands): Don't check asm
14955 operand on register.
14957 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
14959 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
14960 the implied StoreLoad barrier for atomic operations if before.
14962 2013-08-02 Jan Hubicka <jh@suse.cz>
14963 Martin Liska <marxin.liska@gmail.com>
14965 * cgraph.c (cgraph_function_body_availability): Do not check
14967 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
14968 symtab_node_availability): Declare.
14969 * ipa.c (can_replace_by_local_alias): New.
14970 (function_and_variable_visibility): Use it.
14971 * symtab.c (symtab_for_node_and_aliases,
14972 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
14974 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
14976 PR rtl-optimization/57963
14977 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
14978 (lra_constraints): Use them.
14980 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
14982 * config/arm/types.md (define_attr "type"): Add "load_acq"
14984 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
14986 (cortex_a53_store1): Likewise.
14988 2013-08-01 Jan Hubicka <jh@suse.cz>
14990 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
14991 partitions are not needed.
14993 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
14995 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
14996 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
14997 MAYBE_NON_Q_CLASS_P where appropriate.
14999 2013-08-01 Jan Hubicka <jh@suse.cz>
15001 * cgraph.h (release_function_body): Declare.
15002 * tree.c (free_lang_data_in_decl): Free, parameters and return values
15003 of unused delcarations.
15005 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15007 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
15008 RTL form when subtracting a constant.
15010 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15012 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
15013 Generate canonical plus rtx with negated immediate instead of minus
15015 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
15017 2013-08-01 Jan Hubicka <jh@suse.cz>
15019 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
15020 (cgraph_release_function_body): Likewise.
15021 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
15022 * cgraph.h (cgrpah_node): Rename abstract_and_needed
15023 to used_as_abstract_origin.
15024 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
15025 symbols used as abstract origins.
15026 * cgraphunit.c (analyze_functions): Update.
15027 * ipa.c (symtab_remove_unreachable_nodes): Recompute
15028 used_as_abstract_origin.
15029 * tree-inline.c (tree_function_versioning): Update
15030 used_as_abstract_origin; be ready for DECL_RESULT and
15031 DECL_ARGUMENTS to be NULL.
15033 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
15034 for abstract functions.
15035 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
15038 2013-08-01 Jan Hubicka <jh@suse.cz>
15040 * profile.c (compute_value_histograms): Fix thinko.
15042 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15044 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
15045 aarch-common-protos.h to extra_headers.
15046 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15047 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
15048 * config/aarch64/t-aarch64 (aarch-common.o): Define.
15050 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
15052 * config/aarch64/aarch64.md (define_attr "type"): Delete.
15053 Include "../arm/types.md". Define "type" attribute for all patterns.
15054 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
15057 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
15059 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
15060 to support power8 load fusion.
15061 (fusion_gpr_mem_load): Likewise.
15063 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
15065 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
15066 declarations for power8 load fusion.
15067 (emit_fusion_gpr_load): Likewise.
15069 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
15070 tuning for power8, turn on fusion mode by default. Turn on sign
15071 extending fusion mode if normal fusion mode is on, and we are at
15073 (fusion_gpr_load_p): New function, return true if we can fuse an
15074 addis instruction with a dependent load to a GPR.
15075 (emit_fusion_gpr_load): Emit the instructions for power8 load
15078 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
15079 (VSX load fusion peepholes): New peepholes to fuse together an
15080 addi instruction with a VSX load instruction.
15082 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
15083 peepholes to fuse an addis instruction with a load to a GPR base
15084 register. If we are supporting sign extending fusions, convert
15085 sign extending loads to zero extending loads and add an explicit
15088 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15090 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
15091 aarch-common-protos.h to extra_headers.
15092 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
15093 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
15094 (arm_early_store_addr_dep): Likewise.
15095 (arm_no_early_alu_shift_dep): Likewise.
15096 (arm_no_early_alu_shift_value_dep): Likewise.
15097 (arm_no_early_mul_dep): Likewise.
15098 (arm_no_early_store_addr_dep): Likewise.
15099 (arm_mac_accumulator_is_mul_result): Likewise.
15100 (arm_mac_accumulator_is_result): Likewise.
15101 * config/arm/aarch-common.c: ... here. New file.
15102 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
15104 (arm_early_store_addr_dep): Likewise.
15105 (arm_no_early_alu_shift_dep): Likewise.
15106 (arm_no_early_alu_shift_value_dep): Likewise.
15107 (arm_no_early_mul_dep): Likewise.
15108 (arm_no_early_store_addr_dep): Likewise.
15109 (arm_mac_accumulator_is_mul_result): Likewise.
15110 (arm_mac_accumulator_is_result): Likewise.
15111 * config/arm/aarch-common-protos.h: ... here. New file.
15112 * config/arm/t-arm (aarch-common.o): Define.
15114 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
15116 * config/arm/arm.md: Include new file "types.md".
15117 (define_attr "type"): Move from here to ...
15118 (define_attr "mul32"): Likewise.
15119 (define_attr "mul64"): Likewise.
15120 * config/arm/types.md: ... here. New file.
15122 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
15124 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
15125 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
15127 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15129 * gen-pass-instances.awk: Fix offset of substr().
15131 2013-07-31 David Malcolm <dmalcolm@redhat.com>
15133 * Makefile.in (pass-instances.def): New.
15134 (passes.o): Replace dependency on passes.def with one on
15137 * gen-pass-instances.awk: New.
15139 * passes.c (pass_manager::pass_manager): Use pass-instances.def
15140 rather than passes.def, updating local definition of NEXT_PASS
15141 macro to add an extra NUM parameter (currently unused).
15143 2013-07-30 David Malcolm <dmalcolm@redhat.com>
15145 * Makefile.in (PASS_MANAGER_H): New.
15146 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
15147 (passes.o): Likewise.
15148 (statistics.o): Likewise.
15149 (cgraphunit.o): Likewise.
15150 (context.o): Depend on PASS_MANAGER_H.
15152 * pass_manager.h: New.
15154 * cgraphunit.c (cgraph_add_new_function): Update for moves
15155 of globals to fields of pass_manager.
15156 (analyze_function): Likewise.
15157 (expand_function): Likewise.
15158 (ipa_passes): Likewise.
15159 (compile): Likewise.
15161 * context.c (context::context): New.
15162 * context.h (context::context): New.
15163 (context::get_passes): New.
15164 (context::passes_): New.
15166 * lto-cgraph.c (input_node): Update for moves of globals to
15167 fields of pass_manager.
15169 * passes.c (all_passes): Remove, in favor of a field of the
15170 same name within the new class pass_manager.
15171 (all_small_ipa_passes): Likewise.
15172 (all_lowering_passes): Likewise.
15173 (all_regular_ipa_passes): Likewise.
15174 (all_late_ipa_passes): Likewise.
15175 (all_lto_gen_passes): Likewise.
15176 (passes_by_id): Likewise.
15177 (passes_by_id_size): Likewise.
15178 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15179 the new class pass_manager.
15180 (set_pass_for_id): Convert to...
15181 (pass_manager::set_pass_for_id): ...method.
15182 (get_pass_for_id): Convert to...
15183 (pass_manager::get_pass_for_id): ...method.
15184 (register_one_dump_file): Move body of implementation into...
15185 (pass_manager::register_one_dump_file): ...here.
15186 (register_dump_files_1): Convert to...
15187 (pass_manager::register_dump_files_1): ...method.
15188 (register_dump_files): Convert to...
15189 (pass_manager::register_dump_files): ...method.
15190 (create_pass_tab): Update for moves of globals to fields of
15192 (dump_passes): Move body of implementation into...
15193 (pass_manager::dump_passes): ...here.
15194 (register_pass): Move body of implementation into...
15195 (pass_manager::register_pass): ...here.
15196 (init_optimization_passes): Convert into...
15197 (pass_manager::pass_manager): ...constructor for new
15198 pass_manager class, and initialize the pass_lists array.
15199 (check_profile_consistency): Update for moves of globals to
15200 fields of pass_manager.
15201 (dump_profile_report): Move body of implementation into...
15202 (pass_manager::dump_profile_report): ...here.
15203 (ipa_write_summaries_1): Update for moves of pass lists from
15204 being globals to fields of pass_manager.
15205 (ipa_write_optimization_summaries): Likewise.
15206 (ipa_read_summaries): Likewise.
15207 (ipa_read_optimization_summaries): Likewise.
15208 (execute_all_ipa_stmt_fixups): Likewise.
15210 * statistics.c (statistics_fini): Update for moves of globals to
15211 fields of pass_manager.
15213 * toplev.c (general_init): Replace call to
15214 init_optimization_passes with construction of the pass_manager
15217 * tree-pass.h (all_passes): Remove, in favor of a field of the
15218 same name within the new class pass_manager.
15219 (all_small_ipa_passes): Likewise.
15220 (all_lowering_passes): Likewise.
15221 (all_regular_ipa_passes): Likewise.
15222 (all_lto_gen_passes): Likewise.
15223 (all_late_ipa_passes): Likewise.
15224 (passes_by_id): Likewise.
15225 (passes_by_id_size): Likewise.
15226 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
15227 the new class pass_manager.
15228 (get_pass_for_id): Remove.
15230 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15232 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
15235 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
15237 * arm.md (mulhi3): New expand pattern.
15239 2013-07-30 Jan Hubicka <jh@suse.cz>
15240 Martin Liska <marxin.liska@gmail.com>
15242 * profile.c (compute_value_histograms): Do not ICE when
15243 there is mismatch only on some counters.
15245 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15247 PR rtl-optimization/57637
15248 * function.c (move_insn_for_shrink_wrap): Also check the
15249 GEN set of the LIVE problem for the liveness analysis
15250 if it exists, otherwise give up.
15252 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15254 PR tree-optimization/57993
15255 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
15256 replaced statement in the candidate table.
15257 (phi_add_costs): Return infinite cost when the hidden basis does
15258 not dominate all phis on which the candidate is dependent.
15259 (replace_one_candidate): Record replaced statement in the
15262 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
15264 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
15265 (ashlv2si3): New expander.
15266 (*ashlv2si3_i): New define_insn_and_split.
15267 * predicates.md (float_operation): Allow patterns with three
15268 basic sub-patterns.
15270 PR rtl-optimization/58021
15271 * mode-switching.c (create_pre_exit): Always split off preceding
15272 insns if we are not at the basic block head.
15274 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
15276 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
15277 (UCLIBC_DYNAMIC_LINKER): New macro.
15278 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
15280 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
15281 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
15283 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
15284 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
15285 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
15286 for SF and DF modes. Use ieee_quad_format for TF mode.
15287 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
15288 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
15289 (mips_option_override): Handle `-mnan=legacy'.
15290 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
15291 `-mabs=2008' and `-mnan=2008'.
15292 (OPTION_DEFAULT_SPECS): Add "nan" default.
15293 (ASM_SPEC): Handle `-mnan='.
15294 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
15295 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
15296 comment accordingly.
15297 (neg<mode>2): Likewise.
15298 * config/mips/mips.opt (mabs, mnan): New options.
15299 * doc/install.texi (Configuration): Document `--with-nan=' option.
15300 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
15302 (MIPS Options): Document them.
15303 * config.gcc <mips*-*-*>: Handle `--with-nan='.
15304 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
15305 * configure: Regenerate.
15306 * config.in: Regenerate.
15308 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15310 * config/i386/i386.md (float post-reload splitters): Do not check
15311 for subregs of SSE registers.
15313 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
15314 H.J. Lu <hongjiu.lu@intel.com>
15318 * config/i386/i386.md (post-reload splitter
15319 to avoid partial SSE reg dependency stalls): New pattern.
15321 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
15323 * config/s390/s390.md ("movcc"): Swap load and store instructions.
15325 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
15327 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
15328 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
15330 2013-07-26 Cary Coutant <ccoutant@google.com>
15332 * dwarf2out.c (die_checksum_ordered): Don't include template
15333 instantiations in signature.
15334 (is_template_parameter): New function.
15335 (is_template_instantiation): New function.
15336 (generate_skeleton_bottom_up): Don't include template instantiations
15338 (generate_skeleton): Likewise.
15339 (break_out_comdat_types): Move recursive call to break out nested
15341 (prune_unused_types_mark_generic_parms_dies): Call
15342 is_template_parameter.
15344 2013-07-26 Ian Bolton <ian.bolton@arm.com>
15346 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
15347 uses vector registers.
15348 * config/aarch64/iterators.md: Add attributes rtn and vas.
15350 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15351 Richard Earnshaw <richard.earnshaw@arm.com>
15353 * combine.c (simplify_comparison): Re-canonicalize operands
15355 * config/arm/arm.md (movcond_addsi): New splitter.
15357 2013-07-25 Sterling Augustine <saugustine@google.com>
15359 * dwarf2out.c (size_of_pubnames): Move code to...
15360 (include_pubname_in_output): ...here. New.
15361 (want_pubnames): Rearrange.
15362 (output_pubnames): Call include_pubname_in_output. Move assertion.
15364 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15366 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
15368 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
15371 * doc/extend.texi: Remove obsolete builtins. Fix
15372 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
15374 2013-07-25 Jan Hubicka <jh@suse.cz>
15376 * cgraph.c (release_function_body): Break out from ...
15377 (cgraph_release_function_body): ... this one; also release DECL_RESULT
15378 and DECL_ARGUMENTS.
15379 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
15380 old_tree in the map.
15381 (create_specialized_node): Update.
15382 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
15384 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
15385 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
15386 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
15387 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
15390 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15392 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
15393 addsi3_carryin_alt2_<optab>): Correct output template.
15395 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15397 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
15398 Adjust for arm_restrict_it.
15399 Remove trailing whitespace.
15401 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
15403 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
15404 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
15406 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
15408 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
15410 PR rtl-optimization/57960
15411 * lra-constraints.c (process_alt_operands): Use the right mode
15412 when checking strict_low.
15414 2013-07-25 Jan Hubicka <jh@suse.cz>
15416 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
15417 * cgraph.c (cgraph_remove_node): Do not release function body
15418 when in cgraph streaming.
15419 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
15420 in other partitions are not considered reachable; fix handling of
15423 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15425 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
15427 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15432 * config/arm/arm.md ("*sibcall_insn): Replace use of
15433 Ss with US. Adjust output for v5 and v4t.
15434 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
15436 * config/arm/constraints.md ("Ss"): Rename to US.
15438 2013-07-25 Terry Guo <terry.guo@arm.com>
15440 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
15441 shift_add/shift_sub0/shift_sub1 RTXs.
15443 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
15444 Anton Blanchard <anton@au1.ibm.com>
15446 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
15447 (altivec_vpks<VI_char>ss): Likewise.
15448 (altivec_vpks<VI_char>us): Likewise.
15449 (altivec_vpku<VI_char>us): Likewise.
15450 (altivec_vpku<VI_char>um): Likewise.
15452 2013-07-24 David Malcolm <dmalcolm@redhat.com>
15454 Introduce context class.
15456 * Makefile.in (CONTEXT_H): New.
15457 (OBJS): Add context.o.
15458 (toplev.o): Add CONTEXT_H to dependencies.
15461 * toplev.c (general_init): Create the singleton gcc::context instance.
15467 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
15469 PR rtl-optimization/57968
15470 * mode-switching.c (create_pre_exit): Allow instructions that
15471 don't set a return register to need a non-exit mode.
15473 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15474 Anton Blanchard <anton@au1.ibm.com>
15476 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
15477 operands to vperm for little endian.
15478 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
15479 of lvsl to create the control mask for a vperm for little endian.
15481 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15482 Anton Blanchard <anton@au1.ibm.com>
15484 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
15485 two operands for little-endian.
15487 2013-07-23 Steve Ellcey <sellcey@mips.com>
15489 * config/mips/mips.c (mips_case_values_threshold): New.
15490 (TARGET_CASE_VALUES_THRESHOLD): Define.
15492 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15493 Anton Blanchard <anton@au1.ibm.com>
15495 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
15496 selection of field for vector splat in little endian mode.
15498 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
15500 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
15501 expanders to rs6000.md.
15502 (ior<mode>3): Likewise.
15503 (and<mode>3): Likewise.
15504 (one_cmpl<mode>2): Likewise.
15505 (nor<mode>3): Likewise.
15506 (andc<mode>3): Likewise.
15507 (eqv<mode>3): Likewise.
15508 (nand<mode>3): Likewise.
15509 (orc<mode>3): Likewise.
15511 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
15514 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
15515 to split multi-word logical operations.
15516 (rs6000_split_logical_di): Likewise.
15517 (rs6000_split_logical): Likewise.
15519 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
15520 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
15521 and allow TImode operations in 32-bit.
15522 (vsx_and<mode>3_64bit): Likewise.
15523 (vsx_ior<mode>3_32bit): Likewise.
15524 (vsx_ior<mode>3_64bit): Likewise.
15525 (vsx_xor<mode>3_32bit): Likewise.
15526 (vsx_xor<mode>3_64bit): Likewise.
15527 (vsx_one_cmpl<mode>2_32bit): Likewise.
15528 (vsx_one_cmpl<mode>2_64bit): Likewise.
15529 (vsx_nor<mode>3_32bit): Likewise.
15530 (vsx_nor<mode>3_64bit): Likewise.
15531 (vsx_andc<mode>3_32bit): Likewise.
15532 (vsx_andc<mode>3_64bit): Likewise.
15533 (vsx_eqv<mode>3_32bit): Likewise.
15534 (vsx_eqv<mode>3_64bit): Likewise.
15535 (vsx_nand<mode>3_32bit): Likewise.
15536 (vsx_nand<mode>3_64bit): Likewise.
15537 (vsx_orc<mode>3_32bit): Likewise.
15538 (vsx_orc<mode>3_64bit): Likewise.
15540 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
15541 logical types in GPRs.
15543 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
15544 logical insns to rs6000.md, and allow TImode operations in
15546 (altivec_ior<mode>3): Likewise.
15547 (altivec_xor<mode>3): Likewise.
15548 (altivec_one_cmpl<mode>2): Likewise.
15549 (altivec_nor<mode>3): Likewise.
15550 (altivec_andc<mode>3): Likewise.
15552 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
15553 attributes for moving the 128-bit logical operations into
15555 (BOOL_REGS_OUTPUT): Likewise.
15556 (BOOL_REGS_OP1): Likewise.
15557 (BOOL_REGS_OP2): Likewise.
15558 (BOOL_REGS_UNARY): Likewise.
15559 (BOOL_REGS_AND_CR0): Likewise.
15560 (one_cmpl<mode>2): Add support for DI logical operations on
15561 32-bit, splitting the operations to 32-bit.
15562 (anddi3): Likewise.
15563 (iordi3): Likewise.
15564 (xordi3): Likewise.
15565 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
15566 changes to combine the 32/64-bit code, allow logical operations on
15567 TI mode in 32-bit, and to use similar match_operator patterns like
15568 scalar mode uses. Combine the Altivec and VSX code for logical
15569 operations, and move it here.
15570 (ior<mode>3, 128-bit types): Likewise.
15571 (xor<mode>3, 128-bit types): Likewise.
15572 (one_cmpl<mode>3, 128-bit types): Likewise.
15573 (nor<mode>3, 128-bit types): Likewise.
15574 (andc<mode>3, 128-bit types): Likewise.
15575 (eqv<mode>3, 128-bit types): Likewise.
15576 (nand<mode>3, 128-bit types): Likewise.
15577 (orc<mode>3, 128-bit types): Likewise.
15578 (and<mode>3_internal): Likewise.
15579 (bool<mode>3_internal): Likewise.
15580 (boolc<mode>3_internal1): Likewise.
15581 (boolc<mode>3_internal2): Likewise.
15582 (boolcc<mode>3_internal1): Likewise.
15583 (boolcc<mode>3_internal2): Likewise.
15584 (eqv<mode>3_internal1): Likewise.
15585 (eqv<mode>3_internal2): Likewise.
15586 (one_cmpl1<mode>3_internal): Likewise.
15588 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15590 * config/microblaze/microblaze.c (microblaze_expand_prologue):
15591 Rename flag_stack_usage to flag_stack_usage_info.
15593 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
15595 * config/microblaze/sync.md: New file.
15596 * config/microblaze/microblaze.md: Include sync.md
15597 * config/microblaze/microblaze.c: Add print_operand 'y'.
15598 * config/microblaze/constraints.md: Add memory_contraint
15599 'Q' which is a single register.
15601 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
15603 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
15605 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15607 * reload.c (find_reloads): Exit loop once we find this operand
15608 cannot be reloaded somehow for this alternative.
15610 * reload.c (find_reloads): Exit loop once we find a hard register.
15612 * rtlanal.c (computed_jump_p): Exit loop once we find label
15615 * i386.c (ix86_pad_returns): Exit loop after setting replace.
15617 * cfgloopmanip.c (remove_path): Exit loop after setting
15620 * gensupport.c (subst_dup): Avoid loop if code is not
15621 MATCH_DUP nor MATCH_OP_DUP.
15623 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
15625 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
15627 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15629 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
15630 true for SP_REGNUM if mode == ptr_mode.
15631 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
15632 with value R0_REGNUM + 31.
15634 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15636 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
15637 pad pointer-typed argument downward.
15639 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15641 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
15642 and __ILP32__ when the ILP32 model is in use.
15644 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15646 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
15647 (aarch64_load_symref_appropriately): In the case of
15648 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
15649 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
15650 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
15651 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
15652 if 'mode' doesn't equal to 'ptr_mode'.
15653 (aarch64_output_mi_thunk): Add an assertion on the alignment of
15654 'vcall_offset'; change to call aarch64_emit_move differently depending
15655 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
15656 to calculate the upper bound of 'vcall_offset'.
15657 (aarch64_cannot_force_const_mem): Change to also return true if
15659 (aarch64_legitimize_reload_address): In the case of large
15660 displacements, add new local variable 'xmode' and an assertion
15661 based on it; change to use 'xmode' to generate the new rtx and
15663 (aarch64_asm_trampoline_template): Change to generate the template
15664 differently depending on TARGET_ILP32 or not; change to use
15665 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
15666 (aarch64_trampoline_size): Removed.
15667 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
15668 and replace immediate literals with it. Change to use 'ptr_mode'
15669 instead of 'DImode' and call convert_memory_address if the mode
15670 of 'fnaddr' doesn't equal to 'ptr_mode'.
15671 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
15673 (aarch64_elf_asm_destructor): Likewise.
15674 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
15675 on TARGET_ILP32 instead of aarch64_trampoline_size.
15676 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
15677 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
15678 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
15679 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
15680 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
15681 (storewb_pair<GPI:mode>_<P:mode>): ... this.
15682 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
15683 depending on the value of 'mode'.
15684 (add_losym_<mode>): New.
15685 (ldr_got_small_<mode>): New, based on ldr_got_small.
15686 (ldr_got_small): Remove.
15687 (ldr_got_small_sidi): New.
15688 * config/aarch64/iterators.md (P): New.
15689 (PTR): Change to 'ptr_mode' in the condition.
15691 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
15693 * config.gcc (aarch64*-*-*): Support --with-abi.
15694 (aarch64*-*-elf): Support --with-multilib-list.
15695 (aarch64*-*-linux*): Likewise.
15696 (supported_defaults): Add abi to aarch64*-*-*.
15697 * configure.ac: Mention AArch64 for --with-multilib-list.
15698 * configure: Re-generated.
15699 * config/aarch64/biarchilp32.h: New file.
15700 * config/aarch64/biarchlp64.h: New file.
15701 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
15703 (MULTILIB_DEFAULTS): Ditto.
15704 (DRIVER_SELF_SPECS): Ditto.
15705 (ASM_SPEC): Update to also substitute -mabi.
15706 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
15707 file whose name depends on -mabi= and -mbig-endian.
15708 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
15710 (POINTER_SIZE): New define.
15711 (POINTERS_EXTEND_UNSIGNED): Ditto.
15712 (enum aarch64_abi_type): New enumeration tag.
15713 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
15714 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
15715 (TARGET_ILP32): New define.
15716 * config/aarch64/aarch64.opt (mabi): New.
15717 (aarch64_abi): New.
15718 (ilp32, lp64): New values for -mabi.
15719 * config/aarch64/t-aarch64 (comma): New define.
15720 (MULTILIB_OPTIONS): Ditto.
15721 (MULTILIB_DIRNAMES): Ditto.
15722 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
15723 * doc/invoke.texi: Document -mabi for AArch64.
15725 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
15727 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
15729 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
15730 Anton Blanchard <anton@au1.ibm.com>
15732 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
15733 endianness when selecting field to splat.
15735 2013-07-22 Eric Christopher <echristo@gmail.com>
15737 * dwarf2out.c (die_odr_checksum): New function to use
15738 CHECKSUM_ macros and ULEB128 for DIE tag.
15739 (generate_type_signature): Use.
15741 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
15743 * config.gcc (sparc*-*-*): Accept leon3 processor.
15744 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
15745 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
15746 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
15747 * config/sparc/sparc.opt (enum processor_type): Add leon3.
15748 (mfix-ut699): Adjust comment.
15749 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
15750 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
15751 (CPP_CPU_SPEC): Likewise.
15752 (ASM_CPU_SPEC): Likewise.
15753 * config/sparc/sparc.c (leon3_cost): New constant.
15754 (sparc_option_override): Add leon3 support.
15755 (mem_ref): New function.
15756 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
15757 (sparc_do_work_around_errata): Look into the instruction in the delay
15758 slot and adjust accordingly. Add fix for the data cache nullify issues
15759 of the UT699. Change insertion position for the NOP.
15760 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
15761 (leon3_load): New reservation.
15762 (leon_store): Bump latency to 2.
15763 (grfpu): New automaton.
15764 (grfpu_alu): New unit.
15765 (grfpu_ds): Likewise.
15766 (leon_fp_alu): Adjust.
15767 (leon_fp_mult): Delete.
15768 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
15769 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
15770 * config/sparc/sparc.md (cpu): Add leon3.
15771 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
15772 (swapsi): Likewise.
15773 (atomic_test_and_set): Likewise.
15774 (ldstub): Likewise.
15776 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
15778 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
15779 default for R5900 targets.
15780 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
15781 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
15782 * config/mips/mips.c (mips_option_override): Report an error for
15783 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
15784 for -march=r5900 -mhard-float.
15786 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
15788 * df-problems.c (can_move_insns_across): Exit loop once we
15789 find a non-fixed, non-global register.
15791 * ipa-pure-const.c (propagate_nothrow): Exit loop after
15794 * omega.c (omega_eliminate_red): Break after setting red_found.
15795 (omega_problem_has_red_equations): Similarly after setting found.
15796 (omega_query_variable): Similarly after setting coupled.
15798 2013-07-22 Marek Polacek <polacek@redhat.com>
15800 * gimplify.c: Don't include gimple.h twice.
15802 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15804 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
15805 instead of TARGET_THUMB1.
15806 (Pz): New constraint.
15807 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
15809 (compare_negsi_si): Likewise.
15810 (compare_addsi2_op0): Likewise.
15811 (compare_addsi2_op1): Likewise.
15812 (addsi3_carryin_<optab>): Likewise.
15813 (addsi3_carryin_alt2_<optab>): Likewise.
15814 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
15815 for arm_restrict_it.
15816 (subsi3_carryin): Likewise.
15817 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
15818 (minmax_arithsi): Disable for arm_restrict_it.
15819 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
15820 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
15821 (satsi_<SAT:code>_shift): Likewise.
15822 (arm_shiftsi3): Add alternative for 16-bit encoding.
15823 (arm32_movhf): Disable for arm_restrict_it.
15824 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
15825 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
15827 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
15829 * config/arm/arm.md (attribute "insn"): Delete.
15830 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
15831 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
15832 (not_shiftsi): Update for attribute change.
15833 (not_shiftsi_compare0): Likewise.
15834 (not_shiftsi_compare0_scratch): Likewise.
15835 (arm_one_cmplsi2): Likewise.
15836 (thumb1_one_cmplsi2): Likewise.
15837 (notsi_compare0): Likewise.
15838 (notsi_compare0_scratch): Likewise.
15839 (thumb1_movdi_insn): Likewise.
15840 (arm_movsi_insn): Likewise.
15841 (movhi_insn_arch4): Likewise.
15842 (movhi_bytes): Likewise.
15843 (arm_movqi_insn): Likewise.
15844 (thumb1_movqi_insn): Likewise.
15845 (arm32_movhf): Likewise.
15846 (thumb1_movhf): Likewise.
15847 (arm_movsf_soft_insn): Likewise.
15848 (thumb1_movsf_insn): Likewise.
15849 (thumb_movdf_insn): Likewise.
15850 (movsicc_insn): Likewise.
15851 (movsfcc_soft_insn): Likewise.
15852 (and_scc): Likewise.
15853 (cond_move): Likewise.
15854 (if_move_not): Likewise.
15855 (if_not_move): Likewise.
15856 (if_shift_move): Likewise.
15857 (if_move_shift): Likewise.
15858 (if_shift_shift): Likewise.
15859 (if_not_arith): Likewise.
15860 (if_arith_not): Likewise.
15861 (cond_move_not): Likewise.
15862 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
15863 (neon_mov<mode>): Likewise.
15864 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
15865 (thumb2_movsi_vfp): Likewise.
15866 (movsf_vfp): Likewise.
15867 (thumb2_movsf_vfp): Likewise.
15868 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15870 (cortexa7_older_only): Likewise.
15871 (cortexa7_younger): Likewise.
15872 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
15873 (1020alu_shift_op): Likewise.
15874 (1020alu_shift_reg_op): Likewise.
15875 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
15876 (alu_shift_op): Likewise.
15877 (alu_shift_reg_op): Likewise.
15878 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
15879 (11_alu_shift_op): Likewise.
15880 (11_alu_shift_reg_op): Likewise.
15881 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
15882 (9_alu_shift_reg_op): Likewise.
15883 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15885 (cortex_a15_alu_shift): Likewise.
15886 (cortex_a15_alu_shift_reg): Likewise.
15887 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
15888 (cortex_a5_alu_shift): Likewise.
15889 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
15891 (cortex_a53_alu_shift): Likewise.
15892 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15894 (cortex_a7_alu_reg): Likewise.
15895 (cortex_a7_alu_shift): Likewise.
15896 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15897 (cortex_a8_alu_shift): Likewise.
15898 (cortex_a8_alu_shift_reg): Likewise.
15899 (cortex_a8_mov): Likewise.
15900 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
15901 (cortex_a9_dp_shift): Likewise.
15902 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
15903 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
15904 (cortex_r4_mov): Likewise.
15905 (cortex_r4_alu_shift): Likewise.
15906 (cortex_r4_alu_shift_reg): Likewise.
15907 * config/arm/fa526.md (526_alu_op): Update for attribute change.
15908 (526_alu_shift_op): Likewise.
15909 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
15910 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
15911 (626te_alu_shift_op): Likewise.
15912 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
15913 (726te_alu_op): Likewise.
15914 (726te_alu_shift_op): Likewise.
15915 (726te_alu_shift_reg_op): Likewise.
15916 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
15917 (mp626_alu_shift_op): Likewise.
15918 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
15919 (pj4_alu_e1_conds): Likewise.
15920 (pj4_alu): Likewise.
15921 (pj4_alu_conds): Likewise.
15922 (pj4_shift): Likewise.
15923 (pj4_shift_conds): Likewise.
15924 (pj4_alu_shift): Likewise.
15925 (pj4_alu_shift_conds): Likewise.
15927 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15929 * config/arm/predicates.md (shiftable_operator_strict_it):
15931 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
15932 Disable cond_exec version for arm_restrict_it.
15933 (thumb2_smaxsi3): Convert to generate cond_exec.
15934 (thumb2_sminsi3): Likewise.
15935 (thumb32_umaxsi3): Likewise.
15936 (thumb2_uminsi3): Likewise.
15937 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
15938 (thumb2_neg_abssi2): Likewise.
15939 (thumb2_mov_scc): Add alternative for 16-bit encoding.
15940 (thumb2_movsicc_insn): Adjust alternatives.
15941 (thumb2_mov_negscc): Disable for arm_restrict_it.
15942 (thumb2_mov_negscc_strict_it): New pattern.
15943 (thumb2_mov_notscc_strict_it): New pattern.
15944 (thumb2_mov_notscc): Disable for arm_restrict_it.
15945 (thumb2_ior_scc): Likewise.
15946 (thumb2_ior_scc_strict_it): New pattern.
15947 (thumb2_cond_move): Adjust for arm_restrict_it.
15948 (thumb2_cond_arith): Disable for arm_restrict_it.
15949 (thumb2_cond_arith_strict_it): New pattern.
15950 (thumb2_cond_sub): Adjust for arm_restrict_it.
15951 (thumb2_movcond): Likewise.
15952 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
15953 (thumb2_zero_extendhisi2_v6): Likewise.
15954 (thumb2_zero_extendqisi2_v6): Likewise.
15955 (orsi_notsi_si): Likewise.
15956 (orsi_not_shiftsi_si): Likewise.
15958 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
15960 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
15961 instruction sequence is 1 byte shorter.
15963 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
15965 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
15966 it is not needed after split.
15968 2013-07-20 Iain Sandoe <iain@codesourcery.com>
15971 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
15972 second label for nonlocal goto receivers. Don't output pic base labels
15973 unless we're producing PIC; mark that action unreachable().
15974 (ix86_save_reg): If the function contains a nonlocal label, save the
15976 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
15977 * config/darwin.c (emitted_pic_label_num): New GTY.
15978 (update_pic_label_number_if_needed): New.
15979 (machopic_output_function_base_name): Adjust for nonlocal receiver
15981 (machopic_should_output_picbase_label): New.
15982 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
15983 (nonlocal_goto_receiver): New insn and split.
15985 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
15987 * config/aarch64/aarch64-builtins.c
15988 (aarch64_fold_builtin): Fold abs in all modes.
15989 * config/aarch64/aarch64-simd-builtins.def
15990 (abs): Enable for all modes.
15991 * config/aarch64/arm_neon.h
15992 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
15993 (vabs_f64): Add missing intrinsic.
15995 2013-07-19 Ian Bolton <ian.bolton@arm.com>
15997 * config/aarch64/arm_neon.h (vabs_s64): New function
15999 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
16002 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
16003 * config/avr/avr.md (adjust_len): Add `round'.
16004 * config/avr/avr-protos.h (avr_out_round): New prototype.
16005 (avr_out_plus): Add `out_label' argument.
16006 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
16007 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
16008 Handle the case where `insn' is just a pattern.
16009 (avr_out_bitop): Handle the case where `insn' is just a pattern.
16010 (avr_out_round): New function.
16011 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
16013 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
16015 * config/microblaze/microblaze.c (microblaze_expand_prologue):
16016 Add check for flag_stack_usage to handle -fstack-usage support
16018 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
16020 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
16021 interaction for new Power8 flags and VSX.
16023 2013-07-18 Sriraman Tallam <tmsriram@google.com>
16025 PR middle-end/57698
16026 * tree-inline.c (expand_call_inline): Emit errors during
16027 early_inlining only if optimization is not turned on.
16029 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16033 * passes.c (init_optimization_passes): Move the construction of
16034 the pass hierarchy into a new passes.def file.
16036 * Makefile.in (passes.o): Add dependency on passes.def.
16038 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16040 * passes.c (init_optimization_passes): Introduce macros for
16041 constructing the tree of passes (INSERT_PASSES_AFTER,
16042 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
16043 TERMINATE_PASS_LIST).
16045 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
16046 Wei Mi <wmi@google.com>
16048 PR rtl-optimization/57878
16049 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
16051 (reload_pseudo_compare_func): Check nregs first for reload
16054 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16056 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
16058 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
16060 * read-rtl.c (validate_const_int): Once an invalid character is
16061 seen, quit the loop.
16063 * gengtype.c (write_roots): Similarly once we find the "deletable"
16064 or "if_marked" option.
16066 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16068 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
16069 "xtab" and "sat". Move value "clz" from here to ...
16070 (attriubte "type"): ... here.
16071 (satsi_<SAT:code>): Delete "insn" attribute.
16072 (satsi_<SAT:code>_shift): Likewise.
16073 (arm_zero_extendqisi2addsi): Likewise.
16074 (arm_extendqisi2addsi): Likewise.
16075 (clzsi2): Update for attribute changes.
16076 (rbitsi2): Likewise.
16077 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
16079 (arm_usatsihi): Likewise.
16080 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
16082 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
16084 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
16085 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
16086 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
16087 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
16088 in alphabetical order.
16089 (attribute "core_cycles"): Update for attribute changes.
16090 (arm_addsi3): Likewise.
16091 (addsi3_compare0): Likewise.
16092 (addsi3_compare0_scratch): Likewise.
16093 (addsi3_compare_op1): Likewise.
16094 (addsi3_compare_op2): Likewise.
16095 (compare_addsi2_op0): Likewise.
16096 (compare_addsi2_op1): Likewise.
16097 (addsi3_carryin_shift_<optab>): Likewise.
16098 (subsi3_carryin_shift): Likewise.
16099 (rsbsi3_carryin_shift): Likewise.
16100 (arm_subsi3_insn): Likewise.
16101 (subsi3_compare0): Likewise.
16102 (subsi3_compare): Likewise.
16103 (arm_andsi3_insn): Likewise.
16104 (thumb1_andsi3_insn): Likewise.
16105 (andsi3_compare0): Likewise.
16106 (andsi3_compare0_scratch): Likewise.
16107 (zeroextractsi_compare0_scratch
16108 (andsi_not_shiftsi_si): Likewise.
16109 (iorsi3_insn): Likewise.
16110 (iorsi3_compare0): Likewise.
16111 (iorsi3_compare0_scratch): Likewise.
16112 (arm_xorsi3): Likewise.
16113 (thumb1_xorsi3_insn): Likewise.
16114 (xorsi3_compare0): Likewise.
16115 (xorsi3_compare0_scratch): Likewise.
16116 (satsi_<SAT:code>_shift): Likewise.
16118 (arm_shiftsi3): Likewise.
16119 (shiftsi3_compare0): Likewise.
16120 (not_shiftsi): Likewise.
16121 (not_shiftsi_compare0): Likewise.
16122 (not_shiftsi_compare0_scratch): Likewise.
16123 (arm_one_cmplsi2): Likewise.
16124 (thumb_one_complsi2): Likewise.
16125 (notsi_compare0): Likewise.
16126 (notsi_compare0_scratch): Likewise.
16127 (thumb1_zero_extendhisi2): Likewise.
16128 (arm_zero_extendhisi2): Likewise.
16129 (arm_zero_extendhisi2_v6): Likewise.
16130 (arm_zero_extendhisi2addsi): Likewise.
16131 (thumb1_zero_extendqisi2): Likewise.
16132 (thumb1_zero_extendqisi2_v6): Likewise.
16133 (arm_zero_extendqisi2): Likewise.
16134 (arm_zero_extendqisi2_v6): Likewise.
16135 (arm_zero_extendqisi2addsi): Likewise.
16136 (thumb1_extendhisi2): Likewise.
16137 (arm_extendhisi2): Likewise.
16138 (arm_extendhisi2_v6): Likewise.
16139 (arm_extendqisi): Likewise.
16140 (arm_extendqisi_v6): Likewise.
16141 (arm_extendqisi2addsi): Likewise.
16142 (thumb1_extendqisi2): Likewise.
16143 (thumb1_movdi_insn): Likewise.
16144 (arm_movsi_insn): Likewise.
16145 (movsi_compare0): Likewise.
16146 (movhi_insn_arch4): Likewise.
16147 (movhi_bytes): Likewise.
16148 (arm_movqi_insn): Likewise.
16149 (thumb1_movqi_insn): Likewise.
16150 (arm32_movhf): Likewise.
16151 (thumb1_movhf): Likewise.
16152 (arm_movsf_soft_insn): Likewise.
16153 (thumb1_movsf_insn): Likewise.
16154 (movdf_soft_insn): Likewise.
16155 (thumb_movdf_insn): Likewise.
16156 (arm_cmpsi_insn): Likewise.
16157 (cmpsi_shiftsi): Likewise.
16158 (cmpsi_shiftsi_swp): Likewise.
16159 (arm_cmpsi_negshiftsi_si): Likewise.
16160 (movsicc_insn): Likewise.
16161 (movsfcc_soft_insn): Likewise.
16162 (arith_shiftsi): Likewise.
16163 (arith_shiftsi_compare0
16164 (arith_shiftsi_compare0_scratch
16165 (sub_shiftsi): Likewise.
16166 (sub_shiftsi_compare0
16167 (sub_shiftsi_compare0_scratch
16168 (and_scc): Likewise.
16169 (cond_move): Likewise.
16170 (if_plus_move): Likewise.
16171 (if_move_plus): Likewise.
16172 (if_move_not): Likewise.
16173 (if_not_move): Likewise.
16174 (if_shift_move): Likewise.
16175 (if_move_shift): Likewise.
16176 (if_shift_shift): Likewise.
16177 (if_not_arith): Likewise.
16178 (if_arith_not): Likewise.
16179 (cond_move_not): Likewise.
16180 (thumb1_ashlsi3): Set type attribute.
16181 (thumb1_ashrsi3): Likewise.
16182 (thumb1_lshrsi3): Likewise.
16183 (thumb1_rotrsi3): Likewise.
16184 (shiftsi3_compare0_scratch): Likewise.
16185 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
16186 (neon_mov<mode>): Likewise.
16187 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
16189 (thumb2_movsi_insn): Likewise.
16190 (thumb2_cmpsi_neg_shiftsi): Likewise.
16191 (thumb2_extendqisi_v6): Likewise.
16192 (thumb2_zero_extendhisi2_v6): Likewise.
16193 (thumb2_zero_extendqisi2_v6): Likewise.
16194 (thumb2_shiftsi3_short): Likewise.
16195 (thumb2_addsi3_compare0_scratch): Likewise.
16196 (orsi_not_shiftsi_si): Likewise.
16197 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
16198 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
16200 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
16201 (1020alu_shift_op): Likewise.
16202 (1020alu_shift_reg_op): Likewise.
16203 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
16204 (alu_shift_op): Likewise.
16205 (alu_shift_reg_op): Likewise.
16206 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
16207 (11_alu_shift_op): Likewise.
16208 (11_alu_shift_reg_op): Likewise.
16209 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
16210 (9_alu_shift_reg_op): Likewise.
16211 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
16213 (cortex_a15_alu_shift): Likewise.
16214 (cortex_a15_alu_shift_reg): Likewise.
16215 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
16217 (cortex_a5_alu_shift): Likewise.
16218 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
16220 (cortex_a53_alu_shift): Likewise.
16221 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
16223 (cortex_a7_alu_reg): Likewise.
16224 (cortex_a7_alu_shift): Likewise.
16225 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
16227 (cortex_a8_alu_shift): Likewise.
16228 (cortex_a8_alu_shift_reg): Likewise.
16229 (cortex_a8_mov): Likewise.
16230 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
16231 (cortex_a9_dp_shift): Likewise.
16232 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
16234 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
16236 (cortex_r4_mov): Likewise.
16237 (cortex_r4_alu_shift): Likewise.
16238 (cortex_r4_alu_shift_reg): Likewise.
16239 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
16240 (526_alu_shift_op): Likewise.
16241 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
16242 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
16243 (626te_alu_shift_op): Likewise.
16244 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
16245 (726te_alu_op): Likewise.
16246 (726te_alu_shift_op): Likewise.
16247 (726te_alu_shift_reg_op): Likewise.
16248 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
16249 (mp626_alu_shift_op): Likewise.
16250 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
16251 (pj4_alu_e1_conds): Likewise.
16252 (pj4_alu): Likewise.
16253 (pj4_alu_conds): Likewise.
16254 (pj4_shift): Likewise.
16255 (pj4_shift_conds): Likewise.
16256 (pj4_alu_shift): Likewise.
16257 (pj4_alu_shift_conds): Likewise.
16258 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
16260 (cortexa7_older_only): Likewise.
16261 (cortexa7_younger): Likewise.
16263 2013-07-18 David Malcolm <dmalcolm@redhat.com>
16265 * ipa-pure-const.c (generate_summary): Rename to...
16266 (pure_const_generate_summary): ... this.
16268 2013-07-17 Iain Sandoe <iain@codesourcery.com>
16270 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
16272 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
16275 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
16278 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16280 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
16281 enabled without -march=zEC12.
16282 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
16285 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16287 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
16288 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
16289 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
16290 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
16291 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
16292 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
16293 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
16295 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
16296 and ISA_HAS_NMADD3_NMSUB3.
16297 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
16298 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
16299 (nmsub4<mode>, nmsub3<mode>): Likewise.
16300 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
16302 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16304 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
16305 TARGET_MIPS5400 checking.
16307 2013-07-16 Jakub Jelinek <jakub@redhat.com>
16308 Peter Bergner <bergner@vnet.ibm.com>
16310 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
16311 registers in the comment.
16312 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
16313 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
16314 rather than FIRST_PSEUDO_REGISTERS.
16316 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
16318 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
16319 enable extra ISA flags with TARGET_HTM.
16321 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
16323 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
16326 2013-07-15 Cong Hou <congh@google.com>
16328 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
16329 in compare function for sorting.
16331 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
16333 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
16334 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
16335 * config/rs6000/rs6000.opt: Add -mhtm option.
16336 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
16337 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
16338 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
16339 __HTM__ if the HTM instructions are available.
16340 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
16341 htm_spr_reg_operand): New define_predicates.
16342 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
16343 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
16345 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
16346 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
16347 HTM builtin functions.
16348 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
16349 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
16350 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
16351 (rs6000_builtin_mask_calculate): Likewise.
16352 (rs6000_option_override_internal): Likewise.
16353 (bdesc_htm): Add new HTM builtin support.
16354 (htm_spr_num): New function.
16355 (htm_spr_regno): Likewise.
16356 (rs6000_htm_spr_icode): Likewise.
16357 (htm_expand_builtin): Likewise.
16358 (htm_init_builtins): Likewise.
16359 (rs6000_expand_builtin): Add support for HTM builtin functions.
16360 (rs6000_init_builtins): Likewise.
16361 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
16363 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
16364 (TARGET_HTM, MASK_HTM): Define macros.
16365 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
16366 (FIXED_REGISTERS): Likewise.
16367 (CALL_USED_REGISTERS): Likewise.
16368 (CALL_REALLY_USED_REGISTERS): Likewise.
16369 (REG_ALLOC_ORDER): Likewise.
16370 (enum reg_class): Likewise.
16371 (REG_CLASS_NAMES): Likewise.
16372 (REG_CLASS_CONTENTS): Likewise.
16373 (REGISTER_NAMES): Likewise.
16374 (ADDITIONAL_REGISTER_NAMES): Likewise.
16375 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
16376 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
16377 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
16378 * config/rs6000/htm.md: New file.
16379 * config/rs6000/htmintrin.h: New file.
16380 * config/rs6000/htmxlintrin.h: New file.
16382 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
16384 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
16385 Define SYMBOL_TINY_GOT, update comment.
16386 * config/aarch64/aarch64.c
16387 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
16388 (aarch64_expand_mov_immediate): Likewise.
16389 (aarch64_print_operand): Likewise.
16390 (aarch64_classify_symbol): Likewise.
16391 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
16392 (ldr_got_tiny): Define.
16394 2013-07-13 Tobias Grosser <tobias@grosser.es>
16396 PR tree-optimization/54094
16397 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
16398 scheduling dimension for the parallelism check from the polyhedral
16399 information in the AST.
16400 * graphite-dependences.c (carries_deps): Do not assume the schedule is
16403 2013-07-13 Jason Merrill <jason@redhat.com>
16405 * print-tree.c (debug_vec_tree): Use debug_raw.
16406 (debug_raw (vec<tree, va_gc> &)): New.
16407 (debug_raw (vec<tree, va_gc> *)): New.
16408 * tree.h: Declare them.
16410 2013-07-13 Bin Cheng <bin.cheng@arm.com>
16412 * ifcvt.c (ifcvt_after_combine): New static variable.
16413 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
16415 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
16416 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
16417 rest_of_handle_if_after_reload): Pass new argument for if_convert.
16419 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
16421 * config/mips/mips.c (mips_expand_call): Remove empty statement.
16423 2013-07-12 Michael Matz <matz@suse.de>
16425 PR middle-end/55771
16426 * convert.c (convert_to_real): Reject non-float inner types.
16428 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
16430 * config/aarch64/aarch64-protos.h
16431 (aarch64_simd_immediate_valid_for_move): Remove.
16432 * config/aarch64/aarch64.c (simd_immediate_info): New member.
16433 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
16435 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
16437 2013-07-11 Steve Ellcey <sellcey@mips.com>
16439 * config/mips/mips.c (mips_conditional_register_usage): Do not
16440 use t[0-7] registers in MIPS16 mode when optimizing for size.
16442 2013-07-11 Sriraman Tallam <tmsriram@google.com>
16444 * config/i386/i386.c (dispatch_function_versions): Fix array
16445 indexing of function_version_info to match actual_versions.
16447 2013-07-11 Teresa Johnson <tejohnson@google.com>
16449 * vec.h (struct va_gc): Move release out-of-line.
16450 (va_gc::release): Call ggc_free on released vec.
16452 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
16454 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
16455 Require GOT register as additional operand in UNSPEC.
16456 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
16457 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16458 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
16459 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
16460 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
16461 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
16462 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
16464 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
16467 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
16468 name seen by assembler/linker rather if available.
16470 2013-07-11 Andreas Schwab <schwab@suse.de>
16472 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
16474 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
16476 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
16478 2013-07-10 Joseph Myers <joseph@codesourcery.com>
16480 * doc/tm.texi.in: Move hook documentation to ....
16481 * target.def: ... here.
16483 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
16484 text on @hook line.
16485 * doc/tm.texi: Regenerate.
16487 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
16490 * doc/invoke.texi: Document -Wconditionally-supported.
16492 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16495 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
16498 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16501 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
16502 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
16503 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
16504 Remove duplicate devices.
16505 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
16506 * config/avr/t-multilib: Regenerate.
16507 * config/avr/avr-tables.opt: Regenerate.
16508 * doc/avr-mmcu.texi: Regenerate.
16510 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
16513 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
16515 2013-07-10 Graham Stott <graham.stott@btinternet.com>
16517 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
16518 the cost of MULT when optimizing for size.
16520 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
16522 * config/cr16/cr16-protos.h: Don't include target.h.
16524 2013-07-09 Joseph Myers <joseph@codesourcery.com>
16526 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
16527 adjust register size for TDmode and TFmode for VSX registers.
16529 2013-07-08 Kai Tietz <ktietz@redhat.com>
16532 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
16533 hook_bool_const_tree_true.
16535 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16537 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
16538 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
16539 * config/s390/s390.md: Define FPR*_REGNUM constants.
16540 Fix FPR2_REGNUM constant (18 -> 17).
16541 ("*trunc<BFP:mode><DFP_ALL:mode>2")
16542 ("*trunc<DFP_ALL:mode><BFP:mode>2")
16543 ("trunc<BFP:mode><DFP_ALL:mode>2")
16544 ("trunc<DFP_ALL:mode><BFP:mode>2")
16545 ("*extend<BFP:mode><DFP_ALL:mode>2")
16546 ("*extend<DFP_ALL:mode><BFP:mode>2")
16547 ("extend<BFP:mode><DFP_ALL:mode>2")
16548 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
16551 2013-07-08 Graham Stott <graham.stott@btinternet.com>
16553 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
16555 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16557 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
16558 and cfun_fpr_bit_p to cfun_fpr_save_p.
16559 (s390_frame_area, s390_register_info, s390_frame_info)
16560 (s390_emit_prologue, s390_emit_epilogue)
16561 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
16563 * config/s390/s390.h: Define *_REGNUM macros for floating point
16566 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
16568 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
16570 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
16572 PR rtl-optimization/57786
16573 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
16574 and break out of the loop when it is set to false.
16576 2013-07-08 Jakub Jelinek <jakub@redhat.com>
16579 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
16580 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
16581 (const_int 63)) 0)).
16582 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
16583 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
16584 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
16586 PR rtl-optimization/57829
16587 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
16588 mask bits outside of mode are just sign-extension from mode to HWI.
16590 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
16592 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
16593 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
16594 adjust_address instead of change_address to keep info about alignment.
16595 (emit_strmov): Remove.
16596 (emit_memmov): New function.
16597 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
16598 (expand_movmem_epilogue): Likewise and return updated rtx for
16600 (expand_constant_movmem_prologue): Likewise and return updated rtx for
16601 destination and source.
16602 (decide_alignment): Refactor, handle vector_loop.
16603 (ix86_expand_movmem): Likewise.
16604 (ix86_expand_setmem): Likewise.
16605 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
16607 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
16609 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
16610 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
16612 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16614 * config/i386/sse.md (sse_movlhps): Change alternative 3
16615 of operand 2 to "m".
16617 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
16620 * config/i386/sse.md (iptr): New mode attribute.
16621 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
16622 (<sse>_vm<plusminus_insn><mode>3): Ditto.
16623 (<sse>_vmmul<mode>3): Ditto.
16624 (<sse>_vmdiv<mode>3): Ditto.
16625 (sse_vmrcpv4sf2): Ditto.
16626 (<sse>_vmsqrt<mode>2): Ditto.
16627 (sse_vmrsqrtv4sf2): Ditto.
16628 (<sse>_vm<code><mode>3): Ditto.
16629 (avx_vmcmp<mode>3): Ditto.
16630 (<sse>_vmmaskcmp<mode>3): Ditto.
16631 (<sse>_comi): Ditto.
16632 (<sse>_ucomi): Ditto.
16633 (*xop_vmfrcz_<mode>): Ditto.
16634 (*fmai_fmadd_<mode>): Ditto.
16635 (*fmai_fmsub_<mode>): Ditto.
16636 (*fmai_fnmadd_<mode>): Ditto.
16637 (*fmai_fnmsub_<mode>): Ditto.
16638 (*fma4i_vmfmadd_<mode>): Ditto.
16639 (*fma4i_vmfmsub_<mode>): Ditto.
16640 (*fma4i_vmfnmadd_<mode>): Ditto.
16641 (*fma4i_vmfnmsub_<mode>): Ditto.
16642 (*xop_vmfrcz_<mode>): Ditto.
16643 (sse_cvtps2pi): Ditto.
16644 (sse_cvttps2pi): Ditto.
16645 (sse_cvtss2si): Ditto.
16646 (sse_cvtss2si_2): Ditto.
16647 (sse_cvtss2siq_2): Ditto.
16648 (sse_cvttss2si): Ditto.
16649 (sse_cvttss2siq): Ditto.
16650 (sse_cvtsd2si): Ditto.
16651 (sse_cvtsd2si_2): Ditto.
16652 (sse_cvtsd2siq_2): Ditto.
16653 (sse_cvttsd2si): Ditto.
16654 (sse_cvttsd2siq): Ditto.
16655 (sse_cvtsd2ss): Ditto.
16656 (sse_cvtss2sd): Ditto.
16657 (avx2_pbroadcast<mode>): Ditto.
16658 (avx2_pbroadcast<mode>_1): Ditto.
16659 (*avx_vperm_broadcast_v4sf): Ditto.
16661 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
16662 (sse_movlhps): Ditto.
16663 (sse_storehps): Ditto.
16664 (sse_loadhps): Ditto.
16665 (sse_storelps): Ditto.
16666 (sse_loadlps): Ditto.
16667 (*vec_concatv4sf): Ditto.
16668 (*vec_interleave_highv2df): Ditto.
16669 (*vec_interleave_lowv2df): Ditto.
16670 (*vec_extractv2df_1_sse): Ditto.
16671 (*vec_extractv2df_0_sse): Ditto.
16672 (sse2_storelpd): Ditto.
16673 (sse2_loadlpd): Ditto.
16674 (sse2_movsd): Ditto.
16675 (*vec_concatv4si): Ditto.
16676 (vec_concatv2di): Ditto.
16678 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
16679 for Intel asm dialect.
16680 (mmx_punpcklwd): Ditto.
16681 (mmx_punpckldq): Ditto.
16683 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
16684 for intel assembler dialect.
16686 2013-07-06 Jakub Jelinek <jakub@redhat.com>
16689 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
16690 for BUILT_IN_C{LZ,LRSB}*.
16691 * tree.h (CASE_INT_FN): Add FN##IMAX case.
16692 * tree-vrp.c (extract_range_basic): Handle
16693 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
16694 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
16695 fall thru to code calling set_value*.
16696 * builtins.c (expand_builtin): Remove *IMAX cases.
16697 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
16698 if width is bigger than 2*HWI.
16700 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
16702 PR rtl-optimization/55342
16703 * lra-int.h (lra_subreg_reload_pseudos): New.
16704 * lra.c: Add undoing optional reloads to the block diagram.
16705 (lra_subreg_reload_pseudos): New.
16706 (lra_optional_reload_pseudos): Change comments.
16707 (lra): Init and clear lra_subreg_reload_pseudos. Clear
16708 lra_optional_reload_pseudos after undo transformations.
16709 * lra-assigns.c (pseudo_prefix_title): New.
16710 (lra_setup_reg_renumber): Use it.
16711 (spill_for): Ditto. Check subreg reload pseudos too.
16712 (assign_by_spills): Consider subreg reload pseudos too.
16713 * lra-constraints.c (simplify_operand_subreg): Use
16714 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
16715 (curr_insn_transform): Recognize and do optional reloads.
16716 (undo_optional_reloads): New.
16717 (lra_undo_inheritance): Call undo_optional_reloads.
16719 2013-07-05 Thomas Quinot <quinot@adacore.com>
16721 * tree-complex.c (expand_complex_operations_1): Fix typo.
16723 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
16725 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
16726 (tune_params): New member 'const vec_costs'.
16727 * config/aarch64/aarch64.c (generic_vector_cost): New.
16728 (generic_tunings): New member 'generic_vector_cost'.
16729 (aarch64_builtin_vectorization_cost): New.
16730 (aarch64_add_stmt_cost): New.
16731 (TARGET_VECTORIZE_ADD_STMT_COST): New.
16732 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
16734 2013-07-03 Jakub Jelinek <jakub@redhat.com>
16737 * config/i386/predicates.md (vsib_address_operand): Disallow
16738 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
16740 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
16742 PR middle-end/55030
16743 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
16744 expand_builtin_setjmp_receiver.
16745 (expand_label): Adjust, call expand_builtin_setjmp_receiver
16746 with NULL for the label parameter.
16747 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
16748 the frame-pointer. Adjust comments.
16749 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
16750 only if LABEL is non-NULL.
16752 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
16754 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
16755 (ARM_ABI_AAPCS64): Ditto.
16757 (ARM_DEFAULT_ABI): Ditto.
16759 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
16761 * config/aarch64/aarch64-builtins.c
16762 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
16763 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
16765 * config/aarch64/aarch64-simd.md
16766 (aarch64_ld1<VALL:mode>): New.
16767 (aarch64_st1<VALL:mode>): Likewise.
16768 * config/aarch64/arm_neon.h
16769 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
16771 2013-07-02 Sriraman Tallam <tmsriram@google.com>
16773 * config/i386/i386.c (gate_insert_vzeroupper): Check if
16775 (ix86_option_override_internal):Turn on all -mavx target flags by
16776 default as they are dependent on AVX anyway.
16778 2013-07-02 Cary Coutant <ccoutant@google.com>
16780 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
16781 deterministic hash.
16782 (loc_checksum_ordered): Likewise.
16783 (hash_loc_operands): Remove inline keyword.
16785 2013-07-02 Jakub Jelinek <jakub@redhat.com>
16787 PR tree-optimization/57741
16788 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
16789 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
16790 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
16791 Allow REAL_CST step_exprs if flag_associative_math.
16792 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
16794 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16796 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
16798 2013-07-02 Ian Bolton <ian.bolton@arm.com>
16800 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
16802 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16804 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
16806 (iorsi3_insn): Likewise.
16807 (arm_xorsi3): Likewise.
16809 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
16811 * arm.md (attribute "wtype"): Delete. Move attribute values from here
16813 (attribute "type"): ... here, and prefix with "wmmx_".
16814 (attribute "core_cycles"): Update for attribute changes.
16815 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
16816 (tbcstv4hi): Likewise.
16817 (tbcstv2si): Likewise.
16818 (iwmmxt_iordi3): Likewise.
16819 (iwmmxt_xordi3): Likewise.
16820 (iwmmxt_anddi3): Likewise.
16821 (iwmmxt_nanddi3): Likewise.
16822 (iwmmxt_arm_movdi): Likewise.
16823 (iwmmxt_movsi_insn): Likewise.
16824 (mov<mode>_internal): Likewise.
16825 (and<mode>3_iwmmxt): Likewise.
16826 (ior<mode>3_iwmmxt): Likewise.
16827 (xor<mode>3_iwmmxt): Likewise.
16828 (add<mode>3_iwmmxt): Likewise.
16829 (ssaddv8qi3): Likewise.
16830 (ssaddv4hi3): Likewise.
16831 (ssaddv2si3): Likewise.
16832 (usaddv8qi3): Likewise.
16833 (usaddv4hi3): Likewise.
16834 (usaddv2si3): Likewise.
16835 (sub<mode>3_iwmmxt): Likewise.
16836 (sssubv8qi3): Likewise.
16837 (sssubv4hi3): Likewise.
16838 (sssubv2si3): Likewise.
16839 (ussubv8qi3): Likewise.
16840 (ussubv4hi3): Likewise.
16841 (ussubv2si3): Likewise.
16842 (mulv4hi3_iwmmxt): Likewise.
16843 (smulv4hi3_highpart): Likewise.
16844 (umulv4hi3_highpart): Likewise.
16845 (iwmmxt_wmacs): Likewise.
16846 (iwmmxt_wmacsz): Likewise.
16847 (iwmmxt_wmacu): Likewise.
16848 (iwmmxt_wmacuz): Likewise.
16849 (iwmmxt_clrdi): Likewise.
16850 (iwmmxt_clrv8qi): Likewise.
16851 (iwmmxt_clr4hi): Likewise.
16852 (iwmmxt_clr2si): Likewise.
16853 (iwmmxt_uavgrndv8qi3): Likewise.
16854 (iwmmxt_uavgrndv4hi3): Likewise.
16855 (iwmmxt_uavgv8qi3): Likewise.
16856 (iwmmxt_uavgv4hi3): Likewise.
16857 (iwmmxt_tinsrb): Likewise.
16858 (iwmmxt_tinsrh): Likewise.
16859 (iwmmxt_tinsrw): Likewise.
16860 (iwmmxt_textrmub): Likewise.
16861 (iwmmxt_textrmsb): Likewise.
16862 (iwmmxt_textrmuh): Likewise.
16863 (iwmmxt_textrmsh): Likewise.
16864 (iwmmxt_textrmw): Likewise.
16865 (iwmxxt_wshufh): Likewise.
16866 (eqv8qi3): Likewise.
16867 (eqv4hi3): Likewise.
16868 (eqv2si3): Likewise.
16869 (gtuv8qi3): Likewise.
16870 (gtuv4hi3): Likewise.
16871 (gtuv2si3): Likewise.
16872 (gtv8qi3): Likewise.
16873 (gtv4hi3): Likewise.
16874 (gtv2si3): Likewise.
16875 (smax<mode>3_iwmmxt): Likewise.
16876 (umax<mode>3_iwmmxt): Likewise.
16877 (smin<mode>3_iwmmxt): Likewise.
16878 (umin<mode>3_iwmmxt): Likewise.
16879 (iwmmxt_wpackhss): Likewise.
16880 (iwmmxt_wpackwss): Likewise.
16881 (iwmmxt_wpackdss): Likewise.
16882 (iwmmxt_wpackhus): Likewise.
16883 (iwmmxt_wpackwus): Likewise.
16884 (iwmmxt_wpackdus): Likewise.
16885 (iwmmxt_wunpckihb): Likewise.
16886 (iwmmxt_wunpckihh): Likewise.
16887 (iwmmxt_wunpckihw): Likewise.
16888 (iwmmxt_wunpckilb): Likewise.
16889 (iwmmxt_wunpckilh): Likewise.
16890 (iwmmxt_wunpckilw): Likewise.
16891 (iwmmxt_wunpckehub): Likewise.
16892 (iwmmxt_wunpckehuh): Likewise.
16893 (iwmmxt_wunpckehuw): Likewise.
16894 (iwmmxt_wunpckehsb): Likewise.
16895 (iwmmxt_wunpckehsh): Likewise.
16896 (iwmmxt_wunpckehsw): Likewise.
16897 (iwmmxt_wunpckelub): Likewise.
16898 (iwmmxt_wunpckeluh): Likewise.
16899 (iwmmxt_wunpckeluw): Likewise.
16900 (iwmmxt_wunpckelsb): Likewise.
16901 (iwmmxt_wunpckelsh): Likewise.
16902 (iwmmxt_wunpckelsw): Likewise.
16903 (ror<mode>3): Likewise.
16904 (ashr<mode>3_iwmmxt): Likewise.
16905 (lshr<mode>3_iwmmxt): Likewise.
16906 (ashl<mode>3_iwmmxt): Likewise.
16907 (ror<mode>3_di): Likewise.
16908 (ashr<mode>3_di): Likewise.
16909 (lshr<mode>3_di): Likewise.
16910 (ashl<mode>3_di): Likewise.
16911 (iwmmxt_wmadds): Likewise.
16912 (iwmmxt_wmaddu): Likewise.
16913 (iwmmxt_tmia): Likewise.
16914 (iwmmxt_tmiaph): Likewise.
16915 (iwmmxt_tmiabb): Likewise.
16916 (iwmmxt_tmiatb): Likewise.
16917 (iwmmxt_tmiabt): Likewise.
16918 (iwmmxt_tmiatt): Likewise.
16919 (iwmmxt_tmovmskb): Likewise.
16920 (iwmmxt_tmovmskh): Likewise.
16921 (iwmmxt_tmovmskw): Likewise.
16922 (iwmmxt_waccb): Likewise.
16923 (iwmmxt_wacch): Likewise.
16924 (iwmmxt_waccw): Likewise.
16925 (iwmmxt_waligni): Likewise.
16926 (iwmmxt_walignr): Likewise.
16927 (iwmmxt_walignr0): Likewise.
16928 (iwmmxt_walignr1): Likewise.
16929 (iwmmxt_walignr2): Likewise.
16930 (iwmmxt_walignr3): Likewise.
16931 (iwmmxt_wsadb): Likewise.
16932 (iwmmxt_wsadh): Likewise.
16933 (iwmmxt_wsadbz): Likewise.
16934 (iwmmxt_wsadhz): Likewise.
16935 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
16936 (iwmmxt_wabsdiffb): Likewise.
16937 (iwmmxt_wabsdiffh): Likewise.
16938 (iwmmxt_wabsdiffw): Likewise.
16939 (iwmmxt_waddsubhx): Likewise
16940 (iwmmxt_wsubaddhx): Likewise.
16941 (addc<mode>3): Likewise.
16942 (iwmmxt_avg4): Likewise.
16943 (iwmmxt_avg4r): Likewise.
16944 (iwmmxt_wmaddsx): Likewise.
16945 (iwmmxt_wmaddux): Likewise.
16946 (iwmmxt_wmaddsn): Likewise.
16947 (iwmmxt_wmaddun): Likewise.
16948 (iwmmxt_wmulwsm): Likewise.
16949 (iwmmxt_wmulwum): Likewise.
16950 (iwmmxt_wmulsmr): Likewise.
16951 (iwmmxt_wmulumr): Likewise.
16952 (iwmmxt_wmulwsmr): Likewise.
16953 (iwmmxt_wmulwumr): Likewise.
16954 (iwmmxt_wmulwl): Likewise.
16955 (iwmmxt_wqmulm): Likewise.
16956 (iwmmxt_wqmulwm): Likewise.
16957 (iwmmxt_wqmulmr): Likewise.
16958 (iwmmxt_wqmulwmr): Likewise.
16959 (iwmmxt_waddbhusm): Likewise.
16960 (iwmmxt_waddbhusl): Likewise.
16961 (iwmmxt_wqmiabb): Likewise.
16962 (iwmmxt_wqmiabt): Likewise.
16963 (iwmmxt_wqmiatb): Likewise.
16964 (iwmmxt_wqmiatt): Likewise.
16965 (iwmmxt_wqmiabbn): Likewise.
16966 (iwmmxt_wqmiabtn): Likewise.
16967 (iwmmxt_wqmiatbn): Likewise.
16968 (iwmmxt_wqmiattn): Likewise.
16969 (iwmmxt_wmiabb): Likewise.
16970 (iwmmxt_wmiabt): Likewise.
16971 (iwmmxt_wmiatb): Likewise.
16972 (iwmmxt_wmiatt): Likewise.
16973 (iwmmxt_wmiabbn): Likewise.
16974 (iwmmxt_wmiabtn): Likewise.
16975 (iwmmxt_wmiatbn): Likewise.
16976 (iwmmxt_wmiattn): Likewise.
16977 (iwmmxt_wmiawbb): Likewise.
16978 (iwmmxt_wmiawbt): Likewise.
16979 (iwmmxt_wmiawtb): Likewise.
16980 (iwmmxt_wmiawtt): Likewise.
16981 (iwmmxt_wmiawbbn): Likewise.
16982 (iwmmxt_wmiawbtn): Likewise.
16983 (iwmmxt_wmiawtbn): Likewise.
16984 (iwmmxt_wmiawttn): Likewise.
16985 (iwmmxt_wmerge): Likewise.
16986 (iwmmxt_tandc<mode>3): Likewise.
16987 (iwmmxt_torc<mode>3): Likewise.
16988 (iwmmxt_torvsc<mode>3): Likewise.
16989 (iwmmxt_textrc<mode>3): Likewise.
16990 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
16991 (wmmxt_pack): Likewise.
16992 (wmmxt_mult_c1): Likewise.
16993 (wmmxt_mult_c2): Likewise.
16994 (wmmxt_alu_c1): Likewise.
16995 (wmmxt_alu_c2): Likewise.
16996 (wmmxt_alu_c3): Likewise.
16997 (wmmxt_transfer_c1): Likewise.
16998 (wmmxt_transfer_c2): Likewise.
16999 (wmmxt_transfer_c3): Likewise.
17000 (marvell_f_iwmmxt_wstr): Likewise.
17001 (marvell_f_iwmmxt_wldr): Likewise.
17003 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
17005 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
17007 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17010 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17011 * lra-constraints.c (need_for_split_p): Check call used hard regs
17012 living through calls.
17014 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
17015 call used regs for call insn.
17017 2013-06-28 Jakub Jelinek <jakub@redhat.com>
17020 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
17021 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
17023 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
17025 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
17027 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
17029 * lra-constraints.c (need_for_split_p): Check call used hard regs
17030 living through calls.
17032 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
17035 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
17036 to tie with any other modes. Eliminate Altivec vector mode tests,
17037 since these are a subset of ALTIVEC or VSX vector modes. Simplify
17038 code, to return 0 if testing MODE2 for a condition, if we've
17039 already tested MODE1 for the same condition.
17041 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17043 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
17046 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17048 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
17049 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
17051 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
17053 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
17055 (aarch64_symbolic_constant_p): Remove.
17056 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
17057 static. Fix line length and white space.
17058 (aarch64_symbolic_constant_p): Remove.
17059 * config/aarch64/predicates.md (aarch64_valid_symref):
17060 Use aarch64_classify_symbol_expression.
17062 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17064 * config/arm/constraints.md (Ts): New constraint.
17065 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
17067 (compare_scc): Use "Ts" constraint for operand 0.
17068 (ior_scc_scc): Likewise.
17069 (and_scc_scc): Likewise.
17070 (and_scc_scc_nodom): Likewise.
17071 (ior_scc_scc_cmp): Likewise for operand 7.
17072 (and_scc_scc_cmp): Likewise.
17073 * config/arm/thumb2.md (thumb2_movsi_insn):
17074 Add alternatives for 16-bit encodings.
17075 (thumb2_movhi_insn): Likewise.
17076 (thumb2_movsicc_insn): Likewise.
17077 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
17078 (thumb2_negscc): Use "Ts" constraint.
17079 Move mvn instruction outside cond_exec block.
17080 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
17081 for 16-bit encodings.
17083 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17085 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
17087 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
17088 (mulsi3subsi): Likewise.
17089 (mulsidi3adddi): Likewise.
17090 (mulsidi3_v6): Likewise.
17091 (umulsidi3_v6): Likewise.
17092 (umulsidi3adddi_v6): Likewise.
17093 (smulsi3_highpart_v6): Likewise.
17094 (umulsi3_highpart_v6): Likewise.
17095 (mulhisi3tb): Likewise.
17096 (mulhisi3bt): Likewise.
17097 (mulhisi3tt): Likewise.
17098 (maddhisi4): Likewise.
17099 (maddhisi4tb): Likewise.
17100 (maddhisi4tt): Likewise.
17101 (maddhidi4): Likewise.
17102 (maddhidi4tb): Likewise.
17103 (maddhidi4tt): Likewise.
17104 (zeroextractsi_compare0_scratch): Likewise.
17105 (insv_zero): Likewise.
17106 (insv_t2): Likewise.
17107 (anddi_notzesidi_di): Likewise.
17108 (anddi_notsesidi_di): Likewise.
17109 (andsi_notsi_si): Likewise.
17110 (iordi_zesidi_di): Likewise.
17111 (xordi_zesidi_di): Likewise.
17112 (andsi_iorsi3_notsi): Likewise.
17113 (smax_0): Likewise.
17114 (smax_m1): Likewise.
17115 (smin_0): Likewise.
17116 (not_shiftsi): Likewise.
17117 (unaligned_loadsi): Likewise.
17118 (unaligned_loadhis): Likewise.
17119 (unaligned_loadhiu): Likewise.
17120 (unaligned_storesi): Likewise.
17121 (unaligned_storehi): Likewise.
17122 (extv_reg): Likewise.
17123 (extzv_t2): Likewise.
17124 (divsi3): Likewise.
17125 (udivsi3): Likewise.
17126 (arm_zero_extendhisi2addsi): Likewise.
17127 (arm_zero_extendqisi2addsi): Likewise.
17128 (compareqi_eq0): Likewise.
17129 (arm_extendhisi2_v6): Likewise.
17130 (arm_extendqisi2addsi): Likewise.
17131 (arm_movt): Likewise.
17132 (thumb2_ldrd): Likewise.
17133 (thumb2_ldrd_base): Likewise.
17134 (thumb2_ldrd_base_neg): Likewise.
17135 (thumb2_strd): Likewise.
17136 (thumb2_strd_base): Likewise.
17137 (thumb2_strd_base_neg): Likewise.
17138 (arm_negsi2): Add alternative for 16-bit encoding.
17139 (arm_one_cmplsi2): Likewise.
17141 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17143 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
17144 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
17145 (movdfcc): Likewise.
17146 * config/arm/vfp.md (*thumb2_movsf_vfp):
17147 Disable predication for arm_restrict_it.
17148 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
17149 (*thumb2_movdfcc_vfp): Likewise.
17150 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
17151 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
17152 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
17153 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
17154 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
17155 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
17156 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
17157 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
17158 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
17159 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
17160 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
17161 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
17162 Disable predication for arm_restrict_it.
17164 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
17166 * config/i386/bmiintrin.h (_bextr_u32): New.
17167 (_bextr_u64): Ditto.
17169 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
17171 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
17172 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
17173 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
17174 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
17175 * config/mips/n32-elf.h: ...this new file.
17177 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
17180 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
17181 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
17183 2013-06-27 Catherine Moore <clm@codesourcery.com>
17185 * config/mips/mips-tables.opt: Regenerate.
17186 * config/mips/mips-cpus.def: Add m14ke and m14kec.
17187 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
17188 * doc/invoke.texi: Add -m14kc.
17190 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17193 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
17194 constraints of operand 1 and 2.
17197 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
17198 to match RTL canonicalization. Swap predicates and
17199 constraints of operand 1 and 2.
17201 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
17203 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
17204 Process OP_INOUT regs for splitting too.
17206 2013-06-27 Jakub Jelinek <jakub@redhat.com>
17208 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
17209 decl before the loop, initialize to NULL.
17210 (vectorizable_load): Initialize ptr_incr to NULL.
17212 2013-06-27 Martin Jambor <mjambor@suse.cz>
17215 * ipa-ref.h (ipa_maybe_record_reference): Declare.
17216 * ipa-ref.c (ipa_maybe_record_reference): New function.
17217 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
17218 * ipa-cp.c (create_specialized_node): Record potential references from
17220 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
17222 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17224 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
17225 parameter 'mode' of type 'enum machine_mode mode'; change to pass
17226 'mode' to force_reg.
17227 (aarch64_add_offset): Update calls to aarch64_force_temporary.
17228 (aarch64_expand_mov_immediate): Likewise.
17230 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
17232 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
17233 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
17235 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17237 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
17238 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17239 (struct machine_function): Add tbegin_p.
17240 (s390_canonicalize_comparison): Fold CC mode compares to
17241 conditional jump if possible.
17242 (s390_emit_jump): Return the emitted jump.
17243 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
17244 Handle CCRAWmode compares.
17245 (s390_option_override): Default to -mhtm if available.
17246 (s390_reg_clobbered_rtx): Handle floating point regs as well.
17247 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
17248 FPRs instead of df_regs_ever_live_p.
17249 (s390_optimize_nonescaping_tx): New function.
17250 (s390_init_frame_layout): Extend clobbered_regs array to cover
17252 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
17253 (s390_expand_tbegin): New function.
17254 (enum s390_builtin): New enum definition.
17255 (code_for_builtin): New array definition.
17256 (s390_init_builtins): New function.
17257 (s390_expand_builtin): New function.
17258 (TARGET_INIT_BUILTINS): Define.
17259 (TARGET_EXPAND_BUILTIN): Define.
17260 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
17261 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
17262 (s390_alc_comparison): Likewise.
17263 * config/s390/s390-modes.def: Add CCRAWmode.
17264 * config/s390/s390.h (processor_flags): Add PF_TX.
17265 (TARGET_CPU_HTM): Define macro.
17266 (TARGET_HTM): Define macro.
17267 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
17268 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
17269 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
17270 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
17271 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
17272 (TBEGIN_MASK, TBEGINC_MASK): New constants.
17273 ("*cc_to_int"): Move up.
17274 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
17275 constants other than 0.
17276 ("*ccraw_to_int"): New insn and splitter definition.
17277 ("tbegin", "tbegin_nofloat", "tbegin_retry")
17278 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
17279 ("tx_assist"): New expander.
17280 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
17281 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
17282 * config/s390/s390.opt: Add -mhtm option.
17283 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
17284 * config/s390/htmxlintrin.h: New file.
17285 * config/s390/htmintrin.h: New file.
17286 * config/s390/s390intrin.h: New file.
17287 * doc/extend.texi: Document htm builtins.
17288 * config.gcc: Add the new header files to extra_headers.
17290 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
17292 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
17293 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
17295 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
17296 Pat Haugen <pthaugen@us.ibm.com>
17297 Peter Bergner <bergner@vnet.ibm.com>
17299 * config/rs6000/power8.md: New.
17300 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
17301 setting for power8 entry.
17302 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
17303 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
17304 test for Power4/Power5 only.
17305 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
17307 (force_new_group): Adjust comment.
17308 * config/rs6000/rs6000.md: Include power8.md.
17310 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
17312 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
17313 * config/arm/arm-protos.h (arm_max_conditional_execute): New
17315 (tune_params): Update comment.
17316 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
17317 (arm_max_conditional_execute): New function.
17318 (thumb2_final_prescan_insn): Use max_insn_skipped and
17319 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
17321 2013-06-25 Jakub Jelinek <jakub@redhat.com>
17323 PR tree-optimization/57705
17324 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
17325 SSA_NAME step, provided that it is not defined inside the loop.
17326 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
17327 (get_initial_def_for_induction): Handle SSA_NAME IV step.
17329 2013-06-25 Martin Jambor <mjambor@suse.cz>
17331 PR middle-end/57670
17332 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
17333 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
17335 (ipa_note_param_call): Initialize member_ptr flag.
17336 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
17337 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
17338 (ipa_write_indirect_edge_info): Stream member_ptr flag.
17339 (ipa_read_indirect_edge_info): Likewise.
17341 2013-06-25 Richard Biener <rguenther@suse.de>
17343 PR middle-end/56977
17344 * passes.c (init_optimization_passes): Move pass_fold_builtins
17345 and pass_dce earlier with -Og.
17347 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
17349 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
17350 <BIT_FIELD_REF>: Remove trailing TAB.
17351 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
17354 2013-06-24 Martin Jambor <mjambor@suse.cz>
17356 PR tree-optimization/57358
17357 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
17358 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
17359 (ipa_analyze_params_uses): Generate pessimistic info when true.
17361 2013-06-24 Martin Jambor <mjambor@suse.cz>
17363 PR tree-optimization/57539
17364 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
17365 global.inlined_to of the new node to it. All callers changed.
17366 * ipa-inline-transform.c (clone_inlined_nodes): New variable
17367 inlining_into, pass it to cgraph_clone_node.
17368 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
17369 ipa_free_edge_args_substructures.
17370 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
17371 rdesc linked list. Do not assert rdesc edges have inlined caller.
17372 Assert we have found an rdesc in the rdesc list.
17374 2013-06-24 Richard Biener <rguenther@suse.de>
17376 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
17377 (pointer_set_lookup): Declare.
17378 (class pointer_map): New template class implementing a
17379 generic pointer to T map.
17380 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
17381 pointer_map<T>::contains, pointer_map<T>::insert,
17382 pointer_map<T>::traverse): New functions.
17383 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
17384 (pointer_set_lookup): New function.
17385 (pointer_set_contains): Use pointer_set_lookup.
17386 (pointer_set_insert): Likewise.
17387 (insert_aux): Remove.
17388 (struct pointer_map_t): Embed a pointer_set_t.
17389 (pointer_map_create): Adjust.
17390 (pointer_map_destroy): Likewise.
17391 (pointer_map_contains): Likewise.
17392 (pointer_map_insert): Likewise.
17393 (pointer_map_traverse): Likewise.
17394 * tree-streamer.h (struct streamer_tree_cache_d): Use a
17395 pointer_map<unsigned> instead of a pointer_map_t.
17396 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
17397 (streamer_tree_cache_lookup): Likewise.
17398 (streamer_tree_cache_create): Likewise.
17399 (streamer_tree_cache_delete): Likewise.
17400 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
17401 pointer_map<unsigned> instead of a pointer_map_t.
17402 (lto_init_tree_ref_encoder): Adjust.
17403 (lto_destroy_tree_ref_encoder): Likewise.
17404 * lto-section-out.c (lto_output_decl_index): Likewise.
17405 (lto_record_function_out_decl_state): Likewise.
17406 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
17408 2013-06-24 Richard Biener <rguenther@suse.de>
17410 PR tree-optimization/57488
17411 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
17413 2013-06-24 Alan Modra <amodra@gmail.com>
17415 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
17416 (gen_easy_altivec_constant): Likewise.
17417 * config/rs6000/predicates.md (easy_vector_constant_add_self,
17418 easy_vector_constant_msb): Likewise.
17420 2013-06-23 Jakub Jelinek <jakub@redhat.com>
17423 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
17424 add missing return true.
17426 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
17429 * config/sh/predicates.md (general_extend_operand): Invoke
17430 general_movsrc_operand for memory operands.
17431 (general_movsrc_operand): Allow reg+reg addressing, do not use
17432 general_operand for memory operands.
17434 2013-06-23 Sriraman Tallam <tmsriram@google.com>
17436 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
17437 when current target options does not apply.
17438 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
17439 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
17440 * config/i386/bmiintrin.h: Pass appropriate target
17441 attributes to header.
17442 * config/i386/mmintrin.h: Ditto.
17443 * config/i386/nmmintrin.h: Ditto.
17444 * config/i386/avx2intrin.h: Ditto.
17445 * config/i386/fxsrintrin.h: Ditto.
17446 * config/i386/tbmintrin.h: Ditto.
17447 * config/i386/xsaveintrin.h: Ditto.
17448 * config/i386/f16cintrin.h: Ditto.
17449 * config/i386/xtestintrin.h: Ditto.
17450 * config/i386/xsaveoptintrin.h: Ditto.
17451 * config/i386/bmi2intrin.h: Ditto.
17452 * config/i386/lzcntintrin.h: Ditto.
17453 * config/i386/smmintrin.h: Ditto.
17454 * config/i386/wmmintrin.h: Ditto.
17455 * config/i386/x86intrin.h: Remove all header include guards.
17456 * config/i386/prfchwintrin.h: Ditto.
17457 * config/i386/pmmintrin.h: Ditto.
17458 * config/i386/tmmintrin.h: Ditto.
17459 * config/i386/xmmintrin.h: Ditto.
17460 * config/i386/popcntintrin.h: Ditto.
17461 * config/i386/rdseedintrin.h: Ditto.
17462 * config/i386/ammintrin.h: Ditto.
17463 * config/i386/emmintrin.h: Ditto.
17464 * config/i386/immintrin.h: Remove all header include guards.
17465 * config/i386/fma4intrin.h: Ditto.
17466 * config/i386/lwpintrin.h: Ditto.
17467 * config/i386/xopintrin.h: Ditto.
17468 * config/i386/ia32intrin.h: Ditto.
17469 * config/i386/avxintrin.h: Ditto.
17470 * config/i386/rtmintrin.h: Ditto.
17471 * config/i386/fmaintrin.h: Ditto.
17472 * config/i386/mm3dnow.h: Ditto.
17474 2013-06-22 Sriraman Tallam <tmsriram@google.com>
17476 * common/config/i386/i386-common.c: Handle LZCNT.
17478 2013-06-22 Andi Kleen <ak@linux.intel.com>
17480 * doc/extend.texi: Use __atomic_store_n instead of
17481 __atomic_store in HLE example.
17483 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
17485 * config/sh/sh.c: Remove <cstdlib> workaround.
17487 2013-06-21 Andi Kleen <ak@linux.intel.com>
17489 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
17491 2013-06-21 Andi Kleen <ak@linux.intel.com>
17493 * doc/extend.texi: Document that __atomic_clear and
17494 __atomic_test_and_set should only be used with bool.
17496 2013-06-20 Jan Hubicka <jh@suse.cz>
17498 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
17499 types_same_for_odr.
17500 * tree.c (decls_same_for_odr): New function.
17501 (same_for_edr): New function.
17502 (types_same_for_odr): New function.
17503 (get_binfo_at_offset): Use it.
17504 * tree.h (types_same_for_odr): Declare.
17506 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
17507 Jason Merrill <jason@redhat.com>
17509 * system.h: Include <cstdlib> as well as <stdlib.h>.
17511 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
17514 * config/i386/i386.c (construct_container): Report error if
17515 long double is used with disabled x87 float returns.
17517 2013-06-20 Jan Hubicka <jh@suse.cz>
17519 * lto-cgraph.c (input_symtab): Do not set cgraph state.
17521 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
17523 PR rtl-optimization/57425
17524 PR rtl-optimization/57569
17525 * alias.c (write_dependence_p): Remove parameters mem_mode and
17526 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
17527 Changed all callers.
17528 (canon_anti_dependence): Get comments and semantics in sync.
17529 Add parameter mem_canonicalized. Changed all callers.
17530 * rtl.h (canon_anti_dependence): Update prototype.
17532 2013-06-20 Richard Biener <rguenther@suse.de>
17534 * data-streamer-in.c (streamer_read_uhwi): Optimize single
17535 byte case, inline streamer_read_uchar and defer section
17538 2013-06-20 Richard Biener <rguenther@suse.de>
17540 PR tree-optimization/57584
17541 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
17542 SSA names into the expanded expression that take part in
17543 abnormal coalescing.
17545 2013-06-19 Sharad Singhai <singhai@google.com>
17547 * gcov.c (print_usage): Handle new option.
17548 (process_args): Ditto.
17549 (get_gcov_intermediate_filename): New function.
17550 (output_intermediate_file): New function.
17551 (output_gcov_file): New function
17552 (generate_results): Handle new option.
17553 (release_function): Relase demangled name.
17554 (read_graph_file): Handle demangled name.
17555 (output_lines): Ditto.
17556 * doc/gcov.texi: Document gcov intermediate format.
17558 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
17561 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
17562 (lra_emit_add): Use the functions. Add comment about Y as an
17565 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
17568 * collect2.c (collect_atexit): New.
17569 (collect_exit): Delete.
17570 (main): Register collect_atexit with atexit.
17571 (collect_wait): Change collect_exit to exit.
17573 * collect2.h (collect_exit): Delete.
17574 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
17576 2013-06-19 Wei Mi <wmi@google.com>
17578 PR rtl-optimization/57518
17579 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
17580 if regno is used in paradoxical subreg.
17581 (update_equiv_regs): Check pdx_subregs[regno] before
17582 set a reg to be equivalent with a mem.
17584 2013-06-19 Matthias Klose <doko@ubuntu.com>
17587 * file-find.h (find_a_file): Add a mode parameter.
17588 * file-find.c (find_a_file): Likewise.
17589 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
17590 with X_OK for the executables.
17591 * collect2.c (main): Call find_a_file with X_OK.
17593 2013-06-19 Steve Ellcey <sellcey@mips.com>
17596 * config/mips/mips.md (casesi_internal_mips16_<mode>):
17597 Use NEXT_INSN instead of next_real_insn.
17599 2013-06-19 Jan Hubicka <jh@suse.cz>
17601 * cgraph.h (const_value_known_p): Replace by ...
17602 (ctor_for_folding): .. this one.
17603 * cgraphunit.c (process_function_and_variable_attributes): Use it.
17604 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
17605 * expr.c (expand_expr_real_1): Likewise.
17606 (string_constant): Likewise.
17607 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
17608 * ipa.c (process_references): Likewise.
17609 (symtab_remove_unreachable_nodes): Likewise.
17610 * ipa-inline-analysis.c (param_change_prob): Likewise.
17611 * gimple-fold.c (canonicalize_constructor_val): Likewise.
17612 (get_base_constructor): Likwise.
17613 * varpool.c (varpool_remove_node): Likewise.
17614 (varpool_remove_initializer): LIkewise.
17615 (dump_varpool_node): LIkwise.
17616 (const_value_known_p): Rewrite to ...
17617 (ctor_for_folding): ... this one.
17619 2013-06-19 Jakub Jelinek <jakub@redhat.com>
17622 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
17623 PERSONALITY in $PATH derived prefixes.
17625 2013-06-19 Jeff Law <law@redhat.com>
17627 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
17630 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
17631 (simplify_bitwise_binary): Use it to simpify certain binary ops on
17634 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
17636 * config/arm/vfp.md: Move VFP instruction classification documentation
17638 * config/arm/arm.md: ... here. Update instruction classification
17641 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
17643 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
17644 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
17645 pattern. Use more efficient sequences on ARMv5 and Thumb2.
17647 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
17650 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17651 with NEXT_INSN. Use tablejump_p to check for jump table data
17654 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
17657 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
17658 that now in C++ the value is correct per the C++ standards.
17660 2013-06-19 Richard Biener <rguenther@suse.de>
17662 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
17663 for global context.
17665 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17668 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17671 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17672 with next_active_insn.
17674 2013-06-18 Sriraman Tallam <tmsriram@google.com>
17676 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
17677 functions are inlined during failures to flag an error.
17678 * tree-inline.c (expand_call_inline): Allow the error to be flagged
17679 in early inline pass.
17681 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
17683 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
17686 2013-06-18 Julian Brown <julian@codesourcery.com>
17688 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
17689 Permit virtual register pre-reload if !strict.
17690 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
17692 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
17694 * config/arm/neon.md (movmisalign<mode>): Use
17695 neon_perm_struct_or_reg_operand instead of
17696 neon_struct_or_register_operand.
17697 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
17698 neon_permissive_struct_operand instead of neon_struct_operand.
17699 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
17700 neon_vector_mem_operand.
17701 * config/arm/predicates.md (neon_struct_operand): Adjust call to
17702 neon_vector_mem_operand.
17703 (neon_permissive_struct_operand): New.
17704 (neon_struct_or_register_operand): Rename to...
17705 (neon_perm_struct_or_reg_operand): This. Adjust call to
17706 neon_vector_mem_operand.
17708 2013-06-18 Richard Biener <rguenther@suse.de>
17710 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
17711 * lto-streamer.h: Include pointer-set.h.
17712 (struct lto_decl_slot): Remove.
17713 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
17714 Remove next_index entry.
17715 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17716 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17717 (lto_init_tree_ref_encoder): Adjust.
17718 (lto_destroy_tree_ref_encoder): Likewise.
17719 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
17720 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
17721 (lto_output_decl_index): Adjust.
17722 (lto_new_out_decl_state): Likewise.
17723 (lto_record_function_out_decl_state): Likewise.
17724 * lto-streamer-out.c (copy_function): Likewise.
17726 2013-06-18 Richard Biener <rguenther@suse.de>
17728 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
17729 * cgraphunit.c: Include cfgloop.h.
17730 (init_lowered_empty_function): Initialize the loop tree.
17731 (assemble_thunk): Insert new BBs into loops.
17733 2013-06-18 Richard Biener <rguenther@suse.de>
17735 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
17736 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
17737 the map from cache entry to cache index optional.
17738 (streamer_tree_cache_replace_tree): Adjust accordingly.
17739 (streamer_tree_cache_append): Likewise.
17740 (streamer_tree_cache_delete): Likewise.
17741 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
17742 streamer cache map from cache entry to cache index.
17743 * lto-streamer-out.c (create_output_block): Adjust.
17745 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
17747 * config/arm/arm.md (attribute "insn"): Move multiplication and
17748 division attributes to...
17749 (attribute "type"): ... here. Remove mult.
17750 (attribute "mul32"): New attribute.
17751 (attribute "mul64"): Add umaal.
17752 (*arm_mulsi3): Update attributes.
17753 (*arm_mulsi3_v6): Likewise.
17754 (*thumb_mulsi3): Likewise.
17755 (*thumb_mulsi3_v6): Likewise.
17756 (*mulsi3_compare0): Likewise.
17757 (*mulsi3_compare0_v6): Likewise.
17758 (*mulsi_compare0_scratch): Likewise.
17759 (*mulsi_compare0_scratch_v6): Likewise.
17760 (*mulsi3addsi): Likewise.
17761 (*mulsi3addsi_v6): Likewise.
17762 (*mulsi3addsi_compare0): Likewise.
17763 (*mulsi3addsi_compare0_v6): Likewise.
17764 (*mulsi3addsi_compare0_scratch): Likewise.
17765 (*mulsi3addsi_compare0_scratch_v6): Likewise.
17766 (*mulsi3subsi): Likewise.
17767 (*mulsidi3adddi): Likewise.
17768 (*mulsi3addsi_v6): Likewise.
17769 (*mulsidi3adddi_v6): Likewise.
17770 (*mulsidi3_nov6): Likewise.
17771 (*mulsidi3_v6): Likewise.
17772 (*umulsidi3_nov6): Likewise.
17773 (*umulsidi3_v6): Likewise.
17774 (*umulsidi3adddi): Likewise.
17775 (*umulsidi3adddi_v6): Likewise.
17776 (*smulsi3_highpart_nov6): Likewise.
17777 (*smulsi3_highpart_v6): Likewise.
17778 (*umulsi3_highpart_nov6): Likewise.
17779 (*umulsi3_highpart_v6): Likewise.
17780 (mulhisi3): Likewise.
17781 (*mulhisi3tb): Likewise.
17782 (*mulhisi3bt): Likewise.
17783 (*mulhisi3tt): Likewise.
17784 (maddhisi4): Likewise.
17785 (*maddhisi4tb): Likewise.
17786 (*maddhisi4tt): Likewise.
17787 (maddhidi4): Likewise.
17788 (*maddhidi4tb): Likewise.
17789 (*maddhidi4tt): Likewise.
17790 (divsi3): Likewise.
17791 (udivsi3): Likewise.
17792 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
17793 (thumb2_mulsi_short_compare0): Likewise.
17794 (thumb2_mulsi_short_compare0_scratch): Likewise.
17795 * config/arm/arm1020e.md (1020mult1): Update attribute change.
17796 (1020mult2): Likewise.
17797 (1020mult3): Likewise.
17798 (1020mult4): Likewise.
17799 (1020mult5): Likewise.
17800 (1020mult6): Likewise.
17801 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
17803 (cortex_a15_mult64): Likewise.
17804 (cortex_a15_sdiv): Likewise.
17805 (cortex_a15_udiv): Likewise.
17806 * config/arm/arm1026ejs.md (mult1): Update attribute change.
17812 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
17813 (pj4_ir_div): Likewise.
17814 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
17815 (11_mult2): Likewise.
17816 (11_mult3): Likewise.
17817 (11_mult4): Likewise.
17818 (11_mult5): Likewise.
17819 (11_mult6): Likewise.
17820 (11_mult7): Likewise.
17821 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
17822 (cortex_a8_mla): Likewise.
17823 (cortex_a8_mull): Likewise.
17824 (cortex_a8_smulwy): Likewise.
17825 (cortex_a8_smlald): Likewise.
17826 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
17827 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
17828 (cortex_r4_mul_3): Likewise.
17829 (cortex_r4_mla_4): Likewise.
17830 (cortex_r4_mla_3): Likewise.
17831 (cortex_r4_smlald): Likewise.
17832 (cortex_r4_mull): Likewise.
17833 (cortex_r4_sdiv): Likewise.
17834 (cortex_r4_udiv): Likewise.
17835 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
17836 (cortex_a7_idiv): Likewise.
17837 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
17838 (9_mult2): Likewise.
17839 (9_mult3): Likewise.
17840 (9_mult4): Likewise.
17841 (9_mult5): Likewise.
17842 (9_mult6): Likewise.
17843 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
17844 (cortex_a53_sdiv): Likewise.
17845 (cortex_a53_udiv): Likewise.
17846 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
17847 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
17848 (mp626_mult2): Likewise.
17849 (mp626_mult3): Likewise.
17850 (mp626_mult4): Likewise.
17851 * config/arm/fa526.md (526_mult1): Update attribute change.
17852 (526_mult2): Likewise.
17853 * config/arm/arm-generic.md (mult): Update attribute change.
17854 (mult_ldsched_strongarm): Likewise.
17855 (mult_ldsched): Likewise.
17856 (multi_cycle): Likewise.
17857 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
17858 * config/arm/fa606te.md (606te_mult1): Update attribute change.
17859 (606te_mult2): Likewise.
17860 (606te_mult3): Likewise.
17861 (606te_mult4): Likewise.
17862 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
17863 (cortex_a9_mac16): Likewise.
17864 (cortex_a9_multiply): Likewise.
17865 (cortex_a9_mac): Likewise.
17866 (cortex_a9_multiply_long): Likewise.
17867 * config/arm/fa626te.md (626te_mult1): Update attribute change.
17868 (626te_mult2): Likewise.
17869 (626te_mult3): Likewise.
17870 (626te_mult4): Likewise.
17872 2013-06-18 Richard Biener <rguenther@suse.de>
17875 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
17877 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17880 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
17881 with next_active_insn.
17883 2013-06-18 Alan Modra <amodra@gmail.com>
17885 * config/rs6000/rs6000.h (enum data_align): New.
17886 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
17887 (DATA_ABI_ALIGNMENT): Define.
17888 (CONSTANT_ALIGNMENT): Correct comment.
17889 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
17890 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
17892 2013-06-17 David Malcolm <dmalcolm@redhat.com>
17894 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
17895 ATTRIBUTE_UNUSED marking.
17897 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
17899 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
17900 alternative and update.
17901 (aarch64_dup_lanedi): Delete.
17902 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
17903 * config/aarch64/aarch64-simd-builtins.def: Update.
17905 2013-06-17 Richard Biener <rguenther@suse.de>
17907 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
17908 (lto_input_scc): Declare.
17909 (lto_input_tree_1): Likewise.
17910 (struct lto_stats_d): Add num_tree_bodies_output and
17911 num_pickle_refs_output.
17912 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
17913 (lto_read_tree_1): Split out from ...
17914 (lto_read_tree): ... this.
17915 (lto_input_scc): New function.
17916 (lto_input_tree_1): Split out from ...
17917 (lto_input_tree): ... this. Handle LTO_tree_scc.
17918 (lto_data_in_create): Create the streamer cache without hashes.
17919 * lto-streamer-out.c (create_output_block): Create the streamer
17920 cache with hashes when not doing WPA.
17921 (lto_write_tree_1): Split out from ...
17922 (lto_write_tree): ... this.
17923 (get_symbol_initial_value): New function.
17924 (lto_output_tree_1): Split out from ...
17925 (lto_output_tree): ... this. Write trees as series of SCCs
17926 using a DFS walk via DFS_write_tree.
17927 (struct sccs, struct scc_entry): New types.
17928 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
17929 (DFS_write_tree_body): New function.
17930 (DFS_write_tree): Likewise.
17931 (hash_tree): Likewise.
17932 (scc_entry_compare): Likewise.
17933 (hash_scc): Likewise.
17934 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
17935 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
17936 TREE_CHAIN as regular reference.
17937 (streamer_read_integer_cst): Remove.
17938 (streamer_get_pickled_tree): Adjust.
17939 * tree-streamer-out.c (streamer_write_chain): Disable streaming
17940 of DECL_EXTERNALs in BLOCK_VARS for now.
17941 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
17943 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
17944 Add hash value argument and record that if hashes are recorded
17946 (streamer_tree_cache_insert_1): Adjust.
17947 (streamer_tree_cache_insert): Likewise.
17948 (streamer_tree_cache_insert_at): Rename to ...
17949 (streamer_tree_cache_replace_tree): ... this and adjust.
17950 (streamer_tree_cache_append): Adjust.
17951 (record_common_node): Likewise.
17952 (streamer_tree_cache_create): Add argument whether to
17953 record hash values together with trees.
17954 (streamer_tree_cache_delete): Adjust.
17955 * tree-streamer.h (struct streamer_tree_cache_d): Add
17957 (streamer_read_integer_cst): Remove.
17958 (streamer_tree_cache_insert): Adjust.
17959 (streamer_tree_cache_append): Likewise.
17960 (streamer_tree_cache_insert_at): Rename to ...
17961 (streamer_tree_cache_replace_tree): ... this and adjust.
17962 (streamer_tree_cache_create): Add argument whether to record hashes.
17963 (streamer_tree_cache_get): Rename to ...
17964 (streamer_tree_cache_get_tree): ... this.
17965 (streamer_tree_cache_get_hash): New function.
17966 * tree.c (cache_integer_cst): New function.
17967 * tree.h (cache_integer_cst): Declare.
17968 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
17969 * lto-symtab.c (lto_varpool_replace_node): Only release
17970 DECL_INITIAL of non-prevailing decls.
17971 * varpool.c (varpool_remove_initializer): Do not release
17972 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
17974 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
17976 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
17977 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
17978 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
17979 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
17980 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
17981 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
17982 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
17983 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
17984 instead of TARGET_64BIT.
17985 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
17986 Require ISA_HAS_<D>DIV.
17988 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
17990 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
17991 (mips*-*-linux*): Move default with_llsc setting to where other
17993 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
17995 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
17996 Likewise. Remove default with_tune setting. Move default float
17997 setting to its own block. Handle with_llsc in the same block as above.
17999 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
18001 PR rtl-optimization/57425
18002 PR rtl-optimization/57569
18003 * alias.c (write_dependence_p): Add new parameters mem_mode,
18004 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
18005 Changed all callers.
18006 (canon_anti_dependence): New function.
18007 * cse.c (check_dependence): Use canon_anti_dependence.
18008 * cselib.c (cselib_invalidate_mem): Likewise.
18009 * rtl.h (canon_anti_dependence): Declare.
18011 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
18013 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
18014 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
18015 ".set mips3" for 64-bit targets.
18017 2013-06-15 Dehao Chen <dehao@google.com>
18019 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
18020 * gimple-low.c (gimple_check_call_matching_types): Likewise.
18021 (gimple_check_call_args): Likewise.
18022 * value-prof.c (check_ic_target): Likewise.
18023 * ipa-inline.c (early_inliner): Likewise.
18024 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
18025 * cgraph.c (cgraph_create_edge_1): Likewise.
18026 (cgraph_make_edge_direct): Likewise.
18028 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
18031 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
18032 rs6000_output_move_128bit to handle emitting quad memory
18033 operations. Set attribute length to 8 bytes.
18035 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
18037 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
18039 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
18040 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
18041 (aarch64_<su>mlsl<mode>): Likewise.
18043 2013-06-14 Mike Stump <mikestump@comcast.net>
18045 * Makefile.in (TARGET_H): Add insn-codes.h.
18047 2013-06-14 Alan Modra <amodra@gmail.com>
18049 PR middle-end/57134
18050 PR middle-end/57586
18051 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
18052 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
18053 bitfield expansion when EXPAND_MEMORY.
18054 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
18056 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
18058 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
18059 test for clearing quad memory on 32-bit later.
18061 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18063 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
18064 (fold_negate_expr): Likewise.
18065 (fold_real_zero_addition_p): Handle vectors.
18066 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
18068 2013-06-14 Alan Modra <amodra@gmail.com>
18070 * varasm.c (force_const_mem): Revert 2013-06-07 change.
18072 2013-06-13 Jan Hubicka <jh@suse.cz>
18074 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
18075 Local comdats are not externally visible.
18076 * symtab.c (dump_symtab_base): Dump externally visible.
18077 (verify_symtab_base): Verify back links in the symtab hash.
18079 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18081 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
18082 CONVERT_EXPR as equal nodes.
18084 2013-06-13 Bin Cheng <bin.cheng@arm.com>
18086 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
18088 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
18090 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
18091 Generalize to complex and vector.
18092 * tree.c (build_all_ones_cst): New function.
18093 * tree.h (build_all_ones_cst): Declare it.
18095 2013-06-13 Alan Modra <amodra@gmail.com>
18097 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
18098 * config/rs6000/rs6000.md (signbittf2): New insn.
18099 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
18100 (abstf2_internal, cmptf_internal2): Likewise.
18101 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
18103 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
18104 Pat Haugen <pthaugen@us.ibm.com>
18105 Peter Bergner <bergner@vnet.ibm.com>
18107 * config/rs6000/rs6000.c (emit_load_locked): Add support for
18108 power8 byte, half-word, and quad-word atomic instructions.
18109 (emit_store_conditional): Likewise.
18110 (rs6000_expand_atomic_compare_and_swap): Likewise.
18111 (rs6000_expand_atomic_op): Likewise.
18113 * config/rs6000/sync.md (larx): Add new modes for power8.
18115 (AINT): New mode iterator to include TImode as well as normal
18116 integer modes on power8.
18117 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
18118 that VSX registers are not considered. Use AINT mode iterator
18119 instead of INT1 to allow inclusion of quad word atomic operations
18121 (load_locked<mode>): Likewise.
18122 (store_conditional<mode>): Likewise.
18123 (atomic_compare_and_swap<mode>): Likewise.
18124 (atomic_exchange<mode>): Likewise.
18125 (atomic_nand<mode>): Likewise.
18126 (atomic_fetch_<fetchop_name><mode>): Likewise.
18127 (atomic_nand_fetch<mode>): Likewise.
18128 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
18130 (ATOMIC): On power8, add QImode, HImode modes.
18131 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
18132 modes that promote to SImode.
18133 (load_lockedti): Convert TImode arguments to PTImode, so that we
18134 get a guaranteed even/odd register pair.
18135 (load_lockedpti): Likewise.
18136 (store_conditionalti): Likewise.
18137 (store_conditionalpti): Likewise.
18139 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
18140 atomic load/store instructions.
18143 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
18145 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
18147 (insn_count): New attribute, with most cases extracted from...
18148 (length): ...here. Redefine most cases in terms of insn_count.
18149 (single_insn): Delete.
18150 (can_delay): Use insn_count to check for single instructions.
18151 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
18152 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
18153 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
18154 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
18155 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
18156 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
18157 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
18158 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
18159 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
18160 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
18161 rather than "length".
18162 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
18163 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
18164 Use "insn_count" rather than "length".
18165 * config/mips/mips-dsp.md
18166 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
18167 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
18170 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
18172 PR tree-optimization/57361
18173 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
18175 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
18177 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
18179 (aarch64_simd_combine<mode>): New instruction expansion.
18180 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
18181 function prototype.
18182 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
18183 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
18185 2013-06-12 Jan Hubicka <jh@suse.cz>
18187 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
18188 decl has when in streaming stage.
18189 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
18190 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
18192 2013-06-12 Roland Stigge <stigge@antcom.de>
18195 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
18197 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18199 PR tree-optimization/57537
18200 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
18201 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
18203 2013-06-12 Richard Biener <rguenther@suse.de>
18205 * data-streamer.h (streamer_write_char_stream): CSE
18206 obs->current_pointer.
18207 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
18208 streamer_write_char_stream manually and optimize the resulting loop.
18209 (streamer_write_hwi_stream): Likewise.
18211 2013-06-12 Jan Hubicka <jh@suse.cz>
18213 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
18214 * cgraph.h (varpool_create_empty_node): Declare.
18215 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
18217 * symtab.c (symtab_unregister_node): Be lax about missin entries
18219 (symtab_get_node): Update comment.
18220 * varpool.c (varpool_create_empty_node): Break out from ...
18221 (varpool_node_for_decl): ... here.
18222 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
18224 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
18226 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
18227 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
18228 part. Use straight-line flow at the end.
18229 <COMPONENT_REF>: Remove superfluous else.
18230 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
18232 2013-06-12 Jakub Jelinek <jakub@redhat.com>
18235 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
18236 target hook even for !TREE_PUBLIC decls. If no resolution info
18237 is available, return false for common and external decls.
18239 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18241 * config/rl78/constraints.md (U): New constraint.
18242 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
18245 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
18248 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
18249 to allow returning address to AT_PLATFORM name.
18251 2013-06-11 Jan Hubicka <jh@suse.cz>
18253 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
18254 * cgraph.h (symtab_node_base): Add weakref flag.
18255 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
18256 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
18257 (output_weakrefs): Use weakref flag.
18258 * fold-const.c (simple_operand_p): Handle WEAK.
18259 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
18260 * ipa.c (varpool_externally_visible_p): Drop weakref.
18261 (function_and_variable_visibility): Update comment; fix weakref
18262 sanity checks; do not clear DECL_WEAK on them.
18263 * lto-cgraph.c (lto_output_node): update.
18264 (lto_output_varpool_node): Update.
18265 (input_overwrite_node): Update.
18266 (input_node): Update.
18267 (input_varpool_node): Update.
18268 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
18269 (lto_symtab_merge_symbols): Add sanity check.
18270 (lto_symtab_prevailing_decl): Do not special case weakrefs.
18271 * passes.c (rest_of_decl_compilation): Set static flag, too.
18272 * symtab.c (dump_symtab_base): Dump weakref.
18273 (verify_symtab_base): Sanity check weakrefs.
18274 (symtab_make_decl_local): Remove duplicated code.
18275 (symtab_alias_ultimate_target): Simplify.
18276 * varpool.c (varpool_create_variable_alias): Set weakref flag.
18278 2013-06-11 Tom de Vries <tom@codesourcery.com>
18280 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
18281 sequence_vect == NULL.
18283 2013-06-11 DJ Delorie <dj@redhat.com>
18285 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
18286 (rl78_unwind_word_mode): New.
18288 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18290 * final.c (debug_prefix_maps): Make static.
18292 2013-06-11 David Malcolm <dmalcolm@redhat.com>
18294 * function.c (initial_trampoline): Remove stray copy.
18296 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
18298 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
18300 2013-06-11 Martin Jambor <mjambor@suse.cz>
18302 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
18303 within bounds at the beginning of the function.
18305 2013-06-11 Alan Modra <amodra@gmail.com>
18307 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
18309 (get_named_section): Don't NULL !DECL_P decl.
18311 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
18313 * doc/invoke.texi (core-avx2): Document.
18315 (atom): Updated with MOVBE.
18317 2013-06-11 Richard Biener <rguenther@suse.de>
18319 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
18321 2013-06-11 Anton Blanchard <anton@samba.org>
18323 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
18324 correct shift value in little-endian mode.
18326 2013-06-11 Jakub Jelinek <jakub@redhat.com>
18329 * varasm.c (get_variable_align): Move #endif to the right place.
18331 2013-06-10 Cary Coutant <ccoutant@google.com>
18333 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
18334 for hash so that hash table traversal order is deterministic.
18336 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
18337 Pat Haugen <pthaugen@us.ibm.com>
18338 Peter Bergner <bergner@vnet.ibm.com>
18340 * config/rs6000/vector.md (GPR move splitter): Do not split moves
18341 of vectors in GPRS if they are direct moves or quad word load or
18344 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
18346 (direct_move_p): Likewise.
18347 (quad_load_store_p): Likewise.
18349 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
18350 classes into bins based on the physical register type.
18351 (reg_class_to_reg_type): Likewise.
18352 (IS_STD_REG_TYPE): Likewise.
18353 (IS_FP_VECT_REG_TYPE): Likewise.
18354 (reload_fpr_gpr): Arrays to determine what insn to use if we can
18355 use direct move instructions.
18356 (reload_gpr_vsx): Likewise.
18357 (reload_vsx_gpr): Likewise.
18358 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
18359 information that is a simplification of register classes. Also
18360 precalculate direct move reload helpers.
18361 (direct_move_p): New function to return true if the operation can
18362 be done as a direct move instruciton.
18363 (quad_load_store_p): New function to return true if the operation
18364 is a quad memory operation.
18365 (rs6000_legitimize_address): If quad memory, only allow register
18366 indirect for TImode addresses.
18367 (rs6000_legitimate_address_p): Likewise.
18368 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
18369 (rs6000_reload_register_type): Likewise.
18370 (register_to_reg_type): Return register type.
18371 (rs6000_secondary_reload_simple_move): New helper function for
18372 secondary reload and secondary memory needed to identify anything
18373 that is a simple move, and does not need reloading.
18374 (rs6000_secondary_reload_direct_move): New helper function for
18375 secondary reload to identify cases that can be done with several
18376 instructions via the direct move instructions.
18377 (rs6000_secondary_reload_move): New helper function for secondary
18378 reload to identify moves between register types that can be done.
18379 (rs6000_secondary_reload): Add support for quad memory operations
18380 and for direct move.
18381 (rs6000_secondary_memory_needed): Likewise.
18382 (rs6000_debug_secondary_memory_needed): Change argument names.
18383 (rs6000_output_move_128bit): New function to return the move to
18384 use for 128-bit moves, including knowing about the various
18385 limitations of quad memory operations.
18387 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
18388 memory operations. call rs6000_output_move_128bit for the actual
18389 instruciton(s) to generate.
18390 (vsx_movti_64bit): Likewise.
18392 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
18393 (UNSPEC_P8V_MTVSRWZ): Likewise.
18394 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
18395 (UNSPEC_P8V_MTVSRD): Likewise.
18396 (UNSPEC_P8V_XXPERMDI): Likewise.
18397 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
18398 (UNSPEC_FUSION_GPR): Likewise.
18399 (FMOVE128_GPR): New iterator for direct move.
18400 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
18401 (f32_sv): Likewise.
18402 (f32_dm): Likewise.
18403 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
18404 loads and direct move instructions.
18405 (zero_extendsidi2_lfiwzx): Likewise.
18406 (extendsidi2_lfiwax): Likewise.
18407 (extendsidi2_nocell): Likewise.
18408 (floatsi<mode>2_lfiwax): Likewise.
18409 (lfiwax): Likewise.
18410 (floatunssi<mode>2_lfiwzx): Likewise.
18411 (lfiwzx): Likewise.
18412 (fix_trunc<mode>_stfiwx): Likewise.
18413 (fixuns_trunc<mode>_stfiwx): Likewise.
18414 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
18415 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
18416 (parity<mode>2_cmpb): Set length/type attr.
18417 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
18418 for 'mr.' to fast_compare.
18419 (bpermd_<mode>): Change type attr to popcnt.
18420 (p8_fmrgow_<mode>): New insns for power8 direct move support.
18421 (p8_mtvsrwz_1): Likewise.
18422 (p8_mtvsrwz_2): Likewise.
18423 (reload_fpr_from_gpr<mode>): Likewise.
18424 (p8_mtvsrd_1): Likewise.
18425 (p8_mtvsrd_2): Likewise.
18426 (p8_xxpermdi_<mode>): Likewise.
18427 (reload_vsx_from_gpr<mode>): Likewise.
18428 (reload_vsx_from_gprsf): Likewise.
18429 (p8_mfvsrd_3_<mode>): LIkewise.
18430 (reload_gpr_from_vsx<mode>): Likewise.
18431 (reload_gpr_from_vsxsf): Likewise.
18432 (p8_mfvsrd_4_disf): Likewise.
18433 (multi-word GPR splits): Do not split direct moves or quad memory
18436 2013-06-10 David Malcolm <dmalcolm@redhat.com>
18438 * tree-into-ssa.c (interesting_blocks): Make static.
18440 2013-06-10 Jakub Jelinek <jakub@redhat.com>
18443 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
18444 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
18445 Use DATA_ABI_ALIGNMENT for that case instead if defined.
18446 (get_variable_align): New function.
18447 (get_variable_section, emit_bss, emit_common,
18448 assemble_variable_contents, place_block_symbol): Use
18449 get_variable_align instead of DECL_ALIGN.
18450 (assemble_noswitch_variable): Add align argument, use it
18451 instead of DECL_ALIGN.
18452 (assemble_variable): Adjust caller. Use get_variable_align
18453 instead of DECL_ALIGN.
18454 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
18456 (DATA_ABI_ALIGNMENT): Define.
18457 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
18458 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
18459 opt is false, only return the psABI mandated alignment increase.
18460 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
18461 (DATA_ABI_ALIGNMENT): ... this.
18462 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
18463 (DATA_ABI_ALIGNMENT): ... this.
18464 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
18465 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
18466 (DATA_ABI_ALIGNMENT): ... this.
18467 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
18468 * doc/tm.texi: Regenerated.
18470 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
18472 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
18473 cmp_code to construct REG_EQUAL note.
18475 2013-06-09 Jakub Jelinek <jakub@redhat.com>
18478 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
18479 that operands[2] doesn't overlap with operands[0].
18481 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
18482 Jan Hubicka <jh@suse.cz>
18484 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
18485 hack to mark symbols as used.
18487 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
18489 PR rtl-optimization/57559
18490 * lra-constraints.c (process_alt_operands): Don't discourage
18491 memory with known offset for offsetable memory constraint.
18492 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
18494 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
18496 * varasm.c (struct oc_local_state): Reorder fields.
18497 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
18498 and adjust accordingly.
18499 (output_constructor): Reorder initialization code and adjust call to
18500 output_constructor_bitfield.
18502 2013-06-07 Jan Hubicka <jh@suse.cz>
18504 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
18506 2013-06-07 David Malcolm <dmalcolm@redhat.com>
18508 * tree-object-size.c (unknown): Make const.
18510 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
18512 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
18513 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
18514 for last alternative in the cpu_facility attribute.
18516 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18519 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
18520 (xordi3): Change operand 2 constraint to arm_xordi_operand.
18521 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
18522 * config/arm/constraints.md (Dg): New constraint.
18523 * config/arm/neon.md (xordi3_neon): Remove.
18524 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
18525 * config/arm/predicates.md (arm_xordi_operand): New predicate.
18527 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18529 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
18530 Clean up alternatives.
18532 2013-06-07 Alan Modra <amodra@gmail.com>
18534 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
18537 2013-06-07 Alan Modra <amodra@gmail.com>
18539 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
18541 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18543 * config/arm/constraints.md (Df): New constraint.
18544 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
18545 Correct length attribute for last two alternatives.
18547 2013-06-07 Alan Modra <amodra@gmail.com>
18549 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
18550 override user -mfp-in-toc.
18551 (offsettable_ok_by_alignment): Consider just the current access
18552 rather than the whole object, unless BLKmode. Handle
18553 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
18554 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
18555 for -mcmodel=medium.
18556 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
18557 override user -mfp-in-toc or -msum-in-toc. Default to
18558 -mno-fp-in-toc for -mcmodel=medium.
18560 2013-06-06 DJ Delorie <dj@redhat.com>
18562 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
18563 TARGET_VALID_POINTER_MODE.
18565 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
18566 Pat Haugen <pthaugen@us.ibm.com>
18567 Peter Bergner <bergner@vnet.ibm.com>
18569 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
18570 Document new power8 builtins.
18572 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
18573 condition code register, to allow 128-bit logical operations to be
18574 done in the VSX or GPR registers.
18575 (nor<mode>3): Use the canonical form for nor.
18576 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
18577 vclz*, and vpopcnt* vector instructions.
18578 (nand<mode>3): Likewise.
18579 (orc<mode>3): Likewise.
18580 (clz<mode>2): LIkewise.
18581 (popcount<mode>2): Likewise.
18583 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
18584 that only the GPRs are recognized.
18586 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18587 support for new power8 builtins.
18589 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
18591 (xscvdpspn): Likewise.
18597 (vpopcnt): Likewise.
18598 (vpopcntb): Likewise.
18599 (vpopcnth): Likewise.
18600 (vpopcntw): Likewise.
18601 (vpopcntd): Likewise.
18603 (vmrgew): Likewise.
18604 (vmrgow): Likewise.
18606 (eqv_v16qi3): Likewise.
18607 (eqv_v8hi3): Likewise.
18608 (eqv_v4si3): Likewise.
18609 (eqv_v2di3): Likewise.
18610 (eqv_v4sf3): Likewise.
18611 (eqv_v2df3): Likewise.
18613 (nand_v16qi3): Likewise.
18614 (nand_v8hi3): Likewise.
18615 (nand_v4si3): Likewise.
18616 (nand_v2di3): Likewise.
18617 (nand_v4sf3): Likewise.
18618 (nand_v2df3): Likewise.
18620 (orc_v16qi3): Likewise.
18621 (orc_v8hi3): Likewise.
18622 (orc_v4si3): Likewise.
18623 (orc_v2di3): Likewise.
18624 (orc_v4sf3): Likewise.
18625 (orc_v2df3): Likewise.
18627 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
18628 allow power8 quad mode in 64-bit.
18629 (rs6000_builtin_vectorized_function): Add support to vectorize
18630 ISA 2.07 count leading zeros, population count builtins.
18631 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
18632 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
18633 (builtin_function_type): Add vgbbd builtin function which takes an
18635 (altivec_expand_vec_perm_const): Add support for new power8 merge
18638 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
18639 that does not include TImdoe for use with 32-bit.
18640 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
18642 (UNSPEC_VSX_CVDPSPN): Likewise.
18643 (vsx_xscvdpspn): Likewise.
18644 (vsx_xscvspdpn): Likewise.
18645 (vsx_xscvdpspn_scalar): Likewise.
18646 (vsx_xscvspdpn_directmove): Likewise.
18647 (vsx_and<mode>3): Split logical operations into 32-bit and
18648 64-bit. Add support to do logical operations on TImode as well as
18649 VSX vector types. Allow logical operations to be done in either
18650 VSX registers or in general purpose registers in 64-bit mode. Add
18651 splitters if GPRs were used. For AND, add clobber of CCmode to
18652 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
18654 (vsx_and<mode>3_32bit): Likewise.
18655 (vsx_and<mode>3_64bit): Likewise.
18656 (vsx_ior<mode>3): Likewise.
18657 (vsx_ior<mode>3_32bit): Likewise.
18658 (vsx_ior<mode>3_64bit): Likewise.
18659 (vsx_xor<mode>3): Likewise.
18660 (vsx_xor<mode>3_32bit): Likewise.
18661 (vsx_xor<mode>3_64bit): Likewise.
18662 (vsx_one_cmpl<mode>2): Likewise.
18663 (vsx_one_cmpl<mode>2_32bit): Likewise.
18664 (vsx_one_cmpl<mode>2_64bit): Likewise.
18665 (vsx_nor<mode>3): Likewise.
18666 (vsx_nor<mode>3_32bit): Likewise.
18667 (vsx_nor<mode>3_64bit): Likewise.
18668 (vsx_andc<mode>3): Likewise.
18669 (vsx_andc<mode>3_32bit): Likewise.
18670 (vsx_andc<mode>3_64bit): Likewise.
18671 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
18672 and xxlorc instructions.
18673 (vsx_eqv<mode>3_64bit): Likewise.
18674 (vsx_nand<mode>3_32bit): Likewise.
18675 (vsx_nand<mode>3_64bit): Likewise.
18676 (vsx_orc<mode>3_32bit): Likewise.
18677 (vsx_orc<mode>3_64bit): Likewise.
18679 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
18681 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
18683 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
18684 (p8_vmrgow): Likewise.
18685 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
18686 GPRs to be split under VSX.
18687 (p8v_clz<mode>2): Add power8 count leading zero support.
18688 (p8v_popcount<mode>2): Add power8 population count support.
18689 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
18692 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
18695 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
18697 (vec_nand): Likewise.
18698 (vec_vclz): Likewise.
18699 (vec_vclzb): Likewise.
18700 (vec_vclzd): Likewise.
18701 (vec_vclzh): Likewise.
18702 (vec_vclzw): Likewise.
18703 (vec_vgbbd): Likewise.
18704 (vec_vmrgew): Likewise.
18705 (vec_vmrgow): Likewise.
18706 (vec_vpopcnt): Likewise.
18707 (vec_vpopcntb): Likewise.
18708 (vec_vpopcntd): Likewise.
18709 (vec_vpopcnth): Likewise.
18710 (vec_vpopcntw): Likewise.
18712 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18714 PR rtl-optimization/57468
18715 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
18718 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18720 PR rtl-optimization/57459
18721 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
18722 type when setting live regs.
18724 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
18726 * config/s390/s390.opt (mlra): New option.
18727 * config/s390/s390.c (s390_decompose_address): Check displacement
18728 for all registers for LRA.
18729 (s390_secondary_reload): Don't used secondary reloads for LRA.
18730 (s390_lra_p): New function.
18731 (TARGET_LRA_P): Define.
18732 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
18733 of attribute cpu_facility to zarch for the last alternative.
18734 (*cmpmem_short): Ditto.
18736 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
18738 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
18739 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
18740 (arm_expand_prologue): Likewise.
18742 2013-06-06 Teresa Johnson <tejohnson@google.com>
18745 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
18746 as this is now done by redirect_edge_and_branch_force.
18747 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
18748 barriers, and fix interaction with splitting.
18749 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
18750 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
18751 reflect changes made in the routine.
18752 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
18753 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
18754 since this is called in cfglayout mode, and replace partition fixup
18755 with assert as that is now done by force_nonfallthru_and_redirect.
18756 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
18757 already be marked with region crossing note.
18758 (insert_section_boundary_note): Make non-static, gate on flag
18759 has_bb_partition, rewrite to also check for multiple partitions.
18760 (rest_of_handle_reorder_blocks): Remove call to
18761 insert_section_boundary_note, now done later during free_cfg.
18762 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
18763 * bb-reorder.h (insert_section_boundary_note): Declare.
18764 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
18765 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
18766 invoke insert_section_boundary_note.
18767 (try_redirect_by_replacing_jump): Remove unnecessary
18768 check for region crossing note.
18769 (fixup_partition_crossing): New function.
18770 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
18771 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
18772 in non-cfglayout mode.
18773 (force_nonfallthru_and_redirect): Fixup partition boundaries,
18774 remove old code that tried to do this. Emit barrier correctly
18775 when we are in cfglayout mode.
18776 (last_bb_in_partition): New function.
18777 (rtl_split_edge): Correctly fixup partition boundaries.
18778 (commit_one_edge_insertion): Remove old code that tried to
18779 fixup region crossing edge since this is now handled in
18780 split_block, and set up insertion point correctly since
18781 block may now end in a jump.
18782 (verify_hot_cold_block_grouping): Guard against checking when not in
18783 linearized RTL mode.
18784 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
18786 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
18787 rtl_verify_flow_info, so not called in cfglayout mode.
18788 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
18789 (fixup_reorder_chain): Remove old code that attempted to fixup region
18790 crossing note as this is now handled in force_nonfallthru_and_redirect.
18791 (duplicate_insn_chain): Don't duplicate switch section notes.
18792 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
18794 * basic-block.h (emit_barrier_after_bb): Declare.
18796 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18798 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
18799 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
18800 arm_usatsihi): Adjust alternatives for arm_restrict_it.
18802 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18804 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
18806 * config/arm/ldmstm.md: Regenerate.
18808 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18810 * config/arm/sync.md (atomic_loaddi_1):
18811 Disable predication for arm_restrict_it.
18812 (arm_load_exclusive<mode>): Likewise.
18813 (arm_load_exclusivesi): Likewise.
18814 (arm_load_exclusivedi): Likewise.
18815 (arm_load_acquire_exclusive<mode>): Likewise.
18816 (arm_load_acquire_exclusivesi): Likewise.
18817 (arm_load_acquire_exclusivedi): Likewise.
18818 (arm_store_exclusive<mode>): Likewise.
18819 (arm_store_exclusive<mode>): Likewise.
18820 (arm_store_release_exclusivedi): Likewise.
18821 (arm_store_release_exclusive<mode>): Likewise.
18823 2013-06-06 Richard Biener <rguenther@suse.de>
18825 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
18827 (lto_tag_is_tree_code_p): Adjust.
18828 (lto_tag_is_gimple_code_p): Likewise.
18829 (lto_gimple_code_to_tag): Likewise.
18830 (lto_tag_to_gimple_code): Likewise.
18831 (lto_tree_code_to_tag): Likewise.
18832 (lto_tag_to_tree_code): Likewise.
18833 * data-streamer.h (streamer_write_hwi_in_range): Use
18834 uhwi streaming to stream the normalized range.
18835 (streamer_read_hwi_in_range): Likewise.
18837 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18839 * config/arm/arm.md (enabled_for_depr_it): New attribute.
18840 (predicable_short_it): Likewise.
18841 (predicated): Likewise.
18842 (enabled): Handle above.
18843 (define_cond_exec): Set predicated attribute to yes.
18845 2013-06-05 Mike Stump <mikestump@comcast.net>
18847 * gdbinit.in (__FUNCTION__): Add.
18849 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
18851 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
18852 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
18854 2013-06-05 Jan Hubicka <jh@suse.cz>
18856 * varasm.c (mark_decl_referenced): Revert the removal until targets
18859 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
18861 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
18862 instead of mark_decl_referenced.
18864 2013-06-05 Jan Hubicka <jh@suse.cz>
18866 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
18867 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
18868 and symtab_used_from_object_file_p.
18869 (cgraph_make_node_local_1): Clear forced_by_abi.
18870 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
18871 * cgraph.h (symtab_node_base): Add forced_by_abi.
18872 (decide_is_variable_needed): Remove.
18873 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
18874 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
18875 (decide_is_symbol_needed): ... this one; handle symbols in general;
18876 always analyze virtuals; honnor forced_by_abi.
18877 (cgraph_finalize_function): Update.
18878 (varpool_finalize_decl): Update.
18879 (symbol_defined_and_needed): Remove.
18880 (analyze_functions): Update.
18881 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
18882 output_refs, input_overwrite_node): Handle forced_by_abi.
18883 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
18884 (address_taken_from_non_vtable_p): ... this one.
18885 (comdat_can_be_unshared_p_1): New function.
18886 (cgraph_comdat_can_be_unshared_p): Rename to ...
18887 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
18888 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
18889 (function_and_variable_visibility): Clear forced_by_abi as needed.
18890 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
18891 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
18892 * symtab.c (dump_symtab_base): Dump forced_by_abi.
18893 * varpool.c (decide_is_variable_needed): Remove.
18895 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18897 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
18898 (arm_option_override): Override arm_restrict_it where appropriate.
18899 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
18900 * config/arm/arm.opt (mrestrict-it): New command-line option.
18901 * doc/invoke.texi: Document -mrestrict-it.
18903 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18905 * tsan.c (tsan_atomic_table): Make const.
18907 2013-06-05 Richard Biener <rguenther@suse.de>
18909 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
18910 index associated with the tree we are supposed to replace.
18911 * tree-streamer-out.c (pack_ts_base_value_fields): Output
18912 TREE_ASM_WRITTEN as zero for everything but SSA names.
18914 2013-06-05 David Malcolm <dmalcolm@redhat.com>
18916 * tree-ssa-structalias.c (call_stmt_vars): Make static.
18918 2013-06-04 Jan Hubicka <jh@suse.cz>
18920 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
18921 (input_node, input_varpool_node): Handle correctly external same
18923 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
18924 nodes at ltrans stage.
18926 2013-06-04 Jan Hubicka <jh@suse.cz>
18928 * ipa-inline.c (update_caller_keys): Fix availability test.
18929 (update_callee_keys): Likewise.
18930 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
18931 to follow ELF standard.
18933 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
18935 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
18936 (mips64r5900el-*-elf*): New configurations.
18937 * config/mips/mips-cpus.def (r5900): New processor.
18938 * config/mips/mips-tables.opt: Regenerate.
18939 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
18940 (mips_issue_rate): Handle PROCESSOR_R5900.
18941 (mips_reorg_process_insns): Force reorder mode for the R5900.
18942 * config/mips/mips.h (TARGET_MIPS5900): Define.
18943 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
18945 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
18947 * config/mips/mips.md (processor): Add r5900.
18948 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
18950 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18952 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
18953 into function to generate MOVI instruction.
18954 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
18955 (aarch64_preferred_simd_mode): Turn into wrapper.
18956 (aarch64_output_scalar_simd_mov_immediate): New function.
18957 * config/aarch64/aarch64-protos.h: Add prototype for above.
18959 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18961 * config/aarch64/aarch64.c (simd_immediate_info): Remove
18962 element_char member.
18963 (sizetochar): Return signed char.
18964 (aarch64_simd_valid_immediate): Remove elchar and other
18965 unnecessary variables.
18966 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
18967 Calculate element_char as required.
18968 * config/aarch64/aarch64-protos.h: Update and move prototype
18969 for aarch64_output_simd_mov_immediate.
18970 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
18973 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18975 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
18976 information completed by aarch64_simd_valid_immediate.
18977 (aarch64_legitimate_constant_p): Update arguments.
18978 (aarch64_simd_valid_immediate): Work with struct rather than many
18980 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
18981 (aarch64_simd_make_constant): Update arguments.
18982 (aarch64_output_simd_mov_immediate): Work with struct rather than
18983 many pointers. Output immediate directly rather than as operand.
18984 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18986 * config/aarch64/constraints.md (Dn): Update arguments.
18988 2013-06-04 Ian Bolton <ian.bolton@arm.com>
18990 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
18992 (aarch64_simd_immediate_valid_for_move): Remove.
18993 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
18994 (aarch64_simd_make_constant): Update call.
18995 (aarch64_output_simd_mov_immediate): Update call.
18996 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
18998 * config/aarch64/constraints.md (Dn): Update call.
19000 2013-06-04 Ian Bolton <ian.bolton@arm.com>
19002 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
19003 return type to bool for prototype.
19004 (aarch64_legitimate_constant_p): Check for true instead of not -1.
19005 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
19006 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
19008 2013-06-04 Catherine Moore <clm@codesourcery.com>
19010 * config/mips/mips.opt (meva): New.
19011 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
19012 (ASM_SPEC): Handle -meva.
19013 * doc/invoke.texi (meva): Document.
19015 2013-06-04 Alan Modra <amodra@gmail.com>
19017 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
19020 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19022 * rtl.def: Add extra fourth optional field to define_cond_exec.
19023 * gensupport.c (process_one_cond_exec): Process attributes from
19025 * doc/md.texi: Document fourth field in define_cond_exec.
19027 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
19029 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
19030 out the processing order as in store_bit_field_1.
19032 2013-06-04 Jan Hubicka <jh@suse.cz>
19034 PR middle-end/57500
19035 * cgraphunit.c (cgraph_process_same_body_aliases): Create
19036 non-VAR_DECL node if it does not exist yet.
19038 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
19040 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
19041 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
19042 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
19043 target_cpu_default setting.
19045 2013-06-03 Teresa Johnson <tejohnson@google.com>
19047 * dumpfile.c (opt_info_switch_p): Change -fopt-info
19048 default to -fopt-info=optimized instead of all.
19049 * doc/invoke.texi: Ditto.
19050 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
19051 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
19052 (execute_vect_slp): Emit BB vectorization success under
19053 MSG_OPTIMIZED_LOCATIONS.
19054 * tree-vect-slp.c (vect_slp_transform_bb): Change
19055 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
19056 * tree-vect-loop.c (vect_transform_loop): Ditto.
19058 2013-06-03 Jason Merrill <jason@redhat.com>
19061 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
19062 Use TARGET_EXPR for C++.
19064 2013-06-03 Jakub Jelinek <jakub@redhat.com>
19066 PR rtl-optimization/57268
19067 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
19068 if DEBUG_INSN_P (insn).
19071 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19073 PR rtl-optimization/57268
19074 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19075 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19077 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19079 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
19080 (ix86_avoid_lea_for_addr): Likewise.
19081 (exact_dependency_1): Likewise.
19082 (ix86_adjust_cost): Likewise.
19083 (swap_top_of_ready_list): Fix formatting and !reload_completed check
19085 (do_reorder_for_imul): Fix typo, formatting and
19086 !reload_completed check removed.
19087 (ix86_sched_reorder): Fix typo and formatting.
19088 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
19091 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
19093 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
19095 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
19097 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
19098 <STRING_CST>: Likewise.
19099 <VECTOR_CST>: Likewise.
19101 2013-06-01 Janus Weil <janus@gcc.gnu.org>
19102 Mikael Morin <mikael@gcc.gnu.org>
19104 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
19105 * config.in: Regenerated.
19106 * configure: Regenerated.
19108 2013-06-01 Jan Hubicka <jh@suse.cz>
19110 PR middle-end/57366
19111 * cgraphunit.c (compile): When weakref is not supported,
19112 set up transparent aliases before final output pass.
19113 * varasm.c (assemble_alias): Do not try to do it here.
19115 2013-06-01 Jan Hubicka <jh@suse.cz>
19117 PR middle-end/57467
19118 * passes.c (for_per_function): Skip unanalyzed functions.
19120 2013-06-01 Jan Hubicka <jh@suse.cz>
19122 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
19123 (lto_symtab_merge_symbols_1): ... this one.
19124 (lto_symtab_merge_cgraph_nodes): Rename to ...
19125 (lto_symtab_merge_symbols): ... this one; simplify.
19126 * cgraph.c (same_body_aliases_done): Rename to ...
19127 (cpp_implicit_aliases_done): ... this one.
19128 (cgraph_create_function_alias): Update.
19129 (cgraph_same_body_alias): Update.
19130 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
19131 (verify_edge_corresponds_to_fndecl): Simplify.
19132 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
19133 (cgraph_node): Remove same_body_alias.
19134 (varpool_node): Remove alias_of and extra_name_alias.
19135 (same_body_aliases_done): Rename to ..
19136 (cpp_implicit_aliases_done): ... this one.
19137 (symtab_alias_ultimate_target): Add default parameter.
19138 (symtab_resolve_alias): New function.
19139 (fixup_same_cpp_alias_visibility): Declare.
19140 (cgraph_function_node): Add default parameter.
19141 (cgraph_node_asm_name): Likewise.
19142 (cgraph_function_or_thunk_node): Add default parameter; do
19143 not ICE when it is NULL.
19144 (varpool_variable_node): Likewise.
19145 * tree-emutls.c (create_emultls_var): Update.
19146 (ipa_lower_emutls): Update.
19147 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
19148 (cgraph_reset_node): Reset alias info.
19149 (cgraph_finalize_function): Update.
19150 (fixup_same_cpp_alias_visibility): Move to symtab.c.
19151 (analyze_function): Simplify.
19152 (cgraph_process_same_body_aliases): Simplify.
19153 (analyze_functions): Fixup same body aliases.
19154 (handle_alias_pairs): Simplify.
19155 (assemble_thunk): Update.
19156 (assemble_thunks_and_aliases): Update.
19157 (output_weakrefs): Rewrite.
19158 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
19159 (lto_output_varpool_node): Likewise.
19160 (compute_ltrans_boundary): Remve assert.
19161 (get_alias_symbol): New functoin.
19162 (input_node): Rewrite alias handling.
19163 (input_varpool_node): Likewise.
19164 * ipa-pure-const.c (propagate_pure_const): Fix formating.
19165 * ipa.c (process_references): Handle weakrefs correctly.
19166 (symtab_remove_unreachable_nodes): Likewise.
19167 * trans-mem.c (get_cg_data): Update.
19168 (ipa_tm_create_version_alias): Update.
19169 (ipa_tm_execute): Update.
19170 * symtab.c (dump_symtab_base): Dump aliases.
19171 (verify_symtab_base): Verify aliases.
19172 (symtab_node_availability): New function.
19173 (symtab_alias_ultimate_target): Simplify.
19174 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
19175 handle all the fixup cases.
19176 (symtab_resolve_alias): New function.
19177 * passes.c (ipa_write_summaries): Handle weakrefs.
19178 * varpool.c (varpool_analyze_node): Simplify.
19179 (assemble_aliases): Update.
19180 (varpool_create_variable_alias): Simplify.
19181 (varpool_extra_name_alias): Simplify.
19182 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
19183 (lto_symtab_merge_symbols): ... this one.
19185 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
19188 PR rtl-optimization/57268
19189 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19190 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19192 2013-06-01 Tobias Burnus <burnus@net-b.de>
19194 Partially reverted:
19195 2013-05-31 Tobias Burnus <burnus@net-b.de>
19197 PR middle-end/57073
19198 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
19201 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
19203 PR rtl-optimization/57268
19204 * sched-deps.c (sched_analyze_2): Flush dependence lists if
19205 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
19207 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19209 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
19210 unordered comparison operators when -fno-trapping-math is in effect
19212 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
19213 and implement unordered comparison operators properly on the e500.
19215 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
19217 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
19218 for constant scalar integers.
19219 (simplify_relational_operation_1): Likewise.
19221 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
19223 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
19224 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
19227 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19228 Igor Zamyatin <igor.zamyatin@intel.com>
19230 Silvermont (SLM) architecture performance tuning.
19231 * config/i386/i386.h (enum ix86_tune_indices): Add
19232 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
19233 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
19235 * config/i386/i386.c (initial_ix86_tune_features)
19236 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
19237 (ix86_lea_outperforms): Handle Silvermont tuning.
19238 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
19240 (ix86_use_lea_for_mov): Likewise.
19241 (ix86_avoid_lea_for_addr): Likewise.
19242 (ix86_lea_for_add_ok): Likewise.
19243 (exact_dependency_1): New function.
19244 (exact_store_load_dependency): Likewise.
19245 (ix86_adjust_cost): Handle Silvermont tuning.
19246 (do_reoder_for_imul): Likewise.
19247 (swap_top_of_ready_list): New function.
19248 (ix86_sched_reorder): Changed to handle Silvermont tuning.
19250 * config/i386/i386.md (peepholes that split memory operand in fp
19253 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19255 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
19256 Remove un-necessary braces.
19258 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
19260 * config/aarch64/aarch64.c (aarch64_classify_symbol):
19261 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
19263 2013-05-31 Tobias Burnus <burnus@net-b.de>
19265 PR middle-end/57073
19266 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
19268 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19271 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
19272 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
19273 * config/arm/neon.md (iordi3_neon): Remove.
19274 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
19275 * config/arm/predicates.md (imm_for_neon_logic_operand):
19276 Move to earlier in the file.
19277 (neon_logic_op2): Likewise.
19278 (arm_iordi_operand_neon): New predicate.
19280 2013-05-31 Richard Biener <rguenther@suse.de>
19282 PR tree-optimization/57478
19283 PR tree-optimization/57453
19284 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
19287 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19289 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
19290 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
19292 2013-05-30 Tobias Burnus <burnus@net-b.de>
19293 Thomas Koenig <tkoenig@gcc.gnu.org>
19295 PR middle-end/57073
19296 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
19297 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
19299 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
19301 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
19303 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
19305 * target.def (register_usage_leveling_p): New hook.
19306 * targhooks.c (default_register_usage_leveling_p): New.
19307 * targhooks.h (default_register_usage_leveling_p): New prototype.
19308 * lra-assigns.c (register_usage_leveling_p): Use the hook.
19309 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
19310 * doc/tm.texi: Update.
19311 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
19313 2013-05-30 Ian Bolton <ian.bolton@arm.com>
19315 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
19316 (*insv_reg<mode>): New define_insn.
19318 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
19320 PR rtl-optimization/57439
19321 * postreload.c (move2add_valid_value_p): Check that we have
19322 a zero subreg_regno_offset when accessing the register in
19323 the requested mode.
19325 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
19326 Igor Zamyatin <igor.zamyatin@intel.com>
19328 Silvermont (SLM) architecture pipeline model, tuning and
19330 * config.gcc: Add slm config options and target.
19332 * config/i386/slm.md: New.
19334 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
19336 * config/i386/i386-c.c (ix86_target_macros_internal): New case
19338 (ix86_target_macros_internal): Likewise.
19340 * config/i386/i386.c (slm_cost): New cost.
19341 (m_SLM): New macro flag.
19342 (initial_ix86_tune_features): Set m_SLM.
19343 (x86_accumulate_outgoing_args): Likewise.
19344 (x86_arch_always_fancy_math_387): Likewise.
19345 (processor_target_table): Add slm cost.
19346 (cpu_names): Add slm cpu name.
19347 (x86_option_override_internal): Set SLM ISA.
19348 (ix86_issue_rate): New case PROCESSOR_SLM.
19349 (ia32_multipass_dfa_lookahead): Likewise.
19350 (fold_builtin_cpu): Add slm.
19352 * config/i386/i386.h (TARGET_SLM): New target macro.
19353 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
19354 (processor_type): Add PROCESSOR_SLM.
19356 * config/i386/i386.md (cpu): Add new value "slm".
19357 (slm.md): Include slm.md.
19359 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
19360 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19362 * config/arm/arm-protos.h: Add and update function protos.
19363 * config/arm/arm.c (use_simple_return_p): New added.
19364 (thumb2_expand_return): Check simple_return flag.
19365 * config/arm/arm.md: Add simple_return and conditional simple_return.
19366 * config/arm/iterators.md: Add iterator for return and simple_return.
19368 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
19370 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
19371 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
19372 (arm_emit_vfp_multi_reg_pop): Likewise.
19373 (thumb2_emit_ldrd_pop): Likewise.
19374 (arm_expand_epilogue): Add misc REG_CFA notes.
19375 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
19377 2013-05-29 Lawrence Crowl <crowl@google.com>
19379 * config/arm/t-arm: Update for below.
19381 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
19382 Change type to hash_table. Update dependent calls and types.
19384 * config/i386/t-cygming: Update for below.
19386 * config/i386/t-interix: Update for below.
19388 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
19389 Change type to hash_table. Update dependent calls and types.
19390 (i386_find_on_wrapper_list::wrappers): Likewise.
19392 * config/ia64/t-ia64: Update for below.
19394 * config/ia64/ia64.c (bundle_state_table):
19395 Change type to hash_table. Update dependent calls and types.
19397 * config/mips/mips.c (mips_reorg_process_insns::htab):
19398 Change type to hash_table. Update dependent calls and types.
19400 * config/sol2.c (solaris_comdat_htab):
19401 Change type to hash_table. Update dependent calls and types.
19403 * config/t-sol2: Update for above.
19405 2013-05-29 Teresa Johnson <tejohnson@google.com>
19407 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
19408 functions are not yet marked as defined.
19410 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
19411 Pat Haugen <pthaugen@us.ibm.com>
19412 Peter Bergner <bergner@vnet.ibm.com>
19414 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
19418 (vrotl<mode>3): Likewise.
19419 (vashl<mode>3): Likewise.
19420 (vlshr<mode>3): Likewise.
19421 (vashr<mode>3): Likewise.
19423 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
19424 support for power8 V2DI builtins.
19426 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
19427 power8 V2DI builtins.
19428 (vupkhsw): Likewise.
19429 (vupklsw): Likewise.
19430 (vaddudm): Likewise.
19431 (vminsd): Likewise.
19432 (vmaxsd): Likewise.
19433 (vminud): Likewise.
19434 (vmaxud): Likewise.
19435 (vpkudum): Likewise.
19436 (vpksdss): Likewise.
19437 (vpkudus): Likewise.
19438 (vpksdus): Likewise.
19443 (vsubudm): Likewise.
19444 (vcmpequd): Likewise.
19445 (vcmpgtsd): Likewise.
19446 (vcmpgtud): Likewise.
19447 (vcmpequd_p): Likewise.
19448 (vcmpgtsd_p): Likewise.
19449 (vcmpgtud_p): Likewise.
19450 (vupkhsw): Likewise.
19451 (vupklsw): Likewise.
19452 (vaddudm): Likewise.
19453 (vmaxsd): Likewise.
19454 (vmaxud): Likewise.
19455 (vminsd): Likewise.
19456 (vminud): Likewise.
19457 (vpksdss): Likewise.
19458 (vpksdus): Likewise.
19459 (vpkudum): Likewise.
19460 (vpkudus): Likewise.
19465 (vsubudm): Likewise.
19467 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
19468 support for power8 V2DI instructions.
19470 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
19471 power8 V2DI instructions. Combine pack and unpack insns to use an
19472 iterator for each mode. Check whether a particular mode supports
19473 Altivec instructions instead of just checking TARGET_ALTIVEC.
19474 (UNSPEC_VPKUWUM): Likewise.
19475 (UNSPEC_VPKSHSS): Likewise.
19476 (UNSPEC_VPKSWSS): Likewise.
19477 (UNSPEC_VPKUHUS): Likewise.
19478 (UNSPEC_VPKSHUS): Likewise.
19479 (UNSPEC_VPKUWUS): Likewise.
19480 (UNSPEC_VPKSWUS): Likewise.
19481 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
19482 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
19483 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
19484 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
19485 (UNSPEC_VUPKHSB): Likewise.
19486 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
19487 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
19488 (UNSPEC_VUPKHSH): Likewise.
19489 (UNSPEC_VUPKLSB): Likewise.
19490 (UNSPEC_VUPKLSH): Likewise.
19492 (VI_char): Likewise.
19493 (VI_scalar): Likewise.
19494 (VI_unit): Likewise.
19496 (VP_small): Likewise.
19497 (VP_small_lc): Likewise.
19498 (VU_char): Likewise.
19499 (add<mode>3): Likewise.
19500 (altivec_vaddcuw): Likewise.
19501 (altivec_vaddu<VI_char>s): Likewise.
19502 (altivec_vadds<VI_char>s): Likewise.
19503 (sub<mode>3): Likewise.
19504 (altivec_vsubcuw): Likewise.
19505 (altivec_vsubu<VI_char>s): Likewise.
19506 (altivec_vsubs<VI_char>s): Likewise.
19507 (altivec_vavgs<VI_char>): Likewise.
19508 (altivec_vcmpbfp): Likewise.
19509 (altivec_eq<mode>): Likewise.
19510 (altivec_gt<mode>): Likewise.
19511 (altivec_gtu<mode>): Likewise.
19512 (umax<mode>3): Likewise.
19513 (smax<mode>3): Likewise.
19514 (umin<mode>3): Likewise.
19515 (smin<mode>3): Likewise.
19516 (altivec_vpkuhum): Likewise.
19517 (altivec_vpkuwum): Likewise.
19518 (altivec_vpkshss): Likewise.
19519 (altivec_vpkswss): Likewise.
19520 (altivec_vpkuhus): Likewise.
19521 (altivec_vpkshus): Likewise.
19522 (altivec_vpkuwus): Likewise.
19523 (altivec_vpkswus): Likewise.
19524 (altivec_vpks<VI_char>ss): Likewise.
19525 (altivec_vpks<VI_char>us): Likewise.
19526 (altivec_vpku<VI_char>us): Likewise.
19527 (altivec_vpku<VI_char>um): Likewise.
19528 (altivec_vrl<VI_char>): Likewise.
19529 (altivec_vsl<VI_char>): Likewise.
19530 (altivec_vsr<VI_char>): Likewise.
19531 (altivec_vsra<VI_char>): Likewise.
19532 (altivec_vsldoi_<mode>): Likewise.
19533 (altivec_vupkhsb): Likewise.
19534 (altivec_vupkhs<VU_char>): Likewise.
19535 (altivec_vupkls<VU_char>): Likewise.
19536 (altivec_vupkhsh): Likewise.
19537 (altivec_vupklsb): Likewise.
19538 (altivec_vupklsh): Likewise.
19539 (altivec_vcmpequ<VI_char>_p): Likewise.
19540 (altivec_vcmpgts<VI_char>_p): Likewise.
19541 (altivec_vcmpgtu<VI_char>_p): Likewise.
19542 (abs<mode>2): Likewise.
19543 (vec_unpacks_hi_v16qi): Likewise.
19544 (vec_unpacks_hi_v8hi): Likewise.
19545 (vec_unpacks_lo_v16qi): Likewise.
19546 (vec_unpacks_hi_<VP_small_lc>): Likewise.
19547 (vec_unpacks_lo_v8hi): Likewise.
19548 (vec_unpacks_lo_<VP_small_lc>): Likewise.
19549 (vec_pack_trunc_v8h): Likewise.
19550 (vec_pack_trunc_v4si): Likewise.
19551 (vec_pack_trunc_<mode>): Likewise.
19553 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
19555 (vec_vmaxsd): Likewise.
19556 (vec_vmaxud): Likewise.
19557 (vec_vminsd): Likewise.
19558 (vec_vminud): Likewise.
19559 (vec_vpksdss): Likewise.
19560 (vec_vpksdus): Likewise.
19561 (vec_vpkudum): Likewise.
19562 (vec_vpkudus): Likewise.
19563 (vec_vrld): Likewise.
19564 (vec_vsld): Likewise.
19565 (vec_vsrad): Likewise.
19566 (vec_vsrd): Likewise.
19567 (vec_vsubudm): Likewise.
19568 (vec_vupkhsw): Likewise.
19569 (vec_vupklsw): Likewise.
19571 2013-05-29 Jan Hubicka <jh@suse.cz>
19573 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
19574 flags; reorder rest of fields in more consistent way.
19575 (varpool_node): Remove analyzed, finalized and alias.
19576 (cgraph_ndoe): Likewise.
19577 (symtab_alias_ultimate_target): New function.
19578 (cgraph_function_node): Move offline.
19579 (cgraph_reset_node): Declare.
19580 (cgraph_comdat_can_be_unshared_p): Remove.
19581 (varpool_remove_initializer): Declare.
19582 (varpool_first_defined_variable, varpool_next_defined_variable
19583 cgraph_first_defined_function, cgraph_next_defined_function): Update.
19584 (cgraph_function_with_gimple_body_p): Update.
19585 (varpool_all_refs_explicit_p): Update.
19586 (symtab_alias_target): New function.
19587 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
19588 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
19589 (cgraph_function_or_thunk_node): Simplify using
19590 symtab_alias_ultimate_target.
19591 (varpool_variable_node): Likewise.
19592 * cgraph.c (cgraph_create_function_alias): Update.
19593 (cgraph_add_thunk): Update.
19594 (cgraph_remove_node): Update.
19595 (dump_cgraph_node): Do not dump removed flags.
19596 (cgraph_function_body_availability): Update.
19597 (cgraph_propagate_frequency): Update.
19598 (verify_cgraph_node): Check sanity of local flag.
19599 (cgraph_function_node): Move here from cgraph.h; revamp for
19600 cgraph_function_or_thunk_node.
19601 * lto-symtab.c (lto_varpool_replace_node): Update.
19602 (lto_symtab_resolve_can_prevail_p): Update.
19603 (lto_symtab_merge_cgraph_nodes): Update.
19604 * ipa-cp.c (determine_versionability, initialize_node_lattices,
19605 propagate_constants_accross_call, devirtualization_time_bonus,
19606 ipcp_propagate_stage): Update.
19607 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
19608 * ipa-inline-transform.c (clone_inlined_nodes,
19609 preserve_function_body_p): Update.
19610 * ipa-reference.c (propagate): Update.
19611 (write_node_summary_p): Update.
19612 * toplev.c (wrapup_global_declaration_2): Update.
19613 * cgraphunit.c (cgraph_analyze_function): Rename to ...
19614 (analyze_function) ... this one.
19615 (cgraph_process_new_functions): Update.
19616 (cgraph_reset_node): Export.
19617 (cgraph_finalize_function): Update.
19618 (cgraph_add_new_function): Update.
19619 (process_function_and_variable_attributes): Update.
19620 (varpool_finalize_decl): Update.
19621 (symbol_finalized): Remove.
19622 (symbol_finalized_and_needed): Rename to ...
19623 (symbol_defined_and_needed): ... update.
19624 (cgraph_analyze_functions): Update.
19625 (handle_alias_pairs): Update.
19626 (mark_functions_to_output): Update.
19627 (assemble_thunk): Update.
19628 (output_in_order): Update.
19629 (output_weakrefs): Update.
19630 (finalize_compilation_unit): Update.
19631 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
19632 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
19633 input_node, input_varpool_node): Update.
19634 * dbxout.c (dbxout_expand_expr): Update.
19635 * cgraphclones.c (cgraph_clone_node): Update.
19636 (cgraph_copy_node_for_versioning): Update.
19637 (cgraph_materialize_clone): Update.
19638 (cgraph_materialize_all_clones): Update.
19639 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
19640 propagate_pure_const, propagate_nothrow): Update.
19641 * lto-streamer-out.c (lto_output, write_symbol): Update.
19642 * ipa-utils.c (ipa_reverse_postorder): Update.
19643 * ipa-inline.c (can_inline_edge_p): Update.
19644 (update_caller_keys, ipa_inline): Update.
19645 * dwarf2out.c (reference_to_unused,
19646 premark_types_used_by_global_vars_helper): Update.
19647 * tree-eh.c (tree_could_trap_p): Update.
19648 * ipa-split.c (consider_split, execute_split_functions): Update.
19649 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
19650 has_addr_references_p): Update; move ahead in file for better
19652 (process_references): Simplify.
19653 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
19654 bodies are removed.
19655 (cgraph_comdat_can_be_unshared_p): Make static.
19656 (cgraph_externally_visible_p): Update.
19657 (varpool_externally_visible_p): Update.
19658 (function_and_variable_visibility): Update.
19659 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
19660 ipa_tm_mark_force_output_node): Update.
19661 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
19662 estimate_edge_devirt_benefit, inline_generate_summary,
19663 inline_write_summary): Update.
19664 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
19665 * ipa-prop.c (ipa_compute_jump_functions): Update.
19666 (ipa_print_node_params, ipa_prop_read_section,
19667 ipa_update_after_lto_read, read_replacements_section): Update.
19668 * varasm.c (mark_decl_referenced): Update.
19669 (assemble_alias, dump_tm_clone_pairs): Update.
19670 * tree-inline.c (copy_bb): Update.
19671 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
19673 * symtab.c (dump_symtab_base): Print new flags.
19674 (verify_symtab_base): Verify new flags.
19675 (symtab_alias_ultimate_target): New function.
19676 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
19677 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
19679 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
19681 * i386.c (ix86_get_function_versions_dispatcher,
19682 ix86_generate_version_dispatcher_body): Update.
19683 (fold_builtin_cpu): Use varpool_add_new_variable.
19684 * varpool.c (varpool_remove_initializer): Break out from ...
19685 (varpool_remove_node): ... this one.
19686 (dump_varpool_node, varpool_node_for_asm,
19687 cgraph_variable_initializer_availability, varpool_analyze_node,
19688 varpool_assemble_decl, varpool_remove_unreferenced_decls,
19689 varpool_finalize_named_section_flags, varpool_create_variable_alias):
19692 2013-05-29 Jan Hubicka <jh@suse.cz>
19694 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
19696 2013-05-29 Easwaran Raman <eraman@google.com>
19698 PR tree-optimization/57442
19699 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
19700 when control exits the main loop.
19702 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
19704 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
19706 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
19707 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
19708 * rx/t-rx: Add rx100 under multi library matches option for nofpu
19711 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19713 PR tree-optimization/57441
19714 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
19715 Don't limit size of incr_vec to number of candidates.
19717 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
19719 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
19720 and mips16 directories.
19721 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
19722 (MULTILIB_DIRNAMES): Ditto.
19723 (MULTILIB_EXCEPTIONS): Add new exceptions.
19724 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
19725 (MULTILIB_DIRNAMES): Ditto.
19726 (MULTILIB_EXCEPTIONS): Add new exceptions.
19728 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19729 Marcus Shawcroft <marcus.shawcroft@arm.com>
19731 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
19732 SYMBOL_TINY_ABSOLUTE.
19733 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
19734 SYMBOL_TINY_ABSOLUTE.
19735 (aarch64_expand_mov_immediate): Likewise.
19736 (aarch64_classify_symbol): Likewise.
19737 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
19738 Permit SYMBOL_TINY_ABSOLUTE.
19739 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
19741 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
19742 Marcus Shawcroft <marcus.shawcroft@arm.com>
19744 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
19745 Refactor if/switch. Replace gcc_assert with if.
19747 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19749 * config/i386/i386.c (initial_ix86_tune_features): Enable
19750 FP Reassociation for AMD bdver1 and bdver2.
19752 2013-05-29 Martin Jambor <mjambor@suse.cz>
19754 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
19755 and IMAGPART_EXPR do not occur within other handled_components.
19757 2013-05-29 Richard Biener <rguenther@suse.de>
19759 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
19760 access on whether the use is in the BB we currently try to
19762 (vect_bb_vectorization_profitable_p): Pass the BB we currently
19763 vectorize to vect_bb_slp_scalar_cost.
19765 2013-05-29 Richard Biener <rguenther@suse.de>
19767 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
19768 computing scalar cost offsetted by stmts that are kept live
19770 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
19771 for computation of scalar cost.
19773 2013-05-28 Steve Ellcey <sellcey@mips.com>
19775 * config/mips/mips-cpus.def (mips32r2): Change processor type.
19777 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
19779 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
19780 array notation built-in reduction functions.
19781 * doc/passes.texi (Passes): Added documentation about changes done
19783 * doc/invoke.texi (C Dialect Options): Added documentation about
19784 the -fcilkplus flag.
19785 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
19786 (BUILTINS_DEF): Depend on cilkplus.def.
19787 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
19788 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
19789 * cilkplus.def: New file.
19791 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
19793 PR rtl-optimization/57439
19794 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
19796 2013-05-28 Easwaran Raman <eraman@google.com>
19798 PR tree-optimization/57337
19799 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
19800 (find_insert_point): Correctly identify the insertion point
19801 when two statements with the same UID is compared.
19803 2013-05-28 Richard Biener <rguenther@suse.de>
19805 PR tree-optimization/56787
19806 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
19807 from the list of data references.
19808 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
19810 (vect_analyze_loop_operations): Likewise.
19811 (vect_transform_loop): Remove clobbers.
19813 2013-05-28 Martin Jambor <mjambor@suse.cz>
19815 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
19816 and REALPART_EXPRs have scalar type.
19818 2013-05-28 Richard Biener <rguenther@suse.de>
19820 PR tree-optimization/57411
19821 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
19823 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
19824 virtual operand propagation.
19826 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19828 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
19829 destination register for bmasksi_vis.
19830 (vector_init_bshuffle): Likewise.
19831 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
19833 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
19835 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
19836 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
19837 mode if the instruction isn't available in the original mode.
19838 * config/sparc/sparc.opt (mfix-ut699): New option.
19839 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
19840 (divdf3): Turn into expander.
19841 (divdf3_nofix): New insn.
19842 (divdf3_fix): Likewise.
19843 (divsf3): Disable if -mfix-ut699.
19844 (sqrtdf2): Turn into expander.
19845 (sqrtdf2_nofix): New insn.
19846 (sqrtdf2_fix): Likewise.
19847 (sqrtsf2): Disable if -mfix-ut699.
19849 2013-05-27 Richard Biener <rguenther@suse.de>
19851 PR middle-end/57412
19852 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
19853 block for the new loop.
19855 2013-05-27 Richard Biener <rguenther@suse.de>
19857 PR tree-optimization/57343
19858 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
19859 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
19860 (number_of_iterations_cond): Do not build the folded tree.
19862 2013-05-27 Richard Biener <rguenther@suse.de>
19865 PR middle-end/57381
19866 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19867 OEP_CONSTANT_ADDRESS_OF retained.
19869 PR tree-optimization/57417
19870 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
19871 for unchanged base.
19872 (set_ssa_val_to): Compare addresses using
19873 get_addr_base_and_unit_offset.
19875 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
19877 PR rtl-optimization/56833
19878 * postreload.c (move2add_record_mode): New function.
19879 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
19880 (move2add_use_add2_insn): Use move2add_record_sym_value.
19881 (move2add_use_add3_insn): Likewise.
19882 (reload_cse_move2add): Use move2add_valid_value_p and
19883 move2add_record_mode. Invalidate call-clobbered and REG_INC
19884 affected regs by setting reg_mode to VOIDmode.
19885 (move2add_note_store): Don't pretend the inside of a SUBREG is
19886 the actual destination. Invalidate single/leading registers by
19887 setting reg_mode to VOIDmode.
19888 Use move2add_record_sym_value, move2add_valid_value_p and
19889 move2add_record_mode.
19891 2013-05-27 Richard Biener <rguenther@suse.de>
19893 PR tree-optimization/57396
19894 * tree-affine.c (double_int_constant_multiple_p): Properly
19895 return false for val == 0 and div != 0.
19897 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19899 * config/mips/mips.h: Use #elif in preprocessor conditions.
19901 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19904 * config/mips/constraints.md (kl): New constraint.
19905 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
19906 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
19907 constraint for operand 0. Split after CSE for MIPS16. Emit a move
19908 from LO for MIPS16.
19909 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
19911 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
19914 * config/mips/mips.c (mips_can_inline_p): New function.
19915 (TARGET_CAN_INLINE_P): Define.
19917 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
19919 * sched-int.h (ds_t, dw_t): Make unsigned int.
19920 Fix documentation that describes how all the ds_t bits are used.
19921 Reserve the last bit for delayed-branch scheduling.
19922 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
19923 (BITS_PER_DEP_WEAK): Fix definition and documentation.
19924 (gen_dep_weak_1): Remove prototype.
19925 * sched-deps.c (get_dep_weak_1): Make static.
19926 * target.def (speculate_insn, needs_block_p, gen_spec_check,
19927 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
19928 * doc/tm.texi: Regenerate.
19929 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
19931 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
19934 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
19936 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
19937 Sandra Loosemore <sandra@codesourcery.com>
19939 * config.gcc (powerpc-*): Allow native for with-cpu.
19941 2013-05-24 Jeff Law <law@redhat.com>
19943 PR tree-optimization/57124
19944 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
19945 conversion feeding a condition if the range has an overflow
19946 if -fstrict-overflow. Add warnings for when we do make the
19949 2013-05-24 Dehao Chen <dehao@google.com>
19951 * tree-cfg.c (locus_discrim_map): Fix the typo.
19952 (locus_discrim_hasher): Likewise.
19953 (locus_discrim_hasher::hash): Likewise.
19954 (locus_discrim_hasher::equal): Likewise.
19956 2013-05-24 Martin Jambor <mjambor@suse.cz>
19958 PR tree-optimization/57294
19959 * cgraph.h (ipa_record_stmt_references): Declare.
19960 * cgraphbuild.c (ipa_record_stmt_references): New function.
19961 (build_cgraph_edges): Use ipa_record_stmt_references.
19962 (rebuild_cgraph_edges): Likewise.
19963 (cgraph_rebuild_references): Likewise.
19964 * ipa-prop.c (ipa_modify_call_arguments): Discard references
19965 associated with the old statement and build references from the
19966 newly built statements.
19967 * ipa-ref.c (ipa_remove_stmt_references): New function.
19968 * ipa-ref.h (ipa_remove_stmt_references): Declare.
19970 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
19972 * lra-constraints.c (emit_spill_move): Use smaller mode for
19974 (check_and_process_move): Consider mem-reg moves for secondary
19976 (curr_insn_transform): Don't lose insns emitted before for
19977 secondary memory moves.
19978 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
19979 reg set up in the current insn.
19981 2013-05-24 Dehao Chen <dehao@google.com>
19983 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
19985 (locus_descrim_hasher::equal): Likewise.
19986 (build_gimple_cfg): New discriminator assignment algorithm.
19987 (make_edges): Likewise.
19988 (next_discriminator_for_locus): Likewise.
19989 (same_line_p): Likewise.
19990 (assign_discriminators): Likewise.
19991 (make_cond_expr_edges): Likewise.
19992 (make_gimple_switch_edges): Likewise.
19993 (make_goto_expr_edges): Likewise.
19994 (make_gimple_asm_edges): Likewise.
19996 2013-05-24 Ian Bolton <ian.bolton@arm.com>
19998 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
19999 X format specifier to only display bottom 16 bits.
20000 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
20001 immediate to match for operand 2, since it will be masked.
20003 2013-05-24 Richard Biener <rguenther@suse.de>
20005 PR tree-optimization/57287
20006 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
20007 all SSA names that occur in abnormal PHIs.
20009 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
20012 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
20013 that index is not negative.
20015 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
20017 PR rtl-optimization/55177
20018 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
20019 (simplify_byte_swapping_operation): New.
20020 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
20021 (simplify_relational_operation_1): Deal with BSWAP.
20023 2013-05-23 Richard Henderson <rth@redhat.com>
20026 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
20027 (ix86_reorg): Call it.
20029 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
20032 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
20033 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
20034 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
20036 2013-05-23 Christian Bruel <christian.bruel@st.com>
20039 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
20041 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20042 Marcus Shawcroft <marcus.shawcroft@arm.com>
20044 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
20045 * config/aarch64/constraints.md (Usa): Remove.
20046 * doc/md.texi (AArch64 Usa): Remove.
20048 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
20049 Marcus Shawcroft <marcus.shawcroft@arm.com>
20051 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
20052 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
20053 * config/aarch64/predicates.md (aarch64_const_address): Remove.
20054 (aarch64_mov_operand): Use aarch64_mov_operand_p.
20056 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
20058 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
20059 instruction (AdvSIMD).
20060 * config/aarch64/aarch64-builtins.c
20061 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
20062 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
20064 2013-05-23 Martin Jambor <mjambor@suse.cz>
20066 PR middle-end/57347
20067 * tree.h (contains_bitfld_component_ref_p): Declare.
20068 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
20069 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
20071 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
20072 not access a bit-field. Assert all final offsets are byte-aligned.
20074 2013-05-23 Richard Biener <rguenther@suse.de>
20076 PR tree-optimization/57380
20077 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
20078 least one invariant or re-used load.
20079 * passes.c (init_optimization_passes): Move pass_phiprop before
20082 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
20084 * config/aarch64/aarch64-simd.md
20085 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
20087 2013-05-23 Richard Biener <rguenther@suse.de>
20089 PR middle-end/57381
20090 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
20091 OEP_CONSTANT_ADDRESS_OF retained.
20093 2013-05-23 Jakub Jelinek <jakub@redhat.com>
20095 PR middle-end/57344
20096 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
20097 don't lower unit. Handle unit not being always BITS_PER_WORD.
20099 2013-05-23 Richard Biener <rguenther@suse.de>
20101 PR rtl-optimization/57341
20102 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
20103 instead of true_dependence.
20105 2013-05-22 David Malcolm <dmalcolm@redhat.com>
20107 * bb-reorder.c (branch_threshold): Make const.
20108 (exec_threshold): Ditto.
20110 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20111 Pat Haugen <pthaugen@us.ibm.com>
20112 Peter Bergner <bergner@vnet.ibm.com>
20114 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
20115 documentation for the power8 crypto builtins.
20117 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
20119 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
20120 macros for defining power8 builtin functions.
20121 (BU_P8V_AV_2): Likewise.
20122 (BU_P8V_AV_P): Likewise.
20123 (BU_P8V_VSX_1): Likewise.
20124 (BU_P8V_OVERLOAD_1): Likewise.
20125 (BU_P8V_OVERLOAD_2): Likewise.
20126 (BU_CRYPTO_1): Likewise.
20127 (BU_CRYPTO_2): Likewise.
20128 (BU_CRYPTO_3): Likewise.
20129 (BU_CRYPTO_OVERLOAD_1): Likewise.
20130 (BU_CRYPTO_OVERLOAD_2): Likewise.
20131 (XSCVSPDP): Fix typo, point to the correct instruction.
20132 (VCIPHER): Add power8 crypto builtins.
20133 (VCIPHERLAST): Likewise.
20134 (VNCIPHER): Likewise.
20135 (VNCIPHERLAST): Likewise.
20136 (VPMSUMB): Likewise.
20137 (VPMSUMH): Likewise.
20138 (VPMSUMW): Likewise.
20139 (VPERMXOR_V2DI): Likewise.
20140 (VPERMXOR_V4SI: Likewise.
20141 (VPERMXOR_V8HI: Likewise.
20142 (VPERMXOR_V16QI: Likewise.
20143 (VSHASIGMAW): Likewise.
20144 (VSHASIGMAD): Likewise.
20145 (VPMSUM): Likewise.
20146 (VPERMXOR): Likewise.
20147 (VSHASIGMA): Likewise.
20149 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
20150 __CRYPTO__ if the crypto instructions are available.
20151 (altivec_overloaded_builtins): Add support for overloaded power8
20154 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
20155 support for power8 crypto builtins.
20156 (builtin_function_type): Likewise.
20157 (altivec_init_builtins): Add support for builtins that take vector
20158 long long (V2DI) arguments.
20160 * config/rs6000/crypto.md: New file, define power8 crypto
20163 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
20164 Pat Haugen <pthaugen@us.ibm.com>
20165 Peter Bergner <bergner@vnet.ibm.com>
20167 * doc/invoke.texi (Option Summary): Add power8 options.
20168 (RS/6000 and PowerPC Options): Likewise.
20170 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
20171 constraints.md instead of rs6000.h. Reorder w* constraints. Add
20172 wm, wn, wr documentation.
20174 * config/rs6000/constraints.md (wm): New constraint for VSX
20175 registers if direct move instructions are enabled.
20176 (wn): New constraint for no registers.
20177 (wq): New constraint for quad word even GPR registers.
20178 (wr): New constraint if 64-bit instructions are enabled.
20179 (wv): New constraint if power8 vector instructions are enabled.
20180 (wQ): New constraint for quad word memory locations.
20182 * config/rs6000/predicates.md (const_0_to_15_operand): New
20183 constraint for 0..15 for crypto instructions.
20184 (gpc_reg_operand): If VSX allow registers in VSX registers as well
20185 as GPR and floating point registers.
20186 (int_reg_operand): New predicate to match only GPR registers.
20187 (base_reg_operand): New predicate to match base registers.
20188 (quad_int_reg_operand): New predicate to match even GPR registers
20189 for quad memory operations.
20190 (vsx_reg_or_cint_operand): New predicate to allow vector logical
20191 operations in both GPR and VSX registers.
20192 (quad_memory_operand): New predicate for quad memory operations.
20193 (reg_or_indexed_operand): New predicate for direct move support.
20195 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
20196 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
20197 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
20198 (POWERPC_MASKS): Add power8 options.
20199 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
20202 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
20203 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
20205 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
20206 (-mpower8-fusion): New power8 options.
20207 (-mpower8-fusion-sign): Likewise.
20208 (-mpower8-vector): Likewise.
20209 (-mcrypto): Likewise.
20210 (-mdirect-move): Likewise.
20211 (-mquad-memory): Likewise.
20213 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
20214 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
20216 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
20217 (rs6000_debug_vector_unit): Add p8_vector.
20218 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
20219 definitions. Also print fusion state.
20220 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
20221 (rs6000_builtin_mask_calculate): Add power8 builtin support.
20222 (rs6000_option_override_internal): Add support for power8.
20223 (rs6000_common_init_builtins): Add debugging for skipped builtins
20224 if -mdebug=builtin.
20225 (rs6000_adjust_cost): Add power8 support.
20226 (rs6000_issue_rate): Likewise.
20227 (insn_must_be_first_in_group): Likewise.
20228 (insn_must_be_last_in_group): Likewise.
20229 (force_new_group): Likewise.
20230 (rs6000_register_move_cost): Likewise.
20231 (rs6000_opt_masks): Likewise.
20233 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
20234 power8 capable assembler, default to power7 options.
20235 (TARGET_DIRECT_MOVE): Likewise.
20236 (TARGET_CRYPTO): Likewise.
20237 (TARGET_P8_VECTOR): Likewise.
20238 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
20239 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
20240 (VECTOR_MEM_P8_VECTOR_P): Likewise.
20241 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
20242 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
20243 (TARGET_XSCVDPSPN): Likewise.
20244 (TARGET_XSCVSPDPN): Likewsie.
20245 (TARGET_SYNC_HI_QI): Likewise.
20246 (TARGET_SYNC_TI): Likewise.
20247 (MASK_CRYPTO): Likewise.
20248 (MASK_DIRECT_MOVE): Likewise.
20249 (MASK_P8_FUSION): Likewise.
20250 (MASK_P8_VECTOR): Likewise.
20251 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
20252 temporary used by some of the direct move instructions to get two FP
20253 temporary registers does not force creation of a stack frame.
20254 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
20255 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
20256 that any VSX registers are tieable, even if they are also an
20257 Altivec vector mode.
20258 (r6000_reg_class_enum): Add wm, wr, wv constraints.
20259 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
20260 (RS6000_BTM_CRYPTO): Likewise.
20261 (RS6000_BTM_COMMON): Likewise.
20263 * config/rs6000/rs6000.md (cpu attribute): Add power8.
20264 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
20265 (enum rs6000_vector): Add power8 vector support.
20267 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20271 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
20272 (any_sibcall_could_use_r3): this and handle indirect calls.
20273 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
20275 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20277 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
20279 2013-05-22 Richard Biener <rguenther@suse.de>
20281 PR middle-end/57349
20282 * profile.c (branch_prob): Do not split blocks that are
20283 abnormally receiving from ECF_RETURNS_TWICE functions.
20285 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
20287 * recog.c (offsettable_address_addr_space_p): Fix calculation of
20288 address mode. Move pointer mode initialization to the same place.
20290 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
20292 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
20293 while it has any effect.
20295 2013-05-21 Easwaran Raman <eraman@google.com>
20297 PR tree-optimization/57322
20298 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
20299 UID of the statement added to the BB to be 1.
20301 2013-05-21 Jakub Jelinek <jakub@redhat.com>
20303 PR tree-optimization/57331
20304 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
20305 of conversion from pointer type to integral type with integer.
20307 2013-05-21 Martin Jambor <mjambor@suse.cz>
20310 * ipa-prop.c (ipa_read_node_info): Process param_used and
20311 controlled_uses in the same order as when writing.
20313 2013-05-21 Magnus Granberg <baldrick@free.fr>
20316 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
20318 2013-05-21 Richard Biener <rguenther@suse.de>
20320 PR tree-optimization/57318
20321 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
20322 estimate stmts with side-effects as likely eliminated.
20324 2013-05-21 Richard Biener <rguenther@suse.de>
20326 PR tree-optimization/57330
20327 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
20328 preserve the call stmts fntype.
20330 2013-05-21 Richard Biener <rguenther@suse.de>
20332 PR tree-optimization/57303
20333 * tree-ssa-sink.c (statement_sink_location): Improve killing
20334 stmt detection and properly handle self-assignments.
20336 2013-05-21 Christian Bruel <christian.bruel@st.com>
20338 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
20339 spanning registers. LEAF_REG_REMAP is supported only for contiguous
20340 registers. Set register size out of the PARALLEL loop.
20342 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
20345 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
20346 (*fmasf4, *fmasf4_media): New insns.
20348 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20350 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
20351 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
20352 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
20353 (mips_idiv_insns): Update the comments to say that the returned
20354 instruction counts are in units of BASE_INSN_LENGTH.
20355 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
20356 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
20357 using 2 rather than 4 as the length of indirect MIPS16 and
20358 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
20359 length of a NOP. Don't divide MIPS16 lengths by 2.
20360 (mips16_split_long_branches): Assume a branch is long if the
20361 length is greater than 4 rather than 8.
20362 * config/mips/mips.md (length): Give MIPS16 lengths directly,
20363 rather than multiplying them by 2. Multiply instruction counts
20364 by BASE_INSN_LENGTH rather than 4.
20365 (*jump_mips16, tls_get_tp_mips16_<mode>)
20366 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
20368 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20370 * config/mips/mips.md (extended_mips16): Remove branch case.
20371 (length): Remove duplicated extended_mips16 test.
20373 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
20375 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
20377 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
20379 * recog.h (Recog_data): Rename to...
20380 (recog_data_d): ...this.
20381 (recog_data): Update accordingly.
20382 * recog.c (recog_data): Likewise.
20383 * reload.c (save_recog_data): Likewise.
20384 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
20385 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
20387 2013-05-17 Julian Brown <julian@codesourcery.com>
20389 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
20390 found in a REG_EQUAL note, invalidate it.
20392 2013-05-17 Easwaran Raman <eraman@google.com>
20394 * tree-ssa-reassoc.c (find_insert_point): New function.
20395 (insert_stmt_after): Likewise.
20396 (get_def_stmt): Likewise.
20397 (ensure_ops_are_available): Likewise.
20398 (not_dominated_by): Likewise.
20399 (rewrite_expr_tree): Do not move statements beyond what is
20400 necessary. Remove call to swap_ops_for_binary_stmt...
20401 (reassociate_bb): ... and move it here.
20402 (build_and_add_sum): Assign UIDs for new statements.
20403 (linearize_expr): Likewise.
20404 (do_reassoc): Renumber gimple statement UIDs.
20406 2013-05-17 Jan Hubicka <jh@suse.cz>
20408 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
20410 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
20411 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
20413 (output_weakrefs): Update.
20415 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
20416 Martin Jambor <mjambor@suse.cz>
20418 PR middle-end/57276
20419 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
20420 value that corresponds to the given aggval is found in values vector.
20422 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
20424 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
20425 sse, sse2, sse3, ssse3 and sse4a flags to options.
20427 2013-05-17 David Malcolm <dmalcolm@redhat.com>
20429 * gengtype-state.c: (s_expr_writer): New class, to handle
20430 prettifying of output layout of s-expressions.
20431 (state_writer): New class, to write out gtype.state.
20432 (state_written_type_count): Move this variable into member data of
20434 (s_expr_writer::s_expr_writer): New code: constructor for new class
20435 (state_writer::state_writer(): ditto
20436 (s_expr_writer::write_new_line): New function
20437 (s_expr_writer::write_any_indent): ditto
20438 (s_expr_writer::begin_s_expr): ditto
20439 (s_expr_writer::end_s_expr): ditto
20440 (write_state_fileloc): convert to method of state_writer...
20441 (state_writer:: write_state_fileloc): ...and use methods of
20442 s_expr_writer to write indentation into the gtype.state output file
20443 to visually represent the hierarchical structure of the list
20445 (write_state_fields): ditto, renaming to...
20446 (state_writer::write_state_fields)
20447 (write_state_a_string): ditto, renaming to...
20448 (state_writer::write_state_a_string)
20449 (write_state_string_option): ditto, renaming to...
20450 (state_writer::write_state_string_option)
20451 (write_state_type_option): ditto, renaming to...
20452 (state_writer::write_state_type_option)
20453 (write_state_nested_option): ditto, renaming to...
20454 (state_writer::write_state_nested_option)
20455 (write_state_option): ditto, renaming to...
20456 (state_writer::write_state_option)
20457 (write_state_options): ditto, renaming to...
20458 (state_writer::write_state_options)
20459 (write_state_lang_bitmap): ditto, renaming to...
20460 (state_writer::write_state_lang_bitmap)
20461 (write_state_version): ditto, renaming to...
20462 (state_writer::write_state_version)
20463 (write_state_scalar_type): ditto, renaming to...
20464 (state_writer::write_state_scalar_type)
20465 (write_state_string_type): ditto, renaming to...
20466 (state_writer::write_state_string_type)
20467 (write_state_undefined_type): ditto, renaming to...
20468 (state_writer::write_state_undefined_type)
20469 (write_state_struct_union_type): ditto, renaming to...
20470 (state_writer::write_state_struct_union_type)
20471 (write_state_struct_type): ditto, renaming to...
20472 (state_writer::write_state_struct_type)
20473 (write_state_user_struct_type): ditto, renaming to...
20474 (state_writer::write_state_user_struct_type)
20475 (write_state_lang_struct_type): ditto, renaming to...
20476 (state_writer::write_state_lang_struct_type)
20477 (write_state_param_struct_type): ditto, renaming to...
20478 (state_writer::write_state_param_struct_type)
20479 (write_state_pointer_type): ditto, renaming to...
20480 (state_writer::write_state_pointer_type)
20481 (write_state_array_type): ditto, renaming to...
20482 (state_writer::write_state_array_type)
20483 (write_state_gc_used): ditto, renaming to...
20484 (state_writer::write_state_gc_used)
20485 (write_state_common_type_content): ditto, renaming to...
20486 (state_writer::write_state_common_type_content)
20487 (write_state_type): ditto, renaming to...
20488 (state_writer::write_state_type)
20489 (write_state_pair_list): ditto, renaming to...
20490 (state_writer::write_state_pair_list)
20491 (write_state_pair): ditto, renaming to...
20492 (state_writer::write_state_pair)
20493 (write_state_typedefs): ditto, renaming to...
20494 (state_writer::write_state_typedefs)
20495 (write_state_structures): ditto, renaming to...
20496 (state_writer::write_state_structures)
20497 (write_state_param_structs): ditto, renaming to...
20498 (state_writer::write_state_param_structs)
20499 (write_state_variables): ditto, renaming to...
20500 (state_writer::write_state_variables)
20501 (write_state_srcdir): ditto, renaming to...
20502 (state_writer::write_state_srcdir)
20503 (write_state_files_list): ditto, renaming to...
20504 (state_writer::write_state_files_list)
20505 (write_state_languages): ditto, renaming to...
20506 (state_writer::write_state_languages)
20507 (write_state): create a state_writer instance and use it when
20508 writing out the state file
20510 2013-05-17 Mike Stump <mikestump@comcast.net>
20512 PR rtl-optimization/57304
20513 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
20514 accessing DF_REF_REAL_LOC.
20516 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20518 PR rtl-optimization/57281
20519 PR rtl-optimization/57300
20520 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
20521 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
20522 what the other splitter did if the registers are dead.
20524 2013-05-17 Richard Biener <rguenther@suse.de>
20526 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
20529 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20531 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
20534 2013-05-17 Marek Polacek <polacek@redhat.com>
20536 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
20537 length when doing non-zero store of storing '\0' to '\0'.
20539 2013-05-17 Jakub Jelinek <jakub@redhat.com>
20541 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
20542 vect_external_def oprnd1 with loop_vinfo, try to emit
20543 optional cast, negation and and stmts on the loop preheader
20544 edge instead of into the pattern def seq.
20546 PR tree-optimization/57051
20547 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
20548 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
20550 2013-05-16 Nick Clifton <nickc@redhat.com>
20552 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
20553 (rl78_is_naked_func): New function.
20554 (rl78_expand_prologue): Skip prologue generation for naked functions.
20555 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
20556 * doc/extend.texi (naked): Add RL78 to the list of processors
20557 that supports this attribute.
20559 2013-05-16 Jeff Law <law@redhat.com>
20561 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
20563 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20565 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
20566 cache parameters using detect_caches_amd also for CYRIX,
20567 NSC and TM2 signatures.
20569 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
20570 Dzianis Kahanovich <mahatma@eu.by>
20574 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
20575 VIA/Centaur processors and determine their cache parameters
20576 using detect_caches_amd.
20578 2013-05-16 Teresa Johnson <tejohnson@google.com>
20580 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
20581 (rtl_verify_edges): New function.
20582 (rtl_verify_bb_insns): Ditto.
20583 (rtl_verify_bb_pointers): Ditto.
20584 (rtl_verify_bb_insn_chain): Ditto.
20585 (rtl_verify_fallthru): Ditto.
20586 (rtl_verify_bb_layout): Ditto.
20587 (rtl_verify_flow_info_1): Outline checks into new functions.
20588 (rtl_verify_flow_info): Ditto.
20590 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
20592 * cfghooks.c (copy_bbs): Add update_dominance argument.
20593 * cfghooks.h (copy_bbs): Update prototype.
20594 * tree-cfg.c (gimple_duplicate_sese_region):
20595 Add update_dominance argument.
20596 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
20597 * tree-ssa-loop-ch.c (copy_loop_headers): Update
20598 gimple_duplicate_sese_region call.
20599 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
20600 Update copy_bbs call.
20601 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
20602 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
20604 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20606 * tree-vectorizer.h (NUM_PATTERNS): Increment.
20607 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
20608 vect_recog_rotate_pattern.
20609 (vect_recog_rotate_pattern): New function.
20611 2013-05-16 Jason Merrill <jason@redhat.com>
20613 * Makefile.in (LLINKER): New variable.
20614 (mostlyclean): Remove link mutex.
20615 * configure.ac: Handle --enable-link-mutex.
20616 * lock-and-run.sh: New script.
20618 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20621 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
20624 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20626 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
20628 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
20630 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
20631 * config/arm/arm.c (next_consecutive_mem): New function.
20632 (gen_movmem_ldrd_strd): Likewise.
20633 * config/arm/arm.md (movmemqi): Update condition and code.
20634 (unaligned_loaddi, unaligned_storedi): New patterns.
20636 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20638 * config.gcc: Obsolete *-*-solaris2.9*.
20639 * doc/install.texi (Specific, *-*-solaris2*): Document it.
20641 2013-05-16 Richard Biener <rguenther@suse.de>
20643 * passes.c (init_optimization_passes): Move pass_parallelize_loops
20644 earlier, after GRAPHITE transforms and IV canonicalization.
20646 2013-05-16 Jakub Jelinek <jakub@redhat.com>
20648 * omp-low.c (extract_omp_for_data): For collapsed loops,
20649 if at least one of the loops is known at compile time to
20650 iterate zero times, set count to 0.
20651 (expand_omp_regimplify_p): New function.
20652 (expand_omp_for_generic): For collapsed loops, if at least
20653 one of the loops isn't known to iterate at least once,
20654 add runtime check with setting count to 0.
20655 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
20656 For unsigned types if it isn't known at compile time that
20657 the loop will iterate at least once, add runtime check to bypass
20658 the whole loop if initial condition isn't true.
20660 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
20662 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
20664 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
20666 PR middle-end/57286
20667 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
20668 transformations to avoid an infinite loop.
20670 2013-05-16 Marek Polacek <polacek@redhat.com>
20672 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
20674 2013-05-15 Leif Ekblad <leif@rdos.net>
20676 * config/i386/i386.c (ix86_decompose_address): Use
20677 DEFAULT_TLS_SEG_REG to access TLS segment register.
20678 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
20679 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
20680 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
20682 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
20685 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
20686 sibling calls to functions that would normally be lazily bound,
20687 unless $gp is call-clobbered.
20689 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
20691 * config/i386/i386.c (ix86_option_override_internal): Update
20692 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
20693 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
20694 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
20695 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
20697 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
20699 2013-05-15 Andreas Schwab <schwab@suse.de>
20701 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
20702 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
20705 2013-05-15 Teresa Johnson <tejohnson@google.com>
20707 * loop-unroll.c (report_unroll_peel): Check decision before
20708 emitting unroll/peel message.
20710 2013-05-15 Teresa Johnson <tejohnson@google.com>
20712 * function.h (has_bb_partition): New rtl_data flag.
20713 (bb_reorder_complete): Ditto.
20714 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
20715 instead of flag_reorder_blocks_and_partition.
20716 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
20717 with some enhancements.
20718 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
20719 * bb-reorder.c (connect_traces): Check for has_bb_partition
20720 instead of flag_reorder_blocks_and_partition.
20721 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
20722 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
20723 verify_hot_cold_block_grouping.
20724 (partition_hot_cold_basic_blocks): Set has_bb_partition.
20726 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20729 * config/arm/predicates.md (call_insn_operand): New predicate.
20730 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
20731 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
20732 if insn is not a tail call.
20733 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
20735 * config/arm/arm.h (enum reg_class): New caller save register class.
20736 (REG_CLASS_NAMES): Likewise.
20737 (REG_CLASS_CONTENTS): Likewise.
20738 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
20741 2013-05-15 Richard Biener <rguenther@suse.de>
20743 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
20744 of MSG_OPTIMIZED_LOCATIONS.
20745 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
20746 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
20748 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
20749 of MSG_OPTIMIZED_LOCATIONS.
20750 (execute_vect_slp): Likewise.
20751 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
20752 (vect_create_cond_for_alias_checks): Likewise.
20753 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
20754 (vect_recog_widen_mult_pattern): Likewise.
20755 (vect_recog_widen_sum_pattern): Likewise.
20756 (vect_recog_over_widening_pattern): Likewise.
20757 (vect_recog_widen_shift_pattern): Likewise.
20758 (vect_recog_vector_vector_shift_pattern): Likewise.
20759 (vect_recog_divmod_pattern): Likewise.
20760 (vect_recog_mixed_size_cond_pattern): Likewise.
20761 (vect_recog_bool_pattern): Likewise.
20762 (vect_pattern_recog_1): Likewise.
20764 2013-05-15 Martin Jambor <mjambor@suse.cz>
20766 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
20767 non-functions to builtin_unreachable.
20768 * ipa-inline-transform.c (inline_call): Do not assert estimates were
20769 correct when new direct edges were discovered.
20771 2013-05-15 Martin Jambor <mjambor@suse.cz>
20773 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
20774 header, print symbol order instead of node uid, print more information
20775 about indirect edge targets.
20776 (ipa_make_edge_direct_to_target): Print symbol order instead of node
20778 (ipa_make_edge_direct_to_target): Likewise.
20779 (remove_described_reference): Likewise.
20780 (propagate_controlled_uses): Likewise.
20781 (ipa_print_node_params): Also print symbol order.
20782 (ipcp_transform_function): Print symbol order instead of node uids.
20783 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
20784 (cgraph_get_create_real_symbol_node): Likewise.
20785 * ipa-cp.c (print_lattice): Likewise.
20786 (print_all_lattices): Likewise.
20787 (determine_versionability): Likewise.
20788 (initialize_node_lattices): Likewise.
20789 (estimate_local_effects): Likewise.
20790 (update_profiling_info): Likewise.
20791 (create_specialized_node): Likewise.
20792 (perhaps_add_new_callers): Likewise.
20793 (decide_about_value): Likewise.
20794 (decide_whether_version_node): Likewise.
20795 (identify_dead_nodes): Likewise.
20796 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
20797 (dump_inline_summary): Likewise.
20798 (estimate_node_size_and_time): Likewise.
20799 (inline_analyze_function): Likewise.
20800 * ipa-inline.c (report_inline_failed_reason): Likewise.
20801 (want_early_inline_function_p): Likewise.
20802 (edge_badness): Likewise.
20803 (update_edge_key): Likewise.
20804 (inline_small_functions): Likewise. Add dumping of order to two other
20806 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
20807 instead of node uids.
20808 (propagate_pure_const): Likewise.
20809 (propagate_pure_const): Likewise.
20810 * ipa-utils.c (dump_cgraph_node_set): Likewise.
20811 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
20812 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
20814 * tree-pretty-print.c (dump_function_header): Likewise.
20815 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
20816 Print symbol order instead of node uids.
20818 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20820 * config/s390/s390.c (s390_register_move_cost): Don't impose the
20821 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
20823 2013-05-15 Richard Biener <rguenther@suse.de>
20825 PR tree-optimization/57275
20826 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
20827 return value for fail to do runtime alias checks for gather loads.
20829 2013-05-15 Jan Hubicka <jh@suse.cz>
20833 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
20834 weakrefs are not external.
20835 (lto_symtab_merge_decls): Fix thinko when dealing with
20836 non-lto_symtab decls.
20837 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
20838 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
20839 * varpool.c (dump_varpool_node): Dump more flags.
20841 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
20843 * config/i386/i386.c (processor_alias_table): Add instruction
20844 FSGSBASE for AMD bdver3 architecture.
20846 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20848 * tree.c (warn_deprecated_use): Print file:line using locus color.
20849 * diagnostic.c (diagnostic_report_current_module): Print file:line
20850 and file:line:column using locus color.
20852 2013-05-14 Mike Stump <mikestump@comcast.net>
20854 * gdbinit.in: Add __null.
20856 2013-05-14 Mike Stump <mikestump@comcast.net>
20858 * recog.h: Rename struct recog_data to Recog_data.
20859 * recog.c: Likewise.
20860 * reload.c (can_reload_into): Likewise.
20861 * config/picochip/picochip.c: Likewise.
20863 2013-05-14 Mike Stump <mikestump@comcast.net>
20865 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
20867 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
20869 * resource.h (struct resources): Remove unch_memory member.
20870 (CLEAR_RESOURCE): Don't clear unch_memory.
20871 * resource.c (mark_referenced_resources): Don't set it.
20872 (mark_set_resources): Likewise.
20873 (mark_target_live_regs): Don't clear it.
20874 (init_resource_info): Likewise.
20875 * reorg.c (resource_conflicts_p): Don't compare it.
20876 (redundant_insn): Don't set it.
20878 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
20880 * emit-rtl.c (next_label): Remove unused function.
20881 (skip_consecutive_labels, link_cc0_insns): Move to ...
20882 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
20883 only place where these functions are used, and make them static.
20885 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20887 * fold-const.c (fold_negate_expr): Handle vectors.
20888 (fold_truth_not_expr): Make it static.
20889 (fold_invert_truthvalue): New static function.
20890 (invert_truthvalue_loc): Handle vectors. Do not call
20891 fold_truth_not_expr directly.
20892 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
20893 <TRUTH_NOT_EXPR>: Do not cast to boolean.
20894 (fold_comparison): Handle vector constants.
20895 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
20896 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
20897 * tree.h (fold_truth_not_expr): Remove declaration.
20899 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
20901 * config/aarch64/aarch64-simd.md
20902 (aarch64_vcond_internal<mode>): Rename to...
20903 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
20904 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
20905 float modes. Clarify all iterator modes.
20906 (vcond<mode><mode>): Use new name for vcond expanders.
20907 (vcond<v_cmp_result><mode>): Likewise.
20908 (vcondu<mode><mode>: Likewise.
20909 * config/aarch64/iterators.md (VDQF_COND): New.
20911 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
20914 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
20915 variable for the shift amount. Check that we shift by non-negative
20918 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
20921 * config/arm/arm.h (EPILOGUE_USES): Only return true
20922 for LR_REGNUM after epilogue_completed.
20924 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
20926 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
20927 is error_mark_node.
20929 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
20932 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
20933 and Solaris 11+/x86 with gld.
20934 * configure: Regenerate.
20936 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20938 * expmed.c (expand_shift_1): Canonicalize rotates by
20939 constant bitsize / 2 to bitsize - 1.
20940 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
20941 case ROTATERT>: Likewise.
20944 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20946 * config/i386/i386.md (rotateinv): New code attr.
20947 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20948 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20949 roll $31, %eax, etc.
20951 2013-05-14 Richard Biener <rguenther@suse.de>
20953 PR middle-end/57235
20954 * tree-eh.c (sink_clobbers): Give up for successors with
20955 multiple predecessors and no virtual uses.
20957 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
20959 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
20960 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
20962 2013-05-14 Jakub Jelinek <jakub@redhat.com>
20964 PR middle-end/57251
20965 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
20966 the case when both op0 and op1 have VOIDmode.
20968 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
20970 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
20971 in multiply-accumulate mode.
20973 2013-05-13 Guozhi Wei <carrot@google.com>
20975 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
20977 2013-05-13 Kai Tietz <ktietz@redhat.com>
20980 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
20981 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
20982 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
20983 * config/i386/i386.c (ix86_option_override_internal): Likewise.
20984 (ix86_expand_prologue): Likewise.
20985 (ix86_expand_split_stack_prologue): Likewise.
20986 (legitimate_pic_address_disp_p): Likewise.
20987 (legitimize_pic_address): Likewise.
20988 (legitimize_tls_address): Likewise.
20989 (legitimize_pe_coff_symbol): Likewise.
20990 (output_pic_addr_const): Likewise.
20991 (construct_plt_address): Likewise.
20992 (ix86_expand_call): Likewise.
20993 (x86_output_mi_thunk): Likewise.
20994 (x86_function_profiler): Likewise.
20996 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
20998 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
20999 similar switch cases.
21000 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
21001 (aarch64_simd_mov_to_<mode>low): Delete.
21002 (aarch64_simd_mov_to_<mode>high): Delete.
21003 (move_lo_quad_<mode>): Add w<-r alternative.
21004 (aarch64_simd_move_hi_quad_<mode>): Likewise.
21005 (aarch64_simd_mov_from_*): Update type attribute.
21006 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
21009 2013-05-13 Jan Hubicka <jh@suse.cz>
21011 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
21012 * config/i386/i386.c (ix86_compute_frame_layout,
21013 ix86_expand_epilogue, emit_i387_cw_initialization,
21014 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
21015 ix86_local_alignment): Fix use of size/speed predicates.
21017 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21019 PR tree-optimization/45216
21020 PR tree-optimization/57157
21021 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
21022 the (-Y) & (B - 1) variant if OP is |.
21023 * expmed.c (expand_shift_1): For rotations by const0_rtx just
21024 return shifted. Use (-op1) & (prec - 1) as other_amount
21025 instead of prec - op1.
21027 2013-05-13 Martin Jambor <mjambor@suse.cz>
21029 PR middle-end/42371
21030 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
21031 (ipa_constant_data): New type.
21032 (ipa_jump_func): Use ipa_constant_data to hold information about
21033 constant jump functions.
21034 (ipa_get_jf_constant): Adjust to jump function type changes.
21035 (ipa_get_jf_constant_rdesc): New function.
21036 (ipa_param_descriptor): New field controlled_uses.
21037 (ipa_get_controlled_uses): New function.
21038 (ipa_set_controlled_uses): Likewise.
21039 * ipa-ref.h (ipa_find_reference): Declare.
21040 * ipa-prop.c (ipa_cst_ref_desc): New type.
21041 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
21043 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
21044 New parameter cs. Adjust all callers.
21045 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
21046 (remove_described_reference): New function.
21047 (jfunc_rdesc_usable): Likewise.
21048 (try_make_edge_direct_simple_call): Decrement controlled use count,
21049 attempt to remove reference if it hits zero.
21050 (combine_controlled_uses_counters): New function.
21051 (propagate_controlled_uses): Likewise.
21052 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
21053 (ipa_edge_duplication_hook): Duplicate reference descriptions.
21054 (ipa_print_node_params): Print described use counter.
21055 (ipa_write_jump_function): Adjust to jump function type changes.
21056 (ipa_read_jump_function): New parameter CS, pass it to
21057 ipa_set_jf_constant. Adjust caller.
21058 (ipa_write_node_info): Stream controlled use count
21059 (ipa_read_node_info): Likewise.
21060 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
21062 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
21063 count. Remove cloning-added reference if it reaches zero.
21064 * ipa-ref.c (ipa_find_reference): New function.
21066 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
21068 * config/i386/i386.c (processor_target_table): Modified default
21069 alignment values for AMD BD and BT architectures.
21071 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
21073 * tree-vect-generic.c (uniform_vector_p): Move ...
21074 * tree.c (uniform_vector_p): ... here.
21075 * tree.h (uniform_vector_p): Declare it.
21076 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
21079 2013-05-13 Jakub Jelinek <jakub@redhat.com>
21081 PR tree-optimization/57230
21082 * tree-ssa-strlen.c (handle_char_store): Record length for
21083 array store from STRING_CST.
21085 PR tree-optimization/57230
21086 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
21089 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
21091 * config/epiphany/epiphany.c (epiphany_init): Check size of
21092 NUM_MODES_FOR_MODE_SWITCHING.
21093 (epiphany_expand_prologue):
21094 Remove CONFIG_REGNUM initial value handling code.
21095 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
21096 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
21097 (emit_set_fp_mode, epiphany_mode_after): Likewise.
21098 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
21099 Don't return 1 for FP_MODE_NONE.
21100 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
21101 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
21102 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
21103 * config/epiphany/epiphany.md (save_config): New pattern.
21105 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
21107 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
21109 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
21111 * config/i386/i386.md (memory): Handle sseishft1.
21112 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
21113 (*vec_extractv2di_1): Ditto.
21115 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
21117 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
21120 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
21122 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
21123 Add mthumb/march=armv7-a multilib.
21124 Add mthumb/march=armv7-r multilib.
21125 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
21127 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
21129 * config/v850/t-rtems: Add more multilibs.
21131 2013-05-10 Richard Biener <rguenther@suse.de>
21133 PR tree-optimization/57214
21134 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
21135 not propagate from SSA names that occur in abnormal PHI nodes.
21137 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
21139 * stor-layout.c (element_precision): New function.
21140 * machmode.h (element_precision): Declare it.
21141 * tree.c (build_minus_one_cst): New function.
21142 (element_precision): Likewise.
21143 * tree.h (build_minus_one_cst): Declare new function.
21144 (element_precision): Likewise.
21145 * fold-const.c (operand_equal_p): Use element_precision.
21146 (fold_binary_loc): Handle vector types.
21147 * convert.c (convert_to_integer): Use element_precision.
21148 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
21151 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
21153 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
21154 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
21155 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
21156 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21157 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
21158 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
21159 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
21160 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
21161 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
21162 (Uuw8): New constraints.
21163 (Usb4): Move into alphabetical order.
21164 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
21165 (sd8_operand, ub8_operand, uw8_operand): New predicates.
21166 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
21167 previously unnamed patterns.
21168 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
21169 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
21170 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
21171 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
21172 of set_attr_alternative/if_then_else. Use extended_mips16 instead
21173 of specific lengths.
21175 2013-05-10 Jakub Jelinek <jakub@redhat.com>
21177 * config/i386/i386.md (rotateinv): New code attr.
21178 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
21179 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
21180 roll $31, %eax, etc.
21182 PR tree-optimization/45216
21183 PR tree-optimization/57157
21184 * tree-ssa-forwprop.c (simplify_rotate): New function.
21185 (ssa_forward_propagate_and_combine): Call it.
21187 2013-05-10 Richard Biener <rguenther@suse.de>
21189 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
21190 disable peeling when we version for aliasing.
21191 (vector_alignment_reachable_p): Honor explicit user alignment.
21192 (vect_supportable_dr_alignment): Likewise.
21193 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
21194 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
21195 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
21196 then peeling to arrange for the cost-model check to come first.
21198 2013-05-10 Alan Modra <amodra@gmail.com>
21200 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
21201 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
21202 * configure: Regenerate.
21204 2013-05-10 Alan Modra <amodra@gmail.com>
21207 * varasm.c (default_elf_select_section): Move !DECL_P check..
21208 (get_named_section): ..to here before calling get_section_name.
21210 (default_section_type_flags): Add DECL_P check.
21211 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
21212 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
21214 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
21216 * config/epiphany/epiphany.c (epiphany_expand_prologue):
21217 When using gen_stack_adjust_str with a register offset, add a
21218 REG_FRAME_RELATED_EXPR note.
21220 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
21222 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
21223 (*vec_extractv4si_zext_mem): Ditto.
21224 (*vec_extractv2di): Add 0->x and x->x alternatives.
21225 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
21226 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
21228 2013-05-09 Jason Merrill <jason@redhat.com>
21230 N3639 C++1y VLA support
21231 * gimplify.c (gimplify_vla_decl): Don't touch an existing
21234 * tree.c (build_constructor_va): New.
21235 * tree.h: Declare it.
21237 2013-05-09 Martin Jambor <mjambor@suse.cz>
21240 * gimple-fold.c (canonicalize_constructor_val): Call
21241 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
21243 2013-05-09 Jan Hubicka <jh@suse.cz>
21244 Richard Biener <rguenther@suse.de>
21247 * symtab.c (symtab_make_decl_local): Do not add private names.
21249 2013-05-09 Jan Hubicka <jh@suse.cz>
21252 * symtab.c (insert_to_assembler_name_hash): Handle clones.
21253 (unlink_from_assembler_name_hash): Likewise.
21254 (symtab_prevail_in_asm_name_hash, symtab_register_node,
21255 symtab_unregister_node, symtab_initialize_asm_name_hash,
21256 change_decl_assembler_name): Update.
21258 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
21260 * config/aarch64/aarch64.md: New movtf split.
21261 (*movtf_aarch64): Update.
21262 (aarch64_movdi_tilow): Handle TF modes and rename to
21263 aarch64_movdi_<mode>low.
21264 (aarch64_movdi_tihigh): Handle TF modes and rename to
21265 aarch64_movdi_<mode>high
21266 (aarch64_movtihigh_di): Handle TF modes and rename to
21267 aarch64_mov<mode>high_di
21268 (aarch64_movtilow_di): Handle TF modes and rename to
21269 aarch64_mov<mode>low_di
21270 (aarch64_movtilow_tilow): Remove spurious whitespace.
21271 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
21273 (aarch64_print_operand): Update.
21275 2013-05-09 Alan Modra <amodra@gmail.com>
21277 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
21279 * configure: Regenerate.
21281 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21283 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
21284 splitter preparation statements.
21285 * config/i386/sse.md (*vec_extract* splitters): Ditto.
21286 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
21289 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21291 * gimple-ssa-strength-reduction.c (count_candidates): Change
21292 return value to int.
21293 (analyze_candidates_and_replace): Change type of length to int.
21295 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
21297 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
21298 (*vec_extract<mode>): Use VI12_128 mode iterator.
21299 (*vec_extract<mode>_mem): Ditto.
21300 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
21303 2013-05-08 Diego Novillo <dnovillo@google.com>
21308 2012-08-17 Diego Novillo <dnovillo@google.com>
21311 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
21312 * config.in: Regenerate.
21313 * configure: Regenerate.
21314 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
21316 2013-05-08 Jan Hubicka <jh@suse.cz>
21319 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
21320 * cgraph.h (symtab_node_base): Add unique_name.
21321 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
21322 input_overwrite_node, input_varpool_node): Stream unique_name.
21323 * cgraphclones.c (cgraph_create_virtual_clone,
21324 cgraph_function_versioning): Set unique_name.
21325 * ipa.c (function_and_variable_visibility): Set unique_name.
21327 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21329 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
21330 (alloc_cand_and_find_basis): Restrict conditional candidate
21331 processing to CAND_MULTs.
21333 2013-05-08 Jan Hubicka <jh@suse.cz>
21336 lto-symtab.c (lto_symtab_symbol_p): New function.
21337 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
21338 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
21339 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
21340 Skip static symbols.
21342 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
21344 PR tree-optimization/57200
21345 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
21346 Only call inform if the preceding warning_at returns true.
21348 2013-05-07 Han Shen <shenhan@google.com>
21350 * cfgexpand.c (record_or_union_type_has_array_p): New function.
21351 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
21352 * common.opt (fstack-protector-strong): New option.
21353 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
21354 * doc/invoke.texi (Optimization Options): Document
21355 "-fstack-protector-strong".
21356 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
21358 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21360 * config/mips/mips.c (mips_machine_reorg2): Return 0.
21362 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
21364 * ira.c (update_equiv_regs): Add insn having equiv memory even if
21365 it is not lhs of the insn.
21366 (setup_reg_equiv): Remove insn having equiv memory which it is not
21368 * lra-constraints.c (process_address): Try to improve generation
21369 code for address base + disp.
21370 (lra_constraints): Make correct the code for checking insn setting
21371 up backward equivalence. Remove insn only if it is in the init
21373 * lra-eliminations.c (update_reg_eliminate): Change return value.
21374 (lra_eliminate): Use the result.
21376 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
21378 * config/i386/sse.md (ssescalarnummask): New mode attribute.
21379 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
21380 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
21381 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
21382 register target operands.
21383 (*vec_extractv8hi_sse2): New pattern.
21384 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
21385 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
21386 (*vec_extract<mode>_mem): New insn and split pattern.
21388 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
21390 * config/arm/arm.c (arm_asan_shadow_offset): New function.
21391 (TARGET_ASAN_SHADOW_OFFSET): Define.
21392 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
21393 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
21395 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21397 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
21398 (incr_vec_index): Return -1 if increment not found.
21399 (create_add_on_incoming_edge): Assert if increment not found.
21400 (record_increment): Limit number of increments recorded.
21401 (all_phi_incrs_profitable): Return false if an increment not found.
21402 (replace_profitable_candidates): Don't process increments that were
21404 (analyze_candidates_and_replace): Limit size of incr_vec.
21406 2013-05-07 Richard Biener <rguenther@suse.de>
21408 * calls.c (special_function_p): setjmp-like functions are leaf.
21409 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
21410 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
21412 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
21414 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
21415 (aarch64_simd_mov<mode>): New expander.
21416 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
21417 (aarch64_simd_mov_to_<mode>high): Likewise.
21418 (aarch64_simd_mov_from_<mode>low): Likewise.
21419 (aarch64_simd_mov_from_<mode>high): Likewise.
21420 (aarch64_dup_lane<mode>): Update.
21421 (aarch64_dup_lanedi): New instruction pattern.
21422 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
21423 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
21425 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21427 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
21428 (replace_mult_candidate): Remove unnecessary argument; remove
21429 unnecessary parameter from call to introduce_cast_before_cand.
21430 (replace_unconditional_candidate): Remove unnecessary parameter
21431 from call to replace_mult_candidate.
21432 (replace_conditional_candidate): Likewise.
21433 (insert_initializers): Use make_temp_ssa_name.
21434 (introduce_cast_before_cand): Remove unnecessary argument; use
21435 make_temp_ssa_name.
21436 (replace_one_candidate): Remove unnecessary argument; remove
21437 unnecessary parameter from calls to introduce_cast_before_cand.
21438 (replace_profitable_candidates): Remove unnecessary parameters
21439 from calls to replace_one_candidate.
21441 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21443 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
21444 phi def as possibly hiding a basis for a CAND_ADD whose operands
21445 have been commuted in the analysis.
21446 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
21448 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21450 * config/aarch64/aarch64.md
21451 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
21452 shift value between 0-4.
21454 2013-05-07 Richard Biener <rguenther@suse.de>
21456 * double-int.h (rshift): New overload.
21457 * double-int.c (rshift): New function.
21458 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
21459 (create_reference_ops_from_ref): Remove.
21460 (vn_reference_insert): Use shared ops for constructing the
21461 reference and copy it.
21463 2013-05-07 Richard Biener <rguenther@suse.de>
21465 PR middle-end/57190
21466 * tree-eh.c (sink_clobbers): Properly propagate
21467 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
21469 2013-05-07 Jakub Jelinek <jakub@redhat.com>
21471 PR tree-optimization/57149
21472 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
21473 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
21474 collect_phi_def_edges, execute_late_warn_uninitialized): Use
21475 uninit_undefined_value_p instead of ssa_undefined_value_p.
21478 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
21479 for modifier == EXPAND_INITIALIZER.
21481 2013-05-07 Anton Blanchard <anton@samba.org>
21483 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
21484 for powerpc64 little endian.
21485 * configure: Regenerate.
21487 2013-05-06 Graham Stott <grahams@btinternet.com>
21489 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
21490 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
21491 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
21492 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
21494 2013-05-06 Graham Stott <grahams@btinternet.com>
21496 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
21497 codes which allow non-lvalues.
21499 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
21501 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
21502 components are all 1s.
21503 (integer_minus_onep): New function.
21504 * tree.h (integer_minus_onep): Declare it.
21505 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
21506 integer_minus_onep instead of integer_all_onesp.
21508 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21511 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
21512 variations of these patterns.
21514 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21516 * config/i386/i386.md (isa): Add x64_sse4 member.
21517 (enabled): Handle x64_sse4.
21518 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
21519 instruction for 64bit SSE4_1 targets. Update insn attributes.
21520 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
21521 instruction for SSE4_1 targets. Update insn attributes.
21522 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21523 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
21524 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
21526 (*vec_extractv4si): Rename from *sse4_1_pextrd.
21527 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
21528 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
21530 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21533 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
21535 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
21537 * final.c (do_assembler_dialects): Don't handle curly braces and
21538 vertical bar escaped by % as dialect delimiters.
21539 (output_asm_insn): Print curly braces and vertical bar if escaped
21540 by % and ASSEMBLER_DIALECT defined.
21541 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
21542 * doc/tm.texi: Regenerated.
21544 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
21546 * config/mips/mips.c: Include tree-pass.h.
21547 (mips_reorg): Split in pre- and post-dbr_schedule parts.
21548 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
21549 (pass_mips_machine_reorg2): New machine specific pass.
21550 (insert_pass_mips_machine_reorg2): New pass plugin definition.
21551 (mips_option_override): Register the new pass.
21552 * rtl.h (cleanup_barriers): Remove prototype.
21553 (dbr_schedule): Likewise.
21554 * jump.c (cleanup_barriers): Make static.
21555 * reorg.c (dbr_schedule): Likewise.
21557 2013-05-06 Richard Biener <rguenther@suse.de>
21559 PR tree-optimization/57185
21560 * tree-parloops.c (add_field_for_reduction): Handle anonymous
21561 SSA names properly.
21563 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
21566 * config/i386/i386.c (add_parameter_dependencies): Add dependence
21567 between "first_arg" and "insn", not "last" and "insn".
21569 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21571 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
21572 (find_candidates_in_block): Re-enable slsr_process_phi.
21573 (create_phi_basis): Fix double counting of candidate adjustment.
21575 2013-05-06 Richard Biener <rguenther@suse.de>
21577 PR middle-end/57147
21578 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
21579 the edge is also fallthru, preserve it and just clear the
21581 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
21582 also complex, preserve that and just clear the fallthru flag.
21583 * tree-inline.c (update_ssa_across_abnormal_edges): Also
21584 update virtual operands.
21586 2013-05-06 Alan Modra <amodra@gmail.com>
21588 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
21589 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
21590 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
21591 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
21592 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21593 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
21595 2013-05-06 Alan Modra <amodra@gmail.com>
21597 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
21598 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
21599 (DEFAULT_ASM_ENDIAN): Define.
21600 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
21601 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
21602 Update -K PIC clause from sysv4.h.
21603 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
21604 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
21606 2013-05-06 Alan Modra <amodra@gmail.com>
21608 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
21609 twice for little-endian.
21610 (ashrdi3_no_power, ashrdi3): Support little-endian.
21612 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
21615 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
21616 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
21618 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
21619 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
21620 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
21623 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
21624 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
21626 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
21627 * config/pa/pa.opt: Make mbig-switch a no-op.
21628 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
21629 (CASE_VECTOR_MODE): Always return SImode.
21630 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
21631 for the !TARGET_BIG_SWITCH case.
21632 * config/pa/pa-linux.h: Likewise.
21633 * config/pa/pa-openbsd.h: Likewise.
21634 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
21635 * config/pa/pa.md (short_jump): Remove define_insn.
21636 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
21637 (casesi0): Remove define_insn.
21638 (type): Remove btable_branch.
21639 (pa_combine_type): Likewise.
21640 (in_nullified_branch_delay): Likewise.
21641 (in_call_delay): Likewise.
21642 (define_delay): Likewise.
21643 (define_insn_reservation "Z3"): Likewise.
21644 (define_insn_reservation "Z4"): Likewise.
21645 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
21646 (pa_adjust_insn_length): Remove adjustment for btable branches.
21647 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
21650 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
21652 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
21653 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
21654 Add m->r,x alternatives.
21655 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
21656 splitters using SWI48x mode iterator.
21657 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
21658 TARGET_64BIT. Add m->x alternative.
21659 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
21660 Add o->x alternative. Enable for TARGET_SSE.
21661 (sse_storeq): Remove expander.
21662 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
21663 with memory input operand.
21664 (*vec_extractv2di_1 splitter): New.
21665 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
21666 * config/i386/i386.md (ssevecmodelower): New mode attribute.
21668 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21670 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
21671 (INT_LOWPART): Delete.
21672 (extract_MB): Adjust.
21673 (extract_ME): Adjust.
21674 (print_operand): Adjust.
21676 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
21678 * config/rs6000/predicates.md (reg_or_add_cint_operand,
21679 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21680 (reg_or_logical_cint_operand, easy_fp_constant,
21681 logical_const_operand): Delete "CONST_DOUBLE" case.
21682 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
21683 "HOST_BITS_PER_WIDE_INT == 64" test.
21684 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
21685 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
21686 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
21687 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
21689 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
21690 CONST_DOUBLE DImode/VOIDmode case.
21691 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
21692 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21693 CONST_DOUBLE VOIDmode case.
21694 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
21695 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
21696 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
21697 Delete CONST_DOUBLE case.
21698 (splitters for mov FMOVE64 const_double): Delete
21699 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
21700 "HOST_BITS_PER_WIDE_INT >= 64" test.
21701 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
21703 (mov DI const_double): Delete.
21705 2013-05-04 Jakub Jelinek <jakub@redhat.com>
21707 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
21708 on op shows all bits zero in mode of a lowpart subreg, return zero.
21710 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
21713 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
21714 to save TFmode registers and DImode to save TImode registers for
21715 caller save operations.
21716 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
21717 mark being partially clobbered since they only use the first
21720 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
21721 and TDmode only use the upper 64-bits of each VSX register.
21723 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21725 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
21726 (find_candidates_in_block): Disable slsr_process_phi.
21728 2013-05-03 Guozhi Wei <carrot@google.com>
21730 * coverage.c (coverage_obj_init): Move the construction of gcov
21732 (build_init_ctor): ... here.
21734 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
21736 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
21737 (slsr_cand_d): Redefine def_phi.
21738 (stride_status, phi_adjust_status, count_phis_status): New enums.
21739 (find_phi_def): New.
21740 (find_basis_for_base_expr): New.
21741 (find_basis_for_candidate): Handle hidden bases.
21742 (alloc_cand_and_find_basis): Handle phi candidates.
21743 (slsr_process_phi): New.
21744 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
21745 (create_mul_imm_cand): Likewise.
21746 (create_add_ssa_cand): Exclude phi base candidates.
21747 (create_add_imm_cand): Likewise.
21748 (slsr_process_cast): Likewise.
21749 (slsr_process_copy): Likewise.
21750 (find_candidates_in_block): Handle phi candidates.
21751 (dump_candidate): Likewise.
21752 (unconditional_cands): Delete.
21753 (unconditional_cands_with_known_stride_p): Delete.
21754 (phi_dependent_cand_p): New.
21755 (cand_increment): Handle phi-dependent candidates.
21756 (replace_dependent): Delete.
21757 (replace_mult_candidate): New.
21758 (replace_unconditional_candidate): New.
21759 (incr_vec_index): Move to avoid forward reference.
21760 (create_add_on_incoming_edge): New.
21761 (create_phi_basis): New.
21762 (replace_dependents): Delete.
21763 (replace_conditional_candidate): New.
21764 (phi_add_costs): New.
21765 (replace_uncond_cands_and_profitable_phis): New.
21766 (record_increment): Handle phi adjustments.
21767 (record_phi_increments): New.
21768 (record_increments): Handle phi adjustments.
21769 (phi_incr_cost): New.
21770 (lowest_cost_path): Handle phis.
21771 (total_savings): Likewise.
21772 (analyze_increments): Likewise.
21773 (ncd_with_phi): New.
21774 (ncd_of_cand_and_phis): New.
21775 (nearest_common_dominator_for_cands): Handle phi increments.
21776 (all_phi_incrs_profitable): New.
21777 (replace_profitable_candidates): Handle phi-dependent candidates.
21778 (analyze_candidates_and_replace): Likewise.
21780 2013-05-03 Teresa Johnson <tejohnson@google.com>
21783 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
21784 do not exceed REG_BR_PROB_BASE.
21786 2013-05-03 Jeff Law <law@redhat.com>
21788 PR tree-optimization/57144
21789 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
21790 operand of the condition will bit into the new type when eliminating
21791 a cast feeding a condition.
21793 2013-05-03 Jakub Jelinek <jakub@redhat.com>
21795 PR rtl-optimization/57130
21796 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
21797 of COMPARE as in_code to the recursive call if needed.
21799 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
21801 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
21802 (enabled): Handle new members.
21803 * config/i386/sse.md (*vec_concatv2si): Merge from
21804 *vec_concatv2si_sse2 and vec_concatv2si_sse.
21805 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
21807 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
21809 PR tree-optimization/57027
21810 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
21811 for fnms opportunity, check we got the prerequisite kind
21812 of tree / gimple before using accessor functions.
21814 2013-05-03 Richard Biener <rguenther@suse.de>
21816 * double-int.h (lshift): New overload without precision
21817 and arith argument.
21818 (operator *=, operator +=, operator -=): Move ...
21819 * double-int.c (operator *=, operator +=, operator -=): ... here
21820 and implement more efficiently.
21821 (mul_double_with_sign): Remove.
21822 (lshift_double): Adjust to take unsinged shift argument, push
21823 dispatching code to callers.
21824 (mul_double_wide_with_sign): Add early out for callers that
21825 are not interested in high parts or overflow.
21826 (lshift): New function.
21827 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
21828 dispatch code here.
21829 (lrotate, rrotate): Use logical shifts.
21830 * expr.c (get_inner_reference): Use lshift.
21831 * fixed-value.c (do_fixed_divide): Likewise.
21832 * tree-dfa.c (get_ref_base_and_extent): Likewise.
21833 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
21834 (indirect_refs_may_alias_p): Likewise.
21835 (stmt_kills_ref_p_1): Likewise.
21837 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21839 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
21841 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
21843 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
21844 scalar form of FABD instruction.
21846 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
21848 * lra-constraints.c (process_alt_operands): Add checking alt
21849 number to choose the best alternative.
21851 2013-05-02 Richard Biener <rguenther@suse.de>
21853 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
21854 bitmap and its handling.
21855 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
21857 2013-05-02 Richard Biener <rguenther@suse.de>
21859 PR middle-end/57140
21860 * tree-inline.c (copy_loops): Properly handle removed loops.
21861 (copy_cfg_body): Mark destination loops for fixup if source
21862 loops needed fixup.
21864 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
21867 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
21868 generating simple_return for naked functions.
21870 2013-05-02 Martin Jambor <mjambor@suse.cz>
21872 PR middle-end/56988
21873 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
21874 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
21876 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
21877 ipa_agg_replacement_value structures.
21878 (known_aggs_to_agg_replacement_list): Likewise.
21879 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
21880 (read_agg_replacement_chain): Likewise.
21881 (ipcp_transform_function): Also check that by_ref flags match.
21883 2013-05-02 Richard Biener <rguenther@suse.de>
21885 * graphds.h (struct graph): Add obstack member.
21886 * graphds.c (new_graph): Initialize obstack and allocate
21888 (add_edge): Allocate edge from the obstack.
21889 (free_graph): Free the obstack instead of all edges and vertices.
21891 2013-05-02 Teresa Johnson <tejohnson@google.com>
21893 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
21895 * cfg.c (update_bb_profile_for_threading): Ditto.
21896 * tree-inline.c (copy_bb): Ditto.
21897 (copy_edges_for_bb): Ditto.
21898 (initialize_cfun): Ditto.
21899 (copy_cfg_body): Ditto.
21900 (expand_call_inline): Ditto.
21901 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
21902 (estimate_node_size_and_time): Ditto.
21903 (inline_merge_summary): Ditto.
21904 * cgraphclones.c (cgraph_clone_edge): Ditto.
21905 (cgraph_clone_node): Ditto.
21906 * sched-rgn.c (compute_dom_prob_ps): Ditto.
21907 (compute_trg_info): Ditto.
21909 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21911 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
21912 S reg when fp attribute set.
21913 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
21915 2013-05-02 Ian Bolton <ian.bolton@arm.com>
21917 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
21919 (*and_one_cmplsi3_compare0_uxtw): Likewise.
21920 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
21921 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
21923 2013-05-02 Richard Biener <rguenther@suse.de>
21925 * tree-scalar-evolution.c (scev_info_hasher): Remove.
21926 (struct instantiate_cache_entry): New type.
21927 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
21928 (struct instantiate_cache_type): New type.
21929 (set_instantiated_value, get_instantiated_value): Remove.
21930 (get_instantiated_value_entry): New function.
21931 (instantiate_scev_name): Use the new cache and adjust.
21932 (instantiate_scev_poly): Adjust.
21933 (instantiate_scev_binary): Likewise.
21934 (instantiate_array_ref): Likewise.
21935 (instantiate_scev_convert): Likewise.
21936 (instantiate_scev_not): Likewise.
21937 (instantiate_scev_3): Likewise.
21938 (instantiate_scev_2): Likewise.
21939 (instantiate_scev_r): Likewise.
21940 (instantiate_scev): Likewise.
21941 (resolve_mixers): Likewise.
21943 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
21946 * lra-constraints.c (best_small_class_operands_num): Remove.
21947 (process_alt_operands): Remove small_class_operands_num. Take
21948 small classes operands into losers and only if the operand is not
21949 matched. Modify debugging output.
21950 (curr_insn_transform): Remove best_small_class_operands_num.
21953 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21955 * config/aarch64/aarch64-builtins.c
21956 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
21957 * config/aarch64/aarch64-simd-builtins.def
21958 (reduc_splus_): Add new modes.
21959 (reduc_uplus_): New.
21960 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
21961 (reduc_uplus_v4sf): Likewise.
21962 (reduc_splus_v4sf): Likewise.
21963 (aarch64_addv<mode>): Likewise.
21964 (reduc_uplus_<mode>): Likewise.
21965 (reduc_splus_<mode>): Likewise.
21966 (aarch64_addvv2di): Likewise.
21967 (reduc_uplus_v2di): Likewise.
21968 (reduc_splus_v2di): Likewise.
21969 (aarch64_addvv2si): Likewise.
21970 (reduc_uplus_v2si): Likewise.
21971 (reduc_splus_v2si): Likewise.
21972 (reduc_<sur>plus_<mode>): New.
21973 (reduc_<sur>plus_v2di): Likewise.
21974 (reduc_<sur>plus_v2si): Likewise.
21975 (reduc_<sur>plus_v4sf): Likewise.
21976 (aarch64_addpv4sf): Likewise.
21977 * config/aarch64/arm_neon.h
21978 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
21979 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
21980 add UNSPEC_SADDV, UNSPEC_UADDV.
21982 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
21984 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21986 * config/aarch64/arm_neon.h
21987 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
21989 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21991 * config/aarch64/aarch64-builtins
21992 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
21994 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
21996 * config/aarch64/aarch64-simd-builtins.def
21997 (reduc_smax_): New.
21998 (reduc_smin_): Likewise.
21999 (reduc_umax_): Likewise.
22000 (reduc_umin_): Likewise.
22001 (reduc_smax_nan_): Likewise.
22002 (reduc_smin_nan_): Likewise.
22005 (smax): Update for V2SF, V4SF and V2DF modes.
22008 (smin_nan): Likewise.
22009 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
22010 (<su><maxmin><mode>3): ...This, refactor.
22011 (s<maxmin><mode>3): New.
22012 (<maxmin_uns><mode>3): Likewise.
22013 (reduc_<maxmin_uns>_<mode>): Refactor.
22014 (reduc_<maxmin_uns>_v4sf): Likewise.
22015 (reduc_<maxmin_uns>_v2si): Likewise.
22016 (aarch64_<fmaxmin><mode>: Remove.
22017 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
22019 (vmin<q>_f<32,64>): Likewise.
22020 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
22022 (su): Add mappings for smax, smin, umax, umin.
22024 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
22025 (FMAXMIN): Rename as...
22026 (FMAXMIN_UNS): ...This.
22028 (fmaxminv): Likewise.
22029 (fmaxmin): Likewise.
22031 (maxmin_uns_op): Likewise.
22033 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22035 * config/aarch64/arm_neon.h
22036 (vac<ge, gt><sd>_f<32, 64>): Rename to...
22037 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
22038 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
22040 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22042 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
22043 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
22045 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22047 * config/aarch64/aarch64-simd.md
22048 (vcond<mode>_internal): Handle special cases for constant masks.
22049 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
22050 (vcondu<mode><mode>): Likewise.
22051 (vcond<v_cmp_result><mode>): New.
22053 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22055 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
22056 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
22057 * config/aarch64/aarch64-simd-builtins.def
22058 (cmeq): Update to BUILTIN_VALLDI.
22063 * config/aarch64/arm_neon.h
22064 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
22065 to builtins or C as appropriate.
22067 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
22069 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
22071 (cmhi): Rename to...
22073 * config/aarch64/aarch64-simd.md
22074 (simd_mode): Add SF.
22075 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
22076 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
22077 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
22078 (cstore<mode>_neg): ...This.
22079 * config/aarch64/iterators.md
22081 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
22082 (COMPARISONS): New.
22083 (UCOMPARISONS): Likewise.
22084 (optab): Add missing comparisons.
22090 (VCMP_S): Likewise.
22091 (VCMP_U): Likewise.
22092 (V_cmp_result): Add DF, SF modes.
22093 (v_cmp_result): Likewise.
22095 (vmtype): Likewise.
22096 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
22098 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
22100 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
22101 define_insn to define_insn_and_split.
22102 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
22103 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
22104 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
22105 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
22106 (thumb2_negscc): Likewise.
22108 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22110 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
22112 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
22114 * config/arm/thumb2.md: Remove trailing whitespaces.
22116 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
22118 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
22119 Use gen_int_mode rather than GEN_INT.
22121 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
22123 * value-prof.c (stream_in_histogram_value): Remove the strayed
22126 2013-04-30 Richard Biener <rguenther@suse.de>
22128 PR middle-end/57122
22129 * cfghooks.c (split_edge): Properly check for the loop latch edge.
22131 2013-04-30 Richard Biener <rguenther@suse.de>
22133 PR middle-end/57107
22134 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
22136 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
22138 PR rtl-optimization/56957
22139 PR rtl-optimization/57105
22140 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
22141 variable. Use just INSN_UID for determining whether an insn
22142 should be only disconnected from the insn stream.
22143 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
22145 2013-04-30 Jakub Jelinek <jakub@redhat.com>
22147 PR tree-optimization/57104
22148 * tsan.c (instrument_expr): Don't instrument accesses to
22149 DECL_HARD_REGISTER VAR_DECLs.
22151 2013-04-30 Richard Biener <rguenther@suse.de>
22153 * function.h (loops_for_fn): New inline function.
22154 (set_loops_for_fn): Likewise.
22155 * cfgloop.h (place_new_loop): Add struct function parameter.
22156 (get_loop): Likewise.
22157 (get_loops): Likewise.
22158 (number_of_loops): Likewise.
22159 (fel_next): Adjust.
22160 (fel_init): Likewise.
22161 * cfg.c (get_loop_copy): Adjust.
22162 * cfgloop.c (flow_loops_dump): Likewise.
22163 (record_loop_exits): Likewise.
22164 (verify_loop_structure): Likewise.
22165 * cfgloopanal.c (mark_irreducible_loops): Likewise.
22166 (estimate_reg_pressure_cost): Likewise.
22167 (mark_loop_exit_edges): Likewise.
22168 * cfgloopmanip.c (place_new_loop): Likewise.
22169 (add_loop): Likewise.
22170 (duplicate_loop): Likewise.
22171 * graph.c (draw_cfg_nodes): Likewise.
22172 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
22173 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
22174 (extract_affine_chrec): Likewise.
22175 (build_scop_iteration_domain): Likewise.
22176 * graphite.c (graphite_initialize): Likewise.
22177 * ira-build.c (create_loop_tree_nodes): Likewise.
22178 (more_one_region_p): Likewise.
22179 (rebuild_regno_allocno_maps): Likewise.
22180 (mark_loops_for_removal): Likewise.
22181 (mark_all_loops_for_removal): Likewise.
22182 (remove_unnecessary_regions): Likewise.
22183 (ira_build): Likewise.
22184 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
22185 * loop-init.c (fix_loop_structure): Likewise.
22186 (gate_rtl_move_loop_invariants): Likewise.
22187 (gate_rtl_unswitch): Likewise.
22188 (gate_rtl_unroll_and_peel_loops): Likewise.
22189 (rtl_doloop): Likewise.
22190 * lto-streamer-in.c (input_cfg): Likewise.
22191 * lto-streamer-out.c (output_cfg): Likewise.
22192 * modulo-sched.c (sms_schedule): Likewise.
22193 * predict.c (tree_estimate_probability): Likewise.
22194 (tree_estimate_probability_driver): Likewise.
22195 (estimate_loops): Likewise.
22196 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
22197 (move_sese_region_to_fn): Likewise.
22198 (debug_loop_num): Likewise.
22199 * tree-chrec.c (chrec_evaluate): Likewise.
22200 (hide_evolution_in_other_loops_than_loop): Likewise.
22201 (chrec_component_in_loop_num): Likewise.
22202 (reset_evolution_in_loop): Likewise.
22203 (evolution_function_is_invariant_rec_p): Likewise.
22204 * tree-if-conv.c (main_tree_if_conversion): Likewise.
22205 * tree-inline.c (copy_loops): Likewise.
22206 (copy_cfg_body): Likewise.
22207 (tree_function_versioning): Likewise.
22208 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
22209 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
22211 (add_to_evolution_1): Likewise.
22212 (scev_const_prop): Likewise.
22213 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
22214 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
22215 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
22216 (tree_ssa_lim_initialize): Likewise.
22217 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
22218 (verify_loop_closed_ssa): Likewise.
22219 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
22220 (tree_ssa_loop_im): Likewise.
22221 (tree_ssa_loop_unswitch): Likewise.
22222 (tree_vectorize): Likewise.
22223 (check_data_deps): Likewise.
22224 (tree_ssa_loop_ivcanon): Likewise.
22225 (tree_ssa_loop_bounds): Likewise.
22226 (tree_complete_unroll): Likewise.
22227 (tree_complete_unroll_inner): Likewise.
22228 (tree_parallelize_loops): Likewise.
22229 (tree_ssa_loop_prefetch): Likewise.
22230 (tree_ssa_loop_ivopts): Likewise.
22231 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
22232 * tree-vectorizer.c (vectorize_loops): Likewise.
22234 2013-04-29 Mike Frysinger <vapier@gentoo.org>
22236 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
22237 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
22238 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
22239 with EABI_LINK_SPEC.
22241 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22244 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
22247 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
22250 * lra-constraints.c (process_alt_operands): Discourage a bit more
22251 using memory for pseudos. Print cost dump for alternatives.
22252 Modify cost values for conflicts with early clobbers.
22253 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
22255 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22258 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
22260 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22262 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
22263 from/to S register.
22264 (movdi_aarch64): Support LDR/STR from/to D register.
22266 2013-04-29 Ian Bolton <ian.bolton@arm.com>
22268 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
22269 or higher by default.
22271 2013-04-29 Richard Biener <rguenther@suse.de>
22273 PR middle-end/57075
22274 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
22275 even if not adding abnormal edges for calls that can make
22278 2013-04-29 Richard Biener <rguenther@suse.de>
22280 PR middle-end/57103
22281 * tree-cfg.c (move_stmt_op): Fix condition under which to update
22283 (move_stmt_r): Remove redundant checking.
22285 2013-04-29 Teresa Johnson <tejohnson@google.com>
22288 * basic-block.h (apply_scale): New function.
22289 (apply_probability): Use apply_scale.
22290 * gimple-streamer-in.c (input_bb): Ditto.
22291 * lto-streamer-in.c (input_cfg): Ditto.
22292 * lto-cgraph.c (merge_profile_summaries): Ditto.
22293 * tree-optimize.c (execute_fixup_cfg): Ditto.
22294 * tree-inline.c (copy_bb): Update comment to use apply_scale.
22295 (copy_edges_for_bb): Ditto.
22296 (copy_cfg_body): Ditto.
22298 2013-04-29 Tom de Vries <tom@codesourcery.com>
22300 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
22301 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
22302 (tail_merge_optimize): Handle current_loops == NULL.
22304 2013-04-26 Jeff Law <law@redhat.com>
22306 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
22307 (simplify_cond_using_ranges): Generalize code to simplify
22308 COND_EXPRs where one argument is a constant and the other
22309 is an SSA_NAME created by an integral type conversion.
22311 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22313 * config/arm/arm.md (store_minmaxsi): Use only when
22314 optimize_insn_for_size_p.
22316 2013-04-29 Christian Bruel <christian.bruel@st.com>
22319 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
22321 2013-04-29 Richard Biener <rguenther@suse.de>
22323 PR middle-end/57089
22324 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
22325 loop tree make sure to schedule a fixup for the child as well.
22326 (expand_omp_for_generic): Properly add loops.
22327 (expand_omp_for_static_nochunk): Likewise.
22328 (expand_omp_for_static_chunk): Likewise.
22329 (expand_omp_for): For the degenerate case fixup loops.
22330 (expand_omp_sections): Fix default bb placement in loops.
22331 (expand_omp_atomic_pipeline): Properly add loops.
22333 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22335 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
22337 2013-04-29 Tom de Vries <tom@codesourcery.com>
22339 * tree-ssa-tail-merge.c: Update header comment.
22341 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22343 * config/aarch64/arm_neon.h
22344 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
22345 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
22346 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
22347 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
22348 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
22349 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
22350 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
22351 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
22353 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22355 * config/aarch64/aarch64-simd.md
22356 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
22357 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
22358 fix_trunc, fixuns_trunc.
22359 (ftrunc<VDQF:mode>2): New.
22360 * config/aarch64/iterators.md (optab): Add fix, fixuns.
22361 (fix_trunc_optab): New.
22363 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22365 * config/aarch64/aarch64-builtins.c
22366 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
22367 iceilf, lround, iroundf.
22369 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
22372 * config/i386/i386.h (enum ix86_tune_indices)
22373 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
22374 New, split from X86_TUNE_INTER_UNIT_MOVES.
22375 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
22376 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
22377 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
22378 (TARGET_INTER_UNIT_MOVES): Remove.
22379 * config/i386/i386.c (initial_ix86_tune_features): Update.
22380 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
22381 (ix86_expand_convert_uns_didf_sse): Use
22382 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22383 (ix86_expand_vector_init_one_nonzero): Ditto.
22384 (ix86_expand_vector_init_interleave): Ditto.
22385 (inline_secondary_memory_needed): Return true for moves from SSE class
22386 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
22387 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
22388 * config/i386/constraints.md (Yi, Ym): Depend on
22389 TARGET_INTER_UNIT_MOVES_TO_VEC.
22390 (Yj, Yn): New constraints.
22391 * config/i386/i386.md (*movdi_internal): Change constraints of
22392 operand 1 from Yi to Yj and from Ym to Yn.
22393 (*movsi_internal): Ditto.
22394 (*movdf_internal): Ditto.
22395 (*movsf_internal): Ditto.
22396 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
22397 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
22398 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
22399 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
22400 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
22401 * config/i386/sse.md (movdi_to_sse): Ditto.
22402 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
22403 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
22404 TARGET_INTER_UNIT_MOVES.
22405 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
22406 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
22407 instead of TARGET_INTER_UNIT_MOVES.
22408 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
22409 operand 1 from Yi to Yj and from Ym to Yn.
22411 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22413 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
22414 (float_truncate_hi_): Likewise.
22415 (float_extend_lo_): Likewise.
22416 (float_truncate_lo_): Likewise.
22417 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
22418 (aarch64_float_extend_lo_v2df): Likewise.
22419 (vec_unpacks_hi_v4sf): Likewise.
22420 (aarch64_float_truncate_lo_v2sf): Likewise.
22421 (aarch64_float_truncate_hi_v4sf): Likewise.
22422 (vec_pack_trunc_v2df): Likewise.
22423 (vec_pack_trunc_df): Likewise.
22425 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22427 * config/aarch64/aarch64-builtins.c
22428 (aarch64_fold_builtin): Fold float conversions.
22429 * config/aarch64/aarch64-simd-builtins.def
22430 (floatv2si, floatv4si, floatv2di): New.
22431 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
22432 * config/aarch64/aarch64-simd.md
22433 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
22434 * config/aarch64/iterators.md (FLOATUORS): New.
22435 (optab): Add float, floatuns.
22436 (su_optab): Likewise.
22438 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22440 * config/aarch64/aarch64-builtins.c
22441 (aarch64_builtin_vectorized_function): Use new names for
22443 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
22444 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
22445 (fcvtzu): Split as...
22446 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
22447 (fcvtas): Split as...
22448 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
22449 (fcvtau): Split as...
22450 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
22451 (fcvtps): Split as...
22452 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
22453 (fcvtpu): Split as...
22454 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
22455 (fcvtms): Split as...
22456 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
22457 (fcvtmu): Split as...
22458 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
22459 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
22460 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
22461 (lfrintnusf, lfrintnudf): Likewise.
22462 * config/aarch64/aarch64-simd.md
22463 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
22465 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
22466 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
22467 (fcvt_pattern): Likewise.
22469 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22471 * config/aarch64/aarch64-simd.md
22472 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
22473 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
22475 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22477 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
22478 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
22479 (vrnd<a,m,n,p>_f32): Implement using builtins.
22480 (vrnd<i,x><q>_f<32, 64>): New.
22482 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
22484 * config/aarch64/aarch64-builtins.c
22485 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
22486 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
22487 (frintz): Rename to...
22489 (frintp): Rename to...
22491 (frintm): Rename to...
22493 (frinti): Rename to...
22494 (nearbyint): ...this.
22495 (frintx): Rename to...
22497 (frinta): Rename to...
22499 * config/aarch64/aarch64-simd.md
22500 (aarch64_frint<frint_suffix><mode>): Delete.
22501 (<frint_pattern><mode>2): Convert to insn.
22502 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
22503 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
22504 (frint_pattern): Likewise.
22505 (frint_suffix): Likewise.
22507 2013-04-29 Richard Biener <rguenther@suse.de>
22509 PR tree-optimization/57081
22510 * loop-init.c: Include tree-flow.h.
22511 (loop_optimizer_finalize): Free number of iteration estimates.
22512 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
22514 2013-04-29 Jakub Jelinek <jakub@redhat.com>
22516 PR tree-optimization/57083
22517 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
22518 non-singleton shift count range, zero extend low_bound for uns case.
22520 * config/i386/predicates.md (general_vector_operand): New predicate.
22521 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
22522 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
22523 if they aren't nonimmediate operands. If their original values
22524 satisfy const_vector_equal_evenodd_p, don't shift them.
22525 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
22526 predicates. For the SSE4.1 case force operands[{1,2}] into registers
22527 if not nonimmediate_operand.
22528 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
22529 instead of register_operand.
22530 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
22532 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22534 * stor-layout.c (finalize_size_functions): Allocate a structure and
22535 reset cfun before dumping the functions.
22537 2013-04-27 Jakub Jelinek <jakub@redhat.com>
22539 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
22542 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
22543 use xop_pmacsdqh if uns_p.
22544 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
22545 the immediate rotate count.
22547 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
22549 * rtl.h (struct rtx_def): Add comment for field jump.
22550 (LRA_SUBREG_P): New macro.
22551 * recog.c (register_operand): Check LRA_SUBREG_P.
22552 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22554 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22555 (lra_final_code_change): Skip subreg change for operators.
22556 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22557 if there are no operand changes.
22558 * lra-constraints.c (curr_insn_set): New.
22559 (match_reload): Set LRA_SUBREG_P.
22560 (emit_spill_move): Ditto.
22561 (check_and_process_move): Use curr_insn_set. Process only single
22562 set insns. Don't initialize sec_mem_p and change_p.
22563 (simplify_operand_subreg): Use LRA_SUBREG_P.
22564 (reg_in_class_p): New function.
22565 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22566 of #ifdef. Add code to remove cycling.
22567 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22568 non-null disp. Reload inner instead of disp when base and index
22569 are null. Try to put lo_sum into register.
22570 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22571 (check_and_process_move): Move code for move cost check to
22572 simple_move_p. Remove equiv_substitution.
22573 (simple_move_p): New function.
22574 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22575 curr_insn_set. Call check_and_process_move only for single set
22576 insns. Use the new function. Move call of check_and_process_move
22577 after operand equiv substitution and address process.
22579 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22582 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
22583 with nonlocal goto receivers or returns twice calls, ignore
22584 unininitialized values from abnormal edges to nl goto receiver
22585 or returns twice call.
22587 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22589 PR tree-optimization/57051
22590 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
22591 and VEC_RSHIFT_EXPR if shift count is a multiple of element
22594 2013-04-26 Richard Biener <rguenther@suse.de>
22596 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
22597 (expand_omp_taskreg): Likewise. Mark loops for fixup.
22598 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
22599 (fixup_loop_arrays_after_move): New function.
22600 (move_sese_region_to_fn): Properly outline the loop tree parts
22601 of the SESE region.
22603 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
22605 * config/i386/i386.md (type, unit): Fix long lines.
22607 2013-04-26 Richard Biener <rguenther@suse.de>
22609 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
22610 (lto-streamer-out.o): Likewise.
22611 * cfgloop.c (init_loops_structure): Export, add struct function
22612 argument and adjust.
22613 (flow_loops_find): Adjust.
22614 * cfgloop.h (enum loop_estimation): Add EST_LAST.
22615 (init_loops_structure): Declare.
22616 * lto-streamer-in.c: Include cfgloop.h.
22617 (input_cfg): Input the loop tree.
22618 * lto-streamer-out.c: Include cfgloop.h.
22619 (output_cfg): Output the loop tree.
22620 (output_struct_function_base): Do not drop PROP_loops.
22622 2013-03-26 Richard Biener <rguenther@suse.de>
22624 * tree-cfg.c (execute_build_cfg): Build the loop tree.
22625 (pass_build_cfg): Provide PROP_loops.
22626 (move_sese_region_to_fn): Remove loops that are outlined into fn
22628 * tree-inline.c: Include cfgloop.h.
22629 (initialize_cfun): Do not drop PROP_loops.
22630 (copy_loops): New function.
22631 (copy_cfg_body): Copy loop structure.
22632 (tree_function_versioning): Initialize destination loop tree.
22633 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
22634 (pass_parallelize_loops): Do IL verification.
22635 * loop-init.c (loop_optimizer_init): Fixup loops if required.
22636 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
22637 the CFG make sure we fixup loops as well.
22638 * tree-ssa-tail-merge.c: Include cfgloop.h.
22639 (replace_block_by): When merging loop latches mark loops for fixup.
22640 * lto-streamer-out.c (output_struct_function_base): Drop
22641 PROP_loops for now.
22642 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
22643 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
22644 * ipa-split.c: Include cfgloop.h.
22645 (split_function): Add the new return block to the loop tree root.
22646 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
22647 whether we have removed the forwarder block.
22648 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
22649 * cfgloop.h (place_new_loop): Declare.
22650 * cfgloopmanip.c (place_new_loop): Export.
22651 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
22652 (tree-switch-conversion.o): Likewise.
22653 (tree-complex.o): Likewise.
22654 (tree-inline.o): Likewise.
22655 (tree-ssa-tailmerge.o): Likewise.
22656 (ipa-split.o): Likewise.
22657 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
22658 (tree-ssa-copy.o): Likewise.
22659 * tree-switch-conversion.c: Include cfgloop.h
22660 (process_switch): If we emit a bit-test cascade, schedule loops
22662 * tree-complex.c: Include cfgloop.h.
22663 (expand_complex_div_wide): Properly add new basic-blocks to loops.
22664 * asan.c: Include cfgloop.h.
22665 (create_cond_insert_point): Properly add new basic-blocks to
22666 loops, schedule loop fixup.
22667 * cfgloop.c (verify_loop_structure): Check that looks are not
22669 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
22671 (expand_omp_for_generic): Likewise.
22672 (expand_omp_sections): Likewise.
22673 (expand_omp_atomic_pipeline): Schedule loops for fixup.
22674 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
22675 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
22676 is initialized, not when loops are present.
22677 * tree-parloops.c (parallelize_loops): Remove checking here.
22678 * passes.c (init_optimization_passes): Schedule a copy-propagation
22679 pass before complete unrolling of inner loops.
22681 2013-04-26 Jakub Jelinek <jakub@redhat.com>
22683 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
22684 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
22685 (colorize_init): Add argument to _WIN32 version.
22686 * toplev.c: Include diagnostic-color.h.
22687 (process_options): Default to -fdiagnostics-color=auto if
22688 GCC_COLORS env var is in the environment.
22689 * common.opt (fdiagnostics-color=): Add Var and Init.
22690 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
22691 env var is in the environment, the default is auto rather than never.
22693 * diagnostic.h (file_name_as_prefix): Add context argument.
22694 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
22695 the string as locus.
22696 * langhooks.c (lhd_print_error_function): Adjust caller.
22698 2013-04-25 Lawrence Crowl <crowl@google.com>
22700 * var-tracking.c (shared_hash_def::htab):
22701 Change type to hash_table. Update dependent calls and types.
22703 2013-04-25 Lawrence Crowl <crowl@google.com>
22705 * Makefile.in: Update as needed below.
22707 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
22708 Move declaration to after the type's method definitons.
22710 * attribs.c (htab_t scoped_attributes::attribute_hash):
22711 Change type to hash_table. Update dependent calls and types.
22713 * bitmap.c (htab_t bitmap_desc_hash):
22714 Change type to hash_table. Update dependent calls and types.
22716 * cselib.c (htab_t cselib_hash_table):
22717 Change type to hash_table. Update dependent calls and types.
22719 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
22720 (hash_string_slot_node): Move implementation into lto-streamer.h
22721 struct string_slot_hasher.
22722 (eq_string_slot_node): Likewise.
22724 * data-streamer-out.c: Update output_block::string_hash_table
22725 dependent calls and types.
22727 * dwarf2cfi.c (htab_t trace_index):
22728 Change type to hash_table. Update dependent calls and types.
22730 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
22731 Change type to hash_table. Update dependent calls and types.
22732 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
22733 (htab_t optimize_external_refs::map): Likewise.
22734 (htab_t output_comp_unit::extern_map): Likewise.
22735 (htab_t output_comdat_type_unit::extern_map): Likewise.
22736 (htab_t output_macinfo::macinfo_htab): Likewise.
22737 (htab_t optimize_location_lists::htab): Likewise.
22738 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
22740 * except.c (htab_t ehspec_hash_type):
22741 Change type to hash_table. Update dependent calls and types.
22742 (assign_filter_values::ttypes): Likewise.
22743 (assign_filter_values::ehspec): Likewise.
22744 (sjlj_assign_call_site_values::ar_hash): Likewise.
22745 (convert_to_eh_region_ranges::ar_hash): Likewise.
22747 * gcse.c (htab_t pre_ldst_table):
22748 Change type to hash_table. Update dependent calls and types.
22750 * ggc-common.c (htab_t saving_htab):
22751 Change type to hash_table. Update dependent calls and types.
22752 (htab_t loc_hash): Likewise.
22753 (htab_t ptr_hash): Likewise.
22754 (call_count): Rename ggc_call_count.
22755 (call_alloc): Rename ggc_call_alloc.
22756 (loc_descriptor): Rename make_loc_descriptor.
22757 (add_statistics): Rename ggc_add_statistics.
22759 * ggc-common.c (saving_htab):
22760 Change type to hash_table. Update dependent calls and types.
22762 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
22763 (push_gimplify_context): Likewise.
22764 (pop_gimplify_context): Likewise.
22765 (struct gimple_temp_hash_elt): Added.
22766 (struct gimplify_hasher): Likewise.
22767 (struct gimplify_ctx.temp_htab):
22768 Change type to hash_table. Update dependent calls and types.
22770 * gimple-fold.c: Include gimplify-ctx.h.
22772 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
22773 Change type to hash_table. Update dependent calls and types.
22774 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
22775 avoid potential global name collision.
22777 * gimplify.c: Include gimplify-ctx.h.
22778 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
22779 (htab_t gimplify_ctx::temp_htab):
22780 Update dependent calls and types for new type hash_table.
22781 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
22782 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
22784 * gimplify-ctx.h: New.
22785 (struct gimple_temp_hash_elt): Move from gimplify.c.
22786 (class gimplify_hasher): New.
22787 (struct gimplify_ctx): Move from gimple.h.
22788 (htab_t gimplify_ctx::temp_htab):
22789 Change type to hash_table. Update dependent calls and types.
22791 * graphite-clast-to-gimple.c: Include graphite-htab.h.
22792 (htab_t ivs_params::newivs_index):
22793 Change type to hash_table. Update dependent calls and types.
22794 (htab_t ivs_params::params_index): Likewise.
22795 (htab_t print_generated_program::params_index): Likewise.
22796 (htab_t gloog::newivs_index): Likewise.
22797 (htab_t gloog::params_index): Likewise.
22799 * graphite.c: Include graphite-htab.h.
22800 4htab_t graphite_transform_loops::bb_pbb_mapping):
22801 Change type to hash_table. Update dependent calls and types.
22803 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
22804 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
22805 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
22807 * graphite-dependences.c: Include graphite-htab.h.
22808 (loop_is_parallel_p): Change hash table type of parameter.
22810 * graphite-htab.h: New.
22811 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
22812 (extern find_pbb_via_hash): Move from graphite-poly.h.
22813 (extern loop_is_parallel_p): Move from graphite-poly.h.
22814 (extern get_loop_body_pbbs): Move from graphite-poly.h.
22816 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
22817 (extern loop_is_parallel_p): Move to graphite-htab.h.
22818 (extern get_loop_body_pbbs): Move to graphite-htab.h.
22820 * haifa-sched.c (htab_t delay_htab):
22821 Change type to hash_table. Update dependent calls and types.
22822 (htab_t delay_htab_i2): Likewise.
22824 * ira-color.c (htab_t allocno_hard_regs_htab):
22825 Change type to hash_table. Update dependent calls and types.
22827 * ira-costs.c (htab_t cost_classes_htab):
22828 Change type to hash_table. Update dependent calls and types.
22830 * loop-invariant.c (htab_t merge_identical_invariants::eq):
22831 Change type to hash_table. Update dependent calls and types.
22833 * loop-iv.c (htab_t bivs):
22834 Change type to hash_table. Update dependent calls and types.
22836 * loop-unroll.c (htab_t opt_info::insns_to_split):
22837 Change type to hash_table. Update dependent calls and types.
22838 (htab_t opt_info::insns_with_var_to_expand): Likewise.
22840 * lto-streamer.h (struct string_slot): Move from data-streamer.h
22841 (struct string_slot_hasher): New.
22842 (htab_t output_block::string_hash_table):
22843 Change type to hash_table. Update dependent calls and types.
22845 * lto-streamer-in.c (freeing_string_slot_hasher): New.
22846 (htab_t file_name_hash_table):
22847 Change type to hash_table. Update dependent calls and types.
22849 * lto-streamer-out.c: Update output_block::string_hash_table dependent
22852 * lto-streamer.c (htab_t tree_htab):
22853 Change type to hash_table. Update dependent calls and types.
22855 * omp-low.c: Include gimplify-ctx.h.
22857 * passes.c (htab_t name_to_pass_map):
22858 Change type to hash_table. Update dependent calls and types.
22859 (pass_traverse): Rename to passes_pass_traverse.
22861 * plugin.c (htab_t event_tab):
22862 Change type to hash_table. Update dependent calls and types.
22864 * postreload-gcse.c (htab_t expr_table):
22865 Change type to hash_table. Update dependent calls and types.
22866 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
22868 * sese.c (debug_rename_map_1): Make extern.
22869 (htab_t copy_bb_and_scalar_dependences::rename_map):
22870 Change type to hash_table. Update dependent calls and types.
22872 * sese.h (extern debug_rename_map): Move to .c file.
22874 * store-motion.c (htab_t store_motion_mems_table):
22875 Change type to hash_table. Update dependent calls and types.
22877 * trans-mem.c (htab_t tm_new_mem_hash):
22878 Change type to hash_table. Update dependent calls and types.
22880 * tree-browser.c (htab_t TB_up_ht):
22881 Change type to hash_table. Update dependent calls and types.
22883 * tree-cfg.c (htab_t discriminator_per_locus):
22884 Change type to hash_table. Update dependent calls and types.
22886 * tree-complex.c: Include tree-hasher.h
22887 (htab_t complex_variable_components):
22888 Change type to hash_table. Update dependent calls and types.
22890 * tree-eh.c (htab_t finally_tree):
22891 Change type to hash_table. Update dependent calls and types.
22893 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
22894 struct int_tree_hasher.
22895 (extern int_tree_map_eq): Likewise.
22896 (uid_decl_map_hash): Removed.
22897 (extern decl_tree_map_eq): Likewise.
22899 * tree-hasher.h: New.
22900 (struct int_tree_hasher): New.
22901 (typedef int_tree_htab_type): New.
22903 * tree-inline.c: Include gimplify-ctx.h.
22905 * tree-mudflap.c: Include gimplify-ctx.h.
22907 * tree-parloops.c: Include tree-hasher.h.
22908 (htab_t eliminate_local_variables_stmt::decl_address):
22909 Change type to hash_table. Update dependent calls and types.
22910 (htab_t separate_decls_in_region::decl_copies): Likewise.
22912 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
22913 Change type to hash_table. Update dependent calls and types.
22915 * tree-sra.c (candidates):
22916 Change type to hash_table. Update dependent calls and types.
22918 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
22920 (int_tree_map_hash): Likewise.
22922 * tree-ssa-dom.c (htab_t avail_exprs):
22923 Change type to hash_table. Update dependent calls and types.
22925 * tree-ssa-live.c (var_map_base_init::tree_to_index):
22926 Change type to hash_table. Update dependent calls and types.
22928 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
22929 Change type to hash_table. Update dependent calls and types.
22931 * tree-ssa-phiopt.c (seen_ssa_names):
22932 Change type to hash_table. Update dependent calls and types.
22934 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
22935 Change type to hash_table. Update dependent calls and types.
22937 * tree-ssa-uncprop.c (equiv):
22938 Change type to hash_table. Update dependent calls and types.
22940 2013-04-25 Jakub Jelinek <jakub@redhat.com>
22942 PR rtl-optimization/57003
22943 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
22944 call note_stores with kill_clobbered_value callback again after
22945 killing regs_invalidated_by_call.
22947 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22949 * config/aarch64/aarch64-simd.md
22950 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
22951 (aarch64_simd_bsl<mode>): Likewise.
22952 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
22954 2013-04-25 Marek Polacek <polacek@redhat.com>
22956 PR tree-optimization/57066
22957 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
22959 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22961 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
22963 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22965 * config/aarch64/aarch64-builtins.c
22966 (aarch64_fold_builtin): New.
22967 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
22968 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
22969 * config/aarch64/aarch64-simd-builtins.def (abs): New.
22970 * config/aarch64/arm_neon.h
22971 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
22973 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
22974 Tejas Belagod <tejas.belagod@arm.com>
22976 * config/aarch64/aarch64-builtins.c
22977 (aarch64_gimple_fold_builtin): New.
22978 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
22979 * config/aarch64/aarch64-simd-builtins.def (addv): New.
22980 * config/aarch64/aarch64-simd.md (addpv4sf): New.
22981 (addvv4sf): Update.
22982 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
22984 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22986 * config/aarch64/aarch64.md
22987 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
22989 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22991 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
22992 (*ngcsi_uxtw): New pattern.
22994 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22995 Julian Brown <julian@codesourcery.com>
22997 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
22998 (TB_DREG): Add T_V4HF.
22999 (v4hf_UP): New macro.
23000 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23001 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23002 Handle initialisation of V4HF. Adjust initialisation of reinterpret
23004 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
23005 (arm_vector_mode_supported_p): Handle V4HF.
23006 (arm_mangle_map): Handle V4HFmode.
23007 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
23008 * config/arm/arm_neon_builtins.def: Add entries for
23009 vcvtv4hfv4sf, vcvtv4sfv4hf.
23010 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
23011 (neon_vcvtv4hfv4sf): Likewise.
23012 * config/arm/neon-gen.ml: Handle half-precision floating point
23014 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
23015 * config/arm/arm_neon.h: Regenerate.
23016 * config/arm/neon.ml (type elts): Add F16.
23017 (type vectype): Add T_float16x4, T_floatHF.
23018 (type vecmode): Add V4HF.
23019 (type features): Add Requires_FP_bit feature.
23020 (elt_width): Handle F16.
23021 (elt_class): Likewise.
23022 (elt_of_class_width): Likewise.
23023 (mode_of_elt): Refactor.
23024 (type_for_elt): Handle F16, fix error messages.
23025 (vectype_size): Handle T_float16x4.
23026 (vcvt_sh): New function.
23027 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
23028 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
23029 (string_of_mode): Handle V4HF.
23030 * doc/arm-neon-intrinsics.texi: Regenerate.
23032 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
23034 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
23035 format specifier in 'X' case.
23037 2013-04-25 Alan Modra <amodra@gmail.com>
23040 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
23041 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
23042 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
23043 Repeat for many other rotate/shift and mask patterns using subregs.
23044 Name lshiftrt insns.
23045 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
23046 on WORDS_BIG_ENDIAN.
23048 2013-04-25 Alan Modra <amodra@gmail.com>
23050 * config.gcc: Support little-endian powerpc-linux targets.
23051 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
23052 (LINK_OS_LINUX_SPEC): Define.
23053 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
23054 Preserve MASK_LITTLE_ENDIAN.
23055 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
23056 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
23057 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
23058 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
23059 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
23060 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
23061 Correct fp word order for little-endian. Don't shift toc entries
23062 smaller than a word for little-endian.
23063 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
23064 (bswapdi2 splits): Correct low-part subreg for little-endian.
23065 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
23066 low/high where such is correct only for be.
23067 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
23068 little-endian for -mcall-aixdesc.
23070 2013-04-25 Alan Modra <amodra@gmail.com>
23072 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
23073 replace_equiv_address_nv.
23075 2013-04-25 Alan Modra <amodra@gmail.com>
23077 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
23079 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23082 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23083 * rtl.h (struct rtx_def): ...
23085 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23087 PR rtl-optimizations/57046
23088 * lra-constraints (split_reg): Set up lra_risky_transformations_p
23089 for multi-reg splits.
23091 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
23093 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
23095 2013-04-24 Sterling Augustine <saugustine@google.com>
23097 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
23098 (comp_dir_string, debug_str_dwo_section): New.
23099 (DEBUG_STR_DWO_SECTION): Rename to ...
23100 (DEBUG_DWO_STR_SECTION): ... this.
23101 (DEBUG_NORM_STR_SECTION): Delete.
23102 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
23103 (DEBUG_STR_DWO_SECTION_FLAGS): New.
23104 (find_AT_string): Move most logic to ...
23105 (find_AT_string_in_table): ... here. New.
23106 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
23107 add_skeleton_AT_string. Delete logic.
23108 (output_skeleton_debug_sections): Remove call to
23109 add_top_level_skeleton_die_attrs.
23110 (add_comp_dir_attribute): Move logic to comp_dir_string.
23111 (dwarf2out_init): Initialize debug_str_dwo_section.
23112 (output_indirect_string): Call find_string_form.
23113 (output_indirect_strings): Rewrite.
23114 (prune_unused_types): Empty skeleton_debug_str_hash.
23115 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
23116 (dwarf2out_finish): Call output_indirect_strings.
23118 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23120 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
23122 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
23124 * rtl.h (struct rtx_def): Add comment for field jump.
23125 (LRA_SUBREG_P): New macro.
23126 * recog.c (register_operand): Check LRA_SUBREG_P.
23127 * lra.c (lra): Add note at the end of RTL code. Align non-empty
23129 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
23130 (lra_final_code_change): Skip subreg change for operators.
23131 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
23132 if there are no operand changes.
23133 * lra-constraints.c (curr_insn_set): New.
23134 (match_reload): Set LRA_SUBREG_P.
23135 (emit_spill_move): Ditto.
23136 (check_and_process_move): Use curr_insn_set. Process only single
23137 set insns. Don't initialize sec_mem_p and change_p.
23138 (simplify_operand_subreg): Use LRA_SUBREG_P.
23139 (reg_in_class_p): New function.
23140 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
23141 of #ifdef. Add code to remove cycling.
23142 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
23143 non-null disp. Reload inner instead of disp when base and index
23144 are null. Try to put lo_sum into register.
23145 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
23146 (check_and_process_move): Move code for move cost check to
23147 simple_move_p. Remove equiv_substitution.
23148 (simple_move_p): New function.
23149 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
23150 curr_insn_set. Call check_and_process_move only for single set
23151 insns. Use the new function. Move call of check_and_process_move
23152 after operand equiv substitution and address process.
23154 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
23156 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
23157 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
23158 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
23160 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
23162 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
23164 2013-04-24 Marek Polacek <polacek@redhat.com>
23166 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
23167 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
23168 (select_loops_exit_conditions): Likewise.
23169 (number_of_iterations_for_all_loops): Likewise.
23170 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
23171 (scev_analysis): Likewise.
23173 2013-04-02 Catherine Moore <clm@codesourcery.com>
23174 Chao-ying Fu <fu@mips.com>
23176 * config/mips/micromips.md (jraddiusp): New pattern.
23177 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
23178 instruction if possible.
23180 2013-04-24 Alan Modra <amodra@gmail.com>
23182 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
23184 2013-04-24 Julian Brown <julian@codesourcery.com>
23185 Chung-Lin Tang <cltang@codesourcery.com>
23187 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
23188 dependency behavior in enumeration type DIE generation. Add TODO note
23189 to comments about future DW_FORM_sdata/udata re-work of related code.
23191 2013-04-23 Lawrence Crowl <crowl@google.com>
23193 * Makefile.in: Update as needed below.
23195 * hash-table.h (class hash_table):
23196 Correct many methods with parameter types compare_type to the correct
23197 value_type. (Correct code was unlikely to notice the change.)
23198 (hash_table::elements_with_deleted) New.
23199 (class hashtable::iterator): New.
23200 (hashtable::begin()): New.
23201 (hashtable::end()): New.
23202 (FOR_EACH_HASH_TABLE_ELEMENT): New.
23204 * statistics.c (statistics_hashes):
23205 Change type to hash_table. Update dependent calls and types.
23207 * tree-into-ssa.c (var_infos):
23208 Change type to hash_table. Update dependent calls and types.
23210 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
23211 Change type to hash_table. Update dependent calls and types.
23213 * tree-ssa-loop-im.c (struct mem_ref.refs):
23214 Change type to hash_table. Update dependent calls and types.
23216 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
23217 Change type to hash_table. Update dependent calls and types.
23219 * tree-ssa-sccvn.c (vn_tables_s::nary):
23220 Change type to hash_table. Update dependent calls and types.
23221 (vn_tables_s::phis): Likewise.
23222 (vn_tables_s::references): Likewise.
23224 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
23225 (vn_reference_eq): Update parameter and return types.
23227 * tree-ssa-structalias.c (pointer_equiv_class_table):
23228 Change type to hash_table. Update dependent calls and types.
23229 (location_equiv_class_table): Likewise.
23231 * tree-vect-data-refs.c: Consequential changes for making
23232 peeling a hash_table.
23234 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
23235 (destroy_loop_vec_info): Dependent hash_table update.
23237 * tree-vectorizer.h (peeling_htab):
23238 Change type to hash_table. Update dependent calls and types.
23240 2013-04-23 Shiva Chen <shiva0217@gmail.com>
23242 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
23243 to check the register content is equal or not.
23244 * lra-constraints.c (match_reload): Use lra_assign_reg_val
23245 to assign register content record.
23246 * lra-eliminations.c (update_reg_eliminate): Use
23247 lra_update_reg_val_offset to update register content offset.
23248 * lra-int.h (struct lra_reg): Add offset member.
23249 (lra_reg_val_equal_p): New static inline function.
23250 (lra_update_reg_val_offset): New static inline function.
23251 (lra_assign_reg_val): New static inline function.
23252 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
23253 to assign register content record.
23254 (initialize_lra_reg_info_element): Initial offset to zero.
23256 2013-04-23 Catherine Moore <clm@codesourcery.com>
23258 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
23259 operands. Record compression.
23261 2013-04-23 Xinliang David Li <davidxl@google.com>
23263 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
23265 2013-04-23 Richard Biener <rguenther@suse.de>
23267 PR middle-end/57036
23268 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
23269 parameter, only add abnormal goto edges from the copied body
23270 if the call could perform abnormal gotos.
23271 (copy_cfg_body): Adjust.
23273 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
23275 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
23277 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
23279 * coretypes.h (gimple_stmt_iterator): Add struct to make
23282 2013-04-23 Richard Biener <rguenther@suse.de>
23284 PR tree-optimization/57026
23285 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
23286 from SSA names occuring in abnormal PHI nodes.
23288 2013-04-22 Andi Kleen <ak@linux.intel.com>
23290 * lto/lto.c (print_lto_report_1): Fix LTO report names.
23292 2013-04-22 Andi Kleen <ak@linux.intel.com>
23294 * lto/lto.c (print_lto_report_1): Declare early.
23295 (read_cgraph_and_symbols): Call print_lto_report_1 early.
23297 2013-04-22 Andi Kleen <ak@linux.intel.com>
23299 * common.opt (-flto-report-wpa): Add.
23300 * doc/invoke.texi (-flto-report-wpa): Add.
23301 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
23304 2013-04-22 Xinliang David Li <davidxl@google.com>
23306 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
23307 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
23308 * Makefile.in: New dependency
23310 David Daney <ddaney.cavm@gmail.com>
23312 * configure.ac (gcc_cv_as_micromips_support): Use the
23313 --fatal-warnings option.
23314 * configure: Regenerate.
23316 2013-04-22 Marek Polacek <polacek@redhat.com>
23319 * tsan.c (instrument_expr): Don't instrument expression
23320 in case its size is zero.
23322 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
23326 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
23328 * config/alpha/alpha.c (TARGET_LRA_P): New define.
23330 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23332 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
23333 (gimple_stmt_iterator): New typedef.
23334 * gimple.h (gimple_stmt_iterator): Rename to...
23335 (gimple_stmt_iterator_d): ... This.
23336 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
23337 trees be valid for GIMPLE and GENERIC.
23338 (TARGET_GIMPLE_FOLD_BUILTIN): New.
23339 * gimple-fold.c (gimple_fold_call): Call target hook
23340 gimple_fold_builtin.
23341 * hooks.c (hook_bool_gsiptr_false): New.
23342 * hooks.h (hook_bool_gsiptr_false): New.
23343 * target.def (fold_stmt): New.
23344 * doc/tm.texi: Regenerate.
23346 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
23349 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
23350 a set sp if no stack realignment.
23352 2013-04-22 Nick Clifton <nickc@redhat.com>
23354 * config.gcc (tilegx-linux): Extend extra_objs rather than
23356 (tilepro-linux): Likewise.
23358 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23360 * config/aarch64/aarch64-builtins.c
23362 (CF0, CF1, CF2, CF3, CF4, CF10): New.
23363 (VAR<1-12>): Add MAP parameter.
23364 (BUILTIN_*): Likewise.
23365 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
23366 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
23367 (aarch64_ushl_n<mode>): Likewise.
23368 (aarch64_sshr_n<mode>): Likewise.
23369 (aarch64_ushr_n<mode>): Likewise.
23370 (aarch64_<maxmin><mode>): Likewise.
23371 (aarch64_sqrt<mode>): Likewise.
23372 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
23373 (vshr<q>_n_*): Likewise.
23375 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
23377 * config/aarch64/aarch64-builtins.c
23378 (aarch64_simd_builtin_type_mode): Handle SF types.
23380 (BUILTIN_GPF): Define.
23381 (aarch64_init_simd_builtins): Handle SF types.
23382 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
23383 (frecps): Likewise.
23384 (frecpx): Likewise.
23385 * config/aarch64/aarch64-simd.md
23386 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
23387 (aarch64_frecpe<mode>): New.
23388 (aarch64_frecps<mode>): Likewise.
23389 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
23390 (v8type): Add frecp<esx>.
23391 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
23392 (aarch64_frecps<mode>): Likewise.
23393 * config/aarch64/iterators.md (FRECP): New.
23394 (frecp_suffix): Likewise.
23395 * config/aarch64/arm_neon.h
23396 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
23398 2013-04-22 Christian Bruel <christian.bruel@st.com>
23401 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
23402 (REG_CLASS_NAMES): Idem.
23403 (REG_CLASS_CONTENTS): Idem.
23404 (REGCLASS_HAS_FP_REG): Idem.
23405 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
23406 (sh_conditional_register_usage): Idem.
23408 2013-04-21 Jeff Law <law@redhat.com>
23410 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
23411 (ssa_forward_propagate_and_combine): Use it.
23413 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23415 * lra.c: Update the flow chart diagram.
23417 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
23419 PR rtl-optimization/56847
23420 * lra-constraints.c (process_alt_operands): Discourage alternative
23421 with non-matche doffsettable memory constraint fro memory with
23424 2013-04-19 Richard Biener <rguenther@suse.de>
23426 PR tree-optimization/56982
23427 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
23429 * gimplify.c (gimplify_call_expr): Notice special calls.
23430 (gimplify_modify_expr): Likewise.
23431 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
23432 abnormal control flow receivers.
23433 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
23434 in the same way as cfun->has_nonlocal_labels.
23435 (gimple_purge_dead_abnormal_call_edges): Likewise.
23436 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
23437 receivers start a basic-block.
23439 2013-04-19 Richard Biener <rguenther@suse.de>
23441 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
23443 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
23444 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
23445 (SLP_TREE_LOAD_PERMUTATION): Add.
23446 (vect_transform_slp_perm_load): Adjust prototype.
23447 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
23448 (vect_free_slp_instance): Likewise.
23449 (vect_create_new_slp_node): Likewise.
23450 (vect_supported_slp_permutation_p): Remove.
23451 (vect_slp_rearrange_stmts): Adjust.
23452 (vect_supported_load_permutation_p): Likewise. Inline
23453 vect_supported_slp_permutation_p here.
23454 (vect_analyze_slp_instance): Compute load permutations per
23455 slp node instead of per instance.
23456 (vect_get_slp_defs): Adjust.
23457 (vect_transform_slp_perm_load): Likewise.
23458 (vect_schedule_slp_instance): Remove redundant code.
23459 (vect_schedule_slp): Remove hack for PR56270, add it ...
23460 * tree-vect-stmts.c (vectorizable_load): ... here, do not
23461 CSE loads for SLP. Adjust.
23463 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23465 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
23466 spelling in two comments.
23468 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
23471 * config/arm/arm.c (load_multiple_sequence): Require SP
23472 as base register for loads if SP is in the register list.
23474 2013-04-19 Martin Jambor <mjambor@suse.cz>
23476 PR tree-optimization/56718
23477 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
23478 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
23479 and made public. Adjusted all callers.
23480 (ipa_intraprocedural_devirtualization): New function.
23481 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
23482 (ipa_intraprocedural_devirtualization): Likewise.
23483 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
23485 2013-04-19 Richard Biener <rguenther@suse.de>
23487 PR tree-optimization/57000
23488 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
23490 2013-04-19 Terry Guo <terry.guo@arm.com>
23492 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
23494 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
23495 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
23496 (cortex_m4_fmacs): Use new reservations.
23497 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
23499 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
23501 PR rtl-optimization/56999
23502 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
23504 (lra_coalesce): Remove split_origin_bitmap and related code.
23505 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
23506 ranges if necessary.
23508 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
23510 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
23512 (ix86_expand_call): Remove clobbered_registers array and use
23513 x86_64_ms_sysv_extra_clobbered_registers instead.
23514 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
23516 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
23518 * config/i386/i386.md (*call_rex64_ms_sysv): Use
23519 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
23520 (*call_value_rex64_ms_sysv): Ditto.
23522 2013-04-18 Cary Coutant <ccoutant@google.com>
23524 * dwarf2out.c (output_pubnames): Check die_perennial_p of
23525 parent instead of die_mark.
23527 2013-04-18 Diego Novillo <dnovillo@google.com>
23529 * gimple.c (create_gimple_tmp): New.
23530 (get_expr_type): New.
23531 (build_assign): New.
23532 (build_type_cast): New.
23533 * gimple.h (enum ssa_mode): Define.
23534 (gimple_seq_set_location): New.
23535 * asan.c (build_check_stmt): Change some gimple_build_* calls
23536 to use build_assign and build_type_cast.
23538 2013-04-18 Richard Biener <rguenther@suse.de>
23540 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
23541 handle negative step. Remove redundant checks.
23542 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
23543 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
23544 for negative step and grouped loads fail to vectorize.
23546 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23548 * emit-rtl.c (reset_insn_used_flags): New function.
23549 (reset_all_used_flags): Use it.
23550 (verify_insn_sharing): New function.
23551 (verify_rtl_sharing): Fix verification for SEQUENCEs.
23553 2013-04-18 Jakub Jelinek <jakub@redhat.com>
23555 PR tree-optimization/56984
23556 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
23557 and (x >> M) >= N don't register any assertion if N << M is the
23560 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
23562 * lower-subreg.c (resolve_simple_move): If called self-recursive,
23563 do not delete_insn insns that have not yet been emitted, only
23564 unlink them with remove_insn.
23565 * df-scan.c (df_insn_delete): Revert r197492.
23567 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
23569 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
23570 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
23572 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23574 * config/arm/arm.md (movsicc_insn): Convert define_insn into
23575 define_insn_and_split.
23576 (and_scc,ior_scc,negscc): Likewise.
23577 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
23579 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
23581 * config/arm/arm.c (use_return_insn): Return 0 for targets that
23582 can benefit from using a sequence of LDRD instructions in epilogue
23583 instead of a single LDM instruction.
23585 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
23588 * doc/extend.texi: Fix typo.
23590 2013-04-17 Richard Biener <rguenther@suse.de>
23592 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
23593 (vect_build_slp_tree): ... here.
23594 (vect_build_slp_tree_1): Compute which stmts of the SLP group
23595 match. Remove special-casing of mismatched complex loads.
23596 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
23597 re-try the match with swapped commutative operands.
23598 (vect_supported_load_permutation_p): Remove special-casing of
23599 mismatched complex loads.
23600 (vect_analyze_slp_instance): Adjust.
23602 2013-04-17 Richard Biener <rguenther@suse.de>
23604 PR rtl-optimization/56921
23605 * cfgloop.h (struct loop): Add simple_loop_desc member.
23606 (struct niter_desc): Mark with GTY(()).
23607 (simple_loop_desc): Do not use aux field but simple_loop_desc.
23608 * loop-iv.c (get_simple_loop_desc): Likewise.
23609 (free_simple_loop_desc): Likewise.
23612 2013-04-16 Richard Biener <rguenther@suse.de>
23614 PR rtl-optimization/56921
23615 * loop-init.c (pass_rtl_move_loop_invariants): Add
23616 TODO_do_not_ggc_collect to todo_flags_finish.
23617 (pass_rtl_unswitch): Same.
23618 (pass_rtl_unroll_and_peel_loops): Same.
23619 (pass_rtl_doloop): Same.
23621 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
23623 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
23624 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
23625 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
23627 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
23628 * tree-streamer.c (record_common_node): Adjust reference in comment.
23630 2013-04-17 Terry Guo <terry.guo@arm.com>
23632 * config/arm/cortex-m4.md: Add a new bypass.
23634 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23636 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
23638 (*subs_<optab><mode>_multp2): New pattern.
23639 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23640 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
23642 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23644 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
23645 (*subs_mul_imm_<mode>): New pattern.
23647 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
23650 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
23651 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
23652 (vsx_movti_32bit): Change j->wa to O->wa.
23654 2013-04-16 Richard Biener <rguenther@suse.de>
23656 PR rtl-optimization/56921
23657 * loop-init.c (pass_rtl_move_loop_invariants): Add
23658 TODO_do_not_ggc_collect to todo_flags_finish.
23659 (pass_rtl_unswitch): Same.
23660 (pass_rtl_unroll_and_peel_loops): Same.
23661 (pass_rtl_doloop): Same.
23663 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
23665 * config/arm/arm.c (emit_multi_reg_push): New declaration
23666 for an existing function.
23667 (arm_emit_strd_push): New function.
23668 (arm_expand_prologue): Used here.
23669 (arm_emit_ldrd_pop): New function.
23670 (arm_expand_epilogue): Used here.
23671 (arm_get_frame_offsets): Update condition.
23672 (arm_emit_multi_reg_pop): Add a special case for load of a single
23673 register with writeback.
23675 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
23677 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
23680 2013-04-16 Richard Biener <rguenther@suse.de>
23682 PR tree-optimization/56756
23683 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
23684 (first_mem_ref_loc): New.
23685 (execute_sm): Place the load temporarily before a previous
23686 access instead of in the latch edge to ensure its SSA dependencies
23687 are defined at points dominating the load.
23689 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
23691 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
23692 correct fix by moving header and footer insn to the footer of
23693 the merged basic block. Clear BB_END of the merged-away block.
23695 PR middle-end/43631
23696 * emit-rtl.c (make_note_raw): New function.
23697 (link_insn_into_chain): New static inline function.
23698 (add_insn): Use it.
23699 (add_insn_before, add_insn_after): Factor insn chain linking code...
23700 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
23701 using link_insn_into_chain.
23702 (note_outside_basic_block_p): New helper function for emit_note_after
23703 and emit_note_before.
23704 (emit_note_after): Use nobb variant of add_insn_after if the note
23705 should not be contained in a basic block.
23706 (emit_note_before): Use nobb variant of add_insn_before if the note
23707 should not be contained in a basic block.
23708 (emit_note_copy): Use make_note_raw.
23709 (emit_note): Likewise.
23710 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
23711 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
23712 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
23713 the moved barrier the tail of the basic block it follows.
23714 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
23716 2013-04-15 Jakub Jelinek <jakub@redhat.com>
23718 PR tree-optimization/56962
23719 * gimple-ssa-strength-reduction.c (record_increment): Only set
23720 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
23721 either rhs1 or rhs2 is equal to c->base_expr.
23723 2013-04-15 Richard Biener <rguenther@suse.de>
23725 PR tree-optimization/56933
23726 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
23728 (GROUP_READ_WRITE_DEPENDENCE): Remove.
23729 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
23730 * tree-vect-data-refs.c (vect_analyze_group_access): Move
23731 dependence check ...
23732 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
23734 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
23735 GROUP_READ_WRITE_DEPENDENCE.
23737 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23739 * emit-rtl.c (reset_all_used_flags): New function.
23740 (verify_rtl_sharing): Call reset_all_used_flags before and after
23741 performing the checks.
23743 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23745 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
23746 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
23747 * config/arm/constraints.md (De): New constraint.
23748 * config/arm/neon.md (anddi3_neon): Delete.
23749 (neon_vand<mode>): Expand to standard anddi3 pattern.
23750 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
23751 Move earlier in the file.
23752 (neon_inv_logic_op2): Likewise.
23753 (arm_anddi_operand_neon): New predicate.
23755 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
23757 * configure.ac (gcc_cv_ld_as_needed): Set
23758 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
23759 Use -z ignore, -z record on *-*-solaris2*.
23760 (HAVE_LD_AS_NEEDED): Update comment.
23761 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
23762 * configure: Regenerate.
23763 * config.in: Regenerate.
23764 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
23765 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
23766 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
23767 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
23768 equivalents. Fix markup.
23769 * doc/tm.texi: Regenerate.
23771 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
23773 * config/i386/i386.opt: New option mstack-protector-guard=.
23774 * config/i386/i386-opts.h: Add enum stack_protector_guard.
23775 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
23776 TARGET_SSP_TLS_GUARD.
23777 * config/i386/i386.c (ix86_option_override_internal): Set
23778 ix86_stack_protector_guard.
23779 * config/i386/i386.md (stack_protect_set): Enable for
23780 TARGET_SSP_TLS_GUARD only.
23781 (stack_protect_set_<mode>): Ditto.
23782 (stack_protect_test): Ditto.
23783 (stack_protect_test_<mode>): Ditto.
23784 * doc/invoke.texi (i386 Option): Document.
23786 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
23789 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
23790 (S_MODES): Set H_MODE bit.
23791 (SF_MODES): Set only S_MODE and SF_MODE bits.
23792 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
23793 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
23794 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
23795 <MODE_FLOAT>: Likewise.
23797 2013-04-15 Joey Ye <joey.ye@arm.com>
23799 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
23801 2013-04-15 Joey Ye <joey.ye@arm.com>
23803 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
23805 (thumb_far_jump_used_p): Count instruction size and set
23808 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
23810 * reorg.c (fill_simple_delay_slots): Reindent block of code.
23811 * resource.c (mark_target_live_regs): Reformat conditional block.
23813 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23815 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
23816 notes, they are emitted only just before final.
23817 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
23819 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
23821 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
23822 * cfgrtl.c (delete_insn): Call it here instead.
23823 * lra-spills.c (lra_final_code_change): Use delete_insn.
23824 * haifa-sched.c (sched_remove_insn): Likewise.
23825 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
23826 returning to the nop pool.
23827 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
23828 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
23830 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23832 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
23833 * doc/tm.texi: Regenerated.
23835 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
23837 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
23840 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
23842 * df-core.c (df_find_def): Compare register numbers.
23843 (df_find_use): Likewise.
23845 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
23848 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
23849 lra_in_progress for return.
23851 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23853 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
23854 define_insn into define_insn_and_split and emit movsicc patterns.
23856 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
23858 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
23860 2013-04-12 Richard Biener <rguenther@suse.de>
23862 * tree-pass.h (TODO_do_not_ggc_collect): New.
23863 * passes.c (execute_one_ipa_transform_pass): Honor
23864 TODO_do_not_ggc_collect.
23865 (execute_one_pass): Likewise.
23868 2013-04-10 Richard Biener <rguenther@suse.de>
23870 * passes.c (init_optimization_passes): Remove reload pass.
23871 * ira.c (do_reload): Merge into ...
23873 (rest_of_handle_reload): Remove.
23874 (pass_reload): Likewise.
23875 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23876 of reload for vzeroupper pass placement.
23878 2013-04-12 Jakub Jelinek <jakub@redhat.com>
23880 PR tree-optimization/56918
23881 PR tree-optimization/56920
23882 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
23883 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
23884 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
23885 use wide_mul_with_sign method.
23887 2013-04-12 Richard Biener <rguenther@suse.de>
23889 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
23890 not be considered a gimple constant.
23892 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
23894 * fold-const.c (const_binop): Handle vector shifts by a scalar.
23895 (fold_binary_loc): Call const_binop also for mixed vector-scalar
23898 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
23899 Jakub Jelinek <jakub@redhat.com>
23901 * opts.c: Include diagnostic-color.h.
23902 (common_handle_option): Handle OPT_fdiagnostics_color_.
23903 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
23904 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
23905 (diagnostic-color.o): New.
23906 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
23907 (diagnostic_color_rule): New enum.
23908 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
23909 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
23910 the location string.
23911 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
23912 either NULL, or color kind.
23913 * diagnostic-color.c: New file.
23914 * diagnostic-color.h: New file.
23915 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
23917 * doc/invoke.texi (-fdiagnostics-color): Document.
23918 * pretty-print.h (pp_show_color): Define.
23919 (struct pretty_print_info): Add show_color field.
23920 * diagnostic.c: Include diagnostic-color.h.
23921 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
23922 macros. Colorize error:, warning: etc. strings and also the location
23924 (diagnostic_show_locus): Colorize the caret line.
23925 * pretty-print.c: Include diagnostic-color.h.
23926 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
23927 inside of %< %> quotes or quoted through q format modifier.
23929 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23931 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
23933 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23935 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
23936 code in CC_NZ mode.
23937 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
23940 2013-04-11 Marek Polacek <polacek@redhat.com>
23942 PR tree-optimization/48184
23943 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
23945 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
23947 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
23948 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
23949 (skip_simple_arithmetic): Tidy up.
23950 * tree.h (skip_simple_constant_arithmetic): Declare.
23952 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
23954 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
23956 2013-04-11 Richard Biener <rguenther@suse.de>
23958 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23959 generate vector constants.
23961 2013-04-11 Richard Biener <rguenther@suse.de>
23963 PR tree-optimization/56878
23964 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
23965 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
23967 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
23968 Prefer to align the DR with the most invariant base address.
23970 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
23972 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
23975 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
23977 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
23978 floating-point vector comparisons against 0.
23980 2013-04-11 Jakub Jelinek <jakub@redhat.com>
23982 PR tree-optimization/56899
23983 * fold-const.c (extract_muldiv_1): Apply distributive law
23984 only if TYPE_OVERFLOW_WRAPS (ctype).
23986 2013-04-11 Bin Cheng <bin.cheng@arm.com>
23989 * ira-costs.c (scan_one_insn): Check whether the source rtx of
23990 loading has side effect.
23992 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
23994 * config/sparc/sparc.c: Include tree-pass.h.
23995 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
23996 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
23997 head of file. Change return type. Split off gate function.
23998 (sparc_gate_work_around_errata): New function.
23999 (pass_work_around_errata): New pass definition.
24000 (insert_pass_work_around_errata) New pass insert definition to
24001 insert pass_work_around_errata just after delayed-branch scheduling.
24002 (sparc_option_override): Insert the pass.
24003 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
24005 2013-04-10 David S. Miller <davem@davemloft.net>
24007 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
24008 or -mcpu=hypersparc.
24010 * target.def (cstore_mode): New hook.
24011 * target.h: Include insn-codes.h
24012 * targhooks.c: Likewise.
24013 (default_cstore_mode): New function.
24014 * targhooks.h: Declare it.
24015 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
24016 * doc/tm.texi: Rebuild.
24017 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
24018 target hook, rather than inspecting the insn_data.
24019 * config/sparc/sparc.c (sparc_cstore_mode): New function.
24020 (TARGET_CSTORE_MODE): Redefine.
24021 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
24023 * config/sparc/predicates.md (cstore_result_operand): New special
24025 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
24026 Use it for operand 0.
24027 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
24028 (*snesi_special): Likewise.
24029 (*snesi_zero): Likewise.
24030 (*seqsi_zero): Likewise.
24031 (*sltu_insn): Likewise.
24032 (*sgeu_insn): Likewise.
24033 (*seqdi_special): Make operand 0 and comparison operation be of
24035 (*snedi_special): Likewise.
24036 (*snedi_special_vis3): Likewise.
24037 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
24038 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
24039 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
24040 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
24041 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
24042 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
24043 (*sltu_extend_sp64): Likewise.
24044 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
24045 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
24046 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
24047 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
24048 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
24050 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
24052 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
24053 (aarch64_start_file): Use the new function.
24055 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
24056 Jason Merrill <jason@redhat.com>
24058 * common.opt: Add -gdwarf.
24059 * opts.c (common_handle_option): Handle it.
24060 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
24062 2013-04-10 Richard Biener <rguenther@suse.de>
24064 * passes.c (execute_todo): Do not call ggc_collect conditional here.
24065 (execute_one_ipa_transform_pass): But unconditionally here.
24066 (execute_one_pass): And here.
24067 (init_optimization_passes): Remove reload pass.
24068 * tree-pass.h (TODO_ggc_collect): Remove.
24069 (pass_reload): Likewise.
24070 * ira.c (do_reload): Merge into ...
24072 (rest_of_handle_reload): Remove.
24073 (pass_reload): Likewise.
24074 * config/i386/i386.c (ix86_option_override): Refer to ira instead
24075 of reload for vzeroupper pass placement.
24076 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
24077 and todo_flags_finish of all passes.
24079 2013-04-10 Richard Biener <rguenther@suse.de>
24081 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
24082 first_const_oprnd field, rename first_def_type to first_op_type.
24083 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
24084 (vect_get_and_check_slp_defs): Always use the type of the
24085 operand. Allow mixed vect_external_def, vect_constant_def types.
24086 (vect_get_constant_vectors): Handle mixed vect_external_def,
24087 vect_constant_def types.
24089 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
24091 PR tree-optimization/55524
24092 * tree-ssa-math-opts.c
24093 (convert_mult_to_fma): Don't use an fms construct
24094 when we don't have an fms operation, but fnma, and it looks
24095 likely that we'll be able to use the latter.
24097 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
24099 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
24101 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
24102 inline fail caused by overwritable functions.
24104 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
24106 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
24107 unnecessary bits in the constant power of two case.
24109 2013-04-10 Richard Biener <rguenther@suse.de>
24111 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
24112 broken code swapping operands.
24113 (vect_build_slp_tree): Do not compute load permutations here.
24114 (vect_analyze_slp_instance): Compute load permutations here,
24115 after building the SLP tree.
24117 2013-04-09 Christian Bruel <christian.bruel@st.com>
24119 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
24120 of next/prev_real_insn.
24122 2013-04-09 Jan Hubicka <jh@suse.cz>
24124 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
24125 Drop aliased parameter.
24126 (function_and_variable_visibility): Do not handle alias pairs.
24127 * cgraph.c (varpool_externally_visible_p): Update prototype.
24128 * varpool.c (varpool_add_new_variable): Update.
24130 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24132 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
24134 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
24136 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
24138 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
24140 2013-04-09 Marek Polacek <polacek@redhat.com>
24142 PR tree-optimization/48762
24143 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
24145 2013-04-09 Richard Biener <rguenther@suse.de>
24147 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
24149 (vect_build_slp_tree): Likewise.
24150 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
24151 calculating the cost of a SLP instance.
24152 (vect_analyze_slp_instance): Use it from here, after building
24155 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24157 PR middle-end/56883
24158 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
24159 expand_omp_for_static_chunk): Use simple_p = true in
24160 force_gimple_operand_gsi calls when assigning to addressable decls.
24162 2013-04-09 Jeff Law <law@redhat.com>
24164 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
24165 when the boolean was created by converting a wider object which
24166 had a boolean range.
24168 2013-04-09 Richard Biener <rguenther@suse.de>
24170 * tree-vectorizer.h (slp_void_p): Remove.
24171 (slp_tree): Typedef before _slp_tree declaration.
24172 (struct _slp_tree): Use a vector of slp_tree as children.
24173 (vect_get_place_in_interleaving_chain): Remove.
24174 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
24176 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
24178 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
24179 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
24180 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
24181 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
24182 Use slp_node instead of slp_void_p and adjust.
24184 2013-04-09 Richard Biener <rguenther@suse.de>
24186 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
24187 work that is not necessary.
24189 2013-04-09 Jakub Jelinek <jakub@redhat.com>
24191 PR tree-optimization/56854
24192 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
24193 forward into clobber stmts if it would change MEM_REF lhs into
24196 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
24198 * tree.c (type_hash_lookup, type_hash_add): Make static.
24199 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
24201 2013-04-09 Richard Biener <rguenther@suse.de>
24203 * tree.h (unsave_expr_now): Remove.
24204 * tree-inline.c (mark_local_for_remap_r): Remove.
24205 (unsave_expr_1): Likewise.
24206 (unsave_r): Likewise.
24207 (unsave_expr_now): Likewise.
24208 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
24209 (propagate_tree_value): Likewise.
24211 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
24213 * doc/rtl.texi (sequence): Rewrite documentation to match the
24214 current use of SEQUENCE rtl objects.
24215 * rtl.def (SEQUENCE): Likewise.
24217 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
24218 Update documentation.
24219 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
24220 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
24222 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
24224 2013-04-08 Teresa Johnson <tejohnson@google.com>
24226 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
24227 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
24229 (estimate_edge_size_and_time): Add comment to suggest using rounding
24231 (estimate_node_size_and_time): Ditto.
24232 (remap_edge_change_prob): Use helper rounding divide methods.
24233 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
24234 (gimple_mod_pow2_value_transform): Ditto.
24235 (gimple_mod_subtract_transform): Ditto.
24236 (gimple_ic_transform): Ditto.
24237 (gimple_stringops_transform): Ditto.
24238 * stmt.c (conditional_probability): Ditto.
24239 (emit_case_dispatch_table): Ditto.
24240 * lto-cgraph.c (merge_profile_summaries): Ditto.
24241 * tree-optimize.c (execute_fixup_cfg): Ditto.
24242 * cfgcleanup.c (try_forward_edges): Ditto.
24243 * cfgloopmanip.c (scale_loop_profile): Ditto.
24245 (duplicate_loop_to_header_edge): Ditto.
24246 (lv_adjust_loop_entry_edge): Ditto.
24247 * tree-vect-loop.c (vect_transform_loop): Ditto.
24248 * profile.c (compute_branch_probabilities): Ditto.
24249 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
24250 * lto-streamer-in.c (input_cfg): Ditto.
24251 * gimple-streamer-in.c (input_bb): Ditto.
24252 * ipa-cp.c (update_profiling_info): Ditto.
24253 (update_specialized_profile): Ditto.
24254 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
24255 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
24257 * sched-rgn.c (compute_dom_prob_ps): Ditto.
24258 (compute_trg_info): Ditto.
24259 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
24260 (purge_dead_edges): Ditto.
24261 * loop-unswitch.c (unswitch_loop): Ditto.
24262 * cgraphclones.c (cgraph_clone_edge): Ditto.
24263 (cgraph_clone_node): Ditto.
24264 * tree-inline.c (copy_bb): Ditto.
24265 (copy_edges_for_bb): Ditto.
24266 (initialize_cfun): Ditto.
24267 (copy_cfg_body): Ditto.
24268 (expand_call_inline): Ditto.
24270 2013-04-08 Kai Tietz <ktietz@redhat.com>
24272 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
24273 TARGET_CYGWIN64 by TARGET_64BIT.
24275 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24277 * config/epiphany/epiphany.md (GPR_1): New constant.
24278 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
24279 * config/epiphany/epiphany.c (gen_compare_reg):
24280 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
24281 is already in place.
24282 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
24283 Don't require being called during rtl expansion; If y operlaps r0,
24285 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
24286 (epiphany_expand_epilogue): Likewise.
24288 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
24289 Don't use CC_FPmode for ORDERED / UNORDERED.
24290 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
24292 * config/epiphany/constraints.md (CnL): New constraint.
24293 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
24294 * config/epiphany/predicates.md (add_operand): Allow 1024.
24296 * config/epiphany/epiphany.md (logical_op): New code iterator.
24297 (op_mnc): New code attribute.
24298 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
24299 (mov_f+1, mov_f+2): New peephole2 patterns.
24301 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
24302 (cstoresi4): Also allow re-use of zero result when doing a NE
24303 comparison to a non-zero operand.
24304 Use (clobber (scratch)) for first insn if the gpr output is not needed.
24306 * config/epiphany/epiphany.md (<insn_opname>v2si3):
24307 Use gen_addsi3_i / gen_subsi3_i.
24309 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24313 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
24314 contain anything but clobbers, at most one __builtin_stack_restore,
24315 optionally debug stmts and final resx, and if it has at least one
24316 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
24317 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
24318 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
24319 which isn't defaut definition, remove them.
24320 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
24321 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
24322 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
24323 with MEM_REF LHS with SSA_NAME address.
24325 2013-04-08 Jeff Law <law@redhat.com>
24327 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
24329 2013-04-08 Richard Biener <rguenther@suse.de>
24331 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
24333 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
24334 determined vector type.
24335 (vect_analyze_data_refs): Likewise.
24336 (vect_get_new_vect_var): Adjust.
24337 (vect_create_destination_var): Preserve SSA name versions.
24338 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
24339 not dump anything here.
24341 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
24343 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
24344 Add member lr_slot_known.
24345 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
24347 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
24348 Remove code that sets lr_slot_offset according to what a previous
24349 version of epiphany_emit_save_restore used to do.
24350 (epiphany_emit_save_restore): When doing an lr save or restore,
24351 set/verify lr_slot_known and lr_slot_offset.
24353 2013-04-08 Xinyu Qi <xyqi@marvell.com>
24356 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
24359 2013-04-08 Richard Biener <rguenther@suse.de>
24361 * alias.c (find_base_term): Fix thinko in previous change.
24363 2013-04-08 Jakub Jelinek <jakub@redhat.com>
24365 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
24366 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
24367 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
24368 if possible to compute val.
24369 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
24370 For QImode integers don't require anything about precision. Use
24371 const_with_all_bytes_same to find out if the constant doesn't have
24372 repeated bytes in it.
24374 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24376 * config/s390/s390.c (s390_expand_insv): Only accept insertions
24379 2013-04-08 Marek Polacek <polacek@redhat.com>
24381 PR rtl-optimization/48182
24382 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
24385 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24388 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
24389 nuses, make sure we have a label.
24391 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
24394 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
24395 (rs6000_emit_swdiv_low_precision): Remove.
24396 (rs6000_emit_swdiv): Rewrite to handle between one and four
24397 iterations of Newton-Raphson generally; modify required number of
24398 iterations for some cases.
24399 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
24401 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
24403 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
24404 set-but-unused variable.
24406 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
24407 basic blocks of released function bodies garbage-collectable.
24409 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
24410 (struct rtl_opt_pass): Add TODO_df_finish.
24412 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
24414 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24416 * config/arm/constraints.md (q): New constraint.
24417 * config/arm/ldrdstrd.md: New file.
24418 * config/arm/arm.md (ldrdstrd.md) New include.
24419 (arm_movdi): Use "q" instead of "r" constraint
24420 for double-word memory access.
24421 (movdf_soft_insn): Likewise.
24422 * config/arm/vfp.md (movdi_vfp): Likewise.
24423 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
24424 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
24425 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
24426 (mem_ok_for_ldrd_strd): Likewise.
24427 (output_move_double): Update assertion.
24429 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24431 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
24433 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24435 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
24436 define_insn_and_split.
24437 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
24439 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24441 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
24442 define_insn_and_split.
24443 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
24444 (shiftsi3_compare): New pattern.
24445 (rrx): New pattern.
24446 * config/arm/unspecs.md (UNSPEC_RRX): New.
24448 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24450 * config/arm/arm.md (negdi_extendsidi): New pattern.
24451 (negdi_zero_extendsidi): Likewise.
24453 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24455 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
24456 define_insn_and_split.
24457 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
24458 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
24460 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24462 * config/arm/arm.md (arm_subdi3): Convert define_insn into
24463 define_insn_and_split.
24464 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
24465 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
24467 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24469 * config/arm/arm.md (subsi3_carryin): New pattern.
24470 (subsi3_carryin_const): Likewise.
24471 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
24472 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
24474 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24476 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
24478 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
24480 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
24481 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
24483 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24485 * config/arm/arm.c (arm_expand_builtin): Change fcode
24486 type to unsigned int.
24488 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24490 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
24492 2013-04-04 Ian Lance Taylor <iant@google.com>
24494 * doc/standards.texi (Standards): The Go frontend supports the Go 1
24497 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
24499 PR middle-end/56729
24500 * df-scan.c (df_insn_delete): Disable failing assert.
24502 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24504 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
24505 New function prototype.
24506 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
24507 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
24508 (arm_builtin_vectorized_function): New function.
24510 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24512 * config/arm/arm_neon_builtins.def: New file.
24513 * config/arm/arm.c (neon_builtin_data): Move contents to
24514 arm_neon_builtins.def.
24515 (enum arm_builtins): Include neon builtin definitions.
24516 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
24517 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
24519 2013-04-04 Marek Polacek <polacek@redhat.com>
24521 PR tree-optimization/48186
24522 * predict.c (maybe_hot_frequency_p): Return false if
24523 HOT_BB_FREQUENCY_FRACTION is 0.
24524 (cgraph_maybe_hot_edge_p): Likewise.
24526 2013-04-04 Richard Biener <rguenther@suse.de>
24528 PR tree-optimization/56826
24529 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
24532 2013-04-04 Richard Biener <rguenther@suse.de>
24534 PR tree-optimization/56213
24535 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
24536 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
24538 2013-04-04 Richard Biener <rguenther@suse.de>
24540 PR tree-optimization/56837
24541 * tree-loop-distribution.c (classify_partition): For non-zero
24542 values require that the value has the same precision as its
24543 mode to be useful as memset value.
24545 2013-04-03 Nick Clifton <nickc@redhat.com>
24547 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
24548 (fmssf4): Use fmsf.s on E3V5 architectures.
24549 (fnmasf4): Use fnmaf.s on E3V5 architectures.
24550 (fnmssf4): Use fnmsf.s on E3V5 architectures.
24552 2013-04-03 Jeff Law <law@redhat.com>
24554 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
24555 (lra-eliminations.o): Likewise.
24557 2013-04-03 Teresa Johnson <tejohnson@google.com>
24559 * gcov-io.c (compute_working_sets): Moved most of body of old
24560 compute_working_sets here from profile.c.
24561 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
24562 (gcov_working_set_t): Moved typedef here from basic-block.h
24563 (compute_working_set): Declare.
24564 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
24565 (get_working_sets): Renamed from compute_working_set,
24566 replace most of body with call to new compute_working_sets.
24567 (get_exec_counts): Replace call to compute_working_sets
24568 to get_working_sets.
24569 * profile.h (get_working_sets): Renamed from compute_working_set.
24570 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
24571 to get_working_sets.
24572 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
24573 * gcov-dump.c (dump_working_sets): New function.
24575 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
24577 * hwint.c (sext_hwi, zext_hwi): New functions.
24578 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
24579 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
24580 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
24581 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
24582 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
24583 (sext_hwi, zext_hwi): New functions.
24585 2013-04-03 Jeff Law <law@redhat.com>
24587 PR tree-optimization/56799
24588 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
24589 back test for widening conversion erroneously dropped in prior change.
24591 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24594 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
24595 instead of next_real_insn.
24597 2013-04-03 Marek Polacek <polacek@redhat.com>
24600 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
24602 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24605 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
24607 (thumb1_output_casesi): Likewise.
24608 (thumb2_output_casesi): Likewise.
24610 2013-04-03 Richard Biener <rguenther@suse.de>
24612 PR tree-optimization/56817
24613 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
24615 (tree_unroll_loops_completely_1): ... new function to manually
24616 walk the loop tree, properly defering outer loops of unrolled
24617 loops to later iterations.
24619 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24621 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
24622 (vectorizable_load): Likewise.
24623 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
24624 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
24626 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24628 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
24631 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
24633 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
24635 2013-04-03 Bin Cheng <bin.cheng@arm.com>
24637 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
24639 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24641 PR tree-optimization/56790
24642 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
24645 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
24647 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
24649 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
24650 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
24653 2013-04-03 Jakub Jelinek <jakub@redhat.com>
24656 * tree.h (force_folding_builtin_constant_p): New decl.
24657 * builtins.c (force_folding_builtin_constant_p): New variable.
24658 (fold_builtin_constant_p): Fold immediately also if
24659 force_folding_builtin_constant_p.
24661 2013-04-03 Richard Biener <rguenther@suse.de>
24663 PR tree-optimization/56812
24664 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
24665 DRs of the same interleaving chain are independent.
24667 2013-04-02 Jason Merrill <jason@redhat.com>
24669 * gdbinit.in (pbb): Use debug fn.
24671 2013-04-02 Lawrence Crowl <crowl@google.com>
24673 * sese.h (struct ivtype_map_elt_s): Remove unused.
24674 (extern debug_ivtype_map): Remove unused.
24675 (extern eq_ivtype_map_elts): Remove unused.
24676 * sese.c (debug_ivtype_map): Removed unused.
24677 (debug_ivtype_map_1): Removed unused.
24678 (debug_ivtype_elt): Remove unused.
24679 (eq_ivtype_map_elts): Remove unused.
24681 2013-04-02 Kai Tietz <ktietz@redhat.com>
24684 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
24685 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
24686 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
24688 (legitimize_pe_coff_symbol): Likewise.
24689 (is_imported_p): New helper-function.
24690 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
24691 for Windows x64 targets.
24692 (ix86_expand_prologue): Optimize for pe-coff targets.
24693 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
24694 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
24695 medium/large code-model.
24696 (legitimize_pic_address): Likewise.
24697 (legitimize_tls_address): Likewise.
24698 (ix86_expand_call): Likewise.
24699 (x86_output_mi_thunk): Likewise.
24700 (get_dllimport_decl): Add new beimport argument.
24701 (construct_plt_address): Don't assert for x64 pe-coff targets.
24702 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
24704 (SYMBOL_FLAG_STUBVAR): New macro.
24705 (SYMBOL_REF_STUBVAR_P): Likewise.
24706 * config/i386/winnt.c (stub_list): New structure.
24707 (stub_head): New local variable.
24708 (i386_pe_record_stub): New function.
24709 (i386_pe_file_end): Emit refptr-stubs.
24711 2013-04-02 Jakub Jelinek <jakub@redhat.com>
24713 PR rtl-optimization/56745
24714 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
24715 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
24718 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
24719 and both of them are MEM_REFs, just compare first argument for
24720 equality and attempt to deal even with differing offsets.
24723 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
24724 of gimple_clobber_p to be MEM_REF.
24725 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
24726 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
24727 after gimplification.
24728 * asan.c (get_mem_ref_of_assignment): Don't instrument
24729 gimple_clobber_p stmts.
24730 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
24731 gimple_clobber_p stmt if they have MEM_REF lhs and
24732 are dead because of another gimple_clobber_p stmt.
24733 * tree-ssa-live.c (clear_unused_block_pointer): Treat
24734 gimple_clobber_p stmts like debug stmts.
24735 (remove_unused_locals): Remove clobbers with MEM_REF lhs
24736 that refer to unused VAR_DECLs or uninitialized values.
24737 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
24738 gimple_clobber_p stmts if they refer to removed parameters.
24739 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
24742 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
24744 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
24745 using SWI48 mode attribute.
24747 2013-04-02 Wei Mi <wmi@google.com>
24749 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
24750 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
24751 *<rotate_insn><mode>3_mask in i386.md.
24753 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
24755 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
24757 2013-04-02 Richard Biener <rguenther@suse.de>
24759 PR tree-optimization/56778
24760 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
24761 Runtime alias tests are not supported for gather loads.
24762 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
24763 stmts referenced from SSA operands before updating SSA form.
24765 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
24766 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
24768 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
24769 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
24770 * config/arm/cortex-a53.md: New file.
24771 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
24772 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
24773 * config/arm/arm.c (arm_issue_rate): Likewise.
24774 * config/arm/arm-tune.md: Regenerate
24775 * config/arm/arm-tables.opt: Regenerate.
24776 * config/arm/arm-cores.def: Add cortex-a53.
24778 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
24780 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
24783 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24785 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
24786 scalar load/store operations using B/H registers.
24787 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
24789 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
24791 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
24793 * config/aarch64/aarch64.c
24794 (aarch64_simd_scalar_immediate_valid_for_move): New.
24795 * config/aarch64/aarch64-protos.h
24796 (aarch64_simd_scalar_immediate_valid_for_move): New.
24797 * config/aarch64/constraints.md (Dh, Dq): New.
24798 * config/aarch64/iterators.md (hq): New.
24800 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
24802 * reorg.c (get_branch_condition): Deal with conditional returns.
24803 (fill_simple_delay_slots): Remove dead code dealing with jumps.
24805 2013-04-01 Wei Mi <wmi@google.com>
24807 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
24808 Truncate operand 2 using %b asm operand modifier.
24809 (*<shift_insn><mode>3_mask): Ditto.
24810 (*<rotate_insn><mode>3_mask): Ditto.
24812 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
24814 PR middle-end/56798
24815 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
24817 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
24819 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
24821 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
24823 2013-03-30 Lawrence Crowl <crowl@google.com>
24825 * dse.c (clear_alias_sets): Remove never set.
24826 (disqualified_clear_alias_sets): Remove never set.
24827 (clear_alias_mode_pool): Remove never set.
24828 (dse_step0): Remove condition that is never true.
24829 (canon_address): Remove condition that is never true.
24830 (dse_step7): Remove condition that is never true.
24831 (rest_of_handle_dse): Remove condition that is never true.
24832 (rest_of_handle_dse::did_global): Remove never read from above.
24833 (dse_step2_spill): Remove never called from above.
24834 (dse_step5_spill): Remove never called from above.
24836 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
24838 * doc/md.texi (Standard Names) <casesi>: Update documentation for
24839 JUMP_TABLE_DATA changes.
24840 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
24841 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
24842 (Insns) <jump_table_data>: New entry.
24843 * doc/tm.texi: Regenerate.
24845 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
24847 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
24848 for table jump at the end of a basic block using tablejump_p.
24849 * targhooks.c (default_invalid_within_doloop): Likewise.
24850 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
24851 target hook implementation that is identical to the default hook.
24852 (rs6000_invalid_within_doloop): Remove.
24854 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
24855 unused variable from tablejump_p call.
24857 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
24858 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
24859 (INSN_DELETED_P): Likewise.
24860 (emit_jump_table_data): New prototype.
24861 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
24862 after 4th as unused.
24863 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
24864 * sched-vis.c (print_insn): Likewise.
24865 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
24866 insn for compatibility with back ends that use next_active_insn to
24867 identify jump table data.
24868 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
24869 (remove_insn): Likewise.
24870 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
24872 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
24873 (emit_jump_table_data): New function.
24875 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
24876 basic block, a JUMP_TABLE_DATA never is.
24877 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
24878 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
24879 off from code handling real insns.
24880 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
24881 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
24882 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
24883 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
24884 is not a NONDEBUG_INSN_P.
24885 * ira-costs.c (scan_one_insn): Likewise.
24886 * jump.c (mark_all_labels): Likewise.
24887 (mark_jump_label_1): Likewise.
24888 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24889 * lra.c (get_insn_freq): Expect all insns reaching here to be in
24891 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
24892 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
24893 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
24894 JUMP_TABLE_DATA_P insns.
24895 (calculate_elim_costs_all_insns): Likewise.
24896 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
24897 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
24898 (delete_output_reload): Code style fixups.
24899 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
24900 insn flags on this non-insn.
24901 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
24902 as scheduling barriers, for pre-change compatibility.
24903 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
24904 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
24906 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
24907 redundant JUMP_TABLE_DATA_P test.
24908 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
24909 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24910 (frv_for_each_packet): Likewise.
24911 * config/i386/i386.c (min_insn_size): Likewise.
24912 (ix86_avoid_jump_mispredicts): Likewise.
24913 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24914 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
24915 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24916 (mips16_insn_length): Robustify.
24917 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
24918 (mips16_split_long_branches): Likewise.
24919 * config/pa/pa.c (pa_combine_instructions): Likewise.
24920 * config/rs6000/rs6000.c (get_next_active_insn): Treat
24921 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
24922 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
24923 as contributing to pool range lengths.
24924 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
24925 Remove redundant JUMP_TABLE_DATA_P test.
24926 (sh_loop_align): Likewise.
24927 (split_branches): Likewise.
24928 (sh_insn_length_adjustment): Likewise.
24929 * config/spu/spu.c (get_branch_target): Likewise.
24931 2013-03-29 Jan Hubicka <jh@suse.cz>
24933 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
24934 gcov streaming; stream hot bb threshold to ltrans.
24935 * predict.c (get_hot_bb_threshold): Break out from ....
24936 (maybe_hot_count_p): ... here.
24937 (set_hot_bb_threshold): New function.
24938 * lto-section-in.c (lto_section_name): Add profile.
24939 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
24940 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
24941 and data-streamer.h
24942 (histogram_entry): New structure.
24943 (histogram, histogram_pool): New global vars.
24944 (histogram_hash): New structure.
24945 (histogram_hash::hash): New method.
24946 (histogram_hash::equal): Likewise.
24947 (account_time_size): New function.
24948 (cmp_counts): New function.
24949 (dump_histogram): New function.
24950 (ipa_profile_generate_summary): New function.
24951 (ipa_profile_write_summary): New function.
24952 (ipa_profile_read_summary): New function.
24953 (ipa_profile): Decide on threshold.
24954 (pass_ipa_profile): Add ipa_profile_write_summary and
24955 ipa_profile_read_summary.
24956 * Makefile.in (ipa.o): Update dependencies.
24957 * lto-streamer.h (LTO_section_ipa_profile): New section.
24959 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
24961 * tree.h (VAR_P): New.
24963 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
24966 * doc/invoke.texi ([-fwhole-program]): Fix typo.
24968 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
24970 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
24971 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24972 (control_flow_insn_p): Likewise.
24973 * cfgrtl.c (duplicate_insn_chain): Likewise.
24974 * final.c (get_attr_length_1): Likewise.
24975 (shorten_branches): Likewise.
24976 (final_scan_insn): Likewise.
24977 * function.c (instantiate_virtual_regs): Likewise.
24978 * gcse.c (insert_insn_end_basic_block): Likewise.
24979 * ira-costs.c (scan_one_insn): Likewise.
24980 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
24981 * lra.c (check_rtl): Likewise.
24982 * reload1.c (elimination_costs_in_insn): Likewise.
24983 * reorg.c (follow_jumps): Likewise.
24985 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
24986 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
24987 (thumb_far_jump_used_p): Likewise.
24988 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
24989 (workaround_speculation): Likewise.
24990 (add_sched_insns_for_speculation): Likewise.
24991 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
24992 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
24993 (frv_for_each_packet): Likewise.
24994 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
24995 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
24996 (final_emit_insn_group_barriers): Likewise.
24997 * config/m32r/m32r.c (m32r_is_insn): Likewise.
24998 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
24999 (mips16_insn_length): Likewise.
25000 * config/pa/pa.c (pa_reorg): Likewise.
25001 (pa_combine_instructions): Likewise.
25002 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
25003 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
25004 (sh_reorg): Likewise.
25005 (split_branches): Likewise.
25006 * config/spu/spu.c (get_branch_target): Likewise.
25008 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
25011 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
25013 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
25014 Fix declaration name.
25016 2013-03-28 Lawrence Crowl <crowl@google.com>
25018 * graphds.h (struct graph.indicies): Remove unused.
25019 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
25020 (SCOP_ORIGINAL_PDDRS): Remove unused.
25021 * sese.h (extern insert_loop_close_phis): Removed unused.
25022 (extern insert_guard_phis): Removed unused.
25023 (extern ivtype_map_elt_info): Removed unused.
25024 (new_ivtype_map_elt): Removed unused.
25025 * sese.c (ivtype_map_elt_info): Removed unused.
25027 2013-03-28 Lawrence Crowl <crowl@google.com>
25029 * Makefile.in: Add several missing include dependences.
25031 (test-dump.o): New. This object is not added to any executable,
25032 but is present for ad-hoc testing.
25034 (debug (const bitmap_head_def &)): New.
25035 (debug (const bitmap_head_def *)): New.
25037 (extern debug (const bitmap_head_def &)): New.
25038 (extern debug (const bitmap_head_def *)): New.
25040 (debug (edge_def &)): New.
25041 (debug (edge_def *)): New.
25043 (debug (basic_block_def &)): New.
25044 (debug (basic_block_def *)): New.
25046 (dump_node (const_tree, int, FILE *)): Correct source file.
25048 (debug (die_struct &)): New.
25049 (debug (die_struct *)): New.
25051 (extern debug (die_struct &)): New.
25052 (extern debug (die_struct *)): New.
25053 * gimple-pretty-print.c
25054 (debug (gimple_statement_d &)): New.
25055 (debug (gimple_statement_d *)): New.
25056 * gimple-pretty-print.h
25057 (extern debug (gimple_statement_d &)): New.
25058 (extern debug (gimple_statement_d *)): New.
25060 (debug (ira_allocno_copy &)): New.
25061 (debug (ira_allocno_copy *)): New.
25062 (debug (ira_allocno &)): New.
25063 (debug (ira_allocno *)): New.
25065 (extern debug (ira_allocno_copy &)): New.
25066 (extern debug (ira_allocno_copy *)): New.
25067 (extern debug (ira_allocno &)): New.
25068 (extern debug (ira_allocno *)): New.
25070 (debug (live_range &)): New.
25071 (debug (live_range *)): New.
25073 (debug (lra_live_range &)): New.
25074 (debug (lra_live_range *)): New.
25076 (debug (lra_live_range &)): New.
25077 (debug (lra_live_range *)): New.
25079 (debug (omega_pb_d &)): New.
25080 (debug (omega_pb_d *)): New.
25082 (extern debug (omega_pb_d &)): New.
25083 (extern debug (omega_pb_d *)): New.
25085 (debug (const rtx_def &)): New.
25086 (debug (const rtx_def *)): New.
25088 (debug_tree (tree): Move within file.
25089 (debug_raw (const tree_node &)): New.
25090 (debug_raw (const tree_node *)): New.
25091 (dump_tree_via_hooks (const tree_node *, int)): New.
25092 (debug (const tree_node &)): New.
25093 (debug (const tree_node *)): New.
25094 (debug_verbose (const tree_node &)): New.
25095 (debug_verbose (const tree_node *)): New.
25096 (debug_head (const tree_node &)): New.
25097 (debug_head (const tree_node *)): New.
25098 (debug_body (const tree_node &)): New.
25099 (debug_body (const tree_node *)): New.
25100 (debug_vec_tree (tree): Move and reimplement in terms of dump.
25101 (debug (vec<tree, va_gc> &)): New.
25102 (debug (vec<tree, va_gc> *)): New.
25104 (extern debug (const rtx_def &)): New.
25105 (extern debug (const rtx_def *)): New.
25107 (debug_raw (simple_bitmap_def &)): New.
25108 (debug_raw (simple_bitmap_def *)): New.
25109 (debug (simple_bitmap_def &)): New.
25110 (debug (simple_bitmap_def *)): New.
25112 (extern debug (simple_bitmap_def &)): New.
25113 (extern debug (simple_bitmap_def *)): New.
25114 (extern debug_raw (simple_bitmap_def &)): New.
25115 (extern debug_raw (simple_bitmap_def *)): New.
25117 (debug (vinsn_def &)): New.
25118 (debug (vinsn_def *)): New.
25119 (debug_verbose (vinsn_def &)): New.
25120 (debug_verbose (vinsn_def *)): New.
25121 (debug (expr_def &)): New.
25122 (debug (expr_def *)): New.
25123 (debug_verbose (expr_def &)): New.
25124 (debug_verbose (expr_def *)): New.
25125 (debug (vec<rtx> &)): New.
25126 (debug (vec<rtx> *)): New.
25128 (extern debug (vinsn_def &)): New.
25129 (extern debug (vinsn_def *)): New.
25130 (extern debug_verbose (vinsn_def &)): New.
25131 (extern debug_verbose (vinsn_def *)): New.
25132 (extern debug (expr_def &)): New.
25133 (extern debug (expr_def *)): New.
25134 (extern debug_verbose (expr_def &)): New.
25135 (extern debug_verbose (expr_def *)): New.
25136 (extern debug (vec<rtx> &)): New.
25137 (extern debug (vec<rtx> *)): New.
25139 (_list_iter_cond_expr): Make inline instead of static.
25141 (debug (sreal &)): New.
25142 (debug (sreal *)): New.
25144 (extern debug (sreal &)): New.
25145 (extern debug (sreal *)): New.
25147 (extern debug_raw (const tree_node &)): New.
25148 (extern debug_raw (const tree_node *)): New.
25149 (extern debug (const tree_node &)): New.
25150 (extern debug (const tree_node *)): New.
25151 (extern debug_verbose (const tree_node &)): New.
25152 (extern debug_verbose (const tree_node *)): New.
25153 (extern debug_head (const tree_node &)): New.
25154 (extern debug_head (const tree_node *)): New.
25155 (extern debug_body (const tree_node &)): New.
25156 (extern debug_body (const tree_node *)): New.
25157 (extern debug (vec<tree, va_gc> &)): New.
25158 (extern debug (vec<tree, va_gc> *)): New.
25160 (debug (struct loop &)): New.
25161 (debug (struct loop *)): New.
25162 (debug_verbose (struct loop &)): New.
25163 (debug_verbose (struct loop *)): New.
25164 * tree-dump.c: Add header dependence.
25166 (extern debug (struct loop &)): New.
25167 (extern debug (struct loop *)): New.
25168 (extern debug_verbose (struct loop &)): New.
25169 (extern debug_verbose (struct loop *)): New.
25171 (debug (data_reference &)): New.
25172 (debug (data_reference *)): New.
25173 (debug (vec<data_reference_p> &)): New.
25174 (debug (vec<data_reference_p> *)): New.
25175 (debug (vec<ddr_p> &)): New.
25176 (debug (vec<ddr_p> *)): New.
25178 (extern debug (data_reference &)): New.
25179 (extern debug (data_reference *)): New.
25180 (extern debug (vec<data_reference_p> &)): New.
25181 (extern debug (vec<data_reference_p> *)): New.
25182 (extern debug (vec<ddr_p> &)): New.
25183 (extern debug (vec<ddr_p> *)): New.
25185 (debug (pt_solution &)): New.
25186 (debug (pt_solution *)): New.
25188 (extern debug (pt_solution &)): New.
25189 (extern debug (pt_solution *)): New.
25191 (debug (_var_map &)): New.
25192 (debug (_var_map *)): New.
25193 (debug (tree_live_info_d &)): New.
25194 (debug (tree_live_info_d *)): New.
25196 (extern debug (_var_map &)): New.
25197 (extern debug (_var_map *)): New.
25198 (extern debug (tree_live_info_d &)): New.
25199 (extern debug (tree_live_info_d *)): New.
25201 2013-03-28 Jan Hubicka <jh@suse.cz>
25203 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
25205 2013-03-28 Ian Bolton <ian.bolton@arm.com>
25207 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
25208 record only when desired or required.
25210 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
25212 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
25213 *vec_extractv2di_1_rex64. Use x64 isa attribute.
25215 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25217 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
25218 (*andsi3_compare0_uxtw): New pattern.
25219 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
25220 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
25222 2013-03-28 Jan Hubicka <jh@suse.cz>
25224 * data-streamer-in.c (streamer_read_gcov_count): New function.
25225 * gimple-streamer-out.c: Include value-prof.h.
25226 (output_gimple_stmt): Output histogram.
25227 (output_bb): Use streamer_write_gcov_count.
25228 * value-prof.c: Include data-streamer.h
25229 (dump_histogram_value): Add HIST_TYPE_MAX.
25230 (stream_out_histogram_value): New function.
25231 (stream_in_histogram_value): New function.
25232 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
25233 (stream_out_histogram_value, stream_in_histogram_value): Declare.
25234 * data-streamer-out.c (streamer_write_gcov_count): New function.
25235 (streamer_write_gcov_count_stream): New function.
25236 * lto-cgraph.c (lto_output_edge): Update counter streaming.
25237 (lto_output_node): Likewise.
25238 (input_node, input_edge): Likewise.
25239 * lto-streamer-out.c (output_cfg): Update streaming.
25240 * lto-streamer-in.c (input_cfg): Likewise.
25241 * data-streamer.h (streamer_write_gcov_count,
25242 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
25243 * gimple-streamer-in.c: Include value-prof.h
25244 (input_gimple_stmt): Input histograms.
25245 (input_bb): Update profile streaming.
25247 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25249 * genmodes.c (emit_max_int): New function.
25250 (emit_insn_modes_h): Added call to emit_max_function.
25251 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
25253 * machmode.def: Fixed comment.
25255 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
25257 * combine.c (try_combine): Removed useless assert.
25258 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
25260 2013-03-28 Marek Polacek <polacek@redhat.com>
25261 Richard Biener <rguenther@suse.de>
25263 PR tree-optimization/56695
25264 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
25265 build signed result of a vector comparison.
25266 * tree-cfg.c (verify_gimple_comparison): Check that a result
25267 of a vector comparison has signed type.
25269 2013-03-28 Richard Biener <rguenther@suse.de>
25271 PR tree-optimization/37021
25272 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
25273 do not restrict gaps between groups.
25274 * tree-vect-stmts.c (vectorizable_load): Properly account for
25275 a gap between groups.
25277 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
25279 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
25280 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
25283 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
25285 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
25286 * doc/extend.texi (Named Address Spaces): Ditto.
25287 (Variable Attributes): Ditto.
25289 2013-03-27 Kai Tietz <ktietz@redhat.com>
25291 * config.build: Add support for cygwin x64 target.
25292 * config.gcc: Likewise.
25293 * config.host: Likewise.
25294 * configure.ac: Likewise
25295 * configure: Regenerated.
25297 2013-03-27 Kai Tietz <ktietz@redhat.com>
25299 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
25300 * config/i386/t-cygwin-w64: New file.
25301 * config/i386/cygwin-w64.h: New file.
25302 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
25303 and add support for x64-cygwin target.
25304 (CPP_SPEC): Likewise.
25305 (CXX_WRAP_SPEC_LIST): Undefine before define.
25306 (LIBGCJ_SONAME): Use 15 as version.
25308 2013-03-27 Richard Biener <rguenther@suse.de>
25310 PR tree-optimization/56716
25311 * tree-ssa-structalias.c (perform_var_substitution): Adjust
25312 dumping for ref nodes.
25314 2013-03-27 Martin Jambor <mjambor@suse.cz>
25316 PR tree-optimization/55334
25317 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
25318 restricted pointers to arrays.
25320 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
25322 * Makefile.in (.SUFFIXES): Add .cc.
25323 (.c.o): Apply same recipe for implicit rule .cc.o.
25325 2013-03-27 Richard Biener <rguenther@suse.de>
25327 PR tree-optimization/37021
25328 * tree-vect-data-refs.c (vect_check_strided_load): Allow
25329 REALPART/IMAGPART_EXPRs around the supported refs.
25330 * tree-ssa-structalias.c (find_func_aliases): Assume that
25331 floating-point values are not used to transfer pointers.
25333 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
25335 * target.def (TARGET_HAS_IFUNC_P): New target hook.
25336 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
25337 * doc/tm.texi: Regenerate.
25338 * targhooks.h (default_has_ifunc_p): New.
25339 * targhooks.c (default_has_ifunc_p): Ditto.
25340 * config/linux-protos.h: New file.
25341 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
25342 hook for linux which disables support of indirect functions in android.
25343 * config/linux-android.c: New file.
25344 * config/t-linux-android.c: Ditto.
25345 * config.gcc: Added new object file linux-android.o.
25346 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
25347 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
25348 * varasm.c (do_assemble_alias): Likewise.
25349 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
25350 doesn't support indirect functions.
25351 * configure: Regenerate.
25353 2013-03-27 Bin Cheng <bin.cheng@arm.com>
25356 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
25357 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
25360 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25362 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
25364 2013-03-27 Terry Guo <terry.guo@arm.com>
25366 * config/arm/arm-cores.def: Added core cortex-r7.
25367 * config/arm/arm-tune.md: Regenerated.
25368 * config/arm/arm-tables.opt: Regenerated.
25369 * doc/invoke.texi: Added entry for core cortex-r7.
25371 2013-03-27 Walter Lee <walt@tilera.com>
25373 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
25374 double-decrement of next_scratch_regno.
25376 2013-03-27 Walter Lee <walt@tilera.com>
25378 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
25380 (insn_v1mulus): Ditto.
25381 (insn_v2muls): Ditto.
25383 2013-03-27 Walter Lee <walt@tilera.com>
25385 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
25386 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
25388 2013-03-27 Walter Lee <walt@tilera.com>
25390 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
25391 (*sibcall_value): Ditto.
25393 2013-03-27 Walter Lee <walt@tilera.com>
25395 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
25396 (insn_mnz_v8qi): ... this ...
25397 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
25399 (insn_v<n>mnz): Replaced by ...
25400 (insn_v1mnz): ... this ...
25401 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
25403 (insn_mz_<mode>): Replaced by ...
25404 (insn_mz_v8qi): ... this ...
25405 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
25407 (insn_v<n>mz): Replaced by ...
25408 (insn_v1mz): ... this ...
25409 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
25412 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
25414 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
25416 2013-03-26 Roland McGrath <mcgrathr@google.com>
25418 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
25419 than fprintf with a non-constant, non-format string.
25421 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25423 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
25424 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
25425 operand 0 predicate.
25426 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
25427 attribute. Use general_x64nomem_operand as operand 1 predicate.
25428 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
25429 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25430 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
25431 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
25432 (mov<mode>_insv_1): Remove expander. Merge insn with
25433 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
25434 Use general_x64nomem_operand as operand 1 predicate.
25435 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
25436 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
25437 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25438 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
25439 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
25440 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
25441 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
25442 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
25443 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
25444 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
25445 (general_x64nomem_operand): Ditto.
25447 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
25449 * config/rtems.opt: Add -pthread option.
25451 2013-03-26 Richard Biener <rguenther@suse.de>
25453 * alias.c (find_base_term): Avoid redundant and not used recursion.
25454 (base_alias_check): Get the initial base term from the caller.
25455 (true_dependence_1): Compute and pass base terms to base_alias_check.
25456 (write_dependence_p): Likewise.
25457 (may_alias_p): Likewise.
25459 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
25461 * config/aarch64/aarch64.c (aarch64_classify_address): Support
25462 PC-relative load in SI modes and above only.
25464 2013-03-26 Xinyu Qi <xyqi@marvell.com>
25466 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
25467 * config/arm/iwmmxt.md (WCGR0): Update.
25468 (WCGR1, WCGR2, WCGR3): Likewise.
25470 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
25472 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
25473 Use x64 and nox64 isa attributes.
25475 2013-03-26 Richard Biener <rguenther@suse.de>
25477 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
25478 alignment computations and rely on get_object_alignment_1
25479 for the !TYPE_P case.
25480 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
25482 2013-03-26 Walter Lee <walt@tilera.com>
25484 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
25485 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
25487 2013-03-25 Jeff Law <law@redhat.com>
25489 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
25490 check for INTEGRAL_TYPE_P that was missing due to checking in
25491 wrong version of prior patch.
25493 2013-03-25 Walter Lee <walt@tilera.com>
25495 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
25496 TILEGX_INSN_SHUFFLEBYTES1.
25497 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
25499 (tilegx_builtins): Ditto.
25500 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
25502 2013-03-25 Walter Lee <walt@tilera.com>
25504 * config/tilegx/tilegx.md (floatsisf2): New pattern.
25505 (floatunssisf2): New pattern.
25506 (floatsidf2): New pattern.
25507 (floatunssidf2): New pattern.
25509 2013-03-25 Walter Lee <walt@tilera.com>
25511 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
25512 tests for constraint J, K, N, P.
25514 2013-03-25 Walter Lee <walt@tilera.com>
25516 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
25517 Use indirect/pcrel encoding.
25518 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
25521 2013-03-25 Steve Ellcey <sellcey@mips.com>
25523 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
25524 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
25525 * config/mips/mips.c (mips_option_override): Set IMADD default.
25526 * config/mips/mips.h (PTF_AVOID_IMADD): New.
25527 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
25528 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
25529 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
25530 * doc/invoke.texi (-mimadd/-mno-imadd): New.
25532 2013-03-25 Jeff Law <law@redhat.com>
25534 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
25535 slightly to avoid creating and folding useless trees. Simplify
25536 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
25538 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
25540 * config/i386/i386.md (*zero_extendsidi2): Merge with
25541 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
25542 * config/i386/predicates.md (x86_64_zext_operand): Rename from
25543 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
25544 targets. Clarify comment.
25546 2013-03-25 Martin Jambor <mjambor@suse.cz>
25548 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
25549 pass-through jump functions differently.
25550 (ipa_read_jump_function): Likewise. Also use setter functions to set
25553 2013-03-25 Martin Jambor <mjambor@suse.cz>
25555 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
25556 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
25558 (ipa_get_indirect_edge_target): New function.
25559 (devirtualization_time_bonus): New parameter known_aggs, pass it to
25560 ipa_get_indirect_edge_target. Update all callers.
25561 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
25562 ipa_get_indirect_edge_target_1 instead of calling
25563 ipa_get_indirect_edge_target.
25564 (create_specialized_node): Pass aggvlas to
25565 ipcp_discover_new_direct_edges.
25567 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25569 * config/arm/arm.md (f_sels, f_seld): New types.
25570 (*cmov<mode>): New pattern.
25571 * config/arm/predicates.md (arm_vsel_comparison_operator): New
25574 2013-03-25 Kai Tietz <ktietz@redhat.com>
25576 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
25577 POSIX-printf for mingw-hosted builds.
25579 2013-03-25 Richard Biener <rguenther@suse.de>
25581 PR middle-end/56694
25582 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
25583 must-not-throw stmt location.
25585 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25587 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
25588 Emit load-acquire versions when acq is true.
25589 (arm_emit_store_exclusive): Add rel parameter.
25590 Emit store-release versions when rel is true.
25591 (arm_split_compare_and_swap): Use acquire-release instructions
25593 of barriers when appropriate.
25594 (arm_split_atomic_op): Likewise.
25595 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
25596 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
25597 (VUNSPEC_SLX): Likewise.
25598 (VUNSPEC_LDA): Likewise.
25599 (VUNSPEC_STL): Likewise.
25600 * config/arm/sync.md (atomic_load<mode>): New pattern.
25601 (atomic_store<mode>): Likewise.
25602 (arm_load_acquire_exclusive<mode>): Likewise.
25603 (arm_load_acquire_exclusivesi): Likewise.
25604 (arm_load_acquire_exclusivedi): Likewise.
25605 (arm_store_release_exclusive<mode>): Likewise.
25607 2013-03-25 Catherine Moore <clm@codesourcery.com>
25609 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
25610 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
25611 * config/mip/predicates.md (lwsp_swsp_operand,
25612 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
25613 sb16_operand, db4_operand, db7_operand, ib3_operand,
25614 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
25615 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
25616 andi16_operand): New predicates.
25617 * config/mips/mips.md (compression): New attribute.
25618 (enabled): New attribute.
25619 (length): Consider compression in computing length.
25620 (shift_compression): New code attribute.
25621 (*add<mode>3): New operands. Record compression.
25622 (sub<mode>3): Likewise.
25623 (one_cmpl<mode>2): Likewise.
25624 (*and<mode>3): Likewise.
25625 (*ior<mode>3): Likewise.
25626 (unnamed pattern for xor): Likewise.
25627 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
25628 (*<optab><mode>3): Likewise.
25629 (*mov<mode>_internal: Likewise.
25630 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
25631 (mips_unsigned_immediate_p): New.
25632 (umips_lwsp_swsp_address_p): New.
25633 (m16_based_address_p): New.
25634 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
25635 (mips_unsigned_immediate_p): New prototype.
25636 (lwsp_swsp_address_p): New prototype.
25637 (m16_based_address_p): New prototype.
25638 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
25639 (mips_signed_immediate_p): New function.
25640 (m16_based_address_p): New function.
25641 (lwsp_swsp_address_p): New function.
25642 (mips_print_operand_punctuation): Recognize short delay slot insns
25643 for microMIPS.add<mode>3"
25645 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
25648 * config/arm/iterators.md (v_cmp_result): New mode attribute.
25649 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
25651 2013-03-25 Richard Biener <rguenther@suse.de>
25653 PR tree-optimization/56689
25654 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
25657 2013-03-25 Richard Biener <rguenther@suse.de>
25659 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
25661 (memory_references): Likewise.
25662 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
25663 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
25664 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
25665 (gather_mem_refs_in_loops): Fold into ...
25666 (analyze_memory_references): ... this. Move initialization
25667 to tree_ssa_lim_initialize.
25668 (fill_always_executed_in): Rename to ...
25669 (fill_always_executed_in_1): ... this.
25670 (fill_always_executed_in): Move contains_call computation to
25671 this new function from ...
25672 (tree_ssa_lim_initialize): ... here.
25673 (tree_ssa_lim): Call fill_always_executed_in.
25675 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25677 * postreload.c (reload_combine): Fix code detecting returns.
25679 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
25681 * function.c (emit_use_return_register_into_block): On cc0 targets,
25682 do not emit the sequence between cc0 setter and user.
25684 2013-03-25 Kai Tietz <ktietz@redhat.com>
25686 * config/i386/predicates.md (local_symbolic_operand): Interpret
25687 dll-imported symbols as none-local.
25689 2013-03-25 Richard Biener <rguenther@suse.de>
25691 * tree-ssa-loop-im.c (struct depend): Remove.
25692 (struct lim_aux_data): Make depends a vec of gimples.
25693 (free_lim_aux_data): Adjust.
25694 (add_dependency): Likewise.
25695 (set_level): Likewise.
25697 2013-03-25 Richard Biener <rguenther@suse.de>
25699 PR middle-end/56434
25700 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
25701 the pointer returned by calls with ECF_MALLOC set.
25703 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25705 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
25707 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
25709 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
25710 using MMXMODE mode iterator.
25711 (*move<mode>_internal): Merge with *movv2sf_internal and
25712 *movv2sf_internal_rex64 using MMXMODE mode iterator.
25714 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25716 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
25717 (record_last_mem_set_info): Likewise.
25719 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
25720 of XNEWVEC followed by memset.
25721 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
25723 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
25725 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
25726 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
25727 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
25728 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
25729 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
25730 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
25731 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
25732 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
25733 BARRIER_P instead of GET_CODE.
25735 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25737 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
25738 inaccuracy in the probing code.
25740 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
25741 (ctrapdi4): Likewise.
25743 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25745 * calls.c (expand_call): Add missing guard to code handling return
25746 of non-BLKmode structures in MSB.
25747 * function.c (expand_function_end): Likewise.
25749 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
25751 * combine.c (try_combine): Adjust comment. Do not add the set of
25752 insn #0 if the destination indirectly is set or dies in insn #2.
25753 Tidy up code to distribute a new note.
25755 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25757 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
25758 also for alternatives 16 and 17.
25760 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
25762 * config/i386/sse.md (*mov<mode>_internal): Merge with
25763 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
25764 Emit insn template depending on type attribute. Use
25765 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
25766 movd instead of movq mnemonic for interunit moves. Rewrite mode
25767 attribute calculation. Remove unit attribute calculation.
25768 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
25769 Set prefix_data16 attribute for DImode ssemov types.
25770 Use Ym instead of y for SSE-MMX conversion alternatives.
25771 Reorder operand constraints.
25773 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
25775 * df.h (df_insn_delete): Adjust prototype.
25776 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
25777 and let it decide whether mark the basic block dirty.
25778 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
25779 * df-scan.c (df_insn_info_delete): New helper function, split
25780 off from df_insn_delete.
25781 (df_scan_free_bb_info): Use it.
25782 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
25784 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
25785 that the insn is actually an insn and it has a non-NULL basic block.
25786 Do not mark basic block dirty if only deleting a DEBUG_INSN.
25788 2013-03-22 Richard Biener <rguenther@suse.de>
25790 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
25792 (mem_ref_alloc): Do not allocate them.
25793 (refs_independent_p): Do not query or maintain a cache.
25795 2013-03-22 Richard Biener <rguenther@suse.de>
25797 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
25798 (gather_mem_refs_in_loops): Do not compute it.
25799 (analyze_memory_references): Do not allocate it.
25800 (tree_ssa_lim_finalize): Do not free it.
25801 (for_all_locs_in_loop): Do not query all_refs_in_loop.
25803 2013-03-22 Richard Biener <rguenther@suse.de>
25805 * is-a.h (as_a): Use gcc_checking_assert.
25807 2013-03-22 Ian Bolton <ian.bolton@arm.com>
25809 * config/aarch64/aarch64.c (aarch64_print_operand): New
25810 format specifier for printing a constant in hex.
25811 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
25812 format specifier for printing second operand.
25814 2013-03-22 Richard Biener <rguenther@suse.de>
25816 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
25818 (gather_mem_refs_in_loops): Perform store accumulation here.
25819 (create_vop_ref_mapping_loop): Remove.
25820 (create_vop_ref_mapping): Likewise.
25821 (analyze_memory_references): Initialize refs_stored_in_loop.
25822 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
25823 (record_indep_loop): Remove.
25824 (record_dep_loop): New function.
25825 (ref_indep_loop_p_1): Adjust to only walk over references
25826 in the loop, not its subloops.
25827 (ref_indep_loop_p): Rename to ...
25828 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
25829 maintaining a more fine-grained cache.
25830 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
25831 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
25833 2013-03-22 Richard Biener <rguenther@suse.de>
25835 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
25836 (struct mem_ref): Make accesses_in_loop a vec of a vec of
25837 aggregate mem_ref_loc.
25838 (free_mem_ref_locs): Inline into ...
25839 (memref_free): ... this and adjust.
25840 (mem_ref_alloc): Adjust.
25841 (mem_ref_locs_alloc): Remove.
25842 (record_mem_ref_loc): Adjust.
25843 (get_all_locs_in_loop): Rewrite into ...
25844 (for_all_locs_in_loop): ... this iterator.
25845 (rewrite_mem_ref_loc): New functor.
25846 (rewrite_mem_refs): Use for_all_locs_in_loop.
25847 (sm_set_flag_if_changed): New functor.
25848 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
25849 (ref_always_accessed): New functor.
25850 (ref_always_accessed_p): Use for_all_locs_in_loop.
25852 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25854 * tree-pass.h (PROP_gimple_lvec): New.
25855 * passes.c (dump_properties): Handle PROP_gimple_lvec.
25856 (init_optimization_passes): Move pass_lower_vector.
25857 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
25859 (pass_lower_vector): Provide PROP_gimple_lvec.
25860 (pass_lower_vector_ssa): Likewise.
25861 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
25863 2013-03-21 Mark Wielaard <mjw@redhat.com>
25865 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
25867 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
25869 * config/i386/i386.md (*movdi_internal): Disparage slightly
25870 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
25871 conversion alternatives.
25873 2013-03-21 Jakub Jelinek <jakub@redhat.com>
25875 PR middle-end/48087
25876 * diagnostic.def (DK_WERROR): New kind.
25877 * diagnostic.h (werrorcount): Define.
25878 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
25879 promoted to DK_ERROR, increment DK_WERROR counter instead of
25881 * toplev.c (toplev_main): Call print_ignored_options even if
25882 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
25883 even if just werrorcount is non-zero.
25886 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
25888 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
25889 (string_cst_pool_decl): New function.
25890 (optimize_one_addr_into_implicit_ptr): New function.
25891 (resolve_addr_in_expr): Optimize DWARF location expression
25892 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
25893 which doesn't live in memory, but has DW_AT_location or
25894 DW_AT_const_value, or refers to a string literal, into
25895 DW_OP_GNU_implicit_pointer.
25896 (optimize_location_into_implicit_ptr): New function.
25897 (resolve_addr): If removing DW_AT_location of a variable because
25898 it was DW_OP_addr of address of the variable, but the variable doesn't
25899 live in memory, try to emit const value attribute for the initializer.
25901 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25903 * tree.h (VECTOR_TYPE_P): New macro.
25904 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
25905 TYPE_MODE): Use it.
25906 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
25907 VEC_COND_EXPR cannot be lvalues.
25908 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
25910 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
25912 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
25913 Restrict the transformation to equal modes.
25915 2013-03-21 Richard Biener <rguenther@suse.de>
25917 PR tree-optimization/39326
25918 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
25919 (MEM_ANALYZABLE): Adjust.
25920 (record_mem_ref_loc): Move bitmap ops ...
25921 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
25922 unanalyzable refs, do not record locations for it.
25923 (analyze_memory_references): Allocate ref zero as shared
25925 (refs_independent_p): Do not test for unanalyzed mems here.
25926 (ref_indep_loop_p_1): Special-case disambiguation against
25927 the unanalyzed ref.
25928 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
25930 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
25932 * config/arm/arm-protos.h (tune_params): Add
25933 prefer_neon_for_64bits field.
25934 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
25935 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
25936 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
25937 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
25938 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
25939 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
25940 (arm_option_override): Handle -mneon-for-64bits new option.
25941 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
25942 (prefer_neon_for_64bits): Declare new variable.
25943 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
25944 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
25945 (arch_enabled): Handle new arch types. Remove support for onlya8
25947 (one_cmpldi2): Use new arch names.
25948 (zero_extend<mode>di2, extend<mode>di2): Ditto.
25949 * config/arm/arm.opt (mneon-for-64bits): Add option.
25950 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
25951 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
25952 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
25954 * doc/invoke.texi (-mneon-for-64bits): Document.
25956 2013-03-21 Richard Biener <rguenther@suse.de>
25958 PR tree-optimization/39326
25959 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
25960 (sort_bbs_in_loop_postorder_cmp): New function.
25961 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
25963 2013-03-21 Richard Biener <rguenther@suse.de>
25965 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
25966 (vect_insert_into_interleaving_chain): Likewise.
25967 (vect_drs_dependent_in_basic_block): Inline ...
25968 (vect_slp_analyze_data_ref_dependence): ... here. New function,
25970 (vect_analyze_data_ref_dependence): ... here. Simplify.
25971 (vect_check_interleaving): Simplify.
25972 (vect_analyze_data_ref_dependences): Likewise. Split out ...
25973 (vect_slp_analyze_data_ref_dependences): ... this new function.
25974 (dr_group_sort_cmp): New function.
25975 (vect_analyze_data_ref_accesses): Compute data-reference groups
25976 here instead of in vect_analyze_data_ref_dependence. Use
25977 a more efficient algorithm.
25978 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
25979 vect_slp_analyze_data_ref_dependences. Call
25980 vect_analyze_data_ref_accesses earlier.
25981 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
25982 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
25983 (vect_slp_analyze_data_ref_dependences): New prototype.
25985 2013-03-21 Richard Biener <rguenther@suse.de>
25987 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
25988 ref is stored in the loop.
25989 (find_refs_for_sm): Walk only over all stores.
25990 (store_motion_loop): Allocate from lim_bitmap_obstack.
25991 (store_motion): Likewise.
25993 2013-03-21 Richard Biener <rguenther@suse.de>
25995 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
25996 Update virtual SSA form.
25998 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
26000 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
26001 * configure: Regenerate.
26002 * config.in: Regenerate.
26003 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
26004 if !HAVE_LD_EH_FRAME_CIEV3.
26006 2013-03-21 Richard Biener <rguenther@suse.de>
26008 * tree-cfg.c (verify_expr_no_block): New function.
26009 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
26010 nor DECL_VALUE_EXPR have locations with associated blocks.
26011 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
26012 (clear_unused_block_pointer): Remove code dealing with
26013 blocks in DECL_DEBUG_EXPR locations.
26015 2013-03-21 Richard Biener <rguenther@suse.de>
26017 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
26018 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
26019 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
26020 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
26021 instead of DECL_DEBUG_EXPR_IS_FROM.
26022 * gimplify.c (gimplify_modify_expr): Likewise.
26023 * tree-cfg.c (verify_expr_location_1): Likewise.
26024 * tree-complex.c (create_one_component_var): Likewise.
26025 * tree-sra.c (create_access_replacement): Likewise.
26026 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
26027 (clear_unused_block_pointer): Likewise.
26028 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
26029 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
26030 * var-tracking.c (var_debug_decl): Likewise.
26031 (track_expr_p): Likewise.
26032 * tree-inline.c (add_local_variables): Likewise. Set
26033 DECL_HAS_DEBUG_EXPR_P after copying it.
26034 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
26035 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
26037 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
26040 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
26041 * configure: Regenerate.
26042 * config.in: Regenerate.
26043 * config/i386/i386.md (*movdf_internal): Use
26044 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
26045 movd instead of movq mnemonic for interunit moves.
26046 (*movdi_internal): Ditto.
26048 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26050 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
26051 (abd<mode>_3): New pattern.
26052 (aba<mode>_3): New pattern.
26053 (fabd<mode>_3): New pattern.
26055 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
26057 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
26058 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
26059 occurrence of REGISTER_PREFIX as its empty string.
26061 2013-03-20 Jeff Law <law@redhat.com>
26063 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
26064 addititional equivalences for equality comparisons between an SSA_NAME
26065 and a constant where the SSA_NAME was set from a widening conversion.
26067 2013-03-20 Walter Lee <walt@tilera.com>
26069 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
26071 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26073 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
26074 depending on type attribute.
26075 (*movti_internal): Ditto.
26076 (*movtf_internal): Ditto.
26077 (*movxf_internal): Ditto.
26078 (*movdf_internal): Ditto.
26079 (*movsf_internal): Ditto.
26081 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26083 * config/i386/i386.md (*movti_internal): Set prefix attribute to
26084 maybe_vex for sselog1 and ssemov types.
26085 (*movdi_internal): Reorder operand constraints.
26086 (*movsi_internal): Ditto. Set prefix attribute to
26087 maybe_vex for sselog1 and ssemov types.
26088 (*movtf_internal): Set prefix attribute to maybe_vex
26089 for sselog1 and ssemov types.
26090 (*movdf_internal): Ditto. Set prefix_data16 attribute for
26091 DImode ssemov types. Reorder operand constraints.
26092 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
26093 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
26094 attribute for SImode ssemov types. Reorder operand constraints.
26096 2013-03-20 Martin Jambor <mjambor@suse.cz>
26098 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
26099 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
26101 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
26103 * config/rs6000/predicates.md (indexed_address, update_address_mem
26104 update_indexed_address_mem): New predicates.
26105 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
26106 attribute for load/store instructions.
26107 * config/rs6000/dfp.md (movsd_store): Likewise.
26108 (movsd_load): Likewise.
26109 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
26110 (unnamed HI->DI extend define_insn): Likewise.
26111 (unnamed SI->DI extend define_insn): Likewise.
26112 (unnamed QI->SI extend define_insn): Likewise.
26113 (unnamed QI->HI extend define_insn): Likewise.
26114 (unnamed HI->SI extend define_insn): Likewise.
26115 (unnamed HI->SI extend define_insn): Likewise.
26116 (extendsfdf2_fpr): Likewise.
26117 (movsi_internal1): Likewise.
26118 (movsi_internal1_single): Likewise.
26119 (movhi_internal): Likewise.
26120 (movqi_internal): Likewise.
26121 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
26122 attribute for load/store instructions.
26123 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
26125 (mov<mode>_softfloat): Likewise.
26126 (mov<mode>_hardfloat32): Likewise.
26127 (mov<mode>_hardfloat64): Likewise.
26128 (mov<mode>_softfloat64): Likewise.
26129 (movdi_internal32): Likewise.
26130 (movdi_internal64): Likewise.
26131 (probe_stack_<mode>): Likewise.
26133 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
26135 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
26136 floating point, and decimal floating point to reload iterator.
26138 * config/rs6000/constraints.md (wl constraint): New constraints to
26139 return FLOAT_REGS if certain options are used to reduce the number
26140 of separate patterns that exist in the file.
26141 (wx constraint): Likewise.
26142 (wz constraint): Likewise.
26144 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26145 -mdebug=reg, print wg, wl, wx, and wz constraints.
26146 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26147 Initialize the reload functions for 64-bit binary/decimal floating
26149 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26150 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26151 create the buffer on the stack to overcome not having a 32-bit
26153 (rs6000_emit_move): Likewise.
26154 (rs6000_secondary_memory_needed_rtx): Likewise.
26155 (rs6000_alloc_sdmode_stack_slot): Likewise.
26156 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26157 via xxlxor, just like DFmode 0.0.
26159 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
26160 define as 1 if we are running on a power7 or newer.
26161 (enum r6000_reg_class_enum): Add new constraints.
26163 * config/rs6000/dfp.md (movsd): Delete, combine with binary
26164 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26165 with other moves by using conditional constraits (wg). Use LFIWZX
26166 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
26167 (movsd splitter): Likewise.
26168 (movsd_hardfloat): Likewise.
26169 (movsd_softfloat): Likewise.
26171 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
26172 binary and decimal floating point moves.
26173 (fmove_ok): New attributes to combine binary and decimal floating
26174 point moves, and to combine power6x (mfpgpr) moves along normal
26176 (real_value_to_target): Likewise.
26177 (f32_lr): Likewise.
26178 (f32_lm): Likewise.
26179 (f32_li): Likewise.
26180 (f32_sr): Likewise.
26181 (f32_sm): Likewise.
26182 (f32_si): Likewise.
26183 (movsf): Combine binary and decimal floating point moves. Combine
26184 power6x (mfpgpr) moves with other moves by using conditional
26185 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
26186 (mov<mode> for SFmode/SDmode); Likewise.
26187 (SFmode/SDmode splitters): Likewise.
26188 (movsf_hardfloat): Likewise.
26189 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
26190 (movsf_softfloat): Likewise.
26191 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
26193 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
26194 wx and wz constraints.
26196 * config/rs6000/constraints.md (wg constraint): New constraint to
26197 return FLOAT_REGS if -mmfpgpr (power6x) was used.
26199 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
26202 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
26203 -mdebug=reg, print wg, wl, wx, and wz constraints.
26204 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
26205 Initialize the reload functions for 64-bit binary/decimal floating
26207 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
26208 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
26209 create the buffer on the stack to overcome not having a 32-bit
26211 (rs6000_emit_move): Likewise.
26212 (rs6000_secondary_memory_needed_rtx): Likewise.
26213 (rs6000_alloc_sdmode_stack_slot): Likewise.
26214 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
26215 via xxlxor, just like DFmode 0.0.
26217 * config/rs6000/dfp.md (movdd): Delete, combine with binary
26218 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
26219 with other moves by using conditional constraits (wg). Use LFIWZX
26220 and STFIWX for loading SDmode on power7.
26221 (movdd splitters): Likewise.
26222 (movdd_hardfloat32): Likewise.
26223 (movdd_softfloat32): Likewise.
26224 (movdd_hardfloat64_mfpgpr): Likewise.
26225 (movdd_hardfloat64): Likewise.
26226 (movdd_softfloat64): Likewise.
26228 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
26229 64-bit binary and decimal floating point moves.
26230 (FMOVE64X): Likewise.
26231 (movdf): Combine 64-bit binary and decimal floating point moves.
26232 Combine power6x (mfpgpr) moves with other moves by using
26233 conditional constraits (wg).
26234 (mov<mode> for DFmode/DDmode): Likewise.
26235 (DFmode/DDmode splitters): Likewise.
26236 (movdf_hardfloat32): Likewise.
26237 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
26238 (movdf_softfloat32): Likewise.
26239 (movdf_hardfloat64_mfpgpr): Likewise.
26240 (movdf_hardfloat64): Likewise.
26241 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
26242 (movdf_softfloat64): Likewise.
26243 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
26244 (reload_<mode>_load): Move to later in the file so they aren't in
26245 the middle of the floating point move insns.
26246 (reload_<mode>_store): Likewise.
26248 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
26251 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
26252 constraint if -mdebug=reg.
26253 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
26254 Enable using dd reload support if needed.
26256 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
26257 binary and decimal floating point moves in rs6000.md.
26258 (movtd_internal): Likewise.
26260 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
26261 decimal floating point moves.
26263 (movtf_internal): Likewise.
26264 (mov<mode>_internal, TDmode/TFmode): Likewise.
26265 (movtf_softfloat): Likewise.
26266 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
26268 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
26269 movdi_internal64, using wg constraint for move direct operations.
26270 (movdi_internal64): Likewise.
26272 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
26273 MODES_TIEABLE_P for selected modes. Print the numerical value of
26274 the various virtual registers. Use GPR/FPR first/last values,
26275 instead of hard coding the register numbers. Print which modes
26276 have reload functions registered.
26277 (rs6000_option_override_internal): If -mdebug=reg, trace the options
26278 settings before/after setting cpu, target and subtarget settings.
26279 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
26280 and for secondary reload failures in rs6000_secondary_reload_inner.
26281 (rs6000_secondary_reload_fail): Likewise.
26282 (rs6000_secondary_reload_inner): Likewise.
26284 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
26285 macros for first/last GPR and FPR registers.
26286 (LAST_GPR_REGNO): Likewise.
26287 (FIRST_FPR_REGNO): Likewise.
26288 (LAST_FPR_REGNO): Likewise.
26290 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
26291 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
26292 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
26293 (vcond<mode><mode>): Likewise.
26294 (vcondu<mode><mode>): Likewise.
26295 (vector_gtu<mode>): Likewise.
26296 (vector_gte<mode>): Likewise.
26297 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
26298 to prevent the compiler from converting DImode operations to TImode.
26299 (ior<mode>3): Likewise.
26300 (and<mode>3): Likewise.
26301 (one_cmpl<mode>2): Likewise.
26302 (nor<mode>3): Likewise.
26303 (andc<mode>3): Likewise.
26305 * config/rs6000/constraints.md (wt constraint): New constraint
26306 that returns VSX_REGS if TImode is allowed in VSX registers.
26308 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
26309 constant under VSX.
26311 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
26312 similar to TImode, but it is restricted to being in the GPRs.
26314 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
26315 TImode to occupy a single VSX register.
26317 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
26318 -mvsx-timode for power7/power8.
26319 (power7 cpu): Likewise.
26320 (power8 cpu): Likewise.
26322 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
26323 sure that TFmode/TDmode take up two registers if they are ever
26324 allowed in the upper VSX registers.
26325 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
26327 (rs6000_init_hard_regno_mode_ok): Likewise.
26328 (rs6000_debug_reg_global): Add debugging for PTImode and wt
26329 constraint. Print if LRA is turned on.
26330 (rs6000_option_override_internal): Give an error if -mvsx-timode
26331 and VSX is not enabled.
26332 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
26333 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
26334 to reg+offset addressing. Use PTImode when checking offset
26335 addresses for validity.
26336 (reg_offset_addressing_ok_p): Likewise.
26337 (rs6000_legitimate_offset_address_p): Likewise.
26338 (rs6000_legitimize_address): Likewise.
26339 (rs6000_legitimize_reload_address): Likewise.
26340 (rs6000_legitimate_address_p): Likewise.
26341 (rs6000_eliminate_indexed_memrefs): Likewise.
26342 (rs6000_emit_move): Likewise.
26343 (rs6000_secondary_reload): Likewise.
26344 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
26345 reloads to fpr registers to continue to use reg+offset addressing,
26346 but 64-bit reloads to altivec registers need reg+reg addressing.
26347 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
26348 it. Treat LO_SUM like a PLUS operation.
26349 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
26350 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
26351 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
26352 registers to share a register with a smaller sized type, since VSX
26353 puts scalars in the upper 64-bits.
26354 (print_operand): Add support for PTImode.
26355 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
26356 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
26357 registers, but don't have arithmetic support.
26358 (rs6000_memory_move_cost): Add test for VSX.
26359 (rs6000_opt_masks): Add -mvsx-timode.
26361 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
26364 (VSr): Use wt constraint for TImode.
26365 (VSv): Drop TImode support.
26366 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
26367 (vsx_movti_64bit): Likewise.
26368 (vsx_movti_32bit): Likewise.
26369 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
26370 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
26371 one '?' on the appropriate output constraint. Do not allow TImode
26372 logical operations on 32-bit systems.
26373 (vsx_ior<mode>3): Likewise.
26374 (vsx_xor<mode>3): Likewise.
26375 (vsx_one_cmpl<mode>2): Likewise.
26376 (vsx_nor<mode>3): Likewise.
26377 (vsx_andc<mode>3): Likewise.
26378 (vsx_concat_<mode>): Likewise.
26379 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
26381 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
26382 OPTION_MASK_VSX_TIMODE.
26383 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
26384 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
26386 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
26387 (TI2 iterator): New iterator for TImode, PTImode.
26388 (wd mode attribute): Add values for vector types.
26389 (movti_string): Replace TI move operations with operations for TImode
26390 and PTImode. Add support for TImode being allowed in VSX registers.
26391 (mov<mode>_string, TImode/PTImode): Likewise.
26392 (movti_ppc64): Likewise.
26393 (mov<mode>_ppc64, TImode/PTImode): Likewise.
26394 (TI mode splitters): Likewise.
26396 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
26399 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
26401 PR tree-optimization/56355
26402 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
26403 Also handle integers with undefined overflow.
26405 2013-03-20 Catherine Moore <clm@codesourcery.com>
26406 Maciej W. Rozycki <macro@codesourcery.com>
26407 Tom de Vries <tom@codesourcery.com>
26408 Nathan Sidwell <nathan@codesourcery.com>
26409 Iain Sandoe <iain@codesourcery.com>
26410 Nathan Froyd <froydnj@codesourcery.com>
26411 Chao-ying Fu <fu@mips.com>
26413 * doc/extend.texi: (micromips, nomicromips, nocompression):
26414 Document new function attributes.
26415 * doc/invoke.texi (minterlink-compressed, mmicromips,
26416 m14k, m14ke, m14kec): Document new options.
26417 (minterlink-mips16): Update documentation.
26418 * doc/md.texi (ZC, ZD): Document new constraints.
26419 * configure.ac (gcc_cv_as_micromips): Check if linker
26420 supports the .set micromips directive.
26421 * configure: Regenerate.
26422 * config.in: Regenerate.
26423 * config/mips/mips-tables.opt: Regenerate.
26424 * config/mips/micromips.md: New file.
26425 * constraints.md (ZC, ZD): New constraints.
26426 * config/mips/predicates.md (movep_src_register): New predicate.
26427 (movep_src_operand): New predicate.
26428 (non_volatile_mem_operand): New predicate.
26429 * config/mips/mips.md (multimem): New type.
26430 (length): Differentiate between 17-bit and 18-bit branch offsets.
26431 (MOVEP1, MOVEP2): New mode iterator.
26432 (mov_<load>l): Use ZC constraint.
26433 (mov_<load>r): Likewise.
26434 (mov_<store>l): Likewise.
26435 (mov_<store>r): Likewise.
26436 (*branch_equality<mode>_inverted): Add microMIPS support.
26437 (*branch_equality<mode>): Likewise.
26438 (*jump_absolute): Likewise.
26439 (indirect_jump_<mode>): Likewise.
26440 (tablejump_<mode>): Likewise.
26441 (<optab>_internal): Likewise.
26442 (sibcall_internal): Likewise.
26443 (sibcall_value_internal): Likewise.
26444 (prefetch): Use constraint ZD.
26445 * config/mips/mips.opt (minterlink-compressed): New option.
26446 (minterlink-mips16): Now an alias for minterlink-compressed.
26447 (mmicromips): New option.
26448 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26449 (compare_and_swap_12): Likewise.
26450 (sync_add<mode>): Likewise.
26451 (sync_<optab>_12): Likewise.
26452 (sync_old_<optab>_12): Likewise.
26453 (sync_new_<optab>_12): Likewise.
26454 (sync_nand_12): Likewise.
26455 (sync_old_nand_12): Likewise.
26456 (sync_new_nand_12): Likewise.
26457 (sync_sub<mode>): Likewise.
26458 (sync_old_add<mode>): Likewise.
26459 (sync_old_sub<mode>): Likewise.
26460 (sync_new_add<mode>): Likewise.
26461 (sync_new_sub<mode>): Likewise.
26462 (sync_<optab><mode>): Likewise.
26463 (sync_old_<optab><mode>): Likewise.
26464 (sync_new_<optab><mode>): Likewise.
26465 (sync_nand<mode>): Likewise.
26466 (sync_old_nand<mode>): Likewise.
26467 (sync_new_nand<mode>): Likewise.
26468 (sync_lock_test_and_set<mode>): Likewise.
26469 (test_and_set_12): Likewise.
26470 (atomic_compare_and_swap<mode>): Likewise.
26471 (atomic_exchange<mode>_llsc): Likewise.
26472 (atomic_fetch_add<mode>_llsc): Likewise.
26473 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26474 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26475 (umips_save_restore_pattern_p): Likewise.
26476 (umips_load_store_pair_p): Likewise.
26477 (umips_output_load_store_pair): Likewise.
26478 (umips_movep_target_p): Likewise.
26479 (umips_12bit_offset_address_p): Likewise.
26480 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26481 (mips_base_mips16): Rename this...
26482 (mips_base_compression_flags): ...to this. Update all uses.
26483 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26484 (mips_mips16_decl_p): Delete.
26485 (mips_nomips16_decl_p): Delete.
26486 (mips_get_compress_on_flags): New function.
26487 (mips_get_compress_off_flags): New function.
26488 (mips_get_compress_mode): New function.
26489 (mips_get_compress_on_name): New function.
26490 (mips_get_compress_off_name): New function.
26491 (mips_insert_attributes): Support multiple compression types.
26492 (mips_merge_decl_attributes): Likewise.
26493 (umips_12bit_offset_address_p): New function.
26494 (mips_start_function_definition): Emit .set micromips directive.
26495 (mips_call_may_need_jalx_p): New function.
26496 (mips_function_ok_for_sibcall): Add microMIPS support.
26497 (mips_print_operand_punctuation): Support short delay slots and
26499 (umips_swm_mask, umips_swm_encoding): New.
26500 (umips_build_save_restore): New function.
26501 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26502 (was_mips16_p): Remove.
26503 (old_compression_mode): New.
26504 (mips_set_compression_mode): New function.
26505 (mips_set_current_function): Add microMIPS support.
26506 (mips_option_override): Likewise.
26507 (umips_save_restore_pattern_p): New function.
26508 (umips_output_save_restore): New function.
26509 (umips_load_store_pair_p_1): New function.
26510 (umips_load_store_pair_p): New function.
26511 (umips_output_load_store_pair_1): New function.
26512 (umips_output_load_store_pair): New function.
26513 (umips_movep_target_p) New function.
26514 (mips_prepare_pch_save): Add microMIPS support.
26515 * config/mips/mips.h (TARGET_COMPRESSION): New.
26516 (TARGET_CPU_CPP_BUILTINS): Update macro
26517 to use new compression flags and to support microMIPS.
26518 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26519 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26520 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26521 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26522 (ASM_SPEC): Support mmicromips and mno-micromips.
26523 (M16STORE_REG_P): New macro.
26524 (MIPS_CALL): Support TARGET_MICROMIPS.
26525 (MICROMIPS_J): New macro.
26526 (mips_base_mips16): Rename this...
26527 (mips_base_compression_flags): ...to this.
26528 (UMIPS_12BIT_OFFSET_P): New macro.
26529 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26530 (MULTILIB_DIRNAMES): Likewise.
26531 2013-03-20 Richard Biener <rguenther@suse.de>
26533 PR tree-optimization/56661
26534 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
26535 the result does not have to be distinct.
26537 2013-03-20 Richard Biener <rguenther@suse.de>
26539 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
26542 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
26543 Steven Bosscher <steven@gcc.gnu.org>
26545 PR rtl-optimization/56605
26546 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
26548 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
26551 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
26552 that require movd instead of movq.
26554 2013-03-20 Richard Biener <rguenther@suse.de>
26556 * tree-ssa-structalias.c (struct variable_info): Add pointer
26557 to the first field of an aggregate with sub-vars. Make
26558 this and the pointer to the next subfield its ID.
26559 (vi_next): New function.
26560 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
26561 storedanything_id, integer_id): Increment by one.
26562 (new_var_info, get_call_vi, lookup_call_clobber_vi,
26563 get_call_clobber_vi): Adjust.
26564 (solution_set_expand): Simplify and speedup.
26565 (solution_set_add): Inline into ...
26566 (set_union_with_increment): ... this. Adjust accordingly.
26567 (do_sd_constraint): Likewise.
26568 (do_ds_constraint): Likewise.
26569 (do_complex_constraint): Simplify.
26570 (build_pred_graph): Adjust.
26571 (solve_graph): Likewise. Simplify and speedup.
26572 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
26573 get_constraint_for_component_ref, get_constraint_for_1,
26574 first_vi_for_offset, first_or_preceding_vi_for_offset,
26575 create_function_info_for, create_variable_info_for_1,
26576 create_variable_info_for, intra_create_variable_infos): Adjust.
26577 (init_base_vars): Push NULL for ID zero.
26578 (compute_points_to_sets): Adjust.
26580 2013-03-20 Richard Biener <rguenther@suse.de>
26582 * cfgloop.c (verify_loop_structure): Streamline and avoid
26583 ICEing on corrupt loop tree.
26584 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
26587 2013-03-20 Richard Biener <rguenther@suse.de>
26589 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
26590 check whether an SSA update is needed.
26592 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
26594 * config/mips/constraints.md (T): Rename to...
26598 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
26599 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
26601 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26603 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
26604 (*subsi3_carryin_uxtw): Likewise.
26606 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26608 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
26609 (*rorsi3_insn_uxtw): Likewise.
26611 2013-03-19 Ian Bolton <ian.bolton@arm.com>
26613 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
26614 (*extrsi5_insn_uxtw): Likewise.
26616 2013-03-19 Richard Biener <rguenther@suse.de>
26618 PR tree-optimization/56273
26619 * passes.c (init_optimization_passes): Move second VRP after DOM.
26621 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
26623 * config/i386/i386.md (*movti_internal): Merge from
26624 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
26625 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
26626 nox64 isa attributes.
26628 2013-03-18 Richard Biener <rguenther@suse.de>
26630 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
26632 (merge_node_constraints): Likewise.
26633 (build_succ_graph): Likewise.
26634 (valid_graph_edge): Inline into single caller.
26635 (unify_nodes): Likewise. Use bitmap_set_bit return value
26637 (scc_visit): Fix formatting and variable use.
26638 (do_sd_constraint): Use gcc_checking_assert.
26639 (do_ds_constraint): Likewise.
26640 (do_complex_constraint): Likewise.
26641 (condense_visit): Likewise. Cleanup.
26642 (dump_pred_graph): New function.
26643 (perform_var_substitution): Dump the pred-graph before
26644 variable substitution.
26645 (find_equivalent_node): Use gcc_checking_assert.
26646 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
26648 2013-03-18 Richard Biener <rguenther@suse.de>
26650 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
26651 Remove cond_expr_stmt_list argument and do not gimplify the
26653 (vect_loop_versioning): Adjust.
26654 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
26655 Cleanup to use less temporaries.
26656 (vect_create_data_ref_ptr): Cleanup.
26658 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26660 PR tree-optimization/56635
26661 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
26662 require types_compatible_p types.
26664 2013-03-18 Nick Clifton <nickc@redhat.com>
26666 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
26667 spurious backslash.
26669 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
26670 Add missing line to comment describing function.
26672 2013-03-18 Richard Biener <rguenther@suse.de>
26674 PR tree-optimization/56210
26675 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26676 Handle string / character search functions.
26677 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
26679 2013-03-18 Richard Biener <rguenther@suse.de>
26681 PR middle-end/56483
26682 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
26683 and implement properly.
26684 * gimple.h (gimple_cond_single_var_p): Remove.
26686 2013-03-18 Richard Biener <rguenther@suse.de>
26688 * tree-data-ref.h (find_data_references_in_loop): Declare.
26689 * tree-data-ref.c (get_references_in_stmt): Use a stack
26690 vector pre-allocated in the callers.
26691 (find_data_references_in_stmt): Adjust.
26692 (graphite_find_data_references_in_stmt): Likewise.
26693 (create_rdg_vertices): Likewise.
26694 (find_data_references_in_loop): Export.
26695 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
26696 Compute dependences here...
26697 (vect_analyze_data_refs): ...not here. When we encounter
26698 a non-vectorizable data reference in basic-block vectorization
26699 truncate the data reference vector. Do not bother to
26700 fixup data-dependence information for gather loads.
26701 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
26702 of data references, as reported.
26704 2013-03-18 Richard Biener <rguenther@suse.de>
26706 PR tree-optimization/3713
26707 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
26708 has_constants and expr.
26709 (stmt_has_constants): Properly valueize SSA names when deciding
26710 whether the stmt has constants.
26712 2013-03-18 Richard Biener <rguenther@suse.de>
26714 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
26715 whole function when there is nothing to do.
26716 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
26717 * tree-vectorizer.c (vectorize_loops): Update virtual and
26718 loop-closed SSA once.
26719 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
26721 2013-03-18 Richard Biener <rguenther@suse.de>
26723 PR middle-end/56113
26724 * domwalk.c (bb_postorder): New global static.
26725 (cmp_bb_postorder): New function.
26726 (walk_dominator_tree): Replace scheme imposing an order for
26727 visiting dominator sons by one sorting them at the time they
26728 are pushed on the stack.
26730 2013-03-18 Richard Biener <rguenther@suse.de>
26732 PR tree-optimization/39326
26733 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
26734 (struct mem_ref): Replace mem member with ao_ref typed member.
26735 (MEM_ANALYZABLE): Adjust.
26736 (memref_eq): Likewise.
26737 (mem_ref_alloc): Likewise.
26738 (gather_mem_refs_stmt): Likewise.
26739 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
26740 (execute_sm_if_changed_flag_set): Adjust.
26741 (execute_sm): Likewise.
26742 (ref_always_accessed_p): Likewise.
26743 (refs_independent_p): Likewise.
26744 (can_sm_ref_p): Likewise.
26746 2013-03-18 Jakub Jelinek <jakub@redhat.com>
26749 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
26750 return 1 even for !unsignedp.
26752 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26754 * config/i386/i386.md (isa): Add x64 and nox64.
26755 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
26756 (*pushtf): Enable *roF alternative for x64 isa only.
26757 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
26758 mode attribute of integer alternatives to DImode for TARGET_64BIT.
26759 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
26760 (*movtf_internal): Merge from *movtf_internal_rex64 and
26761 *movtf_internal_sse. Use x64 and nox64 isa attributes.
26762 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
26763 nox64 isa attributes.
26764 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
26765 nox64 isa attributes.
26766 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
26768 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
26770 * config/alpha/alpha.c (TARGET_LRA_P): New define.
26772 2013-03-17 Jakub Jelinek <jakub@redhat.com>
26775 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
26776 class names. Remove trailing comma after "ALL_REGS".
26778 2013-03-16 Jan Hubicka <jh@suse.cz>
26780 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
26781 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
26782 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
26783 of cgraph_get_create_node.
26784 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
26786 2013-03-16 Jason Merrill <jason@redhat.com>
26789 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
26790 with DW_AT_default_value.
26792 2013-03-16 Jakub Jelinek <jakub@redhat.com>
26794 * BASE-VER: Set to 4.9.0.
26796 2013-03-14 Andi Kleen <ak@linux.intel.com>
26799 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
26800 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
26801 Document _x* TSX intrinsics.
26803 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
26804 David Holsgrove <david.holsgrove@xilinx.com>
26806 * configure.ac: Add MicroBlaze TLS support detection.
26807 * configure: Regenerate.
26808 * config/microblaze/microblaze-protos.h
26809 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
26810 symbol_mentioned_p, label_mentioned_p): Add prototypes.
26811 * config/microblaze/microblaze.c (microblaze_address_type): Add
26812 ADDRESS_TLS and tls_reloc address types.
26813 (microblaze_address_info): Add tls_reloc.
26814 (TARGET_HAVE_TLS): Define.
26815 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
26816 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
26817 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
26818 load_tls_operand, microblaze_call_tls_get_addr,
26819 microblaze_legitimize_tls_address): New functions.
26820 (microblaze_classify_unspec): Handle UNSPEC_TLS.
26821 (get_base_reg): Use microblaze_tls_symbol_p.
26822 (microblaze_classify_address): Handle TLS.
26823 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
26824 label_mentioned_p and microblaze_tls_referenced_p.
26825 (microblaze_legitimize_address): Handle TLS.
26826 (microblaze_address_insns): Handle ADDRESS_TLS.
26827 (pic_address_needs_scratch): Handle TLS.
26828 (print_operand_address): Handle TLS.
26829 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
26830 (microblaze_expand_move): Handle TLS.
26831 (microblaze_legitimate_constant_p): Check
26832 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
26833 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
26834 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
26835 (PIC_OFFSET_TABLE_REGNUM): Set.
26836 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
26837 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
26838 (addsi3, movsi_internal2, movdf_internal): Update constraints
26839 * config/microblaze/predicates.md (arith_plus_operand): Define
26840 (move_operand): Redefine as move_src_operand,
26841 check microblaze_tls_referenced_p.
26843 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26845 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
26846 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
26848 2013-03-14 Ian Bolton <ian.bolton@arm.com>
26850 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
26853 2013-03-14 Jakub Jelinek <jakub@redhat.com>
26855 PR tree-optimization/53265
26856 * common.opt (Waggressive-loop-optimizations): New option.
26857 * tree-ssa-loop-niter.c: Include tree-pass.h.
26858 (do_warn_aggressive_loop_optimizations): New function.
26859 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
26860 if number_of_latch_executions returned constant.
26861 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
26862 early. If number_of_latch_executions returned constant, set
26863 nb_iterations_upper_bound back to it.
26864 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
26866 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
26867 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
26869 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
26870 (MULTILIB_OSDIRNAMES): Set.
26871 * genmultilib: If defaultosdirname doesn't start with :: , set
26872 defaultosdirname2 instead, clear it and emit two . multilib_raw
26873 entries instead of just one.
26875 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
26877 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26878 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
26879 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
26880 (SUBTARGET_OVERRIDE_OPTIONS): New.
26882 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
26885 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
26886 (musermode): Convert to Var(TARGET_USERMODE).
26887 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
26888 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
26889 * config/sh/sh.c (sh_option_override): Use
26890 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
26891 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
26893 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
26895 (udivsi3_i4_single, divsi3_i4_single): Use
26896 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
26898 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
26900 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
26903 2013-03-13 Richard Biener <rguenther@suse.de>
26905 PR tree-optimization/56608
26906 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
26907 calls when vectorizing basic-blocks.
26909 2013-03-13 Jakub Jelinek <jakub@redhat.com>
26912 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
26915 2013-03-12 Jakub Jelinek <jakub@redhat.com>
26917 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
26919 2013-03-11 Jan Hubicka <jh@suse.cz>
26922 * lto-streamer-out.c (output_symbol_p): Skip references from
26923 constructors of external variables.
26925 2013-03-11 Jan Hubicka <jh@suse.cz>
26927 PR middle-end/56571
26928 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
26930 * emit-rtl.c (verify_rtx_sharing): Likewise.
26931 (copy_insn_1): Likewise.
26932 * rtl.c (copy_rtx): Likewise.
26934 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
26937 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
26938 output_operand_lossage message.
26940 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
26943 * arm.c (shift_op): Validate RTL pattern on the fly.
26944 (arm_print_operand, case 'S'): Don't use shift_operator to validate
26947 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26950 * config/pa/pa.md (call_value): Check for calls to powf and direct to
26951 new call patterns that clobber %fr12.
26952 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
26953 split and postreload patterns.
26954 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
26955 registers %fr12 and %fr12R as call used.
26957 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
26959 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
26960 (canon_address, record_store, replace_read, check_mem_read_rtx,
26961 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
26962 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
26963 rest_of_handle_dse): Likewise.
26965 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
26967 PR middle-end/56524
26968 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
26970 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
26971 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
26972 (save_optabs_if_changed): Replace with...
26973 (init_tree_optimization_optabs): ...this.
26974 * optabs.c (save_optabs_if_changed): Rename to...
26975 (init_tree_optimization_optabs): ...this. Take the optimization node
26976 as argument. Do nothing if the base optabs are already correct.
26977 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
26978 to recompute optabs.
26979 * function.h (function): Remove optabs field.
26980 * function.c (invoke_set_current_function_hook): Call
26981 init_tree_optimization_optabs. Use the result to initialize
26984 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
26986 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
26987 if GTMA_HAS_NO_INSTRUMENTATION.
26988 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
26989 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
26990 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
26991 * gimple-pretty-print.c (dump_gimple_transaction): Handle
26992 GTMA_HAS_NO_INSTRUMENTATION.
26994 2013-03-08 Jakub Jelinek <jakub@redhat.com>
26996 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
26999 2013-03-08 Marek Polacek <polacek@redhat.com>
27000 Jakub Jelinek <jakub@redhat.com>
27002 PR tree-optimization/56478
27003 * predict.c (is_comparison_with_loop_invariant_p): Change the
27004 type of loop_step to tree.
27005 (predict_loops): Adjust.
27006 (predict_iv_comparison): Perform the computations on double_ints.
27008 2013-03-08 Richard Biener <rguenther@suse.de>
27010 PR tree-optimization/56570
27011 * tree-cfg.c (verify_expr_location_1): Verify locations for
27013 * tree-sra.c (create_access_replacement): Strip locations
27014 from DECL_DEBUG_EXPRs.
27016 2013-03-08 Richard Biener <rguenther@suse.de>
27018 * tree-inline.c (expand_call_inline): Do not associate
27019 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
27020 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
27022 2013-03-08 Richard Biener <rguenther@suse.de>
27024 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
27025 or block changes with -Og. Fix for location / block encoding
27026 changes and PHI arguments with locations.
27028 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
27030 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
27032 (struct output_info): Likewise.
27033 (register_overhead): Remove bad gcc_assert.
27034 (bitmap_find_bit): If there is only a single bitmap element, do not
27035 count a miss as a search.
27036 (print_statistics): Update for counter type changes.
27037 (dump_bitmap_statistics): Likewise. Print headers such that they
27038 are properly lined up with the printed counters.
27040 2013-03-07 Jakub Jelinek <jakub@redhat.com>
27042 PR tree-optimization/56559
27043 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
27044 check that it has only a single use.
27046 2013-03-07 Richard Biener <rguenther@suse.de>
27048 * doc/invoke.texi (fwhole-program): Discourage use in combination
27051 2013-03-06 Jakub Jelinek <jakub@redhat.com>
27053 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
27055 PR tree-optimization/56539
27056 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
27057 instead of GSI_CONTINUE_LINKING as last argument to
27058 force_gimple_operand_gsi. Adjust function comment.
27060 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
27063 PR middle-end/56548
27064 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
27065 promoted mode, convert the result back to the original mode.
27067 2013-03-06 Richard Biener <rguenther@suse.de>
27069 PR middle-end/56294
27070 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
27071 (insert_updated_phi_nodes_compare_uids): New function.
27072 (update_ssa): Sort symbols_to_rename after UID before
27073 traversing it to insert PHI nodes.
27075 2013-03-06 Richard Biener <rguenther@suse.de>
27077 PR middle-end/50494
27078 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
27079 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
27082 2013-02-13 Richard Biener <rguenther@suse.de>
27085 * varasm.c (output_constant_def_1): Get the decl representing
27086 the constant as argument.
27087 (output_constant_def): Wrap output_constant_def_1.
27088 (make_decl_rtl): Use output_constant_def_1 with the decl
27089 representing the constant.
27090 (build_constant_desc): Optionally re-use a decl already
27091 representing the constant.
27092 (tree_output_constant_def): Adjust.
27094 2013-03-06 Joey Ye <joey.ye@arm.com>
27097 * gcc.c (convert_white_space): New function.
27098 (main): Handles white space in function name.
27100 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
27103 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
27104 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
27105 to SH_DIV_CALL_TABLE for TARGET_SH2.
27106 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
27108 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
27109 call-table options.
27111 2013-03-05 Sterling Augustine <saugustine@google.com>
27112 Cary Coutant <ccoutant@google.com>
27115 * dwarf2out.c (resolve_addr): Don't call
27116 remove_loc_list_addr_table_entries a second time for the same
27119 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27122 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
27123 (avoid_complex_debug_insns): New function.
27124 (expand_debug_locations): Call it.
27126 PR rtl-optimization/56484
27127 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
27128 lifetimes of hard registers on small register class machines.
27130 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27132 * config/microblaze/microblaze-protos.h: Rename
27133 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
27134 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
27136 (microblaze_fast_interrupt_function_p): New function.
27137 (microblaze_is_interrupt_handler): Rename to
27138 microblaze_is_interrupt_variant and add fast_interrupt check.
27139 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
27140 (save_restore_insns): Likewise.
27141 (compute_frame_size): Likewise.
27142 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
27143 (microblaze_globalize_label): Likewise.
27144 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
27145 * config/microblaze/microblaze.md: Use wrapper
27146 microblaze_is_interrupt_variant.
27148 2013-03-05 Kai Tietz <ktietz@redhat.com>
27150 * sdbout.c (sdbout_one_type): Switch to current function's section
27151 supporting cold/hot.
27153 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
27155 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
27158 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27160 PR middle-end/56461
27161 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
27162 if VALGRIND_GET_VBITS is defined, temporarily make object
27163 memory all defined, and restore previous valgrind addressability
27164 and definability afterwards. Free this_object at the end.
27166 PR middle-end/56461
27167 * lra.c (lra): Call lra_clear_live_ranges if live_p,
27168 right before calling lra_create_live_ranges, also call it
27169 when clearing live_p. Only call lra_clear_live_ranges
27170 at the end if live_p.
27172 PR middle-end/56461
27173 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
27175 2013-03-05 Richard Biener <rguenther@suse.de>
27177 PR tree-optimization/56521
27178 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
27181 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
27184 * except.h (remove_unreachable_eh_regions): New prototype.
27185 * except.c (remove_eh_handler_splicer): New function, split out
27186 of remove_eh_handler.
27187 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
27188 warning about running it on many EH regions one at a time.
27189 (remove_unreachable_eh_regions_worker): New function, walk the
27190 EH tree in depth-first order and remove non-marked regions.
27191 (remove_unreachable_eh_regions): New function.
27192 * tree-eh.c (mark_reachable_handlers): New function, split out
27193 from remove_unreachable_handlers.
27194 (remove_unreachable_handlers): Use mark_reachable_handlers and
27195 remove_unreachable_eh_regions.
27196 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
27197 and remove_unreachable_eh_regions.
27199 2013-03-05 Richard Biener <rguenther@suse.de>
27201 PR middle-end/56525
27202 * loop-init.c (fix_loop_structure): Remove loops in two stages,
27203 not freeing them until the end.
27205 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
27207 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
27209 2013-03-05 Richard Biener <rguenther@suse.de>
27211 PR tree-optimization/56270
27212 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
27213 of loads after scheduling an SLP instance.
27215 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27217 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
27219 (check_gcc_parallelize): Run guality.exp as a separate job from
27220 vect.exp with unsorted.exp and $(dg_target_exps) separately from
27221 struct-layout-1.exp with stackalign.exp.
27223 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
27225 PR middle-end/56461
27226 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
27227 load_index sbitmap even if some bit in it isn't set.
27229 PR middle-end/56461
27230 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
27231 (discover_iteration_bound_by_body_walk): Change queues to
27232 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
27233 spelling in comment. Call safe_push on queues[bound_index] directly.
27234 Release queues[queue_index] in every iteration unconditionally.
27235 Release bounds vector.
27237 PR middle-end/56461
27238 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
27239 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
27240 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
27243 2013-03-05 Richard Biener <rguenther@suse.de>
27246 * tree-inline.c (remap_blocks_to_null): New function.
27247 (expand_call_inline): When expanding a call stmt without
27248 an associated block inline remap all callee blocks to NULL.
27250 2013-03-05 Jakub Jelinek <jakub@redhat.com>
27252 PR rtl-optimization/56494
27253 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
27254 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
27255 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
27257 PR middle-end/56461
27258 * sel-sched-ir.c (free_sched_pools): Release
27259 succs_info_pool.stack[succs_info_pool.max_top] vectors too
27260 if succs_info_pool.max_top isn't -1.
27263 * opts.c (opts_obstack, opts_concat): Moved to...
27264 * opts-common.c (opts_obstack, opts_concat): ... here.
27266 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27268 PR middle-end/56461
27269 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
27271 2013-03-04 Martin Jambor <mjambor@suse.cz>
27273 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
27274 all appropriate places.
27276 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
27278 PR tree-optimization/56424
27279 * ipa-split.c (split_function): Do not set the RSO flag if result is
27280 not by reference and its type is a register type.
27282 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
27284 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
27285 (microblaze_legitimate_pic_operand): Likewise
27286 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
27287 new function microblaze_legitimate_pic_operand
27288 * config/microblaze/microblaze-protos.h
27289 (microblaze_legitimate_pic_operand): Declare.
27291 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27293 * config/microblaze/predicates.md (call_insn_simple_operand):
27294 New predicate for supported rtx code types.
27295 * config/microblaze/microblaze.md (call_internal1): Use
27296 call_insn_simple_operand predicate.
27298 2013-03-04 Jakub Jelinek <jakub@redhat.com>
27300 PR middle-end/56461
27301 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
27302 partitions.ordered_remove.
27304 PR middle-end/56461
27305 * tree-vect-stmts.c (vectorizable_conversion): Don't call
27306 vec_oprnds0.create (1) for modifier == NONE.
27308 PR middle-end/56461
27309 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
27310 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
27311 vec_oprnds1 right before pushing anything to it for
27314 PR middle-end/56461
27315 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
27316 set nbbs to 0 instead of having separate code path.
27317 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
27318 instead of false as last argument if returning NULL.
27320 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
27322 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
27323 the attribute is now called "target" instead of "option".
27324 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
27325 * doc/tm.texi.in (Target Attributes): Likewise document the correct
27326 attribute/pragma name for TARGET_OPTION_VALID_P and
27327 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
27328 * doc/tm.texi: Regenerated.
27330 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
27332 * config/microblaze/microblaze.c:
27333 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
27334 * config/microblaze/microblaze.h: Add -mxl-reorder to
27336 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
27337 instructions emitted if TARGET_REORDER.
27338 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
27339 or 0 for -m/-mno case, but initialises as 2 to detect default use case
27342 2013-03-01 Xinliang David Li <davidxl@google.com>
27344 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
27347 2013-03-01 Jakub Jelinek <jakub@redhat.com>
27349 PR middle-end/56461
27350 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
27351 vector even when returning true. Fix up function comment formatting.
27353 PR middle-end/56461
27354 * ira-build.c (ira_loop_nodes_count): New variable.
27355 (create_loop_tree_nodes): Initialize it.
27356 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
27358 PR middle-end/56461
27359 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
27360 method on dr_chain and result_chain.
27361 * tree-vect-stmts.c (vectorizable_store): Only call
27362 result_chain.create if j == 0.
27364 PR middle-end/56461
27365 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
27366 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
27367 before overwriting it.
27369 2013-03-01 Tobias Burnus <burnus@net-b.de>
27371 * doc/extended.texi (C Extensions): Change order in @menu
27373 (Other MIPS Built-in Functions): Move last MIPS entry before
27374 "picoChip Built-in Functions".
27375 (SH Built-in Functions): Move after RX Built-in Functions.
27376 * doc/gcc.texi (Introduction): Change order in @menu
27378 * doc/md.texi (Constraints): Ditto.
27379 * gty.texi (Type Information): Ditto.
27380 (User-provided marking routines for template types): Make
27382 * doc/invoke.texi (AArch64 Options): Move before
27383 "Adapteva Epiphany Options".
27385 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
27386 Jakub Jelinek <jakub@redhat.com>
27389 * asan.c (gate_asan): Lookup no_sanitize_address instead of
27390 no_address_safety_analysis attribute.
27391 * doc/extend.texi (no_address_safety_attribute): Rename to
27392 no_sanitize_address attribute, mention no_address_safety_analysis
27393 attribute as deprecated alias.
27395 2013-02-28 Jakub Jelinek <jakub@redhat.com>
27397 PR middle-end/56461
27398 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
27399 type to vec<vec<tree> > *.
27400 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
27401 to be vec<tree> instead of vec<tree> *, set vec_defs
27402 to vNULL and call vec_defs.create (number_of_vects), adjust other
27404 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
27405 vectorizable_condition): Adjust vect_get_slp_defs callers.
27407 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27409 * config/aarch64/aarch64.c
27410 (aarch64_float_const_representable): Remove unused variable.
27412 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27414 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
27416 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27418 * config/aarch64/aarch64-builtins.c
27419 (aarch64_init_simd_builtins): Make static.
27421 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
27423 * config/aarch64/aarch64.c
27424 (aarch64_simd_make_constant): Make static.
27426 2013-02-28 Martin Jambor <mjambor@suse.cz>
27428 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
27429 with no initialization to the RHS of debug statements.
27431 2013-02-28 Martin Jambor <mjambor@suse.cz>
27433 PR tree-optimization/56294
27434 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
27436 (get_access_replacement): Do not call create_access_replacement.
27437 Assert a replacement exists.
27438 (get_repl_default_def_ssa_name): Create the replacement declaration
27441 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
27443 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
27444 final_end_function.
27446 2013-02-28 Marek Polacek <polacek@redhat.com>
27448 PR rtl-optimization/56466
27449 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
27450 if we're changing a loop.
27451 (peel_loops_completely): Likewise.
27453 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
27456 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
27458 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27461 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
27462 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
27463 INTX_FTYPE_FX, FX_FTYPE_INTX.
27464 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
27466 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27468 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
27469 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
27470 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
27471 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
27472 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
27473 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
27474 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
27475 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
27476 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
27477 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
27478 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
27479 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
27480 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
27481 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
27482 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
27483 (avrxmega6): Increase max flash segments from 5 to 6.
27484 * config/avr/t-multilib: Regenerate.
27485 * config/avr/avr-tables.opt: Regenerate.
27486 * doc/avr-mmcu.texi: Regenerate.
27488 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
27490 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
27491 (avr_device_to_arch): Rename to avr_device_to_ld.
27492 (avr_device_to_as): New prototype.
27493 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
27494 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
27495 * config/avr/driver-avr.c (avr_device_to_as): New.
27496 (avr_device_to_arch): Rename to avr_device_to_ld.
27498 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27500 PR middle-end/56461
27501 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
27502 method on dr_chain and result_chain.
27504 PR middle-end/56461
27505 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
27506 pointer_set_destroy on not_executed_last_iteration.
27508 PR middle-end/56461
27509 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
27511 PR middle-end/56461
27512 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
27513 FOR_EACH_DEFINED_FUNCTION when freeing state.
27515 PR middle-end/56461
27516 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
27518 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
27521 PR middle-end/56461
27522 * ipa-cp.c (decide_whether_version_node): Call vec_free on
27523 known_aggs[i].items and release known_aggs vector.
27525 PR middle-end/56461
27526 * ipa-reference.c (propagate): Free node_info even for alias nodes.
27528 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27530 * config/microblaze/microblaze.c (microblaze_emit_compare):
27531 Use xor for EQ/NE comparisions.
27532 * config/microblaze/microblaze.md (cstoresf4): Add constraints
27533 (cbranchsf4): Adjust operator to comparison_operator.
27535 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27537 PR middle-end/56461
27538 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
27540 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
27541 vec_safe_push, always update *slot.
27542 (redirect_edge_var_map_clear): Use vec_free.
27543 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
27544 (free_var_map_entry): Use vec_free.
27545 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
27546 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
27548 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
27550 PR middle-end/45472
27551 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
27552 when the may_trap_p bit of the exprs being merged differs.
27553 Reorder tests for speculativeness in the logical and operator.
27555 2013-02-27 Jakub Jelinek <jakub@redhat.com>
27557 * incpath.c (add_standard_paths): Use reconcat instead of concat
27558 where appropriate and avoid leaking memory.
27560 * opts.h: Include obstack.h.
27561 (opts_concat): New prototype.
27562 (opts_obstack): New declaration.
27563 * opts.c (opts_concat): New function.
27564 (opts_obstack): New variable.
27565 (init_options_struct): Call gcc_init_obstack on opts_obstack.
27566 (finish_options): Use opts_concat instead of concat
27567 and XOBNEWVEC instead of XNEWVEC.
27568 * opts-common.c (generate_canonical_option, decode_cmdline_option,
27569 generate_option): Likewise.
27570 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
27571 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
27574 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
27575 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
27577 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27579 PR middle-end/56461
27580 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
27582 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27584 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
27585 (arm_block_move_unaligned_straight): Likewise.
27586 (arm_adjust_block_mem): Likewise.
27588 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
27591 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
27592 temp, cond and label.
27593 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
27596 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
27597 * config/c6x/c6x.h (dbx_register_map): Update declaration.
27600 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
27601 of prologue/epilogue functions.
27604 * config/tilegx/tilegx.c (tilegx_expand_prologue):
27605 Remove unused variable cfa_offset.
27606 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
27609 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
27610 type promotion to unsigned.
27613 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
27614 for HOST_WIDE_INT of 32 bit / same size as int.
27615 (arm_block_move_unaligned_straight): Likewise.
27616 (arm_adjust_block_mem): Likewise.
27619 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
27622 2013-02-26 Marek Polacek <polacek@redhat.com>
27624 PR tree-optimization/56426
27625 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
27627 2013-02-26 Richard Biener <rguenther@suse.de>
27630 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
27631 unused variable loops.
27633 2013-02-26 Jakub Jelinek <jakub@redhat.com>
27635 PR tree-optimization/56448
27636 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
27637 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
27638 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
27639 later operands of the references, or even first operand for
27640 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
27642 PR tree-optimization/56443
27643 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
27644 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
27645 to type_for_mode langhook.
27647 2013-02-25 Matt Turner <mattst88@gmail.com>
27649 * doc/invoke.texi: Document r4700.
27651 2013-02-25 Richard Biener <rguenther@suse.de>
27653 PR tree-optimization/56175
27654 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
27656 (simplify_bitwise_binary): ... here. Also guard the conversion
27657 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
27659 2013-02-25 Catherine Moore <clm@codesourcery.com>
27662 2013-02-24 Catherine Moore <clm@codesourcery.com>
27663 Maciej W. Rozycki <macro@codesourcery.com>
27664 Tom de Vries <tom@codesourcery.com>
27665 Nathan Sidwell <nathan@codesourcery.com>
27666 Iain Sandoe <iain@codesourcery.com>
27667 Nathan Froyd <froydnj@codesourcery.com>
27668 Chao-ying Fu <fu@mips.com>
27670 * doc/extend.texi: (micromips, nomicromips, nocompression):
27671 Document new function attributes.
27672 * doc/invoke.texi (minterlink-compressed, mmicromips,
27673 m14k, m14ke, m14kec): Document new options.
27674 (minterlink-mips16): Update documentation.
27675 * doc/md.texi (ZC, ZD): Document new constraints.
27676 * configure.ac (gcc_cv_as_micromips): Check if linker
27677 supports the .set micromips directive.
27678 * configure: Regenerate.
27679 * config.in: Regenerate.
27680 * config/mips/mips-tables.opt: Regenerate.
27681 * config/mips/micromips.md: New file.
27682 * constraints.md (ZC, AD): New constraints.
27683 * config/mips/predicates.md (movep_src_register): New predicate.
27684 (movep_src_operand): New predicate.
27685 (non_volatile_mem_operand): New predicate.
27686 * config/mips/mips.md (multimem): New type.
27687 (length): Differentiate between 17-bit and 18-bit branch offsets.
27688 (MOVEP1, MOVEP2): New mode iterator.
27689 (mov_<load>l): Use ZC constraint.
27690 (mov_<load>r): Likewise.
27691 (mov_<store>l): Likewise.
27692 (mov_<store>r): Likewise.
27693 (*branch_equality<mode>_inverted): Add microMIPS support.
27694 (*branch_equality<mode>): Likewise.
27695 (*jump_absolute): Likewise.
27696 (indirect_jump_<mode>): Likewise.
27697 (tablejump_<mode>): Likewise.
27698 (<optab>_internal): Likewise.
27699 (sibcall_internal): Likewise.
27700 (sibcall_value_internal): Likewise.
27701 (prefetch): Use constraint ZD.
27702 * config/mips/mips.opt (minterlink-compressed): New option.
27703 (minterlink-mips16): Now an alias for minterlink-compressed.
27704 (mmicromips): New option.
27705 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27706 (compare_and_swap_12): Likewise.
27707 (sync_add<mode>): Likewise.
27708 (sync_<optab>_12): Likewise.
27709 (sync_old_<optab>_12): Likewise.
27710 (sync_new_<optab>_12): Likewise.
27711 (sync_nand_12): Likewise.
27712 (sync_old_nand_12): Likewise.
27713 (sync_new_nand_12): Likewise.
27714 (sync_sub<mode>): Likewise.
27715 (sync_old_add<mode>): Likewise.
27716 (sync_old_sub<mode>): Likewise.
27717 (sync_new_add<mode>): Likewise.
27718 (sync_new_sub<mode>): Likewise.
27719 (sync_<optab><mode>): Likewise.
27720 (sync_old_<optab><mode>): Likewise.
27721 (sync_new_<optab><mode>): Likewise.
27722 (sync_nand<mode>): Likewise.
27723 (sync_old_nand<mode>): Likewise.
27724 (sync_new_nand<mode>): Likewise.
27725 (sync_lock_test_and_set<mode>): Likewise.
27726 (test_and_set_12): Likewise.
27727 (atomic_compare_and_swap<mode>): Likewise.
27728 (atomic_exchange<mode>_llsc): Likewise.
27729 (atomic_fetch_add<mode>_llsc): Likewise.
27730 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27731 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27732 (umips_save_restore_pattern_p): Likewise.
27733 (umips_load_store_pair_p): Likewise.
27734 (umips_output_load_store_pair): Likewise.
27735 (umips_movep_target_p): Likewise.
27736 (umips_12bit_offset_address_p): Likewise.
27737 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27738 (mips_base_mips16): Rename this...
27739 (mips_base_compression_flags): ...to this. Update all uses.
27740 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27741 (mips_mips16_decl_p): Delete.
27742 (mips_nomips16_decl_p): Delete.
27743 (mips_get_compress_on_flags): New function.
27744 (mips_get_compress_off_flags): New function.
27745 (mips_get_compress_mode): New function.
27746 (mips_get_compress_on_name): New function.
27747 (mips_get_compress_off_name): New function.
27748 (mips_insert_attributes): Support multiple compression types.
27749 (mips_merge_decl_attributes): Likewise.
27750 (umips_12bit_offset_address_p): New function.
27751 (mips_start_function_definition): Emit .set micromips directive.
27752 (mips_call_may_need_jalx_p): New function.
27753 (mips_function_ok_for_sibcall): Add microMIPS support.
27754 (mips_print_operand_punctuation): Support short delay slots and
27756 (umips_swm_mask, umips_swm_encoding): New.
27757 (umips_build_save_restore): New function.
27758 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27759 (was_mips16_p): Remove.
27760 (old_compression_mode): New.
27761 (mips_set_compression_mode): New function.
27762 (mips_set_current_function): Add microMIPS support.
27763 (mips_option_override): Likewise.
27764 (umips_save_restore_pattern_p): New function.
27765 (umips_output_save_restore): New function.
27766 (umips_load_store_pair_p_1): New function.
27767 (umips_load_store_pair_p): New function.
27768 (umips_output_load_store_pair_1): New function.
27769 (umips_output_load_store_pair): New function.
27770 (umips_movep_target_p) New function.
27771 (mips_prepare_pch_save): Add microMIPS support.
27772 * config/mips/mips.h (TARGET_COMPRESSION): New.
27773 (TARGET_CPU_CPP_BUILTINS): Update macro
27774 to use new compression flags and to support microMIPS.
27775 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27776 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27777 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27778 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27779 (ASM_SPEC): Support mmicromips and mno-micromips.
27780 (M16STORE_REG_P): New macro.
27781 (MIPS_CALL): Support TARGET_MICROMIPS.
27782 (MICROMIPS_J): New macro.
27783 (mips_base_mips16): Rename this...
27784 (mips_base_compression_flags): ...to this.
27785 (UMIPS_12BIT_OFFSET_P): New macro.
27786 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27787 (MULTILIB_DIRNAMES): Likewise.
27789 2013-02-25 Tom de Vries <tom@codesourcery.com>
27791 PR rtl-optimization/56131
27792 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
27793 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
27794 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
27796 2013-02-25 Tobias Burnus <burnus@net-b.de>
27798 * doc/invoke.texi (-fsanitize=): Move from optimization
27799 to debugging options.
27801 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27803 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
27805 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
27806 Alexander Monakov <amonakov@ispras.ru>
27808 PR middle-end/56077
27809 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
27810 flush pending lists also on non-jumps. Adjust comment.
27812 2013-02-24 Catherine Moore <clm@codesourcery.com>
27813 Maciej W. Rozycki <macro@codesourcery.com>
27814 Tom de Vries <tom@codesourcery.com>
27815 Nathan Sidwell <nathan@codesourcery.com>
27816 Iain Sandoe <iain@codesourcery.com>
27817 Nathan Froyd <froydnj@codesourcery.com>
27818 Chao-ying Fu <fu@mips.com>
27820 * doc/extend.texi: (micromips, nomicromips, nocompression):
27821 Document new function attributes.
27822 * doc/invoke.texi (minterlink-compressed, mmicromips,
27823 m14k, m14ke, m14kec): Document new options.
27824 (minterlink-mips16): Update documentation.
27825 * doc/md.texi (ZC, ZD): Document new constraints.
27826 * configure.ac (gcc_cv_as_micromips): Check if linker
27827 supports the .set micromips directive.
27828 * configure: Regenerate.
27829 * config.in: Regenerate.
27830 * config/mips/mips-tables.opt: Regenerate.
27831 * config/mips/micromips.md: New file.
27832 * constraints.md (ZC, AD): New constraints.
27833 * config/mips/predicates.md (movep_src_register): New predicate.
27834 (movep_src_operand): New predicate.
27835 (non_volatile_mem_operand): New predicate.
27836 * config/mips/mips.md (multimem): New type.
27837 (length): Differentiate between 17-bit and 18-bit branch offsets.
27838 (MOVEP1, MOVEP2): New mode iterator.
27839 (mov_<load>l): Use ZC constraint.
27840 (mov_<load>r): Likewise.
27841 (mov_<store>l): Likewise.
27842 (mov_<store>r): Likewise.
27843 (*branch_equality<mode>_inverted): Add microMIPS support.
27844 (*branch_equality<mode>): Likewise.
27845 (*jump_absolute): Likewise.
27846 (indirect_jump_<mode>): Likewise.
27847 (tablejump_<mode>): Likewise.
27848 (<optab>_internal): Likewise.
27849 (sibcall_internal): Likewise.
27850 (sibcall_value_internal): Likewise.
27851 (prefetch): Use constraint ZD.
27852 * config/mips/mips.opt (minterlink-compressed): New option.
27853 (minterlink-mips16): Now an alias for minterlink-compressed.
27854 (mmicromips): New option.
27855 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
27856 (compare_and_swap_12): Likewise.
27857 (sync_add<mode>): Likewise.
27858 (sync_<optab>_12): Likewise.
27859 (sync_old_<optab>_12): Likewise.
27860 (sync_new_<optab>_12): Likewise.
27861 (sync_nand_12): Likewise.
27862 (sync_old_nand_12): Likewise.
27863 (sync_new_nand_12): Likewise.
27864 (sync_sub<mode>): Likewise.
27865 (sync_old_add<mode>): Likewise.
27866 (sync_old_sub<mode>): Likewise.
27867 (sync_new_add<mode>): Likewise.
27868 (sync_new_sub<mode>): Likewise.
27869 (sync_<optab><mode>): Likewise.
27870 (sync_old_<optab><mode>): Likewise.
27871 (sync_new_<optab><mode>): Likewise.
27872 (sync_nand<mode>): Likewise.
27873 (sync_old_nand<mode>): Likewise.
27874 (sync_new_nand<mode>): Likewise.
27875 (sync_lock_test_and_set<mode>): Likewise.
27876 (test_and_set_12): Likewise.
27877 (atomic_compare_and_swap<mode>): Likewise.
27878 (atomic_exchange<mode>_llsc): Likewise.
27879 (atomic_fetch_add<mode>_llsc): Likewise.
27880 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
27881 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
27882 (umips_save_restore_pattern_p): Likewise.
27883 (umips_load_store_pair_p): Likewise.
27884 (umips_output_load_store_pair): Likewise.
27885 (umips_movep_target_p): Likewise.
27886 (umips_12bit_offset_address_p): Likewise.
27887 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
27888 (mips_base_mips16): Rename this...
27889 (mips_base_compression_flags): ...to this. Update all uses.
27890 (mips_attribute_table): Add micromips, nomicromips and nocompression.
27891 (mips_mips16_decl_p): Delete.
27892 (mips_nomips16_decl_p): Delete.
27893 (mips_get_compress_on_flags): New function.
27894 (mips_get_compress_off_flags): New function.
27895 (mips_get_compress_mode): New function.
27896 (mips_get_compress_on_name): New function.
27897 (mips_get_compress_off_name): New function.
27898 (mips_insert_attributes): Support multiple compression types.
27899 (mips_merge_decl_attributes): Likewise.
27900 (umips_12bit_offset_address_p): New function.
27901 (mips_start_function_definition): Emit .set micromips directive.
27902 (mips_call_may_need_jalx_p): New function.
27903 (mips_function_ok_for_sibcall): Add microMIPS support.
27904 (mips_print_operand_punctuation): Support short delay slots and
27906 (umips_swm_mask, umips_swm_encoding): New.
27907 (umips_build_save_restore): New function.
27908 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
27909 (was_mips16_p): Remove.
27910 (old_compression_mode): New.
27911 (mips_set_compression_mode): New function.
27912 (mips_set_current_function): Add microMIPS support.
27913 (mips_option_override): Likewise.
27914 (umips_save_restore_pattern_p): New function.
27915 (umips_output_save_restore): New function.
27916 (umips_load_store_pair_p_1): New function.
27917 (umips_load_store_pair_p): New function.
27918 (umips_output_load_store_pair_1): New function.
27919 (umips_output_load_store_pair): New function.
27920 (umips_movep_target_p) New function.
27921 (mips_prepare_pch_save): Add microMIPS support.
27922 * config/mips/mips.h (TARGET_COMPRESSION): New.
27923 (TARGET_CPU_CPP_BUILTINS): Update macro
27924 to use new compression flags and to support microMIPS.
27925 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
27926 (MIPS_ARCH_FLOAT_SPEC): Likewise.
27927 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
27928 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
27929 (ASM_SPEC): Support mmicromips and mno-micromips.
27930 (M16STORE_REG_P): New macro.
27931 (MIPS_CALL): Support TARGET_MICROMIPS.
27932 (MICROMIPS_J): New macro.
27933 (mips_base_mips16): Rename this...
27934 (mips_base_compression_flags): ...to this.
27935 (UMIPS_12BIT_OFFSET_P): New macro.
27936 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
27937 (MULTILIB_DIRNAMES): Likewise.
27939 2013-02-24 Jakub Jelinek <jakub@redhat.com>
27942 * target-globals.c (save_target_globals): For init_reg_sets and
27943 target_reinit remporarily set this_fn_optabs to this_target_optabs.
27945 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
27947 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
27948 * config/aarch64/t-aarch64
27949 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
27951 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
27953 PR inline-asm/56148
27954 * lra-constraints.c (process_alt_operands): Reload operand
27955 conflicting with earlier clobber only if no more other conflicting
27958 2013-02-22 Jakub Jelinek <jakub@redhat.com>
27961 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
27962 if not linking a shared library.
27964 2013-02-22 Seth LaForge <sethml@google.com>
27966 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
27968 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
27970 * config/arm/arm.md (split for extendsidi): Update condition.
27971 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
27972 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
27973 (qhs_zextenddi_cstr): Likewise.
27975 2013-02-21 Jakub Jelinek <jakub@redhat.com>
27977 PR middle-end/56420
27978 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
27979 avoid signed wrapping.
27980 (expand_mult): Handle properly multiplication by
27981 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
27982 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
27983 in the compiler if coeff is HOST_WIDE_INT_MIN.
27984 (expand_divmod): Don't make ext_op1 static, change it's type to
27985 uhwi. Avoid undefined behavior in -INTVAL (op1).
27987 PR rtl-optimization/50339
27988 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
27990 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
27991 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
27992 into splitting_ashiftrt field.
27993 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
27995 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
27998 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
28000 PR middle-end/56108
28001 * trans-mem.c (execute_tm_mark): Do not expand transactions that
28002 are sure to go irrevocable.
28004 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
28006 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
28007 scalars are valid operands.
28009 2013-02-21 Martin Jambor <mjambor@suse.cz>
28011 PR tree-optimization/56310
28012 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
28013 only matching indices and non-negative final offsets.
28014 (intersect_aggregates_with_edge): Pass src_idx to
28015 agg_replacements_to_vector. Pass src_idx insstead of index to
28016 intersect_with_agg_replacements.
28018 2013-02-21 Martin Jambor <mjambor@suse.cz>
28020 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
28021 instead of hard-wired defaults.
28023 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
28025 * doc/invoke.texi (MIPS Options): Update documentation of the
28026 floating-point multiply-accumulate instruction restrictions.
28028 2013-02-21 Kostya Serebryany <kcc@google.com>
28030 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
28031 asan_shadow_offset on x86_64 linux.
28033 2013-02-21 Richard Biener <rguenther@suse.de>
28035 PR tree-optimization/56415
28037 2013-02-11 Richard Biener <rguenther@suse.de>
28039 PR tree-optimization/56273
28040 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28043 2013-02-21 Jakub Jelinek <jakub@redhat.com>
28046 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
28049 PR inline-asm/56405
28050 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
28051 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
28053 2013-02-20 Jan Hubicka <jh@suse.cz>
28055 PR tree-optimization/56265
28056 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
28057 when target is referenced for first time.
28059 2013-02-20 Richard Biener <rguenther@suse.de>
28061 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
28062 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
28063 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
28064 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
28065 not return anything.
28066 (rename_ssa_copies): Do not remove unused locals.
28067 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
28068 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
28069 * passes.c (execute_function_todo): Do not schedule unused locals
28070 removal if cleanup_tree_cfg did something.
28071 * tree-ssa-live.c (remove_unused_locals): Dump statistics
28072 about the number of removed locals.
28074 2013-02-20 Richard Biener <rguenther@suse.de>
28076 PR tree-optimization/56398
28077 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
28079 2013-02-20 Martin Jambor <mjambor@suse.cz>
28081 PR tree-optimization/55334
28082 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
28083 restricted pointers to arrays.
28085 2013-02-20 Richard Biener <rguenther@suse.de>
28086 Jakub Jelinek <jakub@redhat.com>
28088 PR tree-optimization/56396
28089 * tree-ssa-ccp.c (n_const_val): New static variable.
28090 (get_value): Return NULL for SSA names we don't have a lattice
28092 (ccp_initialize): Initialize n_const_val.
28093 * tree-ssa-copy.c (n_copy_of): New static variable.
28094 (init_copy_prop): Initialize n_copy_of.
28095 (get_value): Return NULL_TREE for SSA names we don't have a
28098 2013-02-20 Martin Jambor <mjambor@suse.cz>
28100 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
28102 2013-02-20 Richard Biener <rguenther@suse.de>
28104 * genpreds.c (write_lookup_constraint): Do not compare first
28105 letter of the constraint again.
28107 2013-02-20 Richard Biener <rguenther@suse.de>
28109 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
28111 (get_use_iv_cost): Terminate hashtable walk when coming across
28114 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
28116 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
28117 reassociation for avx2 targets.
28119 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28121 * config/microblaze/microblaze.c: microblaze_has_clz = 0
28122 Add version check for v8.10.a to enable microblaze_has_clz
28123 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
28124 version and TARGET_PATTERN_COMPARE check
28125 * config/microblaze/microblaze.md: New clzsi2 instruction
28127 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28129 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
28130 function before branching.
28132 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28134 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
28136 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
28138 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
28140 PR middle-end/55889
28141 * sel-sched.c: Include ira.h.
28142 (implicit_clobber_conflict_p): New function.
28143 (moveup_expr): Use it.
28144 * Makefile.in (sel-sched.o): Depend on ira.h.
28146 2013-02-19 Richard Biener <rguenther@suse.de>
28148 PR tree-optimization/56384
28149 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
28150 (vn_hash_type): Split out from ...
28151 (vn_hash_constant_with_type): ... here.
28152 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
28153 (vn_phi_eq): Compare types from vn_phi_s structure.
28154 (vn_phi_lookup): Populate vn_phi_s type.
28155 (vn_phi_insert): Likewise.
28157 2013-02-19 Jakub Jelinek <jakub@redhat.com>
28159 PR tree-optimization/56350
28160 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
28161 if haven't found reduction or nested cycle operand, rather than
28162 asserting we must find it.
28164 PR tree-optimization/56381
28165 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
28168 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
28169 Jakub Jelinek <jakub@redhat.com>
28172 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
28173 (swap_optab_enable): Same.
28174 (init_all_optabs): Use argument instead of global.
28175 * tree.h (struct tree_optimization_option): New field target_optabs.
28176 * expr.h (init_all_optabs): Add argument to prototype.
28177 (TREE_OPTIMIZATION_OPTABS): New.
28178 (save_optabs_if_changed): Protoize.
28179 * optabs.h: Declare this_fn_optabs.
28180 * optabs.c (save_optabs_if_changed): New.
28181 Declare this_fn_optabs.
28182 (init_optabs): Add argument to init_all_optabs() call.
28183 * function.c (invoke_set_current_function_hook): Handle per
28185 * function.h (struct function): New field optabs.
28186 * config/mips/mips.c (mips_set_mips16_mode): Handle when
28187 optimization_current_node has changed.
28188 * target-globals.h (save_target_globals_default_opts): Protoize.
28189 * target-globals.c (save_target_globals_default_opts): New.
28191 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28194 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
28195 registers %fr12 and %fr12R as call used.
28198 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
28199 and HImode, require all displacements to be an integer multiple of
28201 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
28202 only allow QImode and HImode when reload is in progress and strict is
28203 true. Likewise for symbolic addresses. Use base14_operand to check
28204 displacements in REG+BASE addresses.
28206 2013-02-18 Richard Biener <rguenther@suse.de>
28208 PR tree-optimization/56366
28209 * tree-vect-loop.c (get_initial_def_for_induction): Properly
28210 handle sign-conversion of outer-loop initial induction value.
28212 2013-02-18 Richard Biener <rguenther@suse.de>
28214 PR middle-end/56349
28215 * cfghooks.c (merge_blocks): If we merge a latch into another
28216 block adjust references to it.
28217 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
28218 (verify_loop_structure): Verify that a recorded latch is in fact
28221 2013-02-18 Richard Biener <rguenther@suse.de>
28223 PR tree-optimization/56321
28224 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
28225 order SSA name release and virtual operand unlinking.
28227 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28229 * config/microblaze/microblaze.md (save_stack_block): Define.
28230 (restore_stack_block): Likewise.
28232 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28234 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
28235 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
28236 * config/microblaze/microblaze.c (microblaze_option_override):
28237 Bail out early for PIC modes when target does not support PIC.
28239 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28241 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
28242 Replace with a microblaze version.
28243 (microblaze_trampoline_init): Adapt for microblaze.
28244 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
28247 2013-02-16 Jakub Jelinek <jakub@redhat.com>
28248 Dodji Seketeli <dodji@redhat.com>
28251 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
28252 (instrument_mem_region_access): Do not forget to always put
28253 instrumentation of the of 'base' and 'base + len' in a "if (len !=
28254 0) statement, even for cases where either 'base' or 'base + len'
28255 are not instrumented -- because they have been previously
28256 instrumented. Simplify the logic by putting all the statements
28257 instrument 'base + len' inside a sequence, and then insert that
28258 sequence right before the current insertion point. Then, to
28259 instrument 'base + len', just get an iterator on that statement.
28260 And do not forget to update the pointer to iterator the function
28261 received as argument.
28263 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
28265 PR rtl-optimization/56348
28266 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
28268 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
28270 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
28271 (clean_graph_dump_file): Pass base to start_graph_dump.
28273 2013-02-14 Richard Henderson <rth@redhat.com>
28276 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
28278 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
28280 * collect2-aix.h: Define F_LOADONLY.
28282 2013-02-14 Richard Biener <rguenther@suse.de>
28285 * varasm.c (output_constant_def_1): Get the decl representing
28286 the constant as argument.
28287 (output_constant_def): Wrap output_constant_def_1.
28288 (make_decl_rtl): Use output_constant_def_1 with the decl
28289 representing the constant.
28290 (build_constant_desc): Optionally re-use a decl already
28291 representing the constant.
28292 (tree_output_constant_def): Adjust.
28294 2013-02-14 Dodji Seketeli <dodji@redhat.com>
28297 * asan.c (instrument_builtin_call): Really put the length of the
28298 second source argument into src1_len.
28300 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28302 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
28303 argument. If it is false, don't create edge from then_bb to
28305 (insert_if_then_before_iter): Pass true to it.
28306 (build_check_stmt): Pass false to it.
28307 (transform_statements): Flush hash table only on extended basic
28308 block boundaries, rather than at the beginning of every bb.
28309 Don't flush hash table on nonfreeing_call_p calls.
28310 * tree-flow.h (nonfreeing_call_p): New prototype.
28311 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
28313 2013-02-13 David S. Miller <davem@davemloft.net>
28315 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
28317 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
28320 * ira.c (max_regno_before_ira): Move from ...
28322 (fix_reg_equiv_init): Use max_regno_before_ira instead of
28325 2013-02-13 Jakub Jelinek <jakub@redhat.com>
28327 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
28329 2013-02-13 Richard Biener <rguenther@suse.de>
28332 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
28333 globals in MEM_REFs.
28335 2013-02-13 Richard Biener <rguenther@suse.de>
28337 * loop-init.c (loop_optimizer_init): Clear loop state when
28338 re-initializing preserved loops.
28339 * loop-unswitch.c (unswitch_single_loop): Return whether
28340 we unswitched the loop. Do not verify loop state here.
28341 (unswitch_loops): When we unswitched a loop discover new loops.
28343 2013-02-13 Kostya Serebryany <kcc@google.com>
28345 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
28347 * sanitizer.def: Rename __asan_init to __asan_init_v1.
28349 2013-02-12 Dodji Seketeli <dodji@redhat.com>
28351 Avoid instrumenting duplicated memory access in the same basic block
28352 * Makefile.in (asan.o): Add new dependency on hash-table.h
28353 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
28354 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
28355 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
28356 (free_mem_ref_resources, has_mem_ref_been_instrumented)
28357 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
28358 (get_mem_ref_of_assignment): New functions.
28359 (get_mem_refs_of_builtin_call): Extract from
28360 instrument_builtin_call and tweak a little bit to make it fit with
28362 (instrument_builtin_call): Use the new
28363 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
28364 of is_gimple_builtin_call.
28365 (instrument_derefs, instrument_mem_region_access): Insert the
28366 instrumented memory reference into the hash table.
28367 (maybe_instrument_assignment): Renamed instrument_assignment into
28368 this, and change it to advance the iterator when instrumentation
28369 actually happened and return true in that case. This makes it
28370 homogeneous with maybe_instrument_assignment, and thus give a
28371 chance to callers to be more 'regular'.
28372 (transform_statements): Clear the memory reference hash table
28373 whenever we enter a new BB, when we cross a function call, or when
28374 we are done transforming statements. Use
28375 maybe_instrument_assignment instead of instrumentation. No more
28376 need to special case maybe_instrument_assignment and advance the
28377 iterator after calling it; it's now handled just like
28378 maybe_instrument_call. Update comment.
28380 2013-02-13 Richard Biener <rguenther@suse.de>
28382 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
28383 Fix loop discovery code.
28385 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
28387 PR inline-asm/56148
28388 * lra-constraints.c (process_alt_operands): Match early clobber
28389 operand with itself. Check conflicts with earlyclobber only if
28390 the operand is not reloaded. Prefer to reload conflicting operand
28391 if earlyclobber and matching operands are the same.
28393 2013-02-12 Richard Biener <rguenther@suse.de>
28396 * lto-streamer-out.c (write_symbol): Do not output symbols
28397 for hard register variables.
28399 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
28402 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
28403 (umulsidi3_insn, mulsidi3_insn): New insns.
28405 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
28407 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
28408 (struct tune_params): Add vec_costs field.
28409 * config/arm/arm.c (arm_builtin_vectorization_cost)
28410 (arm_add_stmt_cost): New functions.
28411 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
28412 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
28413 (arm_default_vec_cost): New struct of type cpu_vec_costs.
28414 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
28415 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
28416 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
28417 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
28419 2013-02-12 Richard Biener <rguenther@suse.de>
28422 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
28423 decls again if possible.
28425 2013-02-12 Richard Biener <rguenther@suse.de>
28427 PR middle-end/56288
28428 * tree-ssa.c (verify_ssa_name): Fix check, move
28429 SSA_NAME_IN_FREE_LIST check up.
28431 2013-02-12 Jakub Jelinek <jakub@redhat.com>
28432 Steven Bosscher <steven@gcc.gnu.org>
28434 PR rtl-optimization/56151
28435 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
28436 equal to op0 or op1, and last_insn pattern is CODE operation
28437 with MEM dest and one of the operands matches that MEM.
28439 2013-02-11 Sriraman Tallam <tmsriram@google.com>
28441 * doc/extend.texi: Document Function Multiversioning and "default"
28442 parameter string to target attribute.
28443 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
28444 target attribute parameter is "default".
28445 (ix86_compare_version_priority): Remove checks for target attribute.
28446 (ix86_mangle_function_version_assembler_name): Change error to sorry.
28447 Remove check for target attribute equal to NULL. Add assert.
28448 (ix86_generate_version_dispatcher_body): Change error to sorry.
28450 2013-02-11 Iain Sandoe <iain@codesourcery.com>
28451 Jack Howarth <howarth@bromo.med.uc.edu>
28452 Patrick Marlier <patrick.marlier@gmail.com>
28455 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
28456 define ENDFILE_SPEC as TM_DESTRUCTOR.
28457 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
28459 2013-02-11 Alexander Potapenko <glider@google.com>
28460 Jack Howarth <howarth@bromo.med.uc.edu>
28461 Jakub Jelinek <jakub@redhat.com>
28464 * config/darwin.c (cdtor_record): Rename ctor_record.
28465 (sort_cdtor_records): Rename sort_ctor_records.
28466 (finalize_dtors): New routine to sort destructors by
28467 priority before use in assemble_integer.
28468 (machopic_asm_out_destructor): Use finalize_dtors if needed.
28470 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
28472 PR rtl-optimization/56275
28473 * simplify-rtx.c (avoid_constant_pool_reference): Check that
28474 offset is non-negative and less than cmode size before
28475 calling simplify_subreg.
28477 2013-02-11 Richard Biener <rguenther@suse.de>
28479 PR tree-optimization/56264
28480 * cfgloop.h (fix_loop_structure): Adjust prototype.
28481 * loop-init.c (fix_loop_structure): Return the number of
28482 newly discovered loops.
28483 * tree-cfgcleanup.c (repair_loop_structures): When new loops
28484 are discovered, do a full loop-closed SSA rewrite.
28486 2013-02-11 Richard Biener <rguenther@suse.de>
28488 PR tree-optimization/56273
28489 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
28491 (check_array_ref): Fix missing newline in dumps.
28492 (search_for_addr_array): Likewise.
28494 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
28496 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
28498 2013-02-09 Jakub Jelinek <jakub@redhat.com>
28501 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
28503 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
28505 PR rtl-optimization/56246
28506 * lra-constraints.c (simplify_operand_subreg): Try to reuse
28508 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
28509 constraints are satisfied.
28511 2013-02-08 Jeff Law <law@redhat.com>
28514 * emit-rtl.c (reg_is_parm_p): New function.
28515 * regs.h (reg_is_parm_p): New prototype.
28516 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
28517 callee-clobbered registers.
28519 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
28522 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
28523 If there is no implicit builtin declaration, just return NULL.
28525 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28527 * config/i386/sse.md (FMAMODEM): New mode iterator.
28528 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
28529 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
28531 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28533 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
28534 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
28535 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
28537 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28539 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
28540 (microblaze*-*-elf): Likewise.
28541 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
28543 * config/microblaze/microblaze-c.c: Add builtin defines for
28544 _LITTLE_ENDIAN and _BIG_ENDIAN.
28545 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
28546 add to TARGET_DEFAULT flags.
28547 Expand ASM_SPEC and LINK_SPEC.
28548 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
28549 * config/microblaze/microblaze.md: Update extendsidi2 and
28550 movdi_internal instructions to use low-order / high-order reg
28552 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
28553 options and inversemask / mask of LITTLE_ENDIAN.
28554 * config/microblaze/t-microblaze: Expand multilib options to
28555 include mlittle-endian (le) and update exceptions patterns.
28557 2013-02-08 Jakub Jelinek <jakub@redhat.com>
28559 PR rtl-optimization/56195
28560 * lra-constraints.c (get_reload_reg): Don't reuse regs
28561 if they have smaller mode than requested, if they have
28562 wider mode than requested, try to return a SUBREG.
28564 PR tree-optimization/56250
28565 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
28566 if type is unsigned and code isn't MULT_EXPR.
28568 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28570 PR tree-optimization/56064
28571 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
28572 bits according to mode.
28573 * fixed-value.h (fixed_from_double_int)
28574 (const_fixed_from_double_int): Adjust comments.
28576 2013-02-08 Richard Biener <rguenther@suse.de>
28579 * lto-streamer.h (struct data_in): Remove current_file, current_line
28580 and current_col members.
28581 * lto-streamer-out.c (lto_output_location): Stream changed bits
28582 en-block for efficiency.
28583 * lto-streamer-in.c (clear_line_info): Remove.
28584 (lto_input_location): Cache current file, line and column
28585 globally via local statics. Read changed bits en-block.
28586 (input_function): Do not call clear_line_info.
28587 (lto_read_body): Likewise.
28588 (lto_input_toplevel_asms): Likewise.
28590 2013-02-08 Michael Matz <matz@suse.de>
28592 PR tree-optimization/52448
28593 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
28594 (nt_call_phase): New static.
28595 (add_or_mark_expr): Only mark accesses with newer phase than any
28597 (nonfreeing_call_p): New.
28598 (nt_init_block): Update nt_call_phase, mark blocks as visited.
28599 (nt_fini_block): Keep blocks marked as visited.
28600 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
28602 2013-02-08 Richard Biener <rguenther@suse.de>
28604 * ira.c (ira): Free broken dominator information.
28606 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
28608 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
28610 2013-02-08 Marek Polacek <polacek@redhat.com>
28612 * cfgloop.c (verify_loop_structure): Add more checking of headers.
28614 2013-02-08 Richard Biener <rguenther@suse.de>
28616 PR middle-end/56181
28617 * cfgloop.h (flow_loops_find): Adjust.
28618 (bb_loop_header_p): Declare.
28619 * cfgloop.c (bb_loop_header_p): New function split out from ...
28620 (flow_loops_find): ... here. Adjust function signature,
28621 support incremental loop structure update.
28622 (verify_loop_structure): Cleanup. Verify a loop is a loop.
28623 * cfgloopmanip.c (fix_loop_structure): Move ...
28624 * loop-init.c (fix_loop_structure): ... here.
28625 (apply_loop_flags): Split out from ...
28626 (loop_optimizer_init): ... here.
28627 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
28628 in incremental mode, only remove dead loops here.
28630 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
28633 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
28634 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
28635 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
28636 (*round<mode>3.libgcc): New insns for fixed-modes.
28637 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
28638 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
28639 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
28640 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
28641 implementations. Define to __builtin_avr_absFX,
28642 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
28643 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
28644 __builtin_avr_countlsFX, respectively.
28645 * config/avr/avr-c.c (target.h): Include it.
28646 (enum avr_builtin_id): New enum.
28647 (avr_resolve_overloaded_builtin): New static function.
28648 (avr_register_target_pragmas): Use it to set
28649 targetm.resolve_overloaded_builtin.
28650 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
28651 tree nodes used by DEF_BUILTIN.
28652 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
28653 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
28654 <AVR_BUILTIN_xxBITS>: Same.
28656 2013-02-08 Richard Biener <rguenther@suse.de>
28658 * cfgloop.c (verify_loop_structure): Properly handle
28659 a loop exiting to another loop header.
28660 * ira-int.h (ira_loops): Remove.
28661 * ira.c (ira_loops): Remove.
28662 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
28663 (do_reload): Use loop_optimizer_finalize.
28664 * ira-build.c (create_loop_tree_nodes): Use get_loops and
28665 number_of_loops to access the loop tree.
28666 (more_one_region_p): Likewise.
28667 (finish_loop_tree_nodes): Likewise.
28668 (rebuild_regno_allocno_maps): Likewise.
28669 (mark_loops_for_removal): Likewise.
28670 (mark_all_loops_for_removal): Likewise.
28671 (remove_unnecessary_regions): Likewise.
28672 (ira_build): Likewise.
28673 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
28675 2013-02-08 Richard Biener <rguenther@suse.de>
28677 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
28678 * ipa-pure-const.c (analyze_function): Avoid calling
28679 mark_irreducible_loops twice.
28680 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
28682 2013-02-07 David S. Miller <davem@davemloft.net>
28684 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
28686 * var-tracking.c (vt_add_function_parameter): Test the presence of
28687 HAVE_window_save properly and do not remap argument registers when
28688 we have a leaf function.
28690 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
28693 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
28695 * config/i386/i386.c (ix86_print_operand): Ditto.
28697 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28699 * lra-constraints.c (process_alt_operands): Fix recently added comment.
28701 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
28703 PR rtl-optimization/56225
28704 * lra-constraints.c (process_alt_operands): Check that reload hard
28705 reg can hold value for strict_low_part.
28707 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28710 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
28711 dwarf2out_end_function.
28712 (in_first_function_p, maybe_at_text_label_p,
28713 first_loclabel_num_not_at_text_label): New variables.
28714 (dwarf2out_var_location): In the first function find out
28715 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
28716 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
28719 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
28721 PR rtl-optimization/56178
28722 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
28723 SUBREG of a register. Tidy up related block of code.
28724 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
28725 note if the source is a register or a SUBREG of a register.
28727 2013-02-07 Jakub Jelinek <jakub@redhat.com>
28730 * config/rs6000/rs6000.md (ptrm): New mode attr.
28731 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
28732 call_value_indirect_aix<pttrsize>,
28733 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
28736 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28738 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
28739 if -bnortl. Convert to strcmp and strncmp.
28741 2013-02-07 Alan Modra <amodra@gmail.com>
28744 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
28745 addresses won't wrap when offsetting.
28746 (rs6000_secondary_reload): Provide secondary reloads needed for
28747 wrapping LO_SUM addresses.
28749 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
28751 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
28752 MACH, just __MACH__.
28754 2013-02-06 Richard Biener <rguenther@suse.de>
28756 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
28757 instead of calling fix_loop_structure.
28759 2013-02-06 Jakub Jelinek <jakub@redhat.com>
28761 PR middle-end/56217
28762 * omp-low.c (use_pointer_for_field): Return false if
28763 lower_send_shared_vars doesn't generate any copy-out code.
28765 2013-02-06 Tom de Vries <tom@codesourcery.com>
28767 PR rtl-optimization/56131
28768 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
28769 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
28770 of the label is NULL. Add comment.
28772 2013-02-05 Jakub Jelinek <jakub@redhat.com>
28774 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
28777 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
28778 (STATIC_LIBTSAN_LIBS): Likewise.
28779 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
28780 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
28781 is defined, don't add anything else beyond that.
28782 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
28783 (LINK_COMMAND_SPEC): Use them.
28785 PR tree-optimization/56205
28786 * tree-stdarg.c (check_all_va_list_escapes): Return true if
28787 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
28788 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
28790 2013-02-05 Richard Biener <rguenther@suse.de>
28792 PR tree-optimization/53342
28793 PR tree-optimization/53185
28794 * tree-vectorizer.h (vect_check_strided_load): Remove.
28795 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
28796 not disallow peeling for vectorized strided loads.
28797 (vect_check_strided_load): Make static and simplify.
28798 (vect_analyze_data_refs): Adjust.
28799 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
28800 correctly when vectorizing strided loads.
28802 2013-02-05 Richard Biener <rguenther@suse.de>
28804 * doc/install.texi: Refer to ISL, not PPL.
28806 2013-02-05 Jan Hubicka <jh@suse.cz>
28808 PR tree-optimization/55789
28809 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
28811 2013-02-05 Jan Hubicka <jh@suse.cz>
28813 PR tree-optimization/55789
28814 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
28815 the dead call anyway.
28817 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
28820 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
28822 2013-02-04 Alexander Potapenko <glider@google.com>
28823 Jack Howarth <howarth@bromo.med.uc.edu>
28824 Jakub Jelinek <jakub@redhat.com>
28827 * config/darwin.c (sort_ctor_records): Stabilized qsort
28828 on constructor priority by using original position.
28829 (finalize_ctors): New routine to sort constructors by
28830 priority before use in assemble_integer.
28831 (machopic_asm_out_constructor): Use finalize_ctors if needed.
28833 2013-02-04 Jakub Jelinek <jakub@redhat.com>
28836 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
28837 about visibility on artificial decls.
28838 * config/sol2.c (solaris_assemble_visibility): Likewise.
28840 2013-02-04 Kai Tietz <ktietz@redhat.com>
28843 * config/i386/i386.c (function_value_ms_64): Add additional valtype
28844 argument and improve checking of return-argument types for 16-byte
28846 (ix86_function_value_1): Add additional valtype argument on call
28847 of function_value_64.
28848 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
28849 handling infunction_value_64 function.
28851 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
28853 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
28855 2013-02-04 Richard Biener <rguenther@suse.de>
28857 PR tree-optimization/56188
28858 * tree-ssa-structalias.c (label_visit): Consider case with
28859 initially non-empty points-to set.
28860 (perform_var_substitution): Dump node mapping and clean up.
28862 2013-02-04 Richard Guenther <rguenther@suse.de>
28865 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
28866 node prevail as last resort.
28867 (lto_symtab_merge_decls): Remove guard on LTRANS here.
28868 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
28870 2013-02-04 Richard Biener <rguenther@suse.de>
28872 PR tree-optimization/56113
28873 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
28875 (equiv_class_lookup_or_add): ... this.
28876 (label_visit): Adjust and fix error in previous patch.
28877 (perform_var_substitution): Adjust.
28879 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
28881 * config/sh/divtab.c: Fix formatting and comments throughout the file.
28882 * config/sh/sh4-300.md: Likewise.
28883 * config/sh/sh4a.md: Likewise.
28884 * config/sh/constraints.md: Likewise.
28885 * config/sh/sh.md: Likewise.
28886 * config/sh/netbsd-elf.h: Likewise.
28887 * config/sh/predicates.md: Likewise.
28888 * config/sh/sh-protos.h: Likewise.
28889 * config/sh/ushmedia.h: Likewise.
28890 * config/sh/linux.h: Likewise.
28891 * config/sh/sh.c: Likewise.
28892 * config/sh/superh.h: Likewise.
28893 * config/sh/elf.h: Likewise.
28894 * config/sh/sh4.md: Likewise.
28895 * config/sh/sh.h: Likewise.
28897 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28899 * config/pa/constraints.md: Adjust unused letters. Change "T"
28900 constraint to match_test floating_point_store_memory_operand().
28901 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
28902 (base14_operand): New.
28903 (floating_point_store_memory_operand): New.
28904 (integer_store_memory_operand): Revise to use base14_operand and
28905 reg_plus_base_memory_operand.
28906 (move_dest_operand): Allow symbolic_memory_operands.
28907 (symbolic_memory_operand): Check for LO_SOM.
28908 (symbolic_operand): Change default case to break.
28909 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
28910 CONST_DOUBLE values to be reloaded by putting them into memory when
28911 the destination is a floating point register.
28912 (movdf): Remove code to handle CONST_DOUBLE.
28914 (reload_indf_r1): New.
28915 (reload_insf_r1): New.
28916 Consistently use "Q" and "T" constraints with integer and floating
28917 point move instructions, respectively.
28918 (movdi): Remove FAIL.
28919 Change predicate for source operand unamed DImode move from
28920 general_operand to move_src_operand.
28921 (umulsidi3): Change predicate for destination operand to
28923 Likewise for similar unamed patterns.
28924 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
28925 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
28926 (hppa_legitimize_address): Simplify mask calculation.
28927 (pa_emit_move_sequence): Revised handling of secondary reloads from
28928 REG+D addresses for floating point loads and stores. Directly handle
28929 loading CONST0_RTX (mode) to a floating point register.
28930 (pa_secondary_reload): Handle reloading DF and SFmode constant values
28931 to floating point registers. Don't restrict secondary reloads to
28932 floating point registers to integer modes. Revise some comments and
28934 (TARGET_LEGITIMATE_ADDRESS_P): Define.
28935 (pa_legitimate_address_p): New.
28936 (pa_legitimize_reload_address): New.
28937 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
28938 (STRICT_REG_OK_FOR_BASE_P): New.
28939 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
28940 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
28942 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
28943 Andrew Dixie <andrewd@gentrack.com>
28945 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
28948 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
28950 * expmed.c (extract_bit_field_1): Pass the full width of the
28951 structure to get_best_reg_extraction_insn.
28953 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28956 * configure.ac (use_cxa_atexit): Add AIX.
28957 * configure: Regenerate.
28959 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
28961 2013-02-01 Jakub Jelinek <jakub@redhat.com>
28964 * final.c (need_profile_function): New variable.
28965 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
28966 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
28967 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
28968 notes, targetm.asm_out.function_prologue doesn't emit anything,
28969 HAVE_prologue and profiler should be emitted before prologue,
28970 set need_profile_function instead of emitting it.
28971 (final_scan_insn): If need_profile_function, emit
28972 profile_function on the first NOTE_INSN_BASIC_BLOCK or
28973 NOTE_INSN_FUNCTION_BEG note.
28975 2013-02-01 Richard Henderson <rth@redhat.com>
28977 * config/rs6000/rs6000.md (smulditi3): New.
28980 * config/alpha/alpha.md (umulditi3): New.
28982 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
28984 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
28985 (ASM_OUTPUT_ALIGNED_LOCAL): New.
28987 2013-02-01 Richard Biener <rguenther@suse.de>
28989 PR tree-optimization/56113
28990 * tree-ssa-structalias.c (label_visit): Reduce work for
28991 single-predecessor nodes.
28993 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
28995 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
28996 range isn't testing for zero.
28998 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
29000 PR middle-end/56113
29001 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
29003 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
29004 Nick Clifton <nickc@redhat.com>
29006 * config/v850/constraints.md (Q): Define as a memory constraint.
29007 * config/v850/predicates.md (label_ref_operand): New predicate.
29008 (e3v5_shift_operand): New predicate.
29009 (ior_operator): New predicate.
29010 * config/v850/t-v850: Add e3v5 multilib.
29011 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
29012 (v850_gen_movdi): Prototype.
29013 * config/v850/v850.c: Add support for e3v5 architecture.
29014 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
29016 (construct_save_jarl): Add e3v5 long JARL support.
29017 (v850_adjust_insn_length): New function. Adjust length of call
29018 insns when using e3v5 instructions.
29019 (v850_gen_movdi): New function: Generate instructions to move a
29021 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
29022 (CPP_SPEC): Define __v850e3v5__ as appropriate.
29023 (TARGET_USE_FPU): Enable for e3v5.
29024 (CONST_OK_FOR_W): New macro.
29025 (ADJUST_INSN_LENGTH): Define.
29026 * config/v850/v850.md (UNSPEC_LOOP): Define.
29027 (attr cpu): Add v850e3v5.
29028 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
29029 (movdi): New pattern.
29030 (movdi_internal): New pattern.
29031 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
29032 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
29033 (cstoresf4): Likewise.
29034 (cstoredf4): Likewise.
29035 (insv): New pattern.
29036 (rotlso3_a): New pattern.
29037 (rotlsi3_b): New pattern
29038 (rotlsi3_v850e3v5): New pattern.
29039 (doloop_begin): New pattern.
29040 (fix_loop_counter): New pattern.
29041 (doloop_end): New pattern.
29042 (branch_normal): Add e3v5 long branch support.
29043 (branch_invert): Likewise.
29044 (branch_z_normal): Likewise.
29045 (branch_z_invert): Likewise.
29046 (branch_nz_normal): Likewise.
29047 (branch_nz_invert): Likewise.
29048 (call_internal_short): Add e3v5 register-indirect JARL support.
29049 (call_internal_long): Likewise.
29050 (call_value_internal_short): Likewise.
29051 (call_value_internal_long): Likewise.
29052 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
29053 (mloop): New option.
29054 * config.gcc: Add support for configuring v840e3v5 target.
29055 * doc/invoke.texi: Document new v850 specific command line options.
29057 2013-01-31 Paul Koning <ni1d@arrl.net>
29061 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
29062 children if parent is a class.
29063 (prune_unused_types_prune): Don't add DW_AT_declaration.
29065 2013-01-31 Richard Biener <rguenther@suse.de>
29067 PR tree-optimization/56157
29068 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
29069 match up operand with SLP child.
29071 2013-01-31 Jason Merrill <jason@redhat.com>
29074 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
29075 parameters the first time.
29076 (gen_scheduled_generic_parms_dies): Check completeness here.
29078 2013-01-31 Richard Biener <rguenther@suse.de>
29080 PR middle-end/53073
29081 * common.opt (faggressive-loop-optimizations): New flag,
29082 enabled by default.
29083 * doc/invoke.texi (faggressive-loop-optimizations): Document.
29084 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
29085 infer_loop_bounds_from_undefined by it.
29087 2013-01-31 Richard Biener <rguenther@suse.de>
29089 PR tree-optimization/56150
29090 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
29091 visit virtual operands.
29092 (find_uses_to_rename_bb): Likewise.
29094 2013-01-31 Richard Biener <rguenther@suse.de>
29096 PR tree-optimization/56150
29097 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
29098 mixed store non-store stmts.
29100 2013-01-30 Jakub Jelinek <jakub@redhat.com>
29103 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
29104 LIBASAN_EARLY_SPEC is defined.
29105 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
29106 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
29108 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
29111 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
29112 invalid args instead of ICEing on it.
29113 (ix86_valid_target_attribute_tree): Return error_mark_node if
29114 ix86_valid_target_attribute_inner_p failed.
29115 (ix86_valid_target_attribute_p): Return false only if
29116 ix86_valid_target_attribute_tree returned error_mark_node. Allow
29117 target("default") attribute.
29118 (sorted_attr_string): Change argument from const char * to tree,
29119 merge in all target attribute arguments rather than just one.
29120 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
29121 instead of free. Avoid using strcat.
29122 (ix86_mangle_function_version_assembler_name): Mangle
29123 target("default") as if no target attribute is present. Adjust
29124 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
29125 instead of xmalloc and XDELETEVEC instead of free.
29126 (ix86_function_versions): Don't return true if one of the decls
29127 doesn't have target attribute. If they don't and one of the decls
29128 is DECL_FUNCTION_VERSIONED, report an error. Adjust
29129 sorted_attr_string caller. Use XDELETEVEC instead of free.
29130 (ix86_supports_function_versions): Remove.
29131 (make_name): Fix up formatting.
29132 (make_dispatcher_decl): Remove resolver_name and its initialization.
29133 Avoid leaking memory.
29134 (is_function_default_version): Return true if there is
29135 target("default") attribute rather than no target attribute at all.
29136 (make_resolver_func): Avoid leaking memory.
29137 (ix86_generate_version_dispatcher_body): Likewise.
29138 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
29139 * target.def (supports_function_versions): Remove.
29140 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
29141 * doc/tm.texi: Regenerated.
29143 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
29145 PR rtl-optimization/56144
29146 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
29147 for values with side effects.
29149 2013-01-30 Richard Biener <rguenther@suse.de>
29151 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
29152 (sparseset_pop): Likewise.
29153 * cfganal.c (compute_idf): Likewise. Increase work-stack size
29154 to be able to use quick_push in the worker loop.
29156 2013-01-30 Marek Polacek <polacek@redhat.com>
29158 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
29160 2013-01-30 Richard Biener <rguenther@suse.de>
29163 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
29165 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
29167 PR tree-optimization/56064
29168 * fixed-value.c (fixed_from_double_int): New function.
29169 * fixed-value.h (fixed_from_double_int): New prototype.
29170 (const_fixed_from_double_int): New static inline function.
29171 * fold-const.c (native_interpret_fixed): New static function.
29172 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
29173 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
29174 (native_encode_fixed): New static function.
29175 (native_encode_expr) <FIXED_CST>: Use it.
29176 (native_interpret_int): Move double_int worker code to...
29177 * double-int.c (double_int::from_buffer): ...this new static method.
29178 * double-int.h (double_int::from_buffer): Prototype it.
29180 2013-01-30 Richard Biener <rguenther@suse.de>
29182 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
29183 New pointer-map and obstack.
29184 (init_alias_vars): Allocate pointer-map and obstack.
29185 (delete_points_to_sets): Free them.
29186 (find_what_var_points_to): Cache result.
29187 (find_what_p_points_to): Adjust for changed interface of
29188 find_what_var_points_to.
29189 (compute_points_to_sets): Likewise.
29190 (ipa_pta_execute): Likewise.
29192 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
29194 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
29195 * configure: Regenerate.
29196 * config.in: Regenerate.
29197 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
29198 #nobits/#progbits if supported.
29200 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
29203 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
29204 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
29205 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
29207 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29209 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
29210 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
29212 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29214 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
29215 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
29217 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29219 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
29221 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
29222 * config/arm/cortex-a7.md: New bypasses using
29223 arm_mac_accumulator_is_result.
29225 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29227 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
29228 (cortex_a7_neon_mla): Likewise.
29229 (cortex_a7_fpfmad): New reservation.
29230 (cortex_a7_fpmacs): Use ffmas and update required units.
29231 (cortex_a7_fpmuld): Update required units and latency.
29232 (cortex_a7_fpmacd): Likewise.
29233 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
29234 (cortex_a7_neon). Likewise.
29235 (bypass) Update participating units.
29237 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
29239 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
29240 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
29242 * config/arm/vfp11.md (vfp_farith): Use ffmas.
29243 (vfp_fmul): Use ffmad.
29244 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
29245 (cortex_r4_fmacd): Use ffmad.
29246 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
29247 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
29248 (cortex_a9_fmacd): Use ffmad.
29249 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
29250 (cortex_a8_vfp_macd): Use ffmad.
29251 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
29252 (cortex_a5_fpmacd): Use ffmad.
29253 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
29254 (cortex_a15_vfp_macd): Use ffmad.
29255 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29257 2013-01-29 Jason Merrill <jason@redhat.com>
29260 * varasm.c (default_assemble_visibility): Don't warn about
29261 visibility on artificial decls.
29263 2013-01-29 Richard Biener <rguenther@suse.de>
29265 PR tree-optimization/56113
29266 * tree-ssa-structalias.c (equiv_class_lookup): Also return
29268 (label_visit): Free duplicate bitmaps and record the leader instead.
29269 (perform_var_substitution): Adjust.
29271 2013-01-29 Richard Biener <rguenther@suse.de>
29273 PR tree-optimization/55270
29274 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
29275 the CFG, schedule loops for fixup.
29277 2013-01-29 Nick Clifton <nickc@redhat.com>
29279 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
29282 2013-01-28 Leif Ekblad <leif@rdos.net>
29284 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
29285 * config/i386/i386.h (TARGET_RDOS): New macro.
29286 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
29287 * config/i386/i386.c (ix86_option_override_internal): For 64bit
29288 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
29289 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
29290 DEFAULT_LARGE_SECTION_THRESHOLD.
29291 * config/i386/i386.md (R14_REG, R15_REG): New constants.
29292 * config/i386/rdos.h: New file.
29293 * config/i386/rdos64.h: New file.
29295 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
29298 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
29301 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29303 PR rtl-optimization/56117
29304 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
29305 call cselib_lookup_from_insn on the MEM before calling
29306 add_insn_mem_dependence.
29308 2013-01-28 Richard Biener <rguenther@suse.de>
29310 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
29311 to a stmt that didn't have one.
29312 (copy_phis_for_bb): Likewise for PHI arguments.
29313 (copy_debug_stmt): Likewise for debug stmts.
29315 2013-01-28 Richard Biener <rguenther@suse.de>
29317 PR tree-optimization/56034
29318 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
29319 (partition_builtin_p): Adjust.
29320 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
29321 it is the last partition.
29322 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
29323 up the vertex for the definition.
29324 (classify_partition): Classify whether a partition is a
29325 PKIND_REDUCTION, thus has uses outside of the loop.
29326 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
29327 Merge all PKIND_REDUCTION partitions into the last partition.
29328 (tree_loop_distribution): Seed partitions from reductions as well.
29330 2013-01-28 Jakub Jelinek <jakub@redhat.com>
29332 PR tree-optimization/56125
29333 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
29334 pow(x,c) into sqrt(x) * powi(x, n/2) or
29335 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
29336 optimizing for size.
29337 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
29338 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
29341 PR tree-optimization/56094
29342 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
29343 to UNKNOWN_LOCATION while gimplifying expr.
29345 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
29348 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
29349 operand 0 in movabs insn template for -masm=intel asm alternative.
29350 (*movabs<mode>_2): Ditto for operand 1.
29352 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
29355 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
29358 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29360 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
29363 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29365 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
29366 Undef to avoid warning.
29368 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
29370 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
29371 * configure: Regenerate.
29373 2013-01-25 Jakub Jelinek <jakub@redhat.com>
29375 PR tree-optimization/56098
29376 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
29377 for stmts with volatile ops.
29378 (cond_store_replacement): Don't optimize if assign has volatile ops.
29379 (cond_if_else_store_replacement_1): Don't optimize if either
29380 then_assign or else_assign have volatile ops.
29381 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
29384 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29386 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
29388 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29390 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
29391 missing ':' in asm example.
29393 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
29395 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
29396 entries into lane and laneq entries.
29397 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
29398 Remove AdvSIMD scalar modes.
29399 (aarch64_sq<r>dmulh_laneq<mode>): New.
29400 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
29402 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
29403 builtin implementations to relfect changes in RTL in aarch64-simd.md.
29404 * config/aarch64/iterators.md (VCOND): New.
29407 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
29410 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
29411 Add NULL LIBNAME argument to existing definitions.
29412 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
29413 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
29414 * config/avr/avr.c (DEF_BUILTIN): Same.
29415 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
29416 (avr_expand_builtin): Expand to a vanilla call if a libgcc
29417 implementation is available (DECL_ASSEMBLER_NAME is set).
29418 (avr_fold_absfx): New static function.
29419 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
29420 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
29421 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
29422 AVR_BUILTIN_ABSLLK.
29423 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
29424 (abshk, absk, abslk, absllk): Provide as static inline functions.
29426 2013-01-25 Marek Polacek <polacek@redhat.com>
29428 PR tree-optimization/56035
29429 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
29431 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
29433 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
29434 (*movtf_internal_rex64): Add (!o,C) alternative
29435 (*movxf_internal_rex64): Ditto.
29436 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
29438 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
29440 * doc/invoke.texi: fix typo.
29441 * doc/objc.texi: fix typo.
29443 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
29445 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
29446 for the first two alternatives.
29448 2013-01-24 Diego Novillo <dnovillo@google.com>
29450 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
29451 (ggc-zone.o): Remove.
29452 * configure.ac: Remove option --with-gc.
29453 * configure: Re-generate.
29454 * doc/install.texi: Remove documentation for --with-gc.
29455 * gengtype.c (write_enum_defn): Remove. Update all users.
29456 (write_Types_process_field): Remove generation of gt_e_* argument.
29457 (output_type_enum): Remove. Update all users.
29458 (write_enum_defn): Remove. Update all users.
29459 (enum alloc_zone): Remove. Update all users.
29460 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
29461 * ggc-common.c (ggc_splay_alloc): Remove first argument.
29462 Update all callers.
29463 (struct ptr_data): Remove field TYPE. Update all users.
29464 (gt_pch_note_object): Remove argument TYPE. Update all users.
29465 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
29467 * ggc-none.c (ggc_alloc_typed_stat): Remove.
29468 (struct alloc_zone): Remove.
29469 (ggc_internal_alloc_zone_stat): Remove.
29470 (ggc_internal_cleared_alloc_zone_stat): Remove.
29471 * ggc-page.c (ggc_alloc_typed_stat): Remove.
29472 (ggc_pch_count_object): Remove last argument. Update all users.
29473 (ggc_pch_alloc_object): Remove last argument. Update all users.
29474 (struct alloc_zone): Remove.
29475 * ggc-zone.c: Remove.
29476 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
29477 (struct alloc_zone): Remove.
29478 (ggc_alloc_typed_stat): Remove.
29479 (ggc_alloc_typed): Remove.
29480 (ggc_splay_alloc): Remove first argument.
29481 (rtl_zone): Remove. Update all users.
29482 (tree_zone): Remove. Update all users.
29483 (tree_id_zone): Remove. Update all users.
29484 (ggc_internal_zone_alloc_stat): Remove. Update all users.
29485 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
29486 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
29487 * tree-ssanames.c: Remove references to zone allocator in comments.
29489 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
29491 * config/avr/avr.c (avr_out_fract): Make register numbers that
29492 might be outside of source operand signed.
29494 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
29496 * config/i386/constraints.md (Yf): New constraint.
29497 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
29498 of f constraint to conditionaly disable x87 register preferences.
29499 (*movdf_internal): Ditto.
29500 (*movsf_internal): Ditto.
29502 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
29504 PR inline-asm/55934
29505 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
29506 that have operands with impossible constraints.
29507 Add a FIXME for a speed-up opportunity.
29508 * lra-constraints.c (process_alt_operands): Verify that a class
29509 selected from constraints on asms is valid for the operand mode.
29510 (curr_insn_transform): Remove incorrect comment.
29512 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
29514 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
29515 TOC operand is a valid symbol ref in the constant pool.
29517 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
29519 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
29521 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
29524 * config/avr/stdfix.h: New file.
29525 * t-avr (stdfix-gcc.h): New rule to build it.
29526 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
29528 2013-01-23 Kostya Serebryany <kcc@google.com>
29530 * config/darwin.h: remove dependency on
29531 CoreFoundation (asan on Mac OS).
29533 2013-01-23 Jakub Jelinek <jakub@redhat.com>
29536 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
29537 instead of cmpdi_operand for first comparison operand.
29538 Don't assert that comparison operands aren't both constants.
29540 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29542 * doc/install.texi (Downloading the Source): Update references to
29543 downloading separate components.
29545 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
29547 * doc/extend.texi (__int128): Improve grammar.
29549 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
29552 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
29553 alternative to (o,r).
29554 (*movdi_internal_rex64): Remove (!o,n) alternative.
29555 (DImode immediate->memory splitter): Remove.
29556 (DImode immediate->memory peephole2): Remove.
29557 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
29558 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
29559 alternative to (!o,*r).
29560 (*movtf_internal_sse): New pattern.
29561 (*movxf_internal_rex64): New pattern.
29562 (*movxf_internal): Disable for TARGET_64BIT.
29563 (*movdf_internal_rex64): Remove (!o,F) alternative.
29565 2013-01-22 Jakub Jelinek <jakub@redhat.com>
29567 PR middle-end/56074
29568 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
29569 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
29570 * tree-vect-loop-manip.c (find_loop_location): Also ignore
29571 stmt locations where LOCATION_LOCUS of the stmt location is
29572 UNKNOWN_LOCATION or BUILTINS_LOCATION.
29575 * config/i386/i386.md (UNSPEC_STOS): New.
29576 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
29577 *strsetqi_1): Add UNSPEC_STOS.
29579 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
29582 * doc/invoke.texi: Remove left over -Wsynth example.
29584 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29586 PR tree-optimization/56051
29587 * fold-const.c (fold_binary_loc): Don't fold
29588 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
29589 a narrowing conversion, or widening conversion from signed
29592 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29594 PR rtl-optimization/56023
29595 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
29596 dependent on debug instruction.
29598 2013-01-21 Martin Jambor <mjambor@suse.cz>
29600 PR middle-end/56022
29601 * function.c (allocate_struct_function): Call
29602 invoke_set_current_function_hook earlier.
29604 2013-01-21 Jakub Jelinek <jakub@redhat.com>
29606 * reload1.c (init_reload): Only initialize reload_obstack
29607 during the first call.
29609 2013-01-21 Marek Polacek <polacek@redhat.com>
29611 * cfgloop.c (verify_loop_structure): Fix up grammar.
29613 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
29615 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
29616 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
29618 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29621 * config/arm/marvell-pj4.md: Update copyright year.
29622 Fix up use of alu to alu_reg and simple_alu_imm.
29624 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
29626 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
29628 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
29631 * lra-constraints.c (curr_insn_transform): Don't reuse original
29632 insn for secondary memory move when memory mode should be different.
29634 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29636 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
29637 atomic_storedi_1): New patterns.
29639 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29641 btver2 pipeline descriptions.
29642 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
29644 * config/i386/i386.md (btver2_decode): New type attributes.
29645 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
29647 * config/i386/btver2.md: New file describing btver2 pipelines.
29649 2013-01-19 Andrew Pinski <apinski@cavium.com>
29651 PR tree-optimization/52631
29652 * tree-ssa-sccvn (visit_use): Before looking up the original
29653 statement, try looking up the simplified expression.
29655 2013-01-19 Anthony Green <green@moxielogic.com>
29657 * config/moxie/moxie.c (moxie_expand_prologue): Set
29658 current_function_static_stack_size.
29660 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29662 PR tree-optimization/56029
29663 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
29664 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
29666 2013-01-18 Sharad Singhai <singhai@google.com>
29668 PR tree-optimization/55995
29669 * dumpfile.c (dump_loc): Print location only if available.
29670 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
29672 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
29675 * lra-constraints.c (curr_insn_transform): Reuse original insn for
29676 secondary memory move.
29677 (inherit_reload_reg): Use rclass instead of cl for
29678 check_secondary_memory_needed_p.
29680 2013-01-18 Jakub Jelinek <jakub@redhat.com>
29682 PR middle-end/56015
29683 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
29684 the case where writing real complex part of target modifies op1.
29686 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
29688 * config/aarch64/aarch64-simd.md
29689 (aarch64_vcond_internal<mode>): Handle unordered cases.
29690 * config/aarch64/iterators.md (v_cmp_result): New.
29692 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
29693 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
29695 * config/arm/marvell-pj4.md: New file.
29696 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
29697 * config/arm/arm.md (generic_sched): Add marvell_pj4.
29698 (generic_vfp): Likewise.
29699 * config/arm/arm-cores.def: Add marvell-pj4.
29700 * config/arm/arm-tune.md: Regenerate.
29701 * config/arm/arm-tables.opt: Regenerate.
29702 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
29703 * doc/invoke.texi: Document marvell-pj4.
29705 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
29707 * config/aarch64/arm_neon.h: Map scalar types to standard types.
29709 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29714 * var-tracking.c (negative_power_of_two_p): New.
29715 (global_get_addr_cache, local_get_addr_cache): New.
29716 (get_addr_from_global_cache, get_addr_from_local_cache): New.
29717 (vt_canonicalize_addr): Rewrite using the above. Adjust the
29719 (vt_stack_offset_p): Remove.
29720 (vt_canon_true_dep): Always canonicalize loc's address.
29721 (clobber_overlapping_mems): Make sure we have a MEM.
29722 (local_get_addr_clear_given_value): New.
29723 (val_reset): Clear local cached entries.
29724 (compute_bb_dataflow): Create and release the local cache.
29725 Disable duplicate MEMs clobbering.
29726 (emit_notes_in_bb): Clobber MEMs likewise.
29727 (vt_emit_notes): Create and release the local cache.
29728 (vt_initialize, vt_finalize): Create and release the global
29729 cache, respectively.
29730 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
29732 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29734 PR libmudflap/53359
29735 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
29736 not found in the symtab.
29738 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
29741 PR rtl-optimization/55547
29742 PR rtl-optimization/53827
29745 * alias.c (offset_overlap_p): New, factored out of...
29746 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
29747 the conservative special case for symbolic constants. Don't
29748 adjust zero sizes on alignment.
29750 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
29752 PR rtl-optimization/52573
29753 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
29754 REG_UNUSED for the same register.
29756 2013-01-17 Richard Biener <rguenther@suse.de>
29757 Marek Polacek <polacek@redhat.com>
29759 PR rtl-optimization/55833
29760 * loop-unswitch.c (unswitch_loops): Move loop verification...
29761 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
29762 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
29763 Set it to true when we're removing a loop from hierarchy tree in
29764 an irreducible region.
29765 (fix_bb_placements): Adjust caller.
29766 (fix_loop_placements): Likewise.
29768 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
29770 * config/avr/builtins.def (DEF_BUILTIN): Factor out
29771 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
29772 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
29773 Remove ID. Adjust comments.
29774 * config/avr/avr-c.c (avr_builtin_name): Remove.
29775 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
29776 * config/avr/avr.c (avr_tolower): New static function.
29777 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
29778 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
29779 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
29782 2013-01-17 Jan Hubicka <jh@suse.cz>
29784 PR tree-optimization/55273
29785 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
29787 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
29790 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
29791 store through atomic_store<mode>_1.
29792 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
29794 2013-01-17 Martin Jambor <mjambor@suse.cz>
29796 PR tree-optimizations/55264
29797 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
29798 for virtual methods.
29799 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
29800 virtual methods before inlining is over.
29801 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
29803 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
29806 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
29808 PR rtl-optimization/56005
29809 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
29810 pending reads for prefetch.
29812 2013-01-16 Ian Bolton <ian.bolton@arm.com>
29814 * config/aarch64/aarch64.md
29815 (*cstoresi_neg_uxtw): New pattern.
29816 (*cmovsi_insn_uxtw): New pattern.
29817 (*<optab>si3_uxtw): New pattern.
29818 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
29819 (*<optab>si3_insn_uxtw): New pattern.
29820 (*bswapsi2_uxtw): New pattern.
29822 2013-01-16 Richard Biener <rguenther@suse.de>
29824 * tree-inline.c (tree_function_versioning): Remove set but
29825 never used variable.
29827 2013-01-16 Richard Biener <rguenther@suse.de>
29829 PR tree-optimization/55964
29830 * tree-flow.h (rename_variables_in_loop): Remove.
29831 (rename_variables_in_bb): Likewise.
29832 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
29833 (copy_loop_before): Adjust and delete update-ssa status.
29834 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
29835 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
29836 (rename_variables_in_loop): Remove.
29837 (slpeel_update_phis_for_duplicate_loop): Likewise.
29838 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
29839 use available cfg machinery instead of duplicating it.
29840 Update PHI nodes and perform poor-mans SSA update here.
29841 (slpeel_tree_peel_loop_to_edge): Adjust.
29843 2013-01-16 Richard Biener <rguenther@suse.de>
29845 PR tree-optimization/54767
29846 PR tree-optimization/53465
29847 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
29848 (vrp_visit_phi_node): For PHI arguments coming via backedges
29849 drop all symbolical range information.
29850 (execute_vrp): Compute backedges.
29852 2013-01-16 Richard Biener <rguenther@suse.de>
29854 * doc/install.texi: Update CLooG and ISL requirements to
29857 2013-01-16 Christian Bruel <christian.bruel@st.com>
29860 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
29861 (broken_move): Handle UNSPECV_SP_SWITCH_B.
29862 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
29864 2013-01-16 DJ Delorie <dj@redhat.com>
29866 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
29867 (UNSPECV_SP_SWITCH_E): New.
29868 (sp_switch_1): Change to an unspec.
29869 (sp_switch_2): Change to an unspec. Don't use post-inc when we
29872 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
29874 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
29875 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
29876 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
29877 (expand_mem_thread_fence): Ditto.
29878 (expand_mem_signal_fence): Ditto.
29879 (expand_atomic_load): Ditto.
29880 (expand_atomic_store): Ditto.
29882 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
29884 PR rtl-optimization/55547
29885 PR rtl-optimization/53827
29888 * alias.c (memrefs_conflict_p): Set sizes to negative after
29891 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29894 * function.c (thread_prologue_and_epilogue_insns): Always
29895 add crtl->drap_reg to set_up_by_prologue.set, even if
29896 stack_realign_drap is false.
29898 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
29900 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
29901 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
29902 *call): Fix indention.
29904 2013-01-15 Tom de Vries <tom@codesourcery.com>
29907 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
29910 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
29912 PR rtl-optimization/55153
29913 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
29915 2013-01-15 Martin Jambor <mjambor@suse.cz>
29917 PR tree-optimization/55920
29918 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
29919 accesses as grp_to_be_debug_replaced.
29921 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29923 PR tree-optimization/55920
29924 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
29925 there is non-useless type conversion needed from debug rhs to lhs,
29926 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
29928 2013-01-15 Joseph Myers <joseph@codesourcery.com>
29929 Mikael Pettersson <mikpe@it.uu.se>
29932 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
29934 (ASM_OUTPUT_CASE_LABEL): Remove.
29935 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
29936 * final.c (shorten_branches): Update alignment of labels before
29937 jump tables if CASE_VECTOR_SHORTEN_MODE.
29939 2013-01-15 Richard Biener <rguenther@suse.de>
29942 * system.h: Do not include gmp.h for building host tools.
29944 2013-01-15 Richard Biener <rguenther@suse.de>
29946 PR middle-end/55882
29947 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
29948 account for bitpos when computing alignment.
29950 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
29952 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
29953 (ix86_target_macros_internal): Likewise.
29955 * config/i386/i386.c (m_CORE2I7): Removed.
29956 (m_CORE_HASWELL): New macro.
29957 (m_CORE_ALL): Likewise.
29958 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
29959 (initial_ix86_arch_features): Likewise.
29960 (processor_target_table): Initializations for Core avx2.
29961 (cpu_names): New names "core-avx2".
29962 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
29963 PROCESSOR_CORE_HASWELL.
29964 (ix86_issue_rate): New case.
29965 (ia32_multipass_dfa_lookahead): Likewise.
29966 (ix86_sched_init_global): Likewise.
29968 * config/i386/i386.h (TARGET_HASWELL): New macro.
29969 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
29970 (processor_type): New PROCESSOR_HASWELL.
29972 2013-01-15 Jakub Jelinek <jakub@redhat.com>
29974 PR tree-optimization/55955
29975 * tree-vect-loop.c (vectorizable_reduction): Give up early on
29976 *SHIFT_EXPR and *ROTATE_EXPR codes.
29978 PR tree-optimization/48766
29979 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
29980 -ftrapv disable -fwrapv.
29982 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29985 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
29986 etc. to 1 and not to __flash.
29987 Use LL suffix for __INT24_MAX__ with -mint8.
29988 Use ULL suffix for __UINT24_MAX__ with -mint8.
29990 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
29992 * config/avr/avr-arch.h
29993 (struct base_arch_s): Use typedef avr_arch_t instead.
29994 (struct arch_info_s): Use typedef avr_arch_info_t instead.
29995 (struct mcu_type_s): Use typedef avr_mcu_t instead.
29996 * config/avr/avr.c: Same.
29997 * config/avr/avr-devices.c: Same.
29998 * config/avr/driver-avr.c: Same.
29999 * config/avr/gen-avr-mmcu-texi.c: Same.
30000 * config/avr/avr-mcus.def: Adjust comment.
30002 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
30004 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
30005 * config/aarch64/iterators.md (VALLDI): New.
30007 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
30008 Andi Kleen <ak@linux.intel.com>
30011 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
30012 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
30015 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30017 * config/avr/avr-stdint.h: Remove trailing blanks.
30018 * config/avr/avr-log.h: Same.
30019 * config/avr/avr-arch.h: Same.
30020 * config/avr/avr-devices.c: Same.
30021 * config/avr/avr-dimode.md: Same.
30022 * config/avr/predicates.md: Same.
30023 * config/avr/avr-c.c: Same. And fix typo.
30025 * config/avr/avr-protos.h: Same. And:
30026 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
30027 (init_cumulative_args): Rename to avr_init_cumulative_args.
30028 (expand_prologue): Rename to avr_expand_prologue.
30029 (expand_epilogue): Rename to avr_expand_epilogue.
30030 (adjust_insn_length): Rename to avr_adjust_insn_length.
30031 (notice_update_cc): Rename to avr_notice_update_cc.
30032 (final_prescan_insn): Rename to avr_final_prescan_insn.
30033 * config/avr/avr.c: Same.
30034 * config/avr/avr.h: Same.
30035 * config/avr/avr.md: Remove trailing blanks.
30036 (prologue): Use avr_expand_prologue.
30037 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
30039 2013-01-14 Richard Biener <rguenther@suse.de>
30041 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
30042 verify_location, collect_subblocks): New functions.
30043 (verify_gimple_in_cfg): Verify that locations only reference
30044 BLOCKs in the functions BLOCK tree.
30046 2013-01-14 Richard Biener <rguenther@suse.de>
30048 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
30050 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
30052 (insert_out_of_ssa_copy_on_edge): Likewise.
30053 (rewrite_close_phi_out_of_ssa): Likewise.
30054 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
30056 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
30057 propagated constants.
30058 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
30061 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
30063 * config/avr/avr-modes.def: Add GPL copyright notice.
30065 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
30067 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
30068 MEMMODEL_MASK to determine memory model.
30069 (atomic_store<mode>): Ditto from operands[2].
30070 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
30072 2013-01-13 Jakub Jelinek <jakub@redhat.com>
30075 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
30076 (fold_gimple_assign): Don't call unshare_expr here.
30077 (fold_ctor_reference): Call unshare_expr.
30079 2013-01-13 Terry Guo <terry.guo@arm.com>
30081 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
30082 * doc/fragments.texi: Document MULTILIB_REUSE.
30083 * gcc.c (multilib_reuse): New internal spec.
30084 (set_multilib_dir): Also search multilib from multilib_reuse.
30085 * genmultilib (tmpmultilib3): Refactor code.
30086 (tmpmultilib4): Ditto.
30087 (multilib_reuse): New multilib argument.
30089 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
30091 * Makefile.in: Update copyright.
30093 2013-01-12 Tom de Vries <tom@codesourcery.com>
30095 PR middle-end/55890
30096 * calls.c (expand_call): Check if arg_nr is valid.
30098 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
30100 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
30101 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
30102 documentation. Add missing '__' in front of
30103 __builtin_ia32_packssdw256.
30105 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30108 * config/s390/s390.c (s390_preferred_reload_class): Do not return
30109 NO_REGS for larl operands.
30110 (s390_reload_larl_operand): Use s390_load_address instead of
30113 2013-01-11 Richard Biener <rguenther@suse.de>
30115 * tree-cfg.c (verify_node_sharing_1): Split out from ...
30116 (verify_node_sharing): ... here.
30117 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
30119 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
30121 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
30122 Substitute TREECHECKING.
30123 * configure: Regenerate.
30124 * Makefile.in (TREECHECKING): New.
30126 2013-01-11 Richard Guenther <rguenther@suse.de>
30128 PR tree-optimization/44061
30129 * tree-vrp.c (extract_range_basic): Compute zero as
30130 value-range for __builtin_constant_p of function parameters.
30132 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
30134 Update copyright years.
30136 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
30138 PR rtl-optimization/55672
30139 * lra-eliminations.c (mark_not_eliminable): Permit addition with
30140 const to be eliminable.
30142 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
30144 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
30145 * configure: Regenerate.
30147 2013-01-10 Richard Biener <rguenther@suse.de>
30149 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
30151 2013-01-10 Richard Biener <rguenther@suse.de>
30154 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
30155 locations for virtual PHI arguments.
30156 (rewrite_update_phi_arguments): Likewise.
30158 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
30160 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
30163 2013-01-10 Jakub Jelinek <jakub@redhat.com>
30165 PR tree-optimization/55921
30166 * tree-complex.c (expand_complex_asm): New function.
30167 (expand_complex_operations_1): Call it for GIMPLE_ASM.
30169 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
30172 * config/s390/s390.c (s390_symref_operand_p)
30173 (s390_loadrelative_operand_p): Merge the two functions.
30174 (s390_check_qrst_address, print_operand_address): Add parameters
30175 to s390_loadrelative_operand_p invokation.
30176 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
30177 (s390_reload_larl_operand, s390_secondary_reload): Use
30178 s390_loadrelative_operand_p instead of s390_symref_operand_p.
30179 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
30181 2013-01-09 Mike Stump <mikestump@comcast.net>
30183 * dse.c (record_store): Remove unnecessary assert.
30185 2013-01-09 Jan Hubicka <jh@suse.cz>
30187 PR tree-optimization/55569
30188 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
30189 * cfgloop.h (scale_loop_profile): Likewise.
30191 2013-01-09 Jan Hubicka <jh@suse.cz>
30194 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
30196 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
30198 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
30200 PR middle-end/55114
30201 * expr.h (maybe_emit_group_store): Declare.
30202 * expr.c (maybe_emit_group_store): New function.
30203 * builtins.c (expand_builtin_int_roundingfn): Call it.
30204 (expand_builtin_int_roundingfn_2): Likewise.
30206 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
30208 PR rtl-optimization/55829
30209 * lra-constraints.c (match_reload): Add code for absent output.
30210 (curr_insn_transform): Add code for reloads of matched inputs
30213 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
30215 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
30216 attribute of movddup insn to DF.
30217 (*vec_interleave_lowv2df): Ditto.
30218 (vec_dupv2df): Ditto.
30220 2013-01-09 Jan Hubicka <jh@suse.cz>
30222 PR tree-optimiation/55875
30223 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
30224 EVERY_ITERATION parameter.
30225 (number_of_iterations_exit): Check if exit is executed every iteration.
30226 (idx_infer_loop_bounds): Similarly here.
30227 (n_of_executions_at_most): Simplify
30228 to only test for cases where statement is dominated by the
30229 particular bound; handle correctly the "postdominance" test.
30230 (scev_probably_wraps_p): Use max loop iterations info
30231 as a global bound first.
30233 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
30234 Nick Clifton <nickc@redhat.com>
30236 * config/v850/v850.md (cbranchsf4): New pattern.
30237 (cstoresf4): New pattern.
30238 (cbranchdf4): New pattern.
30239 (cstoredf4): New pattern.
30240 (movsicc): Disallow floating point comparisons.
30241 (cmpsf_le_insn): Fix order of operators.
30242 (cmpsf_lt_insn): Likewise.
30243 (cmpsf_eq_insn): Likewise.
30244 (cmpdf_le_insn): Likewise.
30245 (cmpdf_lt_insn): Likewise.
30246 (cmpdf_eq_insn): Likewise.
30247 (cmpsf_ge_insn): Use LE comparison.
30248 (cmpdf_ge_insn): Likewise.
30249 (cmpsf_gt_insn): Use LT comparison.
30250 (cmpdf_gt_insn): Likewise.
30251 (cmpsf_ne_insn): Delete pattern.
30252 (cmpdf_ne_insn): Delete pattern.
30253 * config/v850/v850.c (v850_gen_float_compare): Use
30254 gen_cmpdf_eq_insn for NE comparison.
30255 (v850_float_z_comparison_operator)
30256 (v850_float_nz_comparison_operator): Move from here ...
30257 * config/v850/predicates.md: ... to here. Move GT and GE
30258 comparisons into v850_float_z_comparison_operator.
30259 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
30261 (v850_float_nz_comparison_operator): Likewise.
30263 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30265 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
30266 with calls to gen_insvsi/gen_insvdi.
30268 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
30270 * config/i386/i386.c (initial_ix86_tune_features): Set up
30271 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
30273 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
30274 Jakub Jelinek <jakub@redhat.com>
30276 PR tree-optimization/48189
30277 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
30278 If nitercst is 0, don't predict the exit edge.
30280 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
30282 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
30283 in asm_fprintf with reg_names.
30284 (aarch64_print_operand_address): Likewise.
30285 (aarch64_return_addr): Likewise.
30286 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
30288 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30290 * config/pa/pa.h (VAL_U6_BITS_P): Define.
30291 (INT_U6_BITS): Likewise.
30292 * config/pa/predicates.md (uint6_operand): New predicate.
30293 (shift5_operand, shift6_operand): Likewise.
30294 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
30296 (lshrdi3): Use shift6_operand.
30297 (shrpsi4, shrpdi4): New insn patterns.
30298 (extzv): Delete expander.
30299 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
30300 predicates in unamed zero extract patterns. Tighten common constraint.
30301 (extv): Delete expander.
30302 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
30303 predicates in unamed sign extract patterns. Tighten common constraint.
30304 (insv): Delete expander.
30305 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
30306 predicates in unamed insert patterns. Tighten common constraint.
30307 Change uint32_operand predicate to uint6_operand predicate in unamed
30308 DImode pattern to insert constant values of type 1...1xxxx.
30310 2013-01-04 Jan Hubicka <jh@suse.cz>
30312 PR tree-optimization/55823
30313 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
30316 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30317 Uros Bizjak <ubizjak@gmail.com>
30319 PR rtl-optimization/55845
30320 * df-problems.c (can_move_insns_across): Stop scanning at
30321 volatile_insn_p source instruction or give up if
30322 across_from .. across_to range contains any volatile_insn_p
30325 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30327 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
30328 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
30330 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
30331 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
30333 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30336 * asan.c (asan_clear_shadow): New function.
30337 (asan_emit_stack_protection): Use it.
30339 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
30341 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
30342 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
30343 with tab instead of space.
30345 2013-01-08 Nick Clifton <nickc@redhat.com>
30347 * config/rl78/rl78.c (rl78_expand_prologue): Always select
30348 register bank 0 at the start of an interrupt handler.
30349 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
30352 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30354 * config/aarch64/aarch64-simd.md
30355 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
30356 (aarch64_simd_bsl): Likewise.
30357 (aarch64_vcond_internal<mode>): Likewise.
30358 (vcond<mode><mode>): Likewise.
30359 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
30360 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
30362 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
30364 * config/aarch64/aarch64-builtins.c
30365 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
30367 2013-01-08 Martin Jambor <mjambor@suse.cz>
30370 * tree-sra.c (analyze_access_subtree): Return true also after
30371 potentially creating a debug-only replacement.
30373 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30375 PR middle-end/55890
30376 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
30378 PR tree-optimization/54120
30379 * tree-vrp.c (range_fits_type_p): Don't allow
30380 src_precision < precision from signed vr to unsigned_p
30381 if vr->min or vr->max is negative.
30382 (simplify_float_conversion_using_ranges): Test can_float_p
30383 against CODE_FOR_nothing.
30385 2013-01-08 Jakub Jelinek <jakub@redhat.com>
30386 Richard Biener <rguenther@suse.de>
30388 PR middle-end/55851
30389 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
30390 types instead of just INTEGER_TYPE types.
30392 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
30394 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
30397 2013-01-07 Steve Ellcey <sellcey@mips.com>
30400 * config/mips/mips.opt: Change mad to mmad to match documentation.
30402 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30405 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
30406 .progmemx.data now.
30408 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30411 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
30412 (avr_addrspace_t): Add .section_name field.
30413 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
30415 (avr_addrspace): Same. Initialize .section_name. Remove last
30416 NULL entry. Put __memx into .progmemx.data.
30417 (progmem_section_prefix): Remove.
30418 (avr_asm_init_sections): No need to initialize progmem_section.
30419 (avr_asm_named_section): Use avr_addrspace[].section_name to get
30420 section name prefix.
30421 (avr_asm_select_section): Ditto. And use get_unnamed_section to
30422 retrieve the progmem section.
30423 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
30424 boundary to run over avr_addrspace[].
30425 (avr_register_target_pragmas): Ditto.
30427 2013-01-06 Jakub Jelinek <jakub@redhat.com>
30429 * varasm.c (output_constant_def_contents): For asan_protect_global
30430 protected strings, adjust DECL_ALIGN if needed, before testing for
30432 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
30433 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
30435 (output_object_block): For asan protected decls, emit asan padding
30436 after their contents.
30437 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
30438 (asan_finish_file): Test it here instead.
30440 2013-01-07 Nick Clifton <nickc@redhat.com>
30441 Matthias Klose <doko@debian.org>
30442 Doug Kwan <dougkwan@google.com>
30443 H.J. Lu <hongjiu.lu@intel.com>
30446 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
30448 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
30450 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
30452 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
30454 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
30456 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30459 * doc/install.texi (Cross-Compiler-Specific Options): Document
30462 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
30464 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
30465 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
30466 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
30467 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
30468 vqmovun_high_s64): Fix source operand number and update copyright.
30470 2013-01-07 Richard Biener <rguenther@suse.de>
30472 PR middle-end/55890
30473 * gimple.h (gimple_call_builtin_p): New overload.
30474 * gimple.c (validate_call): New function.
30475 (gimple_call_builtin_p): Likewise.
30476 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
30477 Use gimple_call_builtin_p.
30478 (find_func_clobbers): Likewise.
30479 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
30480 (strlen_optimize_stmt): Likewise.
30482 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30484 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
30485 (vld1q_dup_*): Likewise.
30486 (vld1_*): Likewise.
30487 (vld1q_*): Likewise.
30488 (vld1_lane_*): Likewise.
30489 (vld1q_lane_*): Likewise.
30491 2013-01-07 Richard Biener <rguenther@suse.de>
30493 * lto-streamer.h (LTO_minor_version): Bump to 2.
30495 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
30497 * config/aarch64/aarch64-protos.h
30498 (aarch64_const_double_zero_rtx_p): Rename to...
30499 (aarch64_float_const_zero_rtx_p): ...this.
30500 (aarch64_float_const_representable_p): New.
30501 (aarch64_output_simd_mov_immediate): Likewise.
30502 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
30503 move immediate case.
30504 * config/aarch64/aarch64.c
30505 (aarch64_const_double_zero_rtx_p): Rename to...
30506 (aarch64_float_const_zero_rtx_p): ...this.
30507 (aarch64_print_operand): Allow printing of new constants.
30508 (aarch64_valid_floating_const): New.
30509 (aarch64_legitimate_constant_p): Check for valid floating-point
30511 (aarch64_simd_valid_immediate): Likewise.
30512 (aarch64_vect_float_const_representable_p): New.
30513 (aarch64_float_const_representable_p): Likewise.
30514 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
30515 (aarch64_output_simd_mov_immediate): New.
30516 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
30517 (*movdf_aarch64): Likewise.
30518 * config/aarch64/constraints.md (Ufc): New.
30519 (Y): call aarch64_float_const_zero_rtx.
30520 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
30522 2013-01-07 Richard Biener <rguenther@suse.de>
30524 PR tree-optimization/55888
30525 PR tree-optimization/55862
30526 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
30527 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
30528 not if it is contained therein.
30530 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30532 * config/avr/t-avr: Typo.
30534 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
30537 * config/avr/t-avr: Don't automatically rebuild
30538 $(srcdir)/config/avr/t-multilib
30539 $(srcdir)/config/avr/avr-tables.opt
30540 $(srcdir)/doc/avr-mmcu.texi
30541 (avr-mcus): New phony target to build them on request.
30542 (s-avr-mlib, s-avr-mmcu-texi): Remove.
30543 * avr/avr-mcus.def: Adjust comments.
30545 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
30547 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
30549 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30551 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
30553 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
30555 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
30557 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
30559 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
30560 to generate profiling.
30561 * config/rs6000/aix64.h (LIB_SPEC): Same.
30563 2013-01-04 Andrew Pinski <apinski@cavium.com>
30565 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
30567 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
30569 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
30571 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
30573 (ix86_expand_move): Ditto.
30574 (ix86_zero_extend_to_Pmode): Ditto.
30575 (ix86_expand_call): Ditto.
30576 (ix86_expand_special_args_builtin): Ditto.
30577 (ix86_expand_builtin): Ditto.
30579 2013-01-04 Richard Biener <rguenther@suse.de>
30581 PR tree-optimization/55862
30582 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
30583 translating them through PHI nodes.
30585 2013-01-04 Martin Jambor <mjambor@suse.cz>
30587 PR tree-optimization/55755
30588 * tree-sra.c (sra_modify_assign): Do not check that an access has no
30589 children when trying to avoid producing a VIEW_CONVERT_EXPR.
30591 2013-01-04 Marek Polacek <polacek@redhat.com>
30593 PR middle-end/55859
30594 * opts.c (default_options_optimization): Clarify error message.
30596 2013-01-04 Richard Biener <rguenther@suse.de>
30598 PR middle-end/55863
30599 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
30602 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30605 * config/pa/pa.md (movsi): Revert previous change.
30606 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
30609 2013-01-03 Richard Henderson <rth@redhat.com>
30611 * config/i386/i386.c (ix86_expand_move): Always assign to op1
30612 after eliminating TLS symbols.
30614 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
30617 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
30618 * graphite-poly.c (debug_gmp_value): Likewise.
30620 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
30623 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
30624 selected code model, define __code_mode_small__, __code_model_medium__,
30625 __code_model_large__, __code_model_32__ or __code_model_kernel__.
30626 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
30627 xchg temporary register with %k. Declare temporary register as
30629 [__x86_64__]: For medium and large code models, preserve %rbx register.
30631 2013-01-03 Richard Biener <rguenther@suse.de>
30633 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
30634 (dump_subscript): Adjust.
30635 (finalize_ddr_dependent): Do not dump redundant info.
30636 (analyze_siv_subscript): Adjust.
30637 (subscript_dependence_tester): Likewise.
30638 (compute_affine_dependence): Likewise.
30640 2013-01-03 Richard Biener <rguenther@suse.de>
30643 2013-01-03 Richard Biener <rguenther@suse.de>
30645 PR tree-optimization/55857
30646 * tree-vect-stmts.c (vectorizable_load): Do not setup
30647 re-alignment for invariant loads.
30649 2013-01-02 Richard Biener <rguenther@suse.de>
30651 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30652 invariant load do not generate a vector load from the scalar location.
30654 2013-01-03 Richard Biener <rguenther@suse.de>
30656 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
30657 for not vectorizing.
30658 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
30659 not build INDIRECT_REFs, call get_name once only.
30660 (vect_create_data_ref_ptr): Likewise. Dump base object kind
30661 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
30663 2013-01-03 Richard Biener <rguenther@suse.de>
30665 PR tree-optimization/55857
30666 * tree-vect-stmts.c (vectorizable_load): Do not setup
30667 re-alignment for invariant loads.
30669 2013-01-03 Richard Biener <rguenther@suse.de>
30672 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
30673 prefer a built-in decl.
30675 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30677 * gcc.c (process_command): Update copyright notice dates.
30678 * gcov.c (print_version): Likewise.
30679 * gcov-dump.c (print_version): Likewise.
30681 PR rtl-optimization/55838
30682 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
30683 iv0.step, iv1.step and step.
30685 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30686 Marc Glisse <marc.glisse@inria.fr>
30688 PR tree-optimization/55832
30689 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
30690 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
30691 integer_{one,zero}_node.
30693 2013-01-03 Jakub Jelinek <jakub@redhat.com>
30696 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
30697 * var-tracking.c (reverse_op): Don't add reverse ops to
30698 VALUEs that have already
30699 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
30701 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
30703 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
30705 2013-01-02 Teresa Johnson <tejohnson@google.com>
30707 * dumpfile.c (dump_loc): Print filename with location.
30708 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
30709 new location_t parameter to emit complete unroll message with
30710 new dump framework.
30711 (canonicalize_loop_induction_variables): Compute loops location
30712 and pass to try_unroll_loop_completely.
30713 * loop-unroll.c (report_unroll_peel): New function.
30714 (peel_loops_completely): Use new dump format with location
30715 for main dumpfile message, and invoke report_unroll_peel on success.
30716 (decide_unrolling_and_peeling): Ditto.
30717 (decide_peel_once_rolling): Remove old dumpfile message subsumed
30718 by report_unroll_peel.
30719 (decide_peel_completely): Ditto.
30720 (decide_unroll_constant_iterations): Ditto.
30721 (decide_unroll_runtime_iterations): Ditto.
30722 (decide_peel_simple): Ditto.
30723 (decide_unroll_stupid): Ditto.
30724 * cfgloop.c (get_loop_location): New function.
30725 * cfgloop.h (get_loop_location): Declare.
30727 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30729 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
30732 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
30734 PR middle-end/55198
30735 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
30736 BLKmode objects when EXPAND_MEMORY is specified.
30738 2013-01-02 Sriraman Tallam <tmsriram@google.com>
30740 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
30742 (fold_builtin_cpu): Do not share cpu model decls across statements.
30744 2013-01-02 Jason Merrill <jason@redhat.com>
30747 * tree.c (build_array_type_1): Revert earlier change.
30749 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
30751 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
30753 * config/aarch64/aarch64-tune.md: Re-generate.
30755 2013-01-02 Richard Biener <rguenther@suse.de>
30757 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
30758 invariant load do not generate a vector load from the scalar location.
30760 2013-01-02 Richard Biener <rguenther@suse.de>
30763 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
30764 * configure: Regenerate.
30766 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30768 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
30769 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
30770 (expand_builtin_int_roundingfn_2): Keep the original target around
30771 for the fallback case.
30773 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
30775 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
30776 to be clear for sign changes.
30778 2013-01-01 Jan Hubicka <jh@suse.cz>
30780 * ipa-inline-analysis.c: Fix formatting.
30782 2013-01-01 Jakub Jelinek <jakub@redhat.com>
30784 PR tree-optimization/55831
30785 * tree-vect-loop.c (get_initial_def_for_induction): Use
30786 gsi_after_labels instead of gsi_start_bb.
30788 Copyright (C) 2013 Free Software Foundation, Inc.
30790 Copying and distribution of this file, with or without modification,
30791 are permitted in any medium without royalty provided the copyright
30792 notice and this notice are preserved.