tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from gimple.h and the...
[gcc.git] / gcc / ChangeLog
1 2013-11-05 Andrew MacLeod <amacleod@redhat.com>
2
3 * tree-outof-ssa.c (queue_phi_copy_p): Combine phi_ssa_name_p from
4 gimple.h and the rest of the condition in eliminate_build.
5 (eliminate_build): Call new routine.
6 * gimple.h (phi_ssa_name_p): Delete.
7
8 2013-11-05 Trevor Saunders <tsaunders@mozilla.com>
9
10 * vec.c (vec_prefix::calculate_allocation): Don't try to handle the
11 case of no prefix and reserving zero slots, because when that's the
12 case we'll never get here.
13 * vec.h (va_heap::reserve): Don't try and handle
14 vec_prefix::calculate_allocation returning zero because that should
15 never happen.
16
17 2013-11-05 Richard Biener <rguenther@suse.de>
18
19 PR middle-end/58941
20 * tree-dfa.c (get_ref_base_and_extent): Merge common code
21 in MEM_REF and TARGET_MEM_REF handling. Make sure to
22 process trailing array detection before diving into the
23 view-converted object (and possibly apply some extra offset).
24
25 2013-11-05 Joseph Myers <joseph@codesourcery.com>
26
27 * config/i386/i386.c (ix86_float_exceptions_rounding_supported_p):
28 New function.
29 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Define.
30
31 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
32
33 PR tree-optimization/58958
34 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Use
35 get_addr_base_and_unit_offset instead of get_ref_base_and_extent.
36
37 2013-11-05 Marc Glisse <marc.glisse@inria.fr>
38
39 * tree-ssa-alias.h (ranges_overlap_p): Handle negative offsets.
40 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Likewise.
41
42 2013-11-05 Jakub Jelinek <jakub@redhat.com>
43
44 PR tree-optimization/58984
45 * ipa-prop.c (ipa_load_from_parm_agg_1): Add SIZE_P argument,
46 set *SIZE_P if non-NULL on success.
47 (ipa_load_from_parm_agg, ipa_analyze_indirect_call_uses): Adjust
48 callers.
49 (ipcp_transform_function): Likewise. Punt if size of access
50 is different from TYPE_SIZE on v->value's type.
51
52 2013-11-05 Tobias Burnus <burnus@net-b.de>
53
54 * doc/invoke.texi (-fopenmp-simd): Document new option.
55 * gimplify.c (gimplify_body): Accept -fopenmp-simd.
56 * omp-low.c (execute_expand_omp, execute_lower_omp): Ditto.
57 * tree.c (attribute_value_equal): Ditto.
58
59 2013-11-04 Wei Mi <wmi@google.com>
60
61 * sched-rgn.c (add_branch_dependences): Keep insns in
62 a SCHED_GROUP at the end of BB to remain their location.
63
64 2013-11-04 Wei Mi <wmi@google.com>
65
66 * gcc/config/i386/i386.c (memory_address_length): Extract a part
67 of code to rip_relative_addr_p.
68 (rip_relative_addr_p): New Function.
69 (ix86_macro_fusion_p): Ditto.
70 (ix86_macro_fusion_pair_p): Ditto.
71 * gcc/config/i386/i386.h: Add new tune features about macro-fusion.
72 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
73 * gcc/doc/tm.texi: Generated.
74 * gcc/doc/tm.texi.in: Ditto.
75 * gcc/haifa-sched.c (try_group_insn): New Function.
76 (group_insns_for_macro_fusion): Ditto.
77 (sched_init): Call group_insns_for_macro_fusion.
78 * gcc/target.def: Add two hooks: macro_fusion_p and
79 macro_fusion_pair_p.
80
81 2013-11-04 Kostya Serebryany <kcc@google.com>
82
83 Update to match the changed asan API.
84 * asan.c (asan_function_start): New function.
85 (asan_emit_stack_protection): Update the string stored in the
86 stack red zone to match new API. Store the PC of the current
87 function in the red zone.
88 (asan_global_struct): Update the __asan_global definition to match
89 the new API.
90 (asan_add_global): Ditto.
91 * asan.h (asan_function_start): New prototype.
92 * final.c (final_start_function): Call asan_function_start.
93 * sanitizer.def (BUILT_IN_ASAN_INIT): Rename __asan_init_v1 to __asan_init_v3.
94
95 2013-11-04 Wei Mi <wmi@google.com>
96
97 * gcc/config/i386/i386-c.c (ix86_target_macros_internal): Separate
98 PROCESSOR_COREI7_AVX out from PROCESSOR_COREI7.
99 * gcc/config/i386/i386.c (ix86_option_override_internal): Ditto.
100 (ix86_issue_rate): Ditto.
101 (ix86_adjust_cost): Ditto.
102 (ia32_multipass_dfa_lookahead): Ditto.
103 (ix86_sched_init_global): Ditto.
104 (get_builtin_code_for_version): Ditto.
105 * gcc/config/i386/i386.h (enum target_cpu_default): Ditto.
106 (enum processor_type): Ditto.
107 * gcc/config/i386/x86-tune.def (DEF_TUNE): Ditto.
108
109 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
110
111 PR rtl-optimization/58967
112 * config/rs6000/rs6000.c (legitimate_lo_sum_address_p): Remove
113 !lra_in_progress for mode sizes bigger word.
114
115 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
116
117 * config/rs6000/altivec.md (vec_widen_umult_hi_v16qi): Swap
118 arguments to merge instruction for little endian.
119 (vec_widen_umult_lo_v16qi): Likewise.
120 (vec_widen_smult_hi_v16qi): Likewise.
121 (vec_widen_smult_lo_v16qi): Likewise.
122 (vec_widen_umult_hi_v8hi): Likewise.
123 (vec_widen_umult_lo_v8hi): Likewise.
124 (vec_widen_smult_hi_v8hi): Likewise.
125 (vec_widen_smult_lo_v8hi): Likewise.
126
127 2013-11-04 Ian Lance Taylor <iant@google.com>
128
129 * builtins.def (ATTR_NOTHROWCALL_LEAF_LIST): Define.
130 * sync-builtins.def: Use ATTR_NOTHROWCALL_LEAF_LIST for all sync
131 builtins that take pointers.
132 * lto-opts.c (lto_write_options): Write -fnon-call-exceptions
133 if set.
134 * lto-wrapper.c (merge_and_complain): Collect
135 OPT_fnon_call_exceptions.
136 (run_gcc): Pass -fnon-call-exceptions.
137
138 2013-11-04 Jakub Jelinek <jakub@redhat.com>
139
140 * optabs.c (expand_vec_perm): Revert one incorrect line from
141 2013-10-31 change.
142
143 PR tree-optimization/58978
144 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): Don't modify
145 use_stmt by single_imm_use directly. Only call single_imm_use
146 on SSA_NAMEs.
147
148 2013-11-04 Vladimir Makarov <vmakarov@redhat.com>
149
150 PR rtl-optimization/58968
151 * lra-spills.c (return_regno_p): New function.
152 (lra_final_code_change): Use it.
153
154 2013-11-04 Joseph Myers <joseph@codesourcery.com>
155
156 * doc/cpp.texi (__GCC_IEC_559, __GCC_IEC_559_COMPLEX): Document
157 macros.
158 * target.def (float_exceptions_rounding_supported_p): New hook.
159 * targhooks.c (default_float_exceptions_rounding_supported_p): New
160 function.
161 * targhooks.h (default_float_exceptions_rounding_supported_p):
162 Declare.
163 * doc/tm.texi.in (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P):
164 New @hook.
165 * doc/tm.texi: Regenerate.
166 * config.gcc (powerpc*-*-linux*): Set extra_objs.
167 * config/rs6000/rs6000-linux.c: New file.
168 * config/rs6000/rs6000-protos.h
169 (rs6000_linux_float_exceptions_rounding_supported_p): Declare.
170 * config/rs6000/linux.h
171 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): New macro.
172 * config/rs6000/linux64.h
173 (TARGET_FLOAT_EXCEPTIONS_ROUNDING_SUPPORTED_P): Likewise.
174 * config/rs6000/t-linux (rs6000-linux.o): New rule.
175 * config/rs6000/t-linux64 (rs6000-linux.o): Likewise.
176
177 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
178
179 * config/rs6000/vsx.md (*vsx_le_perm_store_<mode> for VSX_D):
180 Replace the define_insn_and_split with a define_insn and two
181 define_splits, with the split after reload re-permuting the source
182 register to its original value.
183 (*vsx_le_perm_store_<mode> for VSX_W): Likewise.
184 (*vsx_le_perm_store_v8hi): Likewise.
185 (*vsx_le_perm_store_v16qi): Likewise.
186
187 2013-11-04 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
188
189 * config/rs6000/vector.md (vec_pack_trunc_v2df): Adjust for
190 little endian.
191
192 2013-11-04 Jakub Jelinek <jakub@redhat.com>
193
194 PR tree-optimization/58946
195 * tree-ssa-reassoc.c (maybe_optimize_range_tests): Update all
196 bbs with bbinfo[idx].op != NULL before all blocks with
197 bbinfo[idx].op == NULL.
198
199 2013-11-04 Richard Sandiford <rdsandiford@googlemail.com>
200
201 * config/avr/avr-log.c (avr_double_int_pop_digit): Delete.
202 (avr_dump_double_int_hex): Likewise.
203 (avr_log_vadump): Remove %D and %X handling.
204 * config/avr/avr.c (avr_double_int_push_digit): Delete.
205 (avr_map_op_t): Change map from double_int to unsigned int.
206 (avr_map_op): Update accordingly.
207 (avr_map, avr_map_metric, avr_has_nibble_0xf, avr_map_decompose)
208 (avr_move_bits, avr_out_insert_bits, avr_fold_builtin): Operate on
209 unsigned ints rather than double_ints.
210
211 2013-11-03 Marek Polacek <polacek@redhat.com>
212
213 Implement -fsanitize=vla-bound.
214 * opts.c (common_handle_option): Handle vla-bound.
215 * sanitizer.def (BUILT_IN_UBSAN_HANDLE_VLA_BOUND_NOT_POSITIVE):
216 Define.
217 * flag-types.h (enum sanitize_code): Add SANITIZE_VLA.
218 * asan.c (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR.
219
220 2013-11-02 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
221
222 * config/rs6000/rs6000.c (rs6000_expand_vector_set): Adjust for
223 little endian.
224
225 2013-11-02 Uros Bizjak <ubizjak@gmail.com>
226
227 * config/i386/constraints.md (Ts, Tv): New address constrains.
228 * config/i386/i386.md (*lea<mode>, *<mode>_<bndcheck>): Use Ts
229 constraint for address_no_seg_operand.
230 * config/i386/sse.md (*avx512pf_gatherpf<mode>_mask)
231 (*avx512pf_gatherpf<mode>, *avx512pf_scatterpf<mode>_mask)
232 (*avx512pf_scatterpf<mode>, *avx2_gathersi<mode>)
233 (*avx2_gathersi<mode>_2, *avx2_gatherdi<mode>, *avx2_gatherdi<mode>_2)
234 (*avx2_gatherdi<mode>_3, *avx2_gatherdi<mode>_4)
235 (*avx512f_gathersi<mode>, *avx512f_gathersi<mode>_2)
236 (*avx512f_gatherdi<mode>, *avx512f_gatherdi<mode>_2)
237 (*avx512f_scattersi<mode> *avx512f_scatterdi<mode>): Use Tv
238 constraint for vsib_address_operand.
239
240 2013-11-02 Steven Bosscher <steven@gcc.gnu.org>
241
242 * gcse.c (pre_delete): Remove references to regmove from comments.
243 * recog.c: (validate_replace_rtx_1): Likewise.
244 * config/rl78/rl78.c: Likewise.
245 * config/v850/v850.h: Likewise, and remove unused ENABLE_REGMOVE_PASS.
246 * common/config/m32r/m32r-common.c: Don't manipulate OPT_fregmove.
247 * common/config/mmix/mmix-common.c: Likewise.
248
249 2013-11-01 Trevor Saunders <tsaunders@mozilla.com>
250
251 * function.c (reorder_blocks): Convert block_stack to a stack_vec.
252 * gimplify.c (gimplify_compound_lval): Likewise.
253 * graphite-clast-to-gimple.c (gloog): Likewise.
254 * graphite-dependences.c (loop_is_parallel_p): Likewise.
255 * graphite-scop-detection.c (scopdet_basic_block_info): Likewise.
256 (limit_scop); Likewise.
257 (build_scops): Likewise.
258 (dot_scop): Likewise.
259 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
260 (build_scop_drs): Likewise.
261 (insert_stmts): Likewise.
262 (insert_out_of_ssa_copy): Likewise.
263 (remove_phi): Likewise.
264 (rewrite_commutative_reductions_out_of_ssa_close_phi): Likewise.
265 * hw-doloop.c (discover_loop): Likewise.
266 * tree-call-cdce.c (shrink_wrap_one_built_in_call): Likewise.
267 * tree-dfa.c (dump_enumerated_decls): Likewise.
268 * tree-if-conv.c (if_convertable_loop_p): Likewise.
269 * tree-inline.c (tree_function_versioning): Likewise.
270 * tree-loop-distribution.c (build_rdg): Likewise.
271 (rdg_flag_vertex_and_dependent): Likewise.
272 (distribute_loop): Likewise.
273 * tree-parloops.c (loop_parallel_p): Likewise.
274 (eliminate_local_variables): Likewise.
275 (separate_decls_in_region): Likewise.
276 * tree-predcom.c (tree_predictive_commoning_loop): Likewise.
277 * tree-ssa-phiopt.c (cond_if_else_store_replacement): Likewise.
278 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
279 * tree-vect-loop.c (vect_analyze_scaler_cycles_1): Likewise.
280 * tree-vect-patterns.c (vect_pattern_recog): Likewise.
281 * tree-vect-stmts.c (vect_mark_stmts_to_be_vectorized): Likewise.
282 (vectorizable_condition): Likewise.
283
284 2013-11-01 Uros Bizjak <ubizjak@gmail.com>
285
286 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): Always define as 0/1.
287 * configure: Regenerate.
288 * config/i386/i386.md (*movdi_internal): Change
289 HAVE_AS_IX86_INTERUNIT_MOVQ to runtime check.
290 (*movdf_internal): Ditto.
291 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
292 * config/i386/sse.md (vec_concatv2di): Output interunit movq
293 for HAVE_AS_IX86_INTERUNIT_MOVQ targets.
294
295 2013-10-31 Robert Suchanek <Robert.Suchanek@imgtec.com>
296
297 * lra-spills.c (assign_spill_hard_regs): Remove statement terminator
298 after comment.
299
300 2013-10-31 David Malcolm <dmalcolm@redhat.com>
301
302 Automated part of renaming of symtab_node_base to symtab_node.
303
304 Patch autogenerated by rename_symtab.py from
305 https://github.com/davidmalcolm/gcc-refactoring-scripts
306 revision 58bb219cc090b2f4516a9297d868c245495ee622
307 with ChangeLog entry fixed up by hand.
308
309 * cgraph.c (x_cgraph_nodes_queue): Rename symtab_node_base to
310 symtab_node.
311 (cgraph_node_for_asm): Likewise.
312 * cgraph.h (symtab_node_base): Likewise.
313 (cgraph_node): Likewise.
314 (varpool_node): Likewise.
315 (is_a_helper <cgraph_node>::test): Likewise.
316 (is_a_helper <varpool_node>::test): Likewise.
317 (symtab_nodes): Likewise.
318 (symtab_register_node): Likewise.
319 (symtab_unregister_node): Likewise.
320 (symtab_remove_node): Likewise.
321 (symtab_get_node): Likewise.
322 (symtab_node_for_asm): Likewise.
323 (symtab_node_asm_name): Likewise.
324 (symtab_node_name): Likewise.
325 (symtab_insert_node_to_hashtable): Likewise.
326 (symtab_add_to_same_comdat_group): Likewise.
327 (symtab_dissolve_same_comdat_group_list): Likewise.
328 (dump_symtab_node): Likewise.
329 (debug_symtab_node): Likewise.
330 (dump_symtab_base): Likewise.
331 (verify_symtab_node): Likewise.
332 (verify_symtab_base): Likewise.
333 (symtab_used_from_object_file_p): Likewise.
334 (symtab_alias_ultimate_target): Likewise.
335 (symtab_resolve_alias): Likewise.
336 (fixup_same_cpp_alias_visibility): Likewise.
337 (symtab_for_node_and_aliases): Likewise.
338 (symtab_nonoverwritable_alias): Likewise.
339 (availability symtab_node_availability): Likewise.
340 (symtab_semantically_equivalent_p): Likewise.
341 (fixup_same_cpp_alias_visibility): Likewise.
342 (symtab_prevail_in_asm_name_hash): Likewise.
343 (cgraph): Likewise.
344 (varpool): Likewise.
345 (varpool_first_variable): Likewise.
346 (varpool_next_variable): Likewise.
347 (varpool_first_static_initializer): Likewise.
348 (varpool_next_static_initializer): Likewise.
349 (varpool_first_defined_variable): Likewise.
350 (varpool_next_defined_variable): Likewise.
351 (cgraph_first_defined_function): Likewise.
352 (cgraph_next_defined_function): Likewise.
353 (cgraph_first_function): Likewise.
354 (cgraph_next_function): Likewise.
355 (cgraph_first_function_with_gimple_body): Likewise.
356 (cgraph_next_function_with_gimple_body): Likewise.
357 (symtab_alias_target): Likewise.
358 (symtab_real_symbol_p): Likewise.
359 (symtab_can_be_discarded): Likewise.
360 * cgraphbuild.c (mark_address): Likewise.
361 (mark_load): Likewise.
362 (mark_store): Likewise.
363 * cgraphunit.c (decide_is_symbol_needed): Likewise.
364 (first): Likewise.
365 (enqueue_node): Likewise.
366 (referred_to_p): Likewise.
367 (cgraph_process_same_body_aliases): Likewise.
368 (analyze_functions): Likewise.
369 (handle_alias_pairs): Likewise.
370 (output_weakrefs): Likewise.
371 (compile): Likewise.
372 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
373 * ipa-inline-analysis.c (inline_write_summary): Likewise.
374 * ipa-prop.c (remove_described_reference): Likewise.
375 (try_decrement_rdesc_refcount): Likewise.
376 (ipa_edge_duplication_hook): Likewise.
377 * ipa-ref.c (ipa_record_reference): Likewise.
378 (ipa_maybe_record_reference): Likewise.
379 (ipa_clone_ref): Likewise.
380 (ipa_clone_references): Likewise.
381 (ipa_clone_referring): Likewise.
382 (ipa_find_reference): Likewise.
383 (ipa_remove_stmt_references): Likewise.
384 (ipa_clear_stmts_in_references): Likewise.
385 * ipa-ref.h (symtab_node_base): Likewise.
386 (ipa_ref): Likewise.
387 (ipa_record_reference): Likewise.
388 (ipa_maybe_record_reference): Likewise.
389 (ipa_clone_references): Likewise.
390 (ipa_clone_referring): Likewise.
391 (ipa_clone_ref): Likewise.
392 (ipa_find_reference): Likewise.
393 (ipa_remove_stmt_references): Likewise.
394 (ipa_clear_stmts_in_references): Likewise.
395 * ipa-reference.c (ipa_reference_write_optimization_summary):
396 Likewise.
397 * ipa.c (enqueue_node): Likewise.
398 (process_references): Likewise.
399 (walk_polymorphic_call_targets): Likewise.
400 (symtab_remove_unreachable_nodes): Likewise.
401 (address_taken_from_non_vtable_p): Likewise.
402 (comdat_can_be_unshared_p_1): Likewise.
403 (comdat_can_be_unshared_p): Likewise.
404 (can_replace_by_local_alias): Likewise.
405 (function_and_variable_visibility): Likewise.
406 * is-a.h: Likewise (within example in comment).
407 * lto-cgraph.c (input_cgraph_opt_summary): Likewise.
408 (lto_symtab_encoder_encode): Likewise.
409 (lto_symtab_encoder_delete_node): Likewise.
410 (lto_symtab_encoder_in_partition_p): Likewise.
411 (lto_set_symtab_encoder_in_partition): Likewise.
412 (output_refs): Likewise.
413 (compute_ltrans_boundary): Likewise.
414 (output_symtab): Likewise.
415 (input_node): Likewise.
416 (input_ref): Likewise.
417 (input_edge): Likewise.
418 (input_cgraph_1): Likewise.
419 (input_refs): Likewise.
420 (output_cgraph_opt_summary): Likewise.
421 (input_node_opt_summary): Likewise.
422 (input_cgraph_opt_section): Likewise.
423 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
424 Likewise.
425 * lto-streamer-out.c (lto_output): Likewise.
426 (output_symbol_p): Likewise.
427 (produce_symtab): Likewise.
428 * lto-streamer.h (lto_encoder_entry): Likewise.
429 (lto_free_function_in_decl_state_for_node): Likewise.
430 (lto_symtab_encoder_encode): Likewise.
431 (lto_symtab_encoder_delete_node): Likewise.
432 (lto_symtab_encoder_in_partition_p): Likewise.
433 (lto_set_symtab_encoder_in_partition): Likewise.
434 (lto_symtab_encoder_lookup): Likewise.
435 (lsei_node): Likewise.
436 (lto_symtab_encoder_deref): Likewise.
437 * symtab.c (symtab_hash): Likewise.
438 (assembler_name_hash): Likewise.
439 (symtab_nodes): Likewise.
440 (hash_node): Likewise.
441 (eq_node): Likewise.
442 (hash_node_by_assembler_name): Likewise.
443 (eq_assembler_name): Likewise.
444 (insert_to_assembler_name_hash): Likewise.
445 (unlink_from_assembler_name_hash): Likewise.
446 (symtab_prevail_in_asm_name_hash): Likewise.
447 (symtab_register_node): Likewise.
448 (symtab_insert_node_to_hashtable): Likewise.
449 (symtab_unregister_node): Likewise.
450 (symtab_get_node): Likewise.
451 (symtab_remove_node): Likewise.
452 (symtab_initialize_asm_name_hash): Likewise.
453 (symtab_node_for_asm): Likewise.
454 (symtab_add_to_same_comdat_group): Likewise.
455 (symtab_dissolve_same_comdat_group_list): Likewise.
456 (symtab_node_asm_name): Likewise.
457 (symtab_node_name): Likewise.
458 (dump_symtab_base): Likewise.
459 (dump_symtab_node): Likewise.
460 (dump_symtab): Likewise.
461 (debug_symtab_node): Likewise.
462 (verify_symtab_base): Likewise.
463 (verify_symtab_node): Likewise.
464 (verify_symtab): Likewise.
465 (symtab_used_from_object_file_p): Likewise.
466 (symtab_node_availability): Likewise.
467 (symtab_alias_ultimate_target): Likewise.
468 (fixup_same_cpp_alias_visibility): Likewise.
469 (symtab_resolve_alias): Likewise.
470 (symtab_for_node_and_aliases): Likewise.
471 (symtab_for_node_and_aliases): Likewise.
472 (symtab_nonoverwritable_alias_1): Likewise.
473 (symtab_nonoverwritable_alias): Likewise.
474 (symtab_semantically_equivalent_p): Likewise.
475 * value-prof.c (init_node_map): Likewise.
476 * varasm.c (find_decl): Likewise.
477 * varpool.c (varpool_node_for_asm): Likewise.
478 (varpool_remove_unreferenced_decls): Likewise.
479
480 2013-10-31 David Malcolm <dmalcolm@redhat.com>
481
482 Manual part of renaming of symtab_node_base to symtab_node.
483
484 * ipa-ref.h (symtab_node): Remove typedef to pointer type, as it
485 clashes with the preferred name for the base class.
486 (const_symtab_node): Remove redundant typedef.
487
488 2013-10-31 Jakub Jelinek <jakub@redhat.com>
489
490 * optabs.c (expand_vec_perm): Avoid vector mode punning
491 SUBREGs in SET_DEST.
492 * expmed.c (store_bit_field_1): Likewise.
493 * config/i386/sse.md (movdi_to_sse, vec_pack_sfix_trunc_v2df,
494 vec_pack_sfix_v2df, vec_shl_<mode>, vec_shr_<mode>,
495 vec_interleave_high<mode>, vec_interleave_low<mode>): Likewise.
496 * config/i386/i386.c (ix86_expand_vector_move_misalign,
497 ix86_expand_sse_movcc, ix86_expand_int_vcond, ix86_expand_vec_perm,
498 ix86_expand_sse_unpack, ix86_expand_args_builtin,
499 ix86_expand_vector_init_duplicate, ix86_expand_vector_set,
500 emit_reduc_half, expand_vec_perm_blend, expand_vec_perm_pshufb,
501 expand_vec_perm_interleave2, expand_vec_perm_pshufb2,
502 expand_vec_perm_vpshufb2_vpermq,
503 expand_vec_perm_vpshufb2_vpermq_even_odd, expand_vec_perm_even_odd_1,
504 expand_vec_perm_broadcast_1, expand_vec_perm_vpshufb4_vpermq2,
505 ix86_expand_sse2_mulv4si3, ix86_expand_pinsr): Likewise.
506 (expand_vec_perm_palignr): Likewise. Modify a copy of *d rather
507 than *d itself.
508
509 2013-10-31 Uros Bizjak <ubizjak@gmail.com>
510
511 * config/i386/i386.c (ix86_expand_sse2_abs): Rename function arguments.
512 Use gcc_unreachable for unhandled modes. Do not check results of
513 expand_simple_binop. If not expanded to target, move the result.
514
515 2013-10-31 Chung-Ju Wu <jasonwucj@gmail.com>
516 Shiva Chen <shiva0217@gmail.com>
517
518 * config.gcc (nds32*-*-*): Add nds32 target.
519 * config/nds32/nds32.c: New file.
520 * config/nds32/nds32.h: New file.
521 * config/nds32/nds32.md: New file.
522 * config/nds32/constants.md: New file.
523 * config/nds32/constraints.md: New file.
524 * config/nds32/iterators.md: New file.
525 * config/nds32/nds32-doubleword.md: New file.
526 * config/nds32/nds32-intrinsic.md: New file.
527 * config/nds32/nds32_intrinsic.h: New file.
528 * config/nds32/nds32-modes.def: New file.
529 * config/nds32/nds32-multiple.md: New file.
530 * config/nds32/nds32.opt: New file.
531 * config/nds32/nds32-opts.h: New file.
532 * config/nds32/nds32-protos.h: New file.
533 * config/nds32/nds32-peephole2.md: New file.
534 * config/nds32/pipelines.md: New file.
535 * config/nds32/predicates.md: New file.
536 * config/nds32/t-mlibs: New file.
537 * common/config/nds32: New directory and files.
538
539 * doc/invoke.texi (NDS32 options): Document nds32 specific options.
540 * doc/md.texi (NDS32 family): Document nds32 specific constraints.
541 * doc/install.texi (Cross-Compiler-Specific Options): Document
542 --with-nds32-lib for nds32 target.
543 * doc/extend.texi (Function Attributes, Target Builtins): Document
544 nds32 specific attributes.
545
546 2013-10-31 Vladimir Makarov <vmakarov@redhat.com>
547
548 * lra-constraints (process_alt_operands): Use the result
549 elimination register for operand when matching constraints.
550
551 2013-10-31 Jakub Jelinek <jakub@redhat.com>
552
553 * tree-vrp.c (maybe_set_nonzero_bits): New function.
554 (remove_range_assertions): Call it.
555
556 * tree.c (tree_ctz): New function.
557 * tree.h (tree_ctz): New prototype.
558 * tree-ssanames.h (get_range_info, get_nonzero_bits): Change
559 first argument from tree to const_tree.
560 * tree-ssanames.c (get_range_info, get_nonzero_bits): Likewise.
561 * tree-vectorizer.h (vect_generate_tmps_on_preheader): New prototype.
562 * tree-vect-loop-manip.c (vect_generate_tmps_on_preheader): No longer
563 static.
564 * expr.c (highest_pow2_factor): Reimplemented using tree_ctz.
565 * tree-vect-loop.c (vect_analyze_loop_operations,
566 vect_transform_loop): Don't force scalar loop for bound just because
567 number of iterations is unknown, only do it if it is not known to be
568 a multiple of vectorization_factor.
569 * builtins.c (get_object_alignment_2): Use tree_ctz on offset.
570
571 * gimple-pretty-print.c (dump_ssaname_info): Print newline also
572 in case of VR_VARYING. Print get_nonzero_bits if not all ones.
573 * tree-ssanames.h (struct range_info_def): Add nonzero_bits field.
574 (set_nonzero_bits, get_nonzero_bits): New prototypes.
575 * tree-ssa-ccp.c (get_default_value): Use get_range_info to see if
576 a default def isn't partially constant.
577 (ccp_finalize): If after IPA, set_range_info if integral SSA_NAME
578 is known to be partially zero.
579 (evaluate_stmt): If we'd return otherwise VARYING, use get_range_info
580 to see if a default def isn't partially constant.
581 * tree-ssanames.c (set_range_info): Initialize nonzero_bits upon
582 creation of a range, if VR_RANGE, try to improve nonzero_bits from
583 the range.
584 (set_nonzero_bits, get_nonzero_bits): New functions.
585
586 * tree-cfg.c (assert_unreachable_fallthru_edge_p): New function.
587 * tree-cfg.h (assert_unreachable_fallthru_edge_p): New prototype.
588 * tree-vrp.c (all_imm_uses_in_stmt_or_feed_cond): New function.
589 (remove_range_assertions): If ASSERT_EXPR_VAR has no other immediate
590 uses but in the condition and ASSERT_EXPR and the other successor of
591 the predecessor bb is __builtin_unreachable (), set_range_info of the
592 ASSERT_EXPR_VAR to the range info of the ASSERT_EXPR's lhs.
593
594 2013-10-31 Martin Jambor <mjambor@suse.cz>
595
596 PR rtl-optimization/58934
597 Revert:
598 2013-10-30 Martin Jambor <mjambor@suse.cz>
599 PR rtl-optimization/10474
600 * ira.c (find_moveable_pseudos): Do not calculate dominance info
601 nor df analysis.
602 (interesting_dest_for_shprep): New function.
603 (split_live_ranges_for_shrink_wrap): Likewise.
604 (ira): Calculate dominance info and df analysis. Call
605 split_live_ranges_for_shrink_wrap.
606
607 2013-10-31 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
608 Yury Gribov <y.gribov@samsung.com>
609
610 PR sanitizer/58543
611 * asan.c (asan_clear_shadow): Allocate a new vreg for temporary
612 shadow pointer to avoid clobbering the main one.
613
614 2013-10-31 Zhenqiang Chen <zhenqiang.chen@linaro.org>
615
616 * lower-subreg.c (resolve_simple_move): Copy REG_INC note.
617
618 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
619
620 PR bootstrap/58933
621 * ira-color.c (update_costs_from_copies): Add new parameter. Use
622 it for calling update_costs_from_allocno.
623 (assign_hard_reg): Call restore_costs_from_copies only for
624 !retry_p. Pass new argument to update_costs_from_copies.
625 (color_pass): Pass new argument to update_costs_from_copies.
626 (ira_mark_allocation_change): Ditto.
627
628 2013-10-30 Sharad Singhai <singhai@google.com>
629
630 PR middle-end/58134
631 * opts.c (common_handle_option): Remove deprecated option
632 -ftree-vectorizer-verbose.
633 * doc/invoke.texi (Debugging Options): Ditto.
634 * opts-global.c (handle_common_deferred_options): Ditto.
635 (dump_remap_tree_vectorizer_verbose): Delete.
636 * common.opt: Set -ftree-vectorizer-verbose as an ignored option.
637
638 2013-10-30 DJ Delorie <dj@redhat.com>
639
640 * config/rx/rx.c (ADD_RX_BUILTIN0): New macro, used for builtins
641 that take no arguments.
642
643 2013-10-30 Joern Rennecke <joern.rennecke@embecosm.com>
644
645 PR other/58545
646 * reload1.c (update_eliminables_and_spill): New function, broken
647 out of reload.
648 (reload): Use it. Check for frame size change after frame size
649 alignment, and call update_eliminables_and_spill first if continue-ing.
650
651 2013-10-30 Cong Hou <congh@google.com>
652
653 PR target/58762
654 * config/i386/i386-protos.h (ix86_expand_sse2_abs): New function.
655 * config/i386/i386.c (ix86_expand_sse2_abs): New function.
656 * config/i386/sse.md: Add SSE2 support to abs (8/16/32-bit-int).
657
658 2013-10-18 Mikael Pettersson <mikpelinux@gmail.com>
659
660 PR rtl-optimization/58369
661 * reload1.c (compute_reload_subreg_offset): New function.
662 (choose_reload_regs): Use it to pass endian-correct
663 offset to subreg_regno_offset.
664
665 2013-10-30 Tobias Burnus <burnus@net-b.de>
666
667 PR other/33426
668 * gcc/tree-cfg.c (replace_loop_annotate): Replace warning by
669 warning_at.
670
671 2013-10-30 Jason Merrill <jason@redhat.com>
672
673 * configure.ac (loose_warn): Add -Wno-format if
674 --disable-build-format-warnings.
675
676 2013-10-30 David Malcolm <dmalcolm@redhat.com>
677
678 * cgraphunit.c (analyze_functions): Split symtab_node declarations
679 onto multiple lines to make things easier for rename_symtab.py.
680
681 * symtab.c (symtab_dissolve_same_comdat_group_list): Likewise.
682 (symtab_semantically_equivalent_p): Likewise.
683
684 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
685
686 PR target/58784
687 * lra.c (check_rtl): Remove address check before LRA work.
688
689 2013-10-30 Marc Glisse <marc.glisse@inria.fr>
690
691 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for a
692 POINTER_PLUS_EXPR in the defining statement.
693
694 2013-10-30 Vladimir Makarov <vmakarov@redhat.com>
695
696 * regmove.c: Remove.
697 * tree-pass.h (make_pass_regmove): Remove.
698 * timevar.def (TV_REGMOVE): Remove.
699 * passes.def (pass_regmove): Remove.
700 * opts.c (default_options_table): Remove entry for regmove.
701 * doc/passes.texi: Remove regmove pass description.
702 * doc/invoke.texi (-foptimize-register-move, -fregmove): Remove
703 options.
704 (-fdump-rtl-regmove): Ditto.
705 * common.opt (foptimize-register-move, fregmove): Ignore.
706 * Makefile.in (OBJS): Remove regmove.o.
707 * regmove.c: Remove.
708 * ira-int.h (struct ira_allocno_pref, ira_pref_t): New structure
709 and type.
710 (struct ira_allocno) New member allocno_prefs.
711 (ALLOCNO_PREFS): New macro.
712 (ira_prefs, ira_prefs_num): New external vars.
713 (ira_setup_alts, ira_get_dup_out_num, ira_debug_pref): New prototypes.
714 (ira_debug_prefs, ira_debug_allocno_prefs, ira_create_pref): Ditto.
715 (ira_add_allocno_pref, ira_remove_pref, ira_remove_allocno_prefs):
716 Ditto.
717 (ira_add_allocno_copy_to_list): Remove prototype.
718 (ira_swap_allocno_copy_ends_if_necessary): Ditto.
719 (ira_pref_iterator): New type.
720 (ira_pref_iter_init, ira_pref_iter_cond): New functions.
721 (FOR_EACH_PREF): New macro.
722 * ira.c (commutative_constraint_p): Move from ira-conflicts.c.
723 (ira_get_dup_out_num): Ditto. Rename from get_dup_num. Modify the
724 code.
725 (ira_setup_alts): New function.
726 (decrease_live_ranges_number): New function.
727 (ira): Call the above function.
728 * ira-build.c (ira_prefs, ira_prefs_num): New global vars.
729 (ira_create_allocno): Initialize allocno prefs.
730 (pref_pool, pref_vec): New static vars.
731 (initiate_prefs, find_allocno_pref, ira_create_pref): New
732 functions.
733 (add_allocno_pref_to_list, ira_add_allocno_pref, print_pref): Ditto.
734 (ira_debug_pref, print_prefs, ira_debug_prefs): Ditto.
735 (print_allocno_prefs, ira_debug_allocno_prefs, finish_pref): Ditto.
736 (ira_remove_pref, ira_remove_allocno_prefs, finish_prefs): Ditto.
737 (ira_add_allocno_copy_to_list): Make static. Rename to
738 add_allocno_copy_to_list.
739 (ira_swap_allocno_copy_ends_if_necessary): Make static. Rename to
740 swap_allocno_copy_ends_if_necessary.
741 (remove_unnecessary_allocnos, remove_low_level_allocnos): Call
742 ira_remove_allocno_prefs.
743 (ira_flattening): Ditto.
744 (ira_build): Call initiate_prefs, print_prefs.
745 (ira_destroy): Call finish_prefs.
746 * ira-color.c (struct update_cost_record): New.
747 (struct allocno_color_data): Add new member update_cost_records.
748 (update_cost_record_pool): New static var.
749 (init_update_cost_records, get_update_cost_record): New functions.
750 (free_update_cost_record_list, finish_update_cost_records): Ditto.
751 (struct update_cost_queue_elem): Add member from.
752 (initiate_cost_update): Call init_update_cost_records.
753 (finish_cost_update): Call finish_update_cost_records.
754 (queue_update_cost, get_next_update_cost): Add new param from.
755 (Update_allocno_cost, update_costs_from_allocno): New functions.
756 (update_costs_from_prefs): Ditto.
757 (update_copy_costs): Rename to update_costs_from_copies.
758 (restore_costs_from_copies): New function.
759 (update_conflict_hard_regno_costs): Don't go back.
760 (assign_hard_reg): Call restore_costs_from_copies. Add printing
761 more debug info.
762 (pop_allocnos): Add priniting more debug info.
763 (color_allocnos): Remove prefs for conflicting hard regs.
764 Call update_costs_from_prefs.
765 * ira-conflicts.c (commutative_constraint_p): Move to ira.c
766 (get_dup_num): Rename, modify, and move to ira.c
767 (process_regs_for_copy): Add prefs.
768 (add_insn_allocno_copies): Put src as first arg of
769 process_regs_for_copy. Remove dead code. Call ira_setup_alts.
770 * ira-costs.c (record_reg_classes): Modify and move code into
771 record_operands_costs.
772 (find_costs_and_classes): Create prefs for the hard reg of small
773 reg class.
774 (process_bb_node_for_hard_reg_moves): Add prefs.
775
776 2013-10-30 Richard Biener <rguenther@suse.de>
777
778 PR middle-end/57100
779 * basic-block.h (pre_and_rev_post_order_compute_fn): New function.
780 * cfganal.c (pre_and_rev_post_order_compute_fn): New function
781 as worker for ...
782 (pre_and_rev_post_order_compute): ... which now wraps it.
783 * graph.c (draw_cfg_nodes_no_loops): Use
784 pre_and_rev_post_order_compute_fn to avoid ICEing and dependence
785 on cfun.
786
787 2013-10-30 Christian Bruel <christian.bruel@st.com>
788
789 * gcc/config/sh/sh-mem.cc (sh_expand_cmpnstr): New function.
790 (sh_expand_cmpstr): Handle known align and schedule improvements.
791 * gcc/config/sh/sh-protos.h (sh_expand_cmpstrn): Declare.
792 * gcc/config/sh/sh.md (cmpstrnsi): New pattern.
793
794 2013-10-30 Martin Jambor <mjambor@suse.cz>
795
796 PR rtl-optimization/10474
797 * ira.c (find_moveable_pseudos): Do not calculate dominance info
798 nor df analysis.
799 (interesting_dest_for_shprep): New function.
800 (split_live_ranges_for_shrink_wrap): Likewise.
801 (ira): Calculate dominance info and df analysis. Call
802 split_live_ranges_for_shrink_wrap.
803
804 2013-10-30 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
805
806 PR target/58854
807 * config/arm/arm.c (arm_expand_epilogue_apcs_frame): Emit blockage.
808
809 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
810
811 * tree-core.h (tree_index): Add TI_POINTER_BOUNDS_TYPE.
812 * tree.h (POINTER_BOUNDS_P): New.
813 (BOUNDED_TYPE_P): New.
814 (BOUNDED_P): New.
815 (pointer_bounds_type_node): New.
816 * tree.c (build_common_tree_nodes): Initialize
817 pointer_bounds_type_node.
818 * gimple.h (gimple_call_get_nobnd_arg_index): New.
819 (gimple_call_num_nobnd_args): New.
820 (gimple_call_nobnd_arg): New.
821 (gimple_return_retbnd): New.
822 (gimple_return_set_retbnd): New
823 * gimple.c (gimple_build_return): Increase number of ops
824 for return statement.
825 (gimple_call_get_nobnd_arg_index): New.
826 * gimple-pretty-print.c (dump_gimple_return): Print second op.
827
828 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
829
830 * ipa.c (cgraph_build_static_cdtor_1): Support contructors
831 with "chkp ctor" and "bnd_legacy" attributes.
832 * gimplify.c (gimplify_init_constructor): Avoid infinite
833 loop during gimplification of bounds initializer.
834
835 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
836
837 * c-family/c-common.c (handle_bnd_variable_size_attribute): New.
838 (handle_bnd_legacy): New.
839 (c_common_attribute_table): Add bnd_variable_size and bnd_legacy.
840 * doc/extend.texi: Document bnd_variable_size and bnd_legacy
841 attributes.
842
843 2013-10-29 Ilya Enkovich <ilya.enkovich@intel.com>
844
845 * builtin-types.def (BT_FN_VOID_CONST_PTR): New.
846 (BT_FN_PTR_CONST_PTR): New.
847 (BT_FN_CONST_PTR_CONST_PTR): New.
848 (BT_FN_PTR_CONST_PTR_SIZE): New.
849 (BT_FN_PTR_CONST_PTR_CONST_PTR): New.
850 (BT_FN_VOID_PTRPTR_CONST_PTR): New.
851 (BT_FN_VOID_CONST_PTR_SIZE): New.
852 (BT_FN_PTR_CONST_PTR_CONST_PTR_SIZE): New.
853 * chkp-builtins.def: New.
854 * builtins.def: include chkp-builtins.def.
855 (DEF_CHKP_BUILTIN): New.
856 * builtins.c (expand_builtin): Support BUILT_IN_CHKP_INIT_PTR_BOUNDS,
857 BUILT_IN_CHKP_NULL_PTR_BOUNDS, BUILT_IN_CHKP_COPY_PTR_BOUNDS,
858 BUILT_IN_CHKP_CHECK_PTR_LBOUNDS, BUILT_IN_CHKP_CHECK_PTR_UBOUNDS,
859 BUILT_IN_CHKP_CHECK_PTR_BOUNDS, BUILT_IN_CHKP_SET_PTR_BOUNDS,
860 BUILT_IN_CHKP_NARROW_PTR_BOUNDS, BUILT_IN_CHKP_STORE_PTR_BOUNDS,
861 BUILT_IN_CHKP_GET_PTR_LBOUND, BUILT_IN_CHKP_GET_PTR_UBOUND,
862 BUILT_IN_CHKP_BNDMK, BUILT_IN_CHKP_BNDSTX, BUILT_IN_CHKP_BNDCL,
863 BUILT_IN_CHKP_BNDCU, BUILT_IN_CHKP_BNDLDX, BUILT_IN_CHKP_BNDRET,
864 BUILT_IN_CHKP_INTERSECT, BUILT_IN_CHKP_ARG_BND, BUILT_IN_CHKP_NARROW,
865 BUILT_IN_CHKP_EXTRACT_LOWER, BUILT_IN_CHKP_EXTRACT_UPPER.
866 * common.opt (fcheck-pointer-bounds): New.
867 * toplev.c (process_options): Check Pointer Bounds Checker is
868 supported.
869 * doc/extend.texi: Document Pointer Bounds Checker built-in functions.
870
871 2013-10-30 Ilya Enkovich <ilya.enkovich@intel.com>
872
873 * target.def (builtin_chkp_function): New.
874 (chkp_bound_type): New.
875 (chkp_bound_mode): New.
876 (fn_abi_va_list_bounds_size): New.
877 (load_bounds_for_arg): New.
878 (store_bounds_for_arg): New.
879 * targhooks.h (default_load_bounds_for_arg): New.
880 (default_store_bounds_for_arg): New.
881 (default_fn_abi_va_list_bounds_size): New.
882 (default_chkp_bound_type): New.
883 (default_chkp_bound_mode): New.
884 (default_builtin_chkp_function): New.
885 * targhooks.c (default_load_bounds_for_arg): New.
886 (default_store_bounds_for_arg): New.
887 (default_fn_abi_va_list_bounds_size): New.
888 (default_chkp_bound_type): New.
889 (default_chkp_bound_mode); New.
890 (default_builtin_chkp_function): New.
891 * doc/tm.texi.in (TARGET_FN_ABI_VA_LIST_BOUNDS_SIZE): New.
892 (TARGET_LOAD_BOUNDS_FOR_ARG): New.
893 (TARGET_STORE_BOUNDS_FOR_ARG): New.
894 (TARGET_BUILTIN_CHKP_FUNCTION): New.
895 (TARGET_CHKP_BOUND_TYPE): New.
896 (TARGET_CHKP_BOUND_MODE): New.
897 * doc/tm.texi: Regenerated.
898 * langhooks.h (lang_hooks): Add chkp_supported field.
899 * langhooks-def.h (LANG_HOOKS_CHKP_SUPPORTED): New.
900 (LANG_HOOKS_INITIALIZER); Add LANG_HOOKS_CHKP_SUPPORTED.
901
902 2013-10-29 Andrew Pinski <apinski@cavium.com>
903
904 * tree-ssa-ifcombine.c: Include rtl.h and tm_p.h.
905 (ifcombine_ifandif): Handle cases where maybe_fold_and_comparisons
906 fails, combining the branches anyways.
907 (tree_ssa_ifcombine): Inverse the order of the basic block walk,
908 increases the number of combinings.
909 * gimple.h (gsi_start_nondebug_after_labels_bb): New function.
910
911 2013-10-29 Mike Stump <mikestump@comcast.net>
912
913 * machmode.def (PARTIAL_INT_MODE): Add precision and name.
914 * genmodes.c (PARTIAL_INT_MODE): Add precision and name.
915 (make_vector_mode): Increase namebuf to 16.
916 (emit_insn_modes_h): When processing BImode, don't
917 also match partial int modes.
918 (emit_class_narrowest_mode): Likewise.
919
920 * config/bfin/bfin-modes.def: Add precision to PDI.
921 * config/m32c/m32c-modes.def: Add precision to PSI.
922 * config/msp430/msp430-modes.def: Add precision to PSI.
923 * config/rs6000/rs6000-modes.def: Add precision to PTI.
924 * config/sh/sh-modes.def: Add precision to PSI and PDI.
925
926 2013-10-29 Oleg Endo <olegendo@gcc.gnu.org>
927
928 PR target/54236
929 * config/sh/sh.md (*addc): Rename existing variations to ...
930 (*addc_r_r_1, *addc_2r_1, *addc_r_1): ... these.
931 (*addc_r_lsb, *addc_r_r_lsb, *addc_r_lsb_r, *addc_2r_lsb, *addc_r_msb,
932 *addc_r_r_msb, *addc_2r_msb): New insn_and_split patterns.
933 * config/sh/sh.c (addsubcosts): Handle some addc special cases.
934
935 2013-10-29 Teresa Johnson <tejohnson@google.com>
936
937 PR ipa/58862
938 * tree-ssa-tail-merge.c (replace_block_by): Tolerate profile
939 insanities when updating probabilities.
940
941 2013-10-29 David Malcolm <dmalcolm@redhat.com>
942
943 * gdbhooks.py (CGraphNodePrinter.to_string): Update gdb
944 prettyprinter for cgraph_node to reflect the conversion of the
945 symtable types to a C++ class hierarchy: it now *is* a
946 symtab_node_base, rather than having one (named "symbol").
947
948 2013-10-29 Balaji V. Iyer <balaji.v.iyer@intel.com>
949
950 * builtins.c (is_builtin_name): Added a check for __cilkrts_detach and
951 __cilkrts_pop_frame. If matched, then return true for built-in
952 function name.
953 (expand_builtin): Added BUILT_IN_CILK_DETACH and
954 BUILT_IN_CILK_POP_FRAME case.
955 * langhooks-def.h (lhd_install_body_with_frame_cleanup): New prototype.
956 (lhs_cilk_detect_spawn): Likewise.
957 (LANG_HOOKS_DECLS): Added LANG_HOOKS_CILKPLUS.
958 (LANG_HOOKS_CILKPLUS_DETECT_SPAWN_AND_UNWRAP): New #define.
959 (LANG_HOOKS_CILKPLUS_FRAME_CLEANUP): Likewise.
960 (LANG_HOOKS_CILKPLUS_GIMPLIFY_SPAWN): Likewise.
961 (LANG_HOOKS_CILKPLUS): Likewise.
962 * tree.h (CILK_SPAWN_FN): Likewise.
963 * builtin.def (DEF_CILK_BUILTIN_STUB): Likewise.
964 * Makefile.in (C_COMMON_OBJS): Added c-family/cilk.o.
965 (OBJS): Added cilk-common.o.
966 (BUILTINS_DEF): Added cilk-builtins.def.
967 * langhooks.c (lhd_install_body_with_frame_cleanup): New function.
968 (lhd_cilk_detect_spawn): Likewise.
969 * langhooks.h (lang_hooks_for_cilkplus): New struct.
970 (struct lang_hooks): Added new field called "cilkplus."
971 * cilk-common.c: New file.
972 * cilk.h: Likewise.
973 * cilk-builtins.def: Likewise.
974 * cppbuiltin.c (define_builtin_macros_for_compilation_flags): Added
975 "__cilk" macro and set it to 200.
976 * function.h (struct function::cilk_frame_decl): New field.
977 (struct function::is_cilk_function): Likewise.
978 (struct function::calls_cilk_spawn): Likewise.
979 * gimplify.c (gimplify_call_expr): Added a check if the function call
980 being gimplified is a spawn detach point. If so, then add pop_frame
981 and detach function calls.
982 (gimplify_expr): Added a CILK_SPAWN_STMT and CILK_SYNC_STMT case
983 for gimplifying _Cilk_spawn and _Cilk_sync statements.
984 (gimplify_return_expr): Added a check for _Cilk_spawn usage in
985 function. If so, added a _Cilk_sync and gimplified it.
986 (gimplify_modify_expr): Added a check for _Cilk_spawn in MODIFY and
987 INIT_EXPRs. If so, then call gimplify_cilk_spawn.
988 * ipa-inline-analysis (initialize_inline_failed): Prevent inlining of
989 spawner function.
990 (can_inline_edge_p): Prevent inling of spawnee function.
991 * ira.c (ira_setup_eliminable_regset): Force usage of frame pointer
992 for functions that use Cilk keywords.
993 * tree-inline.h (struct copy_body_data::remap_var_for_cilk): New field.
994 * tree-pretty-print.c (dump_generic_node): Added CILK_SPAWN_STMT and
995 CILK_SYNC_STMT cases.
996 * tree.def (DEFTREECODE): Added CILK_SPAWN_STMT and CILK_SYNC_STMT
997 trees.
998 * generic.texi (CILK_SPAWN_STMT): Added documentation for _Cilk_spawn.
999 (CILK_SYNC_STMT): Added documentation for _Cilk_sync.
1000 * passes.texi (Cilk Keywords): New section that describes the compiler
1001 code changes for handling Cilk Keywords.
1002
1003 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1004
1005 Patch autogenerated by refactor_symtab.py from
1006 https://github.com/davidmalcolm/gcc-refactoring-scripts
1007 revision 58bb219cc090b2f4516a9297d868c245495ee622
1008
1009 * asan.c (asan_finish_file): Update for conversion of symtab types to
1010 a true class hierarchy.
1011 * cfgexpand.c (estimated_stack_frame_size): Likewise.
1012 * cgraph.c (cgraph_get_body): Likewise.
1013 (cgraph_get_create_real_symbol_node): Likewise.
1014 (verify_cgraph_node): Likewise.
1015 (verify_edge_corresponds_to_fndecl): Likewise.
1016 (verify_edge_count_and_frequency): Likewise.
1017 (cgraph_will_be_removed_from_program_if_no_direct_calls): Likewise.
1018 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
1019 (cgraph_can_remove_if_no_direct_calls_and_refs_p): Likewise.
1020 (cgraph_node_cannot_return): Likewise.
1021 (cgraph_set_pure_flag_1): Likewise.
1022 (cgraph_set_const_flag_1): Likewise.
1023 (cgraph_set_nothrow_flag_1): Likewise.
1024 (cgraph_make_node_local_1): Likewise.
1025 (cgraph_for_node_and_aliases): Likewise.
1026 (cgraph_for_node_thunks_and_aliases): Likewise.
1027 (cgraph_node_can_be_local_p): Likewise.
1028 (cgraph_node_cannot_be_local_p_1): Likewise.
1029 (cgraph_function_body_availability): Likewise.
1030 (dump_cgraph_node): Likewise.
1031 (cgraph_rtl_info): Likewise.
1032 (cgraph_mark_address_taken_node): Likewise.
1033 (cgraph_remove_node): Likewise.
1034 (cgraph_release_function_body): Likewise.
1035 (cgraph_update_edges_for_call_stmt_node): Likewise.
1036 (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
1037 (cgraph_make_edge_direct): Likewise.
1038 (cgraph_resolve_speculation): Likewise.
1039 (cgraph_speculative_call_info): Likewise.
1040 (cgraph_turn_edge_to_speculative): Likewise.
1041 (cgraph_create_edge_1): Likewise.
1042 (cgraph_set_call_stmt): Likewise.
1043 (cgraph_node_for_asm): Likewise.
1044 (cgraph_add_thunk): Likewise.
1045 (cgraph_same_body_alias): Likewise.
1046 (cgraph_create_function_alias): Likewise.
1047 (cgraph_create_node): Likewise.
1048 (cgraph_create_empty_node): Likewise.
1049 (record_function_versions): Likewise.
1050 (used_from_object_file_p): Likewise.
1051 * cgraph.h (symtab_can_be_discarded): Likewise.
1052 (symtab_real_symbol_p): Likewise.
1053 (cgraph_mark_force_output_node): Likewise.
1054 (cgraph_edge_recursive_p): Likewise.
1055 (symtab_alias_target): Likewise.
1056 (varpool_all_refs_explicit_p): Likewise.
1057 (varpool_can_remove_if_no_refs): Likewise.
1058 (cgraph_only_called_directly_or_aliased_p): Likewise.
1059 (cgraph_next_function_with_gimple_body): Likewise.
1060 (cgraph_first_function_with_gimple_body): Likewise.
1061 (cgraph_function_with_gimple_body_p): Likewise.
1062 (cgraph_next_function): Likewise.
1063 (cgraph_first_function): Likewise.
1064 (cgraph_next_defined_function): Likewise.
1065 (cgraph_first_defined_function): Likewise.
1066 (varpool_next_defined_variable): Likewise.
1067 (varpool_first_defined_variable): Likewise.
1068 (varpool_next_static_initializer): Likewise.
1069 (varpool_first_static_initializer): Likewise.
1070 (varpool_next_variable): Likewise.
1071 (varpool_first_variable): Likewise.
1072 (varpool_node_name): Likewise.
1073 (varpool): Likewise.
1074 (cgraph): Likewise.
1075 (is_a_helper <varpool_node>::test): Likewise.
1076 (is_a_helper <cgraph_node>::test): Likewise.
1077 (varpool_variable_node): Likewise.
1078 (cgraph_function_or_thunk_node): Likewise.
1079 (varpool_alias_target): Likewise.
1080 (cgraph_alias_target): Likewise.
1081 (cgraph_node_name): Likewise.
1082 (varpool_node_asm_name): Likewise.
1083 (cgraph_node_asm_name): Likewise.
1084 * cgraphbuild.c (remove_cgraph_callee_edges): Likewise.
1085 (cgraph_rebuild_references): Likewise.
1086 (rebuild_cgraph_edges): Likewise.
1087 (record_eh_tables): Likewise.
1088 (build_cgraph_edges): Likewise.
1089 (mark_store): Likewise.
1090 (mark_load): Likewise.
1091 (mark_address): Likewise.
1092 (record_type_list): Likewise.
1093 (record_reference): Likewise.
1094 * cgraphclones.c (cgraph_materialize_all_clones): Likewise.
1095 (cgraph_materialize_clone): Likewise.
1096 (cgraph_function_versioning): Likewise.
1097 (cgraph_copy_node_for_versioning): Likewise.
1098 (update_call_expr): Likewise.
1099 (cgraph_find_replacement_node): Likewise.
1100 (cgraph_create_virtual_clone): Likewise.
1101 (cgraph_clone_node): Likewise.
1102 * cgraphunit.c (compile): Likewise.
1103 (output_weakrefs): Likewise.
1104 (output_in_order): Likewise.
1105 (expand_function): Likewise.
1106 (assemble_thunks_and_aliases): Likewise.
1107 (expand_thunk): Likewise.
1108 (mark_functions_to_output): Likewise.
1109 (handle_alias_pairs): Likewise.
1110 (analyze_functions): Likewise.
1111 (walk_polymorphic_call_targets): Likewise.
1112 (varpool_finalize_decl): Likewise.
1113 (process_function_and_variable_attributes): Likewise.
1114 (cgraph_process_same_body_aliases): Likewise.
1115 (analyze_function): Likewise.
1116 (cgraph_add_new_function): Likewise.
1117 (cgraph_finalize_function): Likewise.
1118 (referred_to_p): Likewise.
1119 (cgraph_reset_node): Likewise.
1120 (cgraph_process_new_functions): Likewise.
1121 (enqueue_node): Likewise.
1122 (decide_is_symbol_needed): Likewise.
1123 * coverage.c (coverage_compute_profile_id): Likewise.
1124 * dbxout.c (dbxout_expand_expr): Likewise.
1125 * dwarf2out.c (premark_types_used_by_global_vars_helper): Likewise.
1126 (reference_to_unused): Likewise.
1127 * gimple-fold.c (can_refer_decl_in_current_unit_p): Likewise.
1128 * gimplify.c (unvisit_body): Likewise.
1129 (unshare_body): Likewise.
1130 * ipa-cp.c (ipcp_generate_summary): Likewise.
1131 (ipcp_decision_stage): Likewise.
1132 (identify_dead_nodes): Likewise.
1133 (decide_whether_version_node): Likewise.
1134 (decide_about_value): Likewise.
1135 (perhaps_add_new_callers): Likewise.
1136 (create_specialized_node): Likewise.
1137 (update_profiling_info): Likewise.
1138 (ipcp_propagate_stage): Likewise.
1139 (estimate_local_effects): Likewise.
1140 (good_cloning_opportunity_p): Likewise.
1141 (devirtualization_time_bonus): Likewise.
1142 (propagate_constants_accross_call): Likewise.
1143 (initialize_node_lattices): Likewise.
1144 (ipcp_cloning_candidate_p): Likewise.
1145 (determine_versionability): Likewise.
1146 (print_all_lattices): Likewise.
1147 (print_lattice): Likewise.
1148 (ipcp_discover_new_direct_edges): Likewise.
1149 * ipa-devirt.c (ipa_devirt): Likewise.
1150 (likely_target_p): Likewise.
1151 (update_type_inheritance_graph): Likewise.
1152 (possible_polymorphic_call_target_p): Likewise.
1153 (dump_possible_polymorphic_call_targets): Likewise.
1154 (devirt_variable_node_removal_hook): Likewise.
1155 (record_binfo): Likewise.
1156 (maybe_record_node): Likewise.
1157 (build_type_inheritance_graph): Likewise.
1158 * ipa-inline-analysis.c (inline_write_summary): Likewise.
1159 (inline_generate_summary): Likewise.
1160 (inline_analyze_function): Likewise.
1161 (do_estimate_growth): Likewise.
1162 (simple_edge_hints): Likewise.
1163 (estimate_node_size_and_time): Likewise.
1164 (estimate_edge_devirt_benefit): Likewise.
1165 (compute_inline_parameters): Likewise.
1166 (estimate_function_body_sizes): Likewise.
1167 (compute_bb_predicates): Likewise.
1168 (initialize_inline_failed): Likewise.
1169 (dump_inline_summary): Likewise.
1170 (dump_inline_edge_summary): Likewise.
1171 * ipa-inline-transform.c (inline_transform): Likewise.
1172 (preserve_function_body_p): Likewise.
1173 (save_inline_function_body): Likewise.
1174 (inline_call): Likewise.
1175 (clone_inlined_nodes): Likewise.
1176 (can_remove_node_now_p): Likewise.
1177 (can_remove_node_now_p_1): Likewise.
1178 * ipa-inline.c (early_inliner): Likewise.
1179 (early_inline_small_functions): Likewise.
1180 (inline_always_inline_functions): Likewise.
1181 (ipa_inline): Likewise.
1182 (flatten_function): Likewise.
1183 (inline_small_functions): Likewise.
1184 (speculation_useful_p): Likewise.
1185 (recursive_inlining): Likewise.
1186 (update_caller_keys): Likewise.
1187 (reset_edge_caches): Likewise.
1188 (update_edge_key): Likewise.
1189 (edge_badness): Likewise.
1190 (relative_time_benefit): Likewise.
1191 (want_inline_self_recursive_call_p): Likewise.
1192 (want_inline_small_function_p): Likewise.
1193 (want_early_inline_function_p): Likewise.
1194 (num_calls): Likewise.
1195 (can_early_inline_edge_p): Likewise.
1196 (can_inline_edge_p): Likewise.
1197 (report_inline_failed_reason): Likewise.
1198 * ipa-profile.c (ipa_profile): Likewise.
1199 (ipa_propagate_frequency): Likewise.
1200 (ipa_propagate_frequency_1): Likewise.
1201 (ipa_profile_generate_summary): Likewise.
1202 * ipa-prop.c (ipcp_transform_function): Likewise.
1203 (read_replacements_section): Likewise.
1204 (ipa_prop_read_section): Likewise.
1205 (ipa_modify_call_arguments): Likewise.
1206 (ipa_print_node_params): Likewise.
1207 (propagate_controlled_uses): Likewise.
1208 (update_indirect_edges_after_inlining): Likewise.
1209 (remove_described_reference): Likewise.
1210 (ipa_make_edge_direct_to_target): Likewise.
1211 (ipa_analyze_node): Likewise.
1212 (ipa_analyze_params_uses): Likewise.
1213 (ipa_compute_jump_functions): Likewise.
1214 (ipa_get_callee_param_type): Likewise.
1215 (ipa_print_node_jump_functions): Likewise.
1216 (ipa_initialize_node_params): Likewise.
1217 (ipa_populate_param_decls): Likewise.
1218 (ipa_func_spec_opts_forbid_analysis_p): Likewise.
1219 (write_agg_replacement_chain): Likewise.
1220 (ipa_write_node_info): Likewise.
1221 (ipa_edge_duplication_hook): Likewise.
1222 (try_decrement_rdesc_refcount): Likewise.
1223 * ipa-pure-const.c (propagate_nothrow): Likewise.
1224 (propagate_pure_const): Likewise.
1225 (pure_const_read_summary): Likewise.
1226 (pure_const_write_summary): Likewise.
1227 (analyze_function): Likewise.
1228 * ipa-ref-inline.h (ipa_ref_referred_ref_list): Likewise.
1229 (ipa_ref_referring_ref_list): Likewise.
1230 * ipa-ref.c (ipa_clear_stmts_in_references): Likewise.
1231 (ipa_remove_stmt_references): Likewise.
1232 (ipa_find_reference): Likewise.
1233 (ipa_dump_referring): Likewise.
1234 (ipa_dump_references): Likewise.
1235 (ipa_record_reference): Likewise.
1236 * ipa-reference.c (ipa_reference_read_optimization_summary): Likewise.
1237 (ipa_reference_write_optimization_summary): Likewise.
1238 (write_node_summary_p): Likewise.
1239 (propagate): Likewise.
1240 (read_write_all_from_decl): Likewise.
1241 (generate_summary): Likewise.
1242 (analyze_function): Likewise.
1243 (propagate_bits): Likewise.
1244 (ipa_reference_get_not_written_global): Likewise.
1245 (ipa_reference_get_not_read_global): Likewise.
1246 * ipa-split.c (execute_split_functions): Likewise.
1247 (split_function): Likewise.
1248 * ipa-utils.c (ipa_merge_profiles): Likewise.
1249 (dump_cgraph_node_set): Likewise.
1250 (ipa_reverse_postorder): Likewise.
1251 (ipa_edge_within_scc): Likewise.
1252 (ipa_get_nodes_in_cycle): Likewise.
1253 (ipa_free_postorder_info): Likewise.
1254 (ipa_reduced_postorder): Likewise.
1255 (searchc): Likewise.
1256 (recursive_call_p): Likewise.
1257 * ipa.c (ipa_cdtor_merge): Likewise.
1258 (record_cdtor_fn): Likewise.
1259 (function_and_variable_visibility): Likewise.
1260 (varpool_externally_visible_p): Likewise.
1261 (cgraph_externally_visible_p): Likewise.
1262 (comdat_can_be_unshared_p): Likewise.
1263 (comdat_can_be_unshared_p_1): Likewise.
1264 (address_taken_from_non_vtable_p): Likewise.
1265 (ipa_discover_readonly_nonaddressable_vars): Likewise.
1266 (symtab_remove_unreachable_nodes): Likewise.
1267 (walk_polymorphic_call_targets): Likewise.
1268 (process_references): Likewise.
1269 (enqueue_node): Likewise.
1270 (has_addr_references_p): Likewise.
1271 (cgraph_non_local_node_p_1): Likewise.
1272 * is-a.h (varpool_analyze_node): Likewise.
1273 * lto-cgraph.c (input_symtab): Likewise.
1274 (merge_profile_summaries): Likewise.
1275 (input_cgraph_1): Likewise.
1276 (input_edge): Likewise.
1277 (input_varpool_node): Likewise.
1278 (input_node): Likewise.
1279 (input_overwrite_node): Likewise.
1280 (compute_ltrans_boundary): Likewise.
1281 (output_refs): Likewise.
1282 (lto_output_varpool_node): Likewise.
1283 (lto_output_node): Likewise.
1284 (reachable_from_other_partition_p): Likewise.
1285 (referenced_from_other_partition_p): Likewise.
1286 (lto_output_edge): Likewise.
1287 (output_node_opt_summary): Likewise.
1288 (add_node_to): Likewise.
1289 (reachable_from_this_partition_p): Likewise.
1290 (lto_set_symtab_encoder_in_partition): Likewise.
1291 (lto_symtab_encoder_in_partition_p): Likewise.
1292 (lto_set_symtab_encoder_encode_initializer): Likewise.
1293 (lto_symtab_encoder_encode_initializer_p): Likewise.
1294 (lto_set_symtab_encoder_encode_body): Likewise.
1295 (lto_symtab_encoder_encode_body_p): Likewise.
1296 * lto-section-in.c (lto_free_function_in_decl_state_for_node):
1297 Likewise.
1298 * lto-streamer-in.c (lto_read_body): Likewise.
1299 (fixup_call_stmt_edges): Likewise.
1300 (fixup_call_stmt_edges_1): Likewise.
1301 * lto-streamer-out.c (produce_symtab): Likewise.
1302 (output_symbol_p): Likewise.
1303 (write_symbol): Likewise.
1304 (lto_output): Likewise.
1305 (copy_function): Likewise.
1306 (output_function): Likewise.
1307 * passes.c (function_called_by_processed_nodes_p): Likewise.
1308 (ipa_write_optimization_summaries): Likewise.
1309 (ipa_write_summaries): Likewise.
1310 (do_per_function_toporder): Likewise.
1311 (do_per_function): Likewise.
1312 (dump_passes): Likewise.
1313 * symtab.c (symtab_semantically_equivalent_p): Likewise.
1314 (symtab_nonoverwritable_alias): Likewise.
1315 (symtab_nonoverwritable_alias_1): Likewise.
1316 (symtab_for_node_and_aliases): Likewise.
1317 (symtab_resolve_alias): Likewise.
1318 (fixup_same_cpp_alias_visibility): Likewise.
1319 (symtab_alias_ultimate_target): Likewise.
1320 (symtab_used_from_object_file_p): Likewise.
1321 (verify_symtab_base): Likewise.
1322 (dump_symtab_base): Likewise.
1323 (symtab_node_name): Likewise.
1324 (symtab_node_asm_name): Likewise.
1325 (symtab_dissolve_same_comdat_group_list): Likewise.
1326 (symtab_add_to_same_comdat_group): Likewise.
1327 (symtab_unregister_node): Likewise.
1328 (symtab_insert_node_to_hashtable): Likewise.
1329 (symtab_register_node): Likewise.
1330 (unlink_from_assembler_name_hash): Likewise.
1331 (insert_to_assembler_name_hash): Likewise.
1332 (eq_assembler_name): Likewise.
1333 (hash_node_by_assembler_name): Likewise.
1334 (eq_node): Likewise.
1335 (hash_node): Likewise.
1336 * toplev.c (wrapup_global_declaration_2): Likewise.
1337 * trans-mem.c (ipa_tm_execute): Likewise.
1338 (ipa_tm_transform_clone): Likewise.
1339 (ipa_tm_transform_transaction): Likewise.
1340 (ipa_tm_transform_calls_redirect): Likewise.
1341 (ipa_tm_insert_gettmclone_call): Likewise.
1342 (ipa_tm_insert_irr_call): Likewise.
1343 (ipa_tm_create_version): Likewise.
1344 (ipa_tm_create_version_alias): Likewise.
1345 (ipa_tm_mark_forced_by_abi_node): Likewise.
1346 (ipa_tm_mark_force_output_node): Likewise.
1347 (ipa_tm_diagnose_tm_safe): Likewise.
1348 (ipa_tm_mayenterirr_function): Likewise.
1349 (ipa_tm_scan_irr_function): Likewise.
1350 (ipa_tm_note_irrevocable): Likewise.
1351 (ipa_tm_scan_calls_clone): Likewise.
1352 (get_cg_data): Likewise.
1353 * tree-eh.c (tree_could_trap_p): Likewise.
1354 * tree-emutls.c (ipa_lower_emutls): Likewise.
1355 (create_emultls_var): Likewise.
1356 (lower_emutls_function_body): Likewise.
1357 (gen_emutls_addr): Likewise.
1358 (emutls_decl): Likewise.
1359 (new_emutls_decl): Likewise.
1360 * tree-inline.c (tree_function_versioning): Likewise.
1361 (optimize_inline_calls): Likewise.
1362 (expand_call_inline): Likewise.
1363 (estimate_num_insns): Likewise.
1364 (copy_bb): Likewise.
1365 (delete_unreachable_blocks_update_callgraph): Likewise.
1366 * tree-nested.c (gimplify_all_functions): Likewise.
1367 (create_nesting_tree): Likewise.
1368 (check_for_nested_with_variably_modified): Likewise.
1369 * tree-pretty-print.c (dump_function_header): Likewise.
1370 * tree-profile.c (tree_profiling): Likewise.
1371 * tree-sra.c (ipa_sra_preliminary_function_checks): Likewise.
1372 (modify_function): Likewise.
1373 (convert_callers): Likewise.
1374 (convert_callers_for_node): Likewise.
1375 * tree-ssa-structalias.c (ipa_pta_execute): Likewise.
1376 (associate_varinfo_to_alias): Likewise.
1377 (create_variable_info_for): Likewise.
1378 (get_constraint_for_ssa_var): Likewise.
1379 * tree-vectorizer.c (increase_alignment): Likewise.
1380 * tree.c (find_decls_types_in_var): Likewise.
1381 (find_decls_types_in_node): Likewise.
1382 (free_lang_data_in_decl): Likewise.
1383 * value-prof.c (gimple_ic_transform): Likewise.
1384 (gimple_ic): Likewise.
1385 (check_ic_target): Likewise.
1386 (init_node_map): Likewise.
1387 * varasm.c (decl_binds_to_current_def_p): Likewise.
1388 (default_binds_local_p_1): Likewise.
1389 (dump_tm_clone_pairs): Likewise.
1390 (assemble_alias): Likewise.
1391 (find_decl): Likewise.
1392 (mark_decl_referenced): Likewise.
1393 * varpool.c (varpool_for_node_and_aliases): Likewise.
1394 (varpool_extra_name_alias): Likewise.
1395 (varpool_create_variable_alias): Likewise.
1396 (add_new_static_var): Likewise.
1397 (varpool_finalize_named_section_flags): Likewise.
1398 (varpool_remove_unreferenced_decls): Likewise.
1399 (enqueue_node): Likewise.
1400 (varpool_assemble_decl): Likewise.
1401 (assemble_aliases): Likewise.
1402 (varpool_analyze_node): Likewise.
1403 (cgraph_variable_initializer_availability): Likewise.
1404 (varpool_add_new_variable): Likewise.
1405 (ctor_for_folding): Likewise.
1406 (dump_varpool_node): Likewise.
1407 (varpool_remove_initializer): Likewise.
1408 (varpool_remove_node): Likewise.
1409 (varpool_node_for_decl): Likewise.
1410 (varpool_create_empty_node): Likewise.
1411 * config/i386/i386.c (ix86_generate_version_dispatcher_body): Likewise.
1412 (ix86_get_function_versions_dispatcher): Likewise.
1413
1414 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1415
1416 * cgraph.h (symtab_node_base): Convert to a class;
1417 add GTY((desc ("%h.type"), tag ("SYMTAB_SYMBOL"))), and take
1418 chain_next/prev from symtab_node_def.
1419 (cgraph_node): Inherit from symtab_node; add GTY option
1420 tag ("SYMTAB_FUNCTION").
1421 (varpool_node): Inherit from symtab_node; add GTY option
1422 tag ("SYMTAB_VARIABLE").
1423 (symtab_node_def): Remove.
1424 (is_a_helper <cgraph_node>::test (symtab_node_def *)): Convert to...
1425 (is_a_helper <cgraph_node>::test (symtab_node_base *)): ...this.
1426 (is_a_helper <varpool_node>::test (symtab_node_def *)): Convert to...
1427 (is_a_helper <varpool_node>::test (symtab_node_base *)): ...this.
1428
1429 * ipa-ref.h (symtab_node_def): Drop.
1430 (symtab_node): Change underlying type from symtab_node_def to
1431 symtab_node_base.
1432 (const_symtab_node): Likwise.
1433
1434 * is-a.h: Update examples in comment.
1435
1436 * symtab.c (symtab_hash): Change symtab_node_def to symtab_node_base.
1437 (assembler_name_hash): Likewise.
1438
1439 2013-10-29 Martin Liska <marxin.liska@gmail.com>
1440
1441 * doc/tree-ssa.texi (gimple_phi_result): Document.
1442 (gimple_phi_num_args, gimple_phi_arg): Likewise.
1443 (gimple_phi_arg_edge, gimple_phi_arg_def): Likewise.
1444 (PHI_RESULT, PHI_NUM_ARGS): Remove.
1445 (PHI_ARG_ELT, PHI_ARG_EDGE, PHI_ARG_DEF): Likewise.
1446
1447 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1448
1449 * expr.h: Revert change and include tree-core.h.
1450 * rtl.h: Revert change and don't include tree-core.h.
1451
1452 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1453
1454 * config/darwin.c: Include gimple.h.
1455 * config/i386/winnt.c: Likewise.
1456
1457 2013-10-29 Marc Glisse <marc.glisse@inria.fr>
1458
1459 PR tree-optimization/19831
1460 * tree-ssa-alias.c (stmt_kills_ref_p_1): Handle BUILT_IN_FREE.
1461
1462 2013-10-29 Andrew MacLeod <amacleod@redhat.com>
1463
1464 * tree-outof-ssa.h: Remove include files.
1465 * tree-outof-ssa.c: Add required include files from tree-outof-ssa.h.
1466 * expr.c: Likewise.
1467 * tree-ssa-coalesce.c: Likewise.
1468 * cfgexpand.c: Likewise.
1469 * tree-ssa-ter.c: Likewise.
1470 * ipa-prop.h: Remove gimple.h and tree-core.h from include list.
1471 * lto-streamer.h: Likewise.
1472 * cgraphbuild.c: Add gimple.h to include list.
1473 * data-streamer-in.c: Likewise.
1474 * ipa-cp.c: Likewise.
1475 * tree-streamer.c: Likewise.
1476 * lto-compress.c: Likewise.
1477 * ipa-reference.c: Likewise.
1478 * data-streamer-out.c: Likewise.
1479 * lto-cgraph.c: Likewise.
1480 * cgraphclones.c: Likewise.
1481 * ipa-utils.c: Likewise.
1482 * data-streamer.c: Likewise.
1483 * ipa-split.c: Likewise.
1484 * lto-section-in.c: Likewise.
1485 * tree-streamer-out.c: Likewise.
1486 * ipa-prop.c: Likewise.
1487 * tree-streamer-in.c: Likewise.
1488 * symtab.c: Likewise.
1489 * opts-global.c: Likewise.
1490 * lto-opts.c: Likewise.
1491 * lto-section-out.c: Likewise.
1492 * lto-streamer.c: Likewise.
1493 * rtl.h: Add tree-core.h to include list.
1494 * expr.h: Remove tree-core.h from include list.
1495 * gimple.h: Likewise.
1496 * ipa-utils.h: Likewise.
1497 * streamer-hooks.h: Likewise.
1498 * streamer-hooks.c: Include input.h.
1499
1500 2013-10-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1501
1502 * config/arm/arm.c (cortexa7_extra_costs): New table.
1503 (arm_cortex_a7_tune): New.
1504 * config/arm/arm-cores.def: Use cortex_a7 tuning for cortex-a7.
1505
1506 2013-10-29 Eric Botcazou <ebotcazou@adacore.com>
1507
1508 * expr.c (expand_expr_real_1) <MEM_EXPR>: Eliminate small redundancy.
1509
1510 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1511
1512 * doc/gty.texi ("Inheritance and GTY"): Make it clear that
1513 to use autogenerated markers for a class-hierarchy, every class
1514 must have a GTY marker.
1515 * gengtype.h (struct type): Add linked list of subclasses to
1516 the "s" member of the union.
1517 (add_subclass): New decl.
1518 * gengtype-state.c (read_state_struct_type): Set up subclass
1519 linked list.
1520 * gengtype.c (get_ultimate_base_class): New.
1521 (add_subclass): New.
1522 (new_structure): Set up subclass linked list.
1523 (set_gc_used_type): Propagate usage information to subclasses.
1524 (output_mangled_typename): Use get_ultimate_base_class.
1525 (walk_subclasses): Use the subclass linked list, avoiding an
1526 O(N^2) when writing out all types.
1527 (walk_type): Issue an error if the base class is missing a tag,
1528 rather than generating bogus C code. Add a gcc_unreachable
1529 default case, in case people omit tags from concrete subclasses,
1530 or get the values wrong.
1531 (write_func_for_structure): Issue an error for subclasses for
1532 which the base doesn't have a "desc", since otherwise the
1533 autogenerated routines for the base would silently fail to visit
1534 any subclass fields.
1535 (write_root): Use get_ultimate_base_class, tweaking constness of
1536 tp to match that function's signature.
1537
1538 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1539
1540 * doc/gty.texi (GTY Options): Add note about inheritance to
1541 description of desc and tag.
1542 (Inheritance and GTY): New.
1543
1544 2013-10-29 David Malcolm <dmalcolm@redhat.com>
1545
1546 * gengtype-parse.c (opts_have): Drop "static" so that
1547 we can use this from gengtype.c.
1548 * gengtype.c (set_gc_used_type): Mark any base class as used;
1549 update field traversal to visit inherited fields.
1550 (output_mangled_typename): Convert references to classes within
1551 an inheritance hierarchy to reference the ultimate base class,
1552 since only it will have gt_ functions.
1553 (get_string_option): New.
1554 (walk_subclasses): New.
1555 (walk_type): Treat GTY structs that have a "desc" as being the
1556 root of an inheritance hierarchy. Generate a switch on it
1557 within the marking function which walks all subclasses, adding
1558 cases for them via walk_subclasses. For subclasses, visit all
1559 fields of the type (including inherited ones).
1560 (write_func_for_structure): Don't write fns for subclasses, only
1561 for the ultimate base class within an inheritance hierarchy.
1562 Subclasses-marking will be handled by the base class marking functions.
1563 (write_types): Likewise.
1564 (write_local_func_for_structure): Likewise.
1565 (USED_BY_TYPED_GC_P): Emit allocators for subclasses that have
1566 a "tag" option (and are thus concrete subclasses).
1567 (write_root): Use the marker function for the ultimate base class.
1568 * gengtype.h (FOR_ALL_INHERITED_FIELDS): New.
1569 (opts_have): Add declaration.
1570
1571 2013-10-28 Vladimir Makarov <vmakarov@redhat.com>
1572
1573 * lra-spills.c (lra_final_code_change): Remove useless move insns
1574 originated from moves of pseudos.
1575
1576 2013-10-28 Jeff Law <law@redhat.com>
1577
1578 * lower-subreg.c (resolve_simple_move): Fix comment typo.
1579
1580 2013-10-28 Trevor Saunders <tsaunders@mozilla.com>
1581
1582 * df-scan.c (df_collection_rec): Adjust.
1583 (copy_defs): New constant.
1584 (copy_uses): Likewise.
1585 (copy_eq_uses): Likewise.
1586 (copy_mw): Likewise.
1587 (copy_all): Likewise.
1588 (df_insn_rescan): Adjust.
1589 (df_notes_rescan): Likewise.
1590 (df_swap_refs): Likewise.
1591 (df_sort_and_compress_refs): Likewise.
1592 (df_sort_and_compress_mws): Likewise.
1593 (df_install_refs): Likewise.
1594 (df_install_mws): Likewise.
1595 (df_refs_add_to_chains): Add flags parameter controlling which vectors
1596 are coppied.
1597 (df_bb_refs_record): Adjust.
1598 (df_record_entry_block_defs): Likewise.
1599 (df_record_exit_block_defs): Likewise.
1600 (df_refs_verify): Likewise.
1601 (df_mws_verify): Likewise.
1602 (df_insn_refs_verify): Likewise.
1603 (df_bb_verify): Likewise.
1604 * ipa-pure-const.c (finish_state): Remove.
1605 (propagate): Adjust.
1606 * tree-data-ref.c tree-ssa-alias.c tree-ssa-loop-ivcanon.c
1607 tree-ssa-threadedge.c tree-vect-loop-manip.c tree-vect-slp.c
1608 var-tracking.c: Adjust.
1609 * vec.c (stack_vecs): Remove.
1610 (register_stack_vec): Likewise.
1611 (stack_vec_register_index): Likewise.
1612 (unregister_stack_vec): Likewise.
1613 * vec.h (struct va_stack): Remove.
1614 (struct vec<T, A, vl_ptr>): Specialize as
1615 struct vec<T, va_heap, vl_ptr> instead since va_heap is the only
1616 allocation strategy compatable with the vl_ptr layout.
1617 (struct vec<T, va_gc, vl_ptr>): Remove because it now gets an empty
1618 specialization anyway.
1619 (class stack_vec): New class.
1620 (vec_stack_alloc): Remove.
1621 (vec<T, va_heap, vl_ptr>::using_auto_storage): New method.
1622
1623 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1624 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1625 Sergey Lega <sergey.s.lega@intel.com>
1626 Anna Tikhonova <anna.tikhonova@intel.com>
1627 Ilya Tocar <ilya.tocar@intel.com>
1628 Andrey Turetskiy <andrey.turetskiy@intel.com>
1629 Ilya Verbin <ilya.verbin@intel.com>
1630 Kirill Yukhin <kirill.yukhin@intel.com>
1631 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1632
1633 * config/i386/i386.md (prefetch): Allow TARGET_AVX512PF.
1634 (*prefetch_avx512pf_<mode>): New.
1635 * config/i386/sse.md (avx512f_vmcmp<mode>3): Ditto.
1636 (avx512f_maskcmp<mode>3): Ditto.
1637 (vashrv16si3): Ditto.
1638
1639 2013-10-28 Alexander Ivchenko <alexander.ivchenko@intel.com>
1640 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1641 Sergey Lega <sergey.s.lega@intel.com>
1642 Anna Tikhonova <anna.tikhonova@intel.com>
1643 Ilya Tocar <ilya.tocar@intel.com>
1644 Andrey Turetskiy <andrey.turetskiy@intel.com>
1645 Ilya Verbin <ilya.verbin@intel.com>
1646 Kirill Yukhin <kirill.yukhin@intel.com>
1647 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1648
1649 * config/i386/i386.md (any_truncate): New.
1650 (trunsuffix): Ditto.
1651 * config/i386/predicates.md (const_8_to_9_operand): New.
1652 (const_10_to_11_operand): Ditto.
1653 (const_12_to_13_operand): Ditto.
1654 (const_14_to_15_operand): Ditto.
1655 (const_16_to_19_operand): Ditto.
1656 (const_20_to_23_operand): Ditto.
1657 (const_24_to_27_operand): Ditto.
1658 (const_28_to_31_operand): Ditto.
1659 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_FIX_NOTRUNC.
1660 (cvtusi2<ssescalarmodesuffix>32): New.
1661 (cvtusi2<ssescalarmodesuffix>64): Ditto.
1662 (ufloatv16siv16sf2): Ditto.
1663 (avx512f_fix_notruncv16sfv16si): Ditto.
1664 (avx512f_ufix_notruncv16sfv16si): Ditto.
1665 (avx512f_vcvtss2usi): Ditto.
1666 (avx512f_vcvtss2usiq): Ditto.
1667 (avx512f_vcvttss2usi): Ditto.
1668 (avx512f_vcvttss2usiq): Ditto.
1669 (avx512f_vcvtsd2usi): Ditto.
1670 (avx512f_vcvtsd2usiq): Ditto.
1671 (avx512f_vcvttsd2usi): Ditto.
1672 (avx512f_vcvttsd2usiq): Ditto.
1673 (ufloatv8siv8df): Ditto.
1674 (avx512f_cvtdq2pd512_2): Ditto.
1675 (avx512f_cvtpd2dq512): Ditto.
1676 (avx512f_ufix_notruncv8dfv8si): Ditto.
1677 (avx512f_cvtpd2ps512): Ditto.
1678 (vec_unpacks_lo_v16sf): Ditto.
1679 (vec_unpacks_hi_v16sf): Ditto.
1680 (vec_unpacks_float_hi_v16si): Ditto.
1681 (vec_unpacks_float_lo_v16si): Ditto.
1682 (avx512f_unpckhps512): Ditto.
1683 (avx512f_unpcklps512): Ditto.
1684 (avx512f_movshdup512): Ditto.
1685 (avx512f_movsldup512): Ditto.
1686 (vec_extract_lo_v32hi): Ditto.
1687 (vec_extract_hi_v32hi): Ditto.
1688 (vec_extract_lo_v64qi): Ditto.
1689 (vec_extract_hi_v64qi): Ditto.
1690 (avx512f_unpckhpd512): Ditto.
1691 (avx512f_movddup512): Ditto.
1692 (avx512f_unpcklpd512): Ditto.
1693 (*avx512f_unpcklpd512): Ditto.
1694 (avx512f_shufps512_1): Ditto.
1695 (avx512f_shufpd512_1): Ditto.
1696 (avx512f_interleave_highv8di): Ditto.
1697 (avx512f_interleave_lowv8di): Ditto.
1698 (PMOV_DST_MODE): Ditto.
1699 (pmov_src_mode): Ditto.
1700 (pmov_src_lower): Ditto.
1701 (pmov_suff): Ditto.
1702 (*avx512f_<code><pmov_src_lower><mode>2): Ditto.
1703 (*avx512f_<code>v8div16qi2): Ditto.
1704 (*avx512f_<code>v8div16qi2_store): Ditto.
1705 (vec_widen_umult_even_v16si): Ditto.
1706 (*vec_widen_umult_even_v16si): Ditto.
1707 (vec_widen_smult_even_v16si): Ditto.
1708 (*vec_widen_smult_even_v16si): Ditto.
1709 (avx512f_interleave_highv16si): Ditto.
1710 (avx512f_interleave_lowv16si): Ditto.
1711 (avx512f_<code>v16qiv16si2): Ditto.
1712 (avx512f_<code>v16hiv16si2): Ditto.
1713 (avx512f_<code>v8qiv8di2): Ditto.
1714 (avx512f_<code>v8hiv8di2): Ditto.
1715 (avx512f_<code>v8siv8di2): Ditto.
1716 (avx512cd_maskb_vec_dupv8di): Ditto.
1717 (avx512cd_maskw_vec_dupv16si): Ditto.
1718 (avx512f_vcvtph2ps512): Ditto.
1719 (avx512f_vcvtps2ph512): Ditto.
1720 (VEC_EXTRACT_MODE): Extened with wider modes.
1721 (VEC_PERM_AVX2): Ditto.
1722 (VEC_PERM_CONST): Ditto.
1723
1724 2013-10-28 Joern Rennecke <joern.rennecke@embecosm.com>
1725
1726 * config/arc/arc.c (arc_ccfsm_post_advance):
1727 Add comment about TYPE_RETURN.
1728
1729 2013-10-28 Bin Cheng <bin.cheng@arm.com>
1730
1731 * tree-ssa-loop-ivopts.c (strip_offset_1): Change parameter type.
1732 Count DECL_FIELD_BIT_OFFSET for COMPONENT_REF.
1733 (strip_offset): Convert offset to unsigned number.
1734
1735 2013-10-27 Tom de Vries <tom@codesourcery.com>
1736
1737 * cfgexpand.c (gimple_expand_cfg): Remove test for parm_birth_insn.
1738 Don't commit insertions after NOTE_INSN_FUNCTION_BEG.
1739
1740 2013-10-27 Oleg Endo <olegendo@gcc.gnu.org>
1741
1742 * config/sh/sh.c (MSW, LSW): Move and rename macros to...
1743 * config/sh/sh.h (SH_REG_MSW_OFFSET, SH_REG_LSW_OFFSET): ... here.
1744 (TARGET_BIG_ENDIAN): New macro.
1745 * config/sh/sh.md: Use it instead of !TARGET_LITTLE_ENDIAN.
1746 Use SH_REG_MSW_OFFSET and SH_REG_LSW_OFFSET.
1747 * config/sh/sh.c: Likewise.
1748 * config/sh/sh.h: Likewise.
1749
1750 2013-10-27 Hans-Peter Nilsson <hp@axis.com>
1751
1752 * config/cris/cris.c (cris_emit_trap_for_misalignment): Replace the
1753 removed PRED_MUDFLAP with PRED_NORETURN. Correct file-path in comment.
1754
1755 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1756
1757 * config/sh/sh.md (movmemsi): Remove empty constraints and predicates.
1758 Fix formatting.
1759 (cmpstr_t, cmpstrsi): Fix formatting.
1760
1761 2013-10-26 Oleg Endo <olegendo@gcc.gnu.org>
1762
1763 PR target/52483
1764 * config/sh/predicates.md (general_movdst_operand): Allow reg+reg
1765 addressing, do not use general_operand for memory operands.
1766
1767 2013-10-26 Vladimir Makarov <vmakarov@redhat.com>
1768
1769 Revert:
1770 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1771 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1772
1773 2013-10-26 Jeff Law <law@redhat.com>
1774
1775 * predict.c (PRED_MUDFLAP): Remove.
1776 * targhooks.c (build_va_arg_indirect_ref): Remove mudflap support.
1777
1778 * Makefile.in (C_COMMON_OBJS): Remove tree-mudflap.
1779 (OBJS): Remove tree-nomudflap.o
1780 (GTFILES): Remove tree-mudflap.c
1781 * builtins.c (expand_builtin_alloc): Remove mudflap support.
1782 * gcc.c (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1783 (mfwrap_spec, mflib_spec): Likewise.
1784 (cpp_unique_options, cc1_options, static_specs): Likewise.
1785 * gimplify (gimplify_vla_decl, build_va_arg_indirect_ref): Likewise.
1786 * passes.def: Likewise.
1787 * toplev.c (compile_file, process_options): Likewise.
1788 * tree-inline.c (copy_tree_r): Likewise.
1789 * tree-pass.,h (make_pass_mudflap_1, make_pass_mudflap_2): Likewise.
1790 * varasm.c (make_decl_rtl, make_decl_rtl_for_debug): Likewise.
1791 (build_constant_desc, output_constant_def_contents): Likewise.
1792 (categorize_decl_for_section): Likewise.
1793 * tree-mudflap.c: Removed.
1794 * tree-mudflap.h: Removed.
1795 * tree-nomudflap.c: Removed.
1796 * bfin/uclinux.h (MFWRAP_SPEC): Remove.
1797 * moxie/uclinux.h (MFWRAP_SPEC): Likewise.
1798 * rs6000/aix.h (MFWRAP_SPEC, MFLIB_SPEC): Likewise.
1799 * config/sol2.h (MFLIB_SPEC): Likewise.
1800 * doc/install.texi: Remove mudflap references.
1801 * doc/passes.texi: Similarly.
1802 * doc/sourcebuild.texi: Similarly.
1803 * doc/invoke.texi: Remove mudlfap related options.
1804
1805 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1806
1807 PR rtl-optimization/58759
1808 * lra-constraints.c (lra_constraints): Remove wrong condition to
1809 remove insn setting up an equivalent pseudo.
1810
1811 2013-10-25 Vladimir Makarov <vmakarov@redhat.com>
1812
1813 * config/rs6000/rs6000-protos.h
1814 (rs6000_secondary_memory_needed_mode): New prototype.
1815 * config/rs6000/rs6000.c: Include ira.h.
1816 (TARGET_LRA_P): Redefine.
1817 (rs6000_legitimate_offset_address_p): Call
1818 legitimate_constant_pool_address_p in strict mode for LRA.
1819 (rs6000_legitimate_address_p): Ditto.
1820 (legitimate_lo_sum_address_p): Add code for LRA. Use lra_in_progress.
1821 (rs6000_emit_move): Add LRA version of code to generate load/store
1822 of SDmode values.
1823 (rs6000_secondary_memory_needed_mode): New.
1824 (rs6000_alloc_sdmode_stack_slot): Do nothing for LRA.
1825 (rs6000_secondary_reload_class): Return NO_REGS for LRA for
1826 constants, memory, and FP registers.
1827 (rs6000_lra_p): New.
1828 * config/rs6000/rs6000.h (SECONDARY_MEMORY_NEEDED_MODE): New macro.
1829 * config/rs6000/rs6000.opt (mlra): New option.
1830 * lra-spills.c (lra_final_code_change): Remove useless move insns.
1831
1832 2013-10-25 Yufeng Zhang <yufeng.zhang@arm.com>
1833
1834 * tree-ssa-math-opts.c (convert_plusminus_to_widen): Call
1835 has_single_use () and not do the conversion if has_single_use ()
1836 returns false for the multiplication result.
1837
1838 2013-10-25 David Malcolm <dmalcolm@redhat.com>
1839
1840 * tree.h (EXCEPTIONAL_CLASS_P): Rename parameter from "CODE"
1841 to "NODE", since this works on a "tree", not an
1842 "enum tree_code".
1843 (CONSTANT_CLASS_P): Likewise.
1844 (TYPE_P): Likewise.
1845 (DECL_P): Likewise.
1846 (INDIRECT_REF_P): Likewise.
1847 (REFERENCE_CLASS_P): Likewise.
1848 (COMPARISON_CLASS_P): Likewise.
1849 (UNARY_CLASS_P): Likewise.
1850 (BINARY_CLASS_P): Likewise.
1851 (STATEMENT_CLASS_P): Likewise.
1852 (VL_EXP_CLASS_P): Likewise.
1853 (EXPRESSION_CLASS_P): Likewise.
1854 (IS_TYPE_OR_DECL_P): Likewise.
1855
1856 2013-10-25 Marc Glisse <marc.glisse@inria.fr>
1857
1858 * tree-ssa-alias.c (ao_ref_init_from_ptr_and_size): Look for an
1859 ADDR_EXPR in the defining statement.
1860
1861 2013-10-25 Richard Biener <rguenther@suse.de>
1862
1863 PR tree-optimization/58626
1864 * tree-loop-distribution.c (enum rdg_dep_type): Remove
1865 anti_dd, output_dd and input_dd.
1866 (struct rdg_edge): Remove level and relation members.
1867 (RDGE_LEVEL, RDGE_RELATION): Remove.
1868 (dot_rdg_1): Adjust.
1869 (create_rdg_edge_for_ddr): Remove.
1870 (create_rdg_edges_for_scalar): Adjust.
1871 (create_edge_for_control_dependence): Likewise.
1872 (create_rdg_edges): Split into ...
1873 (create_rdg_flow_edges): ... this
1874 (create_rdg_cd_edges): ... and this.
1875 (free_rdg): Adjust.
1876 (build_rdg): Likewise, do not compute data dependences or
1877 add edges for them.
1878 (pg_add_dependence_edges): New function.
1879 (pgcmp): Likewise.
1880 (distribute_loop): First apply all non-dependence based
1881 partition mergings. Then compute dependences between partitions
1882 and merge and order partitions according to them.
1883
1884 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1885
1886 PR rtl-optimization/58831
1887 * alias.c (init_alias_analysis): At the beginning of each iteration,
1888 set the reg_seen[N] bit if static_reg_base_value[N] is non-null.
1889
1890 2013-10-25 Eric Botcazou <ebotcazou@adacore.com>
1891
1892 * recog.c (search_ofs): New static variable moved from...
1893 (peep2_find_free_register): ...here.
1894 (peephole2_optimize): Initialize it.
1895
1896 2013-10-25 Tobias Burnus <burnus@net-b.de>
1897
1898 * doc/invoke.texi (fopenmp): Change supported OpenMP version to 4.0.
1899
1900 2013-10-25 Uros Bizjak <ubizjak@gmail.com>
1901
1902 * config/i386/i386.h (TARGET_MPX): New define.
1903 (TARGET_MPX_P): Ditto.
1904
1905 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
1906
1907 * config/i386/constraints.md (B): New.
1908 (Ti): New.
1909 (Tb): New.
1910 * config/i386/i386-c.c (ix86_target_macros_internal): Add __MPX__.
1911 * config/i386/i386-modes.def (BND32): New.
1912 (BND64): New.
1913 * config/i386/i386-protos.h (ix86_bnd_prefixed_insn_p): New.
1914 * config/i386/i386.c (isa_opts): Add mmpx.
1915 (regclass_map): Add bound registers.
1916 (dbx_register_map): Likewise.
1917 (dbx64_register_map): Likewise.
1918 (svr4_dbx_register_map): Likewise.
1919 (PTA_MPX): New.
1920 (ix86_option_override_internal): Support MPX ISA.
1921 (ix86_conditional_register_usage): Support bound registers.
1922 (print_reg): Likewise.
1923 (ix86_code_end): Add MPX bnd prefix.
1924 (output_set_got): Likewise.
1925 (ix86_output_call_insn): Likewise.
1926 (ix86_print_operand): Add '!' (MPX bnd) print prefix support.
1927 (ix86_print_operand_punct_valid_p): Likewise.
1928 (ix86_print_operand_address): Support UNSPEC_BNDMK_ADDR and
1929 UNSPEC_BNDMK_ADDR.
1930 (ix86_class_likely_spilled_p): Add bound regs support.
1931 (ix86_hard_regno_mode_ok): Likewise.
1932 (x86_order_regs_for_local_alloc): Likewise.
1933 (ix86_bnd_prefixed_insn_p): New.
1934 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Fix to new value.
1935 (FIXED_REGISTERS): Add bound registers.
1936 (CALL_USED_REGISTERS): Likewise.
1937 (REG_ALLOC_ORDER): Likewise.
1938 (HARD_REGNO_NREGS): Likewise.
1939 (TARGET_MPX): New.
1940 (VALID_BND_REG_MODE): New.
1941 (FIRST_BND_REG): New.
1942 (LAST_BND_REG): New.
1943 (reg_class): Add BND_REGS.
1944 (REG_CLASS_NAMES): Likewise.
1945 (REG_CLASS_CONTENTS): Likewise.
1946 (BND_REGNO_P): New.
1947 (ANY_BND_REG_P): New.
1948 (BNDmode): New.
1949 (HI_REGISTER_NAMES): Add bound registers.
1950 * config/i386/i386.md (UNSPEC_BNDMK): New.
1951 (UNSPEC_BNDMK_ADDR): New.
1952 (UNSPEC_BNDSTX): New.
1953 (UNSPEC_BNDLDX): New.
1954 (UNSPEC_BNDLDX_ADDR): New.
1955 (UNSPEC_BNDCL): New.
1956 (UNSPEC_BNDCU): New.
1957 (UNSPEC_BNDCN): New.
1958 (UNSPEC_MPX_FENCE): New.
1959 (BND0_REG): New.
1960 (BND1_REG): New.
1961 (type): Add mpxmov, mpxmk, mpxchk, mpxld, mpxst.
1962 (length_immediate): Likewise.
1963 (prefix_0f): Likewise.
1964 (memory): Likewise.
1965 (prefix_rep): Check for bnd prefix.
1966 (length_nobnd): New.
1967 (length): Use length_nobnd if specified.
1968 (BND): New.
1969 (bnd_ptr): New.
1970 (BNDCHECK): New.
1971 (bndcheck): New.
1972 (*jcc_1): Add bnd prefix and rename length attr to length_nobnd.
1973 (*jcc_2): Likewise.
1974 (jump): Likewise.
1975 (simple_return_internal): Likewise.
1976 (simple_return_pop_internal): Likewise.
1977 (*indirect_jump): Add MPX bnd prefix.
1978 (*tablejump_1): Likewise.
1979 (simple_return_internal_long): Likewise.
1980 (simple_return_indirect_internal): Likewise.
1981 (<mode>_mk): New.
1982 (*<mode>_mk): New.
1983 (mov<mode>): New.
1984 (*mov<mode>_internal_mpx): New.
1985 (<mode>_<bndcheck>): New.
1986 (*<mode>_<bndcheck>): New.
1987 (<mode>_ldx): New.
1988 (*<mode>_ldx): New.
1989 (<mode>_stx): New.
1990 (*<mode>_stx): New.
1991 * config/i386/predicates.md (lea_address_operand): Rename to...
1992 (address_no_seg_operand): ... this.
1993 (address_mpx_no_base_operand): New.
1994 (address_mpx_no_index_operand): New.
1995 (bnd_mem_operator): New.
1996 * config/i386/i386.opt (mmpx): New.
1997 * doc/invoke.texi: Add documentation for the flags -mmpx, -mno-mpx.
1998 * doc/rtl.texi Add documentation for BND32mode and BND64mode.
1999
2000 2013-10-24 Ilya Enkovich <ilya.enkovich@intel.com>
2001
2002 * mode-classes.def (MODE_POINTER_BOUNDS): New.
2003 * tree.def (POINTER_BOUNDS_TYPE): New.
2004 * genmodes.c (complete_mode): Support MODE_POINTER_BOUNDS.
2005 (POINTER_BOUNDS_MODE): New.
2006 (make_pointer_bounds_mode): New.
2007 * machmode.h (POINTER_BOUNDS_MODE_P): New.
2008 * stor-layout.c (int_mode_for_mode): Support MODE_POINTER_BOUNDS.
2009 (layout_type): Support POINTER_BOUNDS_TYPE.
2010 * tree-pretty-print.c (dump_generic_node): Support POINTER_BOUNDS_TYPE.
2011 * tree.c (build_int_cst_wide): Support POINTER_BOUNDS_TYPE.
2012 (type_contains_placeholder_1): Likewise.
2013 * tree.h (POINTER_BOUNDS_TYPE_P): New.
2014 * varasm.c (output_constant): Support POINTER_BOUNDS_TYPE.
2015 * doc/rtl.texi (MODE_POINTER_BOUNDS): New.
2016
2017 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2018
2019 * expr.c (expand_expr_real_1): Use mode of memory reference rather than
2020 mode of address computation when calling memory_address_addr_space.
2021
2022 2013-08-24 Richard Henderson <rth@twiddle.net>
2023
2024 PR rtl/58542
2025 * optabs.c (maybe_emit_atomic_exchange): Use create_input_operand
2026 instead of create_convert_operand_to.
2027 (maybe_emit_sync_lock_test_and_set): Likewise.
2028 (expand_atomic_compare_and_swap): Likewise.
2029 (maybe_emit_compare_and_swap_exchange_loop): Don't convert_modes.
2030
2031 2013-08-24 Sriraman Tallam <tmsriram@google.com>
2032
2033 * cgraph.c (cgraph_fnver_htab): Move GTY((...)) to be before htab_t.
2034 Change param_is to use the struct and not the pointer to the struct.
2035
2036 2013-10-24 Andrew MacLeod <amacleod@redhat.com>
2037
2038 * builtins.c (dummy_object, gimplify_va_arg_expr): Move to gimplify.c.
2039 * gimplify.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2040 Move to targhooks.c.
2041 (dummy_object, gimplify_va_arg_expr): Relocate from builtins.c.
2042 * targhooks.c (build_va_arg_indirect_ref, std_gimplify_va_arg_expr):
2043 Relocate from gimplify.c.
2044 * targhooks.h: Add 2 prototypes.
2045 * tree.h: Delete 2 prototypes.
2046
2047 2013-10-24 Igor Shevlyakov <igor.shevlyakov@gmail.com>
2048
2049 * tree-ssa-loop-ivopts.c (multiplier_allowed_in_address_p ): Check both
2050 [reg+mult*reg] and [mult*reg] to determine if multiplier is allowed.
2051
2052 2013-10-24 Cong Hou <congh@google.com>
2053
2054 * convert.c (convert_to_real): Guard those unsafe math function
2055 convertions with flag_unsafe_math_optimizations. Handle sqrt()
2056 specially.
2057
2058 2013-10-24 Markus Trippelsdorf <markus@trippelsdorf.de>
2059
2060 PR ipa/58712
2061 * cgraph.c (cgraph_create_edge_1): Add indirect_unknown_callee
2062 as argument.
2063 (cgraph_create_edge): Use the new argument.
2064 (cgraph_create_indirect_edge): Likewise.
2065
2066 2013-10-24 Joern Rennecke <joern.rennecke@embecosm.com>
2067
2068 * gcc/config/arc/arc.c (arc_ccfsm_post_advance): Also handle
2069 TYPE_UNCOND_BRANCH.
2070 (arc_ifcvt) <case 1 and 2>: Check that arc_ccfsm_post_advance
2071 changes statep->state.
2072
2073 2013-10-24 Tobias Burnus <burnus@net-b.de>
2074
2075 PR other/33426
2076 * tree-cfg.c (replace_loop_annotate): New function.
2077 (execute_build_cfg): Call it.
2078 * gimplify.c (gimple_boolify, gimplify_expr): Handle ANNOTATE_EXPR.
2079 * internal-fn.c (expand_ANNOTATE): New function.
2080 * internal-fn.def (ANNOTATE): Define as new internal function.
2081 * tree-core.h (tree_node_kind): Add annot_expr_ivdep_kind.
2082 * tree-pretty-print.c (dump_generic_node): Handle ANNOTATE_EXPR.
2083 * tree.def (ANNOTATE_EXPR): New DEFTREECODE.
2084 * doc/extend.texi (Pragmas): Document #pragma ivdep.
2085 * doc/generic.texi (Expressions): Document ANNOTATE_EXPR.
2086
2087 2013-10-17 Ian Bolton <ian.bolton@arm.com>
2088 Marcus Shawcroft <marcus.shawcroft@arm.com>
2089
2090 * config/aarch64/aarch64.c (aarch64_preferred_reload_class):
2091 Special case reload SP+C into none GENERAL_REGS.
2092
2093 2013-10-24 Michael Matz <matz@suse.de>
2094
2095 * gengtype.c (is_file_equal): Check that files will be same length.
2096
2097 2013-10-25 Christian Bruel <christian.bruel@st.com>
2098
2099 * config.gcc (sh-*): Add sh-mem.o to extra_obj.
2100 * gcc/config/sh/t-sh (sh-mem.o): New rule.
2101 * gcc/config/sh/sh-mem.cc (expand_block_move): Moved here.
2102 (sh_expand_cmpstr): New function.
2103 * gcc/config/sh/sh.c (force_into, expand_block_move): Move to sh-mem.c.
2104 * gcc/config/sh/sh-protos.h (sh_expand_cmpstr): Declare.
2105 * gcc/config/sh/sh.md (cmpstrsi, cmpstr_t): New patterns.
2106 (rotlhi3_8): Rename.
2107
2108 2013-10-24 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2109
2110 * configure.ac (ZW_PROG_COMPILER_DEPENDENCIES): Use CXX instead of CC.
2111 * Makefile.in (CXXDEPMODE): Assign and change users.
2112 (CCDEPMODE): Delete.
2113 * configure: Regenerate.
2114
2115 2013-10-23 David Malcolm <dmalcolm@redhat.com>
2116
2117 * gengtype-parse.c (require_without_advance): New.
2118 (type): For GTY-marked types that are not GTY((user)), parse any
2119 base classes, requiring them to be single-inheritance, and not
2120 be templates. For non-GTY-marked types and GTY((user)),
2121 continue to skip over any C++ inheritance specification.
2122 * gengtype-state.c (state_writer::write_state_struct_type):
2123 Write base class of type (if any).
2124 (read_state_struct_type): Read base class of type (if any).
2125 * gengtype.c (new_structure): Add a "base_class" parameter.
2126 (create_optional_field_): Update for new parameter to new_structure.
2127 (adjust_field_rtx_def): Likewise.
2128 (adjust_field_tree_exp): Likewise.
2129 * gengtype.h (struct type): Add "base_class" field to the s
2130 union field.
2131 (new_structure): Add "base" parameter.
2132
2133 2013-10-23 Sriraman Tallam <tmsriram@google.com>
2134
2135 PR target/57756
2136 * config/i386/i386.c (ix86_option_override_internal):
2137 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...)
2138 (ix86_valid_target_attribute_tree):
2139 Change TARGET_64BIT to TARGET_64BIT_P (opts->...)
2140 Change TARGET_SSE to TARGET_SSE_P (opts->...)
2141
2142 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2143
2144 * tree-ssa-loop.h: Remove include files.
2145 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2146 * cfgloopmanip.c: Move required includes from tree-ssa-loop.h.
2147 * graphite-clast-to-gimple.c: Likewise.
2148 * graphite-scop-detection.c: Likewise.
2149 * graphite-sese-to-poly.c: Likewise.
2150 * ipa-inline-analysis.c: Likewise.
2151 * ipa-pure-const.c: Likewise.
2152 * loop-init.c: Likewise.
2153 * passes.c: Likewise.
2154 * predict.c: Likewise.
2155 * tree-cfg.c: Likewise.
2156 * tree-cfgcleanup.c: Likewise.
2157 * tree-chrec.c: Likewise.
2158 * tree-data-ref.c: Likewise.
2159 * tree-loop-distribution.c: Likewise.
2160 * tree-parloops.c: Likewise.
2161 * tree-predcom.c: Likewise.
2162 * tree-scalar-evolution.c: Likewise.
2163 * tree-ssa-address.c: Likewise.
2164 * tree-ssa.c: Likewise.
2165 * tree-ssa-dce.c: Likewise.
2166 * tree-ssa-loop.c: Likewise.
2167 * tree-ssa-loop-im.c: Likewise.
2168 * tree-ssa-loop-ivcanon.c: Likewise.
2169 * tree-ssa-loop-ivopts.c: Likewise.
2170 * tree-ssa-loop-manip.c: Likewise.
2171 * tree-ssa-loop-niter.c: Likewise.
2172 * tree-ssa-loop-prefetch.c: Likewise.
2173 * tree-ssa-loop-unswitch.c: Likewise.
2174 * tree-ssa-reassoc.c: Likewise.
2175 * tree-vect-data-refs.c: Likewise.
2176 * tree-vect-loop.c: Likewise.
2177 * tree-vect-loop-manip.c: Likewise.
2178 * tree-vectorizer.c: Likewise.
2179 * tree-vect-stmts.c: Likewise.
2180 * tree-vrp.c: Likewise.
2181
2182 2013-10-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2183
2184 * config/rs6000/altivec.md (mulv8hi3): Adjust for little endian.
2185
2186 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2187
2188 PR tree-optimization/58775
2189 PR tree-optimization/58791
2190 * tree-ssa-reassoc.c (reassoc_stmt_dominates_stmt_p): New function.
2191 (insert_stmt_after): Rewritten, don't move the stmt, but really
2192 insert it.
2193 (get_stmt_uid_with_default): Remove.
2194 (build_and_add_sum): Use insert_stmt_after and
2195 reassoc_stmt_dominates_stmt_p. Fix up uid if bb contains only labels.
2196 (update_range_test): Set uid on stmts added by
2197 force_gimple_operand_gsi. Don't immediately modify statements
2198 in inter-bb optimization, just update oe->op values.
2199 (optimize_range_tests): Return bool whether any changed have been made.
2200 (update_ops): New function.
2201 (struct inter_bb_range_test_entry): New type.
2202 (maybe_optimize_range_tests): Perform statement changes here.
2203 (not_dominated_by, appears_later_in_bb, get_def_stmt,
2204 ensure_ops_are_available): Remove.
2205 (find_insert_point): Rewritten.
2206 (rewrite_expr_tree): Remove MOVED argument, add CHANGED argument,
2207 return LHS of the (new resp. old) stmt. Don't call
2208 ensure_ops_are_available, don't reuse SSA_NAMEs, recurse first
2209 instead of last, move new stmt at the right place.
2210 (linearize_expr, repropagate_negates): Don't reuse SSA_NAMEs.
2211 (negate_value): Likewise. Set uids.
2212 (break_up_subtract_bb): Initialize uids.
2213 (reassociate_bb): Adjust rewrite_expr_tree caller.
2214 (do_reassoc): Don't call renumber_gimple_stmt_uids.
2215
2216 2013-10-23 David Edelsohn <dje.gcc@gmail.com>
2217
2218 PR target/58838
2219 * config/rs6000/rs6000.md (mulsi3_internal1 and splitter): Add
2220 TARGET_32BIT final condition.
2221 (mulsi3_internal2 and splitter): Same.
2222
2223 2013-10-23 Jeff Law <law@redhat.com>
2224
2225 * tree-ssa-threadedge.c (thread_across_edge): Do not allow threading
2226 through joiner blocks with abnormal outgoing edges.
2227
2228 * tree-ssa-threadupdate.c (thread_block_1): Renamed from thread_block.
2229 Add parameter JOINERS, to allow/disallow threading through joiner
2230 blocks.
2231 (thread_block): New. Call thread_block_1.
2232 (mark_threaded_blocks): Remove code to filter out certain cases
2233 of threading through joiner blocks.
2234 (thread_through_all_blocks): Document how we can have a dangling
2235 edge AUX field and clear it.
2236
2237 2013-10-23 Ian Lance Taylor <iant@google.com>
2238
2239 * doc/invoke.texi (Option Summary): Remove -fno-default-inline.
2240 (C++ Dialect Options): Likewise.
2241 (Optimize Options): Likewise.
2242
2243 2013-10-23 Tom de Vries <tom@codesourcery.com>
2244
2245 PR tree-optimization/58805
2246 * tree-ssa-tail-merge.c (stmt_local_def): Add gimple_vdef check.
2247
2248 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2249
2250 * tree-vect-patterns.c (vect_recog_divmod_pattern): Optimize
2251 sequence based on get_range_info returned range.
2252
2253 2013-10-23 Andrew MacLeod <amacleod@redhat.com>
2254
2255 * tree-ssa.h: Remove all #include's
2256 * gengtype.c (open_base_files): Adjust include list for gtype-desc.c.
2257 * alias.c: Move required includes from tree-ssa.h.
2258 * asan.c: Likewise.
2259 * builtins.c: Likewise.
2260 * calls.c: Likewise.
2261 * cfgexpand.c: Likewise.
2262 * cfghooks.c: Likewise.
2263 * cfgloop.c: Likewise.
2264 * cfgloopmanip.c: Likewise.
2265 * cgraph.c: Likewise.
2266 * cgraphbuild.c: Likewise.
2267 * cgraphclones.c: Likewise.
2268 * cgraphunit.c: Likewise.
2269 * dse.c: Likewise.
2270 * except.c: Likewise.
2271 * expr.c: Likewise.
2272 * final.c: Likewise.
2273 * fold-const.c: Likewise.
2274 * ggc-page.c: Likewise.
2275 * gimple-builder.c: Likewise.
2276 * gimple-fold.c: Likewise.
2277 * gimple-iterator.c: Likewise.
2278 * gimple-low.c: Likewise.
2279 * gimple-pretty-print.c: Likewise.
2280 * gimple-ssa-strength-reduction.c: Likewise.
2281 * gimple-streamer-in.c: Likewise.
2282 * gimple-streamer-out.c: Likewise.
2283 * gimplify.c: Likewise.
2284 * graphite-blocking.c: Likewise.
2285 * graphite-clast-to-gimple.c: Likewise.
2286 * graphite-dependences.c: Likewise.
2287 * graphite-interchange.c: Likewise.
2288 * graphite-optimize-isl.c: Likewise.
2289 * graphite-poly.c: Likewise.
2290 * graphite-scop-detection.c: Likewise.
2291 * graphite-sese-to-poly.c: Likewise.
2292 * graphite.c: Likewise.
2293 * ipa-cp.c: Likewise.
2294 * ipa-inline-analysis.c: Likewise.
2295 * ipa-inline-transform.c: Likewise.
2296 * ipa-inline.c: Likewise.
2297 * ipa-prop.c: Likewise.
2298 * ipa-pure-const.c: Likewise.
2299 * ipa-reference.c: Likewise.
2300 * ipa-split.c: Likewise.
2301 * ipa-utils.c: Likewise.
2302 * loop-init.c: Likewise.
2303 * lto-cgraph.c: Likewise.
2304 * lto-section-in.c: Likewise.
2305 * lto-section-out.c: Likewise.
2306 * lto-streamer-in.c: Likewise.
2307 * lto-streamer-out.c: Likewise.
2308 * lto-streamer.c: Likewise.
2309 * omp-low.c: Likewise.
2310 * passes.c: Likewise.
2311 * predict.c: Likewise.
2312 * print-tree.c: Likewise.
2313 * profile.c: Likewise.
2314 * sese.c: Likewise.
2315 * targhooks.c: Likewise.
2316 * tracer.c: Likewise.
2317 * trans-mem.c: Likewise.
2318 * tree-call-cdce.c: Likewise.
2319 * tree-cfg.c: Likewise.
2320 * tree-cfgcleanup.c: Likewise.
2321 * tree-chrec.c: Likewise.
2322 * tree-complex.c: Likewise.
2323 * tree-data-ref.c: Likewise.
2324 * tree-dfa.c: Likewise.
2325 * tree-eh.c: Likewise.
2326 * tree-emutls.c: Likewise.
2327 * tree-if-conv.c: Likewise.
2328 * tree-inline.c: Likewise.
2329 * tree-into-ssa.c: Likewise.
2330 * tree-loop-distribution.c: Likewise.
2331 * tree-mudflap.c: Likewise.
2332 * tree-nested.c: Likewise.
2333 * tree-nrv.c: Likewise.
2334 * tree-object-size.c: Likewise.
2335 * tree-outof-ssa.c: Likewise.
2336 * tree-parloops.c: Likewise.
2337 * tree-phinodes.c: Likewise.
2338 * tree-predcom.c: Likewise.
2339 * tree-pretty-print.c: Likewise.
2340 * tree-profile.c: Likewise.
2341 * tree-scalar-evolution.c: Likewise.
2342 * tree-sra.c: Likewise.
2343 * tree-ssa-address.c: Likewise.
2344 * tree-ssa-alias.c: Likewise.
2345 * tree-ssa-ccp.c: Likewise.
2346 * tree-ssa-coalesce.c: Likewise.
2347 * tree-ssa-copy.c: Likewise.
2348 * tree-ssa-copyrename.c: Likewise.
2349 * tree-ssa-dce.c: Likewise.
2350 * tree-ssa-dom.c: Likewise.
2351 * tree-ssa-dse.c: Likewise.
2352 * tree-ssa-forwprop.c: Likewise.
2353 * tree-ssa-ifcombine.c: Likewise.
2354 * tree-ssa-live.c: Likewise.
2355 * tree-ssa-loop-ch.c: Likewise.
2356 * tree-ssa-loop-im.c: Likewise.
2357 * tree-ssa-loop-ivcanon.c: Likewise.
2358 * tree-ssa-loop-ivopts.c: Likewise.
2359 * tree-ssa-loop-manip.c: Likewise.
2360 * tree-ssa-loop-niter.c: Likewise.
2361 * tree-ssa-loop-prefetch.c: Likewise.
2362 * tree-ssa-loop-unswitch.c: Likewise.
2363 * tree-ssa-loop.c: Likewise.
2364 * tree-ssa-math-opts.c: Likewise.
2365 * tree-ssa-operands.c: Likewise.
2366 * tree-ssa-phiopt.c: Likewise.
2367 * tree-ssa-phiprop.c: Likewise.
2368 * tree-ssa-pre.c: Likewise.
2369 * tree-ssa-propagate.c: Likewise.
2370 * tree-ssa-reassoc.c: Likewise.
2371 * tree-ssa-sccvn.c: Likewise.
2372 * tree-ssa-sink.c: Likewise.
2373 * tree-ssa-strlen.c: Likewise.
2374 * tree-ssa-structalias.c: Likewise.
2375 * tree-ssa-tail-merge.c: Likewise.
2376 * tree-ssa-ter.c: Likewise.
2377 * tree-ssa-threadedge.c: Likewise.
2378 * tree-ssa-threadupdate.c: Likewise.
2379 * tree-ssa-uncprop.c: Likewise.
2380 * tree-ssa-uninit.c: Likewise.
2381 * tree-ssa.c: Likewise.
2382 * tree-ssanames.c: Likewise.
2383 * tree-stdarg.c: Likewise.
2384 * tree-streamer-in.c: Likewise.
2385 * tree-switch-conversion.c: Likewise.
2386 * tree-tailcall.c: Likewise.
2387 * tree-vect-data-refs.c: Likewise.
2388 * tree-vect-generic.c: Likewise.
2389 * tree-vect-loop-manip.c: Likewise.
2390 * tree-vect-loop.c: Likewise.
2391 * tree-vect-patterns.c: Likewise.
2392 * tree-vect-slp.c: Likewise.
2393 * tree-vect-stmts.c: Likewise.
2394 * tree-vectorizer.c: Likewise.
2395 * tree-vrp.c: Likewise.
2396 * tree.c: Likewise.
2397 * tsan.c: Likewise.
2398 * value-prof.c: Likewise.
2399 * var-tracking.c: Likewise.
2400 * varpool.c: Likewise.
2401 * vtable-verify.c: Likewise.
2402
2403 2013-10-23 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2404
2405 * config/tilegx/tilegx.c: Include "tree.h".
2406
2407 2013-10-23 Jakub Jelinek <jakub@redhat.com>
2408
2409 * gimple-pretty-print.c (dump_ssaname_info): Always print "# " before
2410 the info, not after it.
2411 (gump_gimple_phi): Add COMMENT argument, if true, print "# " after
2412 dump_ssaname_info call.
2413 (pp_gimple_stmt_1): Adjust caller.
2414 (dump_phi_nodes): Likewise. Don't print "# " here.
2415
2416 2013-10-22 Jan Hubicka <jh@suse.cz>
2417
2418 * i386.h (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES_EPILOGUES): New
2419 tuning flag.
2420 * x86-tune.def (TARGET_MISALIGNED_MOVE_STRING_PROLOGUES): Define it.
2421 * i386.c (expand_small_movmem_or_setmem): New function.
2422 (expand_set_or_movmem_prologue_epilogue_by_misaligned_moves): New
2423 function.
2424 (alg_usable_p): Add support for value ranges; cleanup.
2425 (ix86_expand_set_or_movmem): Add support for misaligned moves.
2426
2427 2013-10-22 Sterling Augustine <saugustine@google.com>
2428
2429 * doc/invoke.texi: Document -ggnu-pubnames.
2430 * common.opt: Add new option -ggnu-pubnames and modify -gpubnames
2431 logic.
2432 * dwarf2out.c: Include gdb/gdb-index.h.
2433 (DEBUG_PUBNAMES_SECTION, DEBUG_PUBTYPES_SECTION): Handle
2434 debug_generate_pub_sections.
2435 (is_java, output_pubtables, output_pubname): New functions.
2436 (include_pubname_in_output): Handle debug_generate_pub_sections at
2437 level 2.
2438 (size_of_pubnames): Use new local space_for_flags based on
2439 debug_generate_pub_sections.
2440 (output_pubnames): Unify pubnames and pubtypes output logic.
2441 Genericize comments. Call output_pubname.
2442 (dwarf2out_finish): Move logic to output_pubnames and call it.
2443
2444 2013-10-22 Uros Bizjak <ubizjak@gmail.com>
2445
2446 PR target/58779
2447 * config/i386/i386.c (put_condition_code) <case GTU, case LEU>:
2448 Remove CCCmode handling.
2449 <case LTU>: Return 'c' suffix for CCCmode.
2450 <case GEU>: Return 'nc' suffix for CCCmode.
2451 (ix86_cc_mode) <case GTU, case LEU>: Do not generate overflow checks.
2452 * config/i386/i386.md (*sub<mode>3_cconly_overflow): Remove.
2453 (*sub<mode>3_cc_overflow): Ditto.
2454 (*subsi3_zext_cc_overflow): Ditto.
2455
2456 2013-10-22 Steve Ellcey <sellcey@mips.com>
2457
2458 * config/mips/mips.c (mips_rtx_costs): Fix cost estimate for nor
2459 (AND (NOT OP1) (NOT OP2)).
2460
2461 2013-10-22 Bill Schmidt <wschmidt@vnet.ibm.com>
2462
2463 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
2464 meaning of merge-high and merge-low masks for little endian; avoid
2465 use of vector-pack masks for little endian for mismatched modes.
2466
2467 2013-10-22 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2468
2469 * config/tilepro/tilepro.c: Include "tree.h".
2470
2471 2013-10-22 Andreas Schwab <schwab@suse.de>
2472
2473 * config/m68k/m68k.c (notice_update_cc): Handle register conflict
2474 with PRE_DEC.
2475
2476 2013-10-22 Paolo Carlini <paolo.carlini@oracle.com>
2477
2478 * doc/contrib.texi ([Fran@,{c}ois Dumont], [Tim Shen],
2479 [Ed Smith-Rowland]): New entries.
2480 ([Stephen M. Webb]): Update.
2481
2482 2013-10-22 Andrew MacLeod <amacleod@redhat.com>
2483
2484 * tree-ssa-ter.h: Remove duplicate copy of file contents.
2485
2486 2013-10-21 Marek Polacek <polacek@redhat.com>
2487
2488 PR middle-end/58809
2489 * fold-const.c (fold_range_test): Return 0 if the type is not
2490 an integral type.
2491
2492 2013-10-21 Richard Sandiford <rdsandiford@googlemail.com>
2493
2494 * system.h: Move hwint.h include further down.
2495 * hwint.h (sext_hwi, zext_hwi): Define unconditionally. Add
2496 gcc_checking_asserts.
2497 * hwint.c (sext_hwi, zext_hwi): Delete ENABLE_CHECKING versions.
2498
2499 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2500
2501 Fix volatile issues in optimize_bit_field_compare.
2502 * fold-const.c (optimize_bit_field_compare): Bail out if
2503 lvolatilep or rvolatilep.
2504
2505 2013-10-21 Bernd Edlinger <bernd.edlinger@hotmail.de>
2506
2507 Fix DECL_BIT_FIELD depencency on flag_strict_volatile_bitfields
2508 and get_inner_reference returning different pmode for non-volatile
2509 bit-field members dependent on flag_strict_volatile_bitfields.
2510 * stor-layout.c (layout_decl): Remove special handling of
2511 flag_strict_volatile_bitfields.
2512 * expr.c (get_inner_reference): Don't use DECL_BIT_FIELD
2513 if flag_strict_volatile_bitfields > 0 and TREE_THIS_VOLATILE.
2514
2515 2013-10-21 Paulo Matos <pmatos@broadcom.com>
2516
2517 * ipa-inline.c (edge_badness): Cap edge->count at max_count for badness
2518 calculations.
2519
2520 2013-10-21 Jeff Law <law@redhat.com>
2521
2522 * tree-ssa-threadedge.c (thread_through_normal_block): New
2523 argument VISITED. Remove VISISTED as a local variable. When we
2524 have a threadable jump, verify the destination of the jump has not
2525 been visised.
2526 (thread_across_edge): Allocate VISITED bitmap once at function
2527 scope and use it throughout. Make sure to set appropriate bits in
2528 VISITED for E (start of jump thread path).
2529 * tree-ssa-threadupdate.c (mark_threaded_blocks): Reject threading
2530 through a joiner if any edge on the path has a recorded jump thread.
2531
2532 2013-10-21 Ian Lance Taylor <iant@google.com>
2533
2534 * doc/invoke.texi (Optimize Options): For -fno-toplevel-reorder,
2535 don't imply that attributes can solve all problems.
2536 (Directory Options): Fix typo.
2537
2538 2013-10-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2539
2540 * config/arm/arm.c (cortexa9_extra_costs): Update mult costs for
2541 extend and extend_add.
2542
2543 2013-10-21 Richard Biener <rguenther@suse.de>
2544
2545 PR tree-optimization/58794
2546 * fold-const.c (operand_equal_p): Compare FIELD_DECL operand
2547 of COMPONENT_REFs with OEP_CONSTANT_ADDRESS_OF left in place.
2548
2549 2013-10-21 Richard Biener <rguenther@suse.de>
2550
2551 PR middle-end/58742
2552 * fold-const.c (fold_binary_loc): Fold ((T) (X /[ex] C)) * C
2553 to (T) X for sign-changing conversions (or no conversion).
2554
2555 2013-10-20 Uros Bizjak <ubizjak@gmail.com>
2556
2557 * config/i386/i386.md (kxnor<mode>): Add FLAGS_REG clobber.
2558
2559 2013-10-20 Jan Hubicka <jh@suse.cz>
2560
2561 * config/i386/i386-tune.def: Add comment; organize into categories
2562
2563 2013-10-21 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
2564
2565 * config/i386/i386.c (expand_set_or_movmem_via_loop): Add issetmem
2566 argument. Update function comment.
2567 (expand_set_or_movmem_via_rep): New function combining
2568 expand_movmem_via_rep_mov and expand_setmem_via_rep_stos.
2569 (expand_movmem_via_rep_mov): Remove.
2570 expand_setmem_via_rep_stos): Remove.
2571 (expand_movmem_epilogue): Update calls correspondingly.
2572 (expand_setmem_epilogue_via_loop): Likewise.
2573 (emit_memset): New.
2574 (expand_setmem_epilogue): Add VEC_VALUE argument, refactor.
2575 (expand_set_or_movmem_prologue): New function combining
2576 expand_movmem_prologue and expand_setmem_prologue.
2577 (expand_movmem_prologue): Remove.
2578 (expand_setmem_prologue): Remove.
2579 (expand_set_or_movmem_constant_prologue): New function combining
2580 expand_constant_movmem_prologue and expand_constant_setmem_prologue.
2581 (expand_constant_movmem_prologue): Remove.
2582 (expand_constant_setmem_prologue): Remove.
2583 (promote_duplicated_reg): Allow vector-const0 value.
2584 (ix86_expand_set_or_movmem): New function combining ix86_expand_movmem
2585 and ix86_expand_setmem.
2586 (ix86_expand_movmem): Call ix86_expand_set_or_movmem.
2587 (ix86_expand_setmem): Call ix86_expand_set_or_movmem.
2588
2589 2013-10-21 Diego Novillo <dnovillo@google.com>
2590
2591 * asan.c: Include tree.h
2592 * bb-reorder.c: Likewise.
2593 * cfgcleanup.c: Likewise.
2594 * cfgloopmanip.c: Likewise.
2595 * data-streamer-in.c: Likewise.
2596 * data-streamer-out.c: Likewise.
2597 * data-streamer.c: Likewise.
2598 * dwarf2cfi.c: Likewise.
2599 * graphite-blocking.c: Likewise.
2600 * graphite-clast-to-gimple.c: Likewise.
2601 * graphite-dependences.c: Likewise.
2602 * graphite-interchange.c: Likewise.
2603 * graphite-optimize-isl.c: Likewise.
2604 * graphite-poly.c: Likewise.
2605 * graphite-scop-detection.c: Likewise.
2606 * graphite-sese-to-poly.c: Likewise.
2607 * graphite.c: Likewise.
2608 * ipa-devirt.c: Likewise.
2609 * ipa-profile.c: Likewise.
2610 * ipa.c: Likewise.
2611 * ira.c: Likewise.
2612 * loop-init.c: Likewise.
2613 * loop-unroll.c: Likewise.
2614 * lower-subreg.c: Likewise.
2615 * lto/lto-object.c: Likewise.
2616 * recog.c: Likewise.
2617 * reginfo.c: Likewise.
2618 * tree-loop-distribution.c: Likewise.
2619 * tree-parloops.c: Likewise.
2620 * tree-ssa-strlen.c: Likewise.
2621 * tree-streamer.c: Likewise.
2622 * value-prof.c: Likewise.
2623 * target-globals.c: Likewise.
2624 * expr.h: Include tree-core.h instead of tree.h.
2625 * gimple.h: Likewise.
2626 * ipa-prop.h: Likewise.
2627 * ipa-utils.h: Likewise.
2628 * lto-streamer.h: Likewise.
2629 * streamer-hooks.h: Likewise.
2630 * ipa-reference.h: Include cgraph.h instead of tree.h.
2631 * cgraph.h: Include basic-block.h instead of tree.h.
2632 * tree-streamer.h: Do not include tree.h.
2633 * genattrtab.c (write_header): Generate inclusion of
2634 tree.h.
2635 * genautomata.c (main): Likewise.
2636 * genemit.c: Likewise.
2637 * genopinit.c: Likewise.
2638 * genoutput.c (output_prologue): Likewise.
2639 * genpeep.c: Likewise.
2640
2641 2013-10-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2642
2643 * config/rs6000/altivec.md (vec_unpacku_hi_v16qi): Adjust for
2644 little endian.
2645 (vec_unpacku_hi_v8hi): Likewise.
2646 (vec_unpacku_lo_v16qi): Likewise.
2647 (vec_unpacku_lo_v8hi): Likewise.
2648
2649 2013-10-20 Jan Hubicka <jh@suse.cz>
2650
2651 * config/i386/x86-tune.def (X86_TUNE_SLOW_IMUL_IMM32_MEM,
2652 X86_TUNE_SLOW_IMUL_IMM8): Keep enabled only for K8 and AMDFAM10.
2653 (X86_TUNE_USE_VECTOR_FP_CONVERTS): Disable for generic.
2654
2655 2013-10-20 Richard Sandiford <rdsandiford@googlemail.com>
2656
2657 * config/mips/mips.h (ISA_HAS_WSBH): Define.
2658 * config/mips/mips.md (UNSPEC_WSBH, UNSPEC_DSBH, UNSPEC_DSHD): New
2659 constants.
2660 (bswaphi2, bswapsi2, bswapdi2, wsbh, dsbh, dshd): New patterns.
2661
2662 2013-10-19 John David Anglin <danglin@gcc.gnu.org>
2663
2664 PR target/58603
2665 * system.h: Undef m_slot.
2666
2667 2013-10-19 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2668
2669 * config/rs6000/rs6000.c (vspltis_constant): Make sure we check
2670 all elements for both endian flavors.
2671
2672 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2673
2674 PR target/58792
2675 * config/i386/i386.c (ix86_function_value_regno): Add DX_REG,
2676 ST1_REG and XMM1_REG for 32bit and 64bit targets. Also add DI_REG
2677 and SI_REG for 64bit SYSV ABI targets.
2678
2679 2013-10-19 Uros Bizjak <ubizjak@gmail.com>
2680
2681 * mode-switching.c (create_pre_exit): Rename maybe_builtin_apply
2682 to multi_reg_return. Clarify that we are skipping USEs of multiple
2683 return registers. Use bool type where appropriate.
2684
2685 2013-10-18 Jan Hubicka <jh@suse.cz>
2686
2687 * config/i386/i386.h (ACCUMULATE_OUTGOING_ARGS): Disable accumulation
2688 for cold functions.
2689 * x86-tune.def (X86_TUNE_USE_LEAVE): Update comment.
2690 (X86_TUNE_PUSH_MEMORY): Likewise.
2691 (X86_TUNE_AVX256_UNALIGNED_LOAD_OPTIMAL,
2692 X86_TUNE_AVX256_UNALIGNED_STORE_OPTIMAL): New.
2693 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS, X86_TUNE_ALWAYS_FANCY_MATH_387):
2694 New.
2695 * i386.c (x86_accumulate_outgoing_args, x86_arch_always_fancy_math_387,
2696 x86_avx256_split_unaligned_load, x86_avx256_split_unaligned_store):
2697 Remove.
2698 (ix86_option_override_internal): Update to use tune features instead
2699 of variables.
2700
2701 2013-10-18 Cong Hou <congh@google.com>
2702
2703 PR tree-optimization/58508
2704 * tree-vect-loop-manip.c (vect_loop_versioning): Hoist loop invariant
2705 statement that contains data refs with zero-step.
2706
2707 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2708
2709 * tree-ssa.h: Don't include gimple-low.h, tree-ssa-address.h,
2710 sbitmap.h, tree-ssa-threadedge.h, tree-ssa-dom.h and tree-cfgcleanup.h.
2711 * gimple-low.c (gimple_check_call_arg,
2712 gimple_check_call_matching_types): Move to cgraph.c.
2713 * gimple-low.h: Remove prototype.
2714 * cgraph.c: (gimple_check_call_arg, gimple_check_call_matching_types):
2715 Relocate from gimple-low.c.
2716 * cgraph.h: Add prototype. Don't include basic-block.h.
2717 * gimplify.c: Add gimple-low to include list.
2718 * omp-low.c: Add gimple-low and tree-cfgcleanup.h to include list.
2719 * tree-eh.c: Add gimple-low to include list.
2720 * tree-nested.c: Likewise.
2721 * cfgexpand.c: Add tree-ssa-address.h to include list.
2722 * expr.c: Likewise.
2723 * gimple-fold.c: Likewise.
2724 * gimple-ssa-strength-reduction.c: Likewise.
2725 * trans-mem.c: Likewise.
2726 * tree-mudflap.c: Likewise.
2727 * tree-ssa-loop-ivopts.c: Likewise.
2728 * tree-ssa-dom.c: Include tree-ssa-threadedge.h and tree-ssa-dom.h.
2729 (degenerate_phi_result): Move to tree-phinodes.c.
2730 * tree-ssa-loop-ch.c: Include tree-ssa-threadedge.h.
2731 * tree-ssa-threadedge.c: Likewise.
2732 * tree-vrp.c: Likewise.
2733 * tree-phinodes.c (degenerate_phi_result): Relocate here.
2734 * tree-ssa-dom.h (degenerate_phi_result): Remove Prototype.
2735 * tree-phinodes.h (degenerate_phi_result): Add prototype.
2736 * tree-ssa-copy.c: Include tree-ssa-dom.h.
2737 * tree-ssa-forwprop.c: Likewise.
2738 * tree-cfgcleanup.c (execute_cleanup_cfg_post_optimizing,
2739 pass_data_cleanup_cfg_post_optimizing,
2740 make_pass_cleanup_cfg_post_optimizing): Relocate from tree-optimize.c.
2741 * tree-optimize.c: Delete File.
2742 * graphite.c: Include tree-cfgcleanup.h.
2743 * passes.c: Likewise.
2744 * tree-cfg.c: Likewise.
2745 * tree-profile.c: Likewise.
2746 * tree-ssa-dse.c: Likewise.
2747 * tree-ssa-loop-ivcanon.c: Likewise.
2748 * tree-switch-conversion.c: Don't include tree-ssa-operands.h.
2749 * tree-outof-ssa.c: Include sbitmap.h.
2750 * tree-ssa-live.c: Likewise.
2751 * tree-ssa-propagate.c: Likewise.
2752 * tree-ssa-structalias.c: Likewise.
2753 * tree-stdarg.c: Likewise.
2754 * Makefile.in (OBJS): Delete tree-optimize.o.
2755 * basic-block.h (gcov_type, gcov_type_unsigned): Move to coretypes.h.
2756 * coretypes.h (gcov_type, gcov_type_unsigned): Relocate here.
2757 * varasm.c: Include basic-block.h.
2758 * cfgloop.h: Include function.h instead of basic-block.h
2759 (bb_loop_depth): Move to cfgloop.c.
2760 * cfgloop.c (bb_loop_depth): Relocate from cfgloop.h.
2761
2762 2013-10-18 Teresa Johnson <tejohnson@google.com>
2763
2764 * predict.c (probably_never_executed): Compare frequency-based
2765 count to number of training runs.
2766 * params.def (UNLIKELY_BB_COUNT_FRACTION): New parameter.
2767
2768 2013-10-18 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2769
2770 * config/arm/arm.c (cortexa9_extra_costs): New table.
2771 (arm_cortex_a9_tune): Use cortexa9_extra_costs.
2772
2773 2013-10-18 Jeff Law <law@redhat.com>
2774
2775 * tree-ssa-threadupdate.c: Do not include "tm.h" or "tm_p.h".
2776
2777 * tree-ssa-threadupdate.c: Include "dbgcnt.h".
2778 (register_jump_thread): Add "registered_jump_thread" debug
2779 counter support.
2780 * dbgcnt.def (registered_jump_thread): New debug counter.
2781
2782 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2783
2784 * config/rs6000/rs6000.c: Include cgraph.h.
2785
2786 2013-10-18 Teresa Johnson <tejohnson@google.com>
2787
2788 * tree-ssa-tail-merge.c (replace_block_by): Update edge
2789 weights during merging.
2790
2791 2013-10-18 Andrew MacLeod <amacleod@redhat.com>
2792
2793 * tree-cfg.h: Rename from tree-flow.h. Remove #includes.
2794 * tree-ssa.h: Relocate required #includes from tree-cfg.h.
2795 * tree-ssa-operands.h: Remove prototype.
2796 * tree-ssa-operands.c (virtual_operand_p): Move to gimple.c.
2797 * gimple.c (virtual_operand_p): Relocate from gimple.c.
2798 * gimple.h: Add prototype.
2799 * gimple-ssa.h: Include tree-ssa-operands.h.
2800 * tree-dump.c: Add tree-cfg.h to include list.
2801 * tree-ssa-alias.c: Add ipa-reference.h to include list.
2802 * config/alpha/alpha.c: Include gimple-ssa.h instead of tree-flow.h.
2803 * config/i386/i386.c: Don't include tree-flow.h.
2804 * config/rs6000/rs6000.c: Likewise.
2805
2806 2013-10-18 Jan-Benedict Glaw <jbglaw@lug-owl.de>
2807
2808 * config/frv/frv.c (frv_init_cumulative_args): Fix wrong cast.
2809
2810 2013-10-18 Richard Biener <rguenther@suse.de>
2811
2812 * stor-layout.c (layout_type): Do not change TYPE_PRECISION
2813 or TYPE_UNSIGNED of integral types.
2814 (set_min_and_max_values_for_integral_type): Leave TYPE_MIN/MAX_VALUE
2815 NULL_TREE for zero-precision integral types.
2816
2817 2013-10-18 James Greenhalgh <james.greenhalgh@arm.com>
2818
2819 * config/aarch64/arm_neon.h
2820 (vcvt<ds>_n_<fsu><32,64>_<fsu><32,64>): Correct argument types.
2821
2822 2013-10-17 Sriraman Tallam <tmsriram@google.com>
2823
2824 PR target/57756
2825 * opth-gen.awk: Define target_flags_explicit.
2826
2827 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2828
2829 * config/rs6000/rs6000.c (enum rs6000_reload_reg_type): Add new
2830 fields to the reg_addr array that describes the valid addressing
2831 mode for any register, general purpose registers, floating point
2832 registers, and Altivec registers.
2833 (FIRST_RELOAD_REG_CLASS): Likewise.
2834 (LAST_RELOAD_REG_CLASS): Likewise.
2835 (struct reload_reg_map_type): Likewise.
2836 (reload_reg_map_type): Likewise.
2837 (RELOAD_REG_VALID): Likewise.
2838 (RELOAD_REG_MULTIPLE): Likewise.
2839 (RELOAD_REG_INDEXED): Likewise.
2840 (RELOAD_REG_OFFSET): Likewise.
2841 (RELOAD_REG_PRE_INCDEC): Likewise.
2842 (RELOAD_REG_PRE_MODIFY): Likewise.
2843 (reg_addr): Likewise.
2844 (mode_supports_pre_incdec_p): New helper functions to say whether
2845 a given mode supports PRE_INC, PRE_DEC, and PRE_MODIFY.
2846 (mode_supports_pre_modify_p): Likewise.
2847 (rs6000_debug_vector_unit): Rearrange the -mdebug=reg output to
2848 print the valid address mode bits for each mode.
2849 (rs6000_debug_print_mode): Likewise.
2850 (rs6000_debug_reg_global): Likewise.
2851 (rs6000_setup_reg_addr_masks): New function to set up the address
2852 mask bits for each type.
2853 (rs6000_init_hard_regno_mode_ok): Use memset to clear arrays.
2854 Call rs6000_setup_reg_addr_masks to set up the address mask bits.
2855 (rs6000_legitimate_address_p): Use mode_supports_pre_incdec_p and
2856 mode_supports_pre_modify_p to determine if PRE_INC, PRE_DEC, and
2857 PRE_MODIFY are supported.
2858 (rs6000_output_move_128bit): Change to use {src,dest}_vmx_p for altivec
2859 registers, instead of {src,dest}_av_p.
2860 (rs6000_print_options_internal): Tweak the debug output slightly.
2861
2862 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2863
2864 * config/i386/sse.md (*vec_widen_smult_even_v8si): Remove
2865 isa attribute.
2866
2867 2013-10-17 Andrew MacLeod <amacleod@redhat.com>
2868
2869 * tree-flow.h (struct omp_region): Move to omp-low.c.
2870 Remove omp_ prototypes and variables.
2871 * gimple.h (omp_reduction_init): Move prototype to omp-low.h.
2872 (copy_var_decl): Relocate prototype from tree-flow.h.
2873 * gimple.c (copy_var_decl): Relocate from omp-low.c.
2874 * tree.h: Move prototype to omp-low.h.
2875 * omp-low.h: New File. Relocate prototypes here.
2876 * omp-low.c (struct omp_region): Make local here.
2877 (root_omp_region): Make static.
2878 (copy_var_decl) Move to gimple.c.
2879 (new_omp_region): Make static.
2880 (make_gimple_omp_edges): New. Refactored from tree-cfg.c make_edges.
2881 * tree-cfg.c: Include omp-low.h.
2882 (make_edges): Factor out OMP specific bits to make_gimple_omp_edges.
2883 * gimplify.c: Include omp-low.h.
2884 * tree-parloops.c: Likewise.
2885
2886 2013-10-17 Uros Bizjak <ubizjak@gmail.com>
2887
2888 * config/i386/i386.c (ix86_fixup_binary_operands): When both source
2889 operands are in memory, prefer to force non-matched operand 1 to
2890 the register.
2891
2892 2013-10-17 Michael Meissner <meissner@linux.vnet.ibm.com>
2893
2894 PR target/58673
2895 * config/rs6000/rs6000.c (rs6000_legitimate_address_p): Only
2896 restrict TImode addresses to single indirect registers if both
2897 -mquad-memory and -mvsx-timode are used.
2898 (rs6000_output_move_128bit): Use quad_load_store_p to determine if
2899 we should emit load/store quad. Remove using %y for quad memory
2900 addresses.
2901
2902 * config/rs6000/rs6000.md (mov<mode>_ppc64, TI/PTImode): Add
2903 constraints to allow load/store quad on machines where TImode is
2904 not allowed in VSX registers. Use 'n' instead of 'F' constraint
2905 for TImode to load integer constants.
2906
2907 2013-10-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2908
2909 * config/aarch64/aarch64.c (aarch64_print_operand): Handle 'c'.
2910
2911 2013-10-17 Marcus Shawcroft <marcus.shawcroft@arm.com>
2912
2913 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Adjust
2914 handling of STACK_REG.
2915
2916 2013-10-17 Richard Biener <rguenther@suse.de>
2917
2918 PR tree-optimization/58143
2919 * tree-ssa-loop-im.c (arith_code_with_undefined_signed_overflow):
2920 New function.
2921 (rewrite_to_defined_overflow): Likewise.
2922 (move_computations_dom_walker::before_dom): Rewrite stmts
2923 with undefined signed overflow that are not always executed
2924 into unsigned arithmetic.
2925
2926 2013-10-16 Michael Meissner <meissner@linux.vnet.ibm.com>
2927
2928 PR target/57756
2929 * config/rs6000/rs6000.opt (rs6000_isa_flags_explicit): Move the
2930 explicit isa flag to be an options variable, instead of using
2931 global_options_set. Remove define from rs6000.h.
2932 * config/rs6000/rs6000.h (rs6000_isa_flags_explicit): Likewise.
2933
2934 * config/rs6000/rs6000.c (rs6000_option_override_internal):
2935 Initialize rs6000_isa_flags_explicit.
2936 (rs6000_function_specific_save): Add gcc_options* parameter, so
2937 that the powerpc builds after the 2013-10-15 changes.
2938 (rs6000_function_specific_restore): Likewise.
2939
2940 2013-10-16 DJ Delorie <dj@redhat.com>
2941
2942 * config/rl78/rl78.c (rl78_alloc_address_registers_macax): Verify
2943 op is a REG before checking REGNO.
2944 (rl78_alloc_physical_registers): Verify pattern is a SET before
2945 checking SET_SRC.
2946
2947 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2948
2949 * gcc/config/rs6000/vector.md (vec_unpacks_hi_v4sf): Correct for
2950 endianness.
2951 (vec_unpacks_lo_v4sf): Likewise.
2952 (vec_unpacks_float_hi_v4si): Likewise.
2953 (vec_unpacks_float_lo_v4si): Likewise.
2954 (vec_unpacku_float_hi_v4si): Likewise.
2955 (vec_unpacku_float_lo_v4si): Likewise.
2956
2957 2013-10-16 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2958
2959 * config/rs6000/vsx.md (vsx_concat_<mode>): Adjust output for LE.
2960 (vsx_concat_v2sf): Likewise.
2961
2962 2013-10-16 James Greenhalgh <james.greenhalgh@arm.com>
2963
2964 * config/aarch64/aarch64.md
2965 (*mov<mode>_aarch64): Fix output template for DUP (element) Scalar.
2966
2967 2013-10-16 Andrew MacLeod <amacleod@redhat.com>
2968
2969 PR tree-optimization/58697
2970 * cfgloop.c (get_estimated_loop_iterations_int): Rename from
2971 estimated_loop_iterations_int.
2972 (max_stmt_executions_int): Call get_max_loop_iterations_int.
2973 (get_max_loop_iterations_int): New. HWINT version of
2974 get_max_loop_iterations.
2975 * cfgloop.h: Add prototypes.
2976 * loop-iv.c (find_simple_exit): call get_estimated_loop_iterations_int.
2977 * loop-unroll.c (decide_peel_once_rolling): Call
2978 get_estimated_loop_iterations_int.
2979 * tree-ssa-loop-niter.c (estimated_loop_iterations_int): Add back.
2980 * tree-ssa-loop-niter.h: Tweak prototypes.
2981
2982 2013-10-16 David Malcolm <dmalcolm@redhat.com>
2983
2984 * gengtype-parse.c (struct_field_seq): Ignore access-control
2985 keywords ("public:" etc).
2986
2987 2013-10-16 Marcus Shawcroft <marcus.shawcroft@arm.com>
2988
2989 * config/aarch64/aarch64.c (aarch64_regno_regclass): Classify
2990 FRAME_POINTER_REGNUM and ARG_POINTER_REGNUM as POINTER_REGS.
2991
2992 2013-10-16 Yvan Roux <yvan.roux@linaro.org>
2993
2994 * config/arm/arm.opt (mlra): New option.
2995 * config/arm/arm.c (arm_lra_p): New function.
2996 (TARGET_LRA_P): Define.
2997
2998 2013-10-16 Paulo Matos <pmatos@broadcom.com>
2999
3000 * tree-core.h (tree_code_name): Remove.
3001 * tree.h (get_tree_code_name): New prototype.
3002 * tree.c (tree_code_name): Make static.
3003 (get_tree_code_name): New function.
3004 (dump_tree_statistics, tree_check_failed, tree_not_check_failed,
3005 tree_class_check_failed, tree_range_check_failed,
3006 tree_not_class_check_failed, omp_clause_check_failed,
3007 tree_contains_struct_check_failed, tree_operand_check_failed): Use new
3008 wrapper get_tree_code_name instead of calling tree_code_name directly.
3009 * tree-vrp.c (dump_asserts_for): Likewise.
3010 * tree-dump.c (dequeue_and_dump): Likewise.
3011 * tree-pretty-print.c (do_niy, dump_generic_node): Likewise.
3012 * tree-pretty-print.h (pp_unsupported_tree): Likewise.
3013 * lto-streamer-out.c (lto_write_tree, DFS_write_tree): Likewise.
3014 * tree-ssa-dom.c (print_expr_hash_elt): Likewise.
3015 * gimple-pretty-print.c (dump_unary_rhs, dump_binary_rhs,
3016 dump_ternary_rhs, dump_gimple_assign, dump_gimple_cond,
3017 dump_gimple_omp_for): Likewise.
3018 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Likewise.
3019 * tree-ssa-pre.c (print_pre_expr): Likewise.
3020 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Likewise.
3021 * print-tree.c (print_node_brief, print_node): Likewise.
3022 * gimple.c (gimple_check_failed): Likewise.
3023 * lto-streamer.c (lto_tag_name, print_lto_report): Likewise.
3024 * config/frv/frv.c (frv_init_cumulative_args): Likewise.
3025 * config/mep/mep.c (mep_validate_vliw): Likewise.
3026 * config/iq2000/iq2000.c (init_cumulative_args): Likewise.
3027 * config/rs6000/rs6000.c (init_cumulative_args): Likewise.
3028
3029 2013-10-16 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
3030
3031 * config/i386/i386.c (ix86_option_override_internal): Enable FMA4
3032 for AMD bdver3.
3033
3034 2013-10-16 Hans-Peter Nilsson <hp@axis.com>
3035
3036 * config/cris/t-elfmulti (MULTILIB_OPTIONS, MULTILIB_DIRNAMES)
3037 (MULTILIB_MATCHES): Add multilib for -march=v8.
3038
3039 2013-10-15 Sriraman Tallam <tmsriram@google.com>
3040
3041 PR target/57756
3042 * optc-save-gen.awk: Add extra parameter to the save and restore
3043 target calls.
3044 * opth-gen.awk: Generate new TARGET_* macros to accept a parameter.
3045 * tree.c (build_optimization_node): New parameter. Add extra parameter
3046 to call to cl_optimization_save.
3047 (build_target_option_node): New parameter. Add extra parameter
3048 to call to cl_target_option_save.
3049 * tree.h (build_optimization_node): New parameter.
3050 (build_target_option_node): New parameter.
3051 * c-family/c-common.c (handle_optimize_attribute): Fix calls to
3052 build_optimization_node and build_target_option_node.
3053 * c-family/c-pragma.c (handle_pragma_optimize): Ditto.
3054 (handle_pragma_push_options): Ditto.
3055 * toplev.c (process_options): Ditto.
3056 * opts.c (init_options_struct): Check for opts_set non-null.
3057 * target.def (target_option.save): New parameter.
3058 (target_option.restore): New parameter.
3059 * tm.texi: Generate.
3060 * config/i386/i386-c.c (ix86_target_macros_internal): Ditto.
3061 (ix86_pragma_target_parse): Ditto.
3062 * config/i386/i386-protos.h (ix86_valid_target_attribute_tree): New
3063 parameters.
3064 * config/rs6000/rs6000.c (rs6000_option_override_internal): Fix calls
3065 to build_optimization_node and build_target_option_node.
3066 (rs6000_valid_attribute_p): Ditto.
3067 (rs6000_pragma_target_parse): Ditto.
3068 * config/i386/i386.opt (x_ix86_target_flags_explicit): New TargetSave
3069 data.
3070 * config/i386/i386.h:
3071 TARGET_64BIT_P: New Macro
3072 TARGET_MMX_P: New Macro.
3073 TARGET_3DNOW_P: New Macro.
3074 TARGET_3DNOW_A_P: New Macro.
3075 TARGET_SSE_P: New Macro.
3076 TARGET_SSE2_P: New Macro.
3077 TARGET_SSE3_P: New Macro.
3078 TARGET_SSSE3_P: New Macro.
3079 TARGET_SSE4_1_P: New Macro.
3080 TARGET_SSE4_2_P: New Macro.
3081 TARGET_AVX_P: New Macro.
3082 TARGET_AVX2_P: New Macro.
3083 TARGET_AVX512F_P: New Macro.
3084 TARGET_AVX512PF_P: New Macro.
3085 TARGET_AVX512ER_P: New Macro.
3086 TARGET_AVX512CD_P: New Macro.
3087 TARGET_FMA_P: New Macro.
3088 TARGET_SSE4A_P: New Macro.
3089 TARGET_FMA4_P: New Macro.
3090 TARGET_XOP_P: New Macro.
3091 TARGET_LWP_P: New Macro.
3092 TARGET_ABM_P: New Macro.
3093 TARGET_BMI_P: New Macro.
3094 TARGET_BMI2_P: New Macro.
3095 TARGET_LZCNT_P: New Macro.
3096 TARGET_TBM_P: New Macro.
3097 TARGET_POPCNT_P: New Macro.
3098 TARGET_SAHF_P: New Macro.
3099 TARGET_MOVBE_P: New Macro.
3100 TARGET_CRC32_P: New Macro.
3101 TARGET_AES_P: New Macro.
3102 TARGET_PCLMUL_P: New Macro.
3103 TARGET_CMPXCHG16B_P: New Macro.
3104 TARGET_FSGSBASE_P: New Macro.
3105 TARGET_RDRND_P: New Macro.
3106 TARGET_F16C_P: New Macro.
3107 TARGET_RTM_P: New Macro.
3108 TARGET_HLE_P: New Macro.
3109 TARGET_RDSEED_P: New Macro.
3110 TARGET_PRFCHW_P: New Macro.
3111 TARGET_ADX_P: New Macro.
3112 TARGET_FXSR_P: New Macro.
3113 TARGET_XSAVE_P: New Macro.
3114 TARGET_XSAVEOPT_P: New Macro.
3115 TARGET_LP64_P: New Macro.
3116 TARGET_X32_P: New Macro.
3117 TARGET_FPMATH_DEFAULT_P: New Macro.
3118 TARGET_FLOAT_RETURNS_IN_80387_P: New Macro.
3119 * config/i386/i386.c (ix86_option_override_internal): New parameters.
3120 opts and opts_set.
3121 Change ix86_tune_string to access opts->x_ix86_tune_string.
3122 Change ix86_isa_flags to access opts->x_ix86_isa_flags.
3123 Change ix86_arch_string to access opts->x_ix86_arch_string.
3124 Change ix86_stringop_alg to access opts->x_ix86_stringop_alg.
3125 Change ix86_pmode to access opts->x_ix86_pmode.
3126 Change ix86_abi to access opts->x_ix86_abi.
3127 Change ix86_cmodel to access opts->x_ix86_cmodel.
3128 Change ix86_asm_dialect to access opts->x_ix86_asm_dialect.
3129 Change ix86_isa_flags_explicit to access
3130 opts->x_ix86_isa_flags_explicit.
3131 Change ix86_dump_tunes to access opts->x_ix86_dump_tunes.
3132 Change ix86_regparm to access opts->x_ix86_regparm.
3133 Change ix86_branch_cost to access opts->x_ix86_branch_cost.
3134 Change ix86_preferred_stack_boundary_arg to access
3135 opts->x_ix86_preferred_stack_boundary_arg.
3136 Change ix86_force_align_arg_pointer to access
3137 opts->x_ix86_force_align_arg_pointer.
3138 Change ix86_incoming_stack_boundar_arg to access
3139 opts->x_ix86_incoming_stack_boundar_arg.
3140 Change ix86_fpmath to access opts->x_ix86_fpmath.
3141 Change ix86_veclibabi_type to access opts->x_ix86_veclibabi_type.
3142 Change ix86_recip_name to access opts->x_ix86_recip_name.
3143 Change ix86_stack_protector_guard to access
3144 opts->x_ix86_stack_protector_guard.
3145 Change ix86_tune_memcpy_strategy to access
3146 opts->x_ix86_tune_memcpy_strategy.
3147 Change ix86_tune_memset_strategy to access
3148 opts->x_ix86_tune_memset_strategy.
3149 Change global_options to access opts.
3150 Change global_options_set to access opts_set.
3151 Change TARGET_64BIT to TARGET_64BIT_P (opts->...).
3152 Change TARGET_MMX to TARGET_MMX_P (opts->...).
3153 Change TARGET_3DNOW to TARGET_3DNOW_P (opts->...).
3154 Change TARGET_3DNOW_A to TARGET_3DNOW_A_P (opts->...).
3155 Change TARGET_SSE to TARGET_SSE_P (opts->...).
3156 Change TARGET_SSE2 to TARGET_SSE2_P (opts->...).
3157 Change TARGET_SSE3 to TARGET_SSE3_P (opts->...).
3158 Change TARGET_SSSE3 to TARGET_SSSE3_P (opts->...).
3159 Change TARGET_SSE4_1 to TARGET_SSE4_1_P (opts->...).
3160 Change TARGET_SSE4_2 to TARGET_SSE4_2_P (opts->...).
3161 Change TARGET_AVX to TARGET_AVX_P (opts->...).
3162 Change TARGET_AVX2 to TARGET_AVX2_P (opts->...).
3163 Change TARGET_AVX512F to TARGET_AVX512F_P (opts->...).
3164 Change TARGET_AVX512PF to TARGET_AVX512PF_P (opts->...).
3165 Change TARGET_AVX512ER to TARGET_AVX512ER_P (opts->...).
3166 Change TARGET_AVX512CD to TARGET_AVX512CD_P (opts->...).
3167 Change TARGET_FMA to TARGET_FMA_P (opts->...).
3168 Change TARGET_SSE4A to TARGET_SSE4A_P (opts->...).
3169 Change TARGET_FMA4 to TARGET_FMA4_P (opts->...).
3170 Change TARGET_XOP to TARGET_XOP_P (opts->...).
3171 Change TARGET_LWP to TARGET_LWP_P (opts->...).
3172 Change TARGET_ABM to TARGET_ABM_P (opts->...).
3173 Change TARGET_BMI to TARGET_BMI_P (opts->...).
3174 Change TARGET_BMI2 to TARGET_BMI2_P (opts->...).
3175 Change TARGET_LZCNT to TARGET_LZCNT_P (opts->...).
3176 Change TARGET_TBM to TARGET_TBM_P (opts->...).
3177 Change TARGET_POPCNT to TARGET_POPCNT_P (opts->...).
3178 Change TARGET_SAHF to TARGET_SAHF_P (opts->...).
3179 Change TARGET_MOVBE to TARGET_MOVBE_P (opts->...).
3180 Change TARGET_CRC32 to TARGET_CRC32_P (opts->...).
3181 Change TARGET_AES to TARGET_AES_P (opts->...).
3182 Change TARGET_PCLMUL to TARGET_PCLMUL_P (opts->...).
3183 Change TARGET_CMPXCHG16B to TARGET_CMPXCHG16B_P (opts->...).
3184 Change TARGET_FSGSBASE to TARGET_FSGSBASE_P (opts->...).
3185 Change TARGET_RDRND to TARGET_RDRND_P (opts->...).
3186 Change TARGET_F16C to TARGET_F16C_P (opts->...).
3187 Change TARGET_RTM to TARGET_RTM_P (opts->...).
3188 Change TARGET_HLE to TARGET_HLE_P (opts->...).
3189 Change TARGET_RDSEED to TARGET_RDSEED_P (opts->...).
3190 Change TARGET_PRFCHW to TARGET_PRFCHW_P (opts->...).
3191 Change TARGET_ADX to TARGET_ADX_P (opts->...).
3192 Change TARGET_FXSR to TARGET_FXSR_P (opts->...).
3193 Change TARGET_XSAVE to TARGET_XSAVE_P (opts->...).
3194 Change TARGET_XSAVEOPT to TARGET_XSAVEOPT_P (opts->...).
3195 Change TARGET_LP64 to TARGET_LP64_P (opts->...).
3196 Change TARGET_X32 to TARGET_X32_P (opts->...).
3197 Change TARGET_FPMATH_DEFAULT to TARGET_FPMATH_DEFAULT_P (opts->...).
3198 Change TARGET_FLOAT_RETURNS_IN_80387 to
3199 TARGET_FLOAT_RETURNS_IN_80387_P (opts->...).
3200 (ix86_function_specific_save): New parameter. Use opts-> fields
3201 to replace global fields.
3202 (ix86_function_specific_restore): Ditto.
3203 (ix86_valid_target_attribute_inner_p): New parameters.
3204 Fix recursive call.
3205 Fix call to ix86_handle_option and set_option.
3206 (ix86_valid_target_attribute_tree): New parameters.
3207 Change global_options to access opts.
3208 Change global_options_set to access opts_set.
3209 Fix call to ix86_valid_target_attribute_inner_p.
3210 Change ix86_tune_string to access opts->x_ix86_tune_string.
3211 Change ix86_arch_string to access opts->x_ix86_arch_string.
3212 Change ix86_fpmath to access opts->x_ix86_fpmath
3213 Fix call to ix86_option_override_internal.
3214 Fix call to ix86_add_new_builtins.
3215 Fix calls to build_optimization_node and build_target_option_node.
3216 (ix86_valid_target_attribute_p): Remove access to global_options.
3217 Use new gcc_options structure func_options.
3218 Fix call to ix86_valid_target_attribute_tree.
3219 Fix call to build_optimization_node.
3220 (get_builtin_code_for_version): Fix call to
3221 ix86_valid_target_attribute_tree.
3222
3223 2013-10-15 David Malcolm <dmalcolm@redhat.com>
3224
3225 * Makefile.in (PICFLAG): New.
3226 (enable_host_shared): New.
3227 (INTERNAL_CFLAGS): Use PICFLAG.
3228 (LIBIBERTY): Use pic build of libiberty.a if configured with
3229 --enable-host-shared.
3230 * configure.ac: Add --enable-host-shared, setting up new
3231 PICFLAG variable.
3232 * configure: Regenerate.
3233 * doc/install.texi (--enable-shared): Add note contrasting it with ...
3234 (--enable-host-shared): New option.
3235
3236 2013-10-15 Richard Biener <rguenther@suse.de>
3237
3238 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
3239 for built-in functions.
3240
3241 2013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
3242
3243 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
3244 (optimize_range_tests_1): New function,
3245 extracted from optimize_range_tests.
3246 (optimize_range_tests_xor): Similarly.
3247 (optimize_range_tests_diff): New function.
3248 (optimize_range_tests): Use optimize_range_tests_1.
3249
3250 2013-10-15 Cong Hou <congh@google.com>
3251
3252 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
3253 requirement of the reduction pattern so that one operand of the
3254 reduction operation can come from outside of the loop.
3255
3256 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3257
3258 * config/arm/neon-schedgen.ml: Remove.
3259 * config/arm/cortex-a9-neon.md: Remove comment regarding
3260 neon-schedgen.ml.
3261
3262 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3263
3264 * config/arm/types: Remove old neon types.
3265
3266 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3267
3268 * config/arm/cortex-a7.md
3269 (cortex_a7_neon_type): New.
3270 (cortex_a7_neon_mul): Update for new types.
3271 (cortex_a7_neon_mla): Likewise.
3272 (cortex_a7_neon): Likewise.
3273
3274 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3275
3276 * config/arm/cortex-a15-neon.md
3277 (cortex_a15_neon_type): New,
3278
3279 (cortex_a15_neon_int_1): Remove.
3280 (cortex_a15_neon_int_2): Likewise.
3281 (cortex_a15_neon_int_3): Likewise.
3282 (cortex_a15_neon_int_4): Likewise.
3283 (cortex_a15_neon_int_5): Likewise.
3284 (cortex_a15_neon_vqneg_vqabs): Likewise.
3285 (cortex_a15_neon_vmov): Likewise.
3286 (cortex_a15_neon_vaba): Likewise.
3287 (cortex_a15_neon_vaba_qqq): Likewise.
3288 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3289 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3290 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3291 Likewise.
3292 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3293 (cortex_a15_neon_mla_qqq_8_16): Likewise.
3294 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
3295 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3296 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3297 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
3298 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3299 (cortex_a15_neon_shift_1): Likewise.
3300 (cortex_a15_neon_shift_2): Likewise.
3301 (cortex_a15_neon_shift_3): Likewise.
3302 (cortex_a15_neon_vshl_ddd): Likewise.
3303 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3304 (cortex_a15_neon_vsra_vrsra): Likewise.
3305 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3306 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3307 (cortex_a15_neon_bp_3cycle): Likewise.
3308 (cortex_a15_neon_ldm_2): Likewise.
3309 (cortex_a15_neon_stm_2): Likewise.
3310 (cortex_a15_neon_mcr): Likewise.
3311 (cortex_a15_neon_mrc): Likewise.
3312 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
3313 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
3314 (cortex_a15_neon_fp_vmul_ddd): Likewise.
3315 (cortex_a15_neon_fp_vmul_qqd): Likewise.
3316 (cortex_a15_neon_fp_vmla_ddd): Likewise.
3317 (cortex_a15_neon_fp_vmla_qqq): Likewise.
3318 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
3319 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
3320 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3321 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3322 (cortex_a15_neon_bp_simple): Likewise.
3323 (cortex_a15_neon_bp_2cycle): Likewise.
3324 (cortex_a15_neon_bp_3cycle): Likewise.
3325 (cortex_a15_neon_vld1_1_2_regs): Likewise.
3326 (cortex_a15_neon_vld1_3_4_regs): Likewise.
3327 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3328 (cortex_a15_neon_vld2_4_regs): Likewise.
3329 (cortex_a15_neon_vld3_vld4): Likewise.
3330 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3331 (cortex_a15_neon_vst1_3_4_regs): Likewise.
3332 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
3333 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
3334 (cortex_a15_neon_vst3_vst4): Rename to...
3335 (cortex_a15_neon_vst4): This, update for new attributes.
3336 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
3337 (cortex_a15_neon_vld3_vld4_lane): Likewise.
3338 (cortex_a15_neon_vst1_vst2_lane): Likewise.
3339 (cortex_a15_neon_vst3_vst4_lane): Likewise.
3340 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
3341 (cortex_a15_neon_ldm_2): Likewise.
3342 (cortex_a15_neon_stm_2): Likewise.
3343 (cortex_a15_neon_mcr): Likewise.
3344 (cortex_a15_neon_mcr_2_mcrr): Likewise.
3345 (cortex_a15_neon_mrc): Likewise.
3346 (cortex_a15_neon_mrrc): Likewise.
3347
3348 (cortex_a15_neon_abd): New.
3349 (cortex_a15_neon_abd_q): Likewise.
3350 (cortex_a15_neon_aba): Likewise.
3351 (cortex_a15_neon_aba_q): Likewise.
3352 (cortex_a15_neon_acc): Likewise.
3353 (cortex_a15_neon_acc_q): Likewise.
3354 (cortex_a15_neon_arith_basic): Likewise.
3355 (cortex_a15_neon_arith_complex): Likewise.
3356 (cortex_a15_neon_multiply): Likewise.
3357 (cortex_a15_neon_multiply_q): Likewise.
3358 (cortex_a15_neon_mla): Likewise.
3359 (cortex_a15_neon_mla_q): Likewise.
3360 (cortex_a15_neon_sat_mla_long): Likewise.
3361 (cortex_a15_neon_shift_acc): Likewise.
3362 (cortex_a15_neon_shift_imm_basic): Likewise.
3363 (cortex_a15_neon_shift_imm_complex): Likewise.
3364 (cortex_a15_neon_shift_reg_basic): Likewise.
3365 (cortex_a15_neon_shift_reg_basic_q): Likewise.
3366 (cortex_a15_neon_shift_reg_complex): Likewise.
3367 (cortex_a15_neon_shift_reg_complex_q): Likewise.
3368 (cortex_a15_neon_fp_negabs): Likewise
3369 (cortex_a15_neon_fp_arith): Likewise
3370 (cortex_a15_neon_fp_arith_q): Likewise
3371 (cortex_a15_neon_fp_cvt_int): Likewise
3372 (cortex_a15_neon_fp_cvt_int_q): Likewise
3373 (cortex_a15_neon_fp_cvt_16): Likewise
3374 (cortex_a15_neon_fp_mul): Likewise
3375 (cortex_a15_neon_fp_mul_q): Likewise
3376 (cortex_a15_neon_fp_mla): Likewise
3377 (cortex_a15_neon_fp_mla_q): Likewise
3378 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
3379 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
3380 (cortex_a15_neon_bitops): Likewise.
3381 (cortex_a15_neon_bitops_q): Likewise.
3382 (cortex_a15_neon_from_gp): Likewise.
3383 (cortex_a15_neon_from_gp_q): Likewise.
3384 (cortex_a15_neon_tbl3_tbl4): Likewise.
3385 (cortex_a15_neon_zip_q): Likewise.
3386 (cortex_a15_neon_to_gp): Likewise.
3387 (cortex_a15_neon_load_a): Likewise.
3388 (cortex_a15_neon_load_b): Likewise.
3389 (cortex_a15_neon_load_c): Likewise.
3390 (cortex_a15_neon_load_d): Likewise.
3391 (cortex_a15_neon_load_e): Likewise.
3392 (cortex_a15_neon_load_f): Likewise.
3393 (cortex_a15_neon_store_a): Likewise.
3394 (cortex_a15_neon_store_b): Likewise.
3395 (cortex_a15_neon_store_c): Likewise.
3396 (cortex_a15_neon_store_d): Likewise.
3397 (cortex_a15_neon_store_e): Likewise.
3398 (cortex_a15_neon_store_f): Likewise.
3399 (cortex_a15_neon_store_g): Likewise.
3400 (cortex_a15_neon_store_h): Likewise.
3401 (cortex_a15_vfp_to_from_gp): Likewise.
3402
3403 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3404
3405 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
3406
3407 (cortex_a9_neon_vshl_ddd): Remove.
3408 (cortex_a9_neon_vst3_vst4): Likewise.
3409 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3410
3411 (cortex_a9_neon_bit_ops_q): New.
3412
3413 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
3414 (cortex_a9_neon_int_2): Likewise.
3415 (cortex_a9_neon_int_3): Likewise.
3416 (cortex_a9_neon_int_4): Likewise.
3417 (cortex_a9_neon_int_5): Likewise.
3418 (cortex_a9_neon_vqneg_vqabs): Likewise.
3419 (cortex_a9_neon_vmov): Likewise.
3420 (cortex_a9_neon_vaba): Likewise.
3421 (cortex_a9_neon_vaba_qqq): Likewise.
3422 (cortex_a9_neon_shift_1): Likewise.
3423 (cortex_a9_neon_shift_2): Likewise.
3424 (cortex_a9_neon_shift_3): Likewise.
3425 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3426 (cortex_a9_neon_vsra_vrsra): Likewise.
3427 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3428 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3429 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3430 Likewise.
3431 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3432 (cortex_a9_neon_mla_qqq_8_16): Likewise.
3433 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3434 Likewise.
3435 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3436 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3437 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
3438 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3439 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
3440 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
3441 (cortex_a9_neon_fp_vsum): Likewise.
3442 (cortex_a9_neon_fp_vmul_ddd): Likewise.
3443 (cortex_a9_neon_fp_vmul_qqd): Likewise.
3444 (cortex_a9_neon_fp_vmla_ddd): Likewise.
3445 (cortex_a9_neon_fp_vmla_qqq): Likewise.
3446 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
3447 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
3448 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3449 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3450 (cortex_a9_neon_bp_simple): Likewise.
3451 (cortex_a9_neon_bp_2cycle): Likewise.
3452 (cortex_a9_neon_bp_3cycle): Likewise.
3453 (cortex_a9_neon_ldr): Likewise.
3454 (cortex_a9_neon_str): Likewise.
3455 (cortex_a9_neon_vld1_1_2_regs): Likewise.
3456 (cortex_a9_neon_vld1_3_4_regs): Likewise.
3457 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3458 (cortex_a9_neon_vld2_4_regs): Likewise.
3459 (cortex_a9_neon_vld3_vld4): Likewise.
3460 (cortex_a9_neon_vld1_vld2_lane): Likewise.
3461 (cortex_a9_neon_vld3_vld4_lane): Likewise.
3462 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
3463 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3464 (cortex_a9_neon_vst1_3_4_regs): Likewise.
3465 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
3466 (cortex_a9_neon_vst1_vst2_lane): Likewise.
3467 (cortex_a9_neon_vst3_vst4_lane): Likewise.
3468 (cortex_a9_neon_mcr): Likewise.
3469 (cortex_a9_neon_mcr_2_mcrr): Likewise.
3470 (cortex_a9_neon_mrc): Likewise.
3471 (cortex_a9_neon_mrrc): Likewise.
3472
3473 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3474
3475 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
3476
3477 (cortex_a8_neon_vshl_ddd): Remove.
3478 (cortex_a8_neon_vst3_vst4): Likewise.
3479 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
3480
3481 (cortex_a8_neon_bit_ops_q): New.
3482
3483 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
3484 (cortex_a8_neon_int_2): Likewise..
3485 (cortex_a8_neon_int_3): Likewise.
3486 (cortex_a8_neon_int_5): Likewise.
3487 (cortex_a8_neon_vqneg_vqabs): Likewise.
3488 (cortex_a8_neon_int_4): Likewise.
3489 (cortex_a8_neon_vaba): Likewise.
3490 (cortex_a8_neon_vaba_qqq): Likewise.
3491 (cortex_a8_neon_shift_1): Likewise.
3492 (cortex_a8_neon_shift_2): Likewise.
3493 (cortex_a8_neon_shift_3): Likewise.
3494 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
3495 (cortex_a8_neon_vsra_vrsra): Likewise.
3496 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3497 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
3498 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
3499 Likewise.
3500 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
3501 (cortex_a8_neon_mla_qqq_8_16): Likewise.
3502 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
3503 Likewise.
3504 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
3505 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
3506 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
3507 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
3508 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
3509 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
3510 (cortex_a8_neon_fp_vsum): Likewise.
3511 (cortex_a8_neon_fp_vmul_ddd): Likewise.
3512 (cortex_a8_neon_fp_vmul_qqd): Likewise.
3513 (cortex_a8_neon_fp_vmla_ddd): Likewise.
3514 (cortex_a8_neon_fp_vmla_qqq): Likewise.
3515 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
3516 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
3517 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
3518 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
3519 (cortex_a8_neon_bp_simple): Likewise.
3520 (cortex_a8_neon_bp_2cycle): Likewise.
3521 (cortex_a8_neon_bp_3cycle): Likewise.
3522 (cortex_a8_neon_ldr): Likewise.
3523 (cortex_a8_neon_str): Likewise.
3524 (cortex_a8_neon_vld1_1_2_regs): Likewise.
3525 (cortex_a8_neon_vld1_3_4_regs): Likewise.
3526 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
3527 (cortex_a8_neon_vld2_4_regs): Likewise.
3528 (cortex_a8_neon_vld3_vld4): Likewise.
3529 (cortex_a8_neon_vld1_vld2_lane): Likewise.
3530 (cortex_a8_neon_vld3_vld4_lane): Likewise.
3531 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
3532 (cortex_a8_neon_vst1_3_4_regs): Likewise.
3533 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
3534 (cortex_a8_neon_vst1_vst2_lane): Likewise.
3535 (cortex_a8_neon_vst3_vst4_lane): Likewise.
3536 (cortex_a8_neon_mcr): Likewise.
3537 (cortex_a8_neon_mcr_2_mcrr): Likewise.
3538 (cortex_a8_neon_mrc): Likewise.
3539 (cortex_a8_neon_mrrc): Likewise.
3540
3541 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3542
3543 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
3544 (fp): New.
3545 * config/aarch64/aarch64-simd.md (neon_type): Remove.
3546 (aarch64_simd_dup<mode>): Add "type" attribute.
3547 (aarch64_dup_lane<mode>): Likewise.
3548 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
3549 (*aarch64_simd_mov<mode>): Likewise.
3550 (aarch64_simd_mov_from_<mode>low): Likewise.
3551 (aarch64_simd_mov_from_<mode>high): Likewise.
3552 (orn<mode>3): Likewise.
3553 (bic<mode>3): Likewise.
3554 (add<mode>3): Likewise.
3555 (sub<mode>3): Likewise.
3556 (mul<mode>3): Likewise.
3557 (*aarch64_mul3_elt<mode>): Likewise.
3558 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
3559 (*aarch64_mul3_elt_to_128df): Likewise.
3560 (*aarch64_mul3_elt_to_64v2df): Likewise.
3561 (neg<mode>2): Likewise.
3562 (abs<mode>2): Likewise.
3563 (abd<mode>_3): Likewise.
3564 (aba<mode>_3): Likewise.
3565 (fabd<mode>_3): Likewise.
3566 (*fabd_scalar<mode>3): Likewise.
3567 (and<mode>3): Likewise.
3568 (ior<mode>3): Likewise.
3569 (xor<mode>3): Likewise.
3570 (one_cmpl<mode>2): Likewise.
3571 (aarch64_simd_vec_set<mode>): Likewise.
3572 (aarch64_simd_lshr<mode>): Likewise.
3573 (aarch64_simd_ashr<mode>): Likewise.
3574 (aarch64_simd_imm_shl<mode>): Likewise.
3575 (aarch64_simd_reg_sshl<mode): Likewise.
3576 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
3577 (aarch64_simd_reg_shl<mode>_signed): Likewise.
3578 (aarch64_simd_vec_setv2di): Likewise.
3579 (aarch64_simd_vec_set<mode>): Likewise.
3580 (aarch64_mla<mode>): Likewise.
3581 (*aarch64_mla_elt<mode>): Likewise.
3582 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
3583 (aarch64_mls<mode>): Likewise.
3584 (*aarch64_mls_elt<mode>): Likewise.
3585 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
3586 (<su><maxmin><mode>3): Likewise.
3587 (move_lo_quad_<mode>): Likewise.
3588 (aarch64_simd_move_hi_quad_<mode>): Likewise.
3589 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
3590 (vec_pack_trunc_<mode>): Likewise.
3591 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
3592 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
3593 (*aarch64_<su>mlal_lo<mode>): Likewise.
3594 (*aarch64_<su>mlal_hi<mode>): Likewise.
3595 (*aarch64_<su>mlsl_lo<mode>): Likewise.
3596 (*aarch64_<su>mlsl_hi<mode>): Likewise.
3597 (*aarch64_<su>mlal<mode>): Likewise.
3598 (*aarch64_<su>mlsl<mode>): Likewise.
3599 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
3600 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
3601 (add<mode>3): Likewise.
3602 (sub<mode>3): Likewise.
3603 (mul<mode>3): Likewise.
3604 (div<mode>3): Likewise.
3605 (neg<mode>2): Likewise.
3606 (abs<mode>2): Likewise.
3607 (fma<mode>4): Likewise.
3608 (*aarch64_fma4_elt<mode>): Likewise.
3609 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
3610 (*aarch64_fma4_elt_to_128df): Likewise.
3611 (*aarch64_fma4_elt_to_64v2df): Likewise.
3612 (fnma<mode>4): Likewise.
3613 (*aarch64_fnma4_elt<mode>): Likewise.
3614 (*aarch64_fnma4_elt_<vswap_width_name><mode>
3615 (*aarch64_fnma4_elt_to_128df): Likewise.
3616 (*aarch64_fnma4_elt_to_64v2df): Likewise.
3617 (<frint_pattern><mode>2): Likewise.
3618 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
3619 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
3620 (vec_unpacks_lo_v4sf): Likewise.
3621 (aarch64_float_extend_lo_v2df): Likewise.
3622 (vec_unpacks_hi_v4sf): Likewise.
3623 (aarch64_float_truncate_lo_v2sf): Likewise.
3624 (aarch64_float_truncate_hi_v4sf): Likewise.
3625 (aarch64_vmls<mode>): Likewise.
3626 (<su><maxmin><mode>3): Likewise.
3627 (<maxmin_uns><mode>3): Likewise.
3628 (reduc_<sur>plus_<mode>): Likewise.
3629 (reduc_<sur>plus_v2di): Likewise.
3630 (reduc_<sur>plus_v2si): Likewise.
3631 (reduc_<sur>plus_<mode>): Likewise.
3632 (aarch64_addpv4sf): Likewise.
3633 (clz<mode>2): Likewise.
3634 (reduc_<maxmin_uns>_<mode>): Likewise.
3635 (reduc_<maxmin_uns>_v2di): Likewise.
3636 (reduc_<maxmin_uns>_v2si): Likewise.
3637 (reduc_<maxmin_uns>_<mode>): Likewise.
3638 (reduc_<maxmin_uns>_v4sf): Likewise.
3639 (aarch64_simd_bsl<mode>_internal): Likewise.
3640 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
3641 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
3642 (aarch64_get_lane<mode>): Likewise.
3643 (*aarch64_combinez<mode>): Likewise.
3644 (aarch64_combine<mode>): Likewise.
3645 (aarch64_simd_combine<mode>): Likewise.
3646 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
3647 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
3648 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
3649 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
3650 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
3651 (aarch64_<sur>h<addsub><mode>): Likewise.
3652 (aarch64_<sur><addsub>hn<mode>): Likewise.
3653 (aarch64_<sur><addsub>hn2<mode>): Likewise.
3654 (aarch64_pmul<mode>): Likewise.
3655 (aarch64_<su_optab><optab><mode>): Likewise.
3656 (aarch64_<sur>qadd<mode>): Likewise.
3657 (aarch64_sqmovun<mode>): Likewise.
3658 (aarch64_<sur>qmovn<mode>): Likewise.
3659 (aarch64_s<optab><mode>): Likewise.
3660 (aarch64_sq<r>dmulh<mode>): Likewise.
3661 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3662 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
3663 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
3664 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
3665 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3666 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
3667 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
3668 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
3669 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
3670 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
3671 (aarch64_sqdmull<mode>): Likewise.
3672 (aarch64_sqdmull_lane<mode>_internal): Likewise.
3673 (aarch64_sqdmull_n<mode>): Likewise.
3674 (aarch64_sqdmull2<mode>_internal): Likewise.
3675 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
3676 (aarch64_sqdmull2_n<mode>_internal): Likewise.
3677 (aarch64_<sur>shl<mode>): Likewise.
3678 (aarch64_<sur>q<r>shl<mode>
3679 (aarch64_<sur>shll_n<mode>): Likewise.
3680 (aarch64_<sur>shll2_n<mode>): Likewise.
3681 (aarch64_<sur>shr_n<mode>): Likewise.
3682 (aarch64_<sur>sra_n<mode>): Likewise.
3683 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
3684 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
3685 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
3686 (aarch64_cm<optab><mode>): Likewise.
3687 (aarch64_cm<optab>di): Likewise.
3688 (aarch64_cm<optab><mode>): Likewise.
3689 (aarch64_cm<optab>di): Likewise.
3690 (aarch64_cmtst<mode>): Likewise.
3691 (aarch64_cmtstdi): Likewise.
3692 (aarch64_cm<optab><mode>): Likewise.
3693 (*aarch64_fac<optab><mode>): Likewise.
3694 (aarch64_addp<mode>): Likewise.
3695 (aarch64_addpdi): Likewise.
3696 (sqrt<mode>2): Likewise.
3697 (vec_load_lanesoi<mode>): Likewise.
3698 (vec_store_lanesoi<mode>): Likewise.
3699 (vec_load_lanesci<mode>): Likewise.
3700 (vec_store_lanesci<mode>): Likewise.
3701 (vec_load_lanesxi<mode>): Likewise.
3702 (vec_store_lanesxi<mode>): Likewise.
3703 (*aarch64_mov<mode>): Likewise.
3704 (aarch64_ld2<mode>_dreg): Likewise.
3705 (aarch64_ld2<mode>_dreg): Likewise.
3706 (aarch64_ld3<mode>_dreg): Likewise.
3707 (aarch64_ld3<mode>_dreg): Likewise.
3708 (aarch64_ld4<mode>_dreg): Likewise.
3709 (aarch64_ld4<mode>_dreg): Likewise.
3710 (aarch64_tbl1<mode>): Likewise.
3711 (aarch64_tbl2v16qi): Likewise.
3712 (aarch64_combinev16qi): Likewise.
3713 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
3714 (aarch64_st2<mode>_dreg): Likewise.
3715 (aarch64_st2<mode>_dreg): Likewise.
3716 (aarch64_st3<mode>_dreg): Likewise.
3717 (aarch64_st3<mode>_dreg): Likewise.
3718 (aarch64_st4<mode>_dreg): Likewise.
3719 (aarch64_st4<mode>_dreg): Likewise.
3720 (*aarch64_simd_ld1r<mode>): Likewise.
3721 (aarch64_frecpe<mode>): Likewise.
3722 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
3723 (aarch64_frecps<mode>): Likewise.
3724
3725 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3726
3727 * config/arm/iterators.md (V_elem_ch): New.
3728 (q): Likewise.
3729 (VQH_type): Likewise.
3730 * config/arm/arm.md (is_neon_type): New.
3731 (conds): Use is_neon_type.
3732 (anddi3_insn): Update type attribute.
3733 (xordi3_insn): Likewise.
3734 (one_cmpldi2): Likewise.
3735 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
3736 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
3737 (*movmisalign<mode>_neon_store): Likewise.
3738 (*movmisalign<mode>_neon_load): Likewise.
3739 (vec_set<mode>_internal): Likewise.
3740 (vec_set<mode>_internal): Likewise.
3741 (vec_setv2di_internal): Likewise.
3742 (vec_extract<mode>): Likewise.
3743 (vec_extract<mode>): Likewise.
3744 (vec_extractv2di): Likewise.
3745 (*add<mode>3_neon): Likewise.
3746 (adddi3_neon): Likewise.
3747 (*sub<mode>3_neon): Likewise.
3748 (subdi3_neon): Likewise.
3749 (fma<VCVTF:mode>4): Likewise.
3750 (fma<VCVTF:mode>4_intrinsic): Likewise.
3751 (*fmsub<VCVTF:mode>4): Likewise.
3752 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
3753 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
3754 (ior<mode>3): Likewise.
3755 (and<mode>3): Likewise.
3756 (orn<mode>3_neon): Likewise.
3757 (orndi3_neon): Likewise.
3758 (bic<mode>3_neon): Likewise.
3759 (bicdi3_neon): Likewise.
3760 (xor<mode>3): Likewise.
3761 (one_cmpl<mode>2): Likewise.
3762 (abs<mode>2): Likewise.
3763 (neg<mode>2): Likewise.
3764 (negdi2_neon): Likewise.
3765 (*umin<mode>3_neon): Likewise.
3766 (*umax<mode>3_neon): Likewise.
3767 (*smin<mode>3_neon): Likewise.
3768 (*smax<mode>3_neon): Likewise.
3769 (vashl<mode>3): Likewise.
3770 (vashr<mode>3_imm): Likewise.
3771 (vlshr<mode>3_imm): Likewise.
3772 (ashl<mode>3_signed): Likewise.
3773 (ashl<mode>3_unsigned): Likewise.
3774 (neon_load_count): Likewise.
3775 (ashldi3_neon_noclobber): Likewise.
3776 (ashldi3_neon): Likewise.
3777 (signed_shift_di3_neon): Likewise.
3778 (unsigned_shift_di3_neon): Likewise.
3779 (ashrdi3_neon_imm_noclobber): Likewise.
3780 (lshrdi3_neon_imm_noclobber): Likewise.
3781 (<shift>di3_neon): Likewise.
3782 (widen_ssum<mode>3): Likewise.
3783 (widen_usum<mode>3): Likewise.
3784 (quad_halves_<code>v4si): Likewise.
3785 (quad_halves_<code>v4sf): Likewise.
3786 (quad_halves_<code>v8hi): Likewise.
3787 (quad_halves_<code>v16qi): Likewise.
3788 (reduc_splus_v2di): Likewise.
3789 (neon_vpadd_internal<mode>): Likewise.
3790 (neon_vpsmin<mode>): Likewise.
3791 (neon_vpsmax<mode>): Likewise.
3792 (neon_vpumin<mode>): Likewise.
3793 (neon_vpumax<mode>): Likewise.
3794 (*ss_add<mode>_neon): Likewise.
3795 (*us_add<mode>_neon): Likewise.
3796 (*ss_sub<mode>_neon): Likewise.
3797 (*us_sub<mode>_neon): Likewise.
3798 (neon_vadd<mode>_unspec): Likewise.
3799 (neon_vaddl<mode>): Likewise.
3800 (neon_vaddw<mode>): Likewise.
3801 (neon_vhadd<mode>): Likewise.
3802 (neon_vqadd<mode>): Likewise.
3803 (neon_vaddhn<mode>): Likewise.
3804 (neon_vmul<mode>): Likewise.
3805 (neon_vfms<VCVTF:mode>): Likewise.
3806 (neon_vmlal<mode>): Likewise.
3807 (neon_vmls<mode>): Likewise.
3808 (neon_vmlsl<mode>): Likewise.
3809 (neon_vqdmulh<mode>): Likewise.
3810 (neon_vqdmlal<mode>): Likewise.
3811 (neon_vqdmlsl<mode>): Likewise.
3812 (neon_vmull<mode>): Likewise.
3813 (neon_vqdmull<mode>): Likewise.
3814 (neon_vsub<mode>_unspec): Likewise.
3815 (neon_vsubl<mode>): Likewise.
3816 (neon_vsubw<mode>): Likewise.
3817 (neon_vqsub<mode>): Likewise.
3818 (neon_vhsub<mode>): Likewise.
3819 (neon_vsubhn<mode>): Likewise.
3820 (neon_vceq<mode>): Likewise.
3821 (neon_vcge<mode>): Likewise.
3822 (neon_vcgeu<mode>): Likewise.
3823 (neon_vcgt<mode>): Likewise.
3824 (neon_vcgtu<mode>): Likewise.
3825 (neon_vcle<mode>): Likewise.
3826 (neon_vclt<mode>): Likewise.
3827 (neon_vcage<mode>): Likewise.
3828 (neon_vcagt<mode>): Likewise.
3829 (neon_vtst<mode>): Likewise.
3830 (neon_vabd<mode>): Likewise.
3831 (neon_vabdl<mode>): Likewise.
3832 (neon_vaba<mode>): Likewise.
3833 (neon_vabal<mode>): Likewise.
3834 (neon_vmax<mode>): Likewise.
3835 (neon_vmin<mode>): Likewise.
3836 (neon_vpaddl<mode>): Likewise.
3837 (neon_vpadal<mode>): Likewise.
3838 (neon_vpmax<mode>): Likewise.
3839 (neon_vpmin<mode>): Likewise.
3840 (neon_vrecps<mode>): Likewise.
3841 (neon_vrsqrts<mode>): Likewise.
3842 (neon_vqabs<mode>): Likewise.
3843 (neon_vqneg<mode>): Likewise.
3844 (neon_vcls<mode>): Likewise.
3845 (clz<mode>2): Likewise.
3846 (popcount<mode>2): Likewise.
3847 (neon_vrecpe<mode>): Likewise.
3848 (neon_vrsqrte<mode>): Likewise.
3849 (neon_vget_lane<mode>_sext_internal): Likewise.
3850 (neon_vget_lane<mode>_zext_internal): Likewise.
3851 (neon_vdup_n<mode>): Likewise.
3852 (neon_vdup_n<mode>): Likewise.
3853 (neon_vdup_nv2di): Likewise.
3854 (neon_vdup_lane<mode>_interal): Likewise.
3855 (*neon_vswp<mode>): Likewise.
3856 (neon_vcombine<mode>): Likewise.
3857 (float<mode><V_cvtto>2): Likewise.
3858 (floatuns<mode><V_cvtto>2): Likewise.
3859 (fix_trunc<mode><V_cvtto>2): Likewise.
3860 (fixuns_trunc<mode><V_cvtto>2
3861 (neon_vcvt<mode>): Likewise.
3862 (neon_vcvt<mode>): Likewise.
3863 (neon_vcvtv4sfv4hf): Likewise.
3864 (neon_vcvtv4hfv4sf): Likewise.
3865 (neon_vcvt_n<mode>): Likewise.
3866 (neon_vcvt_n<mode>): Likewise.
3867 (neon_vmovn<mode>): Likewise.
3868 (neon_vqmovn<mode>): Likewise.
3869 (neon_vqmovun<mode>): Likewise.
3870 (neon_vmovl<mode>): Likewise.
3871 (neon_vmul_lane<mode>): Likewise.
3872 (neon_vmul_lane<mode>): Likewise.
3873 (neon_vmull_lane<mode>): Likewise.
3874 (neon_vqdmull_lane<mode>): Likewise.
3875 (neon_vqdmulh_lane<mode>): Likewise.
3876 (neon_vqdmulh_lane<mode>): Likewise.
3877 (neon_vmla_lane<mode>): Likewise.
3878 (neon_vmla_lane<mode>): Likewise.
3879 (neon_vmlal_lane<mode>): Likewise.
3880 (neon_vqdmlal_lane<mode>): Likewise.
3881 (neon_vmls_lane<mode>): Likewise.
3882 (neon_vmls_lane<mode>): Likewise.
3883 (neon_vmlsl_lane<mode>): Likewise.
3884 (neon_vqdmlsl_lane<mode>): Likewise.
3885 (neon_vext<mode>): Likewise.
3886 (neon_vrev64<mode>): Likewise.
3887 (neon_vrev32<mode>): Likewise.
3888 (neon_vrev16<mode>): Likewise.
3889 (neon_vbsl<mode>_internal): Likewise.
3890 (neon_vshl<mode>): Likewise.
3891 (neon_vqshl<mode>): Likewise.
3892 (neon_vshr_n<mode>): Likewise.
3893 (neon_vshrn_n<mode>): Likewise.
3894 (neon_vqshrn_n<mode>): Likewise.
3895 (neon_vqshrun_n<mode>): Likewise.
3896 (neon_vshl_n<mode>): Likewise.
3897 (neon_vqshl_n<mode>): Likewise.
3898 (neon_vqshlu_n<mode>): Likewise.
3899 (neon_vshll_n<mode>): Likewise.
3900 (neon_vsra_n<mode>): Likewise.
3901 (neon_vsri_n<mode>): Likewise.
3902 (neon_vsli_n<mode>): Likewise.
3903 (neon_vtbl1v8qi): Likewise.
3904 (neon_vtbl2v8qi): Likewise.
3905 (neon_vtbl3v8qi): Likewise.
3906 (neon_vtbl4v8qi): Likewise.
3907 (neon_vtbl1v16qi): Likewise.
3908 (neon_vtbl2v16qi): Likewise.
3909 (neon_vcombinev16qi): Likewise.
3910 (neon_vtbx1v8qi): Likewise.
3911 (neon_vtbx2v8qi): Likewise.
3912 (neon_vtbx3v8qi): Likewise.
3913 (neon_vtbx4v8qi): Likewise.
3914 (*neon_vtrn<mode>_insn): Likewise.
3915 (*neon_vzip<mode>_insn): Likewise.
3916 (*neon_vuzp<mode>_insn): Likewise.
3917 (neon_vld1<mode>): Likewise.
3918 (neon_vld1_lane<mode>): Likewise.
3919 (neon_vld1_lane<mode>): Likewise.
3920 (neon_vld1_dup<mode>): Likewise.
3921 (neon_vld1_dup<mode>): Likewise.
3922 (neon_vld1_dupv2di): Likewise.
3923 (neon_vst1<mode>): Likewise.
3924 (neon_vst1_lane<mode>): Likewise.
3925 (neon_vst1_lane<mode>): Likewise.
3926 (neon_vld2<mode>): Likewise.
3927 (neon_vld2<mode>): Likewise.
3928 (neon_vld2_lane<mode>): Likewise.
3929 (neon_vld2_lane<mode>): Likewise.
3930 (neon_vld2_dup<mode>): Likewise.
3931 (neon_vst2<mode>): Likewise.
3932 (neon_vst2<mode>): Likewise.
3933 (neon_vst2_lane<mode>): Likewise.
3934 (neon_vst2_lane<mode>): Likewise.
3935 (neon_vld3<mode>): Likewise.
3936 (neon_vld3qa<mode>): Likewise.
3937 (neon_vld3qb<mode>): Likewise.
3938 (neon_vld3_lane<mode>): Likewise.
3939 (neon_vld3_lane<mode>): Likewise.
3940 (neon_vld3_dup<mode>): Likewise.
3941 (neon_vst3<mode>): Likewise.
3942 (neon_vst3qa<mode>): Likewise.
3943 (neon_vst3qb<mode>): Likewise.
3944 (neon_vst3_lane<mode>): Likewise.
3945 (neon_vst3_lane<mode>): Likewise.
3946 (neon_vld4<mode>): Likewise.
3947 (neon_vld4qa<mode>): Likewise.
3948 (neon_vld4qb<mode>): Likewise.
3949 (neon_vld4_lane<mode>): Likewise.
3950 (neon_vld4_lane<mode>): Likewise.
3951 (neon_vld4_dup<mode>): Likewise.
3952 (neon_vst4<mode>): Likewise.
3953 (neon_vst4qa<mode>): Likewise.
3954 (neon_vst4qb<mode>): Likewise.
3955 (neon_vst4_lane<mode>): Likewise.
3956 (neon_vst4_lane<mode>): Likewise.
3957 (neon_vec_unpack<US>_lo_<mode>): Likewise.
3958 (neon_vec_unpack<US>_hi_<mode>): Likewise.
3959 (neon_vec_<US>mult_lo_<mode>): Likewise.
3960 (neon_vec_<US>mult_hi_<mode>): Likewise.
3961 (neon_vec_<US>shiftl_<mode>): Likewise.
3962 (neon_unpack<US>_<mode>): Likewise.
3963 (neon_vec_<US>mult_<mode>): Likewise.
3964 (vec_pack_trunc_<mode>): Likewise.
3965 (neon_vec_pack_trunc_<mode>): Likewise.
3966 (neon_vabd<mode>_2): Likewise.
3967 (neon_vabd<mode>_3): Likewise.
3968
3969 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3970
3971 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
3972 (load_pair): Update type attribute.
3973 (store_pair): Update type attribute.
3974 * config/aarch64/iterators.md (q): New.
3975
3976 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
3977
3978 * config/arm/types.md: Add new types for Neon insns.
3979
3980 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
3981 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3982 Sergey Lega <sergey.s.lega@intel.com>
3983 Anna Tikhonova <anna.tikhonova@intel.com>
3984 Ilya Tocar <ilya.tocar@intel.com>
3985 Andrey Turetskiy <andrey.turetskiy@intel.com>
3986 Ilya Verbin <ilya.verbin@intel.com>
3987 Kirill Yukhin <kirill.yukhin@intel.com>
3988 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3989
3990 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
3991 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
3992 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
3993 (rcp14<mode>): New.
3994 (srcp14<mode>): Ditto.
3995 (rsqrt14<mode>): Ditto.
3996 (rsqrt14<mode>): Ditto.
3997 (avx512f_vmscalef<mode>): Ditto.
3998 (avx512f_scalef<mode>): Ditto.
3999 (avx512f_getexp<mode>): Ditto.
4000 (avx512f_sgetexp<mode>): Ditto.
4001 (avx512f_fixupimm<mode>): Ditto.
4002 (avx512f_sfixupimm<mode>): Ditto.
4003 (avx512f_rndscale<mode>): Ditto.
4004 (*avx512er_exp2<mode>): Ditto.
4005 (*avx512er_rcp28<mode>): Ditto.
4006 (avx512er_rsqrt28<mode>): Ditto.
4007 (avx512f_getmant<mode>): Ditto.
4008 (avx512f_getmant<mode>): Ditto.
4009 (avx512f_rndscale<mode>): Fix formatting.
4010
4011 2013-10-15 Martin Jambor <mjambor@suse.cz>
4012
4013 * ipa-utils.h (ipa_edge_within_scc): Declare.
4014 * ipa-cp.c (edge_within_scc): Moved...
4015 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
4016
4017 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4018 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4019 Sergey Lega <sergey.s.lega@intel.com>
4020 Anna Tikhonova <anna.tikhonova@intel.com>
4021 Ilya Tocar <ilya.tocar@intel.com>
4022 Andrey Turetskiy <andrey.turetskiy@intel.com>
4023 Ilya Verbin <ilya.verbin@intel.com>
4024 Kirill Yukhin <kirill.yukhin@intel.com>
4025 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4026
4027 * config/i386/predicates.md (const_8_to_15_operand): New.
4028 (const_16_to_31_operand): Ditto.
4029 * config/i386/sse.md (V8FI): New.
4030 (V16FI): Ditto.
4031 (reduc_splus_v8df): Ditto.
4032 (reduc_splus_v16sf): Ditto.
4033 (avx512f_vextract<shuffletype>32x4_1): Ditto.
4034 (vec_extract_hi_<mode>): Ditto.
4035 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
4036 (vec_set_lo_<mode>): Ditto.
4037 (vec_set_hi_<mode>): Ditto.
4038 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
4039 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
4040 (avx512f_pshufd_1): Ditto.
4041 (avx512f_broadcast<mode>): Ditto.
4042 (avx512f_broadcast<mode>): Ditto.
4043 (define_split): Split vec_extract_lo into move.
4044 (ssequartermode): Ditto.
4045 (ssedoublemode): Extened with wider modes.
4046 (vec_extract_lo_<mode>): Ditto.
4047
4048 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4049 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4050 Sergey Lega <sergey.s.lega@intel.com>
4051 Anna Tikhonova <anna.tikhonova@intel.com>
4052 Ilya Tocar <ilya.tocar@intel.com>
4053 Andrey Turetskiy <andrey.turetskiy@intel.com>
4054 Ilya Verbin <ilya.verbin@intel.com>
4055 Kirill Yukhin <kirill.yukhin@intel.com>
4056 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4057
4058 * config/i386/predicates.md (register_or_constm1_operand): New.
4059 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
4060 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
4061 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
4062 UNSPEC_SCATTER_PREFETCH
4063 (VI48_512): New.
4064 (avx512f_ucmp<mode>3): Ditto.
4065 (avx512f_vternlog<mode>): Ditto.
4066 (avx512f_align<mode>): Ditto.
4067 (<shift_insn><mode>3): Ditto.
4068 (avx512f_<rotate>v<mode>): Ditto.
4069 (avx512f_<rotate><mode>): Ditto.
4070 (avx512f_eq<mode>3): Ditto.
4071 (avx512f_eq<mode>3_1): Ditto.
4072 (avx512f_gt<mode>3): Ditto.
4073 (avx512f_testm<mode>3): Ditto.
4074 (avx512f_testnm<mode>3): Ditto.
4075 (avx512pf_gatherpf<mode>): Ditto.
4076 (*avx512pf_gatherpf<mode>_mask): Ditto.
4077 (*avx512pf_gatherpf<mode>): Ditto.
4078 (avx512pf_scatterpf<mode>): Ditto.
4079 (*avx512pf_scatterpf<mode>_mask): Ditto.
4080 (*avx512pf_scatterpf<mode>): Ditto.
4081 (avx512f_vec_dup_gpr<mode>): Ditto.
4082 (clz<mode>2): Ditto.
4083 (conflict<mode>): Ditto.
4084 (REDUC_SMINMAX_MODE): Extened with wider modes.
4085 (reduc_<code>_<mode>): Ditto.
4086 (vlshr<mode>3): Ditto.
4087 (vashl<mode>3): Ditto.
4088
4089 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
4090 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4091 Sergey Lega <sergey.s.lega@intel.com>
4092 Anna Tikhonova <anna.tikhonova@intel.com>
4093 Ilya Tocar <ilya.tocar@intel.com>
4094 Andrey Turetskiy <andrey.turetskiy@intel.com>
4095 Ilya Verbin <ilya.verbin@intel.com>
4096 Kirill Yukhin <kirill.yukhin@intel.com>
4097 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4098
4099 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
4100 UNSPEC_SCATTER.
4101 (VI48F_512): New.
4102 (avx512fmaskmode): Ditto.
4103 (bcstscalarsuff): Ditto.
4104 (avx512f_blendm<mode>): Ditto.
4105 (cmp_imm_predicate): Ditto.
4106 (avx512f_cmp<mode>3): Ditto.
4107 (avx512f_vec_dup<mode>): Ditto.
4108 (avx512f_vec_dup_mem<mode>): Ditto.
4109 (avx512f_vpermi2var<mode>3): Ditto.
4110 (avx512f_vpermt2var<mode>3): Ditto.
4111 (vec_init<mode>): Ditto.
4112 (avx512f_gathersi<mode>): Ditto.
4113 (*avx512f_gathersi<mode>): Ditto.
4114 (*avx512f_gathersi<mode>_2): Ditto.
4115 (avx512f_gatherdi<mode>): Ditto.
4116 (*avx512f_gatherdi<mode>): Ditto.
4117 (*avx512f_gatherdi<mode>_2): Ditto.
4118 (avx512f_scattersi<mode>): Ditto.
4119 (*avx512f_scattersi<mode>): Ditto.
4120 (avx512f_scatterdi<mode>): Ditto.
4121 (*avx512f_scatterdi<mode>): Ditto.
4122 (sseintprefix): Extened with wider modes.
4123 (VEC_GATHER_IDXSI): Ditto.
4124 (VEC_GATHER_IDXDI): Ditto.
4125 (VEC_GATHER_SRCDI): Ditto.
4126
4127 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
4128 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
4129
4130 * config/arm/t-aprofile: New file.
4131 * config.gcc: Handle --with-multilib-list option.
4132
4133 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
4134
4135 * reload1.c (reloads_unique_chain_p): Ensure that r1 is
4136 the input for r2.
4137
4138 2013-10-15 Richard Biener <rguenther@suse.de>
4139
4140 * tree-loop-distribution.c (build_empty_rdg): Inline into
4141 single user.
4142 (rdg_flag_vertex): Inline into single user.
4143 (rdg_flag_vertex_and_dependent): Likewise.
4144 (build_rdg_partition_for_vertex): Remove processed bitmap.
4145 (rdg_build_partitions): Simplify.
4146
4147 2013-10-15 Richard Biener <rguenther@suse.de>
4148
4149 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
4150 Restructure forwarding through conversions and copies to
4151 avoid performing copy-propagation the wrong way. Adjust
4152 recursion invocations.
4153 (forward_propagate_addr_expr): Add argument stating if we
4154 are recursing from a single-use.
4155 (ssa_forward_propagate_and_combine): Adjust.
4156
4157 2013-10-14 David Malcolm <dmalcolm@redhat.com>
4158
4159 * dumpfile.h (gcc::dump_manager): New class, to hold state
4160 relating to dumpfile management.
4161 (get_dump_file_name): Remove in favor of method of dump_manager.
4162 (dump_initialized_p): Likewise.
4163 (dump_start): Likewise.
4164 (dump_finish): Likewise.
4165 (dump_switch_p): Likewise.
4166 (dump_register): Likewise.
4167 (get_dump_file_info): Likewise.
4168 * context.c (gcc::context::context): Construct the dump_manager
4169 instance.
4170 * context.h (gcc::context::get_dumps): New.
4171 (gcc::context::m_dumps): New.
4172 * coverage.c (coverage_init): Port to dump_manager API.
4173 * dumpfile.c (extra_dump_files): Convert to field of gcc::dump_manager.
4174 (extra_dump_files_in_use): Likewise.
4175 (extra_dump_files_alloced): Likewise.
4176 (gcc::dump_manager::dump_manager): New.
4177 (dump_register): Convert to...
4178 (gcc::dump_manager::dump_register): ...method, replacing
4179 function-static next_dump with m_next_dump field.
4180 (get_dump_file_info): Convert to...
4181 (gcc::dump_manager::get_dump_file_info): ...method.
4182 (get_dump_file_name): Convert to...
4183 (gcc::dump_manager::get_dump_file_name): ...method.
4184 (dump_start): Convert to...
4185 (gcc::dump_manager::dump_start): ...method.
4186 (dump_finish): Convert to...
4187 (gcc::dump_manager::dump_finish): ...method.
4188 (dump_begin): Replace body with...
4189 (gcc::dump_manager::dump_begin): ...new method.
4190 (dump_phase_enabled_p): Convert to...
4191 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4192 (dump_phase_enabled_p): Convert to...
4193 (gcc::dump_manager::dump_phase_enabled_p): ...method.
4194 (dump_initialized_p): Convert to...
4195 (gcc::dump_manager::dump_initialized_p): ...method.
4196 (dump_flag_name): Replace body with...
4197 (gcc::dump_manager::dump_flag_name): ...new method.
4198 (dump_enable_all): Convert to...
4199 (gcc::dump_manager::dump_enable_all): ...new method.
4200 (opt_info_enable_passes): Convert to...
4201 (gcc::dump_manager::opt_info_enable_passes): ...new method.
4202 (dump_switch_p_1): Convert to...
4203 (gcc::dump_manager::dump_switch_p_1): ...new method.
4204 (dump_switch_p): Convert to...
4205 (gcc::dump_manager::dump_switch_p): ...new method.
4206 (opt_info_switch_p): Port to dump_manager API.
4207 (enable_rtl_dump_file): Likewise.
4208 * opts-global.c (handle_common_deferred_options): Port to new
4209 dump_manager API.
4210 * passes.c (pass_manager::finish_optimization_passes): Likewise.
4211 (pass_manager::register_one_dump_file): Likewise.
4212 (pass_manager::register_pass): Likewise.
4213 (pass_init_dump_file): Likewise.
4214 (pass_fini_dump_file): Likewise.
4215 * statistics.c (statistics_early_init): Likewise.
4216
4217 2013-10-14 Richard Biener <rguenther@suse.de>
4218
4219 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
4220 iterative_hash_canonical_type, gimple_canonical_type_hash,
4221 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
4222 gimple_register_canonical_type, print_gimple_types_stats,
4223 free_gimple_type_tables): Move to lto/lto.c
4224 (gt-gimple.h): Do not include.
4225 * gimple.h (gimple_register_canonical_type,
4226 print_gimple_types_stats, free_gimple_type_tables): Remove.
4227 * Makefile.in (GTFILES): Remove gimple.c.
4228
4229 2013-10-14 Travis Snoozy <quandary@remstate.com>
4230
4231 PR target/58716
4232 * config/msp430/msp430.c (msp430_option_override): Correct thinko
4233 scanning for msp430x targets.
4234
4235 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
4236
4237 PR bootstrap/58509
4238 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
4239 (registers_ok_for_ldd_peep): Move around.
4240 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
4241 * config/sparc/sparc.md (widening peepholes): Use it.
4242
4243 2013-10-14 Richard Biener <rguenther@suse.de>
4244
4245 PR middle-end/58712
4246 PR middle-end/55358
4247 * gimple.c (iterative_hash_canonical_type): Make sure to
4248 record the hash into the correct hashtable slot.
4249
4250 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
4251
4252 PR rtl-optimization/58662
4253 * combine.c (try_combine): Take into account death nodes on I2 when
4254 splitting a PARALLEL of two independent SETs. Fix dump message.
4255
4256 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
4257
4258 PR target/51244
4259 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
4260 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
4261 * config/sh/t-sh (sh_treg_combine.o): New entry.
4262 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
4263 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
4264 (register_sh_passes): New function. Register sh_treg_combine pass.
4265 (sh_option_override): Invoke it.
4266 (sh_canonicalize_comparison): Handle op0_preserve_value.
4267 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
4268 opportunities. Canonicalize branch condition.
4269 (nott): Allow only if pseudos can be created for non-SH2A.
4270
4271 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
4272
4273 PR target/58690
4274 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
4275 (ix86_expand_movmem): Replace copy_addr_to_reg with
4276 ix86_copy_addr_to_reg.
4277 (ix86_expand_setmem): Likewise.
4278
4279 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
4280
4281 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
4282 provided by ix86_fp_compare_mode instead of CCFPUmode.
4283
4284 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
4285
4286 * config/aarch64/arm_neon.h
4287 (vtbx<1,3>_<psu>8): Fix register constriants.
4288
4289 2013-10-11 Jeff Law <law@redhat.com>
4290
4291 PR tree-optimization/58640
4292 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
4293 threading paths that cross over two loop entry points.
4294
4295 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4296
4297 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
4298 handle vector float as well.
4299 (*vsx_le_perm_load_v4si): Likewise.
4300 (*vsx_le_perm_store_v2di): Likewise.
4301 (*vsx_le_perm_store_v4si): Likewise.
4302
4303 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
4304
4305 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
4306 directly to circumvent subtract from splat{31} workaround.
4307 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
4308 prototype.
4309 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
4310 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
4311 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
4312 (altivec_vperm_<mode>): Convert to define_insn_and_split to
4313 separate big and little endian logic.
4314 (*altivec_vperm_<mode>_internal): New define_insn.
4315 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
4316 separate big and little endian logic.
4317 (*altivec_vperm_<mode>_uns_internal): New define_insn.
4318 (vec_permv16qi): Add little endian logic.
4319
4320 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
4321
4322 * doc/extend.texi (returns_nonnull): Remove arguments.
4323
4324 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4325 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4326 Sergey Lega <sergey.s.lega@intel.com>
4327 Anna Tikhonova <anna.tikhonova@intel.com>
4328 Ilya Tocar <ilya.tocar@intel.com>
4329 Andrey Turetskiy <andrey.turetskiy@intel.com>
4330 Ilya Verbin <ilya.verbin@intel.com>
4331 Kirill Yukhin <kirill.yukhin@intel.com>
4332 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4333
4334 * config/i386/sse.md (VI48F_256_512): New.
4335 (avx2_permvar<mode>): Change to ...
4336 (<avx2_avx512f>_permvar<mode>): This.
4337
4338 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4339 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4340 Sergey Lega <sergey.s.lega@intel.com>
4341 Anna Tikhonova <anna.tikhonova@intel.com>
4342 Ilya Tocar <ilya.tocar@intel.com>
4343 Andrey Turetskiy <andrey.turetskiy@intel.com>
4344 Ilya Verbin <ilya.verbin@intel.com>
4345 Kirill Yukhin <kirill.yukhin@intel.com>
4346 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4347
4348 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
4349 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
4350 * config/i386/sse.md (VI4_AVX): New.
4351 (sf2simodelower): Ditto.
4352 (sse2_cvtps2dq): Change to ...
4353 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
4354
4355 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4356 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4357 Sergey Lega <sergey.s.lega@intel.com>
4358 Anna Tikhonova <anna.tikhonova@intel.com>
4359 Ilya Tocar <ilya.tocar@intel.com>
4360 Andrey Turetskiy <andrey.turetskiy@intel.com>
4361 Ilya Verbin <ilya.verbin@intel.com>
4362 Kirill Yukhin <kirill.yukhin@intel.com>
4363 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4364
4365 * config/i386/sse.md (V_512): New.
4366 (VI_512): Ditto.
4367 (vcond<V_512:mode><VF_512:mode>): Ditto.
4368 (vcond<V_512:mode><VI_512:mode>): Ditto.
4369 (vcondu<V_512:mode><VI_512:mode>): Ditto.
4370
4371 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4372 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4373 Sergey Lega <sergey.s.lega@intel.com>
4374 Anna Tikhonova <anna.tikhonova@intel.com>
4375 Ilya Tocar <ilya.tocar@intel.com>
4376 Andrey Turetskiy <andrey.turetskiy@intel.com>
4377 Ilya Verbin <ilya.verbin@intel.com>
4378 Kirill Yukhin <kirill.yukhin@intel.com>
4379 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4380
4381 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
4382 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
4383 (FMAMODE): Ditto.
4384 (fma<mode>4): Removed condition.
4385 (fms<mode>4): Ditto.
4386 (fnma<mode>4): Ditto.
4387 (fnms<mode>4): Ditto.
4388 (fma4i_fmadd_<mode>): Ditto.
4389 (*fma_fmadd_<mode>): Ditto.
4390 (*fma_fmsub_<mode>): Ditto.
4391 (*fma_fnmadd_<mode>): Ditto.
4392 (*fma_fnmsub_<mode>): Ditto.
4393 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
4394 (*fma_fmaddsub_<mode>): Ditto.
4395 (*fma_fmsubadd_<mode>): Ditto.
4396 (*fmai_fmadd_<mode>): Ditto.
4397 (*fmai_fmsub_<mode>): Ditto.
4398 (*fmai_fnmadd_<mode>): Ditto.
4399 (*fmai_fnmsub_<mode>): Ditto.
4400
4401 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4402 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4403 Sergey Lega <sergey.s.lega@intel.com>
4404 Anna Tikhonova <anna.tikhonova@intel.com>
4405 Ilya Tocar <ilya.tocar@intel.com>
4406 Andrey Turetskiy <andrey.turetskiy@intel.com>
4407 Ilya Verbin <ilya.verbin@intel.com>
4408 Kirill Yukhin <kirill.yukhin@intel.com>
4409 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4410
4411 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
4412 (VI124_256): Changed to ...
4413 (VI124_256_48_512): This.
4414 (ssepackmode): Extended with wider modes.
4415 (<code><mode>3): Changed iterator.
4416 (*avx2_<code><mode>3): Ditto.
4417 (vec_pack_trunc_<mode>): Ditto.
4418
4419 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4420 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4421 Sergey Lega <sergey.s.lega@intel.com>
4422 Anna Tikhonova <anna.tikhonova@intel.com>
4423 Ilya Tocar <ilya.tocar@intel.com>
4424 Andrey Turetskiy <andrey.turetskiy@intel.com>
4425 Ilya Verbin <ilya.verbin@intel.com>
4426 Kirill Yukhin <kirill.yukhin@intel.com>
4427 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4428
4429 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
4430 (VI8F_256_512): Ditto.
4431 (abs<mode>2): Changed iterator.
4432 (avx2_perm<mode>): Changed to ...
4433 (<avx2_avx512f>_perm<mode>): This.
4434 (avx2_perm<mode>_1): Changed to ...
4435 (<avx2_avx512f>_perm<mode>_1): This.
4436
4437 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4438 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4439 Sergey Lega <sergey.s.lega@intel.com>
4440 Anna Tikhonova <anna.tikhonova@intel.com>
4441 Ilya Tocar <ilya.tocar@intel.com>
4442 Andrey Turetskiy <andrey.turetskiy@intel.com>
4443 Ilya Verbin <ilya.verbin@intel.com>
4444 Kirill Yukhin <kirill.yukhin@intel.com>
4445 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4446
4447 * config/i386/sse.md (VI48_AVX512F): New.
4448 (VI48_AVX2): Changed to ...
4449 (VI48_AVX2_48_AVX512F): This.
4450 (avx2_ashrv<mode>): Changed to ...
4451 (<avx2_avx512f>_ashrv<mode>): This.
4452 (avx2_<shift_insn>v<mode>): Changed to ...
4453 (<avx2_avx512f>_<shift_insn>v<mode>): This.
4454
4455 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4456 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4457 Sergey Lega <sergey.s.lega@intel.com>
4458 Anna Tikhonova <anna.tikhonova@intel.com>
4459 Ilya Tocar <ilya.tocar@intel.com>
4460 Andrey Turetskiy <andrey.turetskiy@intel.com>
4461 Ilya Verbin <ilya.verbin@intel.com>
4462 Kirill Yukhin <kirill.yukhin@intel.com>
4463 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4464
4465 * config/i386/sse.md (VI4_AVX512F): New.
4466 (VI8_AVX2_AVX512F): Ditto.
4467 (mul<mode>3): Extended with wider modes.
4468 (*<sse4_1_avx2>_mul<mode>3): Ditto.
4469 (mul<mode>3): Ditto.
4470 (vec_widen_<s>mult_odd_<mode>): Ditto.
4471
4472 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4473 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4474 Sergey Lega <sergey.s.lega@intel.com>
4475 Anna Tikhonova <anna.tikhonova@intel.com>
4476 Ilya Tocar <ilya.tocar@intel.com>
4477 Andrey Turetskiy <andrey.turetskiy@intel.com>
4478 Ilya Verbin <ilya.verbin@intel.com>
4479 Kirill Yukhin <kirill.yukhin@intel.com>
4480 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4481
4482 * config/i386/sse.md (VI2_AVX512F): New.
4483 (VI124_AVX512F): Ditto.
4484 (sseunpackmode): Extended with wider modes.
4485 (sseunpackfltmode): Ditto.
4486 (vec_unpacks_float_hi_<mode>): Ditto.
4487 (vec_unpacks_float_lo_<mode>): Ditto.
4488 (vec_unpacku_float_hi_<mode>): Ditto.
4489 (vec_unpacku_float_lo_<mode>): Ditto.
4490 (vec_unpacks_lo_<mode>): Ditto.
4491 (vec_unpacks_hi_<mode>): Ditto.
4492 (vec_unpacku_lo_<mode>): Ditto.
4493 (vec_unpacku_hi_<mode>): Ditto.
4494
4495 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4496 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4497 Sergey Lega <sergey.s.lega@intel.com>
4498 Anna Tikhonova <anna.tikhonova@intel.com>
4499 Ilya Tocar <ilya.tocar@intel.com>
4500 Andrey Turetskiy <andrey.turetskiy@intel.com>
4501 Ilya Verbin <ilya.verbin@intel.com>
4502 Kirill Yukhin <kirill.yukhin@intel.com>
4503 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4504
4505 * config/i386/i386.md (multdiv): New.
4506 (multdiv_mnemonic): Ditto.
4507 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
4508 (<sse>_vm<multdiv_mnemonic><mode>3): This.
4509 (<sse>_vmdiv<mode>3): Removed.
4510
4511 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4512 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4513 Sergey Lega <sergey.s.lega@intel.com>
4514 Anna Tikhonova <anna.tikhonova@intel.com>
4515 Ilya Tocar <ilya.tocar@intel.com>
4516 Andrey Turetskiy <andrey.turetskiy@intel.com>
4517 Ilya Verbin <ilya.verbin@intel.com>
4518 Kirill Yukhin <kirill.yukhin@intel.com>
4519 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4520
4521 * config/i386/sse.md (V): Extended with wider modes.
4522 (VF2): Ditto.
4523 (ssehalfvecmode): Ditto.
4524 (i128): Ditto.
4525 (ssepackfltmode): Ditto.
4526 (avx_vec_concat<mode>): Ditto.
4527 (V_256_512): New iterator.
4528 (VF2_512_256): Ditto.
4529 (si2dfmode): New attribute.
4530 (si2dfmodelower): Ditto.
4531 (sf2dfmode): Ditto.
4532 (concat_tg_mode): Ditto.
4533 (floatv4siv4df2): Changed to ...
4534 (float<si2dfmodelower><mode>2): This.
4535 (avx_cvtps2pd256): Changed to ...
4536 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
4537 (vec_pack_trunc_v4df): Changed to ...
4538 (vec_pack_trunc_<mode>): This.
4539 (avx_vpermil<mode>): Changed to ...
4540 (<sse2_avx_avx512f>_vpermil<mode>): This.
4541 (<fixsuffix>fix_truncv8dfv8si2): New.
4542 (vec_pack_sfix_trunc_v8df): Ditto.
4543 (avx512f_rndscale<mode>): Ditto.
4544 (avx512f_roundpd512): Ditto.
4545 (vec_pack_ufix_trunc_<mode>): Updated iterator.
4546
4547 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4548 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4549 Sergey Lega <sergey.s.lega@intel.com>
4550 Anna Tikhonova <anna.tikhonova@intel.com>
4551 Ilya Tocar <ilya.tocar@intel.com>
4552 Andrey Turetskiy <andrey.turetskiy@intel.com>
4553 Ilya Verbin <ilya.verbin@intel.com>
4554 Kirill Yukhin <kirill.yukhin@intel.com>
4555 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4556
4557 * config/i386/i386.md (any_fix): New iterator.
4558 (fixsuffix): New attribute.
4559 * config/i386/sse.md (VF1): Extened with wider modes.
4560 (VI): Ditto.
4561 (VI_AVX2): Ditto.
4562 (VI8): Ditto.
4563 (sseintvecmodelower): Ditto.
4564 (ssescalarmode): Ditto.
4565 (ssescalarnum): Ditto.
4566 (VF1_128_256): New.
4567 (ssexmmmode): Ditto.
4568 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
4569 (<sse>_rcp<mode>2): Change iterator.
4570 (rsqrt<mode>2): Ditto.
4571 (<sse>_rsqrt<mode>2): Ditto.
4572 (avx2_vec_dup<mode>): Ditto.
4573 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
4574 (round<mode>2_sfix): Ditto.
4575 (avx2_pbroadcast<mode>): Ditto.
4576 (*andnot<mode>3): Handle XI mode.
4577 (*<code><mode>3): Ditto.
4578 (AVXTOSSEMODE): Removed.
4579 (avx_vpermil<mode>): Changed to ...
4580 (<sse2_avx_avx512f>_vpermil<mode>): This.
4581
4582 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
4583 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
4584 Sergey Lega <sergey.s.lega@intel.com>
4585 Anna Tikhonova <anna.tikhonova@intel.com>
4586 Ilya Tocar <ilya.tocar@intel.com>
4587 Andrey Turetskiy <andrey.turetskiy@intel.com>
4588 Ilya Verbin <ilya.verbin@intel.com>
4589 Kirill Yukhin <kirill.yukhin@intel.com>
4590 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
4591
4592 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
4593 (<sse>_comi): Ditto.
4594 (<sse>_ucomi): Ditto.
4595 (sse_cvtss2siq_2): Ditto.
4596 (sse2_cvtsd2si): Ditto.
4597 (sse2_cvtsd2siq): Ditto.
4598 (sse2_cvttsd2si): Ditto.
4599 (sse2_cvttsd2siq): Ditto.
4600 (<shift_insn><mode>3): Ditto.
4601 (sse2_cvtsi2sdq): Update constraint and prefix.
4602 (sse_cvtsi2ss): Update prefix.
4603 (sse_cvtsi2ssq): Ditto.
4604
4605 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4606
4607 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
4608 ignore internal calls.
4609
4610 2013-10-11 Richard Biener <rguenther@suse.de>
4611
4612 * tree-pretty-print.c (dump_generic_node): Allow to dump both (D)
4613 and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with (OVF) if
4614 TREE_OVERFLOW is set.
4615
4616 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
4617
4618 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
4619
4620 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
4621 * gimple.h: Likewise.
4622
4623 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
4624
4625 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
4626 macros.
4627
4628 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
4629
4630 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
4631 * configure: Regenerate.
4632
4633 2013-10-11 Jakub Jelinek <jakub@redhat.com>
4634
4635 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
4636 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
4637 formatting fixes, use pp_colon instead of pp_character (..., ':'),
4638 similarly pp_right_paren.
4639 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
4640 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
4641 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
4642 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
4643 clauses.
4644 (omp_declare_simd_clauses_equal,
4645 omp_remove_redundant_declare_simd_attrs): New functions.
4646 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
4647 (walk_tree_1): Handle new OpenMP 4.0 clauses.
4648 * tree.h (OMP_LOOP_CHECK): Define.
4649 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
4650 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
4651 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
4652 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
4653 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
4654 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
4655 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
4656 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
4657 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
4658 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
4659 OMP_CLAUSE_SIMDLEN_EXPR): Define.
4660 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
4661 (omp_remove_redundant_declare_simd_attrs): New prototype.
4662 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
4663 GIMPLE_OMP_TEAMS): New codes.
4664 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
4665 * omp-low.c (struct omp_context): Add cancel_label and cancellable
4666 fields.
4667 (target_nesting_level): New variable.
4668 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
4669 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
4670 for collapse > 1 static schedule unless ordered.
4671 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
4672 (determine_parallel_type): Adjust get_ws_args_for caller.
4673 (install_var_field): Handle mask & 4 for double indirection.
4674 (scan_sharing_clauses): Ignore shared clause on teams construct.
4675 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
4676 (create_omp_child_function): If inside target or declare target
4677 constructs, set "omp declare target" attribute on the child function.
4678 (find_combined_for): New function.
4679 (scan_omp_parallel): Handle combined loops.
4680 (scan_omp_target, scan_omp_teams): New functions.
4681 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
4682 restrictions and set ctx->cancellable for cancellable constructs.
4683 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
4684 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
4685 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
4686 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
4687 (omp_clause_aligned_alignment): New function.
4688 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR on decls.
4689 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
4690 on teams constructs. Handle user defined reductions and new
4691 OpenMP 4.0 clauses.
4692 (lower_reduction_clauses): Don't set placeholder to address of ref
4693 if it has already the right type.
4694 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
4695 (expand_parallel_call): Use the new non-_start suffixed builtins,
4696 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
4697 and GOMP_parallel_end after the call.
4698 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
4699 (expand_omp_for_init_counts): Handle combined loops.
4700 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
4701 loops.
4702 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
4703 end of cancellable loops.
4704 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
4705 Likewise. Handle collapse > 1 loops.
4706 (expand_omp_simd): Handle combined loops.
4707 (expand_omp_for): Add inner_stmt argument, adjust callers of
4708 expand_omp_for* functions, use expand_omp_for_static*chunk even
4709 for collapse > 1 unless ordered.
4710 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
4711 of cancellable sections.
4712 (expand_omp_single): Remove need_barrier variable, just rely on
4713 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
4714 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
4715 (expand_omp_atomic_load, expand_omp_atomic_store,
4716 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
4717 (expand_omp_target): New function.
4718 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
4719 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
4720 (build_omp_regions_1): Immediately close region for
4721 GF_OMP_TARGET_KIND_UPDATE.
4722 (maybe_add_implicit_barrier_cancel): New function.
4723 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
4724 cancellation.
4725 (lower_omp_single): Likewise. Add clobber after the barrier.
4726 (lower_omp_taskgroup): New function.
4727 (lower_omp_for): Handle combined loops. Adjust
4728 lower_rec_input_clauses caller. Handle cancellation.
4729 (lower_depend_clauses): New function.
4730 (lower_omp_taskreg): Lower depend clauses. Adjust
4731 lower_rec_input_clauses caller. Add clobber after the call. Handle
4732 cancellation.
4733 (lower_omp_target, lower_omp_teams): New functions.
4734 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
4735 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
4736 and GOMP_cancellation_point calls.
4737 (lower_omp): Fold stmts inside of target region.
4738 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
4739 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4740 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
4741 (BT_FN_VOID_OMPFN_PTR_UINT,
4742 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
4743 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
4744 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
4745 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
4746 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
4747 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
4748 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
4749 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
4750 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
4751 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
4752 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
4753 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
4754 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
4755 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
4756 BUILT_IN_GOMP_PARALLEL_END.
4757 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
4758 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
4759 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
4760 GF_OMP_FOR_KIND_DISTRIBUTE.
4761 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
4762 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
4763 (dump_gimple_omp_return): Print lhs if it has any.
4764 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
4765 gimple_omp_atomic_seq_cst_p.
4766 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
4767 and GIMPLE_OMP_TEAMS.
4768 * langhooks.c (lhd_omp_mappable_type): New function.
4769 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
4770 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
4771 hook.
4772 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
4773 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
4774 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and ORT_TARGET.
4775 (struct gimplify_omp_ctx): Add combined_loop field.
4776 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
4777 on stmts inside of target region.
4778 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
4779 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
4780 ORT_TARGET and ORT_TARGET_DATA.
4781 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
4782 Handle GOVD_MAP.
4783 (omp_notice_threadprivate_variable): Complain about threadprivate
4784 variables in target region.
4785 (omp_notice_variable): Complain about vars with non-mappable type
4786 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
4787 (omp_check_private): Ignore ORT_TARGET* regions.
4788 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
4789 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
4790 (find_combined_omp_for): New function.
4791 (gimplify_omp_for): Handle gimplification of combined loops.
4792 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
4793 OMP_TEAMS.
4794 (gimplify_omp_target_update): New function.
4795 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
4796 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
4797 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
4798 (gimplify_body): If fndecl has "omp declare target" attribute, add
4799 implicit ORT_TARGET context around it.
4800 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
4801 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
4802 * tree-nested.c (convert_nonlocal_reference_stmt,
4803 convert_local_reference_stmt, convert_gimple_call): Handle
4804 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4805 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
4806 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
4807 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
4808 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
4809 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
4810 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
4811 BUILT_IN_GOMP_LOOP_END_CANCEL,
4812 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
4813 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
4814 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
4815 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
4816 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
4817 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
4818 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
4819 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
4820 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
4821 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
4822 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
4823 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
4824 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
4825 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
4826 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
4827 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4828 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
4829 gimple_build_omp_teams): New functions.
4830 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
4831 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
4832 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
4833 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4834 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
4835 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
4836 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
4837 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
4838 GF_OMP_ATOMIC_SEQ_CST): New.
4839 (gimple_build_omp_taskgroup, gimple_build_omp_target,
4840 gimple_build_omp_teams): New prototypes.
4841 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
4842 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4843 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
4844 GIMPLE_OMP_SINGLE as end of range.
4845 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
4846 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
4847 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
4848 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
4849 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
4850 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
4851 gimple_omp_target_kind, gimple_omp_target_set_kind,
4852 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
4853 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
4854 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
4855 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
4856 gimple_omp_teams_set_clauses): New inlines.
4857 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
4858 and GIMPLE_OMP_TASKGROUP.
4859 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
4860 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
4861 enum omp_clause_proc_bind_kind): New.
4862 (union omp_clause_subcode): Add depend_kind, map_kind and
4863 proc_bind_kind fields.
4864 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
4865 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
4866 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
4867 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
4868 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
4869
4870 2013-10-10 Teresa Johnson <tejohnson@google.com>
4871
4872 * predict.c (tree_estimate_probability): Add new parameter
4873 for estimate_bb_frequencies.
4874 (estimate_bb_frequencies): Add new parameter to force estimation.
4875 (rebuild_frequencies): When max frequency in function is small,
4876 recompute counts from frequencies.
4877 * predict.h (estimate_bb_frequencies): New parameter.
4878
4879 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4880
4881 * ipa-inline.c (ipa_inline): Fix leak of "order" when
4882 optimizations are disabled.
4883
4884 2013-10-10 David Malcolm <dmalcolm@redhat.com>
4885
4886 * coverage.c (coverage_finish): Fix leak of da_file_name.
4887
4888 2013-10-10 Jan Hubicka <jh@suse.cz>
4889
4890 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
4891 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
4892 Bobcat and generic.
4893
4894 2013-10-10 Jakub Jelinek <jakub@redhat.com>
4895
4896 PR middle-end/58670
4897 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
4898 if any labels are in FALLTHRU_BB, use a special label emitted
4899 immediately after the asm goto insn rather than label_rtx
4900 of the LABEL_DECL.
4901 (expand_asm_stmt): Adjust caller.
4902 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
4903 edge if the last insn in predecessor is a jump with single successor,
4904 but it isn't simplejump_p.
4905
4906 2013-10-10 Richard Biener <rguenther@suse.de>
4907
4908 PR tree-optimization/58656
4909 * tree-ssa-pre.c (phi_translate): Do not cache failed translations.
4910
4911 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4912
4913 * gimplify.c: Include expr.h and tm_p.h for targets with special
4914 va-arg padding requirements.
4915
4916 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4917
4918 * tree-flow.h: Move some prototypes to gimple.h.
4919 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
4920 * gimple.h: Relocate some prototypes from tree-flow.h
4921 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
4922 Move to gimplify.c.
4923 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
4924 (build_va_arg_indirect_ref): Relocate and make static.
4925 (std_gimplify_va_arg_expr): Relocate here.
4926 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
4927 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
4928
4929 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4930
4931 * doc/md.texi: Document the mnemonic attribute.
4932
4933 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
4934
4935 PR target/57377
4936 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
4937 (set_attr_alternative x ...) when searching for user defined
4938 mnemonic attribute.
4939
4940 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
4941
4942 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
4943
4944 2013-10-09 Easwaran Raman <eraman@google.com>
4945
4946 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
4947 * cfgexpand.c (defer_stack_allocation): ...use here
4948 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
4949
4950 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
4951
4952 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
4953 (operand_equal_for_value_replacement): New function, extracted from
4954 value_replacement and enhanced to catch more cases.
4955 (value_replacement): Use operand_equal_for_value_replacement.
4956
4957 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4958
4959 * loop-doloop.c (doloop_modify, doloop_optimize): Use
4960 get_max_loop_iterations.
4961
4962 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4963
4964 * config/arm/aarch-common.c (arm_early_load_addr_dep):
4965 Place comment above function.
4966
4967 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
4968
4969 * tree-flow.h: Remove all remaining prototypes, enums and structs that
4970 are not related to tree-cfg.c.
4971 * tree-ssa-address.h: New file. Relocate prototypes.
4972 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
4973 (addr_for_mem_ref): New. Combine call to get_address_description and
4974 return addr_for_mem_ref.
4975 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
4976 * tree-ssa-live.h: Adjust prototypes.
4977 * passes.c: Include tree-ssa-live.h.
4978 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
4979 * graphite.c (graphite_transform_loops): Make static.
4980 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
4981 make_pass_graphite, pass_data_graphite_transforms,
4982 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
4983 * ipa-pure-const.c (warn_function_noreturn): Make static.
4984 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4985 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4986 Relocate from tree-cfg.c
4987 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
4988 static.
4989 (execute_warn_function_noreturn, gate_warn_function_noreturn,
4990 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
4991 Move to ipa-pure-const.c.
4992 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
4993 Relocate from tree-optimize.c.
4994 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
4995 make_pass_fixup_cfg): Move to tree-cfg.c.
4996 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
4997 Relocate some prototypes.
4998 * tree-data-ref.h (tree_check_data_deps) Add prototype.
4999 * tree-dump.c (dump_function_to_file): Remove prototype.
5000 Add tree-flow.h to the include file.
5001 * tree-dump.h: Remove prototype.
5002 * tree-parloops.h: New File. Add prototypes.
5003 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
5004 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
5005 from tree-ssa-loop.c.
5006 * tree-predcom.c (run_tree_predictive_commoning,
5007 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
5008 Relocate here from tree-ssa-loop.c.
5009 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
5010 ssa_name_values.release ().
5011 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
5012 (ssa_name_values): Relocate from tree-flow.h.
5013 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
5014 * tree-ssa-loop.c (run_tree_predictive_commoning,
5015 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
5016 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
5017 make_pass_graphite, pass_data_graphite_transforms,
5018 make_pass_graphite_transforms, gate_tree_parallelize_loops,
5019 tree_parallelize_loops, pass_data_parallelize_loops,
5020 make_pass_parallelize_loops): Move to other files.
5021 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
5022 moved here.
5023 * tree.h: Remove prototypes from tree-address.c.
5024
5025 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5026
5027 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
5028 (struct int_tree_map): Move to tree-hasher.h
5029 (SCALE, LABEL, PERCENT): Move to gimple.h
5030 * tree-flow-inline.h: Delete. Move functions to other files.
5031 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
5032 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
5033 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
5034 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
5035 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
5036 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
5037 * tree-hasher.h: Don't include tree-flow.h.
5038 (struct int_tree_map): Relocate from tree-flow.h.
5039 * tree-sra.c (contains_view_convert_expr_p): Relocate from
5040 tree-flow-inline.h and make static.
5041 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
5042 tree-flow-inline.h.
5043 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
5044 tree-flow-inline.h and make static.
5045 * tree.h (is_global_var, may_be_aliased): Relocate from
5046 tree-flow-inline.h.
5047 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
5048 * value-prof.c: No longer include tree-flow-inline.h.
5049 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
5050
5051 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
5052
5053 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
5054 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
5055 (enum move_pos): Move to tree-ssa-loop-im.h
5056 * cfgloop.h: Move some prototypes.
5057 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
5058 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
5059 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
5060 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
5061 (loop_containing_stmt): Relocate from tree-flow-inline.h.
5062 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
5063 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
5064 (enum move_pos): Relocate here.
5065 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
5066 tree-ssa-loop.c.
5067 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
5068 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
5069 make_pass_lim): Relocate here from tree-ssa-loop.c.
5070 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
5071 tree-ssa-loop.c.
5072 (loop_edge_to_cancel, unloop_loops): Make static.
5073 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5074 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
5075 (tree_complete_unroll, gate_tree_complete_unroll,
5076 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
5077 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5078 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
5079 * tree-ssa-loop-ivopts.c: Remove local prototypes.
5080 (stmt_invariant_in_loop_p): Remove unused function.
5081 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
5082 * tree-ssa-loop-manip.h: New file. Add prototypes.
5083 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
5084 (gcov_type_to_double_int): Move to cfgloop.h.
5085 (double_int_cmp, bound_index,
5086 estimate_numbers_of_iterations_loop): Make static.
5087 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
5088 (max_loop_iterations): Factor out get_max_loop_iterations.
5089 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
5090 cfgloop.c.
5091 * tree-ssa-loop-niter.h: New file. Add prototypes.
5092 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
5093 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
5094 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
5095 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
5096 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
5097 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
5098 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
5099 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
5100 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
5101 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
5102 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
5103 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
5104 pass_data_complete_unroll, make_pass_complete_unroll,
5105 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
5106 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
5107 tree-ssa-loop-ivcanon.c.
5108 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
5109 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
5110 tree-ssa-loop-prefetch.c.
5111 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
5112 tree-ssa-loop-im.c.
5113 (get_lsm_tmp_name): Relocate and add suffix parameter.
5114 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
5115 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
5116 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
5117 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
5118 (get_estimated_loop_iterations): Factor out accessor from
5119 estimated_loop_iterations in tree-ssa-loop-niter.c.
5120 (get_max_loop_iterations): Factor out accessor from
5121 _max_loop_iterations in tree-ssa-niter.c.
5122 * loop-unroll.c (decide_unroll_constant_iterations,
5123 decide_unroll_runtime_iterations, decide_peel_simple,
5124 decide_unroll_stupid): Use new get_* accessors.
5125
5126 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
5127
5128 PR tree-optimization/20318
5129 * doc/extend.texi (returns_nonnull): New function attribute.
5130 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
5131 attribute.
5132 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
5133 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
5134
5135 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
5136
5137 PR middle-end/58570
5138 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
5139 false if both components are bitfields.
5140
5141 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5142
5143 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
5144 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
5145 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
5146 * config/aarch64/aarch64.h
5147 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
5148 * config/aarch64/aarch64-simd-builtins.def
5149 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
5150
5151 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5152
5153 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
5154
5155 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5156
5157 * config/aarch64/arm_neon.h (vdiv_f64): Added.
5158
5159 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
5160
5161 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
5162 (vneg_f64): New intrinsic.
5163 (vneg_s8): Asm replaced with C.
5164 (vneg_s16): Likewise.
5165 (vneg_s32): Likewise.
5166 (vneg_s64): New intrinsic.
5167 (vnegq_f32): Asm replaced with C.
5168 (vnegq_f64): Likewise.
5169 (vnegq_s8): Likewise.
5170 (vnegq_s16): Likewise.
5171 (vnegq_s32): Likewise.
5172 (vnegq_s64): Likewise.
5173
5174 2013-10-09 Renlin Li <Renlin.Li@arm.com>
5175
5176 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
5177
5178 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5179
5180 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
5181 packed stack special handling.
5182 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
5183 back to fixed stack slots for FPRs saved due to stdarg.
5184
5185 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5186
5187 * config/s390/s390.c (s390_frame_info): Restructure function.
5188
5189 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5190
5191 * config/s390/s390.c (struct s390_frame_layout): New field
5192 gpr_save_slots.
5193 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
5194 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
5195 regs_ever_clobbered to char*.
5196 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
5197 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
5198 (s390_register_info): Enable FPR save slots. Move/Copy some
5199 functionality into ...
5200 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
5201 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
5202 function.
5203 (s390_frame_info): Do gpr slot allocation here now. stdarg does
5204 not imply a stack frame.
5205 (s390_init_frame_layout): Remove variable clobbered_regs.
5206 (s390_update_register_info): Remove function.
5207 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
5208 cannot be used for register renaming.
5209 (s390_hard_regno_scratch_ok): New function.
5210 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
5211 (s390_initial_elimination_offset): Change offset calculation of
5212 the return address pointer.
5213 (save_gprs): Deal with only r6 being saved from the call-saved regs.
5214 (restore_gprs): Set frame related flag.
5215 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
5216 (s390_emit_prologue): Call s390_register_info instead of
5217 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
5218 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
5219 (s390_optimize_prologue): Call s390_optimize_register_info.
5220 Try to remove also FPR slot save/restore INSNs. Remove frame
5221 related flags from restore INSNs.
5222
5223 2013-10-08 DJ Delorie <dj@redhat.com>
5224
5225 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
5226 (movhi): Likewise.
5227
5228 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
5229 avoid conflict with the MI use of %c.
5230 * config/rl78/rl78-real.md: change %c to %C throughout.
5231 * config/rl78/rl78-virt.md: Likewise.
5232
5233 2013-10-08 Jan Hubicka <jh@suse.cz>
5234
5235 * config/i386/i386.c (ix86_option_override_internal): Switch
5236 to SSE math for -ffast-math when target ISA supports SSE2.
5237
5238 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
5239
5240 * tree-flow.h: Remove some prototypes.
5241 * tree.h: Remove some protypes, add a couple.
5242 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
5243 using_eh_for_cleanups_p): Add interface routines for front ends.
5244 * tree-eh.h: New file. Add protoptyes.
5245 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
5246 (add_stmt_to_eh_lp_fn): Make static.
5247 (lower_try_finally): Use new using_eh_for_cleanups_p.
5248 * emit-rtl.c: Include tree-eh.h.
5249 * gimple.h: Include tree-eh.h.
5250
5251 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
5252
5253 PR tree-optimization/58480
5254 * tree-vrp.c (infer_nonnull_range): New function.
5255 (infer_value_range): Call infer_nonnull_range.
5256
5257 2013-10-08 Dehao Chen <dehao@google.com>
5258
5259 PR tree-optimization/58619
5260 * tree-inline.c (copy_phis_for_bb): Combine location data
5261 only if non-null.
5262
5263 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
5264
5265 PR target/58423
5266 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
5267 RTX_FRAME_RELATED_P on INSN.
5268
5269 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5270
5271 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
5272 (altivec_expand_vec_perm_const): Call it.
5273
5274 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5275
5276 * config/rs6000/vector.md (mov<mode>): Emit permuted move
5277 sequences for LE VSX loads and stores at expand time.
5278 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
5279 prototype.
5280 * config/rs6000/rs6000.c (rs6000_const_vec): New.
5281 (rs6000_gen_le_vsx_permute): New.
5282 (rs6000_gen_le_vsx_load): New.
5283 (rs6000_gen_le_vsx_store): New.
5284 (rs6000_gen_le_vsx_move): New.
5285 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
5286 (*vsx_le_perm_load_v4si): New.
5287 (*vsx_le_perm_load_v8hi): New.
5288 (*vsx_le_perm_load_v16qi): New.
5289 (*vsx_le_perm_store_v2di): New.
5290 (*vsx_le_perm_store_v4si): New.
5291 (*vsx_le_perm_store_v8hi): New.
5292 (*vsx_le_perm_store_v16qi): New.
5293 (*vsx_xxpermdi2_le_<mode>): New.
5294 (*vsx_xxpermdi4_le_<mode>): New.
5295 (*vsx_xxpermdi8_le_V8HI): New.
5296 (*vsx_xxpermdi16_le_V16QI): New.
5297 (*vsx_lxvd2x2_le_<mode>): New.
5298 (*vsx_lxvd2x4_le_<mode>): New.
5299 (*vsx_lxvd2x8_le_V8HI): New.
5300 (*vsx_lxvd2x16_le_V16QI): New.
5301 (*vsx_stxvd2x2_le_<mode>): New.
5302 (*vsx_stxvd2x4_le_<mode>): New.
5303 (*vsx_stxvd2x8_le_V8HI): New.
5304 (*vsx_stxvd2x16_le_V16QI): New.
5305
5306 2013-10-07 Renlin Li <Renlin.Li@arm.com>
5307
5308 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
5309
5310 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5311
5312 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
5313 loop to work also for 31bit ABI.
5314 Save the stack pointer for frame_size > 0.
5315
5316 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5317
5318 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
5319 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
5320 constraint letters from expanders.
5321 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
5322 retry count to general_operand.
5323 ("tabort"): Give operand 0 a mode.
5324 ("tabort_1"): Add mode and constraint letter for operand 0.
5325 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
5326
5327 2013-10-04 Jeff Law <law@redhat.com>
5328
5329 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
5330
5331 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
5332 out of ...
5333 (thread_across_edge): Here. Call it.
5334
5335 2013-10-04 Cary Coutant <ccoutant@google.com>
5336
5337 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
5338 discarding a location list expression (or a piece of one).
5339
5340 2013-10-03 Jan Hubicka <jh@suse.cz>
5341
5342 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
5343 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
5344 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
5345
5346 2013-10-03 Jan Hubicka <jh@suse.cz>
5347
5348 * config/i386/i386.c (ix86_option_override_internal): Do not enable
5349 accumulate-outgoing-args when producing unwind info.
5350
5351 2013-10-03 Wei Mi <wmi@google.com>
5352
5353 * lra-constraints.c (insert_move_for_subreg): New function
5354 extracted from simplify_operand_subreg.
5355 (simplify_operand_subreg): Add reload for paradoxical subreg.
5356
5357 2013-10-03 Rong Xu <xur@google.com>
5358
5359 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
5360 the candidate of builtin_expect such that we should fix the
5361 size/time estimation.
5362 (estimate_function_body_sizes): Do the acutally size/time fix-up
5363 for builtin_expect.
5364
5365 2013-10-03 Rong Xu <xur@google.com>
5366
5367 * predict.c (tree_predict_by_opcode): Get the probability
5368 for builtin_expect from param builtin_expect_probability.
5369 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
5370 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
5371 * doc/invoke.texi: Add documentation for builtin-expect-probability.
5372
5373 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
5374
5375 PR c++/19476
5376 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
5377 * calls.c (alloca_call_p): Use get_callee_fndecl.
5378 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
5379 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
5380 Likewise.
5381 (vrp_visit_stmt): Remove duplicated code.
5382
5383 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
5384
5385 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
5386 ceildf2, btruncdf2, instead of vsx_* name.
5387
5388 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
5389 iterators to only do V2DF and V4SF here. Move the DF code to
5390 rs6000.md where it is combined with SF mode. Replace <VSv> with
5391 just 'v' since only vector operations are handled with these insns
5392 after moving the DF support to rs6000.md.
5393 (vsx_sub<mode>3): Likewise.
5394 (vsx_mul<mode>3): Likewise.
5395 (vsx_div<mode>3): Likewise.
5396 (vsx_fre<mode>2): Likewise.
5397 (vsx_neg<mode>2): Likewise.
5398 (vsx_abs<mode>2): Likewise.
5399 (vsx_nabs<mode>2): Likewise.
5400 (vsx_smax<mode>3): Likewise.
5401 (vsx_smin<mode>3): Likewise.
5402 (vsx_sqrt<mode>2): Likewise.
5403 (vsx_rsqrte<mode>2): Likewise.
5404 (vsx_fms<mode>4): Likewise.
5405 (vsx_nfma<mode>4): Likewise.
5406 (vsx_copysign<mode>3): Likewise.
5407 (vsx_btrunc<mode>2): Likewise.
5408 (vsx_floor<mode>2): Likewise.
5409 (vsx_ceil<mode>2): Likewise.
5410 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
5411 (vsx_sminsf3): Likewise.
5412 (vsx_fmadf4): Likewise.
5413 (vsx_fmsdf4): Likewise.
5414 (vsx_nfmadf4): Likewise.
5415 (vsx_nfmsdf4): Likewise.
5416 (vsx_cmpdf_internal1): Likewise.
5417
5418 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
5419 simpler to select whether a target has SPE or traditional floating
5420 point support in iterators.
5421 (TARGET_DF_SPE): Likewise.
5422 (TARGET_SF_FPR): Likewise.
5423 (TARGET_DF_FPR): Likewise.
5424 (TARGET_SF_INSN): Macros to say whether floating point support
5425 exists for a given operation for expanders.
5426 (TARGET_DF_INSN): Likewise.
5427
5428 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
5429 combining of SF/DF mode operations, using both traditional and VSX
5430 registers.
5431 (Fvsx): Likewise.
5432 (Ff): Likewise.
5433 (Fv): Likewise.
5434 (Fs): Likewise.
5435 (Ffre): Likewise.
5436 (FFRE): Likewise.
5437 (abs<mode>2): Combine SF/DF modes using traditional floating point
5438 instructions. Add support for using the upper DF registers with
5439 VSX support, and SF registers with power8-vector support. Update
5440 expanders for operations supported by both the SPE and traditional
5441 floating point units.
5442 (abs<mode>2_fpr): Likewise.
5443 (nabs<mode>2): Likewise.
5444 (nabs<mode>2_fpr): Likewise.
5445 (neg<mode>2): Likewise.
5446 (neg<mode>2_fpr): Likewise.
5447 (add<mode>3): Likewise.
5448 (add<mode>3_fpr): Likewise.
5449 (sub<mode>3): Likewise.
5450 (sub<mode>3_fpr): Likewise.
5451 (mul<mode>3): Likewise.
5452 (mul<mode>3_fpr): Likewise.
5453 (div<mode>3): Likewise.
5454 (div<mode>3_fpr): Likewise.
5455 (sqrt<mode>3): Likewise.
5456 (sqrt<mode>3_fpr): Likewise.
5457 (fre<Fs>): Likewise.
5458 (rsqrt<mode>2): Likewise.
5459 (cmp<mode>_fpr): Likewise.
5460 (smax<mode>3): Likewise.
5461 (smin<mode>3): Likewise.
5462 (smax<mode>3_vsx): Likewise.
5463 (smin<mode>3_vsx): Likewise.
5464 (negsf2): Delete SF operations that are merged with DF.
5465 (abssf2): Likewise.
5466 (addsf3): Likewise.
5467 (subsf3): Likewise.
5468 (mulsf3): Likewise.
5469 (divsf3): Likewise.
5470 (fres): Likewise.
5471 (fmasf4_fpr): Likewise.
5472 (fmssf4_fpr): Likewise.
5473 (nfmasf4_fpr): Likewise.
5474 (nfmssf4_fpr): Likewise.
5475 (sqrtsf2): Likewise.
5476 (rsqrtsf_internal1): Likewise.
5477 (smaxsf3): Likewise.
5478 (sminsf3): Likewise.
5479 (cmpsf_internal1): Likewise.
5480 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
5481 (negdf2): Delete DF operations that are merged with SF.
5482 (absdf2): Likewise.
5483 (nabsdf2): Likewise.
5484 (adddf3): Likewise.
5485 (subdf3): Likewise.
5486 (muldf3): Likewise.
5487 (divdf3): Likewise.
5488 (fred): Likewise.
5489 (rsqrtdf_internal1): Likewise.
5490 (fmadf4_fpr): Likewise.
5491 (fmsdf4_fpr): Likewise.
5492 (nfmadf4_fpr): Likewise.
5493 (nfmsdf4_fpr): Likewise.
5494 (sqrtdf2): Likewise.
5495 (smaxdf3): Likewise.
5496 (smindf3): Likewise.
5497 (cmpdf_internal1): Likewise.
5498 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
5499 (btrunc<mode>2): Delete separate expander, and combine with the
5500 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
5501 (btrunc<mode>2_fpr): Likewise.
5502 (ceil<mode>2): Likewise.
5503 (ceil<mode>2_fpr): Likewise.
5504 (floor<mode>2): Likewise.
5505 (floor<mode>2_fpr): Likewise.
5506 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
5507 Add support for using the upper registers with VSX and
5508 power8-vector. Move insns to be closer to the define_expands. On
5509 VSX systems, prefer the traditional form of FMA over the VSX
5510 version, since the traditional form allows the target not to
5511 overlap with the inputs.
5512 (fms<mode>4_fpr): Likewise.
5513 (nfma<mode>4_fpr): Likewise.
5514 (nfms<mode>4_fpr): Likewise.
5515
5516 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5517 Richard Earnshaw <richard.earnshaw@arm.com>
5518
5519 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
5520 (struct mult_cost_table): Likewise.
5521 (struct mem_cost_table): Likewise.
5522 (struct fp_cost_table): Likewise.
5523 (struct vector_cost_table): Likewise.
5524 (cpu_cost_table): Likewise.
5525 * config/arm/arm.opt (mold-rts-costs): New option.
5526 (mnew-generic-costs): Likewise.
5527 * config/arm/arm.c (generic_extra_costs): New table.
5528 (cortexa15_extra_costs): Likewise.
5529 (arm_slowmul_tune): Use NULL as new costs.
5530 (arm_fastmul_tune): Likewise.
5531 (arm_strongarm_tune): Likewise.
5532 (arm_xscale_tune): Likewise.
5533 (arm_9e_tune): Likewise.
5534 (arm_v6t2_tune): Likewise.
5535 (arm_cortex_a5_tune): Likewise.
5536 (arm_cortex_a9_tune): Likewise.
5537 (arm_v6m_tune): Likewise.
5538 (arm_fa726te_tune): Likewise.
5539 (arm_cortex_a15_tune): Use cortex15_extra_costs.
5540 (arm_cortex_tune): Use generict_extra_costs.
5541 (shifter_op_p): New function.
5542 (arm_unspec_cost): Likewise.
5543 (LIBCALL_COST): Define.
5544 (arm_new_rtx_costs): New function.
5545 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
5546 table is available. Use old costs otherwise unless mnew-generic-costs
5547 is specified.
5548 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
5549 (cpu_cost_table): Declare.
5550
5551 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
5552
5553 PR target/58460
5554 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
5555 (*subs_mul_imm_<mode>)
5556 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
5557 (*sub_<shift>_<mode>)
5558 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
5559 Remove k constraint.
5560
5561 2013-10-03 Ian Bolton <ian.bolton@arm.com>
5562
5563 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
5564 code.
5565 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
5566
5567 2013-10-02 Teresa Johnson <tejohnson@google.com>
5568
5569 * predict.c (probably_never_executed): New function.
5570 (probably_never_executed_bb_p): Invoke probably_never_executed.
5571 (probably_never_executed_edge_p): Ditto.
5572 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
5573 Treat profile insanities conservatively.
5574
5575 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
5576
5577 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
5578
5579 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
5580
5581 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
5582 use it. Use smaller increase for scratch. Don't increase reject
5583 for early clobber scratch.
5584 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
5585 setting eliminated regs except setting fp from hfp.
5586 (lra_eliminate): Check lra_insn_recog_data on NULL.
5587
5588 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
5589
5590 PR target/58587
5591 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
5592 setting -mvsx-timode by default until the underlying problem is fixed.
5593 (RS6000_CPU, power7 defaults): Likewise.
5594
5595 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5596
5597 * config/x-linux (host-linux.o): Remove header dependencies.
5598 Use $(COMPILE) and $(POSTCOMPILE).
5599 * config/t-linux-android (linux-android.o): Ditto.
5600
5601 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5602
5603 * Makefile.in (expmed.o-warn): Remove.
5604
5605 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5606
5607 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
5608 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
5609
5610 2013-10-02 Teresa Johnson <tejohnson@google.com>
5611
5612 * dojump.c (do_jump_1): Divide probability between
5613 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
5614
5615 2013-10-02 Tom Tromey <tromey@redhat.com>
5616
5617 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
5618
5619 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5620
5621 * tree-flow.h: Remove some prototypes.
5622 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
5623 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
5624 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
5625 mark_virtual_phi_result_for_renaming): Relocate here.
5626 * tree-into-ssa.h: Add prototypes.
5627 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
5628 single_pred_before_succ_order.
5629 (blocks_in_phiopt_order): Rename and move to cfganal.c.
5630 (nonfreeing_call_p) Move to gimple.c.
5631 * cfganal.c (single_pred_before_succ_order): Move and renamed from
5632 tree-ssa-phiopt.c.
5633 * basic-block.h (single_pred_before_succ_order): Add prototype.
5634 * gimple.c (nonfreeing_call_p): Relocate here.
5635 * gimple.h: Add prototype.
5636 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
5637 * tree-ssa-dom.h: New file. Relocate prototypes here.
5638 * tree-ssa.h: Include tree-ssa-dom.h.
5639
5640 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
5641
5642 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
5643 Use $(COMPILE) and $(POSTCOMPILE).
5644
5645 * config/alpha/x-alpha (driver-alpha.o): Ditto.
5646
5647 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5648
5649 * tree-flow.h: Remove some prototypes.
5650 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
5651 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5652 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
5653 propagate_tree_value*): Move from here to...
5654 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
5655 propagate_tree_value*): Relocate here.
5656 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
5657 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
5658 * gimple-fold.c: Remove gimple-fold.h from include list.
5659 * tree-vrp.c: Remove gimple-fold.h from include list.
5660 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
5661 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
5662 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
5663 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
5664 * sese.c: Add tree-ssa-propagate.h to include list.
5665
5666 2013-10-02 Richard Biener <rguenther@suse.de>
5667
5668 * tree-loop-distribution.c: Include tree-vectorizer.h for
5669 find_loop_location.
5670 (enum partition_kind): Remove PKIND_REDUCTION.
5671 (struct partition_s): Remove has_writes member, add reduction_p member.
5672 (partition_alloc): Adjust.
5673 (partition_builtin_p): Likewise.
5674 (partition_has_writes): Remove.
5675 (partition_reduction_p): New function.
5676 (partition_merge_into): Likewise.
5677 (generate_code_for_partition): Commonize builtin partition
5678 handling tail.
5679 (rdg_cannot_recompute_vertex_p): Remove.
5680 (already_processed_vertex_p): Likewise.
5681 (rdg_flag_vertex): Do not set has_writes.
5682 (classify_partition): Adjust.
5683 (rdg_build_partitions): Do not set has_writes, treat all
5684 partitions as useful.
5685 (distribute_loop): Record number of library calls generated. Adjust.
5686 (tree_loop_distribution): Report number of loops and library
5687 calls generated as opt-info.
5688
5689 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
5690
5691 * tree-flow.h: Include new .h files. Move prototypes.
5692 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
5693 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
5694 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
5695 * tree-pretty-print.h: Add prototypes from tree-flow.h.
5696 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
5697 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
5698 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
5699 * tree.h (get_ref_base_and_extent): Move prototype out.
5700 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
5701 tree-dfa.h.
5702 * gimple-low.h: New File. Add prototypes from tree-flow.h.
5703 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
5704 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
5705 * tree-scalar-evolution.c: Include tree.h.
5706 * sese.c: Include tree.h.
5707 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
5708 * dwarf2out.c: Include tree-dfa.h.
5709 * tree-chrec.c: Include tree.h.
5710 * tree-data-ref.c: Include tree.h.
5711
5712 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
5713
5714 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
5715 Fix whitespace.
5716
5717 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
5718
5719 * config/t-sol2 (sol2-c.o): Remove header dependencies.
5720 Use $(COMPILE) and $(POSTCOMPILE).
5721 (sol2-cxx.o): Likewise.
5722 (sol2-stubs.o): Likewise.
5723 (sol2.o): Likewise.
5724 * config/x-solaris (host-solaris.o): Likewise.
5725
5726 * config/sparc/t-sparc (sparc.o): Remove.
5727 (sparc-c.o): Remove header dependencies.
5728 Use $(COMPILE) and $(POSTCOMPILE).
5729 * config/sparc/x-sparc: Likewise.
5730
5731 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
5732
5733 * config/arc/arc-opts.h: Add 2013 to Copyright years.
5734 * config/arc/arc700.md: Likewise.
5735 * config/arc/arc-modes.def: Likewise.
5736 * config/arc/arc-simd.h: Likewise.
5737 * config/arc/t-arc-uClibc: Likewise.
5738 * config/arc/t-arc-newlib: Likewise.
5739
5740 2013-10-02 Renlin Li <renlin.li@arm.com>
5741
5742 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
5743 plus_constant.
5744 (aarch64_expand_epilogue): Likewise.
5745
5746 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
5747 Yufeng Zhang <yufeng.zhang@arm.com>
5748
5749 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
5750 declaration.
5751 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
5752 'base_in' represent a conversion and legal_cast_p_1 holds; set
5753 'base_in' with the returned value from get_unwidened.
5754
5755 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5756
5757 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
5758 plus_constant is not used.
5759
5760 2013-10-01 Wei Mi <wmi@google.com>
5761
5762 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
5763 * config/i386/i386.md: Add define_peephole2 to
5764 break partial reg stall for cvtss2sd/cvtsd2ss.
5765
5766 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5767
5768 * config/arc/arc.c (pass_arc_ifcvt::clone):
5769 Update for ctxt_ -> m_ctxt change.
5770
5771 2013-10-01 Jeff Law <law@redhat.com>
5772
5773 * tree-ssa-threadupdate.c (struct redirection_data): Delete
5774 outgoing_edge and intermediate_edge fields. Instead store the path.
5775 (redirection_data::hash): Hash on the last edge's destination index.
5776 (redirection_data::equal): Check the entire thread path.
5777 (lookup_redirectio_data): Corresponding changes.
5778 (create_edge_and_update_destination_phis): Likewise.
5779 (thread_single_edge): Likewise.
5780
5781 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5782 Diego Novillo <dnovillo@google.com>
5783
5784 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
5785 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
5786 (vld32wh_insn, vld32wl_insn): Delete commented-out old
5787 versions of these patterns.
5788
5789 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
5790 (__builtin_arc_aligned): Likewise.
5791
5792 * config/arc/arc.md: Expand adc_0 comment stating the intended
5793 purpose and why it isn't ready.
5794 Replace commented out call_value_via_label_mixed with a
5795 plain comment about bl_s.
5796
5797 * config/arc/arc.c (stdio.h): Don't include directly.
5798 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
5799 Remove if (1) condition.
5800 (arc_encode_section_info): Fix comment.
5801
5802 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5803
5804 * config/arc/arc.c (arc_conditional_register_usage):
5805 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
5806 Also set reg_alloc_order for DMA config regs.
5807
5808 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5809 Jeremy Bennett <jeremy.bennett@embecosm.com>
5810
5811 * doc/install.texi (--with-cpu): Mention ARC.
5812 (arc-*-elf32): New paragraph.
5813 (arc-linux-uclibc): Likewise.
5814 * doc/md.texi (Machine Constraints): Add ARC part.
5815 * doc/invoke.texi: (menu): Add ARC Options.
5816 (Machine Dependent Options) <ARC Options>: Add synopsis.
5817 (node ARC Options): Add.
5818 * doc/extend.texi (long_call / short_call attribute): Add ARC.
5819 (ARC Built-in Functions): New section defining
5820 generic ARC built-in functions.
5821 (ARC SIMD Built-in Functions): New section defining SIMD specific
5822 built-in functions.
5823 (Declaring Attributes of Functions): Extended
5824 description of short_call and long_call attributes for ARC and
5825 added index entries.
5826
5827 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
5828 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
5829 Joern Rennecke <joern.rennecke@embecosm.com>
5830 Muhammad Khurram Riaz <khurram.riaz@arc.com>
5831 Brendan Kehoe <brendan@zen.org>
5832 Michael Eager <eager@eagercon.com>
5833 Simon Cook <simon.cook@embecosm.com>
5834 Jeremy Bennett <jeremy.bennett@embecosm.com>
5835
5836 * config/arc, common/config/arc: New directories.
5837
5838 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
5839 Brendan Kehoe <brendan@zen.org>
5840 Simon Cook <simon.cook@embecosm.com>
5841
5842 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
5843
5844 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5845
5846 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
5847 * tree-ssa-coalesce.h: New. Move prototype to here.
5848 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
5849 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
5850 (gimple_can_coalesce_p): Move to...
5851 * gimple.c (gimple_can_coalesce_p): Here.
5852
5853 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
5854
5855 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
5856 (dump_decl_set): Move to gimple.c.
5857 * gimple.h: Don't include tree-ssa-operands.h.
5858 (dump_decl_set): Add prototype.
5859 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5860 Move to gimple-ssa.h.
5861 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
5862 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
5863 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
5864 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
5865 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
5866 than PHI_ARG_DEF.
5867 (dump_decl_set): Relocate here.
5868 * gimple-ssa.h: New file.
5869 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
5870 Relocate from gimple.h.
5871 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
5872 * tree-ssa-operands.c (swap_ssa_operands): Rename from
5873 swap_tree_operands and remove non-ssa path.
5874 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
5875 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
5876 swap_ssa_operands.
5877 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
5878 vect_is_simple_reduction_1): Use swap_ssa_operands.
5879 * tree-flow.h: Move various prototypes to tree-phinodes.h.
5880 (enum need_phi_state): Move to tree-into-ssa.c.
5881 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5882 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
5883 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
5884 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
5885 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5886 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5887 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5888 num_imm_uses): Move to ssa-iterators.h.
5889 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
5890 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
5891 tree-phinodes.h.
5892 (op_iter_done, op_iter_next_def, op_iter_next_tree,
5893 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
5894 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
5895 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
5896 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
5897 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
5898 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
5899 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
5900 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
5901 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
5902 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
5903 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
5904 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
5905 (set_phi_nodes): Move to tree-phinodes.h.
5906 * tree-ssa-operands.h (enum ssa_op_iter_type,
5907 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
5908 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
5909 (dump_decl_set): Remove prototype.
5910 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
5911 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
5912 (set_phi_nodes): Relocate from tree-flow-inline.h.
5913 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
5914 tree-flow-inline.h
5915 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
5916 include list. Temporarily add gimple.h to include list.
5917 * ssa-iterators.h: New file.
5918 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
5919 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
5920 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
5921 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
5922 Relocate from tree-ssa-operands.h.
5923 (delink_imm_use, link_imm_use_to_list, link_imm_use,
5924 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
5925 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
5926 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
5927 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
5928 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
5929 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
5930 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
5931 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
5932 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
5933 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
5934 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
5935 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
5936 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
5937 Relocate from tree-flow-inline.h.
5938 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
5939
5940 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
5941
5942 * aarch64-simd.md
5943 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
5944 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
5945 Insert '\t' to output template.
5946 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
5947 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
5948 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
5949 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
5950
5951 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
5952
5953 * doc/install.texi (Host/target specific installation notes for GCC):
5954 Put @anchor before @heading.
5955 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
5956 Use @email for email addresses.
5957
5958 2013-10-01 Jeff Law <law@redhat.com>
5959
5960 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
5961 a vec. Only delete the path if we create one without successfully
5962 registering a jump thread.
5963 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
5964 as a pointer.
5965 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
5966 (paths): New vector of jump threading paths.
5967 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
5968 (THREAD_PATH): New accessor macro for the entire thread path.
5969 (lookup_redirection_data): Get intermediate and final outgoing edge
5970 from the thread path.
5971 (create_edge_and_update_destination_phis): Copy the threading path.
5972 (ssa_fix_duplicate_block_edges): Get edges and block types from the
5973 jump threading path.
5974 (ssa_redirect_edges): Get edges and block types from the jump threading
5975 path. Free the path vector.
5976 (thread_block): Get edges from the jump threading path. Look at the
5977 entire path to see if we thread to a loop exit. If we cancel a jump
5978 thread request, then free the path vector.
5979 (thread_single_edge): Get edges and block types from the jump threading
5980 path. Free the path vector.
5981 (thread_through_loop_header): Get edges and block types from the jump
5982 threading path. Free the path vector.
5983 (mark_threaded_blocks): Iterate over the vector of paths and store
5984 the path on the appropriate edge. Get edges and block types from the
5985 jump threading path.
5986 (mark_threaded_blocks): Get edges and block types from the jump
5987 threading path. Free the path vector.
5988 (thread_through_all_blocks): Use the vector of paths rather than
5989 a vector of 3-edge sets.
5990 (register_jump_thread): Accept pointer to a path vector rather
5991 than the path vector itself. Store the path vector for later use.
5992 Simplify.
5993
5994 2013-10-01 Jakub Jelinek <jakub@redhat.com>
5995 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5996
5997 PR target/58574
5998 * config/s390/s390.c (s390_split_branches): Modify check for table
5999 jump insns.
6000 (s390_chunkify_start): Rearrange table jump insn check in order to
6001 deal with compare and branch insns correctly.
6002
6003 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
6004
6005 PR target/58578
6006 Revert
6007 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
6008 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
6009 define_insn_and_split.
6010 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
6011 (shiftsi3_compare): New pattern.
6012 (rrx): New pattern.
6013 * config/arm/unspecs.md (UNSPEC_RRX): New.
6014
6015 2013-10-01 Alan Modra <amodra@gmail.com>
6016
6017 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
6018 casing inout operands.
6019
6020 2013-10-01 Richard Biener <rguenther@suse.de>
6021
6022 PR tree-optimization/58553
6023 * tree-loop-distribution.c (struct partition_s): Add niter member.
6024 (classify_partition): Populate niter member for the partition
6025 and properly identify whether the relevant store happens before
6026 or after the loop exit.
6027 (generate_memset_builtin): Use niter member from the partition.
6028 (generate_memcpy_builtin): Likewise.
6029
6030 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6031
6032 * vec.h (vec_prefix, vec): Prefix member names with "m_".
6033 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
6034
6035 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
6036
6037 * basic-block.h (edge_list): Prefix member names with "m_".
6038 * context.h (context): Likewise.
6039 * domwalk.h (dom_walker): Likewise.
6040 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
6041 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
6042 * hash-table.h (hash_table): Likewise.
6043 * machmode.h (bit_field_mode_iterator): Likewise.
6044 * pass_manager.h (pass_list): Likewise.
6045 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
6046 * tree-pass.h (pass_data): Likewise.
6047 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
6048 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
6049 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
6050 * asan.c (pass_data_asan): Update accordingly.
6051 * cfganal.c (control_dependences::find_control_dependence): Likewise.
6052 (control_dependences::control_dependences): Likewise.
6053 (control_dependences::~control_dependences): Likewise.
6054 (control_dependences::~control_dependences): Likewise.
6055 (control_dependences::get_edges_dependent_on): Likewise.
6056 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
6057 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
6058 * context.c (gcc::context::context): Likewise.
6059 * cprop.c (pass_rtl_cprop::clone): Likewise.
6060 * domwalk.c (dom_walker::walk): Likewise.
6061 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
6062 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
6063 * mode-switching.c (pass_mode_switching::clone): Likewise.
6064 * passes.c (opt_pass::opt_pass): Likewise.
6065 (pass_manager::pass_manager): Likewise.
6066 * predict.c (pass_strip_predict_hints::clone): Likewise.
6067 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
6068 (pass_split_all_insns::clone): Likewise.
6069 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
6070 Likewise.
6071 (bit_field_mode_iterator::next_mode): Likewise.
6072 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
6073 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
6074 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
6075 * tree-complex.c (pass_lower_complex::clone): Likewise.
6076 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
6077 * tree-object-size.c (pass_object_sizes::clone): Likewise.
6078 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
6079 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
6080 (pass_fold_builtins::clone): Likewise.
6081 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
6082 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
6083 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
6084 (pass_cd_dce::clone): Likewise.
6085 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
6086 (pass_phi_only_cprop::clone): Likewise.
6087 * tree-ssa-dse.c (pass_dse::clone): Likewise.
6088 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
6089 * tree-ssa-loop.c (pass_lim::clone): Likewise.
6090 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
6091 * tree-ssa-pre.c (pass_fre::clone): Likewise.
6092 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
6093 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
6094 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
6095 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
6096 * tree-vrp.c (pass_vrp::clone): Likewise.
6097 * tsan.c (pass_tsan::clone): Likewise.
6098
6099 2013-09-30 Jakub Jelinek <jakub@redhat.com>
6100
6101 PR middle-end/58564
6102 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
6103 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
6104
6105 PR middle-end/58564
6106 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
6107 optimization, punt if sign_bit_p looked through any zero extension.
6108
6109 2013-09-30 Teresa Johnson <tejohnson@google.com>
6110
6111 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
6112 Update redirected out edge count in joiner case.
6113 (ssa_redirect_edges): Common the joiner and non-joiner cases
6114 so that joiner case gets profile updates.
6115
6116 2013-09-30 Richard Biener <rguenther@suse.de>
6117
6118 PR tree-optimization/58554
6119 * tree-loop-distribution.c (classify_partition): Require
6120 unconditionally executed stores for memcpy and memset recognition.
6121 (tree_loop_distribution): Calculate dominance info.
6122
6123 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
6124
6125 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
6126 (NO_PROFILE_COUNTERS): Likewise.
6127 (PROFILE_HOOK): Likewise.
6128 (FUNCTION_PROFILER): Likewise.
6129 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
6130
6131 2013-09-30 Iain Sandoe <iain@codesourcery.com>
6132
6133 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
6134 calls and defines in TARGET_MACHO conditional.
6135 (load_macho_picbase_di): Likewise.
6136 (reload_macho_picbase): Likewise.
6137 (reload_macho_picbase_si): Likewise.
6138 (reload_macho_picbase_di): Likewise.
6139 (nonlocal_goto_receiver): Likewise.
6140
6141 2013-09-30 Nick Clifton <nickc@redhat.com>
6142
6143 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
6144 that use the MSP430X ISA.
6145 (msp430_option_override): Scan -mmcu command line option for any
6146 MCU name that supports the MSP430X ISA.
6147 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
6148 -mmcu options which enable the MSP430X ISA.
6149
6150 2013-09-30 Richard Biener <rguenther@suse.de>
6151
6152 PR middle-end/58532
6153 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
6154 before looking for setjmp-like calls.
6155
6156 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6157
6158 PR target/10901
6159 * config/darwin-protos.h (machopic_get_function_picbase): New.
6160 * config/darwin.c (machopic_get_function_picbase): New.
6161 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
6162 label for a new func. (load_macho_picbase_di): Likewise.
6163 (reload_macho_picbase): New expand.
6164 (reload_macho_picbase_si): New insn.
6165 (reload_macho_picbase_di): New insn.
6166 (nonlocal_goto_receiver): New define and split.
6167 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
6168 (unspecv enum): Add UNSPECV_NLGR.
6169
6170 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6171
6172 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
6173 that altivec registers are correctly sized on Darwin.
6174
6175 2013-09-29 Iain Sandoe <iain@codesourcery.com>
6176
6177 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
6178 darwin-driver.o): Use COMPILE and POSTCOMPILE.
6179 * config/x-darwin (host-darwin.o): Likewise.
6180 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
6181 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
6182 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
6183
6184 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
6185
6186 * doc/invoke.texi: Fix usage of @tie{} command.
6187
6188 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
6189
6190 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
6191
6192 2013-09-28 Jan Hubicka <jh@suse.cz>
6193
6194 * config/i386/x86-tune.def: Add documentation for each of the options;
6195 add whitespace.
6196
6197 2013-09-28 Jan Hubicka <jh@suse.cz>
6198
6199 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
6200 generic.
6201 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
6202 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
6203 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
6204 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
6205 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
6206
6207 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
6208
6209 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
6210 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
6211 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
6212 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
6213 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
6214 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
6215 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
6216 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
6217 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
6218 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
6219 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
6220 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
6221 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
6222 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
6223 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
6224 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
6225 gimple.h, godump.c, graphite-clast-to-gimple.c,
6226 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
6227 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
6228 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
6229 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
6230 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
6231 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
6232 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
6233 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
6234 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
6235 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
6236 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
6237 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
6238 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
6239 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
6240 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
6241 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
6242 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
6243 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
6244 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
6245 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
6246 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
6247 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
6248 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
6249 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
6250 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
6251 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
6252 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
6253 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
6254 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
6255 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
6256 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
6257 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
6258 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
6259 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
6260 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
6261 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
6262 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
6263 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
6264 whitespace before "(".
6265
6266 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
6267
6268 * expr.h (extract_bit_field): Remove packedp parameter.
6269 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
6270 from forward declaration.
6271 (store_split_bit_field): Remove packedp arg from calls to
6272 extract_fixed_bit_field.
6273 (extract_bit_field_1): Remove packedp parameter and packedp
6274 argument from recursive calls and calls to extract_fixed_bit_field.
6275 (extract_bit_field): Remove packedp parameter and corresponding
6276 arg to extract_bit_field_1.
6277 (extract_fixed_bit_field): Remove packedp parameter. Remove code
6278 to issue warnings.
6279 (extract_split_bit_field): Remove packedp arg from call to
6280 extract_fixed_bit_field.
6281 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
6282 (copy_blkmode_from_reg): Likewise.
6283 (copy_blkmode_to_reg): Likewise.
6284 (read_complex_part): Likewise.
6285 (store_field): Likewise.
6286 (expand_expr_real_1): Likewise.
6287 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
6288 to extract_bit_field.
6289 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
6290 call to extract_bit_field.
6291 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
6292 call to extract_bit_field.
6293 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
6294 and special packedp behavior from -fstrict-volatile-bitfields
6295 documentation.
6296
6297 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
6298
6299 * lra-eliminations.c (init_elim_table): Guard value_p.
6300
6301 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
6302
6303 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
6304 DFmode, DImode, and SFmode in the upper VSX registers based on the
6305 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
6306 if -mpower8-vector. Combine -mvsx-timode handling with the rest
6307 of the VSX register handling.
6308
6309 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
6310 (f32_sv): Likewise.
6311 (zero_extendsidi2_lfiwzx): Add support for loading into the
6312 Altivec registers with -mpower8-vector. Use wu/wv constraints to
6313 only do VSX memory options on Altivec registers.
6314 (extendsidi2_lfiwax): Likewise.
6315 (extendsfdf2_fpr): Likewise.
6316 (mov<mode>_hardfloat, SF/SD modes): Likewise.
6317 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
6318 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
6319 (movdi_internal64): Likewise.
6320
6321 2013-09-27 Xinliang David Li <davidxl@google.com>
6322
6323 * opts.c (finish_options): Adjust parameters
6324 according to vect cost model.
6325 (common_handle_option): Set dynamic vect cost
6326 model for FDO.
6327 targhooks.c (default_add_stmt_cost): Compute stmt cost
6328 unconditionally.
6329 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
6330 Use helper function.
6331 * tree-vectorizer.h (unlimited_cost_model): New function.
6332 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
6333 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
6334 function.
6335 (vect_enhance_data_refs_alignment): Ditto.
6336 * flag-types.h: New enum.
6337 * common/config/i386/i386-common.c (ix86_option_init_struct):
6338 No need to initialize vect_cost_model flag.
6339 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
6340 unconditionally.
6341
6342 2013-09-27 Diego Novillo <dnovillo@google.com>
6343
6344 * gimple.h (enum ssa_mode): Remove.
6345
6346 2013-09-27 Paulo Matos <pmatos@broadcom.com>
6347
6348 * cfgloop.h (number_of_loops): Fix typo in check for null.
6349
6350 2013-09-27 Jakub Jelinek <jakub@redhat.com>
6351
6352 PR middle-end/58551
6353 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
6354 are children of outermost saved_cfun's loop, and set it up to
6355 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
6356 if loop != loop0 and SESE region contains bbs that belong to loop0.
6357
6358 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6359
6360 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
6361 (binary_scale_code_p, get_base_term, get_index_term): New functions.
6362 (set_address_segment, set_address_base, set_address_index)
6363 (set_address_disp): Accept the argument unconditionally.
6364 (baseness): Remove must_be_base_p and must_be_index_p checks.
6365 (decompose_normal_address): Classify as much as possible in the
6366 main loop.
6367
6368 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
6369
6370 * cse.c (count_reg_usage): Handle INT_LIST.
6371 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
6372 * reginfo.c (reg_scan_mark_refs): Likewise.
6373 * reload1.c (eliminate_regs_1): Likewise.
6374
6375 2013-09-27 Iain Sandoe <iain@codesourcery.com>
6376
6377 PR middle-end/58547
6378 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
6379 signed.
6380
6381 2013-09-27 Richard Biener <rguenther@suse.de>
6382
6383 PR tree-optimization/58459
6384 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
6385 restriction not propagating into loops.
6386
6387 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
6388
6389 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
6390 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
6391 * doc/tree-ssa.texi (Walking use-def chains): Delete.
6392
6393 2013-09-26 Richard Biener <rguenther@suse.de>
6394
6395 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
6396
6397 2013-09-26 Richard Biener <rguenther@suse.de>
6398
6399 * alias.h (component_uses_parent_alias_set): Rename to ...
6400 (component_uses_parent_alias_set_from): ... this.
6401 * alias.c (component_uses_parent_alias_set): Rename to ...
6402 (component_uses_parent_alias_set_from): ... this and return
6403 the desired parent.
6404 (reference_alias_ptr_type_1): Use the result from
6405 component_uses_parent_alias_set_from instead of stripping
6406 components one at a time.
6407 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
6408
6409 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6410
6411 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
6412 Move prototypes to...
6413 * tree-ssa-ter.h: New File. Move prototypes here.
6414 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
6415 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
6416 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
6417 from is_replaceable_p.
6418 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
6419 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
6420 (process_replaceable): Use ter_is_replaceable_p.
6421 (find_replaceable_in_bb): Use ter_is_replaceable_p.
6422 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
6423 newly refactored ssa_is_replaceable_p.
6424 * cfgexpand.c: Include tree-outof-ssa.h.
6425 * ssaexpand.h: Delete.
6426
6427 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
6428
6429 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
6430 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
6431 tree-ssa.c
6432 (create_gimple_tmp): Delete.
6433 (get_expr_type, build_assign, build_type_cast): Move to...
6434 * gimple-builder.c: New File.
6435 (get_expr_type): Relocate from gimple.c.
6436 (build_assign, build_type_cast): Change to only create ssanames.
6437 * gimple.h: Move prototypes to...
6438 * gimple-builder.h: New File. Here.
6439 * tree-ssa.h: And here.
6440 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
6441 count_uses_and_derefs): Relocate from gimple.c.
6442 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
6443 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
6444 * tree-ssa-math-opts (execute_cse_reciprocals): Use
6445 gimple_replace_ssa_lhs.
6446 * asan.c: Include gimple-builder.h.
6447 * Makefile.in: Add gimple-builder.o.
6448
6449 2013-09-26 Richard Biener <rguenther@suse.de>
6450
6451 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
6452 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
6453 (loe_visit_block): Use gcc_checking_assert.
6454 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
6455 gimple_assign_ssa_name_copy_p.
6456 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
6457 change.
6458
6459 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
6460
6461 * config/rs6000/t-rs6000 (rs6000.o): Remove.
6462 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
6463
6464 2013-09-26 Richard Biener <rguenther@suse.de>
6465
6466 PR tree-optimization/58539
6467 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
6468 the fact that debug statements are not taking part in loop-closed
6469 SSA construction.
6470
6471 2013-09-26 Nick Clifton <nickc@redhat.com>
6472
6473 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
6474 time warning message.
6475 (msp430_print_operand_raw): Delete unused letter parameter.
6476 (TARGET_PRINT_OPERAND_ADDRESS): Define.
6477 (msp430_print_operand_address): New function.
6478 (msp430_print_operand): Move address printing code from here to
6479 new function.
6480 * config/msp430/msp430.md (movsipsi2): Add comment in generated
6481 assembler.
6482 (zero_extendpsisi2): Likewise.
6483 (extendpsisi2): New pattern.
6484 (andneghi3): New pattern.
6485
6486 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
6487
6488 * config/aarch64/aarch64.opt (mlra): New option.
6489 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
6490 (TARGET_LRA_P): Define.
6491
6492 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
6493
6494 * expr.c (expand_assignment): Remove obsolete comment.
6495
6496 2013-09-25 Jeff Law <law@redhat.com>
6497
6498 * tree-flow.h (thread_through_all_blocks): Prototype moved into
6499 tree-ssa-threadupdate.h.
6500 (register_jump_thread): Similarly.
6501 * tree-ssa-threadupdate.h: New header file.
6502 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
6503 * tree-vrp.c: Likewise.
6504 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
6505 (thread_around_empty_blocks): Change type of path vector argument to
6506 an edge,type pair from just an edge. Initialize both elements when
6507 appending to a jump threading path. Tweak references to elements
6508 appropriately.
6509 (thread_across_edge): Similarly. Release memory for the elements
6510 as needed.
6511 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
6512 (dump_jump_thread_path): New function broken out from
6513 register_jump_thread.
6514 (register_jump_thread): Use dump_jump_thread_path. Change type of
6515 path vector entries. Search the path for NULL edges and dump
6516 the path if one is found. Tweak the conversion of path to 3-edge
6517 form to use the block copy type information embedded in the path.
6518
6519 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6520
6521 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
6522
6523 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
6524 Vladimir Makarov <vmakarov@redhat.com>
6525
6526 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
6527 from the least significant bit.
6528 (strip_address_mutations): Add bitfield operations handling.
6529 (must_be_index_p): Add shifting and rotate operations handling.
6530 (set_address_base): Use must_be_base_p predicate.
6531 (set_address_index): Use must_be_index_p predicate.
6532
6533 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
6534 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
6535 Sergey Lega <sergey.s.lega@intel.com>
6536 Anna Tikhonova <anna.tikhonova@intel.com>
6537 Ilya Tocar <ilya.tocar@intel.com>
6538 Andrey Turetskiy <andrey.turetskiy@intel.com>
6539 Ilya Verbin <ilya.verbin@intel.com>
6540 Kirill Yukhin <kirill.yukhin@intel.com>
6541 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
6542
6543 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
6544 Use new names.
6545 (ix86_expand_vector_move_misalign): Support new unaligned load and
6546 stores and use new names.
6547 (CODE_FOR_sse2_storedqu): Rename to ...
6548 (CODE_FOR_sse2_storedquv16qi): ... this.
6549 (CODE_FOR_sse2_loaddqu): Rename to ...
6550 (CODE_FOR_sse2_loaddquv16qi): ... this.
6551 (CODE_FOR_avx_loaddqu256): Rename to ...
6552 (CODE_FOR_avx_loaddquv32qi): ... this.
6553 (CODE_FOR_avx_storedqu256): Rename to ...
6554 (CODE_FOR_avx_storedquv32qi): ... this.
6555 * config/i386/i386.md (fpint_logic): New.
6556 * config/i386/sse.md (VMOVE): Extend for AVX512.
6557 (VF): Ditto.
6558 (VF_128_256): New.
6559 (VF_512): Ditto.
6560 (VI_UNALIGNED_LOADSTORE): Ditto.
6561 (sse2_avx_avx512f): Ditto.
6562 (sse2_avx2): Extend for AVX512.
6563 (sse4_1_avx2): Ditto.
6564 (avx2_avx512f): New.
6565 (sse): Extend for AVX512.
6566 (sse2): Ditto.
6567 (sse4_1): Ditto.
6568 (avxsizesuffix): Ditto.
6569 (sseintvecmode): Ditto.
6570 (ssePSmode): Ditto.
6571 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
6572 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
6573 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6574 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
6575 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
6576 (<sse2_avx_avx512f>_storedqu<mode): ... this.
6577 (<sse>_movnt<mode>): Replace constraint "x" with "v".
6578 (STORENT_MODE): Extend for AVX512.
6579 (*absneg<mode>2): Replace constraint "x" with "v".
6580 (*mul<mode>3): Ditto.
6581 (*ieee_smin<mode>3): Ditto.
6582 (*ieee_smax<mode>3): Ditto.
6583 (avx_cmp<mode>3): Replace VF with VF_128_256.
6584 (*<sse>_maskcmp<mode>3_comm): Ditto.
6585 (<sse>_maskcmp<mode>3): Ditto.
6586 (<sse>_andnot<mode>3): Extend for AVX512.
6587 (<code><mode>3, anylogic): Replace VF with VF_128_256.
6588 (<code><mode>3, fpint_logic): New.
6589 (*<code><mode>3): Extend for AVX512.
6590 (avx512flogicsuff): New.
6591 (avx512f_<logic><mode>): Ditto.
6592 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
6593 VF_128_256.
6594 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
6595 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
6596 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
6597 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
6598 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
6599 (xop_vpermil2<mode>3): Ditto.
6600 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
6601 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
6602 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
6603 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
6604
6605 2013-09-25 Tom Tromey <tromey@redhat.com>
6606
6607 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
6608 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
6609 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
6610 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
6611 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
6612 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
6613 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
6614 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
6615 (GRAPHITE_HTAB_H): Remove.
6616
6617 2013-09-25 Tom Tromey <tromey@redhat.com>
6618
6619 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
6620
6621 2013-09-25 Tom Tromey <tromey@redhat.com>
6622
6623 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
6624
6625 2013-09-25 Tom Tromey <tromey@redhat.com>
6626
6627 * config/i386/t-i386 (i386.o): Remove.
6628 (i386-c.o): Use COMPILE and POSTCOMPILE.
6629
6630 2013-09-25 Tom Tromey <tromey@redhat.com>
6631
6632 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
6633
6634 2013-09-25 Tom Tromey <tromey@redhat.com>
6635
6636 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
6637 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
6638 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
6639 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
6640 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
6641 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
6642 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
6643 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
6644 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
6645 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
6646 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
6647 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
6648 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
6649 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
6650 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
6651 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
6652 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
6653 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
6654 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
6655 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
6656 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
6657 (tree-ssa-pre.o, tree-ssa-sccvn.o)
6658 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
6659 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
6660 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
6661 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
6662 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
6663 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
6664 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
6665 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
6666 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
6667 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
6668 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
6669 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
6670 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
6671 (graphite-clast-to-gimple.o, graphite-dependences.o)
6672 (graphite-interchange.o, graphite-poly.o)
6673 (graphite-scop-detection.o, graphite-sese-to-poly.o)
6674 (graphite-optimize-isl.o, tree-vect-loop.o)
6675 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
6676 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
6677 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
6678 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
6679 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
6680 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
6681 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
6682 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
6683 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
6684 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
6685 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
6686 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
6687 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
6688 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
6689 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
6690 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
6691 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
6692 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
6693 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
6694 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
6695 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
6696 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
6697 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
6698 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
6699 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
6700 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
6701 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
6702 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
6703 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
6704 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
6705 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
6706 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
6707 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
6708 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
6709 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
6710 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
6711 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
6712 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
6713 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
6714 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
6715 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
6716 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
6717 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
6718 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
6719 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
6720 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
6721 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
6722 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
6723 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
6724 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
6725 (gcov-dump.o): Remove.
6726 (default-c.o): Use COMPILE and POSTCOMPILE.
6727 (CFLAGS-gcc.o): New variable.
6728 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
6729
6730 2013-09-25 Tom Tromey <tromey@redhat.com>
6731
6732 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
6733 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
6734 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
6735 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
6736 (c-family/c-pragma.o, c-family/c-pretty-print.o)
6737 (c-family/c-semantics.o, c-family/c-ada-spec.o)
6738 (c-family/array-notation-common.o, c-family/stub-objc.o)
6739 (c-family/c-ubsan.o): Remove.
6740
6741 2013-09-25 Tom Tromey <tromey@redhat.com>
6742
6743 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
6744
6745 2013-09-25 Tom Tromey <tromey@redhat.com>
6746
6747 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
6748 to add -DENABLE_SHARED_LIBGCC.
6749 (gcc.o): Don't use subshell.
6750
6751 2013-09-25 Tom Tromey <tromey@redhat.com>
6752
6753 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
6754 * configure.ac: Don't invoke AM_PROG_CC_C_O.
6755 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
6756 * configure, config.in: Rebuild.
6757
6758 2013-09-25 Tom Tromey <tromey@redhat.com>
6759
6760 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
6761 (COMPILE, POSTCOMPILE): New variables.
6762 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
6763 (DEPFILES): New variable.
6764 Include ".Po" files.
6765 * configure.ac: Add checks for dependency checking.
6766 * configure, aclocal.m4: Regenerate.
6767
6768 2013-09-25 Tom Tromey <tromey@redhat.com>
6769
6770 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
6771 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
6772
6773 2013-09-25 Tom Tromey <tromey@redhat.com>
6774
6775 * Makefile.in (generated_files): Add options.h,
6776 target-hooks-def.h, insn-opinit.h,
6777 common/common-target-hooks-def.h, pass-instances.def,
6778 c-family/c-target-hooks-def.h.
6779
6780 2013-09-25 Jeff Law <law@redhat.com>
6781
6782 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
6783 than foo[foo.length () - 1] to access last member in a vec.
6784 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
6785
6786 2013-09-25 Richard Biener <rguenther@suse.de>
6787
6788 PR middle-end/58521
6789 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
6790
6791 2013-09-25 Jan Hubicka <jh@suse.cz>
6792
6793 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
6794 test.
6795
6796 2013-09-25 Marek Polacek <polacek@redhat.com>
6797
6798 PR sanitizer/58420
6799 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
6800 when determining the type name.
6801
6802 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
6803
6804 * config/sh/sh.md: Fix formatting.
6805
6806 2013-09-24 Xinliang David Li <davidxl@google.com>
6807
6808 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
6809 max peel iterations parameter.
6810 * param.def: New parameter.
6811 * doc/invoke.texi: Document New parameter.
6812
6813 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
6814
6815 * gimple-pretty-print.c: Various whitespace tweaks.
6816 * tree-core.h: Likewise.
6817 * tree-pretty-print.c: Likewise.
6818 * tree-ssa-alias.c: Likewise.
6819 * tree-ssa-copy.c: Likewise.
6820 * tree-ssanames.c: Likewise.
6821 * tree-ssanames.h: Likewise.
6822 * tree-vrp.c: Likewise.
6823
6824 2013-09-24 Alan Modra <amodra@gmail.com>
6825
6826 PR middle-end/57134
6827 PR middle-end/57586
6828 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
6829 for output operands that disallow regs. Don't use EXPAND_WRITE on
6830 inout operands.
6831
6832 2013-09-24 Richard Biener <rguenther@suse.de>
6833
6834 PR middle-end/58513
6835 * tree.c (reference_alias_ptr_type): Move ...
6836 * alias.c (reference_alias_ptr_type): ... here and implement
6837 in terms of the new reference_alias_ptr_type_1.
6838 (ref_all_alias_ptr_type_p): New helper.
6839 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
6840 use ref_all_alias_ptr_type_p.
6841 (get_deref_alias_set): Add flag_strict_aliasing check here.
6842 (reference_alias_ptr_type_1): New function, split out from ...
6843 (get_alias_set): ... here.
6844 (alias_ptr_types_compatible_p): New function.
6845 * alias.h (reference_alias_ptr_type): Declare.
6846 (alias_ptr_types_compatible_p): Likewise.
6847 * tree.h (reference_alias_ptr_type): Remove.
6848 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
6849 to compare MEM_REF alias types.
6850
6851 2013-09-24 Richard Biener <rguenther@suse.de>
6852
6853 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
6854
6855 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
6856
6857 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
6858 reload helper function arrays into a single array reg_addr.
6859 (reload_fpr_gpr): Likewise.
6860 (reload_gpr_vsx): Likewise.
6861 (reload_vsx_gpr): Likewise.
6862 (struct rs6000_reg_addr): Likewise.
6863 (reg_addr): Likewise.
6864 (rs6000_debug_reg_global): Change rs6000_vector_reload,
6865 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
6866 (rs6000_init_hard_regno_mode_ok): Likewise.
6867 (rs6000_secondary_reload_direct_move): Likewise.
6868 (rs6000_secondary_reload): Likewise.
6869
6870 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
6871 constraints: wu, ww, and wy. Repurpose wv constraint added during
6872 power8 changes. Put wg constraint in alphabetical order.
6873
6874 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
6875 for future work to add ISA 2.07 VSX single precision support.
6876 (-mvsx-scalar-double): Change default from -1 to 1, update
6877 documentation comment.
6878 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
6879 (-mupper-regs-df): New debug switch to control whether DF values
6880 can go in the traditional Altivec registers.
6881 (-mupper-regs-sf): New debug switch to control whether SF values
6882 can go in the traditional Altivec registers.
6883
6884 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
6885 and wy constraints.
6886 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
6887 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
6888 Add new constraints, wu/ww/wy. Repurpose wv constraint.
6889 (rs6000_debug_legitimate_address_p): Print if we are running
6890 before, during, or after reload.
6891 (rs6000_secondary_reload): Add a comment.
6892 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
6893
6894 * config/rs6000/constraints.md (wa constraint): Sort w<x>
6895 constraints. Update documentation string.
6896 (wd constraint): Likewise.
6897 (wf constraint): Likewise.
6898 (wg constraint): Likewise.
6899 (wn constraint): Likewise.
6900 (ws constraint): Likewise.
6901 (wt constraint): Likewise.
6902 (wx constraint): Likewise.
6903 (wz constraint): Likewise.
6904 (wu constraint): New constraint for ISA 2.07 SFmode scalar
6905 instructions.
6906 (ww constraint): Likewise.
6907 (wy constraint): Likewise.
6908 (wv constraint): Repurpose ISA 2.07 constraint that we did not use
6909 in the previous submissions.
6910 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
6911
6912 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
6913
6914 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
6915 (REG_BR_PROB): Say that the probability is stored in an int_list.
6916 * reg-notes.def: Update commentary to mention INT_LIST.
6917 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
6918 (INT_LIST): New rtx.
6919 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
6920 * rtlanal.c (int_reg_note_p): New function.
6921 (alloc_reg_note): Assert that the note does not have an int argument.
6922 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
6923 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
6924 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
6925 rather than an INSN_LIST. Handle INT_LIST.
6926 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
6927 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
6928 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
6929 Manipulate them as ints rather than rtxes.
6930 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
6931 * regmove.c (copy_src_to_dest): Likewise.
6932 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
6933
6934 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
6935 into the cases that need it.
6936 * config/arm/arm.c (arm_unwind_emit): Likewise.
6937
6938 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
6939 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
6940 * loop-doloop.c (add_test, doloop_modify): Likewise.
6941 * loop-unswitch.c (compare_and_jump_seq): Likewise.
6942 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
6943 * predict.c (combine_predictions_for_insn): Likewise.
6944 * print-rtl.c (print_rtx): Handle INT_LIST.
6945 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
6946 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
6947 * config/arm/arm.c (emit_unlikely_jump): Likewise.
6948 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
6949 (ix86_split_fp_branch, predict_jump): Likewise.
6950 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
6951 * config/sh/sh.c (expand_cbranchsi4): Likewise.
6952 * config/spu/spu.c (ea_load_store_inline): Likewise.
6953
6954 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
6955 value of a REG_BR_PROB note.
6956 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
6957 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
6958 * emit-rtl.c (try_split): Likewise.
6959 * predict.c (br_prob_note_reliable_p): Likewise.
6960 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
6961 * reorg.c (mostly_true_jump): Likewise.
6962 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
6963 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
6964 * config/i386/i386.c (ix86_print_operand): Likewise.
6965 * config/ia64/ia64.c (ia64_print_operand): Likewise.
6966 * config/mmix/mmix.c (mmix_print_operand): Likewise.
6967 * config/rs6000/rs6000.c (output_cbranch): Likewise.
6968 * config/s390/s390.c (s390_expand_tbegin): Likewise.
6969 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
6970 * config/sparc/sparc.c (output_cbranch): Likewise.
6971 * config/spu/spu.c (get_branch_target): Likewise.
6972 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
6973 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
6974
6975 2013-09-23 Jan Hubicka <jh@suse.cz>
6976
6977 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
6978 for ipa-devirt.
6979 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
6980 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
6981 of external calls
6982 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
6983 (gimple_fold_call): Dump inconsistent devirtualizations; add
6984 sanity check for type based devirtualizations.
6985 * ipa-prop.c: Include ipa-utils.h
6986 (ipa_intraprocedural_devirtualization): Add sanity check.
6987 (try_make_edge_direct_virtual_call): Likewise.
6988
6989 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
6990
6991 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
6992 assignment statements.
6993
6994 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
6995
6996 * gimple-pretty-print.c (dump_ssaname_info): New function.
6997 (dump_gimple_phi): Call it.
6998 (pp_gimple_stmt_1): Likewise.
6999 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
7000 (range_info_def): Declare.
7001 * tree-pretty-print.c (pp_double_int): New function.
7002 (dump_generic_node): Call it.
7003 * tree-pretty-print.h (pp_double_int): Declare.
7004 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
7005 * tree-ssanames.h (range_info_def): New structure.
7006 (value_range_type): Move definition here.
7007 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
7008 Declare.
7009 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
7010 initialization.
7011 (set_range_info): New function.
7012 (get_range_info): Likewise.
7013 (duplicate_ssa_name_range_info): Likewise.
7014 (duplicate_ssa_name_fn): Check pointer type and call
7015 duplicate_ssa_name_range_info.
7016 * tree-ssa-copy.c (fini_copy_prop): Likewise.
7017 * tree-vrp.c (value_range_type): Remove definition, now in
7018 tree-ssanames.h.
7019 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
7020 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
7021 (SSA_NAME_RANGE_INFO): New macro.
7022
7023 2013-09-23 Richard Biener <rguenther@suse.de>
7024
7025 PR tree-optimization/58464
7026 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7027 (phi_trans_add): Change to add conditionally on being not present.
7028 (phi_translate_1): Remove recursion detection here.
7029 (phi_translate): Pre-seed the cache with NULL to catch
7030 recursion here in a more generic way.
7031 (bitmap_find_leader): Adjust comment.
7032 (get_representative_for): Dump value-numbers.
7033 (create_expression_by_pieces): Likewise.
7034 (insert_into_preds_of_block): Likewise.
7035
7036 2013-09-23 Christian Bruel <christian.bruel@st.com>
7037
7038 PR target/58475
7039 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
7040 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
7041
7042 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
7043
7044 Revert r202780:
7045 2013-09-20 Renlin Li <renlin.li@arm.com>
7046
7047 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
7048 plus_constant.
7049 (aarch64_expand_epilogue): Likewise.
7050 (aarch64_legitimize_reload_address): Likewise.
7051
7052 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
7053
7054 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
7055 NULL_TREE before pushing them onto the vector. Likewise for labels.
7056
7057 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
7058
7059 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
7060 comparison operators when -fno-trapping-math is in effect.
7061 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
7062 comparison operators in TFmode and assert that unsupported operators
7063 cannot reach here.
7064 (ia64_print_operand): Likewise.
7065
7066 2013-09-21 Jan Hubicka <jh@suse.cz>
7067
7068 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
7069 (sse_typeless_stores): Enable for core
7070 (sse_load0_by_pxor): Likewise.
7071 (four_jump_limit): Disable for core.
7072 (pad_returns): Likewise.
7073 (avoid_vector_decode): Likewise.
7074 (fuse_cmp_and_branch): Enable for cores.
7075 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
7076
7077 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7078
7079 PR middle-end/56791
7080 * config/pa/pa.c (pa_option_override): Disable auto increment and
7081 decrement instructions until reload is completed.
7082
7083 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
7084 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
7085 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
7086
7087 2013-09-20 DJ Delorie <dj@redhat.com>
7088 Nick Clifton <nickc@redhat.com>
7089
7090 * config/rl78/rl78.c: Various whitespace and comment tweaks.
7091 (need_to_save): Save bank 0 on interrupts.
7092 (characterize_address): Strip far address wrappers.
7093 (rl78_as_legitimate_address): Likewise.
7094 (transcode_memory_rtx): Likewise.
7095 (rl78_peep_movhi_p): Disable this peephole after devirt.
7096 (rl78_propogate_register_origins): Forget all origins when a
7097 CLOBBER is seen.
7098 * config/rl78/rl78-virt.md: Various whitespace tweaks.
7099 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
7100 peephole2's.
7101 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
7102 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
7103 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
7104 relocating.
7105 * config/rl78/constraints.md: Various whitespace and paren tweaks.
7106
7107 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
7108
7109 * config/pa/pa.md: In "scc" insn patterns, change output template to
7110 handle const0_rtx in reg_or_0_operand operands.
7111
7112 2013-09-20 Martin Husemann <martin@NetBSD.org>
7113
7114 PR target/56875
7115 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
7116 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
7117
7118 2013-09-20 Richard Biener <rguenther@suse.de>
7119
7120 PR middle-end/58484
7121 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
7122 remembering SSA name version and block index.
7123 (new_scev_info_str): Adjust.
7124 (hash_scev_info): Likewise. Also hash the block index.
7125 (eq_scev_info): Adjust.
7126 (find_var_scev_info): Likewise.
7127 (struct instantiate_cache_entry): Remove.
7128 (struct instantiate_cache_type): Use a htab to map name, block
7129 to chrec.
7130 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7131 (get_instantiated_value_entry): Likewise.
7132 (hash_idx_scev_info, eq_idx_scev_info): New functions.
7133 (instantiate_scev_name): Adjust.
7134
7135 2013-09-20 Jeff Law <law@redhat.com>
7136
7137 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
7138
7139 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
7140
7141 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
7142 Call aarch64_simd_expand_args to update op[argc].
7143
7144 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7145
7146 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
7147 plugin argument.
7148
7149 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
7150
7151 * gengtype.c (file_rules): Added rule for *.cc files.
7152 (get_output_file_with_visibility): Give fatal message when no
7153 rules found.
7154
7155 2013-09-20 Renlin Li <renlin.li@arm.com>
7156
7157 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
7158 (aarch64_expand_epilogue): Likewise.
7159 (aarch64_legitimize_reload_address): Likewise.
7160
7161 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
7162
7163 PR middle-end/57748
7164 * expr.c (expand_assignment): Remove misalignp code path.
7165
7166 2013-09-20 Marek Polacek <polacek@redhat.com>
7167
7168 PR sanitizer/58413
7169 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
7170 TYPE_PRECISION. Add asserts.
7171
7172 2013-09-20 Richard Biener <rguenther@suse.de>
7173
7174 PR tree-optimization/58453
7175 * tree-loop-distribution.c (distribute_loop): Apply the cost
7176 model for -ftree-loop-distribute-patterns, too.
7177
7178 2013-09-20 Richard Biener <rguenther@suse.de>
7179
7180 PR middle-end/58473
7181 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
7182 make type comparison less strict.
7183
7184 2013-09-20 Alan Modra <amodra@gmail.com>
7185
7186 * configure: Regenerate.
7187 * aclocal.m4: Regenerate.
7188
7189 2013-09-20 Marek Polacek <polacek@redhat.com>
7190
7191 PR other/58467
7192 * doc/extend.texi: Document that attribute used is meant to be used
7193 on variables with static storage duration.
7194
7195 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7196
7197 PR tree-optimization/58472
7198 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
7199 simd_lane_access set inv_p = false.
7200 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
7201 the simduid magic VAR_DECL.
7202
7203 2013-09-19 Jan Hubicka <jh@suse.cz>
7204
7205 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
7206
7207 2013-09-17 Jeff Law <law@redhat.com>
7208
7209 * tree-ssa-dom.c (record_temporary_equivalences): New function
7210 split out of dom_opt_dom_walker::after_dom_children.
7211 (dom_opt_dom_walker::thread_across_edge): Move common code
7212 in here from dom_opt_dom_walker::after_dom_children.
7213 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
7214
7215 2013-09-19 Jan Hubicka <jh@suse.cz>
7216
7217 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
7218 (TARGET_GENERIC): Use PROCESOR_GENERIC
7219 (enum processor_type): Unify generic32 and 64.
7220 * i386.md (cpu): Likewise.
7221 * x86-tune.def (use_leave): Enable for generic32.
7222 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
7223 * athlon.md: Change generic64 to generic in all occurences.
7224 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
7225 (ix86_target_macros_internal): Likewise.
7226 * driver-i386.c (host_detect_local_cpu): Likewise.
7227 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
7228 to ..
7229 (generic_memcpy, generic_memset, generic_cost): This one.
7230 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
7231 (m_GENERIC32, m_GENERIC64): Remove.
7232 (m_GENERIC): Turn into one flag.
7233 (processor_target): Unify generic tunnings.
7234 (ix86_option_override_internal): Replace generic32/64 by generic.
7235 (ix86_issue_rate): Likewise.
7236 (ix86_adjust_cost): Likewise.
7237
7238 2013-09-19 Jan Hubicka <jh@suse.cz>
7239
7240 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
7241 of speculative flag.
7242
7243 2013-09-19 Jakub Jelinek <jakub@redhat.com>
7244
7245 * omp-low.c (expand_omp_sections): Always pass len - 1 to
7246 GOMP_sections_start, even if !exit_reachable.
7247
7248 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7249
7250 * lra-constraints.c (need_for_all_save_p): Use macro
7251 HARD_REGNO_CALL_PART_CLOBBERED.
7252 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
7253 set up pseudo conflict hard regs.
7254
7255 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
7256
7257 PR target/58452
7258 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
7259 operands.
7260
7261 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
7262
7263 PR rtl-optimization/58438
7264 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
7265 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
7266 from previous subpasses.
7267
7268 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
7269
7270 * arm.c (arm_get_frame_offsets): Validate architecture supports
7271 LDRD/STRD before accepting the tuning preference.
7272 (arm_expand_prologue): Likewise.
7273 (arm_expand_epilogue): Likewise.
7274
7275 2013-09-18 Richard Biener <rguenther@suse.de>
7276
7277 PR tree-optimization/58417
7278 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
7279 have chrecs with symbols defined in the loop as operands.
7280 (chrec_fold_multiply): Likewise.
7281 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
7282 parameters before folding binary operations.
7283 (struct instantiate_cache_entry_hasher): Remove.
7284 (struct instantiate_cache_type): Use a pointer-map.
7285 (instantiate_cache_type::instantiate_cache_type): New function.
7286 (instantiate_cache_type::get): Likewise.
7287 (instantiate_cache_type::set): Likewise.
7288 (instantiate_cache_type::~instantiate_cache_type): Adjust.
7289 (get_instantiated_value_entry): Likewise.
7290 (global_cache): New global.
7291 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
7292 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
7293 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
7294 (instantiate_scev_name): Adjust.
7295 (instantiate_scev): Construct global instead of local cache.
7296 (resolve_mixers): Likewise.
7297
7298 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
7299 Paolo Carlini <paolo.carlini@oracle.com>
7300
7301 PR c++/58458
7302 * doc/implement-cxx.texi: Fix references to the C++ standards.
7303
7304 2013-09-18 Jakub Jelinek <jakub@redhat.com>
7305
7306 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
7307 * tree-vect-data-refs.c (vect_analyze_data_refs): For
7308 simd_lane_access drs, update also DR_ALIGNED_TO.
7309
7310 2013-09-18 Marek Polacek <polacek@redhat.com>
7311
7312 PR sanitizer/58411
7313 * doc/extend.texi: Document no_sanitize_undefined attribute.
7314 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
7315 no_sanitize_undefined attribute.
7316
7317 2013-09-18 Nick Clifton <nickc@redhat.com>
7318
7319 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
7320 (ASM_DECLARE_FUNCTION_NAME): Define.
7321
7322 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
7323
7324 * compare-elim.c (find_comparison_dom_walker): New class
7325 (find_comparisons_in_bb): Rename to
7326 find_comparison_dom_walker::before_dom_children
7327 (find_comparisons): Adjust
7328 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
7329 adjust.
7330 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
7331 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
7332 (init_walk_dominator_tree): Remove declaration.
7333 (fini_walk_dominator_tree): Remove declaration.
7334 * fwprop.c (single_def_use_dom_walker): New class
7335 (single_def_use_enter_block): Convert to
7336 single_def_use_dom_walker::before_dom_children.
7337 (single_def_use_leave_block): Convert to
7338 single_def_use_dom_walker::after_dom_children.
7339 (build_single_def_use_links): Adjust.
7340 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
7341 class.
7342 (find_candidates_in_block): Convert to
7343 find_candidates_dom_walker::before_dom_children.
7344 (execute_strength_reduction): Adjust.
7345 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
7346 (sese_dom_walker): New class.
7347 (sese_dom_walker::sese_dom_walker): New constructor.
7348 (sese_dom_walker::~sese_dom_walker): New destructor.
7349 (build_sese_conditions_before): Convert to
7350 sese_dom_walker::before_dom_children.
7351 (build_sese_conditions_after): Convert to
7352 sese_dom_walker::after_dom_children.
7353 (build_poly_scop): Adjust
7354 * tree-into-ssa.c (rewrite_dom_walker): New class
7355 (rewrite_enter_block): Convert to
7356 rewrite_dom_walker::before_dom_children.
7357 (rewrite_leave_block): Convert to
7358 rewrite_dom_walker::after_dom_children.
7359 (rewrite_update_dom_walker): New class.
7360 (rewrite_update_enter_block): Convert to
7361 rewrite_update_dom_walker::before_dom_children.
7362 (rewrite_update_leave_block): Convert to
7363 rewrite_update_dom_walker::after_dom_children.
7364 (rewrite_blocks, rewrite_into_ssa): Adjust.
7365 (mark_def_dom_walker): New class.
7366 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
7367 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
7368 (mark_def_sites_blocks): Convert to
7369 mark_def_dom_walker::before_dom_children.
7370 (mark_def_site_blocks): Remove.
7371 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
7372 (tree_ssa_dominator_optimize): Adjust.
7373 (dom_thread_across_edge): Convert to method
7374 dom_opt_dom_walker::thread_across_edge.
7375 (dom_opt_enter_block): Convert to member function
7376 dom_opt_dom_walker::before_dom_children.
7377 (dom_opt_leave_block): Convert to member function
7378 dom_opt_dom_walker::after_dom_children.
7379 * tree-ssa-dse.c (dse_dom_walker): New class.
7380 (dse_enter_block): Convert to member function
7381 dse_dom_walker::before_dom_children.
7382 (tree_ssa_dse): Adjust.
7383 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
7384 (determine_invariantness_stmt): Convert to method
7385 invariantness_dom_walker::before_dom_children.
7386 (determine_invariantness): Remove
7387 (move_computations_dom_walker): New class.
7388 (move_computations_stmt): Convert to method
7389 move_computations_dom_walker::before_dom_children.
7390 (move_computations, tree_ssa_lim): Adjust.
7391 * tree-ssa-phiopt.c (nontrapping_dom_walker): New class.
7392 (nt_init_block): Convert to method
7393 notrappping_dom_walker::before_dom_children.
7394 (nt_fini_block): Convert to method
7395 method nontrapping_dom_walker::after_dom_children.
7396 (get_non_trapping): Adjust.
7397 * tree-ssa-pre.c (eliminate_dom_walker): New class.
7398 (eliminate_bb): Convert to method
7399 eliminate_dom_walker::before_dom_children.
7400 (eliminate_leave_block): Convert to method
7401 eliminate_dom_walker::after_dom_children.
7402 (eliminate): Adjust.
7403 * tree-ssa-strlen.c (strlen_dom_walker): New class.
7404 (strlen_enter_block): Convert to method
7405 strlen_dom_walker::before_dom_children.
7406 (strlen_leave_block): Convert to method
7407 method strlen_dom_walker::after_dom_children.
7408 (tree_ssa_strlen): Adjust.
7409 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
7410 (tree_ssa_uncprop): Adjust.
7411 (uncprop_leave_block): Convert to method
7412 uncprop_dom_walker::after_dom_children.
7413 (uncprop_leave_block): Convert to method
7414 uncprop_dom_walker::before_dom_children.
7415
7416 2013-09-18 Bin Cheng <bin.cheng@arm.com>
7417
7418 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
7419 branch in same basic block. Check both src and dest of the move insn.
7420
7421 2013-09-17 Nick Clifton <nickc@redhat.com>
7422
7423 * config/rl78/rl78-real.md (bf): New pattern.
7424 (bt): New pattern.
7425 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
7426 (rl78_print_operand): Do not put a # before a %B.
7427 * config/rl78/rl78.opt: Tweak doc strings.
7428
7429 2013-09-17 DJ Delorie <dj@redhat.com>
7430
7431 * config/rl78/constraints.md (Wcv): Allow up to $r31.
7432 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
7433 (rl78_option_override): Likewise, if -mallregs.
7434 (is_virtual_register): Likewise.
7435 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
7436 (REGNO_OK_FOR_BASE_P): Likewise.
7437 * config/rl78/rl78.opt (-mallregs): New.
7438
7439 2013-09-17 Nick Clifton <nickc@redhat.com>
7440
7441 * config/rl78/rl78.c (need_to_save): Change return type to bool.
7442 For interrupt functions: save all call clobbered registers if the
7443 interrupt handler is not a leaf function.
7444 (rl78_expand_prologue): Always recompute the frame information.
7445 For interrupt functions: only select bank 0 if one of the bank 0
7446 registers is going to be psuhed.
7447
7448 2013-09-17 DJ Delorie <dj@redhat.com>
7449
7450 * config/rl78/constraints.md: For each W* constraint, rename to C*
7451 and create a W* constraint that checks for an optional ES: prefix
7452 pattern also.
7453 * config/rl78/rl78.md (UNS_ES_ADDR): New.
7454 (es_addr): New. Used to wrap far addresses.
7455 * config/rl78/rl78-protos.h (rl78_es_addr): New.
7456 (rl78_es_base): New.
7457 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
7458 wrapped far addresses.
7459 (rl78_print_operand_1): Unwrap far addresses before processing.
7460 (rl78_lo16): Wrap far addresses in unspecs.
7461 (rl78_es_addr): New.
7462 (rl78_es_base): New.
7463 (insn_ok_now): Check for not-yet-wrapped far addresses.
7464 (transcode_memory_rtx): Properly re-wrap far addresses.
7465
7466 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
7467
7468 * config/sparc/t-rtems: Add leon3 multilibs.
7469
7470 2013-09-17 Cong Hou <congh@google.com>
7471
7472 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
7473 when checking the dot production pattern. The type of rhs operand
7474 of multiply is now checked correctly.
7475
7476 2013-09-17 Jeff Law <law@redhat.com>
7477
7478 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
7479 edge implied equivalences into successor phis.
7480 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
7481 here from tree-ssa-threadedge.c.
7482 (mark_threaded_blocks): When threading through a joiner, if both
7483 successors of the joiner's clone reach the same block, verify the
7484 PHI arguments are equal. If not, cancel the jump threading request.
7485 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
7486 tree-ssa-threadupdate.c
7487 (thread_across_edge): Don't check PHI argument equality when
7488 threading through joiner block here.
7489
7490 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
7491
7492 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
7493 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7494 (warn_uninit, warn_uninitialized_vars,
7495 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7496 Move to tree-ssa-uninit.c.
7497 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7498 (has_undefined_value_p): New. Pass dependant parts of
7499 ssa_undefined_value_p.
7500 (uninit_undefined_value_p): Use has_undefined_value_p.
7501 (warn_uninit, warn_uninitialized_vars,
7502 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
7503 Move from tree-ssa.c.
7504 * tree-ssa.h: Adjust prototypes.
7505
7506 2013-09-17 Jan Hubicka <jh@suse.cz>
7507
7508 PR middle-end/58332
7509 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
7510 * ipa-inline.c (can_inline_edge_p): Do not downgrade
7511 FUNCTION_NOT_OPTIMIZED.
7512 * ipa-inline-analysis.c (compute_inline_parameters): Function
7513 not optimized is not inlinable unless it is alwaysinline.
7514 (inline_analyze_function): Force calls in not optimized
7515 function not inlinable.
7516
7517 2013-09-17 Jan Hubicka <jh@suse.cz>
7518
7519 PR middle-end/58329
7520 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
7521 to return NULL.
7522 * ipa.c (function_and_variable_visibility): Likewise.
7523 * ipa-profile.c (ipa_profile): Likewise.
7524
7525 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
7526
7527 PR ipa/58398
7528 * cgraph.c (cgraph_function_body_availability): Check for ifunc
7529 attribute, and don't inline the resolver in this case.
7530
7531 2013-09-17 Teresa Johnson <tejohnson@google.com>
7532
7533 * coverage.c (get_coverage_counts): Add missing newline.
7534
7535 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7536
7537 PR tree-optimization/58088
7538 * fold-const.c (mask_with_trailing_zeros): New function.
7539 (fold_binary_loc): Make sure we don't recurse infinitely
7540 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
7541 Use mask_with_trailing_zeros where appropriate.
7542
7543 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
7544
7545 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
7546 of 'prev' var to get better distance estimation.
7547
7548 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
7549
7550 * tree-inline.h (struct copy_body_data): Add transform_parameter.
7551 * tree-inline.c (is_parameter_of): New predicate.
7552 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
7553 a parameter has been remapped.
7554 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
7555 (optimize_inline_calls): Initialize transform_parameter.
7556 (copy_gimple_seq_and_replace_locals): Likewise.
7557 (tree_function_versioning): Likewise.
7558 (maybe_inline_call_in_expr): Likewise.
7559
7560 2013-09-17 Nick Clifton <nickc@redhat.com>
7561
7562 * config/msp430/msp430-protos.h: Add prototypes for new functions.
7563 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
7564 interrupt handlers.
7565 (is_attr_func): New function.
7566 (msp430_is_interrupt_func): New function.
7567 (is_naked_func): New function.
7568 (is_reentrant_func): New function.
7569 (is_critical_func): New function.
7570 (msp430_start_function): Add annotations for function attributes.
7571 (msp430_attr): New function.
7572 (msp430_attribute_table): New.
7573 (msp430_function_section): New function.
7574 (TARGET_ASM_FUNCTION_SECTION): Define.
7575 (msp430_builtin): New enum.
7576 (msp430_init_builtins): New function.
7577 (msp430_builtin_devl): New function.
7578 (msp430_expand_builtin): New function.
7579 (TARGET_INIT_BUILTINS): Define.
7580 (TARGET_EXPAND_BUILTINS): Define.
7581 (TARGET_BUILTIN_DECL): Define.
7582 (msp430_expand_prologue): Add support for naked, interrupt,
7583 critical and reentrant functions.
7584 (msp430_expand_epilogue): Likewise.
7585 (msp430_print_operand): Handle 'O' character.
7586 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
7587 NO_TRAMPOLINES.
7588 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
7589 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
7590 (pushm): Use a 'n' rather than an 'i' constraint.
7591 (msp_return): Add generation of the interrupt return instruction.
7592 (disable_interrupts): New pattern.
7593 (enable_interrupts): New pattern.
7594 (push_intr_state): New pattern.
7595 (pop_intr_state): New pattern.
7596 (bic_SR): New pattern.
7597 (bis_SR): New pattern.
7598 * doc/extend.texi: Document MSP430 function attributes and builtin
7599 functions.
7600
7601 2013-09-17 Richard Biener <rguenther@suse.de>
7602
7603 PR tree-optimization/58432
7604 * tree-loop-distribution.c (tree_loop_distribution): Also
7605 scan PHIs for outside loop uses and seed a partition from them.
7606
7607 2013-09-17 Bin Cheng <bin.cheng@arm.com>
7608
7609 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
7610 (restructure_reference): Call backtrace_base_for_ref.
7611
7612 2013-09-17 Alan Modra <amodra@gmail.com>
7613
7614 PR target/57589
7615 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
7616 patch.
7617
7618 2013-09-16 DJ Delorie <dj@redhat.com>
7619
7620 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
7621 vregs location for RL78/G10.
7622 (rl78_expand_prologue): Avoid SEL on G10.
7623 (rl78_expand_epilogue): Likewise.
7624 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
7625 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
7626 __RL78_G10__ when appropriate.
7627 (ASM_SPEC): Pass -mg10 along to the assembler.
7628 * config/rl78/rl78.md (sel_rb): Disable for G10.
7629 * config/rl78/rl78.opt: Add -mg10 option.
7630 * config/rl78/t-rl78: Add -mg10 multilib.
7631
7632 2013-09-16 Xinliang David Li <davidxl@google.com>
7633
7634 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
7635 * omp-low.c (omp_max_vf): Ditto.
7636 (expand_omp_simd): Ditto.
7637 * tree-vectorizer.c (vectorize_loops): Ditto.
7638 (gate_vect_slp): Ditto.
7639 (gate_increase_alignment): Ditto.
7640 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
7641 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
7642 (gate_tree_vectorize): Name change.
7643 (tree_vectorize): Ditto.
7644 (pass_vectorize::gate): Call new function.
7645 (pass_vectorize::execute): Ditto.
7646 * opts.c: O3 default setting change.
7647 (finish_options): Check new flag.
7648 * doc/invoke.texi: Document new flags.
7649 * common.opt: New flags.
7650
7651 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
7652
7653 * doc/tm.texi.in (Cond Exec Macros): Remove node.
7654 (Condition Code): Don't reference it.
7655 * doc/tm.texi: Regenerate.
7656
7657 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
7658
7659 PR middle-end/58418
7660 * lra-constraints.c (undo_optional_reloads): Consider all optional
7661 reload even if it did not get a hard reg.
7662
7663 2013-09-16 Teresa Johnson <tejohnson@google.com>
7664
7665 * dumpfile.c (dump_loc): Remove newline emission.
7666 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
7667 emission to dump_printf_loc calls where missing.
7668 (vect_mark_for_runtime_alias_test): Ditto.
7669 (vect_analyze_data_ref_dependence): Ditto.
7670 (vect_analyze_data_ref_dependences): Ditto.
7671 (vect_slp_analyze_data_ref_dependence): Ditto.
7672 (vect_slp_analyze_data_ref_dependences): Ditto.
7673 (vect_compute_data_ref_alignment): Ditto.
7674 (vect_update_misalignment_for_peel): Ditto.
7675 (vect_verify_datarefs_alignment): Ditto.
7676 (vector_alignment_reachable_p): Ditto.
7677 (vect_get_data_access_cost): Ditto.
7678 (vect_enhance_data_refs_alignment): Ditto.
7679 (vect_find_same_alignment_drs): Ditto.
7680 (vect_analyze_data_refs_alignment): Ditto.
7681 (vect_analyze_group_access): Ditto.
7682 (vect_analyze_data_ref_access): Ditto.
7683 (vect_analyze_data_ref_accesses): Ditto.
7684 (vect_prune_runtime_alias_test_list): Ditto.
7685 (vect_analyze_data_refs): Ditto.
7686 (vect_create_addr_base_for_vector_ref): Ditto.
7687 (vect_create_data_ref_ptr): Ditto.
7688 (vect_grouped_store_supported): Ditto.
7689 (vect_grouped_load_supported): Ditto.
7690 * value-prof.c (check_counter): Ditto.
7691 (check_ic_target): Ditto.
7692 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
7693 (vect_recog_widen_mult_pattern): Ditto.
7694 (vect_recog_widen_sum_pattern): Ditto.
7695 (vect_recog_over_widening_pattern): Ditto.
7696 (vect_recog_widen_shift_pattern): Ditto.
7697 (vect_recog_rotate_pattern): Ditto.
7698 (vect_recog_vector_vector_shift_pattern): Ditto.
7699 (vect_recog_divmod_pattern): Ditto.
7700 (vect_recog_mixed_size_cond_pattern): Ditto.
7701 (vect_recog_bool_pattern): Ditto.
7702 (vect_pattern_recog_1): Ditto.
7703 (vect_pattern_recog): Ditto.
7704 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
7705 (vect_is_simple_iv_evolution): Ditto.
7706 (vect_analyze_scalar_cycles_1): Ditto.
7707 (vect_get_loop_niters): Ditto.
7708 (vect_analyze_loop_1): Ditto.
7709 (vect_analyze_loop_form): Ditto.
7710 (vect_analyze_loop_operations): Ditto.
7711 (vect_analyze_loop_2): Ditto.
7712 (vect_analyze_loop): Ditto.
7713 (report_vect_op): Ditto.
7714 (vect_is_slp_reduction): Ditto.
7715 (vect_is_simple_reduction_1): Ditto.
7716 (vect_get_known_peeling_cost): Ditto.
7717 (vect_estimate_min_profitable_iters): Ditto.
7718 (vect_model_reduction_cost): Ditto.
7719 (vect_model_induction_cost): Ditto.
7720 (get_initial_def_for_induction): Ditto.
7721 (vect_create_epilog_for_reduction): Ditto.
7722 (vectorizable_reduction): Ditto.
7723 (vectorizable_induction): Ditto.
7724 (vectorizable_live_operation): Ditto.
7725 (vect_loop_kill_debug_uses): Ditto.
7726 (vect_transform_loop): Ditto.
7727 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
7728 (vect_stmt_relevant_p): Ditto.
7729 (process_use): Ditto.
7730 (vect_mark_stmts_to_be_vectorized): Ditto.
7731 (vect_model_simple_cost): Ditto.
7732 (vect_model_promotion_demotion_cost): Ditto.
7733 (vect_model_store_cost): Ditto.
7734 (vect_get_store_cost): Ditto.
7735 (vect_model_load_cost): Ditto.
7736 (vect_get_load_cost): Ditto.
7737 (vect_init_vector_1): Ditto.
7738 (vect_get_vec_def_for_operand): Ditto.
7739 (vect_finish_stmt_generation): Ditto.
7740 (vectorizable_call): Ditto.
7741 (vectorizable_conversion): Ditto.
7742 (vectorizable_assignment): Ditto.
7743 (vectorizable_shift): Ditto.
7744 (vectorizable_operation): Ditto.
7745 (vectorizable_store): Ditto.
7746 (vectorizable_load): Ditto.
7747 (vectorizable_condition): Ditto.
7748 (vect_analyze_stmt): Ditto.
7749 (vect_transform_stmt): Ditto.
7750 (vect_is_simple_use): Ditto.
7751 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
7752 (vect_can_advance_ivs_p): Ditto.
7753 (vect_update_ivs_after_vectorizer): Ditto.
7754 (vect_do_peeling_for_loop_bound): Ditto.
7755 (vect_gen_niters_for_prolog_loop): Ditto.
7756 (vect_update_inits_of_drs): Ditto.
7757 (vect_create_cond_for_alias_checks): Ditto.
7758 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
7759 (vect_build_slp_tree_1): Ditto.
7760 (vect_supported_load_permutation_p): Ditto.
7761 (vect_analyze_slp_instance): Ditto.
7762 (vect_analyze_slp): Ditto.
7763 (vect_make_slp_decision): Ditto.
7764 (vect_detect_hybrid_slp): Ditto.
7765 (vect_bb_vectorization_profitable_p): Ditto.
7766 (vect_slp_analyze_bb_1): Ditto.
7767 (vect_update_slp_costs_according_to_vf): Ditto.
7768 (vect_get_mask_element): Ditto.
7769 (vect_transform_slp_perm_load): Ditto.
7770 (vect_schedule_slp_instance): Ditto.
7771 (vect_schedule_slp): Ditto.
7772 (vect_slp_transform_bb): Ditto.
7773 * profile.c (read_profile_edge_counts): Ditto.
7774 (compute_branch_probabilities): Ditto.
7775 * coverage.c (get_coverage_counts): Ditto.
7776
7777 2013-09-16 Diego Novillo <dnovillo@google.com>
7778
7779 * tree-core.h: Add missing comment lines from refactoring of tree.h.
7780
7781 2013-09-16 Jan Hubicka <jh@suse.cz>
7782
7783 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7784 abstract functions; for static functions check the presence of body.
7785
7786 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7787
7788 * config/aarch64/aarch64-simd-builtins.def (fma): New.
7789 * config/aarch64/aarch64-simd.md
7790 (aarch64_mla_elt<mode>): New.
7791 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
7792 (aarch64_mls_elt<mode>): Likewise.
7793 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
7794 (aarch64_fma4_elt<mode>): Likewise.
7795 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
7796 (aarch64_fma4_elt_to_128v2df): Likewise.
7797 (aarch64_fma4_elt_to_64df): Likewise.
7798 (fnma<mode>4): Likewise.
7799 (aarch64_fnma4_elt<mode>): Likewise.
7800 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
7801 (aarch64_fnma4_elt_to_128v2df): Likewise.
7802 (aarch64_fnma4_elt_to_64df): Likewise.
7803 * config/aarch64/iterators.md (VDQSF): New.
7804 * config/aarch64/arm_neon.h
7805 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
7806 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
7807
7808 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7809
7810 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
7811 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
7812 (aarch64_mul3_elt_to_128df): Likewise.
7813 (aarch64_mul3_elt_to_64v2df): Likewise.
7814 * config/aarch64/iterators.md (VEL): Also handle DFmode.
7815 (VMUL): New.
7816 (VMUL_CHANGE_NLANES) Likewise.
7817 (h_con): Likewise.
7818 (f): Likewise.
7819 * config/aarch64/arm_neon.h
7820 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
7821
7822 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
7823
7824 * config/aarch64/arm_neon.h
7825 (vcvtx_high_f32_f64): Fix parameters.
7826
7827 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7828 Uros Bizjak <ubizjak@gmail.com>
7829
7830 * config/alpha.c: Include tree-ssanames.h.
7831
7832 2013-09-16 Richard Biener <rguenther@suse.de>
7833
7834 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
7835 (dot_rdg_1): Handle control_dd.
7836 (create_edge_for_control_dependence): New function.
7837 (create_rdg_edges): Add control dependences if asked for.
7838 (build_rdg): Likewise.
7839 (generate_loops_for_partition): If there are not necessary
7840 control stmts remove all their dependencies.
7841 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
7842 (distribute_loop): Pass on control dependences.
7843 (tree_loop_distribution): Compute control dependences and remove
7844 restriction on number of loop nodes.
7845
7846 2013-09-16 Jakub Jelinek <jakub@redhat.com>
7847
7848 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
7849 for internal calls.
7850
7851 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
7852
7853 * cse.c (try_const_anchors): Punt on CC modes.
7854
7855 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
7856
7857 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
7858
7859 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
7860
7861 PR target/58382
7862 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
7863 calls to word_mode.
7864
7865 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7866
7867 PR target/48094
7868 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
7869 seen.
7870 (darwin_objc1_section): Likewise.
7871 (darwin_file_end): Emit Image Info section when required.
7872
7873 2013-09-14 Jan Hubicka <jh@suse.cz>
7874
7875 * tree-into-ssa.c (gate_into_ssa): New.
7876 (pass_data_build_ssa): Use it.
7877 * cgraph.h (expand_thunk): Update prototype.
7878 * cgraphunit.c (analyze_function): Expand thunks early.
7879 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
7880 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
7881 set lowered flag; do not add new function.
7882 (assemble_thunks_and_aliases): Update.
7883 * tree-ssa.c (gate_init_datastructures): New gate.
7884 (pass_data_init_datastructures): Use it.
7885
7886 2013-09-14 Iain Sandoe <iain@codesourcery.com>
7887
7888 PR target/58269
7889 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
7890 xmm register set described in the psABI.
7891
7892 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
7893
7894 * dwarf2out.c (should_emit_struct_debug): Add check
7895 for type_decl variable is not NULL.
7896
7897 2013-09-13 Jacek Caban <jacek@codeweavers.com>
7898
7899 * config.gcc: Use new winnt-c.c target hooks
7900 * config/t-winnt: New file
7901 * config/winnt-c.c: New file
7902 * doc/tm.texi.in: Document new hook
7903 * doc/tm.texi: Regenerated
7904
7905 2013-09-13 Jan Hubicka <jh@suse.cz>
7906
7907 PR middle-end/58094
7908 * ipa-inline.c (check_callers): New function.
7909 (check_caller_edge): Remove.
7910 (want_inline_function_to_all_callers_p): Also permit alises that are
7911 called dirrectly.
7912 (inline_to_all_callers): Terminate the walk when devirtualization
7913 introduce new calls.
7914
7915 2013-09-13 Jan Hubicka <jh@suse.cz>
7916
7917 * ipa-inline-analysis.c (struct growth_data): Add node.
7918 (do_estimate_growth_1): Fix detection of recursion.
7919
7920 2013-09-13 Jakub Jelinek <jakub@redhat.com>
7921
7922 PR tree-optimization/58392
7923 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
7924 to avoid shadowing of outer loop variable. If
7925 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
7926 replace_by_duplicate_decl simduid of loops that have it set and
7927 set dest_cfun->has_simduid_loops and/or
7928 dest_cfun->has_force_vect_loops.
7929 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
7930 instead of maybe_lookup_decl.
7931 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
7932 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
7933 call. Copy over force_vect and copy and remap simduid. Set
7934 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
7935 (copy_cfg_body): Remove blocks_to_copy argument. Use
7936 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
7937 caller. Don't set cfun->has_simduid_loops and/or
7938 cfun->has_force_vect_loops here.
7939 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
7940 caller.
7941 (expand_call_inline, tree_function_versioning): Adjust copy_body
7942 callers.
7943
7944 2013-09-13 Martin Jambor <mjambor@suse.cz>
7945
7946 PR bootstrap/58388
7947 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
7948 the assert if the edge was a speculative one.
7949
7950 2013-09-13 Richard Biener <rguenther@suse.de>
7951
7952 * tree-data-ref.h (known_dependences_p): Move here ...
7953 * tree-loop-distribution.c (known_dependences_p): ... from here.
7954 (dump_rdg_component, debug_rdg_component): Remove.
7955 (dump_rdg): Adjust.
7956 (generate_loops_for_partition): Use gimple_uid instead of
7957 relying on matching stmt visit order.
7958 (rdg_build_partitions): Take starting stmt vector.
7959 (ldist_gen): Merge into ...
7960 (distribute_loop): ... this function. Do not compute starting
7961 vertices vector.
7962 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
7963
7964 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
7965
7966 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
7967 Set type attribute correctly. Set predicable_short_it attribute.
7968 (cmpsi_shiftsi): Remove %? from output template.
7969
7970 2013-09-13 Richard Biener <rguenther@suse.de>
7971
7972 * tree-loop-distribution.c (struct rdg_component,
7973 rdg_defs_used_in_other_loops_p, free_rdg_components,
7974 rdg_build_components): Remove.
7975 (stmts_from_loop): Do not record virtual PHIs.
7976 (generate_loops_for_partition): Skip virtual PHIs.
7977 (build_rdg_partition_for_component): Rename to ...
7978 (build_rdg_partition_for_vertex): ... this and adjust.
7979 (rdg_build_partitions): Take a vector of starting vertices
7980 instead of components. Remove unnecessary leftover handling.
7981 (ldist_gen): Do not build components or record other stores.
7982 (distribute_loop): Do not distribute loops containing stmts
7983 with side-effects.
7984
7985 2013-09-13 Christian Bruel <christian.bruel@st.com>
7986
7987 PR target/58314
7988 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
7989
7990 2013-09-13 Kai Tietz <ktietz@redhat.com>
7991
7992 * config.gcc: Separate cases for mingw and cygwin targets,
7993 and add 64-bit cygwin target case.
7994
7995 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
7996 dll-export inline-functions.
7997 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
7998
7999
8000 2013-09-13 Jeff Law <law@redhat.com>
8001
8002 PR middle-end/58387
8003 Revert:
8004 2013-09-06 Jeff Law <law@redhat.com>
8005
8006 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8007 edge implied equivalences into successor phis.
8008
8009 2013-09-12 DJ Delorie <dj@redhat.com>
8010
8011 * config/rl78/rl78-virt.md: Change from | to \; for asm line
8012 separators.
8013
8014 2013-09-12 Brooks Moses <bmoses@google.com>
8015
8016 PR driver/42955
8017 * Makefile.in: Do not install driver binaries in $(target)/bin.
8018
8019 2013-09-12 DJ Delorie <dj@redhat.com>
8020
8021 * config/rl78/rl78.opt (mrelax): New.
8022 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
8023 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
8024
8025 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
8026 between SP and FP.
8027 (rl78_expand_epilogue): Likewise.
8028
8029 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
8030
8031 PR middle-end/58335
8032 * lra-eliminations.c (remove_reg_equal_offset_note): New.
8033 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
8034 pointer elimination with using remove_reg_equal_offset_note.
8035
8036 2013-09-12 DJ Delorie <dj@redhat.com>
8037
8038 * config/msp430/: New port.
8039 * config.gcc (msp430): Added.
8040 * doc/invoke.texi: Document MSP430 options.
8041 * doc/install.texi: Document msp430-elf
8042 * doc/md.texi: Document msp430-elf
8043 * doc/contrib.texi: Document msp430-elf
8044
8045 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
8046 PSImode.
8047
8048 2013-09-12 Martin Jambor <mjambor@suse.cz>
8049
8050 PR ipa/58389
8051 * ipa-prop.c (remove_described_reference): Give up if the edge in the
8052 reference descriptor is NULL.
8053 (ipa_edge_removal_hook): If owning a reference descriptor, set its
8054 edge to NULL.
8055
8056 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8057
8058 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
8059 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
8060 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
8061 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
8062 make_temp_ssa_name): move to tree-ssanames.h
8063 * tree-ssa-alias.h: Move prototype.
8064 * tree-ssa.h: Include tree-ssanames.h.
8065 * tree-ssanames.c (FREE_SSANAMES): Move to here.
8066 * tree-ssanames.h: New. Move items from tree-flow*.h
8067 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
8068
8069 2013-09-12 Richard Biener <rguenther@suse.de>
8070
8071 PR tree-optimization/58404
8072 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
8073 propagate non-invariant addresses into dereferences wrapped
8074 in component references.
8075
8076 2013-09-12 Richard Biener <rguenther@suse.de>
8077
8078 PR tree-optimization/58402
8079 * passes.def: Move pass_late_warn_uninitialized later.
8080
8081 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
8082
8083 * tree-ssa.h: New. Move content from tree-flow.h and
8084 tree-flow-inline.h.
8085 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
8086 Move prototypes belonging to tree-ssa.c.
8087 * tree-flow-inline.h (redirect_edge_var_map_def,
8088 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
8089 tree-ssa.h.
8090 * gimple.h: Adjust prototypes.
8091 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
8092 to...
8093 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
8094 * tree.h: Move prototype to tree-ssa.h.
8095 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
8096 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
8097 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
8098 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
8099 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
8100 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
8101 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
8102 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
8103 graphite-blocking.c, graphite-clast-to-gimple.c,
8104 graphite-dependences.c, graphite-interchange.c,
8105 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
8106 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
8107 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
8108 ipa-reference.c, ipa-split.c, ipa-utils.c,
8109 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
8110 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
8111 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
8112 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
8113 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
8114 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
8115 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
8116 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
8117 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
8118 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
8119 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
8120 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
8121 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
8122 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
8123 value-prof.c, var-tracking.c,
8124 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
8125
8126 2013-09-12 Richard Biener <rguenther@suse.de>
8127
8128 PR tree-optimization/58396
8129 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
8130 (build_rdg): Take a loop-nest parameter, fix memleaks.
8131 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
8132
8133 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8134
8135 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
8136 for SLM.
8137
8138 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
8139
8140 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
8141 The FMA instruction names should have a 'v' prefix.
8142
8143 2013-09-12 Richard Biener <rguenther@suse.de>
8144
8145 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
8146 (dot_rdg): Use popen instead of system in optional code.
8147 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
8148 (already_processed_vertex_p): Adjust.
8149 (has_anti_or_output_dependence, predecessor_has_mem_write,
8150 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
8151 rdg_flag_uses): Remove.
8152 (rdg_flag_vertex): Simplify.
8153 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
8154 remove recursion.
8155 (build_rdg_partition_for_component): Process the first vertex
8156 of a component only.
8157 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
8158
8159 2013-09-12 Alan Modra <amodra@gmail.com>
8160
8161 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
8162
8163 2013-09-11 DJ Delorie <dj@redhat.com>
8164 Nick Clifton <nickc@redhat.com>
8165
8166 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
8167 (rl78_stack_based_mem): New.
8168 * config/rl78/constraints.md (Iv08): New.
8169 (Iv16): New.
8170 (Iv24): New.
8171 (Is09): New.
8172 (Is17): New.
8173 (Is25): New.
8174 (ISsi): New.
8175 (IShi): New.
8176 (ISqi): New.
8177 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
8178 (movhi): Likewise.
8179 (movsi): Change from expand to insn-and-split.
8180 (ashrsi3): Clobber AX.
8181 (lshrsi3): New.
8182 (ashlsi3): New.
8183 (cbranchsi4): New.
8184 * config/rl78/rl78.md (CC_REG): Fix.
8185 (addsi3): Allow memory and immediate operands.
8186 (addsi3_internal): Split into...
8187 (addsi3_internal_virt): ...new, and ...
8188 (addsi3_internal_real): ...new.
8189 (subsi): New.
8190 (subsi3_internal_virt): New.
8191 (subsi3_internal_real): New.
8192 (mulsi3): Add memory operand.
8193 (mulsi3_rl78): Likewise.
8194 (mulsi3_g13): Likewise.
8195 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
8196 (cbranchqi4_real): Add more constraint options.
8197 (cbranchhi4_real): Expand pattern.
8198 (cbranchhi4_real_signed): New.
8199 (cbranchhi4_real_inverted): New.
8200 (cbranchsi4_real_lt): New.
8201 (cbranchsi4_real_ge): New.
8202 (cbranchsi4_real_signed): New.
8203 (cbranchsi4_real): New.
8204 (peephole2): New.
8205 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
8206 constant shifts.
8207 (lshrsi3_virt): Likewise.
8208 (ashlsi3_virt): Likewise.
8209 (cbranchqi4_virt_signed): New.
8210 (cbranchhi4_virt_signed): New.
8211 (cbranchsi4_virt): New.
8212 * config/rl78/rl78.c: Whitespace fixes throughout.
8213 (move_elim_pass): New.
8214 (pass_data_rl78_move_elim): New.
8215 (pass_rl78_move_elim): New.
8216 (make_pass_rl78_move_elim): New.
8217 (rl78_devirt_info): Run devirt earlier.
8218 (rl78_move_elim_info): New.
8219 (rl78_asm_file_start): Register it.
8220 (rl78_split_movsi): New.
8221 (rl78_as_legitimate_address): Allow virtual base registers when
8222 appropriate.
8223 (rl78_addr_space_convert): Remove spurious debug stuff.
8224 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
8225 (rl78_print_operand): More cases for not printing '#'.
8226 (rl78_expand_compare): Remove most of the logic.
8227 (content_memory): New.
8228 (clear_content_memory): New.
8229 (get_content_index): New.
8230 (get_content_name): New.
8231 (display_content_memory): New.
8232 (update_content): New.
8233 (record_content): New.
8234 (already_contains): New.
8235 (insn_ok_now): Re-recog insns with virtual registers.
8236 (add_postponed_content_update): New.
8237 (process_postponed_content_update): New.
8238 (gen_and_emit_move): New.
8239 (transcode_memory_rtx): Record new location content.
8240 Use gen_and_emit_move.
8241 (force_into_acc): New.
8242 (move_to_acc): Use gen_and_emit_move.
8243 (move_from_acc): Likewise.
8244 (move_acc_to_reg): Likewise.
8245 (move_to_x): Likewise.
8246 (move_to_hl): Likewise.
8247 (move_to_de): Likewise.
8248 (rl78_alloc_physical_registers_op1): Record location content.
8249 (has_constraint): New.
8250 (rl78_alloc_physical_registers_op2): Record location content.
8251 Optimize use of HL.
8252 (rl78_alloc_physical_registers_ro1): Likewise.
8253 (rl78_alloc_physical_registers_cmp): Likewise.
8254 (rl78_alloc_physical_registers_umul): Likewise.
8255 (rl78_alloc_address_registers_macax): New.
8256 (rl78_alloc_physical_registers): Initialize and set location
8257 content memory as needed.
8258 (rl78_reorg): Make sure split2 is called.
8259 (rl78_rtx_costs): New.
8260
8261 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
8262
8263 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
8264 for (not (neg ...)) and (neg (not ...)) cases.
8265
8266 2013-09-11 Richard Biener <rguenther@suse.de>
8267
8268 PR middle-end/58377
8269 * passes.def: Split critical edges before late uninit warning passes.
8270 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
8271
8272 2013-09-11 Jakub Jelinek <jakub@redhat.com>
8273
8274 PR tree-optimization/58385
8275 * fold-const.c (build_range_check): If both low and high are NULL,
8276 use omit_one_operand_loc to preserve exp side-effects.
8277
8278 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8279
8280 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
8281
8282 2013-09-11 Richard Biener <rguenther@suse.de>
8283
8284 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
8285 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
8286 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
8287 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
8288 stmts_from_loop, known_dependences_p, build_empty_rdg,
8289 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
8290 * tree-loop-distribution.c: ... here.
8291 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
8292 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
8293 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
8294 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
8295 * tree-loop-distribution.c: ... here.
8296 * tree-loop-distribution.c: Include gimple-pretty-print.h.
8297 (struct partition_s): Add loops member.
8298 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
8299 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
8300 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
8301
8302 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8303 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
8304 Sergey Lega <sergey.s.lega@intel.com>
8305 Anna Tikhonova <anna.tikhonova@intel.com>
8306 Ilya Tocar <ilya.tocar@intel.com>
8307 Andrey Turetskiy <andrey.turetskiy@intel.com>
8308 Ilya Verbin <ilya.verbin@intel.com>
8309 Kirill Yukhin <kirill.yukhin@intel.com>
8310 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
8311
8312 * config/i386/constraints.md (k): New.
8313 (Yk): Ditto.
8314 * config/i386/i386.c (const regclass_map): Add new mask registers.
8315 (dbx_register_map): Ditto.
8316 (dbx64_register_map): Ditto.
8317 (svr4_dbx_register_map): Ditto.
8318 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
8319 disabled.
8320 (ix86_preferred_reload_class): Disable constants for mask registers.
8321 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
8322 (ix86_hard_regno_mode_ok): Support new mask registers.
8323 (x86_order_regs_for_local_alloc): Ditto.
8324 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
8325 (FIXED_REGISTERS): Add new mask registers.
8326 (CALL_USED_REGISTERS): Ditto.
8327 (REG_ALLOC_ORDER): Ditto.
8328 (VALID_MASK_REG_MODE): New.
8329 (FIRST_MASK_REG): Ditto.
8330 (LAST_MASK_REG): Ditto.
8331 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
8332 (MAYBE_MASK_CLASS_P): New.
8333 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
8334 (REG_CLASS_CONTENTS): Ditto.
8335 (MASK_REGNO_P): New.
8336 (ANY_MASK_REG_P): Ditto.
8337 (HI_REGISTER_NAMES): Add new mask registers.
8338 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
8339 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
8340 mask registers.
8341 (attribute "type"): Add mskmov, msklog.
8342 (attribute "length_immediate"): Support them.
8343 (attribute "memory"): Ditto.
8344 (attribute "prefix_0f"): Ditto.
8345 (*movhi_internal): Support new mask registers.
8346 (*movqi_internal): Ditto.
8347 (define_split): Split out clobber pattern is a logic
8348 insn on mask registers.
8349 (*k<logic><mode>): New.
8350 (*andhi_1): Extend to support mask regs.
8351 (*andqi_1): Extend to support mask regs.
8352 (kandn<mode>): New.
8353 (define_split): Split and-not to and and not if operands
8354 are not mask regs.
8355 (*<code><mode>_1): Separate HI mode to new pattern...
8356 (*<code>hi_1): This.
8357 (*<code>qi_1): Extend to support mask regs.
8358 (kxnor<mode>): New.
8359 (kortestzhi): Ditto.
8360 (kortestchi): Ditto.
8361 (kunpckhi): Ditto.
8362 (*one_cmpl<mode>2_1): Remove HImode and handle it...
8363 (*one_cmplhi2_1): ...Here, now with mask registers support.
8364 (*one_cmplqi2_1): Support new mask registers.
8365 (HI/QImode arithmetics splitter): Don't split if mask registers
8366 are used.
8367 (HI/QImode not splitter): Ditto.
8368 * config/i386/predicated.md (mask_reg_operand): New.
8369 (general_reg_operand): Ditto.
8370
8371 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
8372
8373 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
8374 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
8375
8376 2013-09-10 Jeff Law <law@redhat.com>
8377
8378 PR tree-optimization/58380
8379 * tree-ssa-threadupdate.c (thread_block): Recognize another case
8380 of threading through a buried loop header.
8381
8382 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
8383 return value for single successor case.
8384
8385 2013-09-10 Jan Hubicka <jh@suse.cz>
8386
8387 * ipa-devirt.c (ipa_devirt): Enable with LTO.
8388
8389 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
8390
8391 PR target/58361
8392 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
8393 support conditional execution.
8394 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
8395
8396 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
8397
8398 * lra.c (lra): Clear lra_optional_reload_pseudos before every
8399 constraint pass.
8400 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
8401 Check destination too to check move insn.
8402 (undo_optional_reloads): Add check that the original peudo did not
8403 changed its allocation and the optional reload was inherited on last
8404 inheritance pass. Break loop after deciding to keep optional reload.
8405 (lra_undo_inheritance): Add check that inherited pseudo still in
8406 memory.
8407
8408 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
8409
8410 * config/aarch64/aarch64.md (generic_sched): New.
8411 * config/aarch64/aarch64-generic.md (load): Make conditional
8412 on generic_sched attribute.
8413 (nonload): Likewise.
8414
8415 2013-09-10 Jan Hubicka <jh@suse.cz>
8416
8417 * lto-cgraph.c: Include ipa-utils.h.
8418 (compute_ltrans_boundary): Also add possible targets into the boundary.
8419
8420 2013-09-10 Jan Hubicka <jh@suse.cz>
8421
8422 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
8423 VAR_DECL of vtable rather than full expression.
8424
8425 2013-09-10 Jan Hubicka <jh@suse.cz>
8426 Paolo Carlini <paolo.carlini@oracle.com>
8427
8428 * cgraphunit.c (analyze_functions): Save input_location, set it
8429 to UNKNOWN_LOCATION and restore it at the end.
8430
8431 2013-09-10 Martin Jambor <mjambor@suse.cz>
8432
8433 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
8434 represented by a thunk.
8435
8436 2013-09-10 Jeff Law <law@redhat.com>
8437
8438 PR tree-optimization/58343
8439 * tree-ssa-threadupdate.c (thread_block): Identify and disable
8440 jump threading requests through loop headers buried in the middle
8441 of a jump threading path.
8442
8443 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
8444 in return value/type.
8445
8446 2013-09-10 Jakub Jelinek <jakub@redhat.com>
8447
8448 PR rtl-optimization/58365
8449 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
8450 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
8451 it differs.
8452
8453 2013-09-10 Richard Biener <rguenther@suse.de>
8454
8455 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
8456 * tree-data-ref.c (create_rdg_vertices): Collect all data
8457 references, signal failure to the caller, use data-ref API.
8458 (build_rdg): Compute data references only once. Maintain lifetime
8459 of data references and data dependences from within RDG.
8460 (free_rdg): Free dependence relations.
8461 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
8462 inventing extra dependences.
8463 (distribute_loop): Update for RDG API changes.
8464
8465 2013-09-10 Kai Tietz <ktietz@redhat.com>
8466
8467 * doc/invoke.texi (fms-extensions): Document changed
8468 behavior for ms-abi targets.
8469 * config/i386/i386.c (ix86_option_override_internal):
8470 Set default value of option -fms-extension for ms-abi targets.
8471
8472 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
8473
8474 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
8475
8476 2013-09-10 Alan Modra <amodra@gmail.com>
8477
8478 PR target/58330
8479 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
8480
8481 2013-09-10 Alan Modra <amodra@gmail.com>
8482
8483 * config/rs6000/predicates.md (add_cint_operand): New.
8484 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
8485 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
8486 using add_cint_operand.
8487 (largetoc_high_plus_aix): Likewise.
8488
8489 2013-09-09 Jakub Jelinek <jakub@redhat.com>
8490
8491 PR tree-optimization/58364
8492 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
8493 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
8494 the current range can't be an unconditional true or false.
8495
8496 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
8497
8498 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
8499
8500 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
8501
8502 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
8503
8504 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
8505
8506 PR c++/43452
8507 * doc/invoke.texi (-Wdelete-incomplete): Document it.
8508
8509 2013-09-09 Ian Bolton <ian.bolton@arm.com>
8510
8511 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
8512 NO_REGS for immediate that can't be moved directly into FP_REGS.
8513
8514 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8515
8516 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
8517 comparison with negated operand.
8518 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
8519 RTL form.
8520
8521 2013-09-09 Richard Biener <rguenther@suse.de>
8522
8523 PR middle-end/58326
8524 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
8525 of a subloop record all its block as affecting loop-closed SSA form.
8526
8527 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8528
8529 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
8530 of an rtx/bitpos pair.
8531 (store_fixed_bit_field): Update accordingly.
8532
8533 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8534
8535 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
8536 GEN_INT.
8537 * builtins.c (expand_errno_check): Likewise.
8538 * dwarf2cfi.c (init_return_column_size): Likewise.
8539 * except.c (sjlj_mark_call_sites): Likewise.
8540 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
8541 * lra-constraints.c (emit_inc): Likewise.
8542 * ree.c (combine_set_extension): Likewise.
8543 * regmove.c (fixup_match_2): Likewise.
8544 * reload1.c (inc_for_reload): Likewise.
8545
8546 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8547
8548 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
8549 (force_to_mode, simplify_shift_const_1, simplify_comparison):
8550 Use gen_int_mode with the mode of the associated simplify_* call.
8551 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
8552 * expmed.c (expand_shift_1): Likewise.
8553 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8554 * loop-iv.c (iv_number_of_iterations): Likewise.
8555 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8556 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
8557
8558 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8559
8560 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
8561 of the associated expand_* call.
8562 (asan_emit_stack_protection): Likewise.
8563 * builtins.c (round_trampoline_addr): Likewise.
8564 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
8565 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
8566 (emit_store_flag): Likewise.
8567 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
8568 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
8569 Likewise.
8570 * function.c (instantiate_virtual_regs_in_insn): Likewise.
8571 * ifcvt.c (noce_try_store_flag_constants): Likewise.
8572 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
8573 * modulo-sched.c (generate_prolog_epilog): Likewise.
8574 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
8575 (expand_ctz, expand_ffs, expand_unop): Likewise.
8576
8577 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8578
8579 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
8580 of the associated gen_rtx_* call.
8581 * caller-save.c (init_caller_save): Likewise.
8582 * combine.c (find_split_point, make_extraction): Likewise.
8583 (make_compound_operation): Likewise.
8584 * dwarf2out.c (mem_loc_descriptor): Likewise.
8585 * explow.c (plus_constant, probe_stack_range): Likewise.
8586 * expmed.c (expand_mult_const): Likewise.
8587 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
8588 * reload1.c (init_reload): Likewise.
8589 * valtrack.c (cleanup_auto_inc_dec): Likewise.
8590 * var-tracking.c (adjust_mems): Likewise.
8591 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
8592 rather than gen_rtx_fmt_ee.
8593
8594 2013-09-09 Jan Hubicka <jh@suse.cz>
8595
8596 PR middle-end/58294
8597 * value-prof.c (gimple_ic): Copy also abnormal edges.
8598
8599 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
8600
8601 * asan.c (asan_shadow_cst): Use gen_int_mode.
8602
8603 2013-09-08 Jan Hubicka <jh@suse.cz>
8604
8605 * ipa-profile.c: Add toplevel comment.
8606 (ipa_propagate_frequency_1): Be more conservative when profile is read.
8607 (contains_hot_call_p): New function.
8608 (ipa_propagate_frequency): Set frequencies based on counts when
8609 profile is read.
8610 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
8611 profile; do not tamper with profile after inlining if it is read.
8612
8613 2013-09-08 Jan Hubicka <jh@suse.cz>
8614
8615 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
8616 speculative edges.
8617
8618 2013-09-08 Jan Hubicka <jh@suse.cz>
8619
8620 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
8621 summary generation.
8622
8623 2013-09-08 Jeff Law <law@redhat.com>
8624
8625 PR bootstrap/58340
8626 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
8627 of 'found'.
8628
8629 2013-09-08 Andi Kleen <ak@linux.intel.com>
8630
8631 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
8632
8633 2013-09-08 Jan Hubicka <jh@suse.cz>
8634
8635 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
8636
8637 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
8638
8639 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
8640 for non-debug insns.
8641 * lra.c (new_insn_reg): Take the containing insn as a parameter.
8642 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
8643 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
8644 accordingly.
8645
8646 2013-09-08 Jan Hubicka <jh@suse.cz>
8647
8648 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
8649 targets and devirtualize to BUILT_IN_UNREACHABLE.
8650 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
8651 * ipa.c (walk_polymorphic_call_targets): New function.
8652 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
8653 functions; use the new timevar.
8654 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
8655 was removed from the program.
8656 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
8657 not consider it in the walk when its vtable is dead.
8658 (possible_polymorphic_call_targets_1): Pass anonymous flag to
8659 record_binfo.
8660 (devirt_variable_node_removal_hook): New function.
8661 (possible_polymorphic_call_targets): Also register
8662 devirt_variable_node_removal_hook.
8663 (ipa_devirt): Do not do non-speculative devirtualization.
8664 (gate_ipa_devirt): One execute if devirtualizing speculatively.
8665
8666 2013-09-08 Jan Hubicka <jh@suse.cz>
8667
8668 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
8669 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
8670 varpool_remove_variable_insertion_hook): Declare.
8671 * varpool.c (varpool_node_hook_list): New structure.
8672 (first_varpool_node_removal_hook,
8673 first_varpool_variable_insertion_hook): New variables.
8674 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
8675 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
8676 varpool_remove_variable_insertion_hook,
8677 varpool_call_variable_insertion_hooks): New functions.
8678 (varpool_remove_node): Use it.
8679
8680 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
8681
8682 PR c++/54941
8683 * diagnostic.c (diagnostic_build_prefix): When s.file is
8684 "<built-in>" don't output line and column numbers.
8685
8686 2013-09-06 Jan Hubicka <jh@suse.cz>
8687
8688 * cgraphunit.c (expand_thunk): Get body before touching arguments.
8689 * lto-streamer-out.c: Stream thunks, too.
8690 * lto-streamer-in.c (input_function): Pop cfun here
8691 (lto_read_body): Instead of here.
8692
8693 2013-09-06 Caroline Tice <cmtice@google.com>
8694
8695 * doc/install.texi: Add documentation for the --enable-vtable-verify
8696 and the --disable-libvtv configure options.
8697
8698 2013-09-06 Jeff Law <law@redhat.com>
8699
8700 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
8701 edge implied equivalences into successor phis.
8702
8703 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
8704
8705 * resource.c (mark_referenced_resources): Handle COND_EXEC.
8706
8707 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
8708
8709 * resource.c (mark_target_live_regs): Compute resources taking
8710 into account if a call is predicated or not.
8711
8712 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
8713
8714 * toplev.c (output_stack_usage): Be prepared for suffixes created by
8715 the compiler in the function names.
8716
8717 2013-09-06 Jan Hubicka <jh@suse.cz>
8718
8719 PR middle-end/58094
8720 * ipa-inline.c (has_caller_p): New function.
8721 (want_inline_function_to_all_callers_p): Use it.
8722 (sum_callers, inline_to_all_callers): Break out from ...
8723 (ipa_inline): ... here.
8724
8725 2013-09-06 Jan Hubicka <jh@suse.cz>
8726
8727 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
8728 only when AVX is enabled.
8729
8730 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8731
8732 * config/aarch64/aarch64.md
8733 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
8734 is fpsimd_<load/store>2.
8735 (load_pair<mode>): Likewise.
8736 (store_pair<mode>): Likewise.
8737
8738 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8739
8740 * config/arm/types.md (type): Add "mrs" type.
8741 * config/aarch64/aarch64.md
8742 (aarch64_load_tp_hard): Make type "mrs".
8743 * config/arm/arm.md
8744 (load_tp_hard): Make type "mrs".
8745 * config/arm/cortex-a15.md: Update with new attributes.
8746 * config/arm/cortex-a5.md: Update with new attributes.
8747 * config/arm/cortex-a53.md: Update with new attributes.
8748 * config/arm/cortex-a7.md: Update with new attributes.
8749 * config/arm/cortex-a8.md: Update with new attributes.
8750 * config/arm/cortex-a9.md: Update with new attributes.
8751 * config/arm/cortex-m4.md: Update with new attributes.
8752 * config/arm/cortex-r4.md: Update with new attributes.
8753 * config/arm/fa526.md: Update with new attributes.
8754 * config/arm/fa606te.md: Update with new attributes.
8755 * config/arm/fa626te.md: Update with new attributes.
8756 * config/arm/fa726te.md: Update with new attributes.
8757
8758 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8759
8760 * config/aarch64/aarch64.md
8761 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
8762 (*movtf_aarch64): Likewise.
8763 * config/arm/arm.md
8764 (thumb1_movdi_insn): Use "multiple" for type where more than one
8765 instruction is used for a move.
8766 (*arm32_movhf): Likewise.
8767 (*thumb_movdf_insn): Likewise.
8768
8769 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8770
8771 * config/arm/types.md (type): Rename fcpys to fmov.
8772 * config/arm/vfp.md
8773 (*arm_movsi_vfp): Rename type fcpys as fmov.
8774 (*thumb2_movsi_vfp): Likewise
8775 (*movhf_vfp_neon): Likewise
8776 (*movhf_vfp): Likewise
8777 (*movsf_vfp): Likewise
8778 (*thumb2_movsf_vfp): Likewise
8779 (*movsfcc_vfp): Likewise
8780 (*thumb2_movsfcc_vfp): Likewise
8781 * config/aarch64/aarch64-simd.md
8782 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
8783 * config/aarch64/aarch64.md
8784 (*movsi_aarch64): Replace type mov_reg with fmovs.
8785 (*movdi_aarch64): Likewise
8786 (*movsf_aarch64): Likewise
8787 (*movdf_aarch64): Likewise
8788 * config/arm/arm.c
8789 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
8790 * config/arm/iwmmxt.md
8791 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
8792 * config/arm/arm1020e.md: Update with new attributes.
8793 * config/arm/cortex-a15-neon.md: Update with new attributes.
8794 * config/arm/cortex-a5.md: Update with new attributes.
8795 * config/arm/cortex-a53.md: Update with new attributes.
8796 * config/arm/cortex-a7.md: Update with new attributes.
8797 * config/arm/cortex-a8-neon.md: Update with new attributes.
8798 * config/arm/cortex-a9.md: Update with new attributes.
8799 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8800 * config/arm/cortex-r4f.md: Update with new attributes.
8801 * config/arm/marvell-pj4.md: Update with new attributes.
8802 * config/arm/vfp11.md: Update with new attributes.
8803
8804 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8805
8806 * config/aarch64/aarch64.md
8807 (*madd<mode>): Fix type attribute.
8808 (*maddsi_uxtw): Likewise.
8809 (*msub<mode>): Likewise.
8810 (*msubsi_uxtw): Likewise.
8811 (<su_optab>maddsidi4): Likewise.
8812 (<su_optab>msubsidi4): Likewise.
8813
8814 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8815
8816 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
8817 * config/arm/arm.md (core_cycles): Remove fdiv.
8818 * config/arm/vfp.md:
8819 (*sqrtsf2_vfp): Update for attribute changes.
8820 (*sqrtdf2_vfp): Likewise.
8821 * config/aarch64/aarch64.md:
8822 (sqrt<mode>2): Update for attribute changes.
8823 * config/arm/arm1020e.md: Update with new attributes.
8824 * config/arm/cortex-a15-neon.md: Update with new attributes.
8825 * config/arm/cortex-a5.md: Update with new attributes.
8826 * config/arm/cortex-a53.md: Update with new attributes.
8827 * config/arm/cortex-a7.md: Update with new attributes.
8828 * config/arm/cortex-a8-neon.md: Update with new attributes.
8829 * config/arm/cortex-a9.md: Update with new attributes.
8830 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8831 * config/arm/cortex-r4f.md: Update with new attributes.
8832 * config/arm/marvell-pj4.md: Update with new attributes.
8833 * config/arm/vfp11.md: Update with new attributes.
8834
8835 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8836
8837 * config/arm/types.md
8838 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
8839 * config/aarch64/aarch64.md
8840 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
8841 new attributes.
8842 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
8843 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
8844 (float<GPI:mode><GPF:mode>2): Likewise.
8845 * config/arm/vfp.md
8846 (*truncsisf2_vfp): Update with new attributes.
8847 (*truncsidf2_vfp): Likewise.
8848 (fixuns_truncsfsi2): Likewise.
8849 (fixuns_truncdfsi2): Likewise.
8850 (*floatsisf2_vfp): Likewise.
8851 (*floatsidf2_vfp): Likewise.
8852 (floatunssisf2): Likewise.
8853 (floatunssidf2): Likewise.
8854 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
8855 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
8856 * config/arm/arm1020e.md: Update with new attributes.
8857 * config/arm/cortex-a15-neon.md: Update with new attributes.
8858 * config/arm/cortex-a5.md: Update with new attributes.
8859 * config/arm/cortex-a53.md: Update with new attributes.
8860 * config/arm/cortex-a7.md: Update with new attributes.
8861 * config/arm/cortex-a8-neon.md: Update with new attributes.
8862 * config/arm/cortex-a9.md: Update with new attributes.
8863 * config/arm/cortex-m4-fpu.md: Update with new attributes.
8864 * config/arm/cortex-r4f.md: Update with new attributes.
8865 * config/arm/marvell-pj4.md: Update with new attributes.
8866 * config/arm/vfp11.md: Update with new attributes.
8867
8868 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8869
8870 * config/aarch64/arm_neon.h
8871 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
8872 (vqtbx<1,2,3,4><q>_s8): Likewise.
8873
8874 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
8875
8876 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
8877 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
8878 (add<mode>3): Add type attribute.
8879 (add<mode>3): Likewise.
8880 (usadd<mode>3): Likewise.
8881 (ssadd<mode>3): Likewise.
8882 (sub<mode>3): Likewise.
8883 (sub<mode>3): Likewise.
8884 (ussub<mode>3): Likewise.
8885 (sssub<mode>3): Likewise.
8886 (ssmulsa3): Likewise.
8887 (usmulusa3): Likewise.
8888 (arm_usatsihi): Likewise.
8889 * config/arm/vfp.md
8890 (*movdi_vfp): Add types for all instructions.
8891 (*movdi_vfp_cortexa8): Likewise.
8892 (*movhf_vfp_neon): Likewise.
8893 (*movhf_vfp): Likewise.
8894 (*movdf_vfp): Likewise.
8895 (*thumb2_movdf_vfp): Likewise.
8896 (*thumb2_movdfcc_vfp): Likewise.
8897 * config/arm/arm.md: Add type attribute to all insn patterns.
8898 (*thumb1_adddi3): Add type attribute.
8899 (*arm_adddi3): Likewise.
8900 (*adddi_sesidi_di): Likewise.
8901 (*adddi_zesidi_di): Likewise.
8902 (*thumb1_addsi3): Likewise.
8903 (addsi3_compare0): Likewise.
8904 (*addsi3_compare0_scratch): Likewise.
8905 (*compare_negsi_si): Likewise.
8906 (cmpsi2_addneg): Likewise.
8907 (*addsi3_carryin_<optab>): Likewise.
8908 (*addsi3_carryin_alt2_<optab>): Likewise.
8909 (*addsi3_carryin_clobercc_<optab>): Likewise.
8910 (*subsi3_carryin): Likewise.
8911 (*subsi3_carryin_const): Likewise.
8912 (*subsi3_carryin_compare): Likewise.
8913 (*subsi3_carryin_compare_const): Likewise.
8914 (*arm_subdi3): Likewise.
8915 (*thumb_subdi3): Likewise.
8916 (*subdi_di_zesidi): Likewise.
8917 (*subdi_di_sesidi): Likewise.
8918 (*subdi_zesidi_di): Likewise.
8919 (*subdi_sesidi_di): Likewise.
8920 (*subdi_zesidi_ze): Likewise.
8921 (thumb1_subsi3_insn): Likewise.
8922 (*arm_subsi3_insn): Likewise.
8923 (*anddi3_insn): Likewise.
8924 (*anddi_zesidi_di): Likewise.
8925 (*anddi_sesdi_di): Likewise.
8926 (*ne_zeroextracts): Likewise.
8927 (*ne_zeroextracts): Likewise.
8928 (*ite_ne_zeroextr): Likewise.
8929 (*ite_ne_zeroextr): Likewise.
8930 (*anddi_notdi_di): Likewise.
8931 (*anddi_notzesidi): Likewise.
8932 (*anddi_notsesidi): Likewise.
8933 (andsi_notsi_si): Likewise.
8934 (thumb1_bicsi3): Likewise.
8935 (*iordi3_insn): Likewise.
8936 (*iordi_zesidi_di): Likewise.
8937 (*iordi_sesidi_di): Likewise.
8938 (*thumb1_iorsi3_insn): Likewise.
8939 (*xordi3_insn): Likewise.
8940 (*xordi_zesidi_di): Likewise.
8941 (*xordi_sesidi_di): Likewise.
8942 (*arm_xorsi3): Likewise.
8943 (*andsi_iorsi3_no): Likewise.
8944 (*smax_0): Likewise.
8945 (*smax_m1): Likewise.
8946 (*arm_smax_insn): Likewise.
8947 (*smin_0): Likewise.
8948 (*arm_smin_insn): Likewise.
8949 (*arm_umaxsi3): Likewise.
8950 (*arm_uminsi3): Likewise.
8951 (*minmax_arithsi): Likewise.
8952 (*minmax_arithsi_): Likewise.
8953 (*satsi_<SAT:code>): Likewise.
8954 (arm_ashldi3_1bit): Likewise.
8955 (arm_ashrdi3_1bit): Likewise.
8956 (arm_lshrdi3_1bit): Likewise.
8957 (*arm_negdi2): Likewise.
8958 (*thumb1_negdi2): Likewise.
8959 (*arm_negsi2): Likewise.
8960 (*thumb1_negsi2): Likewise.
8961 (*negdi_extendsid): Likewise.
8962 (*negdi_zero_extend): Likewise.
8963 (*arm_abssi2): Likewise.
8964 (*thumb1_abssi2): Likewise.
8965 (*arm_neg_abssi2): Likewise.
8966 (*thumb1_neg_abss): Likewise.
8967 (one_cmpldi2): Likewise.
8968 (extend<mode>di2): Likewise.
8969 (*compareqi_eq0): Likewise.
8970 (*arm_extendhisi2addsi): Likewise.
8971 (*arm_movdi): Likewise.
8972 (*thumb1_movdi_insn): Likewise.
8973 (*arm_movt): Likewise.
8974 (*thumb1_movsi_insn): Likewise.
8975 (pic_add_dot_plus_four): Likewise.
8976 (pic_add_dot_plus_eight): Likewise.
8977 (tls_load_dot_plus_eight): Likewise.
8978 (*thumb1_movhi_insn): Likewise.
8979 (*thumb1_movsf_insn): Likewise.
8980 (*movdf_soft_insn): Likewise.
8981 (*thumb_movdf_insn): Likewise.
8982 (cbranchsi4_insn): Likewise.
8983 (cbranchsi4_scratch): Likewise.
8984 (*negated_cbranchsi4): Likewise.
8985 (*tbit_cbranch): Likewise.
8986 (*tlobits_cbranch): Likewise.
8987 (*tstsi3_cbranch): Likewise.
8988 (*cbranchne_decr1): Likewise.
8989 (*addsi3_cbranch): Likewise.
8990 (*addsi3_cbranch_scratch): Likewise.
8991 (*arm_cmpdi_insn): Likewise.
8992 (*arm_cmpdi_unsig): Likewise.
8993 (*arm_cmpdi_zero): Likewise.
8994 (*thumb_cmpdi_zero): Likewise.
8995 (*deleted_compare): Likewise.
8996 (*mov_scc): Likewise.
8997 (*mov_negscc): Likewise.
8998 (*mov_notscc): Likewise.
8999 (*cstoresi_eq0_thumb1_insn): Likewise.
9000 (cstoresi_nltu_thumb1): Likewise.
9001 (cstoresi_ltu_thu): Likewise.
9002 (thumb1_addsi3_addgeu): Likewise.
9003 (*arm_jump): Likewise.
9004 (*thumb_jump): Likewise.
9005 (*check_arch2): Likewise.
9006 (arm_casesi_internal): Likewise.
9007 (thumb1_casesi_dispatch): Likewise.
9008 (*arm_indirect_jump): Likewise.
9009 (*thumb1_indirect_jump): Likewise.
9010 (nop): Likewise.
9011 (*and_scc): Likewise.
9012 (*ior_scc): Likewise.
9013 (*compare_scc): Likewise.
9014 (*cond_move): Likewise.
9015 (*cond_arith): Likewise.
9016 (*cond_sub): Likewise.
9017 (*cmp_ite0): Likewise.
9018 (*cmp_ite1): Likewise.
9019 (*cmp_and): Likewise.
9020 (*cmp_ior): Likewise.
9021 (*ior_scc_scc): Likewise.
9022 (*ior_scc_scc_cmp): Likewise.
9023 (*and_scc_scc): Likewise.
9024 (*and_scc_scc_cmp): Likewise.
9025 (*and_scc_scc_nod): Likewise.
9026 (*negscc): Likewise.
9027 (movcond_addsi): Likewise.
9028 (movcond): Likewise.
9029 (*ifcompare_plus_move): Likewise.
9030 (*if_plus_move): Likewise.
9031 (*ifcompare_move_plus): Likewise.
9032 (*if_move_plus): Likewise.
9033 (*ifcompare_arith_arith): Likewise.
9034 (*if_arith_arith): Likewise.
9035 (*ifcompare_arith_move): Likewise.
9036 (*if_arith_move): Likewise.
9037 (*ifcompare_move_arith): Likewise.
9038 (*if_move_arith): Likewise.
9039 (*ifcompare_move_not): Likewise.
9040 (*if_move_not): Likewise.
9041 (*ifcompare_not_move): Likewise.
9042 (*if_not_move): Likewise.
9043 (*ifcompare_shift_move): Likewise.
9044 (*if_shift_move): Likewise.
9045 (*ifcompare_move_shift): Likewise.
9046 (*if_move_shift): Likewise.
9047 (*ifcompare_shift_shift): Likewise.
9048 (*ifcompare_not_arith): Likewise.
9049 (*ifcompare_arith_not): Likewise.
9050 (*if_arith_not): Likewise.
9051 (*ifcompare_neg_move): Likewise.
9052 (*if_neg_move): Likewise.
9053 (*ifcompare_move_neg): Likewise.
9054 (*if_move_neg): Likewise.
9055 (prologue_thumb1_interwork): Likewise.
9056 (*cond_move_not): Likewise.
9057 (*sign_extract_onebit): Likewise.
9058 (*not_signextract_onebit): Likewise.
9059 (stack_tie): Likewise.
9060 (align_4): Likewise.
9061 (align_8): Likewise.
9062 (consttable_end): Likewise.
9063 (consttable_1): Likewise.
9064 (consttable_2): Likewise.
9065 (consttable_4): Likewise.
9066 (consttable_8): Likewise.
9067 (consttable_16): Likewise.
9068 (*thumb1_tablejump): Likewise.
9069 (prefetch): Likewise.
9070 (force_register_use): Likewise.
9071 (thumb_eh_return): Likewise.
9072 (load_tp_hard): Likewise.
9073 (load_tp_soft): Likewise.
9074 (tlscall): Likewise.
9075 (*arm_movtas_ze): Likewise.
9076 (*arm_rev): Likewise.
9077 (*arm_revsh): Likewise.
9078 (*arm_rev16): Likewise.
9079 * config/arm/thumb2.md
9080 (*thumb2_smaxsi3): Likewise.
9081 (*thumb2_sminsi3): Likewise.
9082 (*thumb32_umaxsi3): Likewise.
9083 (*thumb2_uminsi3): Likewise.
9084 (*thumb2_negdi2): Likewise.
9085 (*thumb2_abssi2): Likewise.
9086 (*thumb2_neg_abss): Likewise.
9087 (*thumb2_movsi_insn): Likewise.
9088 (tls_load_dot_plus_four): Likewise.
9089 (*thumb2_movhi_insn): Likewise.
9090 (*thumb2_mov_scc): Likewise.
9091 (*thumb2_mov_negs): Likewise.
9092 (*thumb2_mov_negs): Likewise.
9093 (*thumb2_mov_nots): Likewise.
9094 (*thumb2_mov_nots): Likewise.
9095 (*thumb2_movsicc_): Likewise.
9096 (*thumb2_movsfcc_soft_insn): Likewise.
9097 (*thumb2_indirect_jump): Likewise.
9098 (*thumb2_and_scc): Likewise.
9099 (*thumb2_ior_scc): Likewise.
9100 (*thumb2_ior_scc_strict_it): Likewise.
9101 (*thumb2_cond_move): Likewise.
9102 (*thumb2_cond_arith): Likewise.
9103 (*thumb2_cond_ari): Likewise.
9104 (*thumb2_cond_sub): Likewise.
9105 (*thumb2_negscc): Likewise.
9106 (*thumb2_movcond): Likewise.
9107 (thumb2_casesi_internal): Likewise.
9108 (thumb2_casesi_internal_pic): Likewise.
9109 (*thumb2_alusi3_short): Likewise.
9110 (*thumb2_mov<mode>_shortim): Likewise.
9111 (*thumb2_addsi_short): Likewise.
9112 (*thumb2_subsi_short): Likewise.
9113 (thumb2_addsi3_compare0): Likewise.
9114 (*thumb2_cbz): Likewise.
9115 (*thumb2_cbnz): Likewise.
9116 (*thumb2_one_cmplsi2_short): Likewise.
9117 (*thumb2_negsi2_short): Likewise.
9118 (*orsi_notsi_si): Likewise.
9119 * config/arm/arm1020e.md: Update with new attributes.
9120 * config/arm/arm1026ejs.md: Update with new attributes.
9121 * config/arm/arm1136jfs.md: Update with new attributes.
9122 * config/arm/arm926ejs.md: Update with new attributes.
9123 * config/arm/cortex-a15.md: Update with new attributes.
9124 * config/arm/cortex-a5.md: Update with new attributes.
9125 * config/arm/cortex-a53.md: Update with new attributes.
9126 * config/arm/cortex-a7.md: Update with new attributes.
9127 * config/arm/cortex-a8.md: Update with new attributes.
9128 * config/arm/cortex-a9.md: Update with new attributes.
9129 * config/arm/cortex-m4.md: Update with new attributes.
9130 * config/arm/cortex-r4.md: Update with new attributes.
9131 * config/arm/fa526.md: Update with new attributes.
9132 * config/arm/fa606te.md: Update with new attributes.
9133 * config/arm/fa626te.md: Update with new attributes.
9134 * config/arm/fa726te.md: Update with new attributes.
9135
9136 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
9137
9138 * config/aarch64/aarch64-simd.md
9139 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
9140 <vwx> iterator to ensure correct register choice.
9141 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
9142 (aarch64_sqdmull_n<mode>): Likewise.
9143 (aarch64_sqdmull2_n<mode>_internal): Likewise.
9144 * config/aarch64/arm_neon.h
9145 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
9146 (vml<as><q>_n_<su>16): Likewise.
9147 (vml<as>l_high_lane<q>_<su>16): Likewise.
9148 (vml<as>l_high_n_<su>16): Likewise.
9149 (vml<as>l_lane<q>_<su>16): Likewise.
9150 (vml<as>l_n_<su>16): Likewise.
9151 (vmul<q>_lane<q>_<su>16): Likewise.
9152 (vmul<q>_n_<su>16): Likewise.
9153 (vmull_lane<q>_<su>16): Likewise.
9154 (vmull_n_<su>16): Likewise.
9155 (vmull_high_lane<q>_<su>16): Likewise.
9156 (vmull_high_n_<su>16): Likewise.
9157 (vqrdmulh<q>_n_s16): Likewise.
9158
9159 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
9160
9161 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
9162 have the correct lane parameter.
9163
9164 2013-09-06 Richard Biener <rguenther@suse.de>
9165
9166 * cfganal.c (control_dependences::~control_dependences):
9167 Properly free all of the vector.
9168
9169 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
9170
9171 PR target/58269
9172 * config/i386/i386.c (ix86_conditional_register_usage):
9173 Proper initialize extended SSE registers.
9174
9175 2013-09-06 Jan Hubicka <jh@suse.cz>
9176
9177 PR tree-optimization/58311
9178 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
9179
9180 2013-09-06 Jan Hubicka <jh@suse.cz>
9181
9182 * Makefile.in (tree-sra.o): Update dependencies.
9183 * tree-sra.c: Include ipa-utils.h
9184 (scan_function): Use recursive_call_p.
9185 (has_caller_p): New function.
9186 (cgraph_for_node_and_aliases): Count also callers of aliases.
9187
9188 2013-09-06 Jan Hubicka <jh@suse.cz>
9189
9190 PR middle-end/58094
9191 * cgraph.h (symtab_semantically_equivalent_p): Declare.
9192 * tree-tailcall.c: Include ipa-utils.h.
9193 (find_tail_calls): Use it.
9194 * ipa-pure-const.c (check_call): Likewise.
9195 * ipa-utils.c (recursive_call_p): New function.
9196 * ipa-utils.h (recursive_call_p): Dclare.
9197 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
9198 (symtab_semantically_equivalent_p): New function.
9199 * Makefile.in (tree-tailcall.o): Update dependencies.
9200
9201 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
9202
9203 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
9204 non-inlinable part.
9205
9206 2013-09-06 Richard Biener <rguenther@suse.de>
9207
9208 * lto-streamer.h (lto_global_var_decls): Remove.
9209 * Makefile.in (OBJS): Remove lto-symtab.o.
9210 (lto-symtab.o): Remove.
9211 (GTFILES): Remove lto-symtab.c
9212 * lto-symtab.c: Move to lto/
9213
9214 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9215
9216 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
9217 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
9218 (UNSPEC_FPINT_RINT): New constant definitions.
9219 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
9220 definition with 2 attributes.
9221 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
9222 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
9223 definitions.
9224
9225 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
9226
9227 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
9228 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
9229 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
9230 attribute to "z196_alone".
9231 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
9232 "zEC12_simple".
9233
9234 2013-09-06 Richard Biener <rguenther@suse.de>
9235
9236 * basic-block.h (class control_dependences): New.
9237 * tree-ssa-dce.c (control_dependence_map): Remove.
9238 (cd): New global.
9239 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
9240 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
9241 find_pdom, find_control_dependence, find_all_control_dependences):
9242 Move to cfganal.c.
9243 (mark_control_dependent_edges_necessary,
9244 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
9245 tree_dce_done, perform_tree_ssa_dce): Adjust.
9246 * cfganal.c (set_control_dependence_map_bit,
9247 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
9248 find_all_control_dependences): Move from tree-ssa-dce.c and
9249 implement as methods of control_dependences class.
9250 (control_dependences::control_dependences): New.
9251 (control_dependences::~control_dependences): Likewise.
9252 (control_dependences::get_edges_dependent_on): Likewise.
9253 (control_dependences::get_edge): Likewise.
9254
9255 2013-09-04 Jan Hubicka <jh@suse.cz>
9256
9257 * tree.c (types_same_for_odr): Drop overactive check.
9258 * ipa-devirt.c (hash_type_name): Likewise.
9259
9260 2013-09-04 Jan Hubicka <jh@suse.cz>
9261
9262 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
9263 (analyze_functions): ... here.
9264
9265 2013-09-04 Jan Hubicka <jh@suse.cz>
9266
9267 PR middle-end/58201
9268 * cgraphunit.c (analyze_functions): Clear AUX fields
9269 after processing; initialize assembler name has.
9270
9271 2013-09-05 Jeff Law <law@redhat.com>
9272
9273 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
9274 from thread_around_empty_block. Record threading path into PATH.
9275 Recurse if threading through the initial block is successful.
9276 (thread_across_edge): Corresponding changes to slightly simplify.
9277
9278 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9279
9280 * config/aarch64/aarch64.md
9281 (type): Remove frecpe, frecps, frecpx.
9282 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
9283 fix to be a TARGET_SIMD instruction.
9284 (aarch64_frecps): Remove.
9285 * config/aarch64/aarch64-simd.md
9286 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
9287 (aarch64_frecps<mode>): Handle all float/vector of float modes.
9288
9289 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9290 Sofiane Naci <sofiane.naci@arm.com>
9291
9292 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
9293 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
9294 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
9295 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
9296 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
9297 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
9298 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
9299 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
9300 into "clz, "rbit". Rename "shift" to "shift_imm".
9301 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
9302 changes. Update for attribute changes all occurrences of arlo_* and
9303 shift* types.
9304 * config/arm/arm-fixed.md: Update for attribute changes
9305 all occurrences of arlo_* types.
9306 * config/arm/thumb2.md: Update for attribute changes all occurrences
9307 of arlo_* types.
9308 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
9309 (cortexa7_older_only): Likewise.
9310 (cortexa7_younger): Likewise.
9311 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
9312 (1020alu_shift_op): Likewise.
9313 (1020alu_shift_reg_op): Likewise.
9314 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
9315 (alu_shift_op): Likewise.
9316 (alu_shift_reg_op): Likewise.
9317 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
9318 (11_alu_shift_op): Likewise.
9319 (11_alu_shift_reg_op): Likewise.
9320 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
9321 (9_alu_shift_reg_op): Likewise.
9322 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
9323 attribute changes.
9324 (cortex_a15_alu_shift): Likewise.
9325 (cortex_a15_alu_shift_reg): Likewise.
9326 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
9327 attribute changes.
9328 (cortex_a5_alu_shift): Likewise.
9329 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
9330 attribute changes.
9331 (cortex_a53_alu_shift): Likewise.
9332 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
9333 attribute changes.
9334 (cortex_a7_alu_reg): Likewise.
9335 (cortex_a7_alu_shift): Likewise.
9336 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
9337 attribute changes.
9338 (cortex_a8_alu_shift): Likewise.
9339 (cortex_a8_alu_shift_reg): Likewise.
9340 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
9341 (cortex_a9_dp_shift): Likewise.
9342 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
9343 attribute changes.
9344 * config/arm/cortex-r4.md
9345 (cortex_r4_alu): Update for attribute changes.
9346 (cortex_r4_mov): Likewise.
9347 (cortex_r4_alu_shift_reg): Likewise.
9348 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
9349 (526_alu_shift_op): Likewise.
9350 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
9351 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
9352 (626te_alu_shift_op): Likewise.
9353 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
9354 (726te_alu_shift_op): Likewise.
9355 (726te_alu_shift_reg_op): Likewise.
9356 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
9357 (mp626_alu_shift_op): Likewise.
9358 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
9359 (pj4_alu_conds): Likewise.
9360 (pj4_shift): Likewise.
9361 (pj4_shift_conds): Likewise.
9362 (pj4_alu_shift): Likewise.
9363 (pj4_alu_shift_conds): Likewise.
9364 * config/aarch64/aarch64.md: Update for attribute change
9365 all occurrences of arlo_* and shift* types.
9366
9367 2013-09-05 Mike Stump <mikestump@comcast.net>
9368
9369 * tree.h: Move documentation for tree_function_decl to tree-core.h
9370 with the declaration.
9371
9372 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
9373
9374 PR target/58139
9375 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
9376 looking for widest mode.
9377
9378 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
9379
9380 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
9381
9382 2013-09-05 Richard Biener <rguenther@suse.de>
9383
9384 PR tree-optimization/58137
9385 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
9386 Do not create vectors of pointers.
9387 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
9388 types for the components of the vector initializer.
9389 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
9390 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
9391
9392 2013-09-05 Martin Jambor <mjambor@suse.cz>
9393
9394 * ipa-prop.c (remove_described_reference): Accept missing references,
9395 return false if that hppens, otherwise return true.
9396 (cgraph_node_for_jfunc): New function.
9397 (try_decrement_rdesc_refcount): Likewise.
9398 (try_make_edge_direct_simple_call): Use them.
9399 (ipa_edge_removal_hook): Remove references from rdescs.
9400 (ipa_edge_duplication_hook): Clone rdescs and their references
9401 when the new edge has the same caller as the old one.
9402 * cgraph.c (cgraph_resolve_speculation): Remove speculative
9403 reference before removing any edges.
9404
9405 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
9406
9407 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
9408 initial store.
9409 * thumb2.md (thumb2_storewb_parisi): New pattern.
9410
9411 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
9412
9413 * config/aarch64/aarch64-option-extensions.def: Add
9414 AARCH64_OPT_EXTENSION of 'crc'.
9415 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
9416 (AARCH64_ISA_CRC): Ditto.
9417 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
9418 description of the CRC extension.
9419
9420 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
9421
9422 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
9423 * config/rs6000/linux.h: Ditto.
9424 * alpha/linux.h: Ditto.
9425 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
9426 no_c99_libc_has_function.
9427 * config/c6x/uclinux-elf.h: Ditto.
9428 * config/lm32/uclinux-elf.h: Ditto.
9429 * config/m68k/uclinux.h: Ditto.
9430 * config/moxie/uclinux.h: Ditto.
9431 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
9432 (crisv32-*-linux*, cris-*-linux*): Ditto.
9433 * config/bfin/bfin.c: Include "tm_p.h".
9434
9435 2013-09-05 Richard Biener <rguenther@suse.de>
9436
9437 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
9438 check for a definition without a basic-block.
9439
9440 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
9441 Sofiane Naci <sofiane.naci@arm.com>
9442
9443 * config/aarch64/aarch64.md
9444 (*movti_aarch64): Rename r_2_f and f_2_r.
9445 (*movsf_aarch64): Likewise.
9446 (*movdf_aarch64): Likewise.
9447 (*movtf_aarch64): Likewise.
9448 (aarch64_movdi_<mode>low): Likewise.
9449 (aarch64_movdi_<mode>high): Likewise.
9450 (aarch64_mov<mode>high_di): Likewise.
9451 (aarch64_mov<mode>low_di): Likewise.
9452 (aarch64_movtilow_tilow): Likewise.
9453 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
9454 values to config/arm/types.md
9455 (attribute "conds"): Update for attribute change.
9456 (anddi3_insn): Likewise.
9457 (iordi3_insn): Likewise.
9458 (xordi3_insn): Likewise.
9459 (one_cmpldi2): Likewise.
9460 * config/arm/types.md (type): Add Neon types.
9461 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
9462 use "type" attribute.
9463 (movmisalign<mode>_neon_store): Likewise.
9464 (movmisalign<mode>_neon_load): Likewise.
9465 (vec_set<mode>_internal): Likewise.
9466 (vec_setv2di_internal): Likewise.
9467 (vec_extract<mode>): Likewise.
9468 (vec_extractv2di): Likewise.
9469 (add<mode>3_neon): Likewise.
9470 (adddi3_neon): Likewise.
9471 (sub<mode>3_neon): Likewise.
9472 (subdi3_neon): Likewise.
9473 (mul<mode>3_neon): Likewise.
9474 (mul<mode>3add<mode>_neon): Likewise.
9475 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
9476 (fma<VCVTF:mode>4)): Likewise.
9477 (fma<VCVTF:mode>4_intrinsic): Likewise.
9478 (fmsub<VCVTF:mode>4)): Likewise.
9479 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
9480 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
9481 (ior<mode>3): Likewise.
9482 (and<mode>3): Likewise.
9483 (anddi3_neon): Likewise.
9484 (orn<mode>3_neon): Likewise.
9485 (orndi3_neon): Likewise.
9486 (bic<mode>3_neon): Likewise.
9487 (bicdi3_neon): Likewise.
9488 (xor<mode>3): Likewise.
9489 (one_cmpl<mode>2): Likewise.
9490 (abs<mode>2): Likewise.
9491 (neg<mode>2): Likewise.
9492 (umin<mode>3_neon): Likewise.
9493 (umax<mode>3_neon): Likewise.
9494 (smin<mode>3_neon): Likewise.
9495 (smax<mode>3_neon): Likewise.
9496 (vashl<mode>3): Likewise.
9497 (vashr<mode>3_imm): Likewise.
9498 (vlshr<mode>3_imm): Likewise.
9499 (ashl<mode>3_signed): Likewise.
9500 (ashl<mode>3_unsigned): Likewise.
9501 (neon_load_count): Likewise.
9502 (ashldi3_neon_noclobber): Likewise.
9503 (signed_shift_di3_neon): Likewise.
9504 (unsigned_shift_di3_neon): Likewise.
9505 (ashrdi3_neon_imm_noclobber): Likewise.
9506 (lshrdi3_neon_imm_noclobber): Likewise.
9507 (widen_ssum<mode>3): Likewise.
9508 (widen_usum<mode>3): Likewise.
9509 (quad_halves_<code>v4si): Likewise.
9510 (quad_halves_<code>v4sf): Likewise.
9511 (quad_halves_<code>v8hi): Likewise.
9512 (quad_halves_<code>v16qi): Likewise.
9513 (reduc_splus_v2di): Likewise.
9514 (neon_vpadd_internal<mode>): Likewise.
9515 (neon_vpsmin<mode>): Likewise.
9516 (neon_vpsmax<mode>): Likewise.
9517 (neon_vpumin<mode>): Likewise.
9518 (neon_vpumax<mode>): Likewise.
9519 (ss_add<mode>_neon): Likewise.
9520 (us_add<mode>_neon): Likewise.
9521 (ss_sub<mode>_neon): Likewise.
9522 (us_sub<mode>_neon): Likewise.
9523 (neon_vadd<mode>_unspec): Likewise.
9524 (neon_vaddl<mode>): Likewise.
9525 (neon_vaddw<mode>): Likewise.
9526 (neon_vhadd<mode>): Likewise.
9527 (neon_vqadd<mode>): Likewise.
9528 (neon_vaddhn<mode>): Likewise.
9529 (neon_vmul<mode>): Likewise.
9530 (neon_vmla<mode>): Likewise.
9531 (neon_vmlal<mode>): Likewise.
9532 (neon_vmls<mode>): Likewise.
9533 (neon_vmlsl<mode>): Likewise.
9534 (neon_vqdmulh<mode>): Likewise.
9535 (neon_vqdmlal<mode>): Likewise.
9536 (neon_vqdmlsl<mode>): Likewise.
9537 (neon_vmull<mode>): Likewise.
9538 (neon_vqdmull<mode>): Likewise.
9539 (neon_vsub<mode>_unspec): Likewise.
9540 (neon_vsubl<mode>): Likewise.
9541 (neon_vsubw<mode>): Likewise.
9542 (neon_vqsub<mode>): Likewise.
9543 (neon_vhsub<mode>): Likewise.
9544 (neon_vsubhn<mode>): Likewise.
9545 (neon_vceq<mode>): Likewise.
9546 (neon_vcge<mode>): Likewise.
9547 (neon_vcgeu<mode>): Likewise.
9548 (neon_vcgt<mode>): Likewise.
9549 (neon_vcgtu<mode>): Likewise.
9550 (neon_vcle<mode>): Likewise.
9551 (neon_vclt<mode>): Likewise.
9552 (neon_vcage<mode>): Likewise.
9553 (neon_vcagt<mode>): Likewise.
9554 (neon_vtst<mode>): Likewise.
9555 (neon_vabd<mode>): Likewise.
9556 (neon_vabdl<mode>): Likewise.
9557 (neon_vaba<mode>): Likewise.
9558 (neon_vabal<mode>): Likewise.
9559 (neon_vmax<mode>): Likewise.
9560 (neon_vmin<mode>): Likewise.
9561 (neon_vpaddl<mode>): Likewise.
9562 (neon_vpadal<mode>): Likewise.
9563 (neon_vpmax<mode>): Likewise.
9564 (neon_vpmin<mode>): Likewise.
9565 (neon_vrecps<mode>): Likewise.
9566 (neon_vrsqrts<mode>): Likewise.
9567 (neon_vqabs<mode>): Likewise.
9568 (neon_vqneg<mode>): Likewise.
9569 (neon_vcls<mode>): Likewise.
9570 (clz<mode>2): Likewise.
9571 (popcount<mode>2): Likewise.
9572 (neon_vrecpe): Likewise.
9573 (neon_vrsqrte): Likewise.
9574 (neon_vget_lane<mode>_sext_internal): Likewise.
9575 (neon_vget_lane<mode>_zext_internal): Likewise.
9576 (neon_vdup_n<mode>): Likewise.
9577 (neon_vdup_nv2di): Likewise.
9578 (neon_vdpu_lane<mode>_internal): Likewise.
9579 (neon_vswp<mode>): Likewise.
9580 (float<mode><V_cvtto>2): Likewise.
9581 (floatuns<mode><V_cvtto>2): Likewise.
9582 (fix_trunc<mode><V_cvtto>)2): Likewise
9583 (fixuns_trunc<mode><V_cvtto)2): Likewise.
9584 (neon_vcvt<mode>): Likewise.
9585 (neon_vcvtv4sfv4hf): Likewise.
9586 (neon_vcvtv4hfv4sf): Likewise.
9587 (neon_vcvt_n<mode>): Likewise.
9588 (neon_vmovn<mode>): Likewise.
9589 (neon_vqmovn<mode>): Likewise.
9590 (neon_vqmovun<mode>): Likewise.
9591 (neon_vmovl<mode>): Likewise.
9592 (neon_vmul_lane<mode>): Likewise.
9593 (neon_vmull_lane<mode>): Likewise.
9594 (neon_vqdmull_lane<mode>): Likewise.
9595 (neon_vqdmulh_lane<mode>): Likewise.
9596 (neon_vmla_lane<mode>): Likewise.
9597 (neon_vmlal_lane<mode>): Likewise.
9598 (neon_vqdmlal_lane<mode>): Likewise.
9599 (neon_vmls_lane<mode>): Likewise.
9600 (neon_vmlsl_lane<mode>): Likewise.
9601 (neon_vqdmlsl_lane<mode>): Likewise.
9602 (neon_vext<mode>): Likewise.
9603 (neon_vrev64<mode>): Likewise.
9604 (neon_vrev32<mode>): Likewise.
9605 (neon_vrev16<mode>): Likewise.
9606 (neon_vbsl<mode>_internal): Likewise.
9607 (neon_vshl<mode>): Likewise.
9608 (neon_vqshl<mode>): Likewise.
9609 (neon_vshr_n<mode>): Likewise.
9610 (neon_vshrn_n<mode>): Likewise.
9611 (neon_vqshrn_n<mode>): Likewise.
9612 (neon_vqshrun_n<mode>): Likewise.
9613 (neon_vshl_n<mode>): Likewise.
9614 (neon_vqshl_n<mode>): Likewise.
9615 (neon_vqshlu_n<mode>): Likewise.
9616 (neon_vshll_n<mode>): Likewise.
9617 (neon_vsra_n<mode>): Likewise.
9618 (neon_vsri_n<mode>): Likewise.
9619 (neon_vsli_n<mode>): Likewise.
9620 (neon_vtbl1v8qi): Likewise.
9621 (neon_vtbl2v8qi): Likewise.
9622 (neon_vtbl3v8qi): Likewise.
9623 (neon_vtbl4v8qi): Likewise.
9624 (neon_vtbx1v8qi): Likewise.
9625 (neon_vtbx2v8qi): Likewise.
9626 (neon_vtbx3v8qi): Likewise.
9627 (neon_vtbx4v8qi): Likewise.
9628 (neon_vtrn<mode>_internal): Likewise.
9629 (neon_vzip<mode>_internal): Likewise.
9630 (neon_vuzp<mode>_internal): Likewise.
9631 (neon_vld1<mode>): Likewise.
9632 (neon_vld1_lane<mode>): Likewise.
9633 (neon_vld1_dup<mode>): Likewise.
9634 (neon_vld1_dupv2di): Likewise.
9635 (neon_vst1<mode>): Likewise.
9636 (neon_vst1_lane<mode>): Likewise.
9637 (neon_vld2<mode>): Likewise.
9638 (neon_vld2_lane<mode>): Likewise.
9639 (neon_vld2_dup<mode>): Likewise.
9640 (neon_vst2<mode>): Likewise.
9641 (neon_vst2_lane<mode>): Likewise.
9642 (neon_vld3<mode>): Likewise.
9643 (neon_vld3qa<mode>): Likewise.
9644 (neon_vld3qb<mode>): Likewise.
9645 (neon_vld3_lane<mode>): Likewise.
9646 (neon_vld3_dup<mode>): Likewise.
9647 (neon_vst3<mode>): Likewise.
9648 (neon_vst3qa<mode>): Likewise.
9649 (neon_vst3qb<mode>): Likewise.
9650 (neon_vst3_lane<mode>): Likewise.
9651 (neon_vld4<mode>): Likewise.
9652 (neon_vld4qa<mode>): Likewise.
9653 (neon_vld4qb<mode>): Likewise.
9654 (neon_vld4_lane<mode>): Likewise.
9655 (neon_vld4_dup<mode>): Likewise.
9656 (neon_vst4<mode>): Likewise.
9657 (neon_vst4qa<mode>): Likewise.
9658 (neon_vst4qb<mode>): Likewise.
9659 (neon_vst4_lane<mode>): Likewise.
9660 (neon_vec_unpack<US>_lo_<mode>): Likewise.
9661 (neon_vec_unpack<US>_hi_<mode>): Likewise.
9662 (neon_vec_<US>mult_lo_<mode>): Likewise.
9663 (neon_vec_<US>mult_hi_<mode>): Likewise.
9664 (neon_vec_<US>shiftl_<mode>): Likewise.
9665 (neon_unpack<US>_<mode>): Likewise.
9666 (neon_vec_<US>mult_<mode>): Likewise.
9667 (vec_pack_trunc_<mode>): Likewise.
9668 (neon_vec_pack_trunk_<mode>): Likewise.
9669 (neon_vabd<mode>_2): Likewise.
9670 (neon_vabd<mode>_3): Likewise.
9671 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
9672 (thumb2_movsi_vfp): Likewise.
9673 (movdi_vfp): Likewise.
9674 (movdi_vfp_cortexa8): Likewise.
9675 (movhf_vfp_neon): Likewise.
9676 (movhf_vfp): Likewiwse.
9677 (movsf_vfp): Likewiwse.
9678 (thumb2_movsf_vfp): Likewiwse.
9679 (movdf_vfp): Likewise.
9680 (thumb2_movdf_vfp): Likewise.
9681 (movsfcc_vfp): Likewise.
9682 (thumb2_movsfcc_vfp): Likewise.
9683 (movdfcc_vfp): Likewise.
9684 (thumb2_movdfcc_vfp): Likewise.
9685 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
9686 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
9687 (v10_v2c): Likewise.
9688 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
9689 attribute change.
9690 (cortex_a15_neon_int_2): Likewise.
9691 (cortex_a15_neon_int_3): Likewise.
9692 (cortex_a15_neon_int_4): Likewise.
9693 (cortex_a15_neon_int_5): Likewise.
9694 (cortex_a15_neon_vqneg_vqabs): Likewise.
9695 (cortex_a15_neon_vmov): Likewise.
9696 (cortex_a15_neon_vaba): Likewise.
9697 (cortex_a15_neon_vaba_qqq): Likewise.
9698 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9699 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9700 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
9701 scalar_64_32_long_scalar): Likewise.
9702 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9703 (cortex_a15_neon_mla_qqq_8_16): Likewise.
9704 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9705 lotype_qdd_64_32_long): Likewise.
9706 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9707 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9708 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
9709 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9710 (cortex_a15_neon_shift_1): Likewise.
9711 (cortex_a15_neon_shift_2): Likewise.
9712 (cortex_a15_neon_shift_3): Likewise.
9713 (cortex_a15_neon_vshl_ddd): Likewise.
9714 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9715 (cortex_a15_neon_vsra_vrsra): Likewise.
9716 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
9717 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
9718 (cortex_a15_neon_fp_vmul_ddd): Likewise.
9719 (cortex_a15_neon_fp_vmul_qqd): Likewise.
9720 (cortex_a15_neon_fp_vmla_ddd): Likewise.
9721 (cortex_a15_neon_fp_vmla_qqq): Likewise.
9722 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
9723 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
9724 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9725 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9726 (cortex_a15_neon_bp_simple): Likewise.
9727 (cortex_a15_neon_bp_2cycle): Likewise.
9728 (cortex_a15_neon_bp_3cycle): Likewise.
9729 (cortex_a15_neon_vld1_1_2_regs): Likewise.
9730 (cortex_a15_neon_vld1_3_4_regs): Likewise.
9731 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9732 (cortex_a15_neon_vld2_4_regs): Likewise.
9733 (cortex_a15_neon_vld3_vld4): Likewise.
9734 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9735 (cortex_a15_neon_vst1_3_4_regs): Likewise.
9736 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
9737 (cortex_a15_neon_vst3_vst4): Likewise.
9738 (cortex_a15_neon_vld1_vld2_lane): Likewise.
9739 (cortex_a15_neon_vld3_vld4_lane" 10
9740 (cortex_a15_neon_vst1_vst2_lane): Likewise.
9741 (cortex_a15_neon_vst3_vst4_lane): Likewise.
9742 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
9743 (cortex_a15_neon_ldm_2): Likewise.0
9744 (cortex_a15_neon_stm_2): Likewise.
9745 (cortex_a15_neon_mcr): Likewise.
9746 (cortex_a15_neon_mcr_2_mcrr): Likewise.
9747 (cortex_a15_neon_mrc): Likewise.
9748 (cortex_a15_neon_mrrc): Likewise.
9749 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
9750 change.
9751 (cortex_a15_alu_shift): Likewise.
9752 (cortex_a15_alu_shift_reg): Likewise.
9753 (cortex_a15_mult32): Likewise.
9754 (cortex_a15_mult64): Likewise.
9755 (cortex_a15_block): Likewise.
9756 (cortex_a15_branch): Likewise.
9757 (cortex_a15_load1): Likewise.
9758 (cortex_a15_load3): Likewise.
9759 (cortex_a15_store1): Likewise.
9760 (cortex_a15_store3): Likewise.
9761 (cortex_a15_call): Likewise.
9762 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
9763 (cortex_a5_f2r): Likewise.
9764 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
9765 change.
9766 (cortex_a53_f2r): Likewise.
9767 * config/arm/cortex-a7.md
9768 (cortex_a7_branch): Update for attribute change.
9769 (cortex_a7_call): Likewise.
9770 (cortex_a7_alu_imm): Likewise.
9771 (cortex_a7_alu_reg): Likewise.
9772 (cortex_a7_alu_shift): Likewise.
9773 (cortex_a7_mul): Likewise.
9774 (cortex_a7_load1): Likewise.
9775 (cortex_a7_store1): Likewise.
9776 (cortex_a7_load2): Likewise.
9777 (cortex_a7_store2): Likewise.
9778 (cortex_a7_load3): Likewise.
9779 (cortex_a7_store3): Likewise.
9780 (cortex_a7_load4): Likewise.
9781 (cortex_a7_store4): Likewise.
9782 (cortex_a7_fpalu): Likewise.
9783 (cortex_a7_fconst): Likewise.
9784 (cortex_a7_fpmuls): Likewise.
9785 (cortex_a7_neon_mul): Likewise.
9786 (cortex_a7_fpmacs): Likewise.
9787 (cortex_a7_neon_mla: Likewise.
9788 (cortex_a7_fpmuld: Likewise.
9789 (cortex_a7_fpmacd: Likewise.
9790 (cortex_a7_fpfmad: Likewise.
9791 (cortex_a7_fdivs: Likewise.
9792 (cortex_a7_fdivd: Likewise.
9793 (cortex_a7_r2f: Likewise.
9794 (cortex_a7_f2r: Likewise.
9795 (cortex_a7_f_flags: Likewise.
9796 (cortex_a7_f_loads: Likewise.
9797 (cortex_a7_f_loadd: Likewise.
9798 (cortex_a7_f_stores: Likewise.
9799 (cortex_a7_f_stored: Likewise.
9800 (cortex_a7_neon): Likewise.
9801 * config/arm/cortex-a8-neon.md
9802 (cortex_a8_neon_mrc): Update for attribute change.
9803 (cortex_a8_neon_mrrc): Likewise.
9804 (cortex_a8_neon_int_1): Likewise.
9805 (cortex_a8_neon_int_2): Likewise.
9806 (cortex_a8_neon_int_3): Likewise.
9807 (cortex_a8_neon_int_4): Likewise.
9808 (cortex_a8_neon_int_5): Likewise.
9809 (cortex_a8_neon_vqneg_vqabs): Likewise.
9810 (cortex_a8_neon_vmov): Likewise.
9811 (cortex_a8_neon_vaba): Likewise.
9812 (cortex_a8_neon_vaba_qqq): Likewise.
9813 (cortex_a8_neon_vsma): Likewise.
9814 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9815 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9816 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9817 long_scalar): Likewise.
9818 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9819 (cortex_a8_neon_mla_qqq_8_16): Likewise.
9820 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9821 long_scalar_qdd_64_32_long): Likewise.
9822 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9823 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9824 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
9825 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9826 (cortex_a8_neon_shift_1): Likewise.
9827 (cortex_a8_neon_shift_2): Likewise.
9828 (cortex_a8_neon_shift_3): Likewise.
9829 (cortex_a8_neon_vshl_ddd): Likewise.
9830 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9831 (cortex_a8_neon_vsra_vrsra): Likewise.
9832 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
9833 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
9834 (cortex_a8_neon_fp_vsum): Likewise.
9835 (cortex_a8_neon_fp_vmul_ddd): Likewise.
9836 (cortex_a8_neon_fp_vmul_qqd): Likewise.
9837 (cortex_a8_neon_fp_vmla_ddd): Likewise.
9838 (cortex_a8_neon_fp_vmla_qqq): Likewise.
9839 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
9840 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
9841 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9842 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9843 (cortex_a8_neon_bp_simple): Likewise.
9844 (cortex_a8_neon_bp_2cycle): Likewise.
9845 (cortex_a8_neon_bp_3cycle): Likewise.
9846 (cortex_a8_neon_ldr): Likewise.
9847 (cortex_a8_neon_str): Likewise.
9848 (cortex_a8_neon_vld1_1_2_regs): Likewise.
9849 (cortex_a8_neon_vld1_3_4_regs): Likewise.
9850 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9851 (cortex_a8_neon_vld2_4_regs): Likewise.
9852 (cortex_a8_neon_vld3_vld4): Likewise.
9853 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9854 (cortex_a8_neon_vst1_3_4_regs): Likewise.
9855 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
9856 (cortex_a8_neon_vst3_vst4): Likewise.
9857 (cortex_a8_neon_vld1_vld2_lane): Likewise.
9858 (cortex_a8_neon_vld3_vld4_lane): Likewise.
9859 (cortex_a8_neon_vst1_vst2_lane): Likewise.
9860 (cortex_a8_neon_vst3_vst4_lane): Likewise.
9861 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
9862 (cortex_a8_neon_mcr): Likewise.
9863 (cortex_a8_neon_mcr_2_mcrr): Likewise.
9864 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
9865 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
9866 change.
9867 (ca9_neon_mrrc): Likewise.
9868 (cortex_a9_neon_int_1): Likewise.
9869 (cortex_a9_neon_int_2): Likewise.
9870 (cortex_a9_neon_int_3): Likewise.
9871 (cortex_a9_neon_int_4): Likewise.
9872 (cortex_a9_neon_int_5): Likewise.
9873 (cortex_a9_neon_vqneg_vqabs): Likewise.
9874 (cortex_a9_neon_vmov): Likewise.
9875 (cortex_a9_neon_vaba): Likewise.
9876 (cortex_a9_neon_vaba_qqq): Likewise.
9877 (cortex_a9_neon_vsma): Likewise.
9878 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9879 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
9880 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
9881 long_scalar): Likewise.
9882 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
9883 (cortex_a9_neon_mla_qqq_8_16): Likewise.
9884 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
9885 long_scalar_qdd_64_32_long): Likewise.
9886 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
9887 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
9888 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
9889 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
9890 (cortex_a9_neon_shift_1): Likewise.
9891 (cortex_a9_neon_shift_2): Likewise.
9892 (cortex_a9_neon_shift_3): Likewise.
9893 (cortex_a9_neon_vshl_ddd): Likewise.
9894 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
9895 (cortex_a9_neon_vsra_vrsra): Likewise.
9896 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
9897 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
9898 (cortex_a9_neon_fp_vsum): Likewise.
9899 (cortex_a9_neon_fp_vmul_ddd): Likewise.
9900 (cortex_a9_neon_fp_vmul_qqd): Likewise.
9901 (cortex_a9_neon_fp_vmla_ddd): Likewise.
9902 (cortex_a9_neon_fp_vmla_qqq): Likewise.
9903 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
9904 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
9905 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
9906 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
9907 (cortex_a9_neon_bp_simple): Likewise.
9908 (cortex_a9_neon_bp_2cycle): Likewise.
9909 (cortex_a9_neon_bp_3cycle): Likewise.
9910 (cortex_a9_neon_ldr): Likewise.
9911 (cortex_a9_neon_str): Likewise.
9912 (cortex_a9_neon_vld1_1_2_regs): Likewise.
9913 (cortex_a9_neon_vld1_3_4_regs): Likewise.
9914 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
9915 (cortex_a9_neon_vld2_4_regs): Likewise.
9916 (cortex_a9_neon_vld3_vld4): Likewise.
9917 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
9918 (cortex_a9_neon_vst1_3_4_regs): Likewise.
9919 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
9920 (cortex_a9_neon_vst3_vst4): Likewise.
9921 (cortex_a9_neon_vld1_vld2_lane): Likewise.
9922 (cortex_a9_neon_vld3_vld4_lane): Likewise.
9923 (cortex_a9_neon_vst1_vst2_lane): Likewise.
9924 (cortex_a9_neon_vst3_vst4_lane): Likewise.
9925 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
9926 (cortex_a9_neon_mcr): Likewise.
9927 (cortex_a9_neon_mcr_2_mcrr): Likewise.
9928 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
9929 (cortex_a9_fps): Likewise.
9930 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
9931 change.
9932 (cortex_m4_fmuls): Likewise.
9933 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
9934 change.
9935 (cortex_r4_mrc): Likewise.
9936 * config/arm/iterators.md: Update comment referring to neon_type.
9937 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
9938 (iwmmxt_movsi_insn): Likewise.
9939 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
9940 attribute change.
9941 (pj4_core_to_vfp): Likewise.
9942 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
9943 attribute change.
9944 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
9945 (vfp_fstore): Likewise.
9946 * doc/md.texi: Change references to neon_type to refer to type.
9947
9948 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9949
9950 * tree.h (DECL_BUILT_IN): Fix typo in comment.
9951
9952 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
9953
9954 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
9955 lglobl if not weak.
9956
9957 2013-09-04 Easwaran Raman <eraman@google.com>
9958
9959 PR middle-end/57370
9960 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
9961 (build_and_add_sum): Use it.
9962 (appears_later_in_bb): Simplify code.
9963
9964 2013-09-04 Teresa Johnson <tejohnson@google.com>
9965
9966 * dumpfile.c (dump_finish): Don't close stderr/stdout.
9967
9968 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
9969
9970 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
9971
9972 2013-09-04 Jan Hubicka <jh@suse.cz>
9973
9974 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
9975 * ipa-devirt.c: Include diganostic.h
9976 (odr_type_d): Add types and types_set.
9977 (hash_type_name): Work for types with vtables during LTO.
9978 (odr_hasher::remove): Fix comment; destroy types_set.
9979 (add_type_duplicate): New function,
9980 (get_odr_type): Use it.
9981 (dump_type_inheritance_graph): Dump type duplicates.
9982 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
9983 graph.
9984 * tree.c (types_same_for_odr): Give exact answers on types with
9985 virtual tables.
9986
9987 2013-09-04 Dodji Seketeli <dodji@redhat.com>
9988
9989 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
9990 explaining their differences.
9991
9992 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
9993
9994 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
9995
9996 2013-09-03 Jeff Law <law@redhat.com>
9997
9998 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
9999 when not threading through a joiner block. Pass joiner/no joiner
10000 state to register_jump_thread.
10001 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
10002 state from argument rather than implying on path length.
10003 Dump the entire jump thread path into debugging dump.
10004 * tree-flow.h (register_jump_thread): Update prototype.
10005
10006 2013-08-29 Xinliang David Li <davidxl@google.com>
10007
10008 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10009 Remove a trivial gcc_assert.
10010
10011 2013-08-29 Xinliang David Li <davidxl@google.com>
10012
10013 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
10014 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
10015 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
10016 Delay base decl alignment adjustment.
10017 * tree-vectorizer.c (vect_destroy_datarefs): New function.
10018 * tree-vectorizer.h: New data structure.
10019 (set_dr_misalignment): New function.
10020 (dr_misalignment): Ditto.
10021 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
10022 (vectorizable_load): Ditto.
10023 (ensure_base_align): New function.
10024 (vectorize_loops): Add dbg_cnt support.
10025 (execute_vect_slp): Ditto.
10026 * dbgcnt.def: New debug counter.
10027 * Makefile: New dependency.
10028
10029 2013-09-03 Meador Inge <meadori@codesourcery.com>
10030
10031 Revert:
10032
10033 2013-08-30 Meador Inge <meadori@codesourcery.com>
10034
10035 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10036
10037 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
10038
10039 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
10040 function descriptor.
10041
10042 2013-09-03 Richard Biener <rguenther@suse.de>
10043
10044 * tree-affine.c (add_elt_to_tree): Fix association issue,
10045 avoid useless converts and make sure to always return a
10046 properly typed result.
10047
10048 2013-09-03 Richard Biener <rguenther@suse.de>
10049
10050 PR middle-end/57656
10051 * fold-const.c (negate_expr_p): Fix division case.
10052 (negate_expr): Likewise.
10053
10054 2013-09-03 Richard Biener <rguenther@suse.de>
10055
10056 PR lto/58285
10057 * tree-streamer-out.c: Include tm.h.
10058 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
10059
10060 2013-09-03 Jan Hubicka <jh@suse.cz>
10061
10062 * tree-profile.c (tree_profiling): Cleanup CFG when done.
10063
10064 2013-09-03 Alan Modra <amodra@gmail.com>
10065
10066 * config.gcc (powerpc*-*-linux*): Add support for little-endian
10067 multilibs to big-endian target and vice versa.
10068 * config/rs6000/t-linux64: Use := assignment on all vars.
10069 (MULTILIB_EXTRA_OPTS): Remove fPIC.
10070 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
10071 * config/rs6000/t-linux64le: New file.
10072 * config/rs6000/t-linux64bele: New file.
10073 * config/rs6000/t-linux64lebe: New file.
10074
10075 2013-09-02 Jan Hubicka <jh@suse.cz>
10076
10077 * ipa-inline-transform.c (inline_transform): Do not
10078 optimize_inline_calls when not optimizing.
10079
10080 2013-09-02 Jan Hubicka <jh@suse.cz>
10081
10082 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
10083 duplicated nodes for assembler names.
10084 * symtab.c (symtab_unregister_node): Do not attempt to unlink
10085 hard registers from assembler name hash.
10086
10087 2013-09-02 Jan Hubicka <jh@suse.cz>
10088
10089 * ipa-split.c (execute_split_functions): Split externally visible
10090 functions called once.
10091
10092 2013-09-02 Martin Jambor <mjambor@suse.cz>
10093
10094 PR ipa/58106
10095 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
10096 linked list. When finding the correct duplicate, also consider also
10097 the caller in additon to its inlined_to node.
10098
10099 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
10100
10101 * config/aarch64/aarch64-simd-builtins.def
10102 (dup_lane_scalar): Remove.
10103 * config/aarch64/aarch64-simd.md
10104 (aarch64_simd_dup): Add 'w->w' alternative.
10105 (aarch64_dup_lane<mode>): Allow for VALL.
10106 (aarch64_dup_lane_scalar<mode>): Remove.
10107 (aarch64_dup_lane_<vswap_width_name><mode>): New.
10108 (aarch64_get_lane_signed<mode>): Add w->w altenative.
10109 (aarch64_get_lane_unsigned<mode>): Likewise.
10110 (aarch64_get_lane<mode>): Likewise.
10111 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
10112 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
10113 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
10114 (VCON): Change container of V2SF.
10115 (vswap_width_name): Likewise.
10116 * config/aarch64/arm_neon.h
10117 (__aarch64_vdup_lane_any): New.
10118 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10119 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
10120 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
10121
10122 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
10123
10124 PR middle-end/56382
10125 * expr.c (emit_move_complex): Do not move complex FP values as parts if
10126 the source or the destination is a single hard register.
10127
10128 2013-09-02 Richard Biener <rguenther@suse.de>
10129
10130 PR middle-end/57511
10131 * tree-scalar-evolution.c (instantiate_scev_name): Allow
10132 non-linear SCEVs.
10133
10134 2013-09-02 Richard Biener <rguenther@suse.de>
10135
10136 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
10137 arithmetic to sizetype.
10138
10139 2013-09-02 Bin Cheng <bin.cheng@arm.com>
10140
10141 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
10142 Find auto-increment use both before and after candidate.
10143
10144 2013-09-02 Marek Polacek <polacek@redhat.com>
10145
10146 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
10147
10148 2013-09-01 Jan Hubicka <jh@suse.cz>
10149
10150 * Makefile.in: Add ipa-profile.o
10151 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
10152 * cgraph.c (struct cgraph_propagate_frequency_data,
10153 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
10154 ipa-profile.c; replace cgraph_ by ipa_ prefix.
10155 * cgraph.h (cgraph_propagate_frequency): Remove.
10156 * ipa-inline-analysis.c: Include ipa-utils.h;
10157 drop duplicated cfgloop.h.
10158 (inline_update_callee_summaries): Update.
10159 * ipa-profile.c: New file.
10160 * ipa-utils.h (ipa_propagate_frequency): Declare.
10161 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
10162 data-streamer.h, value-prof.h.
10163 (symtab_remove_unreachable_nodes): Update profile.
10164 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
10165 account_time_size, cmp_counts, dump_histogram,
10166 ipa_profile_generate_summary, ipa_profile_write_summary,
10167 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
10168 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
10169 Move to ipa-profile.c.
10170
10171 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
10172
10173 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
10174
10175 2013-09-01 Jan Hubicka <jh@suse.cz>
10176
10177 * common.opt (fdevirtualize-speculatively): New function.
10178 * invoke.texi (fdevirtualize-speculatively): Document.
10179 * ipa-devirt.c: Include ipa-inline.h
10180 (likely_target_p): New function.
10181 (ipa_devirt): New function.
10182 (gate_ipa_devirt): New function.
10183 (pass_data_ipa_devirt): New static var.
10184 (pass_ipa_devirt): Likewise.
10185 (make_pass_ipa_devirt): New function.
10186 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
10187 (common_handle_option): Disable devirtualization when
10188 value range profiling is available.
10189 * passes.def (pass_ipa_devirt): Add.
10190 * timever.def (TV_IPA_DEVIRT): New timevar.
10191 * tree-pass.h (make_pass_ipa_devirt):
10192
10193 2013-09-01 Iain Sandoe <iain@codesourcery.com>
10194
10195 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
10196 include sanitize(undefined).
10197
10198 2013-08-31 Diego Novillo <dnovillo@google.com>
10199
10200 * Makefile.in (TREE_CORE_H): Define.
10201 (TREE_H): Use.
10202 (GTFILES): Add tree-core.h.
10203 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
10204 size the array.
10205 * tree-core.h: New file.
10206 Move all data structures, enum, typedefs, global
10207 declarations and constants from ...
10208 * tree.h: ... here.
10209
10210 2013-08-31 Jan Hubicka <jh@suse.cz>
10211
10212 * bulitins.c (expand_builtin): Do not early exit for gcov
10213 instrumented functions.
10214
10215 2013-08-31 Marek Polacek <polacek@redhat.com>
10216
10217 * ubsan.c: Include tm_p.h.
10218
10219 2013-08-31 Jan Hubicka <jh@suse.cz>
10220
10221 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
10222 warning.
10223
10224 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
10225 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
10226 * tree-cfg.c (verify_gimple_label): ... here.
10227 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
10228 (ipa_merge_profiles): New function.
10229 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
10230 (lto_input_function_body): Likewise.
10231 * ipa-utils.h (ipa_merge_profiles): Declare.
10232 * lto-streamer.h (lto_input_function_body): Update prototype.
10233 (emit_label_in_global_context_p): Remove.
10234 * lto-symtab.c: Include ipa-utils.h
10235 (lto_cgraph_replace_node): Use ipa_merge_profiles.
10236
10237 2013-08-31 Jan Hubicka <jh@suse.cz>
10238
10239 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
10240
10241 2013-08-31 Jan Hubicka <jh@suse.cz>
10242
10243 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
10244
10245 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
10246
10247 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
10248 "cmp" RTX before signed_comparison_operator check to account
10249 for "code" changes.
10250
10251 2013-08-30 Jan Hubicka <jh@suse.cz>
10252
10253 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
10254 (detect_type_change_1): Rename to ...
10255 (detect_type_change): ... this one; early return on non-polymorphic
10256 types.
10257 (detect_type_change_ssa): Add comp_type parameter; update
10258 use of detect_type_change.
10259 (compute_complex_assign_jump_func): Add param_type parameter;
10260 update use of detect_type_change_ssa.
10261 (compute_complex_ancestor_jump_func): Likewise.
10262 (ipa_get_callee_param_type): New function.
10263 (ipa_compute_jump_functions_for_edge): Compute parameter type;
10264 update calls to the jump function computation functions.
10265
10266 2013-08-30 Teresa Johnson <tejohnson@google.com>
10267 Steven Bosscher <steven@gcc.gnu.org>
10268
10269 * cfgrtl.c (fixup_new_cold_bb): New routine.
10270 (commit_edge_insertions): Invoke fixup_partitions.
10271 (find_partition_fixes): New routine.
10272 (fixup_partitions): Ditto.
10273 (verify_hot_cold_block_grouping): Update comments.
10274 (rtl_verify_edges): Invoke find_partition_fixes.
10275 (rtl_verify_bb_pointers): Update comments.
10276 (rtl_verify_bb_layout): Ditto.
10277 * basic-block.h (probably_never_executed_edge_p): Declare.
10278 (fixup_partitions): Ditto.
10279 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
10280 * bb-reorder.c (sanitize_hot_paths): New function.
10281 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
10282 sanitize_hot_paths.
10283 * predict.c (probably_never_executed_edge_p): New routine.
10284 * cfg.c (check_bb_profile): Add partition insanity warnings.
10285
10286 2013-08-30 Meador Inge <meadori@codesourcery.com>
10287
10288 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
10289
10290 2013-08-30 Marek Polacek <polacek@redhat.com>
10291
10292 * Makefile.in (ubsan.o): Add.
10293 (c-family/c-ubsan.o): Add.
10294 (builtins.o): Add ubsan.h dependency.
10295 * ubsan.h: New file.
10296 * ubsan.c: New file.
10297 * common.opt: Add -fsanitize=undefined option.
10298 (flag_sanitize): Add variable.
10299 (fsanitize=): Add option. Add Driver.
10300 (fsanitize=thread): Remove option.
10301 (fsanitize=address): Likewise.
10302 (static-libubsan): New option.
10303 * doc/invoke.texi: Document the new flag and -static-libubsan.
10304 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
10305 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
10306 * builtin-attrs.def (ATTR_COLD): Define.
10307 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10308 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
10309 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
10310 * flag-types.h (sanitize_code): New enum.
10311 * opts.c (common_handle_option): Parse command line arguments
10312 of -fsanitize=. Add -fsanitize=unreachable option.
10313 * varasm.c (get_variable_section): Adjust.
10314 (assemble_noswitch_variable): Likewise.
10315 (assemble_variable): Likewise.
10316 (output_constant_def_contents): Likewise.
10317 (categorize_decl_for_section): Likewise.
10318 (place_block_symbol): Likewise.
10319 (output_object_block): Likewise.
10320 * builtins.def: Likewise.
10321 * toplev.c (compile_file): Likewise.
10322 (process_options): Likewise.
10323 * cppbuiltin.c: Likewise.
10324 * tsan.c (tsan_pass): Likewise.
10325 (tsan_gate): Likewise.
10326 (tsan_gate_O0): Likewise.
10327 * cfgexpand.c (partition_stack_vars): Likewise.
10328 (expand_stack_vars): Likewise.
10329 (defer_stack_allocation): Likewise.
10330 (expand_used_vars): Likewise.
10331 * cfgcleanup.c (old_insns_match_p): Likewise.
10332 * asan.c (asan_finish_file): Likewise.
10333 (asan_instrument): Likewise.
10334 (gate_asan): Likewise.
10335 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
10336 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
10337 (asan_global_struct): Use pointer_sized_int_node instead
10338 calling build_nonstandard_integer_type.
10339 (initialize_sanitizer_builtins): Likewise.
10340 (asan_finish_file): Likewise.
10341 * gcc.c: Document %{%:function(args):X}.
10342 (static_spec_functions): Add sanitize.
10343 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
10344 store funcval != NULL there.
10345 (do_spec_1): Adjust handle_spec_function caller.
10346 (handle_braces): Allow %:function(args) as condition.
10347 (sanitize_spec_function): New function.
10348 (ADD_STATIC_LIBUBSAN_LIBS): Define.
10349 (LIBUBSAN_SPEC): Likewise.
10350 (LIBUBSAN_EARLY_SPEC): Likewise.
10351 (SANITIZER_SPEC): Handle libubsan.
10352 (SANITIZER_EARLY_SPEC): Likewise.
10353 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
10354 instead of fsanitize=address.
10355 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
10356 instead of fsanitize=address*.
10357 * builtins.c: Include ubsan.h.
10358 (fold_builtin_0): Instrument __builtin_unreachable.
10359 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
10360 instead of flag_asan.
10361 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
10362 (pointer_sized_int_node): Define.
10363 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
10364
10365 2013-08-30 Mike Stump <mikestump@comcast.net>
10366
10367 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
10368 with RE patterns.
10369
10370 2013-08-29 Jan Hubicka <jh@suse.cz>
10371
10372 * cgraph.c (cgraph_function_body_availability): Handle weakref
10373 correctly.
10374 * passes.def: Remove pass_fixup_cfg.
10375 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
10376 track when we need to remove functions.
10377 (gate_ipa_inline): Execute inlining always; add comment why.
10378 (pass_data_ipa_inline): Remove TODO_remove_functions.
10379 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
10380 do not produce summaries.
10381 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
10382 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
10383 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
10384 constant pool and vtable.
10385
10386 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
10387
10388 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
10389 New arm_neon.h's internal macros to specify 64-bit constants.
10390 Avoid using stdint.h's macros.
10391
10392 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
10393
10394 * recog.c (verify_changes): Verify that changes[i].old is non-zero
10395 before applying REG_P.
10396
10397 2013-08-30 Jakub Jelinek <jakub@redhat.com>
10398
10399 PR tree-optimization/58277
10400 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
10401 after seeing too many stmts with vdef in between dombb and current
10402 bb, invalidate everything.
10403
10404 2013-08-30 Richard Biener <rguenther@suse.de>
10405
10406 * fold-const.c (fold_single_bit_test): Fix overflow test.
10407
10408 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
10409
10410 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
10411 and which can live in a register, always retrieve the value on entry.
10412 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
10413 passed by invisible reference specially.
10414 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
10415 (vt_add_function_parameter): Correctly deal with a parameter passed by
10416 invisible reference.
10417
10418 2013-08-30 Jan Hubicka <jh@suse.cz>
10419
10420 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
10421
10422 2013-08-30 Richard Biener <rguenther@suse.de>
10423
10424 PR tree-optimization/58228
10425 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
10426 allow invariant loads in nested loop vectorization.
10427
10428 2013-08-30 Richard Biener <rguenther@suse.de>
10429
10430 PR tree-optimization/58223
10431 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
10432 (has_anti_or_output_dependence): ... this and adjust to also
10433 look for output dependences.
10434 (mark_nodes_having_upstream_mem_writes): Adjust.
10435 (rdg_flag_uses): Likewise.
10436
10437 2013-08-30 Richard Biener <rguenther@suse.de>
10438
10439 PR tree-optimization/58010
10440 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
10441 assert that we have a loop-closed PHI.
10442
10443 2013-08-29 Jan Hubicka <jh@suse.cz>
10444
10445 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
10446 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
10447 * lto-section-in.c (lto_free_function_in_decl_state): New function.
10448 (lto_free_function_in_decl_state_for_node): New function.
10449
10450 2013-08-29 Xinliang David Li <davidxl@google.com>
10451
10452 * loop-unroll.c (report_unroll_peel): Minor message change.
10453 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
10454 Emit alignment peeling message with default -fopt-info.
10455 (vect_loop_versioning): Emit loop version info message.
10456 * tree-vectorizer.c (vectorize_loops): Minor message change.
10457 (execute_vect_slp): Ditto.
10458
10459 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
10460
10461 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
10462 of the clone from the DECL_NAME of the original function.
10463
10464 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
10465
10466 * passes.c (register_pass): Add overload.
10467 * tree-pass.h (register_pass): Forward declare it. Add comment.
10468
10469 2013-08-29 Jan Hubicka <jh@suse.cz>
10470
10471 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
10472 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
10473 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
10474 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10475 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
10476 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
10477 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
10478 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
10479
10480 2013-08-29 Teresa Johnson <tejohnson@google.com>
10481
10482 * dumpfile.c (dump_loc): Output column number.
10483 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
10484 * doc/invoke.texi: Document optall -fopt-info flag.
10485 * profile.c (read_profile_edge_counts): Use new dump framework.
10486 (compute_branch_probabilities): Ditto.
10487 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
10488 when pass not in any opt group.
10489 * pass_manager.h (pass_manager::get_pass_profile): New method.
10490 * value-prof.c (check_counter): Use new dump framework.
10491 (check_ic_target): Ditto.
10492 * coverage.c (get_coverage_counts): Ditto.
10493 (coverage_init): Setup new dump framework.
10494
10495 2013-08-29 Richard Biener <rguenther@suse.de>
10496
10497 PR tree-optimization/58246
10498 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
10499 handle the dominance check inside a basic-block.
10500
10501 2013-08-29 Richard Biener <rguenther@suse.de>
10502
10503 PR middle-end/57287
10504 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
10505 of default defs that appear in abnormal PHI nodes.
10506
10507 2013-08-29 Richard Biener <rguenther@suse.de>
10508
10509 PR tree-optimization/57685
10510 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
10511 single-use operands to avoid exponential complexity.
10512
10513 2013-08-28 Dehao Chen <dehao@google.com>
10514
10515 * ipa-inline.c (edge_badness): Fix integer underflow.
10516
10517 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
10518
10519 * gtm-builtins.def (_ITM_free): Declare leaf.
10520
10521 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10522
10523 PR target/58067
10524 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
10525 (*tls_local_dynamic_base_64_largepic): Likewise.
10526 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
10527 Remove predicate from call operand.
10528 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
10529 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
10530
10531 2013-08-28 Jeff Law <law@redhat.com>
10532
10533 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
10534 checks for the number of predecessors and successors allowed.
10535 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
10536 which require copying a joiner block if there is a request which
10537 is a subpath that requires no joiner block copying.
10538
10539 2013-08-28 Jan Hubicka <jh@suse.cz>
10540
10541 * lto-streamer-out.c (DFS_write_tree_body): Drop
10542 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
10543 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
10544 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
10545 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
10546 Do not read DECL_ERROR_ISSUED.
10547 (unpack_ts_decl_with_vis_value_fields): Do not read
10548 DECL_DEFER_OUTPUT.
10549 (lto_input_ts_binfo_tree_pointers): Do not read
10550 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
10551 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
10552 write DECL_ERROR_ISSUED..
10553 (pack_ts_decl_with_vis_value_fields): Do not write
10554 DECL_DEFER_OUTPUT.
10555 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
10556 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
10557 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
10558 * tree.h (tree_decl_common): Update comment.
10559 (DECL_ERROR_ISSUED): Remove.
10560
10561 2013-08-28 Jakub Jelinek <jakub@redhat.com>
10562
10563 PR middle-end/58257
10564 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
10565
10566 2013-08-28 Jan Hubicka <jh@suse.cz>
10567
10568 * builtins.def (free): Declare leaf.
10569
10570 2013-08-27 David Malcolm <dmalcolm@redhat.com>
10571
10572 * gdbhooks.py: New.
10573 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
10574 * configure: Regenerate.
10575
10576 2013-08-27 Martin Jambor <mjambor@suse.cz>
10577
10578 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
10579 (ipa_ancestor_jf_data): Likewise.
10580 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
10581 (ipa_get_jf_pass_through_type_preserved): New function.
10582 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
10583 (ipa_get_jf_ancestor_type_preserved): New function.
10584 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
10585 (ipa_get_jf_ancestor_result): Likewise.
10586 (propagate_vals_accross_pass_through): Use
10587 ipa_get_jf_pass_through_result to do all the value mappings.
10588 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
10589 type_preserved flag.
10590 (ipa_set_jf_cst_copy): New function.
10591 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
10592 (ipa_set_jf_arith_pass_through): Likewise.
10593 (ipa_set_ancestor_jf): Likewise.
10594 (compute_complex_assign_jump_func): Set type_preserved instead of
10595 punting.
10596 (ipa_compute_jump_functions_for_edge): Likewise.
10597 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
10598 (update_jump_functions_after_inlining): Update type_preserved.
10599 Explicitely create jump functions when combining one with pass_through.
10600 (ipa_write_jump_function): Stream the type_preserved flags.
10601 (ipa_read_jump_function): Likewise.
10602
10603 2013-08-27 Jakub Jelinek <jakub@redhat.com>
10604 Aldy Hernandez <aldyh@redhat.com>
10605
10606 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
10607 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
10608 * function.h (struct function): Add has_force_vect_loops and
10609 has_simduid_loops.
10610 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
10611 * gimple.c (gimple_build_omp_critical): Add KIND argument and
10612 handle it.
10613 * gimple.def: Update CLAUSES comments.
10614 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
10615 (gimple_build_omp_for): Add argument to prototype.
10616 (gimple_omp_for_kind): New.
10617 (gimple_omp_for_set_kind): New.
10618 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
10619 GOVD_DATA_SHARE_CLASS.
10620 (enum omp_region_type): Add ORT_SIMD.
10621 (gimple_add_tmp_var): Handle ORT_SIMD.
10622 (gimplify_var_or_parm_decl): Same.
10623 (is_gimple_stmt): Same.
10624 (omp_firstprivatize_variable): Same.
10625 (omp_add_variable): Only use splay_tree_insert if lookup failed.
10626 (omp_notice_variable): Handle ORT_SIMD.
10627 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
10628 (omp_check_private): Handle ORT_SIMD.
10629 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10630 OMP_CLAUSE_SAFELEN.
10631 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
10632 Handle OMP_CLAUSE_LASTPRIVATE.
10633 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
10634 OMP_CLAUSE_SAFELEN.
10635 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
10636 (gimplify_expr): Handle OMP_SIMD.
10637 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
10638 (expand_GOMP_SIMD_VF): New.
10639 (expand_GOMP_SIMD_LAST_LANE): New.
10640 * internal-fn.def (GOMP_SIMD_LANE): New.
10641 (GOMP_SIMD_VF): New.
10642 (GOMP_SIMD_LAST_LANE): New.
10643 * omp-low.c: Include target.h.
10644 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
10645 OMP_CLAUSE_SAFELEN.
10646 (check_omp_nesting_restrictions): Same.
10647 (omp_max_vf): New.
10648 (lower_rec_simd_input_clauses): New.
10649 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
10650 OMP_CLAUSE_LINEAR.
10651 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
10652 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
10653 (expand_omp_build_assign): New.
10654 (expand_omp_for_init_counts): New.
10655 (expand_omp_for_init_vars): New.
10656 (extract_omp_for_update_vars): New.
10657 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
10658 and rewrite accordingly.
10659 (expand_omp_simd): New.
10660 (expand_omp_for): Use expand_omp_simd.
10661 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
10662 (lower_omp_for): Do not lower the body.
10663 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
10664 in their own loops.
10665 * tree-flow.h (find_omp_clause): Remove prototype.
10666 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
10667 forcing vectorization of the loop, or if flag_tree_vectorize.
10668 (gate_tree_if_conversion): Similarly.
10669 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
10670 gimple_build_omp_for.
10671 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
10672 * tree-parloops (create_parallel_loop): Pass kind argument to
10673 gimple_build_omp_for.
10674 * tree-pretty-print.c (dump_omp_clause): Add cases for
10675 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
10676 OMP_CLAUSE__SIMDUID_.
10677 (dump_generic_node): Handle OMP_SIMD.
10678 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
10679 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
10680 unroll OMP_SIMD loops here.
10681 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
10682 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
10683 loop->safelen.
10684 (vect_analyze_data_refs): Handle simd loops.
10685 * tree-vect-loop.c (vectorizable_live_operation): Handle
10686 IFN_GOMP_SIMD*.
10687 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
10688 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
10689 (vectorizable_load): Same.
10690 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
10691 (struct simduid_to_vf): New.
10692 (simduid_to_vf::hash): New.
10693 (simduid_to-vf::equal): New.
10694 (struct simd_array_to_simduid): New.
10695 (simd_array_to_simduid::hash): New.
10696 (simd_array_to_simduid::equal): New.
10697 (adjust_simduid_builtins): New.
10698 (struct note_simd_array_uses_struct): New.
10699 (note_simd_array_uses_cb): New.
10700 (note_simd_array_uses): New.
10701 (vectorize_loops): Handle simd hints and adjust simd builtins
10702 accordingly.
10703 * tree-vectorizer.h (struct _stmt_vec_info): Add
10704 simd_lane_access_p field.
10705 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
10706 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
10707 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
10708 (omp_clause_code_name): Same.
10709 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
10710 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
10711 * tree.def (OMP_SIMD): New entry.
10712 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
10713 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
10714 (OMP_CLAUSE_DECL): Adjust range for new clauses.
10715 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
10716 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
10717 (OMP_CLAUSE_LINEAR_STEP): New.
10718 (OMP_CLAUSE_SAFELEN_EXPR): New.
10719 (OMP_CLAUSE__SIMDUID__DECL): New.
10720 (find_omp_clause): New prototype.
10721
10722 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
10723
10724 * config/i386/driver-i386.c (host_detect_local_cpu): Update
10725 Haswell processor detection.
10726
10727 2013-08-27 Christian Widmer <shadow@umbrox.de>
10728
10729 PR target/57927
10730 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
10731 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
10732 AVX2 capable processors.
10733
10734 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
10735
10736 * config/aarch64/arm_neon.h: Replace all inline asm implementations
10737 of vget_low_* with implementations in terms of other intrinsics.
10738
10739 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
10740
10741 PR middle-end/57219
10742 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
10743 values to -1, 0 and 1.
10744
10745 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
10746
10747 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
10748 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
10749 (<optab><mode>3_insn): Remove.
10750 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
10751 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
10752 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
10753 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
10754 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
10755 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
10756 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
10757 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
10758 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
10759 (ror<mode>3_insn): Likewise.
10760 * config/aarch64/predicates.md (aarch64_simd_register): New.
10761
10762 2013-08-27 Richard Biener <rguenther@suse.de>
10763
10764 PR tree-optimization/57521
10765 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
10766 one edge is non-critical.
10767 (find_phi_replacement_condition): Make sure to use a non-critical
10768 edge. Cleanup and remove old bug workarounds.
10769 (bb_postdominates_preds): Remove.
10770 (if_convertible_loop_p_1): Do not compute post-dominators.
10771 (combine_blocks): Do not free post-dominators.
10772 (main_tree_if_conversion): Likewise.
10773 (pass_data_if_conversion): Add TODO_verify_ssa.
10774
10775 2013-08-27 DJ Delorie <dj@redhat.com>
10776
10777 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
10778
10779 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
10780
10781 * function.c (assign_parm_find_data_types): Set passed_mode and
10782 nominal_mode to the TYPE_MODE of nominal_type for the built
10783 pointer type in case of the struct-pass-by-reference.
10784
10785 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10786
10787 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
10788 (UINT16_TYPE): Change default to "unsigned int".
10789
10790 * config/avr/avr.opt (mfract-convert-truncate): New option.
10791 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
10792 is set, round negative fractional integers according to n1169
10793 when converting to integer types.
10794
10795 2013-08-26 Jan Hubicka <jh@suse.cz>
10796
10797 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
10798 methods can not be called indirectly when their address is not taken.
10799
10800 2013-08-26 Jan Hubicka <jh@suse.cz>
10801
10802 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
10803 ctor_for_folding.
10804
10805 2013-08-26 Jan Hubicka <jh@suse.cz>
10806
10807 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
10808 can be unshared.
10809
10810 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10811
10812 * reload.c (find_valid_class): Allow classes that do not include
10813 FIRST_PSEUDO_REGISTER - 1.
10814
10815 2013-08-26 Jan Hubicka <jh@suse.cz>
10816
10817 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
10818 fix edge count/frequency when speculation failed; fix type check
10819 for the direct call.
10820
10821 2013-08-26 Jan Hubicka <jh@suse.cz>
10822
10823 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
10824
10825 2013-08-26 Jan Hubicka <jh@suse.cz>
10826
10827 * ipa-inline-transform.c (inline_transform): Be ready for basic block
10828 to be changed by edge redirection.
10829
10830 2013-08-26 Jan Hubicka <jh@suse.cz>
10831
10832 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
10833 formating; add sanity check.
10834 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
10835 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
10836 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
10837 (resolve_noninline_speculation): Update callee keys, too.
10838
10839 2013-08-26 Jan Hubicka <jh@suse.cz>
10840
10841 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
10842 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
10843
10844 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
10845
10846 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
10847 into proper place.
10848
10849 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
10850
10851 * config/i386/i386.c (ix86_debug_options): Remove prototype.
10852 (x86_64_elf_select_section): Ditto.
10853 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
10854 arguments.
10855 (ix86_pass_by_reference): Ditto.
10856 (output_set_got): Ditto.
10857 (ix86_unary_operator_ok): Ditto.
10858 (ix86_expand_builtin): Ditto.
10859
10860 2013-08-23 Jan Hubicka <jh@suse.cz>
10861
10862 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
10863
10864 2013-08-23 Jan Hubicka <jh@suse.cz>
10865
10866 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
10867 (tree_decl_with_vis): Add FINAL field.
10868
10869 2013-08-23 Jeff Law <law@redhat.com>
10870
10871 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
10872 the debugging dump when the expression is fully redundant.
10873
10874 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10875
10876 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
10877 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
10878 * pretty-print.c (pp_formatted_text_data): Likewise.
10879 (pp_write_text_to_stream): Likewise.
10880 (pp_write_text_as_dot_label_to_stream): Likewise.
10881 (pp_append_r): Likewise.
10882 (pp_format): Likewise.
10883 (pp_flush): Likewise.
10884 (pp_clear_output_area): Likewise.
10885 (pp_append_text): Likewise.
10886 (pp_formatted_text): Likewise.
10887 (pp_remaining_character_count_for_line): Likewise.
10888 (pp_newline): Likewise.
10889 (pp_character): Likewise.
10890 (output_buffer::~output_buffer): Define.
10891 (pretty_printer::~pretty_printer): Destruct output buffer.
10892 * pretty-print.h (output_buffer::~output_buffer): Declare.
10893 (pretty_printer::~pretty_printer): Declare virtual.
10894
10895 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
10896
10897 PR other/57324
10898 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
10899 HOST_WIDE_INT_M1U): New macros.
10900 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
10901 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
10902 unsigned -1 for lshift.
10903 * cse.c (cse_insn): Likewise.
10904 * double-int.c (rshift_double, lshift_double): Likewise.
10905 * builtins.c (fold_builtin_bitop): Likewise.
10906 * combine.c (force_to_mode): Likewise.
10907 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
10908 * simplify-rtx.c (simplify_const_unary_operation,
10909 simplify_const_binary_operation): Likewise.
10910 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
10911 check_va_list_escapes): Likewise.
10912 * rtlanal.c (nonzero_bits1): Likewise.
10913 * expmed.c (expand_smod_pow2): Likewise.
10914 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
10915
10916 2013-08-23 Jan Hubicka <jh@suse.cz>
10917
10918 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
10919 as having address taken.
10920
10921 2013-08-23 Jan Hubicka <jh@suse.cz>
10922
10923 * ipa-utils.h (method_class_type): Declare.
10924 * ipa-devirt.c (method_class_type): Export.
10925
10926 * cgraphunit.c (analyze_functions): Do basic devirtualization;
10927 do not walk base classes of anonymous types.
10928
10929 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
10930
10931 PR rtl-optimization/58220
10932 PR regression/58221
10933 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
10934 handle SEQUENCE insns properly.
10935
10936 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
10937
10938 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
10939 definition.
10940 (pp_newline_and_indent): Likewise.
10941 (pp_separate_with): Likewise.
10942 * pretty-print.c (pp_newline_and_flush): Define.
10943 (pp_newline_and_indent): Likewise.
10944 (pp_separate_with): Likewise.
10945
10946 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10947
10948 PR target/58218
10949 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
10950 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
10951
10952 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
10953
10954 * config/i386/predicates.md (ext_sse_reg_operand): New.
10955 * config/i386/i386.md (*movti_internal): Use
10956 predicate to determine if EVEX is needed.
10957 (*movsi_internal): Ditto.
10958 (*movdf_internal): Ditto.
10959 (*movsf_internal): Ditto.
10960 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
10961
10962 2013-08-23 Jakub Jelinek <jakub@redhat.com>
10963
10964 PR tree-optimization/58209
10965 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
10966 (find_tail_calls): Give up for pointer result types if m is non-NULL.
10967 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
10968 emit POINTER_PLUS_EXPR.
10969 (create_tailcall_accumulator): For pointer result type accumulate in
10970 sizetype type.
10971
10972 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
10973
10974 * configure.ac: Add backslashes missing from the last change.
10975 * configure: Regenerate.
10976
10977 2013-08-22 Jan Hubicka <jh@susue.cz>
10978
10979 * ipa.c (function_and_variable_visibility): First remember function
10980 was global and then make it local.
10981
10982 2013-08-22 Julian Brown <julian@codesourcery.com>
10983
10984 * configure.ac: Add aarch64 to list of arches which use "nop" in
10985 debug_line test.
10986 * configure: Regenerate.
10987
10988 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
10989
10990 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
10991 gnu_libc_has_function.
10992 * config/s390/tpf.h: Likewise.
10993
10994 2013-08-22 Jan Hubicka <jh@susue.cz>
10995
10996 * timevar.c (validate_phases): Add cast.
10997
10998 2013-08-22 Jan Hubicka <jh@susue.cz>
10999
11000 * timevar.c (validate_phases): Use size_t for memory.
11001 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
11002
11003 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
11004
11005 * pretty-print.h (output_buffer::output_buffer): Declare.
11006 (pretty_printer::pretty_printer): Likewise.
11007 (pp_construct): Remove.
11008 * pretty-print.c (output_buffer::output_buffer): Define.
11009 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
11010 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
11011 (print_gimple_expr): Likewise.
11012 (print_gimple_seq): Likewise.
11013 (gimple_dump_bb): Likewise.
11014 * sched-vis.c (dump_value_slim): Likewise.
11015 (dump_insn_slim): Likewise.
11016 (dump_rtl_slim): Likewise.
11017 (str_pattern_slim): Likewise.
11018 * tree-mudflap.c (mf_varname_tree): Likewise.
11019 * graph.c (print_graph_cfg): Likewise.
11020 (start_graph_dump): Likewise.
11021 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
11022 placement-new.
11023 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
11024 pretty printer initialization.
11025 * coretypes.h (diagnostic_context): Remove superflous type alias
11026 declaration.
11027 (pretty_printer): Likewise. Declare directly as a class.
11028 (pretty_print_info): Remove declaration as class.
11029 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
11030 and pp_clear_output_area.
11031 (asan_add_global): Likewise.
11032
11033 2013-08-22 Jan Hubicka <jh@suse.cz>
11034
11035 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
11036 * ipa-utils.h (update_type_inheritance_graph): Declare.
11037 (possible_polymorphic_call_target_p): Declare.
11038 (possible_polymorphic_call_target_p): New.
11039 * ipa-devirt.c: Update toplevel comments.
11040 (cached_polymorphic_call_targets): Move up.
11041 (odr_type_d): Move ID down.
11042 (polymorphic_type_binfo_p): Update comment.
11043 (odr_hasher::remove): Likewise;
11044 (get_odr_type): Set anonymous_namespace.
11045 (dump_odr_type): Dump it.
11046 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
11047 (maybe_record_node): Record node in cached_polymorphic_call_targets.
11048 (record_binfo): Add comment.
11049 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
11050 built.
11051 (devirt_node_removal_hook): Do not iCE when cache is freed.
11052 (possible_polymorphic_call_target_p): New predicate.
11053 (update_type_inheritance_graph): New function.
11054
11055 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
11056 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
11057 Sergey Lega <sergey.s.lega@intel.com>
11058 Anna Tikhonova <anna.tikhonova@intel.com>
11059 Ilya Tocar <ilya.tocar@intel.com>
11060 Andrey Turetskiy <andrey.turetskiy@intel.com>
11061 Ilya Verbin <ilya.verbin@intel.com>
11062 Kirill Yukhin <kirill.yukhin@intel.com>
11063 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
11064
11065 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
11066 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
11067 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
11068 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
11069 (OPTION_MASK_ISA_AVX2_UNSET): Update.
11070 (OPTION_MASK_ISA_AVX512F_UNSET): New.
11071 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
11072 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
11073 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
11074 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
11075 OPT_mavx512pf, OPT_mavx512er cases.
11076 * config/i386/constraints.md (v): New constraint.
11077 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
11078 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
11079 (bit_AVX512CD): New.
11080 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
11081 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
11082 * config/i386/i386-c.c (ix86_target_macros_internal):
11083 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
11084 __AVX512PF__.
11085 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
11086 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
11087 * config/i386/i386.c (regclass_map, dbx_register_map)
11088 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
11089 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
11090 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
11091 -mavx512pf options.
11092 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
11093 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
11094 -mavx512pf options. Fix formatting.
11095 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
11096 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
11097 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
11098 -mavx512cd, -mavx512pf options.
11099 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
11100 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
11101 (ix86_preferred_output_reload_class): Replace SSE_REGS with
11102 ALL_SSE_REGS.
11103 (ix86_hard_regno_mode_ok): Support 512-bit registers.
11104 (ix86_set_reg_reg_cost): Ditto.
11105 (x86_order_regs_for_local_alloc): Ditto.
11106 (MAX_VECT_LEN): Extend to 64-byte.
11107 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
11108 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
11109 (TARGET_AVX512ER, TARGET_AVX512CD): New.
11110 (BIGGEST_ALIGNMENT): Extend to 512-bits.
11111 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
11112 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
11113 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
11114 (SSE_REG_MODE_P): Support new modes.
11115 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
11116 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
11117 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
11118 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
11119 (REG_CLASS_CONTENTS): Add new registers.
11120 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
11121 (EXT_REX_SSE_REGNO_P): New.
11122 (HI_REGISTER_NAMES): Add new registers.
11123 * config/i386/i386.md: Define constants for new registers.
11124 (mode): Add new 512-bit modes.
11125 (prefix): Support evex prefix.
11126 (isa): Support avx512f, noavx512f, fma_avx512f.
11127 (ssemodesuffix): Add new 512-bit modes.
11128 (movxi): New.
11129 (*movxi_internal_avx512f): Ditto.
11130 (*movdi_internal): Replace constraint "x" with the new constraint "v".
11131 Support MODE_XI.
11132 (*movsi_internal): Likewise.
11133 (*movdf_internal): Likewise.
11134 (*movsf_internal): Likewise.
11135 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
11136 (<code><mode>3): Likewise.
11137 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
11138 New.
11139 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
11140 with the new constraint "v".
11141 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
11142 modes.
11143 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
11144 with the new constraint "v".
11145 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
11146 (<sse2>_storedqu<avxsizesuffix>): Likewise.
11147 (*<plusminus_insn><mode>3): Likewise.
11148 (<sse>_vm<plusminus_insn><mode>3): Likewise.
11149 (*mul<mode>3): Likewise.
11150 (<sse>_vmmul<mode>3): Likewise.
11151 (<sse>_div<mode>3): Likewise.
11152 (<sse>_vmdiv<mode>3): Likewise.
11153 (<sse>_sqrt<mode>2): Likewise.
11154 (<sse>_vmsqrt<mode>2): Likewise.
11155 (*<code><mode>3_finite): Likewise.
11156 (*<code><mode>3) <smaxmin>: Likewise.
11157 (<sse>_vm<code><mode>3): Likewise.
11158 (*<code><mode>3) <any_logic>: Likewise.
11159 (*fma_fmadd_<mode>): Likewise.
11160 (*fma_fmsub_<mode>): Likewise.
11161 (*fma_fnmadd_<mode>): Likewise.
11162 (*fma_fnmsub_<mode>): Likewise.
11163 (*fma_fmaddsub_<mode>): Likewise.
11164 (*fma_fmsubadd_<mode>): Likewise.
11165 (*fmai_fmadd_<mode>): Likewise.
11166 (*fmai_fmsub_<mode>): Likewise.
11167 (*fmai_fnmadd_<mode>): Likewise.
11168 (*fmai_fnmsub_<mode>): Likewise.
11169 (sse_cvtsi2ss): Likewise.
11170 (sse_cvtsi2ssq): Likewise.
11171 (sse_cvtss2si): Likewise.
11172 (sse_cvtss2si_2): Likewise.
11173 (sse_cvtss2siq): Likewise.
11174 (sse_cvtss2siq_2): Likewise.
11175 (sse_cvttss2si): Likewise.
11176 (sse_cvtss2siq_2): Likewise.
11177 (float<sseintvecmodelower><mode>2): Likewise.
11178 (sse2_cvtsd2si_2): Likewise.
11179 (sse2_cvtsd2siq_2): Likewise.
11180 (*<plusminus_insn><mode>3): Likewise.
11181 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
11182 (*<sse4_1_avx2>_mul<mode>3): Likewise.
11183 (ashr<mode>3): Likewise.
11184 (<shift_insn><mode>3): Likewise.
11185 (avx2_<code><mode>3): Likewise.
11186 (*avx2_<code><mode>3): Likewise.
11187 (*andnot<mode>3): Likewise.
11188 (*<code><mode>3) <any_logic>: Likewise.
11189 (abs<mode>2): Likewise.
11190 (avx2_permvar<mode>): Likewise.
11191 (avx2_perm<mode>_1): Likewise.
11192 (*avx_vpermilp<mode>): Likewise.
11193 (avx_vpermilvar<mode>3): Likewise.
11194 (avx2_ashrv<mode>): Likewise.
11195 (avx2_<shift_insn>v<mode>): Likewise.
11196 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
11197 -mavx512cd.
11198 * doc/rtl.texi: Document XImode.
11199
11200 2013-08-21 Jeff Law <law@redhat.com>
11201
11202 * tree-flow.h (register_jump_thread): Pass vector of edges
11203 instead of each important edge.
11204 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
11205 thread path into a vector and pass that to register_jump_thread.
11206 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
11207 passed in edge vector to the current 3-edge form.
11208
11209 Revert:
11210 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11211
11212 * dce.c (fini_dce): Call df_analyze again just in case
11213 delete_unmarked_insns removed anything.
11214
11215 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
11216
11217 * reload.h (struct reg_equivs): Rename to ..
11218 (struct reg_equivs_s): .. this.
11219
11220 2013-08-20 Martin Liska <marxin.liska@gmail.com>
11221
11222 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
11223
11224 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
11225
11226 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
11227
11228 2013-08-21 Jeff Law <law@redhat.com>
11229
11230 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
11231 simplify assignments too. If the RHS collapses to a singleton
11232 range, then return the value for the range.
11233
11234 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
11235
11236 * config/i386/sse.md (V16): Rename to...
11237 (VMOVE): this.
11238 (mov<mode>): Update iterator name.
11239 (*mov<mode>_internal): Ditto.
11240 (push<mode>1): Ditto.
11241 (movmisalign<mode>): Ditto.
11242
11243 2013-08-20 Jan Hubicka <jh@suse.cz>
11244
11245 PR bootstrap/58186
11246 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
11247 entry for direct edges.
11248 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
11249
11250 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11251
11252 Revert my last two changes, r201865 and r201864:
11253
11254 Revert r201865:
11255 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11256
11257 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11258 instances can own GC refs.
11259
11260 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11261 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11262 (gcc::context::gt_pch_nx): Likewise.
11263 (gcc::context::gt_pch_nx): Likewise.
11264 * ggc.h (gt_ggc_mx <T>): New.
11265 (gt_pch_nx_with_op <T>): New.
11266 (gt_pch_nx <T>): New.
11267 * passes.c (opt_pass::gt_ggc_mx): New.
11268 (opt_pass::gt_pch_nx): New.
11269 (opt_pass::gt_pch_nx_with_op): New.
11270 (pass_manager::gt_ggc_mx): New.
11271 (pass_manager::gt_pch_nx): New.
11272 (pass_manager::gt_pch_nx_with_op): New.
11273 (pass_manager::operator new): Use
11274 ggc_internal_cleared_alloc_stat rather than xcalloc.
11275 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11276 (pass_manager::gt_ggc_mx): New.
11277 (pass_manager::gt_pch_nx): New.
11278 (pass_manager::gt_pch_nx_with_op): New.
11279 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11280 (opt_pass::operator new): New.
11281 (opt_pass::gt_ggc_mx): New.
11282 (opt_pass::gt_pch_nx): New.
11283 (opt_pass::gt_pch_nx_with_op): New.
11284
11285 Revert r201864:
11286 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11287
11288 * Makefile.in (GTFILES): Add context.h.
11289 * context.c (gcc::context::operator new): New.
11290 (gcc::context::gt_ggc_mx): New.
11291 (gcc::context::gt_pch_nx): New.
11292 (gcc::context::gt_pch_nx): New.
11293 * context.h (gcc::context): Add GTY((user)) marking.
11294 (gcc::context::operator new): New.
11295 (gcc::context::gt_ggc_mx): New.
11296 (gcc::context::gt_pch_nx): New.
11297 (gcc::context::gt_pch_nx): New.
11298 (g): Add GTY marking.
11299 (gt_ggc_mx (gcc::context *)): New.
11300 (gt_pch_nx (gcc::context *)): New.
11301 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11302 void *cookie)): New.
11303 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11304
11305 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
11306
11307 * dce.c (fini_dce): Call df_analyze again just in case
11308 delete_unmarked_insns removed anything.
11309
11310 2013-08-20 Teresa Johnson <tejohnson@google.com>
11311
11312 PR rtl-optimizations/57451
11313 * final.c (reemit_insn_block_notes): Prevent lexical blocks
11314 from crossing split section boundaries.
11315
11316 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
11317
11318 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
11319 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
11320 with MULTLIB_DEFAULTS.
11321
11322 2013-08-20 Nick Clifton <nickc@redhat.com>
11323
11324 * target.def (narrow_volatile_bitfield): Note that the default
11325 value is false, not !TARGET_STRICT_ALIGN.
11326 * doc/tm.texi: Regenerate.
11327
11328 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
11329
11330 Fix LIB_SPEC for systems without libpthread.
11331
11332 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
11333 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
11334 for Android.
11335 * config/i386/linux-common.h: Likewise.
11336 * config/mips/linux-common.h: Likewise.
11337
11338 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
11339
11340 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
11341 checks.
11342
11343 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11344
11345 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
11346 instances can own GC refs.
11347
11348 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
11349 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
11350 (gcc::context::gt_pch_nx): Likewise.
11351 (gcc::context::gt_pch_nx): Likewise.
11352 * ggc.h (gt_ggc_mx <T>): New.
11353 (gt_pch_nx_with_op <T>): New.
11354 (gt_pch_nx <T>): New.
11355 * passes.c (opt_pass::gt_ggc_mx): New.
11356 (opt_pass::gt_pch_nx): New.
11357 (opt_pass::gt_pch_nx_with_op): New.
11358 (pass_manager::gt_ggc_mx): New.
11359 (pass_manager::gt_pch_nx): New.
11360 (pass_manager::gt_pch_nx_with_op): New.
11361 (pass_manager::operator new): Use
11362 ggc_internal_cleared_alloc_stat rather than xcalloc.
11363 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
11364 (pass_manager::gt_ggc_mx): New.
11365 (pass_manager::gt_pch_nx): New.
11366 (pass_manager::gt_pch_nx_with_op): New.
11367 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
11368 (opt_pass::operator new): New.
11369 (opt_pass::gt_ggc_mx): New.
11370 (opt_pass::gt_pch_nx): New.
11371 (opt_pass::gt_pch_nx_with_op): New.
11372
11373 2013-08-20 David Malcolm <dmalcolm@redhat.com>
11374
11375 * Makefile.in (GTFILES): Add context.h.
11376 * context.c (gcc::context::operator new): New.
11377 (gcc::context::gt_ggc_mx): New.
11378 (gcc::context::gt_pch_nx): New.
11379 (gcc::context::gt_pch_nx): New.
11380 * context.h (gcc::context): Add GTY((user)) marking.
11381 (gcc::context::operator new): New.
11382 (gcc::context::gt_ggc_mx): New.
11383 (gcc::context::gt_pch_nx): New.
11384 (gcc::context::gt_pch_nx): New.
11385 (g): Add GTY marking.
11386 (gt_ggc_mx (gcc::context *)): New.
11387 (gt_pch_nx (gcc::context *)): New.
11388 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
11389 void *cookie)): New.
11390 * gengtype.c (open_base_files) <ifiles>: Add context.h.
11391
11392 2013-08-20 Alan Modra <amodra@gmail.com>
11393
11394 PR target/57865
11395 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
11396 (rs6000_emit_epilogue): Likewise.
11397
11398 2013-08-19 Dehao Chen <dehao@google.com>
11399
11400 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
11401
11402 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
11403 Jakub Jelinek <jakub@redhat.com>
11404
11405 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
11406 (BUILT_IN_FABSD64): Likewise.
11407 (BUILT_IN_FABSD128): Likewise.
11408 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
11409 (fold_builtin_1): Likewise.
11410 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
11411 destination and source operands.
11412 (*abstd2_fpr): Likewise.
11413 (*nabstd2_fpr): Likewise.
11414
11415 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
11416
11417 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
11418 JUMP_P and INSN_P.
11419
11420 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
11421
11422 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
11423 incomplete.
11424
11425 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
11426
11427 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
11428 * builtins.c (default_libc_has_function): New.
11429 (gnu_libc_has_function): Ditto.
11430 (no_c99_libc_has_function): Ditto.
11431 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
11432 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
11433 (fold_builtin_sincos): Likewise.
11434 (fold_builtin_cexp): Likewise.
11435 * builtins.def (DEF_C94_BUILTIN): Likewise.
11436 (DEF_C99_BUILTIN): Likewise.
11437 (DEF_C99_C90RES_BUILTIN): Likewise.
11438 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
11439 definitions to using this define.
11440 * config/darwin-protos.h (darwin_libc_has_function): New.
11441 * config/darwin.c (darwin_libc_has_function): Ditto.
11442 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
11443 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
11444 * config/darwin.h: Ditto.
11445 * config/elfos.h: Ditto.
11446 * config/freebsd.h: Ditto.
11447 * config/i386/cygming.h: Ditto.
11448 * config/i386/djgpp.h: Ditto.
11449 * config/i386/i386-interix.h: Ditto.
11450 * config/microblaze/microblaze.h: Ditto.
11451 * config/mmix/mmix.h: Ditto.
11452 * config/gnu-user.h: Ditto.
11453 * config/ia64/hpux.h: Ditto.
11454 * config/pa/pa-hpux.h: Ditto.
11455 * config/pdp11/pdp11.h: Ditto.
11456 * config/picochip/picochip.h: Ditto.
11457 * config/linux.h: Ditto.
11458 * config/netbsd.h: Ditto.
11459 * config/openbsd.h: Ditto.
11460 * config/rs6000/aix43.h: Ditto.
11461 * config/rs6000/aix51.h: Ditto.
11462 * config/rs6000/aix52.h: Ditto.
11463 * config/rs6000/aix53.h: Ditto.
11464 * config/rs6000/aix61.h: Ditto.
11465 * config/rs6000/darwin.h: Ditto.
11466 * config/rs6000/linux.h: Ditto.
11467 * config/rs6000/linux64.h: Ditto.
11468 * config/s390/tpf.h: Ditto.
11469 * config/sol2-10.h: Ditto.
11470 * config/sol2.h: Ditto.
11471 * config/vms/vms.h: Ditto.
11472 * config/vxworks.h: Ditto.
11473 * config/linux-android.c (linux_android_libc_has_function):
11474 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
11475 * config/linux-protos.h (linux_android_libc_has_function):
11476 New declaration.
11477 * config/i386/i386.c (ix86_libc_has_function): New.
11478 * config/i386/i386-protos.h
11479 (ix86_libc_has_function): New declaration.
11480 * config/i386/i386.md
11481 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
11482 ("isinf<mode>2): Likewise.
11483 * convert.c (convert_to_integer): Using new target hook
11484 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
11485 TARGET_C99_FUNCTIONS.
11486 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
11487 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
11488 * coretypes.h (function_class): New enum for different
11489 classes of functions.
11490 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
11491 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
11492 (TARGET_HAS_SINCOS): Likewise.
11493 (TARGET_LIBC_HAS_FUNCTION): New.
11494 * doc/tm.texi: Regenerated.
11495 * targhooks.h (default_libc_has_function): New declaration.
11496 (no_c99_libc_has_function): Ditto.
11497 (gnu_libc_has_function): Ditto.
11498 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
11499 and TARGET_HAS_SINCOS.
11500
11501 2013-08-18 Jan Hubicka <jh@suse.cz>
11502
11503 * Makeifle-in (ipa-devirt.o): New.
11504 (GTFILES): Add ipa-utils.h and ipa-devirt.c
11505 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
11506 (analyze_functions): Look into possible targets of polymorphic call.
11507 * dumpfile.c (dump_files): Add type-inheritance dump.
11508 * dumpfile.h (TDI_inheritance): New.
11509 * ipa-devirt.c: New file.
11510 * ipa-utils.h (odr_type_d): Forward declare.
11511 (odr_type): New type.
11512 (build_type_inheritance_graph): Declare.
11513 (possible_polymorphic_call_targets): Declare and introduce inline
11514 variant when only edge is pased.
11515 (dump_possible_polymorphic_call_targets): Likewise.
11516 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
11517 * tree.c (type_in_anonymous_namespace_p): Break out from ...
11518 (types_same_for_odr): ... here.
11519 * tree.h (type_in_anonymous_namespace_p): Declare.
11520
11521 2013-08-18 Jakub Jelinek <jakub@redhat.com>
11522
11523 PR tree-optimization/58006
11524 * tree-parloops.c (take_address_of): Don't ICE if get_name
11525 returns NULL.
11526 (eliminate_local_variables_stmt): Remove clobber stmts.
11527
11528 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
11529
11530 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
11531 error message is issued for an alias to undefined symbol.
11532
11533 2013-08-18 Jan Hubicka <jh@suse.cz>
11534
11535 * cgraph.c (cgraph_create_indirect_edge): Discover
11536 polymorphic calls and record basic info into indirect_info.
11537 * gimple-fold.c (gimple_fold_call): When doing BINFO based
11538 devirtualization, ignore objc function calls.
11539 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
11540 call with no parm index info.
11541 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
11542 * tree.c (virtual_method_call_p): New function.
11543 * tree.h (virtual_method_call_p): Declare.
11544
11545 2013-08-16 Jan Hubicka <jh@suse.cz>
11546
11547 PR middle-end/58179
11548 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
11549
11550 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
11551
11552 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
11553 attribute.
11554
11555 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11556
11557 * gengtype.c (type_for_name): Add special-case support for
11558 locating types within the "gcc::" namespace.
11559 (open_base_files): Emit a "using namespace gcc" directive.
11560
11561 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
11562
11563 PR target/58160
11564 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
11565 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
11566
11567 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
11568 array instead of each individual operand as a separate argument.
11569 (emit_fusion_gpr_load): Likewise.
11570 (expand_fusion_gpr_load): Add new function declaration.
11571
11572 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
11573 signature to have the operands passed as an array, instead of as
11574 separate arguments. Allow ZERO_EXTEND to be in the memory
11575 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
11576 depend on the register live/dead flags when peepholes are run.
11577 (expand_fusion_gpr_load): New function to be called from the
11578 peephole2 pass, to change the register that addis sets to be the
11579 target register.
11580 (emit_fusion_gpr_load): Change the calling signature to have the
11581 operands passed as an array, instead of as separate arguments.
11582 Allow ZERO_EXTEND to be in the memory address, and also
11583 SIGN_EXTEND if -mpower8-fusion-sign.
11584
11585 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
11586 unspec enumeration.
11587 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
11588 adjust the register addis loads up in the peephole2 pass. Do not
11589 depend on the register live/dead state when the peephole pass is done.
11590
11591 2013-08-16 David Malcolm <dmalcolm@redhat.com>
11592
11593 * gengtype.c (create_user_defined_type): Ensure that the kind
11594 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
11595 declaration is seen before the GTY((user)) marking.
11596
11597 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
11598
11599 PR target/58105
11600 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
11601
11602 2013-08-16 Jan Hubicka <jh@suse.cz>
11603
11604 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
11605 arugment expected_type.
11606 (gimple_fold_call): Use it.
11607 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
11608 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
11609 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
11610 (try_make_edge_direct_virtual_call): Likewise.
11611 * tree.c (obj_type_ref_class): New.
11612 * tree.h (obj_type_ref_class): Use it.
11613
11614 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
11615
11616 * sched-vis.c (rtl_slim_pp_initialized): Remove.
11617 (rtl_slim_pp): Likewise.
11618 (init_rtl_slim_pretty_print): Likewise.
11619 (dump_value_slim): Don't call it. Use local pretty printer.
11620 (dump_insn_slim): Likewise.
11621 (dump_rtl_slim): Likewise.
11622 (str_pattern_slim): Likewise.
11623 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
11624 Simplify.
11625
11626 2013-08-16 Jakub Jelinek <jakub@redhat.com>
11627
11628 PR tree-optimization/58164
11629 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
11630 walk gimple_goto_dest of GIMPLE_GOTO.
11631
11632 PR tree-optimization/58165
11633 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
11634 bi_call must be the last stmt in a bb, don't split_block, instead
11635 use fallthru edge from it and give up if there is none.
11636 Release conds vector when returning early.
11637
11638 2013-08-14 Xinliang David Li <davidxl@google.com>
11639
11640 * config/i386/i386.c (ix86_option_override_internal):
11641 Remove unused variable and field.
11642
11643 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11644
11645 PR target/57949
11646 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
11647 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
11648 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
11649 and Linux, correct BLKmode alignment when 128-bit alignment is
11650 required and compatibility flag is not set.
11651 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
11652 for zero-size arguments when compatibility flag is not set.
11653
11654 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11655
11656 PR tree-optimization/58145
11657 * tree-sra.c (build_ref_for_offset): If prev_base has
11658 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
11659
11660 2013-08-14 Xinliang David Li <davidxl@google.com>
11661
11662 * config/i386/i386.c (ix86_option_override_internal):
11663 Fix uninitialized variable error.
11664
11665 2013-08-14 Xinliang David Li <davidxl@google.com>
11666
11667 * config/i386/i386.opt: Define two new options.
11668 * config/i386/x86-tune.def: Add arch selector field in macros.
11669 * config/i386/i386.h: Adjust macro definition.
11670 * config/i386/i386.c (ix86_option_override_internal):
11671 Refactor the code.
11672 (parse_mtune_ctrl_str): New function.
11673 (set_ix86_tune_features): New function.
11674 (ix86_function_specific_restore): Call the new helper function.
11675
11676 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
11677
11678 PR rtl-optimization/57662
11679 * sel-sched.c (code_motion_process_successors): When the current insn
11680 is removed after the recursive traversal, break from the loop.
11681 Add comments and debug printouts.
11682
11683 2013-08-14 Jakub Jelinek <jakub@redhat.com>
11684 Alexandre Oliva <aoliva@redhat.com>
11685
11686 PR target/58067
11687 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
11688 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
11689 there also UNSPEC_PLTOFF.
11690
11691 2013-08-14 Marek Polacek <polacek@redhat.com>
11692
11693 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
11694 NUM_CONDITIONS bit positions.
11695
11696 2013-08-13 Cary Coutant <ccoutant@google.com>
11697
11698 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
11699 (attr_checksum): Hash vector contents instead of pointer.
11700 (attr_checksum_ordered): Likewise.
11701
11702 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
11703
11704 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
11705 when Pmode != word_mode. Add length_address attribute.
11706 (sse3_monitor_<mode>): Merge from sse3_monitor and
11707 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
11708 Pmode != word_mode. Update insn length attribute.
11709 * config/i386/i386.c (ix86_option_override_internal): Update
11710 ix86_gen_monitor selection for merged sse3_monitor insn.
11711
11712 2013-08-13 Julian Brown <julian@codesourcery.com>
11713
11714 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
11715 perform invalid legitimization on greater-than-word-size modes for
11716 TARGET_E500_DOUBLE.
11717
11718 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
11719
11720 * ira.c (setup_class_translate_array): Use aclass instead of cl
11721 for classes not fully covered by allocno classes.
11722
11723 2013-08-13 Jakub Jelinek <jakub@redhat.com>
11724
11725 PR tree-optimization/57661
11726 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
11727 * tree-inline.c (tree_function_versioning): Initialize it.
11728 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
11729 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
11730 that is not being copied.
11731
11732 PR sanitizer/56417
11733 * asan.c (instrument_strlen_call): Fix typo in comment.
11734 Use char * type even for the lhs of POINTER_PLUS_EXPR.
11735
11736 2013-08-13 Steve Ellcey <sellcey@mips.com>
11737
11738 * config/mips/mips.md (prefetch): Use lw instead of ld on
11739 loongson in 32bit mode.
11740
11741 2013-08-13 Nick Clifton <nickc@redhat.com>
11742
11743 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
11744
11745 2013-08-13 Jan Hubicka <jh@suse.cz>
11746
11747 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
11748 introduced edge; fix typo in sanity check.
11749 (cgraph_resolve_speculation): Export; improve diagnostic.
11750 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
11751 speculation at type mismatch.
11752 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
11753 (cgraph_resolve_speculation): Declare.
11754 (symtab_can_be_discarded): New function.
11755 * value-prof.c (gimple_ic_transform): Remove actual transform code.
11756 * ipa-inline-transform.c (speculation_removed): New global var.
11757 (clone_inlined_nodes): See if speculation can be removed.
11758 (inline_call): If speculations was removed, we growths may not match.
11759 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
11760 (speculation_useful_p): New function.
11761 (resolve_noninline_speculation): New function.
11762 (inline_small_functions): Resolve useless speculations.
11763 * ipa-inline.h (speculation_useful_p): Declare
11764 * ipa.c (can_replace_by_local_alias): Simplify.
11765 (ipa_profile): Produce speculative calls in non-lto, too;
11766 add simple cost model; produce local aliases.
11767
11768 2013-08-13 David Malcolm <dmalcolm@redhat.com>
11769
11770 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
11771 PASS_MANAGER_H.
11772
11773 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
11774
11775 * config/i386/i386.c (ix86_function_versions): Use error + inform.
11776
11777 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
11778
11779 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
11780 iterator instead of X87MODEF.
11781
11782 2013-08-12 Perez Read <netfirewall@gmail.com>
11783
11784 PR target/58132
11785 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
11786 operand 0 for intel asm alternative.
11787 (*movabs<mode>_2): Ditto for operand 1.
11788
11789 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
11790
11791 * config/aarch64/arm_none.h
11792 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
11793
11794 2013-08-12 Nick Clifton <nickc@redhat.com>
11795
11796 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
11797
11798 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
11799
11800 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
11801 expand for QI/HImode operand to produce more effictive code for
11802 unsigned char(short) --> float(double) conversion.
11803
11804 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
11805
11806 * doc/invoke.texi: Mention that -ftls-model does not force the final
11807 model.
11808
11809 2013-08-12 Marek Polacek <polacek@redhat.com>
11810 Marc Glisse <marc.glisse@inria.fr>
11811
11812 PR tree-optimization/57980
11813 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
11814 when creating -1 constant.
11815
11816 2013-08-10 Jan Hubicka <jh@suse.cz>
11817
11818 Workaround binutils PR14342.
11819 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
11820 (gimple_init_edge_profiler): Likewise.
11821 (gimple_gen_ic_func_profiler): Likewise.
11822
11823 2013-08-09 Jan Hubicka <jh@suse.cz>
11824
11825 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
11826
11827 2013-08-09 Xinliang David Li <davidxl@google.com>
11828
11829 * config/i386/stringop.def: New file.
11830 * config/i386/stringop.opt: New file.
11831 * config/i386/i386-opts.h: Include stringopt.def.
11832 * config/i386/i386.opt: Include stringopt.opt.
11833 * config/i386/i386.c (ix86_option_override_internal):
11834 Override default size based stringop inline strategies with options.
11835 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
11836 New function.
11837
11838 2013-08-09 Jan Hubicka <jh@suse.cz>
11839
11840 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
11841
11842 2013-08-09 Jan Hubicka <jh@suse.cz>
11843
11844 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
11845 CGRAPH_FREQ_MAX.
11846 (dump_cgraph_node): Dump profile-id.
11847 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
11848 and common_target_probability.
11849 * lto-cgraph.c (lto_output_edge): Stream common targets.
11850 (lto_output_node): Stream profile ids.
11851 (input_node): Stream profile ids.
11852 (input_edge): Stream common targets.
11853 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
11854 * ipa.c: Include value-prof.h
11855 (ipa_profile_generate_summary): Turn indirect call statement histograms
11856 into common targets.
11857 (ipa_profile): Turn common targets into speculative edges.
11858
11859 2013-08-09 Jan Hubicka <jh@suse.cz>
11860
11861 * cgraph.h (cgraph_node): Add profile_id.
11862 * value-prof.c (cgraph_node_map): Turn into pointer_map.
11863 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
11864 (del_node_map): Update.
11865 (find_func_by_funcdef_no): Replace by ...
11866 (find_func_by_profile_id): ... this one.
11867 (gimple_ic_transform): Do not remove useful histograms when
11868 speculation is not done; dump info when indirect call removal
11869 can happen at LTO.
11870 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
11871 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
11872 (__gcov_indirect_call_profiler_v2): .. this one.
11873 * profile.h (init_node_map): Update.
11874 * coverage.c (coverage_compute_profile_id): New function.
11875 * coverage.h (coverage_compute_profile_id): Declare.
11876 * tree-profile.c (init_ic_make_global_vars): Make
11877 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
11878 (gimple_init_edge_profiler): Update prototype of
11879 __gcov_indirect_call_profiler.
11880 (gimple_gen_ic_func_profiler): Simplify.
11881 (tree_profiling): Use init_node_map
11882
11883 2013-08-09 Jan Hubicka <jh@suse.cz>
11884
11885 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
11886 non-speculative refs.
11887 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
11888 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
11889 (cgraph_set_call_stmt): Likewise.
11890 (cgraph_create_edge_1): Fix release checking compilatoin;
11891 clear lto_stmt_uid.
11892 (cgraph_free_edge): Free indirect info.
11893 (cgraph_turn_edge_to_speculative): New function.
11894 (cgraph_speculative_call_info): New function.
11895 (cgraph_make_edge_direct): Return direct edge; handle speculation.
11896 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
11897 (dump_cgraph_node): Dump speculation.
11898 (verify_edge_count_and_frequency): Accept speculative edges.
11899 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
11900 (verify_cgraph_node): Handle speculation.
11901 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
11902 (cgraph_set_call_stmt): Update prototype.
11903 (cgraph_make_edge_direct): Update prototype.
11904 (cgraph_speculative_call_info): Declare.
11905 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
11906 to change; update call of ipa_find_references.
11907 * ipa-ref.c (ipa_record_reference): Fix return value; clear
11908 lto_stmt_uid and speculative flags.
11909 (ipa_dump_references): Dump speculation.
11910 (ipa_clone_references): Clone speculative flag.
11911 (ipa_clone_referring): Likewise.
11912 (ipa_clone_ref): New function.
11913 (ipa_find_reference): Look into lto_stmt_uids
11914 (ipa_clear_stmts_in_references): Do not clear speculative calls.
11915 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
11916 (ipa_find_reference): Update declaration.
11917 (ipa_clone_ref): Declare.
11918 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
11919 stream speculative flag.
11920 (lto_output_ref): Stream statements uids and speculation.
11921 (input_ref): Likewise.
11922 (input_edge): Stream speuclation.
11923 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
11924 (cgraph_set_call_stmt_including_clones): Handle speculation.
11925 * ipa-inline.c (heap_edge_removal_hook): New function.
11926 (inline_small_functions): Register it.
11927 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
11928 also initialize refs.
11929 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
11930 edge to change.
11931 (try_make_edge_direct_simple_call): Likewise.
11932 (try_make_edge_direct_simple_call): Likewise.
11933 (update_indirect_edges_after_inlining): Likewise.
11934 (remove_described_reference): Look proper lto_stmt_uid.
11935 (propagate_controlled_uses): Likewise.
11936 (propagate_controlled_uses): Liekwise.
11937 * tree-inline.c (copy_bb): Copy speculative edges.
11938 (redirect_all_calls): New function.
11939 (copy_cfg_body): Do redirection after loop info is updated.
11940 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
11941
11942 2013-08-09 Jan Hubicka <jh@suse.cz>
11943
11944 * lto-streamer-out.c (output_function): Renumber PHIs.
11945 * lto-streamer-in.c (input_function): Likewise.
11946
11947 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
11948
11949 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
11950 (get_lane_unsigned): Likewise.
11951 (dup_lane_scalar): Likewise.
11952 (get_lane): enable for VALL.
11953 * config/aarch64/aarch64-simd.md
11954 (aarch64_dup_lane_scalar<mode>): Remove.
11955 (aarch64_get_lane_signed<mode>): Likewise.
11956 (aarch64_get_lane_unsigned<mode>): Likewise.
11957 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
11958 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
11959 (aarch64_get_lane<mode>): Enable for all vector modes.
11960 (aarch64_get_lanedi): Remove misleading constraints.
11961 * config/aarch64/arm_neon.h
11962 (__aarch64_vget_lane_any): Define.
11963 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
11964 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
11965 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
11966 * config/aarch64/iterators.md (VDQQH): New.
11967 (VDQQHS): Likewise.
11968 (vwcore): Likewise.
11969
11970 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
11971
11972 * configure.ac: Add GAS check for LEON instructions on SPARC.
11973 * configure: Regenerate.
11974 * config.in: Likewise.
11975 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
11976 sparc*-*-* block.
11977 * config/sparc/sparc.opt (LEON, LEON3): New masks.
11978 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
11979 for LEON or LEON3.
11980 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
11981 (AS_LEON_FLAG): New macro.
11982 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
11983 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
11984 Deal with LEON and LEON3 for the memory model.
11985 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
11986 (atomic_compare_and_swap<mode>_1): Likewise.
11987 (*atomic_compare_and_swap<mode>_1): Likewise.
11988
11989 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
11990
11991 * config/arm/neon.md (vcond): Fix floating-point vector
11992 comparisons against 0.
11993
11994 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
11995
11996 * lra-constraints.c (emit_spill_move): Remove assert.
11997 (process_alt_operands): Add more debugging
11998 output. Increase reject for spilling into memory. Decrease
11999 reject for reloading scratch.
12000 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
12001
12002 2013-08-08 Steve Ellcey <sellcey@mips.com>
12003
12004 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
12005 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
12006 micromips incompatible. Add nan2008.
12007 (MULTILIB_DIRNAMES): Add nan2008.
12008 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12009 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
12010 and micromips incompatible. Add nan2008.
12011 (MULTILIB_DIRNAMES): Add nan2008.
12012 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
12013
12014 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
12015
12016 PR rtl-optimization/58079
12017 * combine.c (combine_simplify_rtx): Avoid using SUBST if
12018 simplify_comparison has widened a comparison with an integer.
12019
12020 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12021
12022 * config/arm/neon.md (movmisalign<mode>): Disable when we
12023 don't allow unaligned accesses.
12024 (*movmisalign<mode>_neon_store): Likewise.
12025 (*movmisalign<mode>_neon_load): Likewise.
12026 (*movmisalign<mode>_neon_store): Likewise.
12027 (*movmisalign<mode>_neon_load): Likewise.
12028
12029 2013-08-08 Jan Hubicka <jh@suse.cz>
12030
12031 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
12032 (make_pass_rebuild_cgraph_edges): Also clear references.
12033 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
12034 * ipa-inline-transform.c (inline_transform): Remove all references
12035 after inlining.
12036 * cgraphunit.c (expand_function): Remove all references after
12037 expansion.
12038 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
12039 (ipa_find_reference): Rewrite to iterator.
12040 (remove_stmt_references): Likewise.
12041 (ipa_clear_stmts_in_references): New function.
12042 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
12043 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
12044 clear references.
12045 * ipa-split.c (split_function): Remove references in split function.
12046
12047 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12048
12049 PR target/57431
12050 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
12051 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
12052
12053 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
12054
12055 PR target/56979
12056 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
12057 suggested mode for the assignment isn't compatible with the
12058 registers required.
12059
12060 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
12061
12062 PR target/58065
12063 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
12064
12065 2013-08-07 Xinliang David Li <davidxl@google.com>
12066
12067 * config/i386/i386.opt: New option -mtune-ctrl=.
12068 * config/i386/x86-tune.def: New file.
12069 * config/i386/i386.h: include x86-tune.def.
12070 * config/i386/i386.c (ix86_option_override_internal):
12071 Parsing -mtune-ctrl= option and set tune features.
12072
12073 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
12074
12075 PR other/12081
12076 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
12077 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
12078 to gen_2arg_fn_t.
12079
12080 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12081
12082 * rtl.h (update_alignments): Declare.
12083 * final.c (grow_label_align): New function extracted from...
12084 (shorten_branches): ...here. Call it.
12085 (update_alignments): New function.
12086 * reorg.c (sibling_labels): New variable.
12087 (get_label_before): Add SIBLING parameter. If it is non-zero, push
12088 the new label along with it onto the sibling_labels vector.
12089 (fill_simple_delay_slots): Adjust call to get_label_before.
12090 (fill_slots_from_thread): Likewise.
12091 (relax_delay_slots): Likewise.
12092 (make_return_insns): Likewise.
12093 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
12094
12095 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
12096
12097 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
12098 document its semantics.
12099 (diagnostic_report_diagnostic): Adjust accordingly.
12100
12101 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12102
12103 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
12104 (sparc_option_override): ...and port to new C++ pass API.
12105 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
12106
12107 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
12108
12109 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
12110
12111 2013-08-06 Caroline Tice <cmtice@google.com>
12112
12113 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
12114 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
12115 * tree-pass.h: Add pass_vtable_verify.
12116 * varasm.c (assemble_variable): Add code to properly set the comdat
12117 section and name for the .vtable_map_vars section.
12118 (assemble_vtyv_preinit_initializer): New function.
12119 (default_sectin_type_flags): Make sure .vtable_map_vars section has
12120 LINK_ONCE flag.
12121 * output.h: Add function decl for assemble_vtv_preinit_initializer.
12122 * vtable-verify.c: New file.
12123 * vtable-verify.h: New file.
12124 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
12125 initialiation levels.
12126 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
12127 * passes.def: Insert pass_vtable_verify.
12128 * aclocal.m4: Reorder includes.
12129 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
12130 -fvtv-counts options.
12131 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
12132 as appropriate, if -fvtable-verify=... is used.
12133 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
12134 -fvtable-verify=... is used.
12135 * Makefile.in (OBJS): Add vtable-verify.o to list.
12136 (vtable-verify.o): Add new build rule.
12137 (GTFILES): Add vtable-verify.c to list.
12138 * common.opt (fvtable-verify=): New flag.
12139 (vtv_priority): Values for fvtable-verify= flag.
12140 (fvtv-counts): New flag.
12141 (fvtv-debug): New flag.
12142 * tree.h (save_vtable_map_decl): New extern function decl.
12143
12144 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12145
12146 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
12147 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
12148 (pass_data_rl78_devirt): ...new pass_data instance and...
12149 (make_pass_rl78_devirt): ...new function.
12150 (rl78_asm_file_start): Port pass registration to new C++ API.
12151
12152 2013-08-07 David Malcolm <dmalcolm@redhat.com>
12153
12154 * coretypes.h (rtl_opt_pass): Add.
12155 (gcc::context): Add.
12156 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
12157 (epiphany_init): Port to new C++ pass API.
12158 (epiphany_optimize_mode_switching): Likewise.
12159 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
12160 (pass_manager::get_pass_mode_switching): New.
12161 (pass_manager::get_pass_peephole2): New.
12162 * mode-switching.c (pass_mode_switching): Add clone method.
12163 * recog.c (pass_peephole2): Add clone method.
12164 (pass_split_all_insns): Add clone method.
12165
12166 2013-08-06 David Malcolm <dmalcolm@redhat.com>
12167
12168 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
12169 (mips_option_override): ...here, porting to new C++ API for passes.
12170
12171 2013-08-06 Jan Hubicka <jh@suse.cz>
12172
12173 * cgraph.c (cgraph_get_body): New function based on lto.c
12174 implementation.
12175 * cgraph.h (cgraph_get_body): Declare.
12176 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
12177 LTO paths.
12178 * cgraphunit.c (expand_function): Get body prior expanding.
12179 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
12180 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
12181 really need.
12182 * passes.c (do_per_function_toporder): Get body.
12183 * tree-inline.c (expand_call_inline): Get body prior inlining it.
12184 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
12185
12186 2013-08-06 Martin Jambor <mjambor@suse.cz>
12187
12188 PR fortran/57987
12189 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
12190 is not re-finalized. Rename second parameter to no_collect.
12191
12192 2013-08-06 Martin Jambor <mjambor@suse.cz>
12193
12194 PR middle-end/58041
12195 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
12196 MEM_REF has proper alignment information.
12197
12198 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
12199
12200 PR other/12081
12201 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
12202 class insn_gen_fn.
12203 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
12204 rtx (*) (rtx, ...) with insn_gen_fn.
12205 * genoutput.c (output_insn_data): Cast gen_? function pointers to
12206 insn_gen_fn::stored_funcptr. Add initializer braces.
12207
12208 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12209
12210 Rewrite how instances of passes are cloned to remove assumptions
12211 about their sizes (thus allowing pass subclasses to have
12212 additional data fields, albeit non-GC-managed ones at this point).
12213
12214 * passes.c (make_pass_instance): Now that passes have clone
12215 methods, rewrite this function to eliminate XNEW and memcpy
12216 calls that used hardcoded sizes. Since this function no longer
12217 creates pass instances, rename it to...
12218 (add_pass_instance): ...this. Document the old way that passes were
12219 numbered and flagged, and rework this function to continue using it.
12220 (next_pass_1): Add an initial_pass argument for use by
12221 add_pass_instance.
12222 (position_pass): When adding multiple instances of a pass, use
12223 the pass's clone method, rather than relying on the XNEW/memcpy
12224 within the former make_pass_instance (now add_pass_instance).
12225 (pass_manager::pass_manager): When invoking next_pass_1, also supply
12226 the initial instance of the current pass within the pass manager.
12227
12228 2013-08-05 David Malcolm <dmalcolm@redhat.com>
12229
12230 This is the automated part of the conversion of passes from C
12231 structs to C++ classes.
12232
12233 Patch autogenerated by refactor_passes.py from
12234 https://github.com/davidmalcolm/gcc-refactoring-scripts
12235 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
12236
12237 * asan.c (pass_asan): Convert from a global struct to a subclass of
12238 gimple_opt_pass along with...
12239 (pass_data_asan): ...new pass_data instance and...
12240 (make_pass_asan): ...new function.
12241 (pass_asan_O0): Convert from a global struct to a subclass of
12242 gimple_opt_pass along with...
12243 (pass_data_asan_O0): ...new pass_data instance and...
12244 (make_pass_asan_O0): ...new function.
12245 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
12246 subclass of rtl_opt_pass along with...
12247 (pass_data_inc_dec): ...new pass_data instance and...
12248 (make_pass_inc_dec): ...new function.
12249 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
12250 a subclass of rtl_opt_pass along with...
12251 (pass_data_reorder_blocks): ...new pass_data instance and...
12252 (make_pass_reorder_blocks): ...new function.
12253 (pass_duplicate_computed_gotos): Convert from a global struct to a
12254 subclass of rtl_opt_pass along with...
12255 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
12256 (make_pass_duplicate_computed_gotos): ...new function.
12257 (pass_partition_blocks): Convert from a global struct to a subclass of
12258 rtl_opt_pass along with...
12259 (pass_data_partition_blocks): ...new pass_data instance and...
12260 (make_pass_partition_blocks): ...new function.
12261 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
12262 struct to a subclass of rtl_opt_pass along with...
12263 (pass_data_branch_target_load_optimize1): ...new pass_data instance
12264 and...
12265 (make_pass_branch_target_load_optimize1): ...new function.
12266 (pass_branch_target_load_optimize2): Convert from a global struct to a
12267 subclass of rtl_opt_pass along with...
12268 (pass_data_branch_target_load_optimize2): ...new pass_data instance
12269 and...
12270 (make_pass_branch_target_load_optimize2): ...new function.
12271 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
12272 of rtl_opt_pass along with...
12273 (pass_data_jump): ...new pass_data instance and...
12274 (make_pass_jump): ...new function.
12275 (pass_jump2): Convert from a global struct to a subclass of
12276 rtl_opt_pass along with...
12277 (pass_data_jump2): ...new pass_data instance and...
12278 (make_pass_jump2): ...new function.
12279 * cfgexpand.c (pass_expand): Convert from a global struct to a
12280 subclass of rtl_opt_pass along with...
12281 (pass_data_expand): ...new pass_data instance and...
12282 (make_pass_expand): ...new function.
12283 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
12284 of rtl_opt_pass along with...
12285 (pass_data_free_cfg): ...new pass_data instance and...
12286 (make_pass_free_cfg): ...new function.
12287 (pass_into_cfg_layout_mode): Convert from a global struct to a
12288 subclass of rtl_opt_pass along with...
12289 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
12290 (make_pass_into_cfg_layout_mode): ...new function.
12291 (pass_outof_cfg_layout_mode): Convert from a global struct to a
12292 subclass of rtl_opt_pass along with...
12293 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
12294 (make_pass_outof_cfg_layout_mode): ...new function.
12295 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
12296 struct to a subclass of gimple_opt_pass along with...
12297 (pass_data_build_cgraph_edges): ...new pass_data instance and...
12298 (make_pass_build_cgraph_edges): ...new function.
12299 (pass_rebuild_cgraph_edges): Convert from a global struct to a
12300 subclass of gimple_opt_pass along with...
12301 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
12302 (make_pass_rebuild_cgraph_edges): ...new function.
12303 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
12304 subclass of gimple_opt_pass along with...
12305 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
12306 and...
12307 (make_pass_remove_cgraph_callee_edges): ...new function.
12308 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
12309 struct to a subclass of rtl_opt_pass along with...
12310 (pass_data_stack_adjustments): ...new pass_data instance and...
12311 (make_pass_stack_adjustments): ...new function.
12312 * combine.c (pass_combine): Convert from a global struct to a subclass
12313 of rtl_opt_pass along with...
12314 (pass_data_combine): ...new pass_data instance and...
12315 (make_pass_combine): ...new function.
12316 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
12317 global struct to a subclass of rtl_opt_pass along with...
12318 (pass_data_compare_elim_after_reload): ...new pass_data instance
12319 and...
12320 (make_pass_compare_elim_after_reload): ...new function.
12321 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
12322 of rtl_opt_pass along with...
12323 (pass_data_rtl_cprop): ...new pass_data instance and...
12324 (make_pass_rtl_cprop): ...new function.
12325 * cse.c (pass_cse): Convert from a global struct to a subclass of
12326 rtl_opt_pass along with...
12327 (pass_data_cse): ...new pass_data instance and...
12328 (make_pass_cse): ...new function.
12329 (pass_cse2): Convert from a global struct to a subclass of
12330 rtl_opt_pass along with...
12331 (pass_data_cse2): ...new pass_data instance and...
12332 (make_pass_cse2): ...new function.
12333 (pass_cse_after_global_opts): Convert from a global struct to a
12334 subclass of rtl_opt_pass along with...
12335 (pass_data_cse_after_global_opts): ...new pass_data instance and...
12336 (make_pass_cse_after_global_opts): ...new function.
12337 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
12338 of rtl_opt_pass along with...
12339 (pass_data_ud_rtl_dce): ...new pass_data instance and...
12340 (make_pass_ud_rtl_dce): ...new function.
12341 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
12342 rtl_opt_pass along with...
12343 (pass_data_fast_rtl_dce): ...new pass_data instance and...
12344 (make_pass_fast_rtl_dce): ...new function.
12345 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
12346 a subclass of rtl_opt_pass along with...
12347 (pass_data_df_initialize_opt): ...new pass_data instance and...
12348 (make_pass_df_initialize_opt): ...new function.
12349 (pass_df_initialize_no_opt): Convert from a global struct to a
12350 subclass of rtl_opt_pass along with...
12351 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
12352 (make_pass_df_initialize_no_opt): ...new function.
12353 (pass_df_finish): Convert from a global struct to a subclass of
12354 rtl_opt_pass along with...
12355 (pass_data_df_finish): ...new pass_data instance and...
12356 (make_pass_df_finish): ...new function.
12357 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
12358 rtl_opt_pass along with...
12359 (pass_data_rtl_dse1): ...new pass_data instance and...
12360 (make_pass_rtl_dse1): ...new function.
12361 (pass_rtl_dse2): Convert from a global struct to a subclass of
12362 rtl_opt_pass along with...
12363 (pass_data_rtl_dse2): ...new pass_data instance and...
12364 (make_pass_rtl_dse2): ...new function.
12365 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
12366 subclass of rtl_opt_pass along with...
12367 (pass_data_dwarf2_frame): ...new pass_data instance and...
12368 (make_pass_dwarf2_frame): ...new function.
12369 * except.c (pass_set_nothrow_function_flags): Convert from a global
12370 struct to a subclass of rtl_opt_pass along with...
12371 (pass_data_set_nothrow_function_flags): ...new pass_data instance
12372 and...
12373 (make_pass_set_nothrow_function_flags): ...new function.
12374 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
12375 subclass of rtl_opt_pass along with...
12376 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
12377 and...
12378 (make_pass_convert_to_eh_region_ranges): ...new function.
12379 * final.c (pass_compute_alignments): Convert from a global struct to a
12380 subclass of rtl_opt_pass along with...
12381 (pass_data_compute_alignments): ...new pass_data instance and...
12382 (make_pass_compute_alignments): ...new function.
12383 (pass_final): Convert from a global struct to a subclass of
12384 rtl_opt_pass along with...
12385 (pass_data_final): ...new pass_data instance and...
12386 (make_pass_final): ...new function.
12387 (pass_shorten_branches): Convert from a global struct to a subclass of
12388 rtl_opt_pass along with...
12389 (pass_data_shorten_branches): ...new pass_data instance and...
12390 (make_pass_shorten_branches): ...new function.
12391 (pass_clean_state): Convert from a global struct to a subclass of
12392 rtl_opt_pass along with...
12393 (pass_data_clean_state): ...new pass_data instance and...
12394 (make_pass_clean_state): ...new function.
12395 * function.c (pass_instantiate_virtual_regs): Convert from a global
12396 struct to a subclass of rtl_opt_pass along with...
12397 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
12398 (make_pass_instantiate_virtual_regs): ...new function.
12399 (pass_leaf_regs): Convert from a global struct to a subclass of
12400 rtl_opt_pass along with...
12401 (pass_data_leaf_regs): ...new pass_data instance and...
12402 (make_pass_leaf_regs): ...new function.
12403 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
12404 subclass of rtl_opt_pass along with...
12405 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
12406 and...
12407 (make_pass_thread_prologue_and_epilogue): ...new function.
12408 (pass_match_asm_constraints): Convert from a global struct to a
12409 subclass of rtl_opt_pass along with...
12410 (pass_data_match_asm_constraints): ...new pass_data instance and...
12411 (make_pass_match_asm_constraints): ...new function.
12412 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
12413 subclass of rtl_opt_pass along with...
12414 (pass_data_rtl_fwprop): ...new pass_data instance and...
12415 (make_pass_rtl_fwprop): ...new function.
12416 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
12417 rtl_opt_pass along with...
12418 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
12419 (make_pass_rtl_fwprop_addr): ...new function.
12420 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
12421 rtl_opt_pass along with...
12422 (pass_data_rtl_pre): ...new pass_data instance and...
12423 (make_pass_rtl_pre): ...new function.
12424 (pass_rtl_hoist): Convert from a global struct to a subclass of
12425 rtl_opt_pass along with...
12426 (pass_data_rtl_hoist): ...new pass_data instance and...
12427 (make_pass_rtl_hoist): ...new function.
12428 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
12429 subclass of gimple_opt_pass along with...
12430 (pass_data_lower_cf): ...new pass_data instance and...
12431 (make_pass_lower_cf): ...new function.
12432 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
12433 from a global struct to a subclass of gimple_opt_pass along with...
12434 (pass_data_strength_reduction): ...new pass_data instance and...
12435 (make_pass_strength_reduction): ...new function.
12436 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
12437 of rtl_opt_pass along with...
12438 (pass_data_rtl_ifcvt): ...new pass_data instance and...
12439 (make_pass_rtl_ifcvt): ...new function.
12440 (pass_if_after_combine): Convert from a global struct to a subclass of
12441 rtl_opt_pass along with...
12442 (pass_data_if_after_combine): ...new pass_data instance and...
12443 (make_pass_if_after_combine): ...new function.
12444 (pass_if_after_reload): Convert from a global struct to a subclass of
12445 rtl_opt_pass along with...
12446 (pass_data_if_after_reload): ...new pass_data instance and...
12447 (make_pass_if_after_reload): ...new function.
12448 * init-regs.c (pass_initialize_regs): Convert from a global struct to
12449 a subclass of rtl_opt_pass along with...
12450 (pass_data_initialize_regs): ...new pass_data instance and...
12451 (make_pass_initialize_regs): ...new function.
12452 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
12453 of ipa_opt_pass_d along with...
12454 (pass_data_ipa_cp): ...new pass_data instance and...
12455 (make_pass_ipa_cp): ...new function.
12456 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
12457 global struct to a subclass of gimple_opt_pass along with...
12458 (pass_data_inline_parameters): ...new pass_data instance and...
12459 (make_pass_inline_parameters): ...new function.
12460 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
12461 subclass of gimple_opt_pass along with...
12462 (pass_data_early_inline): ...new pass_data instance and...
12463 (make_pass_early_inline): ...new function.
12464 (pass_ipa_inline): Convert from a global struct to a subclass of
12465 ipa_opt_pass_d along with...
12466 (pass_data_ipa_inline): ...new pass_data instance and...
12467 (make_pass_ipa_inline): ...new function.
12468 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
12469 struct to a subclass of gimple_opt_pass along with...
12470 (pass_data_local_pure_const): ...new pass_data instance and...
12471 (make_pass_local_pure_const): ...new function.
12472 (pass_ipa_pure_const): Convert from a global struct to a subclass of
12473 ipa_opt_pass_d along with...
12474 (pass_data_ipa_pure_const): ...new pass_data instance and...
12475 (make_pass_ipa_pure_const): ...new function.
12476 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
12477 to a subclass of ipa_opt_pass_d along with...
12478 (pass_data_ipa_reference): ...new pass_data instance and...
12479 (make_pass_ipa_reference): ...new function.
12480 * ipa-split.c (pass_split_functions): Convert from a global struct to
12481 a subclass of gimple_opt_pass along with...
12482 (pass_data_split_functions): ...new pass_data instance and...
12483 (make_pass_split_functions): ...new function.
12484 (pass_feedback_split_functions): Convert from a global struct to a
12485 subclass of gimple_opt_pass along with...
12486 (pass_data_feedback_split_functions): ...new pass_data instance and...
12487 (make_pass_feedback_split_functions): ...new function.
12488 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
12489 global struct to a subclass of simple_ipa_opt_pass along with...
12490 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
12491 instance and...
12492 (make_pass_ipa_function_and_variable_visibility): ...new function.
12493 (pass_ipa_free_inline_summary): Convert from a global struct to a
12494 subclass of simple_ipa_opt_pass along with...
12495 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
12496 (make_pass_ipa_free_inline_summary): ...new function.
12497 (pass_ipa_whole_program_visibility): Convert from a global struct to a
12498 subclass of ipa_opt_pass_d along with...
12499 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
12500 and...
12501 (make_pass_ipa_whole_program_visibility): ...new function.
12502 (pass_ipa_profile): Convert from a global struct to a subclass of
12503 ipa_opt_pass_d along with...
12504 (pass_data_ipa_profile): ...new pass_data instance and...
12505 (make_pass_ipa_profile): ...new function.
12506 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
12507 ipa_opt_pass_d along with...
12508 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
12509 (make_pass_ipa_cdtor_merge): ...new function.
12510 * ira.c (pass_ira): Convert from a global struct to a subclass of
12511 rtl_opt_pass along with...
12512 (pass_data_ira): ...new pass_data instance and...
12513 (make_pass_ira): ...new function.
12514 (pass_reload): Convert from a global struct to a subclass of
12515 rtl_opt_pass along with...
12516 (pass_data_reload): ...new pass_data instance and...
12517 (make_pass_reload): ...new function.
12518 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
12519 subclass of rtl_opt_pass along with...
12520 (pass_data_cleanup_barriers): ...new pass_data instance and...
12521 (make_pass_cleanup_barriers): ...new function.
12522 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
12523 of rtl_opt_pass along with...
12524 (pass_data_loop2): ...new pass_data instance and...
12525 (make_pass_loop2): ...new function.
12526 (pass_rtl_loop_init): Convert from a global struct to a subclass of
12527 rtl_opt_pass along with...
12528 (pass_data_rtl_loop_init): ...new pass_data instance and...
12529 (make_pass_rtl_loop_init): ...new function.
12530 (pass_rtl_loop_done): Convert from a global struct to a subclass of
12531 rtl_opt_pass along with...
12532 (pass_data_rtl_loop_done): ...new pass_data instance and...
12533 (make_pass_rtl_loop_done): ...new function.
12534 (pass_rtl_move_loop_invariants): Convert from a global struct to a
12535 subclass of rtl_opt_pass along with...
12536 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
12537 (make_pass_rtl_move_loop_invariants): ...new function.
12538 (pass_rtl_unswitch): Convert from a global struct to a subclass of
12539 rtl_opt_pass along with...
12540 (pass_data_rtl_unswitch): ...new pass_data instance and...
12541 (make_pass_rtl_unswitch): ...new function.
12542 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
12543 subclass of rtl_opt_pass along with...
12544 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
12545 and...
12546 (make_pass_rtl_unroll_and_peel_loops): ...new function.
12547 (pass_rtl_doloop): Convert from a global struct to a subclass of
12548 rtl_opt_pass along with...
12549 (pass_data_rtl_doloop): ...new pass_data instance and...
12550 (make_pass_rtl_doloop): ...new function.
12551 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
12552 a subclass of rtl_opt_pass along with...
12553 (pass_data_lower_subreg): ...new pass_data instance and...
12554 (make_pass_lower_subreg): ...new function.
12555 (pass_lower_subreg2): Convert from a global struct to a subclass of
12556 rtl_opt_pass along with...
12557 (pass_data_lower_subreg2): ...new pass_data instance and...
12558 (make_pass_lower_subreg2): ...new function.
12559 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
12560 struct to a subclass of ipa_opt_pass_d along with...
12561 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
12562 (make_pass_ipa_lto_gimple_out): ...new function.
12563 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
12564 of ipa_opt_pass_d along with...
12565 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
12566 (make_pass_ipa_lto_finish_out): ...new function.
12567 * mode-switching.c (pass_mode_switching): Convert from a global struct
12568 to a subclass of rtl_opt_pass along with...
12569 (pass_data_mode_switching): ...new pass_data instance and...
12570 (make_pass_mode_switching): ...new function.
12571 * modulo-sched.c (pass_sms): Convert from a global struct to a
12572 subclass of rtl_opt_pass along with...
12573 (pass_data_sms): ...new pass_data instance and...
12574 (make_pass_sms): ...new function.
12575 * omp-low.c (pass_expand_omp): Convert from a global struct to a
12576 subclass of gimple_opt_pass along with...
12577 (pass_data_expand_omp): ...new pass_data instance and...
12578 (make_pass_expand_omp): ...new function.
12579 (pass_lower_omp): Convert from a global struct to a subclass of
12580 gimple_opt_pass along with...
12581 (pass_data_lower_omp): ...new pass_data instance and...
12582 (make_pass_lower_omp): ...new function.
12583 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
12584 of gimple_opt_pass along with...
12585 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
12586 (make_pass_diagnose_omp_blocks): ...new function.
12587 * passes.c (pass_early_local_passes): Convert from a global struct to
12588 a subclass of simple_ipa_opt_pass along with...
12589 (pass_data_early_local_passes): ...new pass_data instance and...
12590 (make_pass_early_local_passes): ...new function.
12591 (pass_all_early_optimizations): Convert from a global struct to a
12592 subclass of gimple_opt_pass along with...
12593 (pass_data_all_early_optimizations): ...new pass_data instance and...
12594 (make_pass_all_early_optimizations): ...new function.
12595 (pass_all_optimizations): Convert from a global struct to a subclass
12596 of gimple_opt_pass along with...
12597 (pass_data_all_optimizations): ...new pass_data instance and...
12598 (make_pass_all_optimizations): ...new function.
12599 (pass_all_optimizations_g): Convert from a global struct to a subclass
12600 of gimple_opt_pass along with...
12601 (pass_data_all_optimizations_g): ...new pass_data instance and...
12602 (make_pass_all_optimizations_g): ...new function.
12603 (pass_rest_of_compilation): Convert from a global struct to a subclass
12604 of rtl_opt_pass along with...
12605 (pass_data_rest_of_compilation): ...new pass_data instance and...
12606 (make_pass_rest_of_compilation): ...new function.
12607 (pass_postreload): Convert from a global struct to a subclass of
12608 rtl_opt_pass along with...
12609 (pass_data_postreload): ...new pass_data instance and...
12610 (make_pass_postreload): ...new function.
12611 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
12612 subclass of rtl_opt_pass along with...
12613 (pass_data_gcse2): ...new pass_data instance and...
12614 (make_pass_gcse2): ...new function.
12615 * postreload.c (pass_postreload_cse): Convert from a global struct to
12616 a subclass of rtl_opt_pass along with...
12617 (pass_data_postreload_cse): ...new pass_data instance and...
12618 (make_pass_postreload_cse): ...new function.
12619 * predict.c (pass_profile): Convert from a global struct to a subclass
12620 of gimple_opt_pass along with...
12621 (pass_data_profile): ...new pass_data instance and...
12622 (make_pass_profile): ...new function.
12623 (pass_strip_predict_hints): Convert from a global struct to a subclass
12624 of gimple_opt_pass along with...
12625 (pass_data_strip_predict_hints): ...new pass_data instance and...
12626 (make_pass_strip_predict_hints): ...new function.
12627 * recog.c (pass_peephole2): Convert from a global struct to a subclass
12628 of rtl_opt_pass along with...
12629 (pass_data_peephole2): ...new pass_data instance and...
12630 (make_pass_peephole2): ...new function.
12631 (pass_split_all_insns): Convert from a global struct to a subclass of
12632 rtl_opt_pass along with...
12633 (pass_data_split_all_insns): ...new pass_data instance and...
12634 (make_pass_split_all_insns): ...new function.
12635 (pass_split_after_reload): Convert from a global struct to a subclass
12636 of rtl_opt_pass along with...
12637 (pass_data_split_after_reload): ...new pass_data instance and...
12638 (make_pass_split_after_reload): ...new function.
12639 (pass_split_before_regstack): Convert from a global struct to a
12640 subclass of rtl_opt_pass along with...
12641 (pass_data_split_before_regstack): ...new pass_data instance and...
12642 (make_pass_split_before_regstack): ...new function.
12643 (pass_split_before_sched2): Convert from a global struct to a subclass
12644 of rtl_opt_pass along with...
12645 (pass_data_split_before_sched2): ...new pass_data instance and...
12646 (make_pass_split_before_sched2): ...new function.
12647 (pass_split_for_shorten_branches): Convert from a global struct to a
12648 subclass of rtl_opt_pass along with...
12649 (pass_data_split_for_shorten_branches): ...new pass_data instance
12650 and...
12651 (make_pass_split_for_shorten_branches): ...new function.
12652 * ree.c (pass_ree): Convert from a global struct to a subclass of
12653 rtl_opt_pass along with...
12654 (pass_data_ree): ...new pass_data instance and...
12655 (make_pass_ree): ...new function.
12656 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
12657 subclass of rtl_opt_pass along with...
12658 (pass_data_stack_regs): ...new pass_data instance and...
12659 (make_pass_stack_regs): ...new function.
12660 (pass_stack_regs_run): Convert from a global struct to a subclass of
12661 rtl_opt_pass along with...
12662 (pass_data_stack_regs_run): ...new pass_data instance and...
12663 (make_pass_stack_regs_run): ...new function.
12664 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
12665 subclass of rtl_opt_pass along with...
12666 (pass_data_cprop_hardreg): ...new pass_data instance and...
12667 (make_pass_cprop_hardreg): ...new function.
12668 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
12669 subclass of rtl_opt_pass along with...
12670 (pass_data_reginfo_init): ...new pass_data instance and...
12671 (make_pass_reginfo_init): ...new function.
12672 * regmove.c (pass_regmove): Convert from a global struct to a subclass
12673 of rtl_opt_pass along with...
12674 (pass_data_regmove): ...new pass_data instance and...
12675 (make_pass_regmove): ...new function.
12676 * regrename.c (pass_regrename): Convert from a global struct to a
12677 subclass of rtl_opt_pass along with...
12678 (pass_data_regrename): ...new pass_data instance and...
12679 (make_pass_regrename): ...new function.
12680 * reorg.c (pass_delay_slots): Convert from a global struct to a
12681 subclass of rtl_opt_pass along with...
12682 (pass_data_delay_slots): ...new pass_data instance and...
12683 (make_pass_delay_slots): ...new function.
12684 (pass_machine_reorg): Convert from a global struct to a subclass of
12685 rtl_opt_pass along with...
12686 (pass_data_machine_reorg): ...new pass_data instance and...
12687 (make_pass_machine_reorg): ...new function.
12688 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
12689 of rtl_opt_pass along with...
12690 (pass_data_sched): ...new pass_data instance and...
12691 (make_pass_sched): ...new function.
12692 (pass_sched2): Convert from a global struct to a subclass of
12693 rtl_opt_pass along with...
12694 (pass_data_sched2): ...new pass_data instance and...
12695 (make_pass_sched2): ...new function.
12696 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
12697 to a subclass of rtl_opt_pass along with...
12698 (pass_data_stack_ptr_mod): ...new pass_data instance and...
12699 (make_pass_stack_ptr_mod): ...new function.
12700 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
12701 to a subclass of rtl_opt_pass along with...
12702 (pass_data_rtl_store_motion): ...new pass_data instance and...
12703 (make_pass_rtl_store_motion): ...new function.
12704 * tracer.c (pass_tracer): Convert from a global struct to a subclass
12705 of gimple_opt_pass along with...
12706 (pass_data_tracer): ...new pass_data instance and...
12707 (make_pass_tracer): ...new function.
12708 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
12709 to a subclass of gimple_opt_pass along with...
12710 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
12711 (make_pass_diagnose_tm_blocks): ...new function.
12712 (pass_lower_tm): Convert from a global struct to a subclass of
12713 gimple_opt_pass along with...
12714 (pass_data_lower_tm): ...new pass_data instance and...
12715 (make_pass_lower_tm): ...new function.
12716 (pass_tm_init): Convert from a global struct to a subclass of
12717 gimple_opt_pass along with...
12718 (pass_data_tm_init): ...new pass_data instance and...
12719 (make_pass_tm_init): ...new function.
12720 (pass_tm_mark): Convert from a global struct to a subclass of
12721 gimple_opt_pass along with...
12722 (pass_data_tm_mark): ...new pass_data instance and...
12723 (make_pass_tm_mark): ...new function.
12724 (pass_tm_edges): Convert from a global struct to a subclass of
12725 gimple_opt_pass along with...
12726 (pass_data_tm_edges): ...new pass_data instance and...
12727 (make_pass_tm_edges): ...new function.
12728 (pass_tm_memopt): Convert from a global struct to a subclass of
12729 gimple_opt_pass along with...
12730 (pass_data_tm_memopt): ...new pass_data instance and...
12731 (make_pass_tm_memopt): ...new function.
12732 (pass_ipa_tm): Convert from a global struct to a subclass of
12733 simple_ipa_opt_pass along with...
12734 (pass_data_ipa_tm): ...new pass_data instance and...
12735 (make_pass_ipa_tm): ...new function.
12736 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
12737 subclass of gimple_opt_pass along with...
12738 (pass_data_call_cdce): ...new pass_data instance and...
12739 (make_pass_call_cdce): ...new function.
12740 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
12741 subclass of gimple_opt_pass along with...
12742 (pass_data_build_cfg): ...new pass_data instance and...
12743 (make_pass_build_cfg): ...new function.
12744 (pass_split_crit_edges): Convert from a global struct to a subclass of
12745 gimple_opt_pass along with...
12746 (pass_data_split_crit_edges): ...new pass_data instance and...
12747 (make_pass_split_crit_edges): ...new function.
12748 (pass_warn_function_return): Convert from a global struct to a
12749 subclass of gimple_opt_pass along with...
12750 (pass_data_warn_function_return): ...new pass_data instance and...
12751 (make_pass_warn_function_return): ...new function.
12752 (pass_warn_function_noreturn): Convert from a global struct to a
12753 subclass of gimple_opt_pass along with...
12754 (pass_data_warn_function_noreturn): ...new pass_data instance and...
12755 (make_pass_warn_function_noreturn): ...new function.
12756 (pass_warn_unused_result): Convert from a global struct to a subclass
12757 of gimple_opt_pass along with...
12758 (pass_data_warn_unused_result): ...new pass_data instance and...
12759 (make_pass_warn_unused_result): ...new function.
12760 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
12761 a subclass of gimple_opt_pass along with...
12762 (pass_data_merge_phi): ...new pass_data instance and...
12763 (make_pass_merge_phi): ...new function.
12764 * tree-complex.c (pass_lower_complex): Convert from a global struct to
12765 a subclass of gimple_opt_pass along with...
12766 (pass_data_lower_complex): ...new pass_data instance and...
12767 (make_pass_lower_complex): ...new function.
12768 (pass_lower_complex_O0): Convert from a global struct to a subclass of
12769 gimple_opt_pass along with...
12770 (pass_data_lower_complex_O0): ...new pass_data instance and...
12771 (make_pass_lower_complex_O0): ...new function.
12772 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
12773 subclass of gimple_opt_pass along with...
12774 (pass_data_lower_eh): ...new pass_data instance and...
12775 (make_pass_lower_eh): ...new function.
12776 (pass_refactor_eh): Convert from a global struct to a subclass of
12777 gimple_opt_pass along with...
12778 (pass_data_refactor_eh): ...new pass_data instance and...
12779 (make_pass_refactor_eh): ...new function.
12780 (pass_lower_resx): Convert from a global struct to a subclass of
12781 gimple_opt_pass along with...
12782 (pass_data_lower_resx): ...new pass_data instance and...
12783 (make_pass_lower_resx): ...new function.
12784 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
12785 of gimple_opt_pass along with...
12786 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
12787 (make_pass_lower_eh_dispatch): ...new function.
12788 (pass_cleanup_eh): Convert from a global struct to a subclass of
12789 gimple_opt_pass along with...
12790 (pass_data_cleanup_eh): ...new pass_data instance and...
12791 (make_pass_cleanup_eh): ...new function.
12792 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
12793 to a subclass of simple_ipa_opt_pass along with...
12794 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
12795 (make_pass_ipa_lower_emutls): ...new function.
12796 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
12797 a subclass of gimple_opt_pass along with...
12798 (pass_data_if_conversion): ...new pass_data instance and...
12799 (make_pass_if_conversion): ...new function.
12800 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
12801 subclass of gimple_opt_pass along with...
12802 (pass_data_build_ssa): ...new pass_data instance and...
12803 (make_pass_build_ssa): ...new function.
12804 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
12805 global struct to a subclass of gimple_opt_pass along with...
12806 (pass_data_loop_distribution): ...new pass_data instance and...
12807 (make_pass_loop_distribution): ...new function.
12808 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
12809 subclass of gimple_opt_pass along with...
12810 (pass_data_mudflap_1): ...new pass_data instance and...
12811 (make_pass_mudflap_1): ...new function.
12812 (pass_mudflap_2): Convert from a global struct to a subclass of
12813 gimple_opt_pass along with...
12814 (pass_data_mudflap_2): ...new pass_data instance and...
12815 (make_pass_mudflap_2): ...new function.
12816 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
12817 subclass of gimple_opt_pass along with...
12818 (pass_data_mudflap_1): ...new pass_data instance and...
12819 (make_pass_mudflap_1): ...new function.
12820 (pass_mudflap_2): Convert from a global struct to a subclass of
12821 gimple_opt_pass along with...
12822 (pass_data_mudflap_2): ...new pass_data instance and...
12823 (make_pass_mudflap_2): ...new function.
12824 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
12825 gimple_opt_pass along with...
12826 (pass_data_nrv): ...new pass_data instance and...
12827 (make_pass_nrv): ...new function.
12828 (pass_return_slot): Convert from a global struct to a subclass of
12829 gimple_opt_pass along with...
12830 (pass_data_return_slot): ...new pass_data instance and...
12831 (make_pass_return_slot): ...new function.
12832 * tree-object-size.c (pass_object_sizes): Convert from a global struct
12833 to a subclass of gimple_opt_pass along with...
12834 (pass_data_object_sizes): ...new pass_data instance and...
12835 (make_pass_object_sizes): ...new function.
12836 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
12837 global struct to a subclass of gimple_opt_pass along with...
12838 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
12839 and...
12840 (make_pass_cleanup_cfg_post_optimizing): ...new function.
12841 (pass_fixup_cfg): Convert from a global struct to a subclass of
12842 gimple_opt_pass along with...
12843 (pass_data_fixup_cfg): ...new pass_data instance and...
12844 (make_pass_fixup_cfg): ...new function.
12845 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
12846 (make_pass_mudflap_1): ...new function.
12847 (pass_mudflap_2): Replace declaration with that of...
12848 (make_pass_mudflap_2): ...new function.
12849 (pass_asan): Replace declaration with that of...
12850 (make_pass_asan): ...new function.
12851 (pass_asan_O0): Replace declaration with that of...
12852 (make_pass_asan_O0): ...new function.
12853 (pass_tsan): Replace declaration with that of...
12854 (make_pass_tsan): ...new function.
12855 (pass_tsan_O0): Replace declaration with that of...
12856 (make_pass_tsan_O0): ...new function.
12857 (pass_lower_cf): Replace declaration with that of...
12858 (make_pass_lower_cf): ...new function.
12859 (pass_refactor_eh): Replace declaration with that of...
12860 (make_pass_refactor_eh): ...new function.
12861 (pass_lower_eh): Replace declaration with that of...
12862 (make_pass_lower_eh): ...new function.
12863 (pass_lower_eh_dispatch): Replace declaration with that of...
12864 (make_pass_lower_eh_dispatch): ...new function.
12865 (pass_lower_resx): Replace declaration with that of...
12866 (make_pass_lower_resx): ...new function.
12867 (pass_build_cfg): Replace declaration with that of...
12868 (make_pass_build_cfg): ...new function.
12869 (pass_early_tree_profile): Replace declaration with that of...
12870 (make_pass_early_tree_profile): ...new function.
12871 (pass_cleanup_eh): Replace declaration with that of...
12872 (make_pass_cleanup_eh): ...new function.
12873 (pass_sra): Replace declaration with that of...
12874 (make_pass_sra): ...new function.
12875 (pass_sra_early): Replace declaration with that of...
12876 (make_pass_sra_early): ...new function.
12877 (pass_early_ipa_sra): Replace declaration with that of...
12878 (make_pass_early_ipa_sra): ...new function.
12879 (pass_tail_recursion): Replace declaration with that of...
12880 (make_pass_tail_recursion): ...new function.
12881 (pass_tail_calls): Replace declaration with that of...
12882 (make_pass_tail_calls): ...new function.
12883 (pass_tree_loop): Replace declaration with that of...
12884 (make_pass_tree_loop): ...new function.
12885 (pass_tree_loop_init): Replace declaration with that of...
12886 (make_pass_tree_loop_init): ...new function.
12887 (pass_lim): Replace declaration with that of...
12888 (make_pass_lim): ...new function.
12889 (pass_tree_unswitch): Replace declaration with that of...
12890 (make_pass_tree_unswitch): ...new function.
12891 (pass_predcom): Replace declaration with that of...
12892 (make_pass_predcom): ...new function.
12893 (pass_iv_canon): Replace declaration with that of...
12894 (make_pass_iv_canon): ...new function.
12895 (pass_scev_cprop): Replace declaration with that of...
12896 (make_pass_scev_cprop): ...new function.
12897 (pass_empty_loop): Replace declaration with that of...
12898 (make_pass_empty_loop): ...new function.
12899 (pass_record_bounds): Replace declaration with that of...
12900 (make_pass_record_bounds): ...new function.
12901 (pass_graphite): Replace declaration with that of...
12902 (make_pass_graphite): ...new function.
12903 (pass_graphite_transforms): Replace declaration with that of...
12904 (make_pass_graphite_transforms): ...new function.
12905 (pass_if_conversion): Replace declaration with that of...
12906 (make_pass_if_conversion): ...new function.
12907 (pass_loop_distribution): Replace declaration with that of...
12908 (make_pass_loop_distribution): ...new function.
12909 (pass_vectorize): Replace declaration with that of...
12910 (make_pass_vectorize): ...new function.
12911 (pass_slp_vectorize): Replace declaration with that of...
12912 (make_pass_slp_vectorize): ...new function.
12913 (pass_complete_unroll): Replace declaration with that of...
12914 (make_pass_complete_unroll): ...new function.
12915 (pass_complete_unrolli): Replace declaration with that of...
12916 (make_pass_complete_unrolli): ...new function.
12917 (pass_parallelize_loops): Replace declaration with that of...
12918 (make_pass_parallelize_loops): ...new function.
12919 (pass_loop_prefetch): Replace declaration with that of...
12920 (make_pass_loop_prefetch): ...new function.
12921 (pass_iv_optimize): Replace declaration with that of...
12922 (make_pass_iv_optimize): ...new function.
12923 (pass_tree_loop_done): Replace declaration with that of...
12924 (make_pass_tree_loop_done): ...new function.
12925 (pass_ch): Replace declaration with that of...
12926 (make_pass_ch): ...new function.
12927 (pass_ccp): Replace declaration with that of...
12928 (make_pass_ccp): ...new function.
12929 (pass_phi_only_cprop): Replace declaration with that of...
12930 (make_pass_phi_only_cprop): ...new function.
12931 (pass_build_ssa): Replace declaration with that of...
12932 (make_pass_build_ssa): ...new function.
12933 (pass_build_alias): Replace declaration with that of...
12934 (make_pass_build_alias): ...new function.
12935 (pass_build_ealias): Replace declaration with that of...
12936 (make_pass_build_ealias): ...new function.
12937 (pass_dominator): Replace declaration with that of...
12938 (make_pass_dominator): ...new function.
12939 (pass_dce): Replace declaration with that of...
12940 (make_pass_dce): ...new function.
12941 (pass_dce_loop): Replace declaration with that of...
12942 (make_pass_dce_loop): ...new function.
12943 (pass_cd_dce): Replace declaration with that of...
12944 (make_pass_cd_dce): ...new function.
12945 (pass_call_cdce): Replace declaration with that of...
12946 (make_pass_call_cdce): ...new function.
12947 (pass_merge_phi): Replace declaration with that of...
12948 (make_pass_merge_phi): ...new function.
12949 (pass_split_crit_edges): Replace declaration with that of...
12950 (make_pass_split_crit_edges): ...new function.
12951 (pass_pre): Replace declaration with that of...
12952 (make_pass_pre): ...new function.
12953 (pass_profile): Replace declaration with that of...
12954 (make_pass_profile): ...new function.
12955 (pass_strip_predict_hints): Replace declaration with that of...
12956 (make_pass_strip_predict_hints): ...new function.
12957 (pass_lower_complex_O0): Replace declaration with that of...
12958 (make_pass_lower_complex_O0): ...new function.
12959 (pass_lower_complex): Replace declaration with that of...
12960 (make_pass_lower_complex): ...new function.
12961 (pass_lower_vector): Replace declaration with that of...
12962 (make_pass_lower_vector): ...new function.
12963 (pass_lower_vector_ssa): Replace declaration with that of...
12964 (make_pass_lower_vector_ssa): ...new function.
12965 (pass_lower_omp): Replace declaration with that of...
12966 (make_pass_lower_omp): ...new function.
12967 (pass_diagnose_omp_blocks): Replace declaration with that of...
12968 (make_pass_diagnose_omp_blocks): ...new function.
12969 (pass_expand_omp): Replace declaration with that of...
12970 (make_pass_expand_omp): ...new function.
12971 (pass_expand_omp_ssa): Replace declaration with that of...
12972 (make_pass_expand_omp_ssa): ...new function.
12973 (pass_object_sizes): Replace declaration with that of...
12974 (make_pass_object_sizes): ...new function.
12975 (pass_strlen): Replace declaration with that of...
12976 (make_pass_strlen): ...new function.
12977 (pass_fold_builtins): Replace declaration with that of...
12978 (make_pass_fold_builtins): ...new function.
12979 (pass_stdarg): Replace declaration with that of...
12980 (make_pass_stdarg): ...new function.
12981 (pass_early_warn_uninitialized): Replace declaration with that of...
12982 (make_pass_early_warn_uninitialized): ...new function.
12983 (pass_late_warn_uninitialized): Replace declaration with that of...
12984 (make_pass_late_warn_uninitialized): ...new function.
12985 (pass_cse_reciprocals): Replace declaration with that of...
12986 (make_pass_cse_reciprocals): ...new function.
12987 (pass_cse_sincos): Replace declaration with that of...
12988 (make_pass_cse_sincos): ...new function.
12989 (pass_optimize_bswap): Replace declaration with that of...
12990 (make_pass_optimize_bswap): ...new function.
12991 (pass_optimize_widening_mul): Replace declaration with that of...
12992 (make_pass_optimize_widening_mul): ...new function.
12993 (pass_warn_function_return): Replace declaration with that of...
12994 (make_pass_warn_function_return): ...new function.
12995 (pass_warn_function_noreturn): Replace declaration with that of...
12996 (make_pass_warn_function_noreturn): ...new function.
12997 (pass_cselim): Replace declaration with that of...
12998 (make_pass_cselim): ...new function.
12999 (pass_phiopt): Replace declaration with that of...
13000 (make_pass_phiopt): ...new function.
13001 (pass_forwprop): Replace declaration with that of...
13002 (make_pass_forwprop): ...new function.
13003 (pass_phiprop): Replace declaration with that of...
13004 (make_pass_phiprop): ...new function.
13005 (pass_tree_ifcombine): Replace declaration with that of...
13006 (make_pass_tree_ifcombine): ...new function.
13007 (pass_dse): Replace declaration with that of...
13008 (make_pass_dse): ...new function.
13009 (pass_nrv): Replace declaration with that of...
13010 (make_pass_nrv): ...new function.
13011 (pass_rename_ssa_copies): Replace declaration with that of...
13012 (make_pass_rename_ssa_copies): ...new function.
13013 (pass_sink_code): Replace declaration with that of...
13014 (make_pass_sink_code): ...new function.
13015 (pass_fre): Replace declaration with that of...
13016 (make_pass_fre): ...new function.
13017 (pass_check_data_deps): Replace declaration with that of...
13018 (make_pass_check_data_deps): ...new function.
13019 (pass_copy_prop): Replace declaration with that of...
13020 (make_pass_copy_prop): ...new function.
13021 (pass_vrp): Replace declaration with that of...
13022 (make_pass_vrp): ...new function.
13023 (pass_uncprop): Replace declaration with that of...
13024 (make_pass_uncprop): ...new function.
13025 (pass_return_slot): Replace declaration with that of...
13026 (make_pass_return_slot): ...new function.
13027 (pass_reassoc): Replace declaration with that of...
13028 (make_pass_reassoc): ...new function.
13029 (pass_rebuild_cgraph_edges): Replace declaration with that of...
13030 (make_pass_rebuild_cgraph_edges): ...new function.
13031 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
13032 (make_pass_remove_cgraph_callee_edges): ...new function.
13033 (pass_build_cgraph_edges): Replace declaration with that of...
13034 (make_pass_build_cgraph_edges): ...new function.
13035 (pass_local_pure_const): Replace declaration with that of...
13036 (make_pass_local_pure_const): ...new function.
13037 (pass_tracer): Replace declaration with that of...
13038 (make_pass_tracer): ...new function.
13039 (pass_warn_unused_result): Replace declaration with that of...
13040 (make_pass_warn_unused_result): ...new function.
13041 (pass_diagnose_tm_blocks): Replace declaration with that of...
13042 (make_pass_diagnose_tm_blocks): ...new function.
13043 (pass_lower_tm): Replace declaration with that of...
13044 (make_pass_lower_tm): ...new function.
13045 (pass_tm_init): Replace declaration with that of...
13046 (make_pass_tm_init): ...new function.
13047 (pass_tm_mark): Replace declaration with that of...
13048 (make_pass_tm_mark): ...new function.
13049 (pass_tm_memopt): Replace declaration with that of...
13050 (make_pass_tm_memopt): ...new function.
13051 (pass_tm_edges): Replace declaration with that of...
13052 (make_pass_tm_edges): ...new function.
13053 (pass_split_functions): Replace declaration with that of...
13054 (make_pass_split_functions): ...new function.
13055 (pass_feedback_split_functions): Replace declaration with that of...
13056 (make_pass_feedback_split_functions): ...new function.
13057 (pass_strength_reduction): Replace declaration with that of...
13058 (make_pass_strength_reduction): ...new function.
13059 (pass_ipa_lower_emutls): Replace declaration with that of...
13060 (make_pass_ipa_lower_emutls): ...new function.
13061 (pass_ipa_function_and_variable_visibility): Replace declaration with
13062 that of...
13063 (make_pass_ipa_function_and_variable_visibility): ...new function.
13064 (pass_ipa_tree_profile): Replace declaration with that of...
13065 (make_pass_ipa_tree_profile): ...new function.
13066 (pass_early_local_passes): Replace declaration with that of...
13067 (make_pass_early_local_passes): ...new function.
13068 (pass_ipa_whole_program_visibility): Replace declaration with that
13069 of...
13070 (make_pass_ipa_whole_program_visibility): ...new function.
13071 (pass_ipa_lto_gimple_out): Replace declaration with that of...
13072 (make_pass_ipa_lto_gimple_out): ...new function.
13073 (pass_ipa_increase_alignment): Replace declaration with that of...
13074 (make_pass_ipa_increase_alignment): ...new function.
13075 (pass_ipa_inline): Replace declaration with that of...
13076 (make_pass_ipa_inline): ...new function.
13077 (pass_ipa_free_lang_data): Replace declaration with that of...
13078 (make_pass_ipa_free_lang_data): ...new function.
13079 (pass_ipa_free_inline_summary): Replace declaration with that of...
13080 (make_pass_ipa_free_inline_summary): ...new function.
13081 (pass_ipa_cp): Replace declaration with that of...
13082 (make_pass_ipa_cp): ...new function.
13083 (pass_ipa_reference): Replace declaration with that of...
13084 (make_pass_ipa_reference): ...new function.
13085 (pass_ipa_pure_const): Replace declaration with that of...
13086 (make_pass_ipa_pure_const): ...new function.
13087 (pass_ipa_pta): Replace declaration with that of...
13088 (make_pass_ipa_pta): ...new function.
13089 (pass_ipa_lto_finish_out): Replace declaration with that of...
13090 (make_pass_ipa_lto_finish_out): ...new function.
13091 (pass_ipa_tm): Replace declaration with that of...
13092 (make_pass_ipa_tm): ...new function.
13093 (pass_ipa_profile): Replace declaration with that of...
13094 (make_pass_ipa_profile): ...new function.
13095 (pass_ipa_cdtor_merge): Replace declaration with that of...
13096 (make_pass_ipa_cdtor_merge): ...new function.
13097 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
13098 of...
13099 (make_pass_cleanup_cfg_post_optimizing): ...new function.
13100 (pass_init_datastructures): Replace declaration with that of...
13101 (make_pass_init_datastructures): ...new function.
13102 (pass_fixup_cfg): Replace declaration with that of...
13103 (make_pass_fixup_cfg): ...new function.
13104 (pass_expand): Replace declaration with that of...
13105 (make_pass_expand): ...new function.
13106 (pass_instantiate_virtual_regs): Replace declaration with that of...
13107 (make_pass_instantiate_virtual_regs): ...new function.
13108 (pass_rtl_fwprop): Replace declaration with that of...
13109 (make_pass_rtl_fwprop): ...new function.
13110 (pass_rtl_fwprop_addr): Replace declaration with that of...
13111 (make_pass_rtl_fwprop_addr): ...new function.
13112 (pass_jump): Replace declaration with that of...
13113 (make_pass_jump): ...new function.
13114 (pass_jump2): Replace declaration with that of...
13115 (make_pass_jump2): ...new function.
13116 (pass_lower_subreg): Replace declaration with that of...
13117 (make_pass_lower_subreg): ...new function.
13118 (pass_cse): Replace declaration with that of...
13119 (make_pass_cse): ...new function.
13120 (pass_fast_rtl_dce): Replace declaration with that of...
13121 (make_pass_fast_rtl_dce): ...new function.
13122 (pass_ud_rtl_dce): Replace declaration with that of...
13123 (make_pass_ud_rtl_dce): ...new function.
13124 (pass_rtl_dce): Replace declaration with that of...
13125 (make_pass_rtl_dce): ...new function.
13126 (pass_rtl_dse1): Replace declaration with that of...
13127 (make_pass_rtl_dse1): ...new function.
13128 (pass_rtl_dse2): Replace declaration with that of...
13129 (make_pass_rtl_dse2): ...new function.
13130 (pass_rtl_dse3): Replace declaration with that of...
13131 (make_pass_rtl_dse3): ...new function.
13132 (pass_rtl_cprop): Replace declaration with that of...
13133 (make_pass_rtl_cprop): ...new function.
13134 (pass_rtl_pre): Replace declaration with that of...
13135 (make_pass_rtl_pre): ...new function.
13136 (pass_rtl_hoist): Replace declaration with that of...
13137 (make_pass_rtl_hoist): ...new function.
13138 (pass_rtl_store_motion): Replace declaration with that of...
13139 (make_pass_rtl_store_motion): ...new function.
13140 (pass_cse_after_global_opts): Replace declaration with that of...
13141 (make_pass_cse_after_global_opts): ...new function.
13142 (pass_rtl_ifcvt): Replace declaration with that of...
13143 (make_pass_rtl_ifcvt): ...new function.
13144 (pass_into_cfg_layout_mode): Replace declaration with that of...
13145 (make_pass_into_cfg_layout_mode): ...new function.
13146 (pass_outof_cfg_layout_mode): Replace declaration with that of...
13147 (make_pass_outof_cfg_layout_mode): ...new function.
13148 (pass_loop2): Replace declaration with that of...
13149 (make_pass_loop2): ...new function.
13150 (pass_rtl_loop_init): Replace declaration with that of...
13151 (make_pass_rtl_loop_init): ...new function.
13152 (pass_rtl_move_loop_invariants): Replace declaration with that of...
13153 (make_pass_rtl_move_loop_invariants): ...new function.
13154 (pass_rtl_unswitch): Replace declaration with that of...
13155 (make_pass_rtl_unswitch): ...new function.
13156 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
13157 (make_pass_rtl_unroll_and_peel_loops): ...new function.
13158 (pass_rtl_doloop): Replace declaration with that of...
13159 (make_pass_rtl_doloop): ...new function.
13160 (pass_rtl_loop_done): Replace declaration with that of...
13161 (make_pass_rtl_loop_done): ...new function.
13162 (pass_web): Replace declaration with that of...
13163 (make_pass_web): ...new function.
13164 (pass_cse2): Replace declaration with that of...
13165 (make_pass_cse2): ...new function.
13166 (pass_df_initialize_opt): Replace declaration with that of...
13167 (make_pass_df_initialize_opt): ...new function.
13168 (pass_df_initialize_no_opt): Replace declaration with that of...
13169 (make_pass_df_initialize_no_opt): ...new function.
13170 (pass_reginfo_init): Replace declaration with that of...
13171 (make_pass_reginfo_init): ...new function.
13172 (pass_inc_dec): Replace declaration with that of...
13173 (make_pass_inc_dec): ...new function.
13174 (pass_stack_ptr_mod): Replace declaration with that of...
13175 (make_pass_stack_ptr_mod): ...new function.
13176 (pass_initialize_regs): Replace declaration with that of...
13177 (make_pass_initialize_regs): ...new function.
13178 (pass_combine): Replace declaration with that of...
13179 (make_pass_combine): ...new function.
13180 (pass_if_after_combine): Replace declaration with that of...
13181 (make_pass_if_after_combine): ...new function.
13182 (pass_ree): Replace declaration with that of...
13183 (make_pass_ree): ...new function.
13184 (pass_partition_blocks): Replace declaration with that of...
13185 (make_pass_partition_blocks): ...new function.
13186 (pass_match_asm_constraints): Replace declaration with that of...
13187 (make_pass_match_asm_constraints): ...new function.
13188 (pass_regmove): Replace declaration with that of...
13189 (make_pass_regmove): ...new function.
13190 (pass_split_all_insns): Replace declaration with that of...
13191 (make_pass_split_all_insns): ...new function.
13192 (pass_fast_rtl_byte_dce): Replace declaration with that of...
13193 (make_pass_fast_rtl_byte_dce): ...new function.
13194 (pass_lower_subreg2): Replace declaration with that of...
13195 (make_pass_lower_subreg2): ...new function.
13196 (pass_mode_switching): Replace declaration with that of...
13197 (make_pass_mode_switching): ...new function.
13198 (pass_sms): Replace declaration with that of...
13199 (make_pass_sms): ...new function.
13200 (pass_sched): Replace declaration with that of...
13201 (make_pass_sched): ...new function.
13202 (pass_ira): Replace declaration with that of...
13203 (make_pass_ira): ...new function.
13204 (pass_reload): Replace declaration with that of...
13205 (make_pass_reload): ...new function.
13206 (pass_clean_state): Replace declaration with that of...
13207 (make_pass_clean_state): ...new function.
13208 (pass_branch_prob): Replace declaration with that of...
13209 (make_pass_branch_prob): ...new function.
13210 (pass_value_profile_transformations): Replace declaration with that
13211 of...
13212 (make_pass_value_profile_transformations): ...new function.
13213 (pass_postreload_cse): Replace declaration with that of...
13214 (make_pass_postreload_cse): ...new function.
13215 (pass_gcse2): Replace declaration with that of...
13216 (make_pass_gcse2): ...new function.
13217 (pass_split_after_reload): Replace declaration with that of...
13218 (make_pass_split_after_reload): ...new function.
13219 (pass_branch_target_load_optimize1): Replace declaration with that
13220 of...
13221 (make_pass_branch_target_load_optimize1): ...new function.
13222 (pass_thread_prologue_and_epilogue): Replace declaration with that
13223 of...
13224 (make_pass_thread_prologue_and_epilogue): ...new function.
13225 (pass_stack_adjustments): Replace declaration with that of...
13226 (make_pass_stack_adjustments): ...new function.
13227 (pass_peephole2): Replace declaration with that of...
13228 (make_pass_peephole2): ...new function.
13229 (pass_if_after_reload): Replace declaration with that of...
13230 (make_pass_if_after_reload): ...new function.
13231 (pass_regrename): Replace declaration with that of...
13232 (make_pass_regrename): ...new function.
13233 (pass_cprop_hardreg): Replace declaration with that of...
13234 (make_pass_cprop_hardreg): ...new function.
13235 (pass_reorder_blocks): Replace declaration with that of...
13236 (make_pass_reorder_blocks): ...new function.
13237 (pass_branch_target_load_optimize2): Replace declaration with that
13238 of...
13239 (make_pass_branch_target_load_optimize2): ...new function.
13240 (pass_leaf_regs): Replace declaration with that of...
13241 (make_pass_leaf_regs): ...new function.
13242 (pass_split_before_sched2): Replace declaration with that of...
13243 (make_pass_split_before_sched2): ...new function.
13244 (pass_compare_elim_after_reload): Replace declaration with that of...
13245 (make_pass_compare_elim_after_reload): ...new function.
13246 (pass_sched2): Replace declaration with that of...
13247 (make_pass_sched2): ...new function.
13248 (pass_stack_regs): Replace declaration with that of...
13249 (make_pass_stack_regs): ...new function.
13250 (pass_stack_regs_run): Replace declaration with that of...
13251 (make_pass_stack_regs_run): ...new function.
13252 (pass_df_finish): Replace declaration with that of...
13253 (make_pass_df_finish): ...new function.
13254 (pass_compute_alignments): Replace declaration with that of...
13255 (make_pass_compute_alignments): ...new function.
13256 (pass_duplicate_computed_gotos): Replace declaration with that of...
13257 (make_pass_duplicate_computed_gotos): ...new function.
13258 (pass_variable_tracking): Replace declaration with that of...
13259 (make_pass_variable_tracking): ...new function.
13260 (pass_free_cfg): Replace declaration with that of...
13261 (make_pass_free_cfg): ...new function.
13262 (pass_machine_reorg): Replace declaration with that of...
13263 (make_pass_machine_reorg): ...new function.
13264 (pass_cleanup_barriers): Replace declaration with that of...
13265 (make_pass_cleanup_barriers): ...new function.
13266 (pass_delay_slots): Replace declaration with that of...
13267 (make_pass_delay_slots): ...new function.
13268 (pass_split_for_shorten_branches): Replace declaration with that of...
13269 (make_pass_split_for_shorten_branches): ...new function.
13270 (pass_split_before_regstack): Replace declaration with that of...
13271 (make_pass_split_before_regstack): ...new function.
13272 (pass_convert_to_eh_region_ranges): Replace declaration with that
13273 of...
13274 (make_pass_convert_to_eh_region_ranges): ...new function.
13275 (pass_shorten_branches): Replace declaration with that of...
13276 (make_pass_shorten_branches): ...new function.
13277 (pass_set_nothrow_function_flags): Replace declaration with that of...
13278 (make_pass_set_nothrow_function_flags): ...new function.
13279 (pass_dwarf2_frame): Replace declaration with that of...
13280 (make_pass_dwarf2_frame): ...new function.
13281 (pass_final): Replace declaration with that of...
13282 (make_pass_final): ...new function.
13283 (pass_rtl_seqabstr): Replace declaration with that of...
13284 (make_pass_rtl_seqabstr): ...new function.
13285 (pass_release_ssa_names): Replace declaration with that of...
13286 (make_pass_release_ssa_names): ...new function.
13287 (pass_early_inline): Replace declaration with that of...
13288 (make_pass_early_inline): ...new function.
13289 (pass_inline_parameters): Replace declaration with that of...
13290 (make_pass_inline_parameters): ...new function.
13291 (pass_update_address_taken): Replace declaration with that of...
13292 (make_pass_update_address_taken): ...new function.
13293 (pass_convert_switch): Replace declaration with that of...
13294 (make_pass_convert_switch): ...new function.
13295 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
13296 to a subclass of simple_ipa_opt_pass along with...
13297 (pass_data_ipa_tree_profile): ...new pass_data instance and...
13298 (make_pass_ipa_tree_profile): ...new function.
13299 * tree-sra.c (pass_sra_early): Convert from a global struct to a
13300 subclass of gimple_opt_pass along with...
13301 (pass_data_sra_early): ...new pass_data instance and...
13302 (make_pass_sra_early): ...new function.
13303 (pass_sra): Convert from a global struct to a subclass of
13304 gimple_opt_pass along with...
13305 (pass_data_sra): ...new pass_data instance and...
13306 (make_pass_sra): ...new function.
13307 (pass_early_ipa_sra): Convert from a global struct to a subclass of
13308 gimple_opt_pass along with...
13309 (pass_data_early_ipa_sra): ...new pass_data instance and...
13310 (make_pass_early_ipa_sra): ...new function.
13311 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
13312 subclass of gimple_opt_pass along with...
13313 (pass_data_ccp): ...new pass_data instance and...
13314 (make_pass_ccp): ...new function.
13315 (pass_fold_builtins): Convert from a global struct to a subclass of
13316 gimple_opt_pass along with...
13317 (pass_data_fold_builtins): ...new pass_data instance and...
13318 (make_pass_fold_builtins): ...new function.
13319 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
13320 subclass of gimple_opt_pass along with...
13321 (pass_data_copy_prop): ...new pass_data instance and...
13322 (make_pass_copy_prop): ...new function.
13323 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
13324 global struct to a subclass of gimple_opt_pass along with...
13325 (pass_data_rename_ssa_copies): ...new pass_data instance and...
13326 (make_pass_rename_ssa_copies): ...new function.
13327 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
13328 subclass of gimple_opt_pass along with...
13329 (pass_data_dce): ...new pass_data instance and...
13330 (make_pass_dce): ...new function.
13331 (pass_dce_loop): Convert from a global struct to a subclass of
13332 gimple_opt_pass along with...
13333 (pass_data_dce_loop): ...new pass_data instance and...
13334 (make_pass_dce_loop): ...new function.
13335 (pass_cd_dce): Convert from a global struct to a subclass of
13336 gimple_opt_pass along with...
13337 (pass_data_cd_dce): ...new pass_data instance and...
13338 (make_pass_cd_dce): ...new function.
13339 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
13340 subclass of gimple_opt_pass along with...
13341 (pass_data_dominator): ...new pass_data instance and...
13342 (make_pass_dominator): ...new function.
13343 (pass_phi_only_cprop): Convert from a global struct to a subclass of
13344 gimple_opt_pass along with...
13345 (pass_data_phi_only_cprop): ...new pass_data instance and...
13346 (make_pass_phi_only_cprop): ...new function.
13347 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
13348 subclass of gimple_opt_pass along with...
13349 (pass_data_dse): ...new pass_data instance and...
13350 (make_pass_dse): ...new function.
13351 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
13352 a subclass of gimple_opt_pass along with...
13353 (pass_data_forwprop): ...new pass_data instance and...
13354 (make_pass_forwprop): ...new function.
13355 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
13356 struct to a subclass of gimple_opt_pass along with...
13357 (pass_data_tree_ifcombine): ...new pass_data instance and...
13358 (make_pass_tree_ifcombine): ...new function.
13359 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
13360 subclass of gimple_opt_pass along with...
13361 (pass_data_ch): ...new pass_data instance and...
13362 (make_pass_ch): ...new function.
13363 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
13364 subclass of gimple_opt_pass along with...
13365 (pass_data_tree_loop): ...new pass_data instance and...
13366 (make_pass_tree_loop): ...new function.
13367 (pass_tree_loop_init): Convert from a global struct to a subclass of
13368 gimple_opt_pass along with...
13369 (pass_data_tree_loop_init): ...new pass_data instance and...
13370 (make_pass_tree_loop_init): ...new function.
13371 (pass_lim): Convert from a global struct to a subclass of
13372 gimple_opt_pass along with...
13373 (pass_data_lim): ...new pass_data instance and...
13374 (make_pass_lim): ...new function.
13375 (pass_tree_unswitch): Convert from a global struct to a subclass of
13376 gimple_opt_pass along with...
13377 (pass_data_tree_unswitch): ...new pass_data instance and...
13378 (make_pass_tree_unswitch): ...new function.
13379 (pass_predcom): Convert from a global struct to a subclass of
13380 gimple_opt_pass along with...
13381 (pass_data_predcom): ...new pass_data instance and...
13382 (make_pass_predcom): ...new function.
13383 (pass_vectorize): Convert from a global struct to a subclass of
13384 gimple_opt_pass along with...
13385 (pass_data_vectorize): ...new pass_data instance and...
13386 (make_pass_vectorize): ...new function.
13387 (pass_graphite): Convert from a global struct to a subclass of
13388 gimple_opt_pass along with...
13389 (pass_data_graphite): ...new pass_data instance and...
13390 (make_pass_graphite): ...new function.
13391 (pass_graphite_transforms): Convert from a global struct to a subclass
13392 of gimple_opt_pass along with...
13393 (pass_data_graphite_transforms): ...new pass_data instance and...
13394 (make_pass_graphite_transforms): ...new function.
13395 (pass_check_data_deps): Convert from a global struct to a subclass of
13396 gimple_opt_pass along with...
13397 (pass_data_check_data_deps): ...new pass_data instance and...
13398 (make_pass_check_data_deps): ...new function.
13399 (pass_iv_canon): Convert from a global struct to a subclass of
13400 gimple_opt_pass along with...
13401 (pass_data_iv_canon): ...new pass_data instance and...
13402 (make_pass_iv_canon): ...new function.
13403 (pass_scev_cprop): Convert from a global struct to a subclass of
13404 gimple_opt_pass along with...
13405 (pass_data_scev_cprop): ...new pass_data instance and...
13406 (make_pass_scev_cprop): ...new function.
13407 (pass_record_bounds): Convert from a global struct to a subclass of
13408 gimple_opt_pass along with...
13409 (pass_data_record_bounds): ...new pass_data instance and...
13410 (make_pass_record_bounds): ...new function.
13411 (pass_complete_unroll): Convert from a global struct to a subclass of
13412 gimple_opt_pass along with...
13413 (pass_data_complete_unroll): ...new pass_data instance and...
13414 (make_pass_complete_unroll): ...new function.
13415 (pass_complete_unrolli): Convert from a global struct to a subclass of
13416 gimple_opt_pass along with...
13417 (pass_data_complete_unrolli): ...new pass_data instance and...
13418 (make_pass_complete_unrolli): ...new function.
13419 (pass_parallelize_loops): Convert from a global struct to a subclass
13420 of gimple_opt_pass along with...
13421 (pass_data_parallelize_loops): ...new pass_data instance and...
13422 (make_pass_parallelize_loops): ...new function.
13423 (pass_loop_prefetch): Convert from a global struct to a subclass of
13424 gimple_opt_pass along with...
13425 (pass_data_loop_prefetch): ...new pass_data instance and...
13426 (make_pass_loop_prefetch): ...new function.
13427 (pass_iv_optimize): Convert from a global struct to a subclass of
13428 gimple_opt_pass along with...
13429 (pass_data_iv_optimize): ...new pass_data instance and...
13430 (make_pass_iv_optimize): ...new function.
13431 (pass_tree_loop_done): Convert from a global struct to a subclass of
13432 gimple_opt_pass along with...
13433 (pass_data_tree_loop_done): ...new pass_data instance and...
13434 (make_pass_tree_loop_done): ...new function.
13435 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
13436 struct to a subclass of gimple_opt_pass along with...
13437 (pass_data_cse_reciprocals): ...new pass_data instance and...
13438 (make_pass_cse_reciprocals): ...new function.
13439 (pass_cse_sincos): Convert from a global struct to a subclass of
13440 gimple_opt_pass along with...
13441 (pass_data_cse_sincos): ...new pass_data instance and...
13442 (make_pass_cse_sincos): ...new function.
13443 (pass_optimize_bswap): Convert from a global struct to a subclass of
13444 gimple_opt_pass along with...
13445 (pass_data_optimize_bswap): ...new pass_data instance and...
13446 (make_pass_optimize_bswap): ...new function.
13447 (pass_optimize_widening_mul): Convert from a global struct to a
13448 subclass of gimple_opt_pass along with...
13449 (pass_data_optimize_widening_mul): ...new pass_data instance and...
13450 (make_pass_optimize_widening_mul): ...new function.
13451 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
13452 subclass of gimple_opt_pass along with...
13453 (pass_data_phiopt): ...new pass_data instance and...
13454 (make_pass_phiopt): ...new function.
13455 (pass_cselim): Convert from a global struct to a subclass of
13456 gimple_opt_pass along with...
13457 (pass_data_cselim): ...new pass_data instance and...
13458 (make_pass_cselim): ...new function.
13459 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
13460 subclass of gimple_opt_pass along with...
13461 (pass_data_phiprop): ...new pass_data instance and...
13462 (make_pass_phiprop): ...new function.
13463 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
13464 subclass of gimple_opt_pass along with...
13465 (pass_data_pre): ...new pass_data instance and...
13466 (make_pass_pre): ...new function.
13467 (pass_fre): Convert from a global struct to a subclass of
13468 gimple_opt_pass along with...
13469 (pass_data_fre): ...new pass_data instance and...
13470 (make_pass_fre): ...new function.
13471 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
13472 subclass of gimple_opt_pass along with...
13473 (pass_data_reassoc): ...new pass_data instance and...
13474 (make_pass_reassoc): ...new function.
13475 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
13476 subclass of gimple_opt_pass along with...
13477 (pass_data_sink_code): ...new pass_data instance and...
13478 (make_pass_sink_code): ...new function.
13479 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
13480 subclass of gimple_opt_pass along with...
13481 (pass_data_strlen): ...new pass_data instance and...
13482 (make_pass_strlen): ...new function.
13483 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
13484 struct to a subclass of gimple_opt_pass along with...
13485 (pass_data_build_alias): ...new pass_data instance and...
13486 (make_pass_build_alias): ...new function.
13487 (pass_build_ealias): Convert from a global struct to a subclass of
13488 gimple_opt_pass along with...
13489 (pass_data_build_ealias): ...new pass_data instance and...
13490 (make_pass_build_ealias): ...new function.
13491 (pass_ipa_pta): Convert from a global struct to a subclass of
13492 simple_ipa_opt_pass along with...
13493 (pass_data_ipa_pta): ...new pass_data instance and...
13494 (make_pass_ipa_pta): ...new function.
13495 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
13496 subclass of gimple_opt_pass along with...
13497 (pass_data_uncprop): ...new pass_data instance and...
13498 (make_pass_uncprop): ...new function.
13499 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
13500 global struct to a subclass of gimple_opt_pass along with...
13501 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
13502 (make_pass_late_warn_uninitialized): ...new function.
13503 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
13504 to a subclass of gimple_opt_pass along with...
13505 (pass_data_init_datastructures): ...new pass_data instance and...
13506 (make_pass_init_datastructures): ...new function.
13507 (pass_early_warn_uninitialized): Convert from a global struct to a
13508 subclass of gimple_opt_pass along with...
13509 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
13510 (make_pass_early_warn_uninitialized): ...new function.
13511 (pass_update_address_taken): Convert from a global struct to a
13512 subclass of gimple_opt_pass along with...
13513 (pass_data_update_address_taken): ...new pass_data instance and...
13514 (make_pass_update_address_taken): ...new function.
13515 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
13516 struct to a subclass of gimple_opt_pass along with...
13517 (pass_data_release_ssa_names): ...new pass_data instance and...
13518 (make_pass_release_ssa_names): ...new function.
13519 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
13520 subclass of gimple_opt_pass along with...
13521 (pass_data_stdarg): ...new pass_data instance and...
13522 (make_pass_stdarg): ...new function.
13523 * tree-switch-conversion.c (pass_convert_switch): Convert from a
13524 global struct to a subclass of gimple_opt_pass along with...
13525 (pass_data_convert_switch): ...new pass_data instance and...
13526 (make_pass_convert_switch): ...new function.
13527 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
13528 to a subclass of gimple_opt_pass along with...
13529 (pass_data_tail_recursion): ...new pass_data instance and...
13530 (make_pass_tail_recursion): ...new function.
13531 (pass_tail_calls): Convert from a global struct to a subclass of
13532 gimple_opt_pass along with...
13533 (pass_data_tail_calls): ...new pass_data instance and...
13534 (make_pass_tail_calls): ...new function.
13535 * tree-vect-generic.c (pass_lower_vector): Convert from a global
13536 struct to a subclass of gimple_opt_pass along with...
13537 (pass_data_lower_vector): ...new pass_data instance and...
13538 (make_pass_lower_vector): ...new function.
13539 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
13540 gimple_opt_pass along with...
13541 (pass_data_lower_vector_ssa): ...new pass_data instance and...
13542 (make_pass_lower_vector_ssa): ...new function.
13543 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
13544 to a subclass of gimple_opt_pass along with...
13545 (pass_data_slp_vectorize): ...new pass_data instance and...
13546 (make_pass_slp_vectorize): ...new function.
13547 (pass_ipa_increase_alignment): Convert from a global struct to a
13548 subclass of simple_ipa_opt_pass along with...
13549 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
13550 (make_pass_ipa_increase_alignment): ...new function.
13551 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
13552 gimple_opt_pass along with...
13553 (pass_data_vrp): ...new pass_data instance and...
13554 (make_pass_vrp): ...new function.
13555 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
13556 subclass of simple_ipa_opt_pass along with...
13557 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
13558 (make_pass_ipa_free_lang_data): ...new function.
13559 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
13560 gimple_opt_pass along with...
13561 (pass_data_tsan): ...new pass_data instance and...
13562 (make_pass_tsan): ...new function.
13563 (pass_tsan_O0): Convert from a global struct to a subclass of
13564 gimple_opt_pass along with...
13565 (pass_data_tsan_O0): ...new pass_data instance and...
13566 (make_pass_tsan_O0): ...new function.
13567 * var-tracking.c (pass_variable_tracking): Convert from a global
13568 struct to a subclass of rtl_opt_pass along with...
13569 (pass_data_variable_tracking): ...new pass_data instance and...
13570 (make_pass_variable_tracking): ...new function.
13571 * web.c (pass_web): Convert from a global struct to a subclass of
13572 rtl_opt_pass along with...
13573 (pass_data_web): ...new pass_data instance and...
13574 (make_pass_web): ...new function.
13575 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
13576 declaration with that of...
13577 (make_pass_mode_switch_use): ...new function.
13578 (pass_resolve_sw_modes): Replace declaration with that of...
13579 (make_pass_resolve_sw_modes): ...new function.
13580 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
13581 from a global struct to a subclass of rtl_opt_pass along with...
13582 (pass_data_mode_switch_use): ...new pass_data instance and...
13583 (make_pass_mode_switch_use): ...new function.
13584 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
13585 from a global struct to a subclass of rtl_opt_pass along with...
13586 (pass_data_resolve_sw_modes): ...new pass_data instance and...
13587 (make_pass_resolve_sw_modes): ...new function.
13588 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
13589 struct to a subclass of rtl_opt_pass along with...
13590 (pass_data_insert_vzeroupper): ...new pass_data instance and...
13591 (make_pass_insert_vzeroupper): ...new function.
13592 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
13593 global struct to a subclass of rtl_opt_pass along with...
13594 (pass_data_work_around_errata): ...new pass_data instance and...
13595 (make_pass_work_around_errata): ...new function.
13596 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
13597 struct to a subclass of rtl_opt_pass along with...
13598 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
13599 (make_pass_mips_machine_reorg2): ...new function.
13600
13601 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13602
13603 * passes.c (pass_manager::operator new): New.
13604
13605 2013-08-05 David Malcolm <dmalcolm@redhat.com>
13606
13607 Handwritten part of conversion of passes to C++ classes.
13608
13609 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
13610 (toplev.o): Add dep on PASS_MANAGER_H.
13611 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
13612 of early local pases to reflect this moving from a global to a
13613 member of gcc::pass_manager.
13614 (cgraph_add_new_function): Likewise.
13615 * lto-cgraph.c (lto_output_node): Update for conversion of
13616 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
13617 * passes.c (opt_pass::clone): New.
13618 (opt_pass::gate): New.
13619 (opt_pass::execute): New.
13620 (opt_pass::opt_pass): New.
13621 (pass_manager::execute_early_local_passes): New.
13622 (pass_manager::execute_pass_mode_switching): new.
13623 (finish_optimization_passes): Convert to...
13624 (pass_manager::finish_optimization_passes): ...this.
13625 (finish_optimization_passes): Update for conversion of passes to
13626 C++ classes.
13627 (register_dump_files_1): Use has_gate since we cannot portably
13628 check a vtable entry against NULL.
13629 (dump_one_pass): Likewise.
13630 (ipa_write_summaries_2): Likewise.
13631 (ipa_write_optimization_summaries_1): Likewise.
13632 (ipa_read_summaries_1): Likewise.
13633 (ipa_read_optimization_summaries_1): Likewise.
13634 (execute_ipa_stmt_fixups): Likewise.
13635 (pass_manager::pass_manager): Rewrite pass-creation, invoking
13636 pass-creation functions rather than wiring up globals, and
13637 storing the results in fields of pass_manager generated using
13638 pass-instances.def.
13639 (pass_manager::dump_profile_report): Update for conversion of
13640 passes to C++ classes.
13641 (pass_manager::execute_ipa_summary_passes): Likewise.
13642 (execute_one_ipa_transform_pass): Likewise.
13643 (execute_one_pass): Use has_gate and has_execute since we cannot
13644 portably check a vtable entry against NULL.
13645 * pass_manager.h (pass_manager::finish_optimization_passes): New.
13646 (pass_manager): Use pass-instances.def to add fields for the
13647 various pass instances.
13648 * toplev.c (finalize): Update for move of
13649 finish_optimization_passes to a method of gcc::pass_manager.
13650 * toplev.h (finish_optimization_passes): Move to method of class
13651 pass_manager.
13652 * tree-pass.h (struct pass_data): New.
13653 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
13654 (opt_pass::gate): Convert to virtual function.
13655 (opt_pass::~opt_pass): New.
13656 (opt_pass::clone): New.
13657 (opt_pass::execute): Convert to virtual function.
13658 (opt_pass::opt_pass): New.
13659 (opt_pass::ctxt_): new.
13660 (gimple_opt_pass): Convert to subclass of opt_pass.
13661 (gimple_opt_pass::gimple_opt_pass): New.
13662 (rtl_opt_pass): Convert to subclass of opt_pass.
13663 (rtl_opt_pass::rtl_opt_pass): New.
13664 (ipa_opt_pass_d): Convert to subclass of opt_pass.
13665 (ipa_opt_pass_d::ipa_opt_pass_d): New.
13666 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
13667 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
13668 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
13669 invocation of pass_mode_switching to reflect this moving from a
13670 global to a member of gcc::pass_manager.
13671 (ix86_option_override): Rework how pass_insert_vzeroupper is
13672 added to the pass_manager to reflect autogenerated changes.
13673 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
13674
13675 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
13676
13677 PR rtl-optimization/57708
13678 * recog.c (peep2_find_free_register): Validate all regs in a
13679 multi-reg mode.
13680
13681 2013-08-05 Jan Hubicka <jh@suse.cz>
13682
13683 PR lto/57602
13684 * cgraph.c (verify_cgraph_node): Accept local flags from other
13685 partitions.
13686 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
13687 (function_and_variable_visibility): Likewise.
13688 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
13689
13690 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13691
13692 * graph.c (init_graph_slim_pretty_print): Remove.
13693 (print_graph_cfg): Do not call it. Use local pretty printer.
13694 (start_graph_dump): Likewise.
13695
13696 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13697
13698 * gimple-pretty-print.c (buffer): Remove.
13699 (initialized): Likewise.
13700 (maybe_init_pretty_print): Likewise.
13701 (print_gimple_stmt): Do not call it. Use non-static local
13702 pretty_printer variable.
13703 (print_gimple_expr): Likewise.
13704 (print_gimple_seq): Likewise.
13705 (gimple_dump_bb): Likewise.
13706
13707 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
13708
13709 * asan.c (asan_pp): Remove.
13710 (asan_pp_initialized): Likewise.
13711 (asan_pp_initialize): Likewise.
13712 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
13713 (asan_emit_stack_protection): Tidy. Use local pretty printer.
13714 (asan_add_global): Likewise.
13715
13716 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
13717
13718 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
13719 * diagnostic.h (diagnostic_flush_buffer): Adjust.
13720 * pretty-print.c (pp_formatted_text_data): Likewise.
13721 (pp_indent): Rename from pp_base_indent.
13722 (pp_format): Rename from pp_base_format.
13723 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
13724 (pp_format_verbatim): Rename from pp_base_format_verbatim.
13725 (pp_flush): Rename from pp_base_flush.
13726 (pp_set_line_maximum_length): Rename from
13727 pp_base_set_line_maximum_length.
13728 (pp_clear_output_area): Rename from pp_base_clear_output_area.
13729 (pp_set_prefix): Rename from pp_base_set_prefix.
13730 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
13731 (pp_emit_prefix): Rename from pp_base_emit_prefix.
13732 (pp_append_text): Rename from pp_base_append_text.
13733 (pp_formatted_text): Rename from pp_base_formatted_text.
13734 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
13735 (pp_remaining_character_count_for_line): Rename from
13736 pp_base_remaining_character_count_for_line.
13737 (pp_newline): Rename from pp_base_newline.
13738 (pp_character): Rename from pp_base_character.
13739 (pp_string): Rename from pp_base_string.
13740 (pp_maybe_space): Rename from pp_base_maybe_space.
13741 * asan.c (asan_pp_string): Adjust.
13742 (asan_emit_stack_protection): Likewise.
13743 (asan_add_global): Likewise.
13744 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
13745 * tree-mudflap.c (mf_varname_tree): Likewise.
13746 * tree-pretty-print.c (pp_tree_identifier): Rename from
13747 pp_base_tree_identifier.
13748 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
13749 Declare as function.
13750
13751 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13752
13753 * pretty-print.h (pp_bar_bar): New.
13754 (pp_ampersand_ampersand): Likewise.
13755 (pp_less_equal): Likewise.
13756 (pp_greater_equal): Likewise.
13757 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
13758 printer functions instead of pp_string or operators and punctuators.
13759 (dump_gimple_call): Likewise.
13760 (dump_gimple_omp_for): Likewise.
13761 (dump_gimple_transaction): Likewise.
13762 (dump_gimple_phi): Likewise.
13763 (pp_gimple_stmt_1): Likewise.
13764 * sched-vis.c (print_insn): Likewise.
13765 * tree-mudflap.c (mf_varname_tree): Likewise.
13766 * tree-pretty-print.c (dump_block_node): Likewise.
13767 (dump_generic_node): Likewise.
13768
13769 2013-08-02 Jan Hubicka <jh@suse.cz>
13770
13771 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
13772 boundaries.
13773 * lto-streamer-out.c (tree_is_indexable): Results decls and
13774 parm decls are not indexable.
13775 (DFS_write_tree_body): Do not follow args and results.
13776 (hash_tree): Likewise.
13777 (output_functions): Rearrange so struct function is needed
13778 only when real body is output; be able to also ouptut abstract
13779 functions; output DECL_ARGUMENTS and DECL_RESULT.
13780 (lto_output): When not in WPA, ale store abstract functions.
13781 (write_symbol): Do not care about RESULT_DECL.
13782 (output_symbol_p): Handle correctly sbtract decls.
13783 * lto-streamer-in.c (input_function): Rearrange so struct
13784 function can be NULL at entry; allow streaming of
13785 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
13786 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
13787 sanity check during LTO.
13788 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
13789 RESULT_DECl and DECL_ARGUMENTS.
13790 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
13791 Likewise.
13792
13793 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
13794
13795 * pretty-print.h (pp_underscore): New.
13796 (pp_comma): Tidy.
13797 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
13798 printer functions instead of pp_character.
13799 (dump_binary_rhs): Likewise.
13800 (dump_ternary_rhs): Likewise.
13801 (dump_gimple_call_args): Likewise.
13802 (pp_points_to_solution): Likewise.
13803 (dump_gimple_call): Likewise.
13804 (dump_gimple_switch): Likewise.
13805 (dump_gimple_cond): Likewise.
13806 (dump_gimple_bind): Likewise.
13807 (dump_gimple_try): Likewise.
13808 (dump_gimple_omp_for): Likewise.
13809 (dump_gimple_omp_continue): Likewise.
13810 (dump_gimple_omp_single): Likewise.
13811 (dump_gimple_omp_sections): Likewise.
13812 (dump_gimple_omp_block): Likewise.
13813 (dump_gimple_omp_critical): Likewise.
13814 (dump_gimple_transaction): Likewise.
13815 (dump_gimple_asm): Likewise.
13816 (dump_gimple_phi): Likewise.
13817 (dump_gimple_omp_parallel): Likewise.
13818 (dump_gimple_omp_task): Likewise.
13819 (dump_gimple_omp_atomic_load): Likewise.
13820 (dump_gimple_omp_atomic_store): Likewise.
13821 (dump_gimple_mem_ops): Likewise.
13822 (pp_gimple_stmt_1): Likewise.
13823 (pp_cfg_jump): Likewise.
13824 (dump_implicit_edges): Likewise.
13825 (gimple_dump_bb_for_graph): Likewise.
13826 * graph.c (draw_cfg_node): Likewise.
13827 * langhooks.c (lhd_print_error_function): Likewise.
13828 * sched-vis.c (print_exp): Likewise.
13829 (print_value): Likewise.
13830 (print_pattern): Likewise.
13831 (print_insn): Likewise.
13832 (rtl_dump_bb_for_graph): Likewise.
13833 * tree-pretty-print.c (dump_function_declaration): Likewise.
13834 (dump_array_domain): Likewise.
13835 (dump_omp_clause): Likewise.
13836 (dump_location): Likewise.
13837 (dump_generic_node): Likewise.
13838 (print_struct_decl): Likewise.
13839 * diagnostic.c (diagnostic_show_locus): Use pp_space.
13840
13841 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
13842
13843 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
13844 candidate table when replacing a candidate statement.
13845 (replace_rhs_if_not_dup): Likewise.
13846 (replace_one_candidate): Likewise.
13847
13848 2013-08-02 Jan Hubicka <jh@suse.cz>
13849 Martin Liska <marxin.liska@gmail.com>
13850
13851 * cgraphunit.c (add_new_function): Fix logic when adding from
13852 late IPA pass.
13853 (assemble_thunk): Rename to ...
13854 (expand_thunk); .. this one; export; get it working with
13855 general functions; make produced gimple valid.
13856 * cgraph.h (expand_thunk): Declare.
13857
13858 2013-08-02 Jan Hubicka <jh@suse.cz>
13859
13860 * ipa-cp.c (gather_context_independent_values): Use
13861 ipa_get_param_move_cost.
13862 (get_replacement_map): Remove PARAM; move parameter folding
13863 into tree-inline.c
13864 (create_specialized_node): Update.
13865 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
13866 assert that we have gimple body; update move_cost.
13867 (count_formal_params): Assert that we have gimple body.
13868 (ipa_dump_param): New function.
13869 (ipa_alloc_node_params): Break out from ...
13870 (ipa_initialize_node_params): ... here.
13871 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
13872 (ipa_write_node_info): Stream move costs.
13873 (ipa_read_node_info): Read move costs.
13874 (ipa_update_after_lto_read): Do not recompute node params.
13875 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
13876 (ipa_get_param): Check we are not in WPA.
13877 (ipa_get_param_move_cost): New.
13878 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
13879 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
13880 parm numbers to be present.
13881
13882 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13883
13884 PR rtl-optimization/58048
13885 * lra-constraints.c (process_alt_operands): Don't check asm
13886 operand on register.
13887
13888 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
13889
13890 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
13891 the implied StoreLoad barrier for atomic operations if before.
13892
13893 2013-08-02 Jan Hubicka <jh@suse.cz>
13894 Martin Liska <marxin.liska@gmail.com>
13895
13896 * cgraph.c (cgraph_function_body_availability): Do not check
13897 cgraph flags.
13898 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
13899 symtab_node_availability): Declare.
13900 * ipa.c (can_replace_by_local_alias): New.
13901 (function_and_variable_visibility): Use it.
13902 * symtab.c (symtab_for_node_and_aliases,
13903 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
13904
13905 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
13906
13907 PR rtl-optimization/57963
13908 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
13909 (lra_constraints): Use them.
13910
13911 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
13912
13913 * config/arm/types.md (define_attr "type"): Add "load_acq"
13914 and "store_rel".
13915 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
13916 changes.
13917 (cortex_a53_store1): Likewise.
13918
13919 2013-08-01 Jan Hubicka <jh@suse.cz>
13920
13921 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
13922 partitions are not needed.
13923
13924 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
13925
13926 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
13927 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
13928 MAYBE_NON_Q_CLASS_P where appropriate.
13929
13930 2013-08-01 Jan Hubicka <jh@suse.cz>
13931
13932 * cgraph.h (release_function_body): Declare.
13933 * tree.c (free_lang_data_in_decl): Free, parameters and return values
13934 of unused delcarations.
13935
13936 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13937
13938 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
13939 RTL form when subtracting a constant.
13940
13941 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
13942
13943 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
13944 Generate canonical plus rtx with negated immediate instead of minus
13945 where appropriate.
13946 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
13947
13948 2013-08-01 Jan Hubicka <jh@suse.cz>
13949
13950 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
13951 (cgraph_release_function_body): Likewise.
13952 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
13953 * cgraph.h (cgrpah_node): Rename abstract_and_needed
13954 to used_as_abstract_origin.
13955 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
13956 symbols used as abstract origins.
13957 * cgraphunit.c (analyze_functions): Update.
13958 * ipa.c (symtab_remove_unreachable_nodes): Recompute
13959 used_as_abstract_origin.
13960 * tree-inline.c (tree_function_versioning): Update
13961 used_as_abstract_origin; be ready for DECL_RESULT and
13962 DECL_ARGUMENTS to be NULL.
13963
13964 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
13965 for abstract functions.
13966 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
13967 real symbols.
13968
13969 2013-08-01 Jan Hubicka <jh@suse.cz>
13970
13971 * profile.c (compute_value_histograms): Fix thinko.
13972
13973 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13974
13975 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
13976 aarch-common-protos.h to extra_headers.
13977 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
13978 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
13979 * config/aarch64/t-aarch64 (aarch-common.o): Define.
13980
13981 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
13982
13983 * config/aarch64/aarch64.md (define_attr "type"): Delete.
13984 Include "../arm/types.md". Define "type" attribute for all patterns.
13985 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
13986 attribute changes.
13987
13988 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
13989
13990 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
13991 to support power8 load fusion.
13992 (fusion_gpr_mem_load): Likewise.
13993
13994 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
13995
13996 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
13997 declarations for power8 load fusion.
13998 (emit_fusion_gpr_load): Likewise.
13999
14000 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
14001 tuning for power8, turn on fusion mode by default. Turn on sign
14002 extending fusion mode if normal fusion mode is on, and we are at
14003 -O2 or -O3.
14004 (fusion_gpr_load_p): New function, return true if we can fuse an
14005 addis instruction with a dependent load to a GPR.
14006 (emit_fusion_gpr_load): Emit the instructions for power8 load
14007 fusion to GPRs.
14008
14009 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
14010 (VSX load fusion peepholes): New peepholes to fuse together an
14011 addi instruction with a VSX load instruction.
14012
14013 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
14014 peepholes to fuse an addis instruction with a load to a GPR base
14015 register. If we are supporting sign extending fusions, convert
14016 sign extending loads to zero extending loads and add an explicit
14017 sign extension.
14018
14019 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14020
14021 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
14022 aarch-common-protos.h to extra_headers.
14023 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
14024 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
14025 (arm_early_store_addr_dep): Likewise.
14026 (arm_no_early_alu_shift_dep): Likewise.
14027 (arm_no_early_alu_shift_value_dep): Likewise.
14028 (arm_no_early_mul_dep): Likewise.
14029 (arm_no_early_store_addr_dep): Likewise.
14030 (arm_mac_accumulator_is_mul_result): Likewise.
14031 (arm_mac_accumulator_is_result): Likewise.
14032 * config/arm/aarch-common.c: ... here. New file.
14033 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
14034 here to ...
14035 (arm_early_store_addr_dep): Likewise.
14036 (arm_no_early_alu_shift_dep): Likewise.
14037 (arm_no_early_alu_shift_value_dep): Likewise.
14038 (arm_no_early_mul_dep): Likewise.
14039 (arm_no_early_store_addr_dep): Likewise.
14040 (arm_mac_accumulator_is_mul_result): Likewise.
14041 (arm_mac_accumulator_is_result): Likewise.
14042 * config/arm/aarch-common-protos.h: ... here. New file.
14043 * config/arm/t-arm (aarch-common.o): Define.
14044
14045 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
14046
14047 * config/arm/arm.md: Include new file "types.md".
14048 (define_attr "type"): Move from here to ...
14049 (define_attr "mul32"): Likewise.
14050 (define_attr "mul64"): Likewise.
14051 * config/arm/types.md: ... here. New file.
14052
14053 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
14054
14055 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
14056 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
14057
14058 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
14059
14060 * gen-pass-instances.awk: Fix offset of substr().
14061
14062 2013-07-31 David Malcolm <dmalcolm@redhat.com>
14063
14064 * Makefile.in (pass-instances.def): New.
14065 (passes.o): Replace dependency on passes.def with one on
14066 pass-instances.def
14067
14068 * gen-pass-instances.awk: New.
14069
14070 * passes.c (pass_manager::pass_manager): Use pass-instances.def
14071 rather than passes.def, updating local definition of NEXT_PASS
14072 macro to add an extra NUM parameter (currently unused).
14073
14074 2013-07-30 David Malcolm <dmalcolm@redhat.com>
14075
14076 * Makefile.in (PASS_MANAGER_H): New.
14077 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
14078 (passes.o): Likewise.
14079 (statistics.o): Likewise.
14080 (cgraphunit.o): Likewise.
14081 (context.o): Depend on PASS_MANAGER_H.
14082
14083 * pass_manager.h: New.
14084
14085 * cgraphunit.c (cgraph_add_new_function): Update for moves
14086 of globals to fields of pass_manager.
14087 (analyze_function): Likewise.
14088 (expand_function): Likewise.
14089 (ipa_passes): Likewise.
14090 (compile): Likewise.
14091
14092 * context.c (context::context): New.
14093 * context.h (context::context): New.
14094 (context::get_passes): New.
14095 (context::passes_): New.
14096
14097 * lto-cgraph.c (input_node): Update for moves of globals to
14098 fields of pass_manager.
14099
14100 * passes.c (all_passes): Remove, in favor of a field of the
14101 same name within the new class pass_manager.
14102 (all_small_ipa_passes): Likewise.
14103 (all_lowering_passes): Likewise.
14104 (all_regular_ipa_passes): Likewise.
14105 (all_late_ipa_passes): Likewise.
14106 (all_lto_gen_passes): Likewise.
14107 (passes_by_id): Likewise.
14108 (passes_by_id_size): Likewise.
14109 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14110 the new class pass_manager.
14111 (set_pass_for_id): Convert to...
14112 (pass_manager::set_pass_for_id): ...method.
14113 (get_pass_for_id): Convert to...
14114 (pass_manager::get_pass_for_id): ...method.
14115 (register_one_dump_file): Move body of implementation into...
14116 (pass_manager::register_one_dump_file): ...here.
14117 (register_dump_files_1): Convert to...
14118 (pass_manager::register_dump_files_1): ...method.
14119 (register_dump_files): Convert to...
14120 (pass_manager::register_dump_files): ...method.
14121 (create_pass_tab): Update for moves of globals to fields of
14122 pass_manager.
14123 (dump_passes): Move body of implementation into...
14124 (pass_manager::dump_passes): ...here.
14125 (register_pass): Move body of implementation into...
14126 (pass_manager::register_pass): ...here.
14127 (init_optimization_passes): Convert into...
14128 (pass_manager::pass_manager): ...constructor for new
14129 pass_manager class, and initialize the pass_lists array.
14130 (check_profile_consistency): Update for moves of globals to
14131 fields of pass_manager.
14132 (dump_profile_report): Move body of implementation into...
14133 (pass_manager::dump_profile_report): ...here.
14134 (ipa_write_summaries_1): Update for moves of pass lists from
14135 being globals to fields of pass_manager.
14136 (ipa_write_optimization_summaries): Likewise.
14137 (ipa_read_summaries): Likewise.
14138 (ipa_read_optimization_summaries): Likewise.
14139 (execute_all_ipa_stmt_fixups): Likewise.
14140
14141 * statistics.c (statistics_fini): Update for moves of globals to
14142 fields of pass_manager.
14143
14144 * toplev.c (general_init): Replace call to
14145 init_optimization_passes with construction of the pass_manager
14146 instance.
14147
14148 * tree-pass.h (all_passes): Remove, in favor of a field of the
14149 same name within the new class pass_manager.
14150 (all_small_ipa_passes): Likewise.
14151 (all_lowering_passes): Likewise.
14152 (all_regular_ipa_passes): Likewise.
14153 (all_lto_gen_passes): Likewise.
14154 (all_late_ipa_passes): Likewise.
14155 (passes_by_id): Likewise.
14156 (passes_by_id_size): Likewise.
14157 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
14158 the new class pass_manager.
14159 (get_pass_for_id): Remove.
14160
14161 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14162
14163 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
14164 configs.
14165
14166 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
14167
14168 * arm.md (mulhi3): New expand pattern.
14169
14170 2013-07-30 Jan Hubicka <jh@suse.cz>
14171 Martin Liska <marxin.liska@gmail.com>
14172
14173 * profile.c (compute_value_histograms): Do not ICE when
14174 there is mismatch only on some counters.
14175
14176 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
14177
14178 PR rtl-optimization/57637
14179 * function.c (move_insn_for_shrink_wrap): Also check the
14180 GEN set of the LIVE problem for the liveness analysis
14181 if it exists, otherwise give up.
14182
14183 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14184
14185 PR tree-optimization/57993
14186 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
14187 replaced statement in the candidate table.
14188 (phi_add_costs): Return infinite cost when the hidden basis does
14189 not dominate all phis on which the candidate is dependent.
14190 (replace_one_candidate): Record replaced statement in the
14191 candidate table.
14192
14193 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
14194
14195 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
14196 (ashlv2si3): New expander.
14197 (*ashlv2si3_i): New define_insn_and_split.
14198 * predicates.md (float_operation): Allow patterns with three
14199 basic sub-patterns.
14200
14201 PR rtl-optimization/58021
14202 * mode-switching.c (create_pre_exit): Always split off preceding
14203 insns if we are not at the basic block head.
14204
14205 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
14206
14207 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
14208 (UCLIBC_DYNAMIC_LINKER): New macro.
14209 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
14210 `-mnan=2008'.
14211 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
14212 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
14213 `-mnan=2008'.
14214 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
14215 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
14216 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
14217 for SF and DF modes. Use ieee_quad_format for TF mode.
14218 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
14219 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
14220 (mips_option_override): Handle `-mnan=legacy'.
14221 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
14222 `-mabs=2008' and `-mnan=2008'.
14223 (OPTION_DEFAULT_SPECS): Add "nan" default.
14224 (ASM_SPEC): Handle `-mnan='.
14225 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
14226 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
14227 comment accordingly.
14228 (neg<mode>2): Likewise.
14229 * config/mips/mips.opt (mabs, mnan): New options.
14230 * doc/install.texi (Configuration): Document `--with-nan=' option.
14231 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
14232 `-mnan=' options.
14233 (MIPS Options): Document them.
14234 * config.gcc <mips*-*-*>: Handle `--with-nan='.
14235 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
14236 * configure: Regenerate.
14237 * config.in: Regenerate.
14238
14239 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14240
14241 * config/i386/i386.md (float post-reload splitters): Do not check
14242 for subregs of SSE registers.
14243
14244 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
14245 H.J. Lu <hongjiu.lu@intel.com>
14246
14247 PR target/57954
14248 PR target/57988
14249 * config/i386/i386.md (post-reload splitter
14250 to avoid partial SSE reg dependency stalls): New pattern.
14251
14252 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
14253
14254 * config/s390/s390.md ("movcc"): Swap load and store instructions.
14255
14256 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
14257
14258 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
14259 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
14260
14261 2013-07-26 Cary Coutant <ccoutant@google.com>
14262
14263 * dwarf2out.c (die_checksum_ordered): Don't include template
14264 instantiations in signature.
14265 (is_template_parameter): New function.
14266 (is_template_instantiation): New function.
14267 (generate_skeleton_bottom_up): Don't include template instantiations
14268 in type unit DIE.
14269 (generate_skeleton): Likewise.
14270 (break_out_comdat_types): Move recursive call to break out nested
14271 types earlier.
14272 (prune_unused_types_mark_generic_parms_dies): Call
14273 is_template_parameter.
14274
14275 2013-07-26 Ian Bolton <ian.bolton@arm.com>
14276
14277 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
14278 uses vector registers.
14279 * config/aarch64/iterators.md: Add attributes rtn and vas.
14280
14281 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14282 Richard Earnshaw <richard.earnshaw@arm.com>
14283
14284 * combine.c (simplify_comparison): Re-canonicalize operands
14285 where appropriate.
14286 * config/arm/arm.md (movcond_addsi): New splitter.
14287
14288 2013-07-25 Sterling Augustine <saugustine@google.com>
14289
14290 * dwarf2out.c (size_of_pubnames): Move code to...
14291 (include_pubname_in_output): ...here. New.
14292 (want_pubnames): Rearrange.
14293 (output_pubnames): Call include_pubname_in_output. Move assertion.
14294
14295 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14296
14297 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
14298
14299 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
14300
14301 PR target/38836
14302 * doc/extend.texi: Remove obsolete builtins. Fix
14303 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
14304
14305 2013-07-25 Jan Hubicka <jh@suse.cz>
14306
14307 * cgraph.c (release_function_body): Break out from ...
14308 (cgraph_release_function_body): ... this one; also release DECL_RESULT
14309 and DECL_ARGUMENTS.
14310 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
14311 old_tree in the map.
14312 (create_specialized_node): Update.
14313 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
14314 into index.
14315 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
14316 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
14317 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
14318 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
14319 DECL_RESULT.
14320
14321 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14322
14323 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
14324 addsi3_carryin_alt2_<optab>): Correct output template.
14325
14326 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14327
14328 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
14329 Adjust for arm_restrict_it.
14330 Remove trailing whitespace.
14331
14332 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
14333
14334 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
14335 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
14336
14337 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
14338
14339 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
14340
14341 PR rtl-optimization/57960
14342 * lra-constraints.c (process_alt_operands): Use the right mode
14343 when checking strict_low.
14344
14345 2013-07-25 Jan Hubicka <jh@suse.cz>
14346
14347 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
14348 * cgraph.c (cgraph_remove_node): Do not release function body
14349 when in cgraph streaming.
14350 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
14351 in other partitions are not considered reachable; fix handling of
14352 clones.
14353
14354 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14355
14356 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
14357
14358 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
14359
14360 PR target/19599
14361 PR target/57731
14362 PR target/57837
14363 * config/arm/arm.md ("*sibcall_insn): Replace use of
14364 Ss with US. Adjust output for v5 and v4t.
14365 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
14366
14367 * config/arm/constraints.md ("Ss"): Rename to US.
14368
14369 2013-07-25 Terry Guo <terry.guo@arm.com>
14370
14371 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
14372 shift_add/shift_sub0/shift_sub1 RTXs.
14373
14374 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
14375 Anton Blanchard <anton@au1.ibm.com>
14376
14377 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
14378 (altivec_vpks<VI_char>ss): Likewise.
14379 (altivec_vpks<VI_char>us): Likewise.
14380 (altivec_vpku<VI_char>us): Likewise.
14381 (altivec_vpku<VI_char>um): Likewise.
14382
14383 2013-07-24 David Malcolm <dmalcolm@redhat.com>
14384
14385 Introduce context class.
14386
14387 * Makefile.in (CONTEXT_H): New.
14388 (OBJS): Add context.o.
14389 (toplev.o): Add CONTEXT_H to dependencies.
14390 (context.o): New.
14391
14392 * toplev.c (general_init): Create the singleton gcc::context instance.
14393
14394 * context.c: New.
14395
14396 * context.h: New.
14397
14398 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
14399
14400 PR rtl-optimization/57968
14401 * mode-switching.c (create_pre_exit): Allow instructions that
14402 don't set a return register to need a non-exit mode.
14403
14404 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14405 Anton Blanchard <anton@au1.ibm.com>
14406
14407 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
14408 operands to vperm for little endian.
14409 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
14410 of lvsl to create the control mask for a vperm for little endian.
14411
14412 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14413 Anton Blanchard <anton@au1.ibm.com>
14414
14415 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
14416 two operands for little-endian.
14417
14418 2013-07-23 Steve Ellcey <sellcey@mips.com>
14419
14420 * config/mips/mips.c (mips_case_values_threshold): New.
14421 (TARGET_CASE_VALUES_THRESHOLD): Define.
14422
14423 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
14424 Anton Blanchard <anton@au1.ibm.com>
14425
14426 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
14427 selection of field for vector splat in little endian mode.
14428
14429 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
14430
14431 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
14432 expanders to rs6000.md.
14433 (ior<mode>3): Likewise.
14434 (and<mode>3): Likewise.
14435 (one_cmpl<mode>2): Likewise.
14436 (nor<mode>3): Likewise.
14437 (andc<mode>3): Likewise.
14438 (eqv<mode>3): Likewise.
14439 (nand<mode>3): Likewise.
14440 (orc<mode>3): Likewise.
14441
14442 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
14443 declaration.
14444
14445 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
14446 to split multi-word logical operations.
14447 (rs6000_split_logical_di): Likewise.
14448 (rs6000_split_logical): Likewise.
14449
14450 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
14451 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
14452 and allow TImode operations in 32-bit.
14453 (vsx_and<mode>3_64bit): Likewise.
14454 (vsx_ior<mode>3_32bit): Likewise.
14455 (vsx_ior<mode>3_64bit): Likewise.
14456 (vsx_xor<mode>3_32bit): Likewise.
14457 (vsx_xor<mode>3_64bit): Likewise.
14458 (vsx_one_cmpl<mode>2_32bit): Likewise.
14459 (vsx_one_cmpl<mode>2_64bit): Likewise.
14460 (vsx_nor<mode>3_32bit): Likewise.
14461 (vsx_nor<mode>3_64bit): Likewise.
14462 (vsx_andc<mode>3_32bit): Likewise.
14463 (vsx_andc<mode>3_64bit): Likewise.
14464 (vsx_eqv<mode>3_32bit): Likewise.
14465 (vsx_eqv<mode>3_64bit): Likewise.
14466 (vsx_nand<mode>3_32bit): Likewise.
14467 (vsx_nand<mode>3_64bit): Likewise.
14468 (vsx_orc<mode>3_32bit): Likewise.
14469 (vsx_orc<mode>3_64bit): Likewise.
14470
14471 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
14472 logical types in GPRs.
14473
14474 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
14475 logical insns to rs6000.md, and allow TImode operations in
14476 32-bit.
14477 (altivec_ior<mode>3): Likewise.
14478 (altivec_xor<mode>3): Likewise.
14479 (altivec_one_cmpl<mode>2): Likewise.
14480 (altivec_nor<mode>3): Likewise.
14481 (altivec_andc<mode>3): Likewise.
14482
14483 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
14484 attributes for moving the 128-bit logical operations into
14485 rs6000.md.
14486 (BOOL_REGS_OUTPUT): Likewise.
14487 (BOOL_REGS_OP1): Likewise.
14488 (BOOL_REGS_OP2): Likewise.
14489 (BOOL_REGS_UNARY): Likewise.
14490 (BOOL_REGS_AND_CR0): Likewise.
14491 (one_cmpl<mode>2): Add support for DI logical operations on
14492 32-bit, splitting the operations to 32-bit.
14493 (anddi3): Likewise.
14494 (iordi3): Likewise.
14495 (xordi3): Likewise.
14496 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
14497 changes to combine the 32/64-bit code, allow logical operations on
14498 TI mode in 32-bit, and to use similar match_operator patterns like
14499 scalar mode uses. Combine the Altivec and VSX code for logical
14500 operations, and move it here.
14501 (ior<mode>3, 128-bit types): Likewise.
14502 (xor<mode>3, 128-bit types): Likewise.
14503 (one_cmpl<mode>3, 128-bit types): Likewise.
14504 (nor<mode>3, 128-bit types): Likewise.
14505 (andc<mode>3, 128-bit types): Likewise.
14506 (eqv<mode>3, 128-bit types): Likewise.
14507 (nand<mode>3, 128-bit types): Likewise.
14508 (orc<mode>3, 128-bit types): Likewise.
14509 (and<mode>3_internal): Likewise.
14510 (bool<mode>3_internal): Likewise.
14511 (boolc<mode>3_internal1): Likewise.
14512 (boolc<mode>3_internal2): Likewise.
14513 (boolcc<mode>3_internal1): Likewise.
14514 (boolcc<mode>3_internal2): Likewise.
14515 (eqv<mode>3_internal1): Likewise.
14516 (eqv<mode>3_internal2): Likewise.
14517 (one_cmpl1<mode>3_internal): Likewise.
14518
14519 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14520
14521 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14522 Rename flag_stack_usage to flag_stack_usage_info.
14523
14524 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
14525
14526 * config/microblaze/sync.md: New file.
14527 * config/microblaze/microblaze.md: Include sync.md
14528 * config/microblaze/microblaze.c: Add print_operand 'y'.
14529 * config/microblaze/constraints.md: Add memory_contraint
14530 'Q' which is a single register.
14531
14532 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
14533
14534 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
14535
14536 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14537
14538 * reload.c (find_reloads): Exit loop once we find this operand
14539 cannot be reloaded somehow for this alternative.
14540
14541 * reload.c (find_reloads): Exit loop once we find a hard register.
14542
14543 * rtlanal.c (computed_jump_p): Exit loop once we find label
14544 reference is used.
14545
14546 * i386.c (ix86_pad_returns): Exit loop after setting replace.
14547
14548 * cfgloopmanip.c (remove_path): Exit loop after setting
14549 irred_invalidated.
14550
14551 * gensupport.c (subst_dup): Avoid loop if code is not
14552 MATCH_DUP nor MATCH_OP_DUP.
14553
14554 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
14555
14556 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
14557
14558 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14559
14560 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
14561 true for SP_REGNUM if mode == ptr_mode.
14562 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
14563 with value R0_REGNUM + 31.
14564
14565 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14566
14567 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
14568 pad pointer-typed argument downward.
14569
14570 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14571
14572 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
14573 and __ILP32__ when the ILP32 model is in use.
14574
14575 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14576
14577 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
14578 (aarch64_load_symref_appropriately): In the case of
14579 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
14580 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
14581 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
14582 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
14583 if 'mode' doesn't equal to 'ptr_mode'.
14584 (aarch64_output_mi_thunk): Add an assertion on the alignment of
14585 'vcall_offset'; change to call aarch64_emit_move differently depending
14586 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
14587 to calculate the upper bound of 'vcall_offset'.
14588 (aarch64_cannot_force_const_mem): Change to also return true if
14589 mode != ptr_mode.
14590 (aarch64_legitimize_reload_address): In the case of large
14591 displacements, add new local variable 'xmode' and an assertion
14592 based on it; change to use 'xmode' to generate the new rtx and
14593 reload.
14594 (aarch64_asm_trampoline_template): Change to generate the template
14595 differently depending on TARGET_ILP32 or not; change to use
14596 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
14597 (aarch64_trampoline_size): Removed.
14598 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
14599 and replace immediate literals with it. Change to use 'ptr_mode'
14600 instead of 'DImode' and call convert_memory_address if the mode
14601 of 'fnaddr' doesn't equal to 'ptr_mode'.
14602 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
14603 to output symbol.
14604 (aarch64_elf_asm_destructor): Likewise.
14605 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
14606 on TARGET_ILP32 instead of aarch64_trampoline_size.
14607 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
14608 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
14609 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
14610 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
14611 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
14612 (storewb_pair<GPI:mode>_<P:mode>): ... this.
14613 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
14614 depending on the value of 'mode'.
14615 (add_losym_<mode>): New.
14616 (ldr_got_small_<mode>): New, based on ldr_got_small.
14617 (ldr_got_small): Remove.
14618 (ldr_got_small_sidi): New.
14619 * config/aarch64/iterators.md (P): New.
14620 (PTR): Change to 'ptr_mode' in the condition.
14621
14622 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
14623
14624 * config.gcc (aarch64*-*-*): Support --with-abi.
14625 (aarch64*-*-elf): Support --with-multilib-list.
14626 (aarch64*-*-linux*): Likewise.
14627 (supported_defaults): Add abi to aarch64*-*-*.
14628 * configure.ac: Mention AArch64 for --with-multilib-list.
14629 * configure: Re-generated.
14630 * config/aarch64/biarchilp32.h: New file.
14631 * config/aarch64/biarchlp64.h: New file.
14632 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
14633 (ABI_SPEC): Ditto.
14634 (MULTILIB_DEFAULTS): Ditto.
14635 (DRIVER_SELF_SPECS): Ditto.
14636 (ASM_SPEC): Update to also substitute -mabi.
14637 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
14638 file whose name depends on -mabi= and -mbig-endian.
14639 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
14640 TARGET_ILP32.
14641 (POINTER_SIZE): New define.
14642 (POINTERS_EXTEND_UNSIGNED): Ditto.
14643 (enum aarch64_abi_type): New enumeration tag.
14644 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
14645 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
14646 (TARGET_ILP32): New define.
14647 * config/aarch64/aarch64.opt (mabi): New.
14648 (aarch64_abi): New.
14649 (ilp32, lp64): New values for -mabi.
14650 * config/aarch64/t-aarch64 (comma): New define.
14651 (MULTILIB_OPTIONS): Ditto.
14652 (MULTILIB_DIRNAMES): Ditto.
14653 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
14654 * doc/invoke.texi: Document -mabi for AArch64.
14655
14656 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
14657
14658 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
14659
14660 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
14661 Anton Blanchard <anton@au1.ibm.com>
14662
14663 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
14664 endianness when selecting field to splat.
14665
14666 2013-07-22 Eric Christopher <echristo@gmail.com>
14667
14668 * dwarf2out.c (die_odr_checksum): New function to use
14669 CHECKSUM_ macros and ULEB128 for DIE tag.
14670 (generate_type_signature): Use.
14671
14672 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
14673
14674 * config.gcc (sparc*-*-*): Accept leon3 processor.
14675 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
14676 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
14677 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
14678 * config/sparc/sparc.opt (enum processor_type): Add leon3.
14679 (mfix-ut699): Adjust comment.
14680 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
14681 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
14682 (CPP_CPU_SPEC): Likewise.
14683 (ASM_CPU_SPEC): Likewise.
14684 * config/sparc/sparc.c (leon3_cost): New constant.
14685 (sparc_option_override): Add leon3 support.
14686 (mem_ref): New function.
14687 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
14688 (sparc_do_work_around_errata): Look into the instruction in the delay
14689 slot and adjust accordingly. Add fix for the data cache nullify issues
14690 of the UT699. Change insertion position for the NOP.
14691 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
14692 (leon3_load): New reservation.
14693 (leon_store): Bump latency to 2.
14694 (grfpu): New automaton.
14695 (grfpu_alu): New unit.
14696 (grfpu_ds): Likewise.
14697 (leon_fp_alu): Adjust.
14698 (leon_fp_mult): Delete.
14699 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
14700 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
14701 * config/sparc/sparc.md (cpu): Add leon3.
14702 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
14703 (swapsi): Likewise.
14704 (atomic_test_and_set): Likewise.
14705 (ldstub): Likewise.
14706
14707 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
14708
14709 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
14710 default for R5900 targets.
14711 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
14712 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
14713 * config/mips/mips.c (mips_option_override): Report an error for
14714 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
14715 for -march=r5900 -mhard-float.
14716
14717 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
14718
14719 * df-problems.c (can_move_insns_across): Exit loop once we
14720 find a non-fixed, non-global register.
14721
14722 * ipa-pure-const.c (propagate_nothrow): Exit loop after
14723 setting can_throw.
14724
14725 * omega.c (omega_eliminate_red): Break after setting red_found.
14726 (omega_problem_has_red_equations): Similarly after setting found.
14727 (omega_query_variable): Similarly after setting coupled.
14728
14729 2013-07-22 Marek Polacek <polacek@redhat.com>
14730
14731 * gimplify.c: Don't include gimple.h twice.
14732
14733 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14734
14735 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
14736 instead of TARGET_THUMB1.
14737 (Pz): New constraint.
14738 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
14739 encodings.
14740 (compare_negsi_si): Likewise.
14741 (compare_addsi2_op0): Likewise.
14742 (compare_addsi2_op1): Likewise.
14743 (addsi3_carryin_<optab>): Likewise.
14744 (addsi3_carryin_alt2_<optab>): Likewise.
14745 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
14746 for arm_restrict_it.
14747 (subsi3_carryin): Likewise.
14748 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
14749 (minmax_arithsi): Disable for arm_restrict_it.
14750 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
14751 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
14752 (satsi_<SAT:code>_shift): Likewise.
14753 (arm_shiftsi3): Add alternative for 16-bit encoding.
14754 (arm32_movhf): Disable for arm_restrict_it.
14755 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
14756 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
14757
14758 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
14759
14760 * config/arm/arm.md (attribute "insn"): Delete.
14761 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
14762 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
14763 (not_shiftsi): Update for attribute change.
14764 (not_shiftsi_compare0): Likewise.
14765 (not_shiftsi_compare0_scratch): Likewise.
14766 (arm_one_cmplsi2): Likewise.
14767 (thumb1_one_cmplsi2): Likewise.
14768 (notsi_compare0): Likewise.
14769 (notsi_compare0_scratch): Likewise.
14770 (thumb1_movdi_insn): Likewise.
14771 (arm_movsi_insn): Likewise.
14772 (movhi_insn_arch4): Likewise.
14773 (movhi_bytes): Likewise.
14774 (arm_movqi_insn): Likewise.
14775 (thumb1_movqi_insn): Likewise.
14776 (arm32_movhf): Likewise.
14777 (thumb1_movhf): Likewise.
14778 (arm_movsf_soft_insn): Likewise.
14779 (thumb1_movsf_insn): Likewise.
14780 (thumb_movdf_insn): Likewise.
14781 (movsicc_insn): Likewise.
14782 (movsfcc_soft_insn): Likewise.
14783 (and_scc): Likewise.
14784 (cond_move): Likewise.
14785 (if_move_not): Likewise.
14786 (if_not_move): Likewise.
14787 (if_shift_move): Likewise.
14788 (if_move_shift): Likewise.
14789 (if_shift_shift): Likewise.
14790 (if_not_arith): Likewise.
14791 (if_arith_not): Likewise.
14792 (cond_move_not): Likewise.
14793 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
14794 (neon_mov<mode>): Likewise.
14795 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
14796 (thumb2_movsi_vfp): Likewise.
14797 (movsf_vfp): Likewise.
14798 (thumb2_movsf_vfp): Likewise.
14799 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
14800 change.
14801 (cortexa7_older_only): Likewise.
14802 (cortexa7_younger): Likewise.
14803 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
14804 (1020alu_shift_op): Likewise.
14805 (1020alu_shift_reg_op): Likewise.
14806 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
14807 (alu_shift_op): Likewise.
14808 (alu_shift_reg_op): Likewise.
14809 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
14810 (11_alu_shift_op): Likewise.
14811 (11_alu_shift_reg_op): Likewise.
14812 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
14813 (9_alu_shift_reg_op): Likewise.
14814 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
14815 change.
14816 (cortex_a15_alu_shift): Likewise.
14817 (cortex_a15_alu_shift_reg): Likewise.
14818 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
14819 (cortex_a5_alu_shift): Likewise.
14820 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
14821 change.
14822 (cortex_a53_alu_shift): Likewise.
14823 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
14824 change.
14825 (cortex_a7_alu_reg): Likewise.
14826 (cortex_a7_alu_shift): Likewise.
14827 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
14828 (cortex_a8_alu_shift): Likewise.
14829 (cortex_a8_alu_shift_reg): Likewise.
14830 (cortex_a8_mov): Likewise.
14831 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
14832 (cortex_a9_dp_shift): Likewise.
14833 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
14834 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
14835 (cortex_r4_mov): Likewise.
14836 (cortex_r4_alu_shift): Likewise.
14837 (cortex_r4_alu_shift_reg): Likewise.
14838 * config/arm/fa526.md (526_alu_op): Update for attribute change.
14839 (526_alu_shift_op): Likewise.
14840 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
14841 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
14842 (626te_alu_shift_op): Likewise.
14843 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
14844 (726te_alu_op): Likewise.
14845 (726te_alu_shift_op): Likewise.
14846 (726te_alu_shift_reg_op): Likewise.
14847 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
14848 (mp626_alu_shift_op): Likewise.
14849 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
14850 (pj4_alu_e1_conds): Likewise.
14851 (pj4_alu): Likewise.
14852 (pj4_alu_conds): Likewise.
14853 (pj4_shift): Likewise.
14854 (pj4_shift_conds): Likewise.
14855 (pj4_alu_shift): Likewise.
14856 (pj4_alu_shift_conds): Likewise.
14857
14858 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14859
14860 * config/arm/predicates.md (shiftable_operator_strict_it):
14861 New predicate.
14862 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
14863 Disable cond_exec version for arm_restrict_it.
14864 (thumb2_smaxsi3): Convert to generate cond_exec.
14865 (thumb2_sminsi3): Likewise.
14866 (thumb32_umaxsi3): Likewise.
14867 (thumb2_uminsi3): Likewise.
14868 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
14869 (thumb2_neg_abssi2): Likewise.
14870 (thumb2_mov_scc): Add alternative for 16-bit encoding.
14871 (thumb2_movsicc_insn): Adjust alternatives.
14872 (thumb2_mov_negscc): Disable for arm_restrict_it.
14873 (thumb2_mov_negscc_strict_it): New pattern.
14874 (thumb2_mov_notscc_strict_it): New pattern.
14875 (thumb2_mov_notscc): Disable for arm_restrict_it.
14876 (thumb2_ior_scc): Likewise.
14877 (thumb2_ior_scc_strict_it): New pattern.
14878 (thumb2_cond_move): Adjust for arm_restrict_it.
14879 (thumb2_cond_arith): Disable for arm_restrict_it.
14880 (thumb2_cond_arith_strict_it): New pattern.
14881 (thumb2_cond_sub): Adjust for arm_restrict_it.
14882 (thumb2_movcond): Likewise.
14883 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
14884 (thumb2_zero_extendhisi2_v6): Likewise.
14885 (thumb2_zero_extendqisi2_v6): Likewise.
14886 (orsi_notsi_si): Likewise.
14887 (orsi_not_shiftsi_si): Likewise.
14888
14889 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
14890
14891 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
14892 instruction sequence is 1 byte shorter.
14893
14894 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
14895
14896 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
14897 it is not needed after split.
14898
14899 2013-07-20 Iain Sandoe <iain@codesourcery.com>
14900
14901 PR target/51784
14902 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
14903 second label for nonlocal goto receivers. Don't output pic base labels
14904 unless we're producing PIC; mark that action unreachable().
14905 (ix86_save_reg): If the function contains a nonlocal label, save the
14906 PIC base reg.
14907 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
14908 * config/darwin.c (emitted_pic_label_num): New GTY.
14909 (update_pic_label_number_if_needed): New.
14910 (machopic_output_function_base_name): Adjust for nonlocal receiver
14911 case.
14912 (machopic_should_output_picbase_label): New.
14913 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
14914 (nonlocal_goto_receiver): New insn and split.
14915
14916 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
14917
14918 * config/aarch64/aarch64-builtins.c
14919 (aarch64_fold_builtin): Fold abs in all modes.
14920 * config/aarch64/aarch64-simd-builtins.def
14921 (abs): Enable for all modes.
14922 * config/aarch64/arm_neon.h
14923 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
14924 (vabs_f64): Add missing intrinsic.
14925
14926 2013-07-19 Ian Bolton <ian.bolton@arm.com>
14927
14928 * config/aarch64/arm_neon.h (vabs_s64): New function
14929
14930 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
14931
14932 PR target/57516
14933 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
14934 * config/avr/avr.md (adjust_len): Add `round'.
14935 * config/avr/avr-protos.h (avr_out_round): New prototype.
14936 (avr_out_plus): Add `out_label' argument.
14937 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
14938 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
14939 Handle the case where `insn' is just a pattern.
14940 (avr_out_bitop): Handle the case where `insn' is just a pattern.
14941 (avr_out_round): New function.
14942 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
14943
14944 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
14945
14946 * config/microblaze/microblaze.c (microblaze_expand_prologue):
14947 Add check for flag_stack_usage to handle -fstack-usage support
14948
14949 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
14950
14951 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
14952 interaction for new Power8 flags and VSX.
14953
14954 2013-07-18 Sriraman Tallam <tmsriram@google.com>
14955
14956 PR middle-end/57698
14957 * tree-inline.c (expand_call_inline): Emit errors during
14958 early_inlining only if optimization is not turned on.
14959
14960 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14961
14962 * passes.def: New.
14963
14964 * passes.c (init_optimization_passes): Move the construction of
14965 the pass hierarchy into a new passes.def file.
14966
14967 * Makefile.in (passes.o): Add dependency on passes.def.
14968
14969 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14970
14971 * passes.c (init_optimization_passes): Introduce macros for
14972 constructing the tree of passes (INSERT_PASSES_AFTER,
14973 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
14974 TERMINATE_PASS_LIST).
14975
14976 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
14977 Wei Mi <wmi@google.com>
14978
14979 PR rtl-optimization/57878
14980 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
14981 top.
14982 (reload_pseudo_compare_func): Check nregs first for reload
14983 pseudos.
14984
14985 2013-07-18 David Malcolm <dmalcolm@redhat.com>
14986
14987 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
14988
14989 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
14990
14991 * read-rtl.c (validate_const_int): Once an invalid character is
14992 seen, quit the loop.
14993
14994 * gengtype.c (write_roots): Similarly once we find the "deletable"
14995 or "if_marked" option.
14996
14997 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
14998
14999 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
15000 "xtab" and "sat". Move value "clz" from here to ...
15001 (attriubte "type"): ... here.
15002 (satsi_<SAT:code>): Delete "insn" attribute.
15003 (satsi_<SAT:code>_shift): Likewise.
15004 (arm_zero_extendqisi2addsi): Likewise.
15005 (arm_extendqisi2addsi): Likewise.
15006 (clzsi2): Update for attribute changes.
15007 (rbitsi2): Likewise.
15008 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
15009 attribute.
15010 (arm_usatsihi): Likewise.
15011 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
15012
15013 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
15014
15015 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
15016 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
15017 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
15018 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
15019 in alphabetical order.
15020 (attribute "core_cycles"): Update for attribute changes.
15021 (arm_addsi3): Likewise.
15022 (addsi3_compare0): Likewise.
15023 (addsi3_compare0_scratch): Likewise.
15024 (addsi3_compare_op1): Likewise.
15025 (addsi3_compare_op2): Likewise.
15026 (compare_addsi2_op0): Likewise.
15027 (compare_addsi2_op1): Likewise.
15028 (addsi3_carryin_shift_<optab>): Likewise.
15029 (subsi3_carryin_shift): Likewise.
15030 (rsbsi3_carryin_shift): Likewise.
15031 (arm_subsi3_insn): Likewise.
15032 (subsi3_compare0): Likewise.
15033 (subsi3_compare): Likewise.
15034 (arm_andsi3_insn): Likewise.
15035 (thumb1_andsi3_insn): Likewise.
15036 (andsi3_compare0): Likewise.
15037 (andsi3_compare0_scratch): Likewise.
15038 (zeroextractsi_compare0_scratch
15039 (andsi_not_shiftsi_si): Likewise.
15040 (iorsi3_insn): Likewise.
15041 (iorsi3_compare0): Likewise.
15042 (iorsi3_compare0_scratch): Likewise.
15043 (arm_xorsi3): Likewise.
15044 (thumb1_xorsi3_insn): Likewise.
15045 (xorsi3_compare0): Likewise.
15046 (xorsi3_compare0_scratch): Likewise.
15047 (satsi_<SAT:code>_shift): Likewise.
15048 (rrx): Likewise.
15049 (arm_shiftsi3): Likewise.
15050 (shiftsi3_compare0): Likewise.
15051 (not_shiftsi): Likewise.
15052 (not_shiftsi_compare0): Likewise.
15053 (not_shiftsi_compare0_scratch): Likewise.
15054 (arm_one_cmplsi2): Likewise.
15055 (thumb_one_complsi2): Likewise.
15056 (notsi_compare0): Likewise.
15057 (notsi_compare0_scratch): Likewise.
15058 (thumb1_zero_extendhisi2): Likewise.
15059 (arm_zero_extendhisi2): Likewise.
15060 (arm_zero_extendhisi2_v6): Likewise.
15061 (arm_zero_extendhisi2addsi): Likewise.
15062 (thumb1_zero_extendqisi2): Likewise.
15063 (thumb1_zero_extendqisi2_v6): Likewise.
15064 (arm_zero_extendqisi2): Likewise.
15065 (arm_zero_extendqisi2_v6): Likewise.
15066 (arm_zero_extendqisi2addsi): Likewise.
15067 (thumb1_extendhisi2): Likewise.
15068 (arm_extendhisi2): Likewise.
15069 (arm_extendhisi2_v6): Likewise.
15070 (arm_extendqisi): Likewise.
15071 (arm_extendqisi_v6): Likewise.
15072 (arm_extendqisi2addsi): Likewise.
15073 (thumb1_extendqisi2): Likewise.
15074 (thumb1_movdi_insn): Likewise.
15075 (arm_movsi_insn): Likewise.
15076 (movsi_compare0): Likewise.
15077 (movhi_insn_arch4): Likewise.
15078 (movhi_bytes): Likewise.
15079 (arm_movqi_insn): Likewise.
15080 (thumb1_movqi_insn): Likewise.
15081 (arm32_movhf): Likewise.
15082 (thumb1_movhf): Likewise.
15083 (arm_movsf_soft_insn): Likewise.
15084 (thumb1_movsf_insn): Likewise.
15085 (movdf_soft_insn): Likewise.
15086 (thumb_movdf_insn): Likewise.
15087 (arm_cmpsi_insn): Likewise.
15088 (cmpsi_shiftsi): Likewise.
15089 (cmpsi_shiftsi_swp): Likewise.
15090 (arm_cmpsi_negshiftsi_si): Likewise.
15091 (movsicc_insn): Likewise.
15092 (movsfcc_soft_insn): Likewise.
15093 (arith_shiftsi): Likewise.
15094 (arith_shiftsi_compare0
15095 (arith_shiftsi_compare0_scratch
15096 (sub_shiftsi): Likewise.
15097 (sub_shiftsi_compare0
15098 (sub_shiftsi_compare0_scratch
15099 (and_scc): Likewise.
15100 (cond_move): Likewise.
15101 (if_plus_move): Likewise.
15102 (if_move_plus): Likewise.
15103 (if_move_not): Likewise.
15104 (if_not_move): Likewise.
15105 (if_shift_move): Likewise.
15106 (if_move_shift): Likewise.
15107 (if_shift_shift): Likewise.
15108 (if_not_arith): Likewise.
15109 (if_arith_not): Likewise.
15110 (cond_move_not): Likewise.
15111 (thumb1_ashlsi3): Set type attribute.
15112 (thumb1_ashrsi3): Likewise.
15113 (thumb1_lshrsi3): Likewise.
15114 (thumb1_rotrsi3): Likewise.
15115 (shiftsi3_compare0_scratch): Likewise.
15116 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
15117 (neon_mov<mode>): Likewise.
15118 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
15119 attribute changes.
15120 (thumb2_movsi_insn): Likewise.
15121 (thumb2_cmpsi_neg_shiftsi): Likewise.
15122 (thumb2_extendqisi_v6): Likewise.
15123 (thumb2_zero_extendhisi2_v6): Likewise.
15124 (thumb2_zero_extendqisi2_v6): Likewise.
15125 (thumb2_shiftsi3_short): Likewise.
15126 (thumb2_addsi3_compare0_scratch): Likewise.
15127 (orsi_not_shiftsi_si): Likewise.
15128 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
15129 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
15130 changes.
15131 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
15132 (1020alu_shift_op): Likewise.
15133 (1020alu_shift_reg_op): Likewise.
15134 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
15135 (alu_shift_op): Likewise.
15136 (alu_shift_reg_op): Likewise.
15137 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
15138 (11_alu_shift_op): Likewise.
15139 (11_alu_shift_reg_op): Likewise.
15140 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
15141 (9_alu_shift_reg_op): Likewise.
15142 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
15143 changes.
15144 (cortex_a15_alu_shift): Likewise.
15145 (cortex_a15_alu_shift_reg): Likewise.
15146 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
15147 changes.
15148 (cortex_a5_alu_shift): Likewise.
15149 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
15150 changes.
15151 (cortex_a53_alu_shift): Likewise.
15152 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
15153 changes.
15154 (cortex_a7_alu_reg): Likewise.
15155 (cortex_a7_alu_shift): Likewise.
15156 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
15157 changes.
15158 (cortex_a8_alu_shift): Likewise.
15159 (cortex_a8_alu_shift_reg): Likewise.
15160 (cortex_a8_mov): Likewise.
15161 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
15162 (cortex_a9_dp_shift): Likewise.
15163 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
15164 changes.
15165 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
15166 changes.
15167 (cortex_r4_mov): Likewise.
15168 (cortex_r4_alu_shift): Likewise.
15169 (cortex_r4_alu_shift_reg): Likewise.
15170 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
15171 (526_alu_shift_op): Likewise.
15172 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
15173 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
15174 (626te_alu_shift_op): Likewise.
15175 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
15176 (726te_alu_op): Likewise.
15177 (726te_alu_shift_op): Likewise.
15178 (726te_alu_shift_reg_op): Likewise.
15179 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
15180 (mp626_alu_shift_op): Likewise.
15181 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
15182 (pj4_alu_e1_conds): Likewise.
15183 (pj4_alu): Likewise.
15184 (pj4_alu_conds): Likewise.
15185 (pj4_shift): Likewise.
15186 (pj4_shift_conds): Likewise.
15187 (pj4_alu_shift): Likewise.
15188 (pj4_alu_shift_conds): Likewise.
15189 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
15190 changes.
15191 (cortexa7_older_only): Likewise.
15192 (cortexa7_younger): Likewise.
15193
15194 2013-07-18 David Malcolm <dmalcolm@redhat.com>
15195
15196 * ipa-pure-const.c (generate_summary): Rename to...
15197 (pure_const_generate_summary): ... this.
15198
15199 2013-07-17 Iain Sandoe <iain@codesourcery.com>
15200
15201 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
15202
15203 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
15204
15205 PR target/57909
15206 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
15207 usage in HI mode.
15208
15209 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15210
15211 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
15212 enabled without -march=zEC12.
15213 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
15214 flags to be set.
15215
15216 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15217
15218 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
15219 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
15220 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
15221 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
15222 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
15223 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
15224 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
15225 ISA_HAS_FP4.
15226 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
15227 and ISA_HAS_NMADD3_NMSUB3.
15228 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
15229 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
15230 (nmsub4<mode>, nmsub3<mode>): Likewise.
15231 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
15232
15233 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15234
15235 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
15236 TARGET_MIPS5400 checking.
15237
15238 2013-07-16 Jakub Jelinek <jakub@redhat.com>
15239 Peter Bergner <bergner@vnet.ibm.com>
15240
15241 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
15242 registers in the comment.
15243 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
15244 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
15245 rather than FIRST_PSEUDO_REGISTERS.
15246
15247 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
15248
15249 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
15250 enable extra ISA flags with TARGET_HTM.
15251
15252 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
15253
15254 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
15255 Fix comment typos.
15256
15257 2013-07-15 Cong Hou <congh@google.com>
15258
15259 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
15260 in compare function for sorting.
15261
15262 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
15263
15264 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
15265 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
15266 * config/rs6000/rs6000.opt: Add -mhtm option.
15267 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
15268 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
15269 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15270 __HTM__ if the HTM instructions are available.
15271 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
15272 htm_spr_reg_operand): New define_predicates.
15273 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
15274 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
15275 Include htm.md.
15276 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
15277 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
15278 HTM builtin functions.
15279 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
15280 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
15281 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
15282 (rs6000_builtin_mask_calculate): Likewise.
15283 (rs6000_option_override_internal): Likewise.
15284 (bdesc_htm): Add new HTM builtin support.
15285 (htm_spr_num): New function.
15286 (htm_spr_regno): Likewise.
15287 (rs6000_htm_spr_icode): Likewise.
15288 (htm_expand_builtin): Likewise.
15289 (htm_init_builtins): Likewise.
15290 (rs6000_expand_builtin): Add support for HTM builtin functions.
15291 (rs6000_init_builtins): Likewise.
15292 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
15293 option.
15294 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
15295 (TARGET_HTM, MASK_HTM): Define macros.
15296 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
15297 (FIXED_REGISTERS): Likewise.
15298 (CALL_USED_REGISTERS): Likewise.
15299 (CALL_REALLY_USED_REGISTERS): Likewise.
15300 (REG_ALLOC_ORDER): Likewise.
15301 (enum reg_class): Likewise.
15302 (REG_CLASS_NAMES): Likewise.
15303 (REG_CLASS_CONTENTS): Likewise.
15304 (REGISTER_NAMES): Likewise.
15305 (ADDITIONAL_REGISTER_NAMES): Likewise.
15306 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
15307 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
15308 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
15309 * config/rs6000/htm.md: New file.
15310 * config/rs6000/htmintrin.h: New file.
15311 * config/rs6000/htmxlintrin.h: New file.
15312
15313 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
15314
15315 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15316 Define SYMBOL_TINY_GOT, update comment.
15317 * config/aarch64/aarch64.c
15318 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
15319 (aarch64_expand_mov_immediate): Likewise.
15320 (aarch64_print_operand): Likewise.
15321 (aarch64_classify_symbol): Likewise.
15322 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
15323 (ldr_got_tiny): Define.
15324
15325 2013-07-13 Tobias Grosser <tobias@grosser.es>
15326
15327 PR tree-optimization/54094
15328 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
15329 scheduling dimension for the parallelism check from the polyhedral
15330 information in the AST.
15331 * graphite-dependences.c (carries_deps): Do not assume the schedule is
15332 in 2D + 1 form.
15333
15334 2013-07-13 Jason Merrill <jason@redhat.com>
15335
15336 * print-tree.c (debug_vec_tree): Use debug_raw.
15337 (debug_raw (vec<tree, va_gc> &)): New.
15338 (debug_raw (vec<tree, va_gc> *)): New.
15339 * tree.h: Declare them.
15340
15341 2013-07-13 Bin Cheng <bin.cheng@arm.com>
15342
15343 * ifcvt.c (ifcvt_after_combine): New static variable.
15344 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
15345 for size.
15346 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
15347 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
15348 rest_of_handle_if_after_reload): Pass new argument for if_convert.
15349
15350 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
15351
15352 * config/mips/mips.c (mips_expand_call): Remove empty statement.
15353
15354 2013-07-12 Michael Matz <matz@suse.de>
15355
15356 PR middle-end/55771
15357 * convert.c (convert_to_real): Reject non-float inner types.
15358
15359 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
15360
15361 * config/aarch64/aarch64-protos.h
15362 (aarch64_simd_immediate_valid_for_move): Remove.
15363 * config/aarch64/aarch64.c (simd_immediate_info): New member.
15364 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
15365 cases.
15366 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
15367
15368 2013-07-11 Steve Ellcey <sellcey@mips.com>
15369
15370 * config/mips/mips.c (mips_conditional_register_usage): Do not
15371 use t[0-7] registers in MIPS16 mode when optimizing for size.
15372
15373 2013-07-11 Sriraman Tallam <tmsriram@google.com>
15374
15375 * config/i386/i386.c (dispatch_function_versions): Fix array
15376 indexing of function_version_info to match actual_versions.
15377
15378 2013-07-11 Teresa Johnson <tejohnson@google.com>
15379
15380 * vec.h (struct va_gc): Move release out-of-line.
15381 (va_gc::release): Call ggc_free on released vec.
15382
15383 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
15384
15385 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
15386 Require GOT register as additional operand in UNSPEC.
15387 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
15388 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15389 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
15390 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
15391 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
15392 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
15393 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
15394
15395 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
15396
15397 PR target/57631
15398 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
15399 name seen by assembler/linker rather if available.
15400
15401 2013-07-11 Andreas Schwab <schwab@suse.de>
15402
15403 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
15404
15405 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
15406
15407 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
15408
15409 2013-07-10 Joseph Myers <joseph@codesourcery.com>
15410
15411 * doc/tm.texi.in: Move hook documentation to ....
15412 * target.def: ... here.
15413
15414 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
15415 text on @hook line.
15416 * doc/tm.texi: Regenerate.
15417
15418 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
15419
15420 PR c++/57869
15421 * doc/invoke.texi: Document -Wconditionally-supported.
15422
15423 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15424
15425 PR target/57844
15426 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
15427 of my_fp.
15428
15429 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15430
15431 PR target/57506
15432 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
15433 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
15434 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
15435 Remove duplicate devices.
15436 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
15437 * config/avr/t-multilib: Regenerate.
15438 * config/avr/avr-tables.opt: Regenerate.
15439 * doc/avr-mmcu.texi: Regenerate.
15440
15441 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
15442
15443 PR target/56987
15444 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
15445
15446 2013-07-10 Graham Stott <graham.stott@btinternet.com>
15447
15448 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
15449 the cost of MULT when optimizing for size.
15450
15451 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
15452
15453 * config/cr16/cr16-protos.h: Don't include target.h.
15454
15455 2013-07-09 Joseph Myers <joseph@codesourcery.com>
15456
15457 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
15458 adjust register size for TDmode and TFmode for VSX registers.
15459
15460 2013-07-08 Kai Tietz <ktietz@redhat.com>
15461
15462 PR target/56892
15463 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
15464 hook_bool_const_tree_true.
15465
15466 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15467
15468 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
15469 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
15470 * config/s390/s390.md: Define FPR*_REGNUM constants.
15471 Fix FPR2_REGNUM constant (18 -> 17).
15472 ("*trunc<BFP:mode><DFP_ALL:mode>2")
15473 ("*trunc<DFP_ALL:mode><BFP:mode>2")
15474 ("trunc<BFP:mode><DFP_ALL:mode>2")
15475 ("trunc<DFP_ALL:mode><BFP:mode>2")
15476 ("*extend<BFP:mode><DFP_ALL:mode>2")
15477 ("*extend<DFP_ALL:mode><BFP:mode>2")
15478 ("extend<BFP:mode><DFP_ALL:mode>2")
15479 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
15480 FPR4_REGNUM.
15481
15482 2013-07-08 Graham Stott <graham.stott@btinternet.com>
15483
15484 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
15485
15486 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
15487
15488 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
15489 and cfun_fpr_bit_p to cfun_fpr_save_p.
15490 (s390_frame_area, s390_register_info, s390_frame_info)
15491 (s390_emit_prologue, s390_emit_epilogue)
15492 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
15493 register numbers.
15494 * config/s390/s390.h: Define *_REGNUM macros for floating point
15495 register numbers.
15496
15497 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
15498
15499 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
15500
15501 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
15502
15503 PR rtl-optimization/57786
15504 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
15505 and break out of the loop when it is set to false.
15506
15507 2013-07-08 Jakub Jelinek <jakub@redhat.com>
15508
15509 PR target/57819
15510 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
15511 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
15512 (const_int 63)) 0)).
15513 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
15514 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
15515 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
15516
15517 PR rtl-optimization/57829
15518 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
15519 mask bits outside of mode are just sign-extension from mode to HWI.
15520
15521 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
15522
15523 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
15524 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
15525 adjust_address instead of change_address to keep info about alignment.
15526 (emit_strmov): Remove.
15527 (emit_memmov): New function.
15528 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
15529 (expand_movmem_epilogue): Likewise and return updated rtx for
15530 destination.
15531 (expand_constant_movmem_prologue): Likewise and return updated rtx for
15532 destination and source.
15533 (decide_alignment): Refactor, handle vector_loop.
15534 (ix86_expand_movmem): Likewise.
15535 (ix86_expand_setmem): Likewise.
15536 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
15537
15538 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
15539
15540 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
15541 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
15542
15543 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15544
15545 * config/i386/sse.md (sse_movlhps): Change alternative 3
15546 of operand 2 to "m".
15547
15548 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
15549
15550 PR target/57807
15551 * config/i386/sse.md (iptr): New mode attribute.
15552 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
15553 (<sse>_vm<plusminus_insn><mode>3): Ditto.
15554 (<sse>_vmmul<mode>3): Ditto.
15555 (<sse>_vmdiv<mode>3): Ditto.
15556 (sse_vmrcpv4sf2): Ditto.
15557 (<sse>_vmsqrt<mode>2): Ditto.
15558 (sse_vmrsqrtv4sf2): Ditto.
15559 (<sse>_vm<code><mode>3): Ditto.
15560 (avx_vmcmp<mode>3): Ditto.
15561 (<sse>_vmmaskcmp<mode>3): Ditto.
15562 (<sse>_comi): Ditto.
15563 (<sse>_ucomi): Ditto.
15564 (*xop_vmfrcz_<mode>): Ditto.
15565 (*fmai_fmadd_<mode>): Ditto.
15566 (*fmai_fmsub_<mode>): Ditto.
15567 (*fmai_fnmadd_<mode>): Ditto.
15568 (*fmai_fnmsub_<mode>): Ditto.
15569 (*fma4i_vmfmadd_<mode>): Ditto.
15570 (*fma4i_vmfmsub_<mode>): Ditto.
15571 (*fma4i_vmfnmadd_<mode>): Ditto.
15572 (*fma4i_vmfnmsub_<mode>): Ditto.
15573 (*xop_vmfrcz_<mode>): Ditto.
15574 (sse_cvtps2pi): Ditto.
15575 (sse_cvttps2pi): Ditto.
15576 (sse_cvtss2si): Ditto.
15577 (sse_cvtss2si_2): Ditto.
15578 (sse_cvtss2siq_2): Ditto.
15579 (sse_cvttss2si): Ditto.
15580 (sse_cvttss2siq): Ditto.
15581 (sse_cvtsd2si): Ditto.
15582 (sse_cvtsd2si_2): Ditto.
15583 (sse_cvtsd2siq_2): Ditto.
15584 (sse_cvttsd2si): Ditto.
15585 (sse_cvttsd2siq): Ditto.
15586 (sse_cvtsd2ss): Ditto.
15587 (sse_cvtss2sd): Ditto.
15588 (avx2_pbroadcast<mode>): Ditto.
15589 (avx2_pbroadcast<mode>_1): Ditto.
15590 (*avx_vperm_broadcast_v4sf): Ditto.
15591
15592 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
15593 (sse_movlhps): Ditto.
15594 (sse_storehps): Ditto.
15595 (sse_loadhps): Ditto.
15596 (sse_storelps): Ditto.
15597 (sse_loadlps): Ditto.
15598 (*vec_concatv4sf): Ditto.
15599 (*vec_interleave_highv2df): Ditto.
15600 (*vec_interleave_lowv2df): Ditto.
15601 (*vec_extractv2df_1_sse): Ditto.
15602 (*vec_extractv2df_0_sse): Ditto.
15603 (sse2_storelpd): Ditto.
15604 (sse2_loadlpd): Ditto.
15605 (sse2_movsd): Ditto.
15606 (*vec_concatv4si): Ditto.
15607 (vec_concatv2di): Ditto.
15608
15609 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
15610 for Intel asm dialect.
15611 (mmx_punpcklwd): Ditto.
15612 (mmx_punpckldq): Ditto.
15613
15614 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
15615 for intel assembler dialect.
15616
15617 2013-07-06 Jakub Jelinek <jakub@redhat.com>
15618
15619 PR target/29776
15620 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
15621 for BUILT_IN_C{LZ,LRSB}*.
15622 * tree.h (CASE_INT_FN): Add FN##IMAX case.
15623 * tree-vrp.c (extract_range_basic): Handle
15624 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
15625 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
15626 fall thru to code calling set_value*.
15627 * builtins.c (expand_builtin): Remove *IMAX cases.
15628 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
15629 if width is bigger than 2*HWI.
15630
15631 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
15632
15633 PR rtl-optimization/55342
15634 * lra-int.h (lra_subreg_reload_pseudos): New.
15635 * lra.c: Add undoing optional reloads to the block diagram.
15636 (lra_subreg_reload_pseudos): New.
15637 (lra_optional_reload_pseudos): Change comments.
15638 (lra): Init and clear lra_subreg_reload_pseudos. Clear
15639 lra_optional_reload_pseudos after undo transformations.
15640 * lra-assigns.c (pseudo_prefix_title): New.
15641 (lra_setup_reg_renumber): Use it.
15642 (spill_for): Ditto. Check subreg reload pseudos too.
15643 (assign_by_spills): Consider subreg reload pseudos too.
15644 * lra-constraints.c (simplify_operand_subreg): Use
15645 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
15646 (curr_insn_transform): Recognize and do optional reloads.
15647 (undo_optional_reloads): New.
15648 (lra_undo_inheritance): Call undo_optional_reloads.
15649
15650 2013-07-05 Thomas Quinot <quinot@adacore.com>
15651
15652 * tree-complex.c (expand_complex_operations_1): Fix typo.
15653
15654 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
15655
15656 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
15657 (tune_params): New member 'const vec_costs'.
15658 * config/aarch64/aarch64.c (generic_vector_cost): New.
15659 (generic_tunings): New member 'generic_vector_cost'.
15660 (aarch64_builtin_vectorization_cost): New.
15661 (aarch64_add_stmt_cost): New.
15662 (TARGET_VECTORIZE_ADD_STMT_COST): New.
15663 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
15664
15665 2013-07-03 Jakub Jelinek <jakub@redhat.com>
15666
15667 PR target/57777
15668 * config/i386/predicates.md (vsib_address_operand): Disallow
15669 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
15670
15671 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
15672
15673 PR middle-end/55030
15674 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
15675 expand_builtin_setjmp_receiver.
15676 (expand_label): Adjust, call expand_builtin_setjmp_receiver
15677 with NULL for the label parameter.
15678 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
15679 the frame-pointer. Adjust comments.
15680 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
15681 only if LABEL is non-NULL.
15682
15683 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
15684
15685 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
15686 (ARM_ABI_AAPCS64): Ditto.
15687 (arm_abi): Ditto.
15688 (ARM_DEFAULT_ABI): Ditto.
15689
15690 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
15691
15692 * config/aarch64/aarch64-builtins.c
15693 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
15694 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
15695 (st1): Likewise.
15696 * config/aarch64/aarch64-simd.md
15697 (aarch64_ld1<VALL:mode>): New.
15698 (aarch64_st1<VALL:mode>): Likewise.
15699 * config/aarch64/arm_neon.h
15700 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
15701
15702 2013-07-02 Sriraman Tallam <tmsriram@google.com>
15703
15704 * config/i386/i386.c (gate_insert_vzeroupper): Check if
15705 target ISA is AVX.
15706 (ix86_option_override_internal):Turn on all -mavx target flags by
15707 default as they are dependent on AVX anyway.
15708
15709 2013-07-02 Cary Coutant <ccoutant@google.com>
15710
15711 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
15712 deterministic hash.
15713 (loc_checksum_ordered): Likewise.
15714 (hash_loc_operands): Remove inline keyword.
15715
15716 2013-07-02 Jakub Jelinek <jakub@redhat.com>
15717
15718 PR tree-optimization/57741
15719 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
15720 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
15721 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
15722 Allow REAL_CST step_exprs if flag_associative_math.
15723 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
15724
15725 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15726
15727 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
15728
15729 2013-07-02 Ian Bolton <ian.bolton@arm.com>
15730
15731 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
15732
15733 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15734
15735 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
15736 encoding.
15737 (iorsi3_insn): Likewise.
15738 (arm_xorsi3): Likewise.
15739
15740 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
15741
15742 * arm.md (attribute "wtype"): Delete. Move attribute values from here
15743 to ...
15744 (attribute "type"): ... here, and prefix with "wmmx_".
15745 (attribute "core_cycles"): Update for attribute changes.
15746 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
15747 (tbcstv4hi): Likewise.
15748 (tbcstv2si): Likewise.
15749 (iwmmxt_iordi3): Likewise.
15750 (iwmmxt_xordi3): Likewise.
15751 (iwmmxt_anddi3): Likewise.
15752 (iwmmxt_nanddi3): Likewise.
15753 (iwmmxt_arm_movdi): Likewise.
15754 (iwmmxt_movsi_insn): Likewise.
15755 (mov<mode>_internal): Likewise.
15756 (and<mode>3_iwmmxt): Likewise.
15757 (ior<mode>3_iwmmxt): Likewise.
15758 (xor<mode>3_iwmmxt): Likewise.
15759 (add<mode>3_iwmmxt): Likewise.
15760 (ssaddv8qi3): Likewise.
15761 (ssaddv4hi3): Likewise.
15762 (ssaddv2si3): Likewise.
15763 (usaddv8qi3): Likewise.
15764 (usaddv4hi3): Likewise.
15765 (usaddv2si3): Likewise.
15766 (sub<mode>3_iwmmxt): Likewise.
15767 (sssubv8qi3): Likewise.
15768 (sssubv4hi3): Likewise.
15769 (sssubv2si3): Likewise.
15770 (ussubv8qi3): Likewise.
15771 (ussubv4hi3): Likewise.
15772 (ussubv2si3): Likewise.
15773 (mulv4hi3_iwmmxt): Likewise.
15774 (smulv4hi3_highpart): Likewise.
15775 (umulv4hi3_highpart): Likewise.
15776 (iwmmxt_wmacs): Likewise.
15777 (iwmmxt_wmacsz): Likewise.
15778 (iwmmxt_wmacu): Likewise.
15779 (iwmmxt_wmacuz): Likewise.
15780 (iwmmxt_clrdi): Likewise.
15781 (iwmmxt_clrv8qi): Likewise.
15782 (iwmmxt_clr4hi): Likewise.
15783 (iwmmxt_clr2si): Likewise.
15784 (iwmmxt_uavgrndv8qi3): Likewise.
15785 (iwmmxt_uavgrndv4hi3): Likewise.
15786 (iwmmxt_uavgv8qi3): Likewise.
15787 (iwmmxt_uavgv4hi3): Likewise.
15788 (iwmmxt_tinsrb): Likewise.
15789 (iwmmxt_tinsrh): Likewise.
15790 (iwmmxt_tinsrw): Likewise.
15791 (iwmmxt_textrmub): Likewise.
15792 (iwmmxt_textrmsb): Likewise.
15793 (iwmmxt_textrmuh): Likewise.
15794 (iwmmxt_textrmsh): Likewise.
15795 (iwmmxt_textrmw): Likewise.
15796 (iwmxxt_wshufh): Likewise.
15797 (eqv8qi3): Likewise.
15798 (eqv4hi3): Likewise.
15799 (eqv2si3): Likewise.
15800 (gtuv8qi3): Likewise.
15801 (gtuv4hi3): Likewise.
15802 (gtuv2si3): Likewise.
15803 (gtv8qi3): Likewise.
15804 (gtv4hi3): Likewise.
15805 (gtv2si3): Likewise.
15806 (smax<mode>3_iwmmxt): Likewise.
15807 (umax<mode>3_iwmmxt): Likewise.
15808 (smin<mode>3_iwmmxt): Likewise.
15809 (umin<mode>3_iwmmxt): Likewise.
15810 (iwmmxt_wpackhss): Likewise.
15811 (iwmmxt_wpackwss): Likewise.
15812 (iwmmxt_wpackdss): Likewise.
15813 (iwmmxt_wpackhus): Likewise.
15814 (iwmmxt_wpackwus): Likewise.
15815 (iwmmxt_wpackdus): Likewise.
15816 (iwmmxt_wunpckihb): Likewise.
15817 (iwmmxt_wunpckihh): Likewise.
15818 (iwmmxt_wunpckihw): Likewise.
15819 (iwmmxt_wunpckilb): Likewise.
15820 (iwmmxt_wunpckilh): Likewise.
15821 (iwmmxt_wunpckilw): Likewise.
15822 (iwmmxt_wunpckehub): Likewise.
15823 (iwmmxt_wunpckehuh): Likewise.
15824 (iwmmxt_wunpckehuw): Likewise.
15825 (iwmmxt_wunpckehsb): Likewise.
15826 (iwmmxt_wunpckehsh): Likewise.
15827 (iwmmxt_wunpckehsw): Likewise.
15828 (iwmmxt_wunpckelub): Likewise.
15829 (iwmmxt_wunpckeluh): Likewise.
15830 (iwmmxt_wunpckeluw): Likewise.
15831 (iwmmxt_wunpckelsb): Likewise.
15832 (iwmmxt_wunpckelsh): Likewise.
15833 (iwmmxt_wunpckelsw): Likewise.
15834 (ror<mode>3): Likewise.
15835 (ashr<mode>3_iwmmxt): Likewise.
15836 (lshr<mode>3_iwmmxt): Likewise.
15837 (ashl<mode>3_iwmmxt): Likewise.
15838 (ror<mode>3_di): Likewise.
15839 (ashr<mode>3_di): Likewise.
15840 (lshr<mode>3_di): Likewise.
15841 (ashl<mode>3_di): Likewise.
15842 (iwmmxt_wmadds): Likewise.
15843 (iwmmxt_wmaddu): Likewise.
15844 (iwmmxt_tmia): Likewise.
15845 (iwmmxt_tmiaph): Likewise.
15846 (iwmmxt_tmiabb): Likewise.
15847 (iwmmxt_tmiatb): Likewise.
15848 (iwmmxt_tmiabt): Likewise.
15849 (iwmmxt_tmiatt): Likewise.
15850 (iwmmxt_tmovmskb): Likewise.
15851 (iwmmxt_tmovmskh): Likewise.
15852 (iwmmxt_tmovmskw): Likewise.
15853 (iwmmxt_waccb): Likewise.
15854 (iwmmxt_wacch): Likewise.
15855 (iwmmxt_waccw): Likewise.
15856 (iwmmxt_waligni): Likewise.
15857 (iwmmxt_walignr): Likewise.
15858 (iwmmxt_walignr0): Likewise.
15859 (iwmmxt_walignr1): Likewise.
15860 (iwmmxt_walignr2): Likewise.
15861 (iwmmxt_walignr3): Likewise.
15862 (iwmmxt_wsadb): Likewise.
15863 (iwmmxt_wsadh): Likewise.
15864 (iwmmxt_wsadbz): Likewise.
15865 (iwmmxt_wsadhz): Likewise.
15866 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
15867 (iwmmxt_wabsdiffb): Likewise.
15868 (iwmmxt_wabsdiffh): Likewise.
15869 (iwmmxt_wabsdiffw): Likewise.
15870 (iwmmxt_waddsubhx): Likewise
15871 (iwmmxt_wsubaddhx): Likewise.
15872 (addc<mode>3): Likewise.
15873 (iwmmxt_avg4): Likewise.
15874 (iwmmxt_avg4r): Likewise.
15875 (iwmmxt_wmaddsx): Likewise.
15876 (iwmmxt_wmaddux): Likewise.
15877 (iwmmxt_wmaddsn): Likewise.
15878 (iwmmxt_wmaddun): Likewise.
15879 (iwmmxt_wmulwsm): Likewise.
15880 (iwmmxt_wmulwum): Likewise.
15881 (iwmmxt_wmulsmr): Likewise.
15882 (iwmmxt_wmulumr): Likewise.
15883 (iwmmxt_wmulwsmr): Likewise.
15884 (iwmmxt_wmulwumr): Likewise.
15885 (iwmmxt_wmulwl): Likewise.
15886 (iwmmxt_wqmulm): Likewise.
15887 (iwmmxt_wqmulwm): Likewise.
15888 (iwmmxt_wqmulmr): Likewise.
15889 (iwmmxt_wqmulwmr): Likewise.
15890 (iwmmxt_waddbhusm): Likewise.
15891 (iwmmxt_waddbhusl): Likewise.
15892 (iwmmxt_wqmiabb): Likewise.
15893 (iwmmxt_wqmiabt): Likewise.
15894 (iwmmxt_wqmiatb): Likewise.
15895 (iwmmxt_wqmiatt): Likewise.
15896 (iwmmxt_wqmiabbn): Likewise.
15897 (iwmmxt_wqmiabtn): Likewise.
15898 (iwmmxt_wqmiatbn): Likewise.
15899 (iwmmxt_wqmiattn): Likewise.
15900 (iwmmxt_wmiabb): Likewise.
15901 (iwmmxt_wmiabt): Likewise.
15902 (iwmmxt_wmiatb): Likewise.
15903 (iwmmxt_wmiatt): Likewise.
15904 (iwmmxt_wmiabbn): Likewise.
15905 (iwmmxt_wmiabtn): Likewise.
15906 (iwmmxt_wmiatbn): Likewise.
15907 (iwmmxt_wmiattn): Likewise.
15908 (iwmmxt_wmiawbb): Likewise.
15909 (iwmmxt_wmiawbt): Likewise.
15910 (iwmmxt_wmiawtb): Likewise.
15911 (iwmmxt_wmiawtt): Likewise.
15912 (iwmmxt_wmiawbbn): Likewise.
15913 (iwmmxt_wmiawbtn): Likewise.
15914 (iwmmxt_wmiawtbn): Likewise.
15915 (iwmmxt_wmiawttn): Likewise.
15916 (iwmmxt_wmerge): Likewise.
15917 (iwmmxt_tandc<mode>3): Likewise.
15918 (iwmmxt_torc<mode>3): Likewise.
15919 (iwmmxt_torvsc<mode>3): Likewise.
15920 (iwmmxt_textrc<mode>3): Likewise.
15921 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
15922 (wmmxt_pack): Likewise.
15923 (wmmxt_mult_c1): Likewise.
15924 (wmmxt_mult_c2): Likewise.
15925 (wmmxt_alu_c1): Likewise.
15926 (wmmxt_alu_c2): Likewise.
15927 (wmmxt_alu_c3): Likewise.
15928 (wmmxt_transfer_c1): Likewise.
15929 (wmmxt_transfer_c2): Likewise.
15930 (wmmxt_transfer_c3): Likewise.
15931 (marvell_f_iwmmxt_wstr): Likewise.
15932 (marvell_f_iwmmxt_wldr): Likewise.
15933
15934 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
15935
15936 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
15937
15938 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15939
15940 Revert:
15941 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15942 * lra-constraints.c (need_for_split_p): Check call used hard regs
15943 living through calls.
15944
15945 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
15946 call used regs for call insn.
15947
15948 2013-06-28 Jakub Jelinek <jakub@redhat.com>
15949
15950 PR target/57736
15951 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
15952 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
15953
15954 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15955
15956 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
15957
15958 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
15959
15960 * lra-constraints.c (need_for_split_p): Check call used hard regs
15961 living through calls.
15962
15963 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
15964
15965 PR target/57744
15966 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
15967 to tie with any other modes. Eliminate Altivec vector mode tests,
15968 since these are a subset of ALTIVEC or VSX vector modes. Simplify
15969 code, to return 0 if testing MODE2 for a condition, if we've
15970 already tested MODE1 for the same condition.
15971
15972 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15973
15974 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
15975 layout.
15976
15977 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15978
15979 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
15980 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
15981
15982 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
15983
15984 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
15985 Define.
15986 (aarch64_symbolic_constant_p): Remove.
15987 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
15988 static. Fix line length and white space.
15989 (aarch64_symbolic_constant_p): Remove.
15990 * config/aarch64/predicates.md (aarch64_valid_symref):
15991 Use aarch64_classify_symbol_expression.
15992
15993 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
15994
15995 * config/arm/constraints.md (Ts): New constraint.
15996 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
15997 16-bit encodings.
15998 (compare_scc): Use "Ts" constraint for operand 0.
15999 (ior_scc_scc): Likewise.
16000 (and_scc_scc): Likewise.
16001 (and_scc_scc_nodom): Likewise.
16002 (ior_scc_scc_cmp): Likewise for operand 7.
16003 (and_scc_scc_cmp): Likewise.
16004 * config/arm/thumb2.md (thumb2_movsi_insn):
16005 Add alternatives for 16-bit encodings.
16006 (thumb2_movhi_insn): Likewise.
16007 (thumb2_movsicc_insn): Likewise.
16008 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
16009 (thumb2_negscc): Use "Ts" constraint.
16010 Move mvn instruction outside cond_exec block.
16011 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
16012 for 16-bit encodings.
16013
16014 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16015
16016 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
16017 encoding.
16018 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
16019 (mulsi3subsi): Likewise.
16020 (mulsidi3adddi): Likewise.
16021 (mulsidi3_v6): Likewise.
16022 (umulsidi3_v6): Likewise.
16023 (umulsidi3adddi_v6): Likewise.
16024 (smulsi3_highpart_v6): Likewise.
16025 (umulsi3_highpart_v6): Likewise.
16026 (mulhisi3tb): Likewise.
16027 (mulhisi3bt): Likewise.
16028 (mulhisi3tt): Likewise.
16029 (maddhisi4): Likewise.
16030 (maddhisi4tb): Likewise.
16031 (maddhisi4tt): Likewise.
16032 (maddhidi4): Likewise.
16033 (maddhidi4tb): Likewise.
16034 (maddhidi4tt): Likewise.
16035 (zeroextractsi_compare0_scratch): Likewise.
16036 (insv_zero): Likewise.
16037 (insv_t2): Likewise.
16038 (anddi_notzesidi_di): Likewise.
16039 (anddi_notsesidi_di): Likewise.
16040 (andsi_notsi_si): Likewise.
16041 (iordi_zesidi_di): Likewise.
16042 (xordi_zesidi_di): Likewise.
16043 (andsi_iorsi3_notsi): Likewise.
16044 (smax_0): Likewise.
16045 (smax_m1): Likewise.
16046 (smin_0): Likewise.
16047 (not_shiftsi): Likewise.
16048 (unaligned_loadsi): Likewise.
16049 (unaligned_loadhis): Likewise.
16050 (unaligned_loadhiu): Likewise.
16051 (unaligned_storesi): Likewise.
16052 (unaligned_storehi): Likewise.
16053 (extv_reg): Likewise.
16054 (extzv_t2): Likewise.
16055 (divsi3): Likewise.
16056 (udivsi3): Likewise.
16057 (arm_zero_extendhisi2addsi): Likewise.
16058 (arm_zero_extendqisi2addsi): Likewise.
16059 (compareqi_eq0): Likewise.
16060 (arm_extendhisi2_v6): Likewise.
16061 (arm_extendqisi2addsi): Likewise.
16062 (arm_movt): Likewise.
16063 (thumb2_ldrd): Likewise.
16064 (thumb2_ldrd_base): Likewise.
16065 (thumb2_ldrd_base_neg): Likewise.
16066 (thumb2_strd): Likewise.
16067 (thumb2_strd_base): Likewise.
16068 (thumb2_strd_base_neg): Likewise.
16069 (arm_negsi2): Add alternative for 16-bit encoding.
16070 (arm_one_cmplsi2): Likewise.
16071
16072 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
16073
16074 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
16075 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
16076 (movdfcc): Likewise.
16077 * config/arm/vfp.md (*thumb2_movsf_vfp):
16078 Disable predication for arm_restrict_it.
16079 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
16080 (*thumb2_movdfcc_vfp): Likewise.
16081 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
16082 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
16083 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
16084 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
16085 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
16086 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
16087 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
16088 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
16089 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
16090 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
16091 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
16092 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
16093 Disable predication for arm_restrict_it.
16094
16095 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
16096
16097 * config/i386/bmiintrin.h (_bextr_u32): New.
16098 (_bextr_u64): Ditto.
16099
16100 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
16101
16102 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
16103 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
16104 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
16105 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
16106 * config/mips/n32-elf.h: ...this new file.
16107
16108 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
16109
16110 PR target/57224
16111 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
16112 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
16113
16114 2013-06-27 Catherine Moore <clm@codesourcery.com>
16115
16116 * config/mips/mips-tables.opt: Regenerate.
16117 * config/mips/mips-cpus.def: Add m14ke and m14kec.
16118 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
16119 * doc/invoke.texi: Add -m14kc.
16120
16121 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16122
16123 PR target/57623
16124 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
16125 constraints of operand 1 and 2.
16126
16127 PR target/57623
16128 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
16129 to match RTL canonicalization. Swap predicates and
16130 constraints of operand 1 and 2.
16131
16132 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
16133
16134 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
16135 Process OP_INOUT regs for splitting too.
16136
16137 2013-06-27 Jakub Jelinek <jakub@redhat.com>
16138
16139 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
16140 decl before the loop, initialize to NULL.
16141 (vectorizable_load): Initialize ptr_incr to NULL.
16142
16143 2013-06-27 Martin Jambor <mjambor@suse.cz>
16144
16145 PR lto/57208
16146 * ipa-ref.h (ipa_maybe_record_reference): Declare.
16147 * ipa-ref.c (ipa_maybe_record_reference): New function.
16148 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
16149 * ipa-cp.c (create_specialized_node): Record potential references from
16150 aggvals.
16151 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
16152
16153 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16154
16155 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
16156 parameter 'mode' of type 'enum machine_mode mode'; change to pass
16157 'mode' to force_reg.
16158 (aarch64_add_offset): Update calls to aarch64_force_temporary.
16159 (aarch64_expand_mov_immediate): Likewise.
16160
16161 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
16162
16163 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
16164 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
16165
16166 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16167
16168 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
16169 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16170 (struct machine_function): Add tbegin_p.
16171 (s390_canonicalize_comparison): Fold CC mode compares to
16172 conditional jump if possible.
16173 (s390_emit_jump): Return the emitted jump.
16174 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
16175 Handle CCRAWmode compares.
16176 (s390_option_override): Default to -mhtm if available.
16177 (s390_reg_clobbered_rtx): Handle floating point regs as well.
16178 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
16179 FPRs instead of df_regs_ever_live_p.
16180 (s390_optimize_nonescaping_tx): New function.
16181 (s390_init_frame_layout): Extend clobbered_regs array to cover
16182 FPRs as well.
16183 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
16184 (s390_expand_tbegin): New function.
16185 (enum s390_builtin): New enum definition.
16186 (code_for_builtin): New array definition.
16187 (s390_init_builtins): New function.
16188 (s390_expand_builtin): New function.
16189 (TARGET_INIT_BUILTINS): Define.
16190 (TARGET_EXPAND_BUILTIN): Define.
16191 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
16192 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
16193 (s390_alc_comparison): Likewise.
16194 * config/s390/s390-modes.def: Add CCRAWmode.
16195 * config/s390/s390.h (processor_flags): Add PF_TX.
16196 (TARGET_CPU_HTM): Define macro.
16197 (TARGET_HTM): Define macro.
16198 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
16199 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
16200 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
16201 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
16202 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
16203 (TBEGIN_MASK, TBEGINC_MASK): New constants.
16204 ("*cc_to_int"): Move up.
16205 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
16206 constants other than 0.
16207 ("*ccraw_to_int"): New insn and splitter definition.
16208 ("tbegin", "tbegin_nofloat", "tbegin_retry")
16209 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
16210 ("tx_assist"): New expander.
16211 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
16212 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
16213 * config/s390/s390.opt: Add -mhtm option.
16214 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
16215 * config/s390/htmxlintrin.h: New file.
16216 * config/s390/htmintrin.h: New file.
16217 * config/s390/s390intrin.h: New file.
16218 * doc/extend.texi: Document htm builtins.
16219 * config.gcc: Add the new header files to extra_headers.
16220
16221 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
16222
16223 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
16224 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
16225
16226 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
16227 Pat Haugen <pthaugen@us.ibm.com>
16228 Peter Bergner <bergner@vnet.ibm.com>
16229
16230 * config/rs6000/power8.md: New.
16231 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
16232 setting for power8 entry.
16233 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
16234 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
16235 test for Power4/Power5 only.
16236 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
16237 support.
16238 (force_new_group): Adjust comment.
16239 * config/rs6000/rs6000.md: Include power8.md.
16240
16241 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
16242
16243 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
16244 * config/arm/arm-protos.h (arm_max_conditional_execute): New
16245 declaration.
16246 (tune_params): Update comment.
16247 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
16248 (arm_max_conditional_execute): New function.
16249 (thumb2_final_prescan_insn): Use max_insn_skipped and
16250 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
16251
16252 2013-06-25 Jakub Jelinek <jakub@redhat.com>
16253
16254 PR tree-optimization/57705
16255 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
16256 SSA_NAME step, provided that it is not defined inside the loop.
16257 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
16258 (get_initial_def_for_induction): Handle SSA_NAME IV step.
16259
16260 2013-06-25 Martin Jambor <mjambor@suse.cz>
16261
16262 PR middle-end/57670
16263 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
16264 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
16265 calls in the dump.
16266 (ipa_note_param_call): Initialize member_ptr flag.
16267 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
16268 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
16269 (ipa_write_indirect_edge_info): Stream member_ptr flag.
16270 (ipa_read_indirect_edge_info): Likewise.
16271
16272 2013-06-25 Richard Biener <rguenther@suse.de>
16273
16274 PR middle-end/56977
16275 * passes.c (init_optimization_passes): Move pass_fold_builtins
16276 and pass_dce earlier with -Og.
16277
16278 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
16279
16280 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
16281 <BIT_FIELD_REF>: Remove trailing TAB.
16282 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
16283 remove blank line.
16284
16285 2013-06-24 Martin Jambor <mjambor@suse.cz>
16286
16287 PR tree-optimization/57358
16288 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
16289 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
16290 (ipa_analyze_params_uses): Generate pessimistic info when true.
16291
16292 2013-06-24 Martin Jambor <mjambor@suse.cz>
16293
16294 PR tree-optimization/57539
16295 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
16296 global.inlined_to of the new node to it. All callers changed.
16297 * ipa-inline-transform.c (clone_inlined_nodes): New variable
16298 inlining_into, pass it to cgraph_clone_node.
16299 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
16300 ipa_free_edge_args_substructures.
16301 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
16302 rdesc linked list. Do not assert rdesc edges have inlined caller.
16303 Assert we have found an rdesc in the rdesc list.
16304
16305 2013-06-24 Richard Biener <rguenther@suse.de>
16306
16307 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
16308 (pointer_set_lookup): Declare.
16309 (class pointer_map): New template class implementing a
16310 generic pointer to T map.
16311 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
16312 pointer_map<T>::contains, pointer_map<T>::insert,
16313 pointer_map<T>::traverse): New functions.
16314 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
16315 (pointer_set_lookup): New function.
16316 (pointer_set_contains): Use pointer_set_lookup.
16317 (pointer_set_insert): Likewise.
16318 (insert_aux): Remove.
16319 (struct pointer_map_t): Embed a pointer_set_t.
16320 (pointer_map_create): Adjust.
16321 (pointer_map_destroy): Likewise.
16322 (pointer_map_contains): Likewise.
16323 (pointer_map_insert): Likewise.
16324 (pointer_map_traverse): Likewise.
16325 * tree-streamer.h (struct streamer_tree_cache_d): Use a
16326 pointer_map<unsigned> instead of a pointer_map_t.
16327 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
16328 (streamer_tree_cache_lookup): Likewise.
16329 (streamer_tree_cache_create): Likewise.
16330 (streamer_tree_cache_delete): Likewise.
16331 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
16332 pointer_map<unsigned> instead of a pointer_map_t.
16333 (lto_init_tree_ref_encoder): Adjust.
16334 (lto_destroy_tree_ref_encoder): Likewise.
16335 * lto-section-out.c (lto_output_decl_index): Likewise.
16336 (lto_record_function_out_decl_state): Likewise.
16337 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
16338
16339 2013-06-24 Richard Biener <rguenther@suse.de>
16340
16341 PR tree-optimization/57488
16342 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
16343
16344 2013-06-24 Alan Modra <amodra@gmail.com>
16345
16346 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
16347 (gen_easy_altivec_constant): Likewise.
16348 * config/rs6000/predicates.md (easy_vector_constant_add_self,
16349 easy_vector_constant_msb): Likewise.
16350
16351 2013-06-23 Jakub Jelinek <jakub@redhat.com>
16352
16353 PR target/57688
16354 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
16355 add missing return true.
16356
16357 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
16358
16359 PR target/52483
16360 * config/sh/predicates.md (general_extend_operand): Invoke
16361 general_movsrc_operand for memory operands.
16362 (general_movsrc_operand): Allow reg+reg addressing, do not use
16363 general_operand for memory operands.
16364
16365 2013-06-23 Sriraman Tallam <tmsriram@google.com>
16366
16367 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
16368 when current target options does not apply.
16369 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
16370 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
16371 * config/i386/bmiintrin.h: Pass appropriate target
16372 attributes to header.
16373 * config/i386/mmintrin.h: Ditto.
16374 * config/i386/nmmintrin.h: Ditto.
16375 * config/i386/avx2intrin.h: Ditto.
16376 * config/i386/fxsrintrin.h: Ditto.
16377 * config/i386/tbmintrin.h: Ditto.
16378 * config/i386/xsaveintrin.h: Ditto.
16379 * config/i386/f16cintrin.h: Ditto.
16380 * config/i386/xtestintrin.h: Ditto.
16381 * config/i386/xsaveoptintrin.h: Ditto.
16382 * config/i386/bmi2intrin.h: Ditto.
16383 * config/i386/lzcntintrin.h: Ditto.
16384 * config/i386/smmintrin.h: Ditto.
16385 * config/i386/wmmintrin.h: Ditto.
16386 * config/i386/x86intrin.h: Remove all header include guards.
16387 * config/i386/prfchwintrin.h: Ditto.
16388 * config/i386/pmmintrin.h: Ditto.
16389 * config/i386/tmmintrin.h: Ditto.
16390 * config/i386/xmmintrin.h: Ditto.
16391 * config/i386/popcntintrin.h: Ditto.
16392 * config/i386/rdseedintrin.h: Ditto.
16393 * config/i386/ammintrin.h: Ditto.
16394 * config/i386/emmintrin.h: Ditto.
16395 * config/i386/immintrin.h: Remove all header include guards.
16396 * config/i386/fma4intrin.h: Ditto.
16397 * config/i386/lwpintrin.h: Ditto.
16398 * config/i386/xopintrin.h: Ditto.
16399 * config/i386/ia32intrin.h: Ditto.
16400 * config/i386/avxintrin.h: Ditto.
16401 * config/i386/rtmintrin.h: Ditto.
16402 * config/i386/fmaintrin.h: Ditto.
16403 * config/i386/mm3dnow.h: Ditto.
16404
16405 2013-06-22 Sriraman Tallam <tmsriram@google.com>
16406
16407 * common/config/i386/i386-common.c: Handle LZCNT.
16408
16409 2013-06-22 Andi Kleen <ak@linux.intel.com>
16410
16411 * doc/extend.texi: Use __atomic_store_n instead of
16412 __atomic_store in HLE example.
16413
16414 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
16415
16416 * config/sh/sh.c: Remove <cstdlib> workaround.
16417
16418 2013-06-21 Andi Kleen <ak@linux.intel.com>
16419
16420 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
16421
16422 2013-06-21 Andi Kleen <ak@linux.intel.com>
16423
16424 * doc/extend.texi: Document that __atomic_clear and
16425 __atomic_test_and_set should only be used with bool.
16426
16427 2013-06-20 Jan Hubicka <jh@suse.cz>
16428
16429 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
16430 types_same_for_odr.
16431 * tree.c (decls_same_for_odr): New function.
16432 (same_for_edr): New function.
16433 (types_same_for_odr): New function.
16434 (get_binfo_at_offset): Use it.
16435 * tree.h (types_same_for_odr): Declare.
16436
16437 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
16438 Jason Merrill <jason@redhat.com>
16439
16440 * system.h: Include <cstdlib> as well as <stdlib.h>.
16441
16442 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
16443
16444 PR target/57655
16445 * config/i386/i386.c (construct_container): Report error if
16446 long double is used with disabled x87 float returns.
16447
16448 2013-06-20 Jan Hubicka <jh@suse.cz>
16449
16450 * lto-cgraph.c (input_symtab): Do not set cgraph state.
16451
16452 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
16453
16454 PR rtl-optimization/57425
16455 PR rtl-optimization/57569
16456 * alias.c (write_dependence_p): Remove parameters mem_mode and
16457 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
16458 Changed all callers.
16459 (canon_anti_dependence): Get comments and semantics in sync.
16460 Add parameter mem_canonicalized. Changed all callers.
16461 * rtl.h (canon_anti_dependence): Update prototype.
16462
16463 2013-06-20 Richard Biener <rguenther@suse.de>
16464
16465 * data-streamer-in.c (streamer_read_uhwi): Optimize single
16466 byte case, inline streamer_read_uchar and defer section
16467 overrun check.
16468
16469 2013-06-20 Richard Biener <rguenther@suse.de>
16470
16471 PR tree-optimization/57584
16472 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
16473 SSA names into the expanded expression that take part in
16474 abnormal coalescing.
16475
16476 2013-06-19 Sharad Singhai <singhai@google.com>
16477
16478 * gcov.c (print_usage): Handle new option.
16479 (process_args): Ditto.
16480 (get_gcov_intermediate_filename): New function.
16481 (output_intermediate_file): New function.
16482 (output_gcov_file): New function
16483 (generate_results): Handle new option.
16484 (release_function): Relase demangled name.
16485 (read_graph_file): Handle demangled name.
16486 (output_lines): Ditto.
16487 * doc/gcov.texi: Document gcov intermediate format.
16488
16489 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
16490
16491 PR bootstrap/57604
16492 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
16493 (lra_emit_add): Use the functions. Add comment about Y as an
16494 address segment.
16495
16496 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
16497
16498 PR driver/57652
16499 * collect2.c (collect_atexit): New.
16500 (collect_exit): Delete.
16501 (main): Register collect_atexit with atexit.
16502 (collect_wait): Change collect_exit to exit.
16503 (do_wait): Same.
16504 * collect2.h (collect_exit): Delete.
16505 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
16506
16507 2013-06-19 Wei Mi <wmi@google.com>
16508
16509 PR rtl-optimization/57518
16510 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
16511 if regno is used in paradoxical subreg.
16512 (update_equiv_regs): Check pdx_subregs[regno] before
16513 set a reg to be equivalent with a mem.
16514
16515 2013-06-19 Matthias Klose <doko@ubuntu.com>
16516
16517 PR driver/57651
16518 * file-find.h (find_a_file): Add a mode parameter.
16519 * file-find.c (find_a_file): Likewise.
16520 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
16521 with X_OK for the executables.
16522 * collect2.c (main): Call find_a_file with X_OK.
16523
16524 2013-06-19 Steve Ellcey <sellcey@mips.com>
16525
16526 PR target/56942
16527 * config/mips/mips.md (casesi_internal_mips16_<mode>):
16528 Use NEXT_INSN instead of next_real_insn.
16529
16530 2013-06-19 Jan Hubicka <jh@suse.cz>
16531
16532 * cgraph.h (const_value_known_p): Replace by ...
16533 (ctor_for_folding): .. this one.
16534 * cgraphunit.c (process_function_and_variable_attributes): Use it.
16535 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
16536 * expr.c (expand_expr_real_1): Likewise.
16537 (string_constant): Likewise.
16538 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
16539 * ipa.c (process_references): Likewise.
16540 (symtab_remove_unreachable_nodes): Likewise.
16541 * ipa-inline-analysis.c (param_change_prob): Likewise.
16542 * gimple-fold.c (canonicalize_constructor_val): Likewise.
16543 (get_base_constructor): Likwise.
16544 * varpool.c (varpool_remove_node): Likewise.
16545 (varpool_remove_initializer): LIkewise.
16546 (dump_varpool_node): LIkwise.
16547 (const_value_known_p): Rewrite to ...
16548 (ctor_for_folding): ... this one.
16549
16550 2013-06-19 Jakub Jelinek <jakub@redhat.com>
16551
16552 PR driver/57651
16553 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
16554 PERSONALITY in $PATH derived prefixes.
16555
16556 2013-06-19 Jeff Law <law@redhat.com>
16557
16558 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
16559 in comment.
16560
16561 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
16562 (simplify_bitwise_binary): Use it to simpify certain binary ops on
16563 booleans.
16564
16565 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
16566
16567 * config/arm/vfp.md: Move VFP instruction classification documentation
16568 to ...
16569 * config/arm/arm.md: ... here. Update instruction classification
16570 documentation.
16571
16572 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
16573
16574 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
16575 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
16576 pattern. Use more efficient sequences on ARMv5 and Thumb2.
16577
16578 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
16579
16580 PR target/57609
16581 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16582 with NEXT_INSN. Use tablejump_p to check for jump table data
16583 insns.
16584
16585 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
16586
16587 PR c++/56544
16588 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
16589 that now in C++ the value is correct per the C++ standards.
16590
16591 2013-06-19 Richard Biener <rguenther@suse.de>
16592
16593 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
16594 for global context.
16595
16596 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16597
16598 Revert:
16599 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16600
16601 PR target/57609
16602 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16603 with next_active_insn.
16604
16605 2013-06-18 Sriraman Tallam <tmsriram@google.com>
16606
16607 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
16608 functions are inlined during failures to flag an error.
16609 * tree-inline.c (expand_call_inline): Allow the error to be flagged
16610 in early inline pass.
16611
16612 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
16613
16614 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
16615 in comments.
16616
16617 2013-06-18 Julian Brown <julian@codesourcery.com>
16618
16619 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
16620 Permit virtual register pre-reload if !strict.
16621 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
16622 change.
16623 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
16624 prototype.
16625 * config/arm/neon.md (movmisalign<mode>): Use
16626 neon_perm_struct_or_reg_operand instead of
16627 neon_struct_or_register_operand.
16628 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
16629 neon_permissive_struct_operand instead of neon_struct_operand.
16630 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
16631 neon_vector_mem_operand.
16632 * config/arm/predicates.md (neon_struct_operand): Adjust call to
16633 neon_vector_mem_operand.
16634 (neon_permissive_struct_operand): New.
16635 (neon_struct_or_register_operand): Rename to...
16636 (neon_perm_struct_or_reg_operand): This. Adjust call to
16637 neon_vector_mem_operand.
16638
16639 2013-06-18 Richard Biener <rguenther@suse.de>
16640
16641 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
16642 * lto-streamer.h: Include pointer-set.h.
16643 (struct lto_decl_slot): Remove.
16644 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
16645 Remove next_index entry.
16646 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16647 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16648 (lto_init_tree_ref_encoder): Adjust.
16649 (lto_destroy_tree_ref_encoder): Likewise.
16650 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
16651 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
16652 (lto_output_decl_index): Adjust.
16653 (lto_new_out_decl_state): Likewise.
16654 (lto_record_function_out_decl_state): Likewise.
16655 * lto-streamer-out.c (copy_function): Likewise.
16656
16657 2013-06-18 Richard Biener <rguenther@suse.de>
16658
16659 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
16660 * cgraphunit.c: Include cfgloop.h.
16661 (init_lowered_empty_function): Initialize the loop tree.
16662 (assemble_thunk): Insert new BBs into loops.
16663
16664 2013-06-18 Richard Biener <rguenther@suse.de>
16665
16666 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
16667 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
16668 the map from cache entry to cache index optional.
16669 (streamer_tree_cache_replace_tree): Adjust accordingly.
16670 (streamer_tree_cache_append): Likewise.
16671 (streamer_tree_cache_delete): Likewise.
16672 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
16673 streamer cache map from cache entry to cache index.
16674 * lto-streamer-out.c (create_output_block): Adjust.
16675
16676 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
16677
16678 * config/arm/arm.md (attribute "insn"): Move multiplication and
16679 division attributes to...
16680 (attribute "type"): ... here. Remove mult.
16681 (attribute "mul32"): New attribute.
16682 (attribute "mul64"): Add umaal.
16683 (*arm_mulsi3): Update attributes.
16684 (*arm_mulsi3_v6): Likewise.
16685 (*thumb_mulsi3): Likewise.
16686 (*thumb_mulsi3_v6): Likewise.
16687 (*mulsi3_compare0): Likewise.
16688 (*mulsi3_compare0_v6): Likewise.
16689 (*mulsi_compare0_scratch): Likewise.
16690 (*mulsi_compare0_scratch_v6): Likewise.
16691 (*mulsi3addsi): Likewise.
16692 (*mulsi3addsi_v6): Likewise.
16693 (*mulsi3addsi_compare0): Likewise.
16694 (*mulsi3addsi_compare0_v6): Likewise.
16695 (*mulsi3addsi_compare0_scratch): Likewise.
16696 (*mulsi3addsi_compare0_scratch_v6): Likewise.
16697 (*mulsi3subsi): Likewise.
16698 (*mulsidi3adddi): Likewise.
16699 (*mulsi3addsi_v6): Likewise.
16700 (*mulsidi3adddi_v6): Likewise.
16701 (*mulsidi3_nov6): Likewise.
16702 (*mulsidi3_v6): Likewise.
16703 (*umulsidi3_nov6): Likewise.
16704 (*umulsidi3_v6): Likewise.
16705 (*umulsidi3adddi): Likewise.
16706 (*umulsidi3adddi_v6): Likewise.
16707 (*smulsi3_highpart_nov6): Likewise.
16708 (*smulsi3_highpart_v6): Likewise.
16709 (*umulsi3_highpart_nov6): Likewise.
16710 (*umulsi3_highpart_v6): Likewise.
16711 (mulhisi3): Likewise.
16712 (*mulhisi3tb): Likewise.
16713 (*mulhisi3bt): Likewise.
16714 (*mulhisi3tt): Likewise.
16715 (maddhisi4): Likewise.
16716 (*maddhisi4tb): Likewise.
16717 (*maddhisi4tt): Likewise.
16718 (maddhidi4): Likewise.
16719 (*maddhidi4tb): Likewise.
16720 (*maddhidi4tt): Likewise.
16721 (divsi3): Likewise.
16722 (udivsi3): Likewise.
16723 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
16724 (thumb2_mulsi_short_compare0): Likewise.
16725 (thumb2_mulsi_short_compare0_scratch): Likewise.
16726 * config/arm/arm1020e.md (1020mult1): Update attribute change.
16727 (1020mult2): Likewise.
16728 (1020mult3): Likewise.
16729 (1020mult4): Likewise.
16730 (1020mult5): Likewise.
16731 (1020mult6): Likewise.
16732 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
16733 change.
16734 (cortex_a15_mult64): Likewise.
16735 (cortex_a15_sdiv): Likewise.
16736 (cortex_a15_udiv): Likewise.
16737 * config/arm/arm1026ejs.md (mult1): Update attribute change.
16738 (mult2): Likewise.
16739 (mult3): Likewise.
16740 (mult4): Likewise.
16741 (mult5): Likewise.
16742 (mult6): Likewise.
16743 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
16744 (pj4_ir_div): Likewise.
16745 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
16746 (11_mult2): Likewise.
16747 (11_mult3): Likewise.
16748 (11_mult4): Likewise.
16749 (11_mult5): Likewise.
16750 (11_mult6): Likewise.
16751 (11_mult7): Likewise.
16752 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
16753 (cortex_a8_mla): Likewise.
16754 (cortex_a8_mull): Likewise.
16755 (cortex_a8_smulwy): Likewise.
16756 (cortex_a8_smlald): Likewise.
16757 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
16758 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
16759 (cortex_r4_mul_3): Likewise.
16760 (cortex_r4_mla_4): Likewise.
16761 (cortex_r4_mla_3): Likewise.
16762 (cortex_r4_smlald): Likewise.
16763 (cortex_r4_mull): Likewise.
16764 (cortex_r4_sdiv): Likewise.
16765 (cortex_r4_udiv): Likewise.
16766 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
16767 (cortex_a7_idiv): Likewise.
16768 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
16769 (9_mult2): Likewise.
16770 (9_mult3): Likewise.
16771 (9_mult4): Likewise.
16772 (9_mult5): Likewise.
16773 (9_mult6): Likewise.
16774 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
16775 (cortex_a53_sdiv): Likewise.
16776 (cortex_a53_udiv): Likewise.
16777 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
16778 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
16779 (mp626_mult2): Likewise.
16780 (mp626_mult3): Likewise.
16781 (mp626_mult4): Likewise.
16782 * config/arm/fa526.md (526_mult1): Update attribute change.
16783 (526_mult2): Likewise.
16784 * config/arm/arm-generic.md (mult): Update attribute change.
16785 (mult_ldsched_strongarm): Likewise.
16786 (mult_ldsched): Likewise.
16787 (multi_cycle): Likewise.
16788 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
16789 * config/arm/fa606te.md (606te_mult1): Update attribute change.
16790 (606te_mult2): Likewise.
16791 (606te_mult3): Likewise.
16792 (606te_mult4): Likewise.
16793 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
16794 (cortex_a9_mac16): Likewise.
16795 (cortex_a9_multiply): Likewise.
16796 (cortex_a9_mac): Likewise.
16797 (cortex_a9_multiply_long): Likewise.
16798 * config/arm/fa626te.md (626te_mult1): Update attribute change.
16799 (626te_mult2): Likewise.
16800 (626te_mult3): Likewise.
16801 (626te_mult4): Likewise.
16802
16803 2013-06-18 Richard Biener <rguenther@suse.de>
16804
16805 PR lto/57334
16806 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
16807
16808 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16809
16810 PR target/57609
16811 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
16812 with next_active_insn.
16813
16814 2013-06-18 Alan Modra <amodra@gmail.com>
16815
16816 * config/rs6000/rs6000.h (enum data_align): New.
16817 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
16818 (DATA_ABI_ALIGNMENT): Define.
16819 (CONSTANT_ALIGNMENT): Correct comment.
16820 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
16821 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
16822
16823 2013-06-17 David Malcolm <dmalcolm@redhat.com>
16824
16825 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
16826 ATTRIBUTE_UNUSED marking.
16827
16828 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
16829
16830 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
16831 alternative and update.
16832 (aarch64_dup_lanedi): Delete.
16833 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
16834 * config/aarch64/aarch64-simd-builtins.def: Update.
16835
16836 2013-06-17 Richard Biener <rguenther@suse.de>
16837
16838 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
16839 (lto_input_scc): Declare.
16840 (lto_input_tree_1): Likewise.
16841 (struct lto_stats_d): Add num_tree_bodies_output and
16842 num_pickle_refs_output.
16843 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
16844 (lto_read_tree_1): Split out from ...
16845 (lto_read_tree): ... this.
16846 (lto_input_scc): New function.
16847 (lto_input_tree_1): Split out from ...
16848 (lto_input_tree): ... this. Handle LTO_tree_scc.
16849 (lto_data_in_create): Create the streamer cache without hashes.
16850 * lto-streamer-out.c (create_output_block): Create the streamer
16851 cache with hashes when not doing WPA.
16852 (lto_write_tree_1): Split out from ...
16853 (lto_write_tree): ... this.
16854 (get_symbol_initial_value): New function.
16855 (lto_output_tree_1): Split out from ...
16856 (lto_output_tree): ... this. Write trees as series of SCCs
16857 using a DFS walk via DFS_write_tree.
16858 (struct sccs, struct scc_entry): New types.
16859 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
16860 (DFS_write_tree_body): New function.
16861 (DFS_write_tree): Likewise.
16862 (hash_tree): Likewise.
16863 (scc_entry_compare): Likewise.
16864 (hash_scc): Likewise.
16865 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
16866 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
16867 TREE_CHAIN as regular reference.
16868 (streamer_read_integer_cst): Remove.
16869 (streamer_get_pickled_tree): Adjust.
16870 * tree-streamer-out.c (streamer_write_chain): Disable streaming
16871 of DECL_EXTERNALs in BLOCK_VARS for now.
16872 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
16873 reference.
16874 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
16875 Add hash value argument and record that if hashes are recorded
16876 in the cache.
16877 (streamer_tree_cache_insert_1): Adjust.
16878 (streamer_tree_cache_insert): Likewise.
16879 (streamer_tree_cache_insert_at): Rename to ...
16880 (streamer_tree_cache_replace_tree): ... this and adjust.
16881 (streamer_tree_cache_append): Adjust.
16882 (record_common_node): Likewise.
16883 (streamer_tree_cache_create): Add argument whether to
16884 record hash values together with trees.
16885 (streamer_tree_cache_delete): Adjust.
16886 * tree-streamer.h (struct streamer_tree_cache_d): Add
16887 vector of hashes.
16888 (streamer_read_integer_cst): Remove.
16889 (streamer_tree_cache_insert): Adjust.
16890 (streamer_tree_cache_append): Likewise.
16891 (streamer_tree_cache_insert_at): Rename to ...
16892 (streamer_tree_cache_replace_tree): ... this and adjust.
16893 (streamer_tree_cache_create): Add argument whether to record hashes.
16894 (streamer_tree_cache_get): Rename to ...
16895 (streamer_tree_cache_get_tree): ... this.
16896 (streamer_tree_cache_get_hash): New function.
16897 * tree.c (cache_integer_cst): New function.
16898 * tree.h (cache_integer_cst): Declare.
16899 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
16900 * lto-symtab.c (lto_varpool_replace_node): Only release
16901 DECL_INITIAL of non-prevailing decls.
16902 * varpool.c (varpool_remove_initializer): Do not release
16903 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
16904
16905 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16906
16907 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
16908 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
16909 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
16910 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
16911 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
16912 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
16913 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
16914 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
16915 instead of TARGET_64BIT.
16916 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
16917 Require ISA_HAS_<D>DIV.
16918
16919 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
16920
16921 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
16922 (mips*-*-linux*): Move default with_llsc setting to where other
16923 defaults are set.
16924 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
16925 with_arch block.
16926 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
16927 Likewise. Remove default with_tune setting. Move default float
16928 setting to its own block. Handle with_llsc in the same block as above.
16929
16930 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
16931
16932 PR rtl-optimization/57425
16933 PR rtl-optimization/57569
16934 * alias.c (write_dependence_p): Add new parameters mem_mode,
16935 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
16936 Changed all callers.
16937 (canon_anti_dependence): New function.
16938 * cse.c (check_dependence): Use canon_anti_dependence.
16939 * cselib.c (cselib_invalidate_mem): Likewise.
16940 * rtl.h (canon_anti_dependence): Declare.
16941
16942 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
16943
16944 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
16945 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
16946 ".set mips3" for 64-bit targets.
16947
16948 2013-06-15 Dehao Chen <dehao@google.com>
16949
16950 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
16951 * gimple-low.c (gimple_check_call_matching_types): Likewise.
16952 (gimple_check_call_args): Likewise.
16953 * value-prof.c (check_ic_target): Likewise.
16954 * ipa-inline.c (early_inliner): Likewise.
16955 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
16956 * cgraph.c (cgraph_create_edge_1): Likewise.
16957 (cgraph_make_edge_direct): Likewise.
16958
16959 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
16960
16961 PR target/57615
16962 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
16963 rs6000_output_move_128bit to handle emitting quad memory
16964 operations. Set attribute length to 8 bytes.
16965
16966 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
16967
16968 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
16969 New pattern.
16970 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
16971 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
16972 (aarch64_<su>mlsl<mode>): Likewise.
16973
16974 2013-06-14 Mike Stump <mikestump@comcast.net>
16975
16976 * Makefile.in (TARGET_H): Add insn-codes.h.
16977
16978 2013-06-14 Alan Modra <amodra@gmail.com>
16979
16980 PR middle-end/57134
16981 PR middle-end/57586
16982 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
16983 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
16984 bitfield expansion when EXPAND_MEMORY.
16985 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
16986
16987 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
16988
16989 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
16990 test for clearing quad memory on 32-bit later.
16991
16992 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
16993
16994 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
16995 (fold_negate_expr): Likewise.
16996 (fold_real_zero_addition_p): Handle vectors.
16997 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
16998
16999 2013-06-14 Alan Modra <amodra@gmail.com>
17000
17001 * varasm.c (force_const_mem): Revert 2013-06-07 change.
17002
17003 2013-06-13 Jan Hubicka <jh@suse.cz>
17004
17005 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
17006 Local comdats are not externally visible.
17007 * symtab.c (dump_symtab_base): Dump externally visible.
17008 (verify_symtab_base): Verify back links in the symtab hash.
17009
17010 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17011
17012 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
17013 CONVERT_EXPR as equal nodes.
17014
17015 2013-06-13 Bin Cheng <bin.cheng@arm.com>
17016
17017 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
17018
17019 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
17020
17021 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
17022 Generalize to complex and vector.
17023 * tree.c (build_all_ones_cst): New function.
17024 * tree.h (build_all_ones_cst): Declare it.
17025
17026 2013-06-13 Alan Modra <amodra@gmail.com>
17027
17028 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
17029 * config/rs6000/rs6000.md (signbittf2): New insn.
17030 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
17031 (abstf2_internal, cmptf_internal2): Likewise.
17032 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
17033
17034 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
17035 Pat Haugen <pthaugen@us.ibm.com>
17036 Peter Bergner <bergner@vnet.ibm.com>
17037
17038 * config/rs6000/rs6000.c (emit_load_locked): Add support for
17039 power8 byte, half-word, and quad-word atomic instructions.
17040 (emit_store_conditional): Likewise.
17041 (rs6000_expand_atomic_compare_and_swap): Likewise.
17042 (rs6000_expand_atomic_op): Likewise.
17043
17044 * config/rs6000/sync.md (larx): Add new modes for power8.
17045 (stcx): Likewise.
17046 (AINT): New mode iterator to include TImode as well as normal
17047 integer modes on power8.
17048 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
17049 that VSX registers are not considered. Use AINT mode iterator
17050 instead of INT1 to allow inclusion of quad word atomic operations
17051 on power8.
17052 (load_locked<mode>): Likewise.
17053 (store_conditional<mode>): Likewise.
17054 (atomic_compare_and_swap<mode>): Likewise.
17055 (atomic_exchange<mode>): Likewise.
17056 (atomic_nand<mode>): Likewise.
17057 (atomic_fetch_<fetchop_name><mode>): Likewise.
17058 (atomic_nand_fetch<mode>): Likewise.
17059 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
17060 each type.
17061 (ATOMIC): On power8, add QImode, HImode modes.
17062 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
17063 modes that promote to SImode.
17064 (load_lockedti): Convert TImode arguments to PTImode, so that we
17065 get a guaranteed even/odd register pair.
17066 (load_lockedpti): Likewise.
17067 (store_conditionalti): Likewise.
17068 (store_conditionalpti): Likewise.
17069
17070 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
17071 atomic load/store instructions.
17072 (HSI): Likewise.
17073
17074 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
17075
17076 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
17077 loads.
17078 (insn_count): New attribute, with most cases extracted from...
17079 (length): ...here. Redefine most cases in terms of insn_count.
17080 (single_insn): Delete.
17081 (can_delay): Use insn_count to check for single instructions.
17082 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
17083 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
17084 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
17085 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
17086 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
17087 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
17088 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
17089 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
17090 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
17091 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
17092 rather than "length".
17093 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
17094 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
17095 Use "insn_count" rather than "length".
17096 * config/mips/mips-dsp.md
17097 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
17098 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
17099 length attributes.
17100
17101 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
17102
17103 PR tree-optimization/57361
17104 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
17105
17106 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
17107
17108 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
17109 to split.
17110 (aarch64_simd_combine<mode>): New instruction expansion.
17111 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
17112 function prototype.
17113 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
17114 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
17115
17116 2013-06-12 Jan Hubicka <jh@suse.cz>
17117
17118 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
17119 decl has when in streaming stage.
17120 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
17121 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
17122
17123 2013-06-12 Roland Stigge <stigge@antcom.de>
17124
17125 PR target/57578
17126 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
17127
17128 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17129
17130 PR tree-optimization/57537
17131 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
17132 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
17133
17134 2013-06-12 Richard Biener <rguenther@suse.de>
17135
17136 * data-streamer.h (streamer_write_char_stream): CSE
17137 obs->current_pointer.
17138 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
17139 streamer_write_char_stream manually and optimize the resulting loop.
17140 (streamer_write_hwi_stream): Likewise.
17141
17142 2013-06-12 Jan Hubicka <jh@suse.cz>
17143
17144 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
17145 * cgraph.h (varpool_create_empty_node): Declare.
17146 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
17147 duplicated nodes.
17148 * symtab.c (symtab_unregister_node): Be lax about missin entries
17149 in node hash.
17150 (symtab_get_node): Update comment.
17151 * varpool.c (varpool_create_empty_node): Break out from ...
17152 (varpool_node_for_decl): ... here.
17153 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
17154
17155 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
17156
17157 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
17158 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
17159 part. Use straight-line flow at the end.
17160 <COMPONENT_REF>: Remove superfluous else.
17161 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
17162
17163 2013-06-12 Jakub Jelinek <jakub@redhat.com>
17164
17165 PR target/56564
17166 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
17167 target hook even for !TREE_PUBLIC decls. If no resolution info
17168 is available, return false for common and external decls.
17169
17170 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
17171
17172 * config/rl78/constraints.md (U): New constraint.
17173 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
17174 valloc attribute.
17175
17176 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
17177
17178 PR target/57589
17179 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
17180 to allow returning address to AT_PLATFORM name.
17181
17182 2013-06-11 Jan Hubicka <jh@suse.cz>
17183
17184 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
17185 * cgraph.h (symtab_node_base): Add weakref flag.
17186 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
17187 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
17188 (output_weakrefs): Use weakref flag.
17189 * fold-const.c (simple_operand_p): Handle WEAK.
17190 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
17191 * ipa.c (varpool_externally_visible_p): Drop weakref.
17192 (function_and_variable_visibility): Update comment; fix weakref
17193 sanity checks; do not clear DECL_WEAK on them.
17194 * lto-cgraph.c (lto_output_node): update.
17195 (lto_output_varpool_node): Update.
17196 (input_overwrite_node): Update.
17197 (input_node): Update.
17198 (input_varpool_node): Update.
17199 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
17200 (lto_symtab_merge_symbols): Add sanity check.
17201 (lto_symtab_prevailing_decl): Do not special case weakrefs.
17202 * passes.c (rest_of_decl_compilation): Set static flag, too.
17203 * symtab.c (dump_symtab_base): Dump weakref.
17204 (verify_symtab_base): Sanity check weakrefs.
17205 (symtab_make_decl_local): Remove duplicated code.
17206 (symtab_alias_ultimate_target): Simplify.
17207 * varpool.c (varpool_create_variable_alias): Set weakref flag.
17208
17209 2013-06-11 Tom de Vries <tom@codesourcery.com>
17210
17211 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
17212 sequence_vect == NULL.
17213
17214 2013-06-11 DJ Delorie <dj@redhat.com>
17215
17216 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
17217 (rl78_unwind_word_mode): New.
17218
17219 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17220
17221 * final.c (debug_prefix_maps): Make static.
17222
17223 2013-06-11 David Malcolm <dmalcolm@redhat.com>
17224
17225 * function.c (initial_trampoline): Remove stray copy.
17226
17227 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
17228
17229 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
17230
17231 2013-06-11 Martin Jambor <mjambor@suse.cz>
17232
17233 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
17234 within bounds at the beginning of the function.
17235
17236 2013-06-11 Alan Modra <amodra@gmail.com>
17237
17238 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
17239 reporting.
17240 (get_named_section): Don't NULL !DECL_P decl.
17241
17242 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
17243
17244 * doc/invoke.texi (core-avx2): Document.
17245 (slm): Likewise.
17246 (atom): Updated with MOVBE.
17247
17248 2013-06-11 Richard Biener <rguenther@suse.de>
17249
17250 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
17251
17252 2013-06-11 Anton Blanchard <anton@samba.org>
17253
17254 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
17255 correct shift value in little-endian mode.
17256
17257 2013-06-11 Jakub Jelinek <jakub@redhat.com>
17258
17259 PR target/56564
17260 * varasm.c (get_variable_align): Move #endif to the right place.
17261
17262 2013-06-10 Cary Coutant <ccoutant@google.com>
17263
17264 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
17265 for hash so that hash table traversal order is deterministic.
17266
17267 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
17268 Pat Haugen <pthaugen@us.ibm.com>
17269 Peter Bergner <bergner@vnet.ibm.com>
17270
17271 * config/rs6000/vector.md (GPR move splitter): Do not split moves
17272 of vectors in GPRS if they are direct moves or quad word load or
17273 store moves.
17274
17275 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
17276 declaration.
17277 (direct_move_p): Likewise.
17278 (quad_load_store_p): Likewise.
17279
17280 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
17281 classes into bins based on the physical register type.
17282 (reg_class_to_reg_type): Likewise.
17283 (IS_STD_REG_TYPE): Likewise.
17284 (IS_FP_VECT_REG_TYPE): Likewise.
17285 (reload_fpr_gpr): Arrays to determine what insn to use if we can
17286 use direct move instructions.
17287 (reload_gpr_vsx): Likewise.
17288 (reload_vsx_gpr): Likewise.
17289 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
17290 information that is a simplification of register classes. Also
17291 precalculate direct move reload helpers.
17292 (direct_move_p): New function to return true if the operation can
17293 be done as a direct move instruciton.
17294 (quad_load_store_p): New function to return true if the operation
17295 is a quad memory operation.
17296 (rs6000_legitimize_address): If quad memory, only allow register
17297 indirect for TImode addresses.
17298 (rs6000_legitimate_address_p): Likewise.
17299 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
17300 (rs6000_reload_register_type): Likewise.
17301 (register_to_reg_type): Return register type.
17302 (rs6000_secondary_reload_simple_move): New helper function for
17303 secondary reload and secondary memory needed to identify anything
17304 that is a simple move, and does not need reloading.
17305 (rs6000_secondary_reload_direct_move): New helper function for
17306 secondary reload to identify cases that can be done with several
17307 instructions via the direct move instructions.
17308 (rs6000_secondary_reload_move): New helper function for secondary
17309 reload to identify moves between register types that can be done.
17310 (rs6000_secondary_reload): Add support for quad memory operations
17311 and for direct move.
17312 (rs6000_secondary_memory_needed): Likewise.
17313 (rs6000_debug_secondary_memory_needed): Change argument names.
17314 (rs6000_output_move_128bit): New function to return the move to
17315 use for 128-bit moves, including knowing about the various
17316 limitations of quad memory operations.
17317
17318 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
17319 memory operations. call rs6000_output_move_128bit for the actual
17320 instruciton(s) to generate.
17321 (vsx_movti_64bit): Likewise.
17322
17323 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
17324 (UNSPEC_P8V_MTVSRWZ): Likewise.
17325 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
17326 (UNSPEC_P8V_MTVSRD): Likewise.
17327 (UNSPEC_P8V_XXPERMDI): Likewise.
17328 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
17329 (UNSPEC_FUSION_GPR): Likewise.
17330 (FMOVE128_GPR): New iterator for direct move.
17331 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
17332 (f32_sv): Likewise.
17333 (f32_dm): Likewise.
17334 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
17335 loads and direct move instructions.
17336 (zero_extendsidi2_lfiwzx): Likewise.
17337 (extendsidi2_lfiwax): Likewise.
17338 (extendsidi2_nocell): Likewise.
17339 (floatsi<mode>2_lfiwax): Likewise.
17340 (lfiwax): Likewise.
17341 (floatunssi<mode>2_lfiwzx): Likewise.
17342 (lfiwzx): Likewise.
17343 (fix_trunc<mode>_stfiwx): Likewise.
17344 (fixuns_trunc<mode>_stfiwx): Likewise.
17345 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
17346 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
17347 (parity<mode>2_cmpb): Set length/type attr.
17348 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
17349 for 'mr.' to fast_compare.
17350 (bpermd_<mode>): Change type attr to popcnt.
17351 (p8_fmrgow_<mode>): New insns for power8 direct move support.
17352 (p8_mtvsrwz_1): Likewise.
17353 (p8_mtvsrwz_2): Likewise.
17354 (reload_fpr_from_gpr<mode>): Likewise.
17355 (p8_mtvsrd_1): Likewise.
17356 (p8_mtvsrd_2): Likewise.
17357 (p8_xxpermdi_<mode>): Likewise.
17358 (reload_vsx_from_gpr<mode>): Likewise.
17359 (reload_vsx_from_gprsf): Likewise.
17360 (p8_mfvsrd_3_<mode>): LIkewise.
17361 (reload_gpr_from_vsx<mode>): Likewise.
17362 (reload_gpr_from_vsxsf): Likewise.
17363 (p8_mfvsrd_4_disf): Likewise.
17364 (multi-word GPR splits): Do not split direct moves or quad memory
17365 operations.
17366
17367 2013-06-10 David Malcolm <dmalcolm@redhat.com>
17368
17369 * tree-into-ssa.c (interesting_blocks): Make static.
17370
17371 2013-06-10 Jakub Jelinek <jakub@redhat.com>
17372
17373 PR target/56564
17374 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
17375 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
17376 Use DATA_ABI_ALIGNMENT for that case instead if defined.
17377 (get_variable_align): New function.
17378 (get_variable_section, emit_bss, emit_common,
17379 assemble_variable_contents, place_block_symbol): Use
17380 get_variable_align instead of DECL_ALIGN.
17381 (assemble_noswitch_variable): Add align argument, use it
17382 instead of DECL_ALIGN.
17383 (assemble_variable): Adjust caller. Use get_variable_align
17384 instead of DECL_ALIGN.
17385 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
17386 caller.
17387 (DATA_ABI_ALIGNMENT): Define.
17388 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
17389 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
17390 opt is false, only return the psABI mandated alignment increase.
17391 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
17392 (DATA_ABI_ALIGNMENT): ... this.
17393 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
17394 (DATA_ABI_ALIGNMENT): ... this.
17395 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
17396 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
17397 (DATA_ABI_ALIGNMENT): ... this.
17398 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
17399 * doc/tm.texi: Regenerated.
17400
17401 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
17402
17403 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
17404 cmp_code to construct REG_EQUAL note.
17405
17406 2013-06-09 Jakub Jelinek <jakub@redhat.com>
17407
17408 PR target/57568
17409 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
17410 that operands[2] doesn't overlap with operands[0].
17411
17412 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
17413 Jan Hubicka <jh@suse.cz>
17414
17415 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
17416 hack to mark symbols as used.
17417
17418 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
17419
17420 PR rtl-optimization/57559
17421 * lra-constraints.c (process_alt_operands): Don't discourage
17422 memory with known offset for offsetable memory constraint.
17423 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
17424
17425 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
17426
17427 * varasm.c (struct oc_local_state): Reorder fields.
17428 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
17429 and adjust accordingly.
17430 (output_constructor): Reorder initialization code and adjust call to
17431 output_constructor_bitfield.
17432
17433 2013-06-07 Jan Hubicka <jh@suse.cz>
17434
17435 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
17436
17437 2013-06-07 David Malcolm <dmalcolm@redhat.com>
17438
17439 * tree-object-size.c (unknown): Make const.
17440
17441 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
17442
17443 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
17444 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
17445 for last alternative in the cpu_facility attribute.
17446
17447 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17448
17449 PR target/56315
17450 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
17451 (xordi3): Change operand 2 constraint to arm_xordi_operand.
17452 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
17453 * config/arm/constraints.md (Dg): New constraint.
17454 * config/arm/neon.md (xordi3_neon): Remove.
17455 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
17456 * config/arm/predicates.md (arm_xordi_operand): New predicate.
17457
17458 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17459
17460 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
17461 Clean up alternatives.
17462
17463 2013-06-07 Alan Modra <amodra@gmail.com>
17464
17465 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
17466 va_list_gpr_size.
17467
17468 2013-06-07 Alan Modra <amodra@gmail.com>
17469
17470 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
17471
17472 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17473
17474 * config/arm/constraints.md (Df): New constraint.
17475 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
17476 Correct length attribute for last two alternatives.
17477
17478 2013-06-07 Alan Modra <amodra@gmail.com>
17479
17480 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
17481 override user -mfp-in-toc.
17482 (offsettable_ok_by_alignment): Consider just the current access
17483 rather than the whole object, unless BLKmode. Handle
17484 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
17485 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
17486 for -mcmodel=medium.
17487 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
17488 override user -mfp-in-toc or -msum-in-toc. Default to
17489 -mno-fp-in-toc for -mcmodel=medium.
17490
17491 2013-06-06 DJ Delorie <dj@redhat.com>
17492
17493 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
17494 TARGET_VALID_POINTER_MODE.
17495
17496 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
17497 Pat Haugen <pthaugen@us.ibm.com>
17498 Peter Bergner <bergner@vnet.ibm.com>
17499
17500 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
17501 Document new power8 builtins.
17502
17503 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
17504 condition code register, to allow 128-bit logical operations to be
17505 done in the VSX or GPR registers.
17506 (nor<mode>3): Use the canonical form for nor.
17507 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
17508 vclz*, and vpopcnt* vector instructions.
17509 (nand<mode>3): Likewise.
17510 (orc<mode>3): Likewise.
17511 (clz<mode>2): LIkewise.
17512 (popcount<mode>2): Likewise.
17513
17514 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
17515 that only the GPRs are recognized.
17516
17517 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
17518 support for new power8 builtins.
17519
17520 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
17521 builtin functions.
17522 (xscvdpspn): Likewise.
17523 (vclz): Likewise.
17524 (vclzb): Likewise.
17525 (vclzh): Likewise.
17526 (vclzw): Likewise.
17527 (vclzd): Likewise.
17528 (vpopcnt): Likewise.
17529 (vpopcntb): Likewise.
17530 (vpopcnth): Likewise.
17531 (vpopcntw): Likewise.
17532 (vpopcntd): Likewise.
17533 (vgbbd): Likewise.
17534 (vmrgew): Likewise.
17535 (vmrgow): Likewise.
17536 (eqv): Likewise.
17537 (eqv_v16qi3): Likewise.
17538 (eqv_v8hi3): Likewise.
17539 (eqv_v4si3): Likewise.
17540 (eqv_v2di3): Likewise.
17541 (eqv_v4sf3): Likewise.
17542 (eqv_v2df3): Likewise.
17543 (nand): Likewise.
17544 (nand_v16qi3): Likewise.
17545 (nand_v8hi3): Likewise.
17546 (nand_v4si3): Likewise.
17547 (nand_v2di3): Likewise.
17548 (nand_v4sf3): Likewise.
17549 (nand_v2df3): Likewise.
17550 (orc): Likewise.
17551 (orc_v16qi3): Likewise.
17552 (orc_v8hi3): Likewise.
17553 (orc_v4si3): Likewise.
17554 (orc_v2di3): Likewise.
17555 (orc_v4sf3): Likewise.
17556 (orc_v2df3): Likewise.
17557
17558 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
17559 allow power8 quad mode in 64-bit.
17560 (rs6000_builtin_vectorized_function): Add support to vectorize
17561 ISA 2.07 count leading zeros, population count builtins.
17562 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
17563 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
17564 (builtin_function_type): Add vgbbd builtin function which takes an
17565 unsigned argument.
17566 (altivec_expand_vec_perm_const): Add support for new power8 merge
17567 instructions.
17568
17569 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
17570 that does not include TImdoe for use with 32-bit.
17571 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
17572 instructions.
17573 (UNSPEC_VSX_CVDPSPN): Likewise.
17574 (vsx_xscvdpspn): Likewise.
17575 (vsx_xscvspdpn): Likewise.
17576 (vsx_xscvdpspn_scalar): Likewise.
17577 (vsx_xscvspdpn_directmove): Likewise.
17578 (vsx_and<mode>3): Split logical operations into 32-bit and
17579 64-bit. Add support to do logical operations on TImode as well as
17580 VSX vector types. Allow logical operations to be done in either
17581 VSX registers or in general purpose registers in 64-bit mode. Add
17582 splitters if GPRs were used. For AND, add clobber of CCmode to
17583 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
17584 encoding.
17585 (vsx_and<mode>3_32bit): Likewise.
17586 (vsx_and<mode>3_64bit): Likewise.
17587 (vsx_ior<mode>3): Likewise.
17588 (vsx_ior<mode>3_32bit): Likewise.
17589 (vsx_ior<mode>3_64bit): Likewise.
17590 (vsx_xor<mode>3): Likewise.
17591 (vsx_xor<mode>3_32bit): Likewise.
17592 (vsx_xor<mode>3_64bit): Likewise.
17593 (vsx_one_cmpl<mode>2): Likewise.
17594 (vsx_one_cmpl<mode>2_32bit): Likewise.
17595 (vsx_one_cmpl<mode>2_64bit): Likewise.
17596 (vsx_nor<mode>3): Likewise.
17597 (vsx_nor<mode>3_32bit): Likewise.
17598 (vsx_nor<mode>3_64bit): Likewise.
17599 (vsx_andc<mode>3): Likewise.
17600 (vsx_andc<mode>3_32bit): Likewise.
17601 (vsx_andc<mode>3_64bit): Likewise.
17602 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
17603 and xxlorc instructions.
17604 (vsx_eqv<mode>3_64bit): Likewise.
17605 (vsx_nand<mode>3_32bit): Likewise.
17606 (vsx_nand<mode>3_64bit): Likewise.
17607 (vsx_orc<mode>3_32bit): Likewise.
17608 (vsx_orc<mode>3_64bit): Likewise.
17609
17610 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
17611
17612 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
17613 instruction.
17614 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
17615 (p8_vmrgow): Likewise.
17616 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
17617 GPRs to be split under VSX.
17618 (p8v_clz<mode>2): Add power8 count leading zero support.
17619 (p8v_popcount<mode>2): Add power8 population count support.
17620 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
17621 support.
17622
17623 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
17624 instruction.
17625
17626 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
17627 builtin functions.
17628 (vec_nand): Likewise.
17629 (vec_vclz): Likewise.
17630 (vec_vclzb): Likewise.
17631 (vec_vclzd): Likewise.
17632 (vec_vclzh): Likewise.
17633 (vec_vclzw): Likewise.
17634 (vec_vgbbd): Likewise.
17635 (vec_vmrgew): Likewise.
17636 (vec_vmrgow): Likewise.
17637 (vec_vpopcnt): Likewise.
17638 (vec_vpopcntb): Likewise.
17639 (vec_vpopcntd): Likewise.
17640 (vec_vpopcnth): Likewise.
17641 (vec_vpopcntw): Likewise.
17642
17643 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17644
17645 PR rtl-optimization/57468
17646 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
17647 spilled pseudos.
17648
17649 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17650
17651 PR rtl-optimization/57459
17652 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
17653 type when setting live regs.
17654
17655 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
17656
17657 * config/s390/s390.opt (mlra): New option.
17658 * config/s390/s390.c (s390_decompose_address): Check displacement
17659 for all registers for LRA.
17660 (s390_secondary_reload): Don't used secondary reloads for LRA.
17661 (s390_lra_p): New function.
17662 (TARGET_LRA_P): Define.
17663 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
17664 of attribute cpu_facility to zarch for the last alternative.
17665 (*cmpmem_short): Ditto.
17666
17667 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
17668
17669 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
17670 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
17671 (arm_expand_prologue): Likewise.
17672
17673 2013-06-06 Teresa Johnson <tejohnson@google.com>
17674
17675 PR c++/53743
17676 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
17677 as this is now done by redirect_edge_and_branch_force.
17678 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
17679 barriers, and fix interaction with splitting.
17680 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
17681 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
17682 reflect changes made in the routine.
17683 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
17684 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
17685 since this is called in cfglayout mode, and replace partition fixup
17686 with assert as that is now done by force_nonfallthru_and_redirect.
17687 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
17688 already be marked with region crossing note.
17689 (insert_section_boundary_note): Make non-static, gate on flag
17690 has_bb_partition, rewrite to also check for multiple partitions.
17691 (rest_of_handle_reorder_blocks): Remove call to
17692 insert_section_boundary_note, now done later during free_cfg.
17693 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
17694 * bb-reorder.h (insert_section_boundary_note): Declare.
17695 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
17696 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
17697 invoke insert_section_boundary_note.
17698 (try_redirect_by_replacing_jump): Remove unnecessary
17699 check for region crossing note.
17700 (fixup_partition_crossing): New function.
17701 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
17702 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
17703 in non-cfglayout mode.
17704 (force_nonfallthru_and_redirect): Fixup partition boundaries,
17705 remove old code that tried to do this. Emit barrier correctly
17706 when we are in cfglayout mode.
17707 (last_bb_in_partition): New function.
17708 (rtl_split_edge): Correctly fixup partition boundaries.
17709 (commit_one_edge_insertion): Remove old code that tried to
17710 fixup region crossing edge since this is now handled in
17711 split_block, and set up insertion point correctly since
17712 block may now end in a jump.
17713 (verify_hot_cold_block_grouping): Guard against checking when not in
17714 linearized RTL mode.
17715 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
17716 notes.
17717 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
17718 rtl_verify_flow_info, so not called in cfglayout mode.
17719 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
17720 (fixup_reorder_chain): Remove old code that attempted to fixup region
17721 crossing note as this is now handled in force_nonfallthru_and_redirect.
17722 (duplicate_insn_chain): Don't duplicate switch section notes.
17723 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
17724 note.
17725 * basic-block.h (emit_barrier_after_bb): Declare.
17726
17727 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17728
17729 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
17730 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
17731 arm_usatsihi): Adjust alternatives for arm_restrict_it.
17732
17733 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17734
17735 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
17736 where appropriate.
17737 * config/arm/ldmstm.md: Regenerate.
17738
17739 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17740
17741 * config/arm/sync.md (atomic_loaddi_1):
17742 Disable predication for arm_restrict_it.
17743 (arm_load_exclusive<mode>): Likewise.
17744 (arm_load_exclusivesi): Likewise.
17745 (arm_load_exclusivedi): Likewise.
17746 (arm_load_acquire_exclusive<mode>): Likewise.
17747 (arm_load_acquire_exclusivesi): Likewise.
17748 (arm_load_acquire_exclusivedi): Likewise.
17749 (arm_store_exclusive<mode>): Likewise.
17750 (arm_store_exclusive<mode>): Likewise.
17751 (arm_store_release_exclusivedi): Likewise.
17752 (arm_store_release_exclusive<mode>): Likewise.
17753
17754 2013-06-06 Richard Biener <rguenther@suse.de>
17755
17756 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
17757 after LTO_null.
17758 (lto_tag_is_tree_code_p): Adjust.
17759 (lto_tag_is_gimple_code_p): Likewise.
17760 (lto_gimple_code_to_tag): Likewise.
17761 (lto_tag_to_gimple_code): Likewise.
17762 (lto_tree_code_to_tag): Likewise.
17763 (lto_tag_to_tree_code): Likewise.
17764 * data-streamer.h (streamer_write_hwi_in_range): Use
17765 uhwi streaming to stream the normalized range.
17766 (streamer_read_hwi_in_range): Likewise.
17767
17768 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17769
17770 * config/arm/arm.md (enabled_for_depr_it): New attribute.
17771 (predicable_short_it): Likewise.
17772 (predicated): Likewise.
17773 (enabled): Handle above.
17774 (define_cond_exec): Set predicated attribute to yes.
17775
17776 2013-06-05 Mike Stump <mikestump@comcast.net>
17777
17778 * gdbinit.in (__FUNCTION__): Add.
17779
17780 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
17781
17782 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
17783 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
17784
17785 2013-06-05 Jan Hubicka <jh@suse.cz>
17786
17787 * varasm.c (mark_decl_referenced): Revert the removal until targets
17788 are fixed.
17789
17790 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
17791
17792 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
17793 instead of mark_decl_referenced.
17794
17795 2013-06-05 Jan Hubicka <jh@suse.cz>
17796
17797 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
17798 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
17799 and symtab_used_from_object_file_p.
17800 (cgraph_make_node_local_1): Clear forced_by_abi.
17801 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
17802 * cgraph.h (symtab_node_base): Add forced_by_abi.
17803 (decide_is_variable_needed): Remove.
17804 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
17805 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
17806 (decide_is_symbol_needed): ... this one; handle symbols in general;
17807 always analyze virtuals; honnor forced_by_abi.
17808 (cgraph_finalize_function): Update.
17809 (varpool_finalize_decl): Update.
17810 (symbol_defined_and_needed): Remove.
17811 (analyze_functions): Update.
17812 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17813 output_refs, input_overwrite_node): Handle forced_by_abi.
17814 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
17815 (address_taken_from_non_vtable_p): ... this one.
17816 (comdat_can_be_unshared_p_1): New function.
17817 (cgraph_comdat_can_be_unshared_p): Rename to ...
17818 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
17819 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
17820 (function_and_variable_visibility): Clear forced_by_abi as needed.
17821 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
17822 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
17823 * symtab.c (dump_symtab_base): Dump forced_by_abi.
17824 * varpool.c (decide_is_variable_needed): Remove.
17825
17826 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17827
17828 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
17829 (arm_option_override): Override arm_restrict_it where appropriate.
17830 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
17831 * config/arm/arm.opt (mrestrict-it): New command-line option.
17832 * doc/invoke.texi: Document -mrestrict-it.
17833
17834 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17835
17836 * tsan.c (tsan_atomic_table): Make const.
17837
17838 2013-06-05 Richard Biener <rguenther@suse.de>
17839
17840 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
17841 index associated with the tree we are supposed to replace.
17842 * tree-streamer-out.c (pack_ts_base_value_fields): Output
17843 TREE_ASM_WRITTEN as zero for everything but SSA names.
17844
17845 2013-06-05 David Malcolm <dmalcolm@redhat.com>
17846
17847 * tree-ssa-structalias.c (call_stmt_vars): Make static.
17848
17849 2013-06-04 Jan Hubicka <jh@suse.cz>
17850
17851 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
17852 (input_node, input_varpool_node): Handle correctly external same
17853 body aliases.
17854 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
17855 nodes at ltrans stage.
17856
17857 2013-06-04 Jan Hubicka <jh@suse.cz>
17858
17859 * ipa-inline.c (update_caller_keys): Fix availability test.
17860 (update_callee_keys): Likewise.
17861 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
17862 to follow ELF standard.
17863
17864 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
17865
17866 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
17867 (mips64r5900el-*-elf*): New configurations.
17868 * config/mips/mips-cpus.def (r5900): New processor.
17869 * config/mips/mips-tables.opt: Regenerate.
17870 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
17871 (mips_issue_rate): Handle PROCESSOR_R5900.
17872 (mips_reorg_process_insns): Force reorder mode for the R5900.
17873 * config/mips/mips.h (TARGET_MIPS5900): Define.
17874 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
17875 TARGET_MIPS5900.
17876 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
17877 TARGET_MIPS5900.
17878 * config/mips/mips.md (processor): Add r5900.
17879 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
17880
17881 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17882
17883 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
17884 into function to generate MOVI instruction.
17885 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
17886 (aarch64_preferred_simd_mode): Turn into wrapper.
17887 (aarch64_output_scalar_simd_mov_immediate): New function.
17888 * config/aarch64/aarch64-protos.h: Add prototype for above.
17889
17890 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17891
17892 * config/aarch64/aarch64.c (simd_immediate_info): Remove
17893 element_char member.
17894 (sizetochar): Return signed char.
17895 (aarch64_simd_valid_immediate): Remove elchar and other
17896 unnecessary variables.
17897 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
17898 Calculate element_char as required.
17899 * config/aarch64/aarch64-protos.h: Update and move prototype
17900 for aarch64_output_simd_mov_immediate.
17901 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
17902 Update arguments.
17903
17904 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17905
17906 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
17907 information completed by aarch64_simd_valid_immediate.
17908 (aarch64_legitimate_constant_p): Update arguments.
17909 (aarch64_simd_valid_immediate): Work with struct rather than many
17910 pointers.
17911 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
17912 (aarch64_simd_make_constant): Update arguments.
17913 (aarch64_output_simd_mov_immediate): Work with struct rather than
17914 many pointers. Output immediate directly rather than as operand.
17915 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17916 Update prototype.
17917 * config/aarch64/constraints.md (Dn): Update arguments.
17918
17919 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17920
17921 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
17922 longer static.
17923 (aarch64_simd_immediate_valid_for_move): Remove.
17924 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
17925 (aarch64_simd_make_constant): Update call.
17926 (aarch64_output_simd_mov_immediate): Update call.
17927 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
17928 Add prototype.
17929 * config/aarch64/constraints.md (Dn): Update call.
17930
17931 2013-06-04 Ian Bolton <ian.bolton@arm.com>
17932
17933 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
17934 return type to bool for prototype.
17935 (aarch64_legitimate_constant_p): Check for true instead of not -1.
17936 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
17937 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
17938
17939 2013-06-04 Catherine Moore <clm@codesourcery.com>
17940
17941 * config/mips/mips.opt (meva): New.
17942 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
17943 (ASM_SPEC): Handle -meva.
17944 * doc/invoke.texi (meva): Document.
17945
17946 2013-06-04 Alan Modra <amodra@gmail.com>
17947
17948 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
17949 constant output.
17950
17951 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17952
17953 * rtl.def: Add extra fourth optional field to define_cond_exec.
17954 * gensupport.c (process_one_cond_exec): Process attributes from
17955 define_cond_exec.
17956 * doc/md.texi: Document fourth field in define_cond_exec.
17957
17958 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
17959
17960 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
17961 out the processing order as in store_bit_field_1.
17962
17963 2013-06-04 Jan Hubicka <jh@suse.cz>
17964
17965 PR middle-end/57500
17966 * cgraphunit.c (cgraph_process_same_body_aliases): Create
17967 non-VAR_DECL node if it does not exist yet.
17968
17969 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
17970
17971 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
17972 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
17973 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
17974 target_cpu_default setting.
17975
17976 2013-06-03 Teresa Johnson <tejohnson@google.com>
17977
17978 * dumpfile.c (opt_info_switch_p): Change -fopt-info
17979 default to -fopt-info=optimized instead of all.
17980 * doc/invoke.texi: Ditto.
17981 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
17982 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
17983 (execute_vect_slp): Emit BB vectorization success under
17984 MSG_OPTIMIZED_LOCATIONS.
17985 * tree-vect-slp.c (vect_slp_transform_bb): Change
17986 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
17987 * tree-vect-loop.c (vect_transform_loop): Ditto.
17988
17989 2013-06-03 Jason Merrill <jason@redhat.com>
17990
17991 PR c++/57415
17992 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
17993 Use TARGET_EXPR for C++.
17994
17995 2013-06-03 Jakub Jelinek <jakub@redhat.com>
17996
17997 PR rtl-optimization/57268
17998 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
17999 if DEBUG_INSN_P (insn).
18000
18001 Reapply
18002 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18003
18004 PR rtl-optimization/57268
18005 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18006 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18007
18008 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18009
18010 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
18011 (ix86_avoid_lea_for_addr): Likewise.
18012 (exact_dependency_1): Likewise.
18013 (ix86_adjust_cost): Likewise.
18014 (swap_top_of_ready_list): Fix formatting and !reload_completed check
18015 removed.
18016 (do_reorder_for_imul): Fix typo, formatting and
18017 !reload_completed check removed.
18018 (ix86_sched_reorder): Fix typo and formatting.
18019 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
18020 list.
18021
18022 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
18023
18024 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
18025
18026 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
18027
18028 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
18029 <STRING_CST>: Likewise.
18030 <VECTOR_CST>: Likewise.
18031
18032 2013-06-01 Janus Weil <janus@gcc.gnu.org>
18033 Mikael Morin <mikael@gcc.gnu.org>
18034
18035 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
18036 * config.in: Regenerated.
18037 * configure: Regenerated.
18038
18039 2013-06-01 Jan Hubicka <jh@suse.cz>
18040
18041 PR middle-end/57366
18042 * cgraphunit.c (compile): When weakref is not supported,
18043 set up transparent aliases before final output pass.
18044 * varasm.c (assemble_alias): Do not try to do it here.
18045
18046 2013-06-01 Jan Hubicka <jh@suse.cz>
18047
18048 PR middle-end/57467
18049 * passes.c (for_per_function): Skip unanalyzed functions.
18050
18051 2013-06-01 Jan Hubicka <jh@suse.cz>
18052
18053 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
18054 (lto_symtab_merge_symbols_1): ... this one.
18055 (lto_symtab_merge_cgraph_nodes): Rename to ...
18056 (lto_symtab_merge_symbols): ... this one; simplify.
18057 * cgraph.c (same_body_aliases_done): Rename to ...
18058 (cpp_implicit_aliases_done): ... this one.
18059 (cgraph_create_function_alias): Update.
18060 (cgraph_same_body_alias): Update.
18061 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
18062 (verify_edge_corresponds_to_fndecl): Simplify.
18063 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
18064 (cgraph_node): Remove same_body_alias.
18065 (varpool_node): Remove alias_of and extra_name_alias.
18066 (same_body_aliases_done): Rename to ..
18067 (cpp_implicit_aliases_done): ... this one.
18068 (symtab_alias_ultimate_target): Add default parameter.
18069 (symtab_resolve_alias): New function.
18070 (fixup_same_cpp_alias_visibility): Declare.
18071 (cgraph_function_node): Add default parameter.
18072 (cgraph_node_asm_name): Likewise.
18073 (cgraph_function_or_thunk_node): Add default parameter; do
18074 not ICE when it is NULL.
18075 (varpool_variable_node): Likewise.
18076 * tree-emutls.c (create_emultls_var): Update.
18077 (ipa_lower_emutls): Update.
18078 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
18079 (cgraph_reset_node): Reset alias info.
18080 (cgraph_finalize_function): Update.
18081 (fixup_same_cpp_alias_visibility): Move to symtab.c.
18082 (analyze_function): Simplify.
18083 (cgraph_process_same_body_aliases): Simplify.
18084 (analyze_functions): Fixup same body aliases.
18085 (handle_alias_pairs): Simplify.
18086 (assemble_thunk): Update.
18087 (assemble_thunks_and_aliases): Update.
18088 (output_weakrefs): Rewrite.
18089 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
18090 (lto_output_varpool_node): Likewise.
18091 (compute_ltrans_boundary): Remve assert.
18092 (get_alias_symbol): New functoin.
18093 (input_node): Rewrite alias handling.
18094 (input_varpool_node): Likewise.
18095 * ipa-pure-const.c (propagate_pure_const): Fix formating.
18096 * ipa.c (process_references): Handle weakrefs correctly.
18097 (symtab_remove_unreachable_nodes): Likewise.
18098 * trans-mem.c (get_cg_data): Update.
18099 (ipa_tm_create_version_alias): Update.
18100 (ipa_tm_execute): Update.
18101 * symtab.c (dump_symtab_base): Dump aliases.
18102 (verify_symtab_base): Verify aliases.
18103 (symtab_node_availability): New function.
18104 (symtab_alias_ultimate_target): Simplify.
18105 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
18106 handle all the fixup cases.
18107 (symtab_resolve_alias): New function.
18108 * passes.c (ipa_write_summaries): Handle weakrefs.
18109 * varpool.c (varpool_analyze_node): Simplify.
18110 (assemble_aliases): Update.
18111 (varpool_create_variable_alias): Simplify.
18112 (varpool_extra_name_alias): Simplify.
18113 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
18114 (lto_symtab_merge_symbols): ... this one.
18115
18116 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
18117
18118 Revert
18119 PR rtl-optimization/57268
18120 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18121 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18122
18123 2013-06-01 Tobias Burnus <burnus@net-b.de>
18124
18125 Partially reverted:
18126 2013-05-31 Tobias Burnus <burnus@net-b.de>
18127
18128 PR middle-end/57073
18129 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
18130 further up.
18131
18132 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
18133
18134 PR rtl-optimization/57268
18135 * sched-deps.c (sched_analyze_2): Flush dependence lists if
18136 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
18137
18138 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18139
18140 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
18141 unordered comparison operators when -fno-trapping-math is in effect
18142 on the e500.
18143 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
18144 and implement unordered comparison operators properly on the e500.
18145
18146 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
18147
18148 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
18149 for constant scalar integers.
18150 (simplify_relational_operation_1): Likewise.
18151
18152 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
18153
18154 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
18155 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
18156 Fix comment.
18157
18158 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18159 Igor Zamyatin <igor.zamyatin@intel.com>
18160
18161 Silvermont (SLM) architecture performance tuning.
18162 * config/i386/i386.h (enum ix86_tune_indices): Add
18163 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
18164 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
18165
18166 * config/i386/i386.c (initial_ix86_tune_features)
18167 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
18168 (ix86_lea_outperforms): Handle Silvermont tuning.
18169 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
18170 call.
18171 (ix86_use_lea_for_mov): Likewise.
18172 (ix86_avoid_lea_for_addr): Likewise.
18173 (ix86_lea_for_add_ok): Likewise.
18174 (exact_dependency_1): New function.
18175 (exact_store_load_dependency): Likewise.
18176 (ix86_adjust_cost): Handle Silvermont tuning.
18177 (do_reoder_for_imul): Likewise.
18178 (swap_top_of_ready_list): New function.
18179 (ix86_sched_reorder): Changed to handle Silvermont tuning.
18180
18181 * config/i386/i386.md (peepholes that split memory operand in fp
18182 converts): New.
18183
18184 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18185
18186 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
18187 Remove un-necessary braces.
18188
18189 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
18190
18191 * config/aarch64/aarch64.c (aarch64_classify_symbol):
18192 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
18193
18194 2013-05-31 Tobias Burnus <burnus@net-b.de>
18195
18196 PR middle-end/57073
18197 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
18198
18199 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18200
18201 PR target/56315
18202 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
18203 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
18204 * config/arm/neon.md (iordi3_neon): Remove.
18205 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
18206 * config/arm/predicates.md (imm_for_neon_logic_operand):
18207 Move to earlier in the file.
18208 (neon_logic_op2): Likewise.
18209 (arm_iordi_operand_neon): New predicate.
18210
18211 2013-05-31 Richard Biener <rguenther@suse.de>
18212
18213 PR tree-optimization/57478
18214 PR tree-optimization/57453
18215 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
18216 are life as well.
18217
18218 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
18219
18220 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
18221 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
18222
18223 2013-05-30 Tobias Burnus <burnus@net-b.de>
18224 Thomas Koenig <tkoenig@gcc.gnu.org>
18225
18226 PR middle-end/57073
18227 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
18228 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
18229
18230 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
18231
18232 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
18233
18234 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
18235
18236 * target.def (register_usage_leveling_p): New hook.
18237 * targhooks.c (default_register_usage_leveling_p): New.
18238 * targhooks.h (default_register_usage_leveling_p): New prototype.
18239 * lra-assigns.c (register_usage_leveling_p): Use the hook.
18240 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
18241 * doc/tm.texi: Update.
18242 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
18243
18244 2013-05-30 Ian Bolton <ian.bolton@arm.com>
18245
18246 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
18247 (*insv_reg<mode>): New define_insn.
18248
18249 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
18250
18251 PR rtl-optimization/57439
18252 * postreload.c (move2add_valid_value_p): Check that we have
18253 a zero subreg_regno_offset when accessing the register in
18254 the requested mode.
18255
18256 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
18257 Igor Zamyatin <igor.zamyatin@intel.com>
18258
18259 Silvermont (SLM) architecture pipeline model, tuning and
18260 insn selection.
18261 * config.gcc: Add slm config options and target.
18262
18263 * config/i386/slm.md: New.
18264
18265 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
18266
18267 * config/i386/i386-c.c (ix86_target_macros_internal): New case
18268 PROCESSOR_SLM.
18269 (ix86_target_macros_internal): Likewise.
18270
18271 * config/i386/i386.c (slm_cost): New cost.
18272 (m_SLM): New macro flag.
18273 (initial_ix86_tune_features): Set m_SLM.
18274 (x86_accumulate_outgoing_args): Likewise.
18275 (x86_arch_always_fancy_math_387): Likewise.
18276 (processor_target_table): Add slm cost.
18277 (cpu_names): Add slm cpu name.
18278 (x86_option_override_internal): Set SLM ISA.
18279 (ix86_issue_rate): New case PROCESSOR_SLM.
18280 (ia32_multipass_dfa_lookahead): Likewise.
18281 (fold_builtin_cpu): Add slm.
18282
18283 * config/i386/i386.h (TARGET_SLM): New target macro.
18284 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
18285 (processor_type): Add PROCESSOR_SLM.
18286
18287 * config/i386/i386.md (cpu): Add new value "slm".
18288 (slm.md): Include slm.md.
18289
18290 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
18291 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18292
18293 * config/arm/arm-protos.h: Add and update function protos.
18294 * config/arm/arm.c (use_simple_return_p): New added.
18295 (thumb2_expand_return): Check simple_return flag.
18296 * config/arm/arm.md: Add simple_return and conditional simple_return.
18297 * config/arm/iterators.md: Add iterator for return and simple_return.
18298
18299 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
18300
18301 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
18302 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
18303 (arm_emit_vfp_multi_reg_pop): Likewise.
18304 (thumb2_emit_ldrd_pop): Likewise.
18305 (arm_expand_epilogue): Add misc REG_CFA notes.
18306 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
18307
18308 2013-05-29 Lawrence Crowl <crowl@google.com>
18309
18310 * config/arm/t-arm: Update for below.
18311
18312 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
18313 Change type to hash_table. Update dependent calls and types.
18314
18315 * config/i386/t-cygming: Update for below.
18316
18317 * config/i386/t-interix: Update for below.
18318
18319 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
18320 Change type to hash_table. Update dependent calls and types.
18321 (i386_find_on_wrapper_list::wrappers): Likewise.
18322
18323 * config/ia64/t-ia64: Update for below.
18324
18325 * config/ia64/ia64.c (bundle_state_table):
18326 Change type to hash_table. Update dependent calls and types.
18327
18328 * config/mips/mips.c (mips_reorg_process_insns::htab):
18329 Change type to hash_table. Update dependent calls and types.
18330
18331 * config/sol2.c (solaris_comdat_htab):
18332 Change type to hash_table. Update dependent calls and types.
18333
18334 * config/t-sol2: Update for above.
18335
18336 2013-05-29 Teresa Johnson <tejohnson@google.com>
18337
18338 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
18339 functions are not yet marked as defined.
18340
18341 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
18342 Pat Haugen <pthaugen@us.ibm.com>
18343 Peter Bergner <bergner@vnet.ibm.com>
18344
18345 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
18346 instructions.
18347 (VEC_A): Likewise.
18348 (VEC_C): Likewise.
18349 (vrotl<mode>3): Likewise.
18350 (vashl<mode>3): Likewise.
18351 (vlshr<mode>3): Likewise.
18352 (vashr<mode>3): Likewise.
18353
18354 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
18355 support for power8 V2DI builtins.
18356
18357 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
18358 power8 V2DI builtins.
18359 (vupkhsw): Likewise.
18360 (vupklsw): Likewise.
18361 (vaddudm): Likewise.
18362 (vminsd): Likewise.
18363 (vmaxsd): Likewise.
18364 (vminud): Likewise.
18365 (vmaxud): Likewise.
18366 (vpkudum): Likewise.
18367 (vpksdss): Likewise.
18368 (vpkudus): Likewise.
18369 (vpksdus): Likewise.
18370 (vrld): Likewise.
18371 (vsld): Likewise.
18372 (vsrd): Likewise.
18373 (vsrad): Likewise.
18374 (vsubudm): Likewise.
18375 (vcmpequd): Likewise.
18376 (vcmpgtsd): Likewise.
18377 (vcmpgtud): Likewise.
18378 (vcmpequd_p): Likewise.
18379 (vcmpgtsd_p): Likewise.
18380 (vcmpgtud_p): Likewise.
18381 (vupkhsw): Likewise.
18382 (vupklsw): Likewise.
18383 (vaddudm): Likewise.
18384 (vmaxsd): Likewise.
18385 (vmaxud): Likewise.
18386 (vminsd): Likewise.
18387 (vminud): Likewise.
18388 (vpksdss): Likewise.
18389 (vpksdus): Likewise.
18390 (vpkudum): Likewise.
18391 (vpkudus): Likewise.
18392 (vrld): Likewise.
18393 (vsld): Likewise.
18394 (vsrad): Likewise.
18395 (vsrd): Likewise.
18396 (vsubudm): Likewise.
18397
18398 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
18399 support for power8 V2DI instructions.
18400
18401 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
18402 power8 V2DI instructions. Combine pack and unpack insns to use an
18403 iterator for each mode. Check whether a particular mode supports
18404 Altivec instructions instead of just checking TARGET_ALTIVEC.
18405 (UNSPEC_VPKUWUM): Likewise.
18406 (UNSPEC_VPKSHSS): Likewise.
18407 (UNSPEC_VPKSWSS): Likewise.
18408 (UNSPEC_VPKUHUS): Likewise.
18409 (UNSPEC_VPKSHUS): Likewise.
18410 (UNSPEC_VPKUWUS): Likewise.
18411 (UNSPEC_VPKSWUS): Likewise.
18412 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
18413 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
18414 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
18415 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
18416 (UNSPEC_VUPKHSB): Likewise.
18417 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
18418 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
18419 (UNSPEC_VUPKHSH): Likewise.
18420 (UNSPEC_VUPKLSB): Likewise.
18421 (UNSPEC_VUPKLSH): Likewise.
18422 (VI2): Likewise.
18423 (VI_char): Likewise.
18424 (VI_scalar): Likewise.
18425 (VI_unit): Likewise.
18426 (VP): Likewise.
18427 (VP_small): Likewise.
18428 (VP_small_lc): Likewise.
18429 (VU_char): Likewise.
18430 (add<mode>3): Likewise.
18431 (altivec_vaddcuw): Likewise.
18432 (altivec_vaddu<VI_char>s): Likewise.
18433 (altivec_vadds<VI_char>s): Likewise.
18434 (sub<mode>3): Likewise.
18435 (altivec_vsubcuw): Likewise.
18436 (altivec_vsubu<VI_char>s): Likewise.
18437 (altivec_vsubs<VI_char>s): Likewise.
18438 (altivec_vavgs<VI_char>): Likewise.
18439 (altivec_vcmpbfp): Likewise.
18440 (altivec_eq<mode>): Likewise.
18441 (altivec_gt<mode>): Likewise.
18442 (altivec_gtu<mode>): Likewise.
18443 (umax<mode>3): Likewise.
18444 (smax<mode>3): Likewise.
18445 (umin<mode>3): Likewise.
18446 (smin<mode>3): Likewise.
18447 (altivec_vpkuhum): Likewise.
18448 (altivec_vpkuwum): Likewise.
18449 (altivec_vpkshss): Likewise.
18450 (altivec_vpkswss): Likewise.
18451 (altivec_vpkuhus): Likewise.
18452 (altivec_vpkshus): Likewise.
18453 (altivec_vpkuwus): Likewise.
18454 (altivec_vpkswus): Likewise.
18455 (altivec_vpks<VI_char>ss): Likewise.
18456 (altivec_vpks<VI_char>us): Likewise.
18457 (altivec_vpku<VI_char>us): Likewise.
18458 (altivec_vpku<VI_char>um): Likewise.
18459 (altivec_vrl<VI_char>): Likewise.
18460 (altivec_vsl<VI_char>): Likewise.
18461 (altivec_vsr<VI_char>): Likewise.
18462 (altivec_vsra<VI_char>): Likewise.
18463 (altivec_vsldoi_<mode>): Likewise.
18464 (altivec_vupkhsb): Likewise.
18465 (altivec_vupkhs<VU_char>): Likewise.
18466 (altivec_vupkls<VU_char>): Likewise.
18467 (altivec_vupkhsh): Likewise.
18468 (altivec_vupklsb): Likewise.
18469 (altivec_vupklsh): Likewise.
18470 (altivec_vcmpequ<VI_char>_p): Likewise.
18471 (altivec_vcmpgts<VI_char>_p): Likewise.
18472 (altivec_vcmpgtu<VI_char>_p): Likewise.
18473 (abs<mode>2): Likewise.
18474 (vec_unpacks_hi_v16qi): Likewise.
18475 (vec_unpacks_hi_v8hi): Likewise.
18476 (vec_unpacks_lo_v16qi): Likewise.
18477 (vec_unpacks_hi_<VP_small_lc>): Likewise.
18478 (vec_unpacks_lo_v8hi): Likewise.
18479 (vec_unpacks_lo_<VP_small_lc>): Likewise.
18480 (vec_pack_trunc_v8h): Likewise.
18481 (vec_pack_trunc_v4si): Likewise.
18482 (vec_pack_trunc_<mode>): Likewise.
18483
18484 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
18485 V2DI builtins.
18486 (vec_vmaxsd): Likewise.
18487 (vec_vmaxud): Likewise.
18488 (vec_vminsd): Likewise.
18489 (vec_vminud): Likewise.
18490 (vec_vpksdss): Likewise.
18491 (vec_vpksdus): Likewise.
18492 (vec_vpkudum): Likewise.
18493 (vec_vpkudus): Likewise.
18494 (vec_vrld): Likewise.
18495 (vec_vsld): Likewise.
18496 (vec_vsrad): Likewise.
18497 (vec_vsrd): Likewise.
18498 (vec_vsubudm): Likewise.
18499 (vec_vupkhsw): Likewise.
18500 (vec_vupklsw): Likewise.
18501
18502 2013-05-29 Jan Hubicka <jh@suse.cz>
18503
18504 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
18505 flags; reorder rest of fields in more consistent way.
18506 (varpool_node): Remove analyzed, finalized and alias.
18507 (cgraph_ndoe): Likewise.
18508 (symtab_alias_ultimate_target): New function.
18509 (cgraph_function_node): Move offline.
18510 (cgraph_reset_node): Declare.
18511 (cgraph_comdat_can_be_unshared_p): Remove.
18512 (varpool_remove_initializer): Declare.
18513 (varpool_first_defined_variable, varpool_next_defined_variable
18514 cgraph_first_defined_function, cgraph_next_defined_function): Update.
18515 (cgraph_function_with_gimple_body_p): Update.
18516 (varpool_all_refs_explicit_p): Update.
18517 (symtab_alias_target): New function.
18518 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
18519 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
18520 (cgraph_function_or_thunk_node): Simplify using
18521 symtab_alias_ultimate_target.
18522 (varpool_variable_node): Likewise.
18523 * cgraph.c (cgraph_create_function_alias): Update.
18524 (cgraph_add_thunk): Update.
18525 (cgraph_remove_node): Update.
18526 (dump_cgraph_node): Do not dump removed flags.
18527 (cgraph_function_body_availability): Update.
18528 (cgraph_propagate_frequency): Update.
18529 (verify_cgraph_node): Check sanity of local flag.
18530 (cgraph_function_node): Move here from cgraph.h; revamp for
18531 cgraph_function_or_thunk_node.
18532 * lto-symtab.c (lto_varpool_replace_node): Update.
18533 (lto_symtab_resolve_can_prevail_p): Update.
18534 (lto_symtab_merge_cgraph_nodes): Update.
18535 * ipa-cp.c (determine_versionability, initialize_node_lattices,
18536 propagate_constants_accross_call, devirtualization_time_bonus,
18537 ipcp_propagate_stage): Update.
18538 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
18539 * ipa-inline-transform.c (clone_inlined_nodes,
18540 preserve_function_body_p): Update.
18541 * ipa-reference.c (propagate): Update.
18542 (write_node_summary_p): Update.
18543 * toplev.c (wrapup_global_declaration_2): Update.
18544 * cgraphunit.c (cgraph_analyze_function): Rename to ...
18545 (analyze_function) ... this one.
18546 (cgraph_process_new_functions): Update.
18547 (cgraph_reset_node): Export.
18548 (cgraph_finalize_function): Update.
18549 (cgraph_add_new_function): Update.
18550 (process_function_and_variable_attributes): Update.
18551 (varpool_finalize_decl): Update.
18552 (symbol_finalized): Remove.
18553 (symbol_finalized_and_needed): Rename to ...
18554 (symbol_defined_and_needed): ... update.
18555 (cgraph_analyze_functions): Update.
18556 (handle_alias_pairs): Update.
18557 (mark_functions_to_output): Update.
18558 (assemble_thunk): Update.
18559 (output_in_order): Update.
18560 (output_weakrefs): Update.
18561 (finalize_compilation_unit): Update.
18562 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
18563 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
18564 input_node, input_varpool_node): Update.
18565 * dbxout.c (dbxout_expand_expr): Update.
18566 * cgraphclones.c (cgraph_clone_node): Update.
18567 (cgraph_copy_node_for_versioning): Update.
18568 (cgraph_materialize_clone): Update.
18569 (cgraph_materialize_all_clones): Update.
18570 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
18571 propagate_pure_const, propagate_nothrow): Update.
18572 * lto-streamer-out.c (lto_output, write_symbol): Update.
18573 * ipa-utils.c (ipa_reverse_postorder): Update.
18574 * ipa-inline.c (can_inline_edge_p): Update.
18575 (update_caller_keys, ipa_inline): Update.
18576 * dwarf2out.c (reference_to_unused,
18577 premark_types_used_by_global_vars_helper): Update.
18578 * tree-eh.c (tree_could_trap_p): Update.
18579 * ipa-split.c (consider_split, execute_split_functions): Update.
18580 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
18581 has_addr_references_p): Update; move ahead in file for better
18582 readability.
18583 (process_references): Simplify.
18584 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
18585 bodies are removed.
18586 (cgraph_comdat_can_be_unshared_p): Make static.
18587 (cgraph_externally_visible_p): Update.
18588 (varpool_externally_visible_p): Update.
18589 (function_and_variable_visibility): Update.
18590 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
18591 ipa_tm_mark_force_output_node): Update.
18592 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
18593 estimate_edge_devirt_benefit, inline_generate_summary,
18594 inline_write_summary): Update.
18595 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
18596 * ipa-prop.c (ipa_compute_jump_functions): Update.
18597 (ipa_print_node_params, ipa_prop_read_section,
18598 ipa_update_after_lto_read, read_replacements_section): Update.
18599 * varasm.c (mark_decl_referenced): Update.
18600 (assemble_alias, dump_tm_clone_pairs): Update.
18601 * tree-inline.c (copy_bb): Update.
18602 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
18603 Update.
18604 * symtab.c (dump_symtab_base): Print new flags.
18605 (verify_symtab_base): Verify new flags.
18606 (symtab_alias_ultimate_target): New function.
18607 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
18608 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
18609 Update.
18610 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
18611 Update.
18612 * i386.c (ix86_get_function_versions_dispatcher,
18613 ix86_generate_version_dispatcher_body): Update.
18614 (fold_builtin_cpu): Use varpool_add_new_variable.
18615 * varpool.c (varpool_remove_initializer): Break out from ...
18616 (varpool_remove_node): ... this one.
18617 (dump_varpool_node, varpool_node_for_asm,
18618 cgraph_variable_initializer_availability, varpool_analyze_node,
18619 varpool_assemble_decl, varpool_remove_unreferenced_decls,
18620 varpool_finalize_named_section_flags, varpool_create_variable_alias):
18621 Update.
18622
18623 2013-05-29 Jan Hubicka <jh@suse.cz>
18624
18625 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
18626
18627 2013-05-29 Easwaran Raman <eraman@google.com>
18628
18629 PR tree-optimization/57442
18630 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
18631 when control exits the main loop.
18632
18633 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
18634
18635 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
18636 and RX600.
18637 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
18638 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
18639 * rx/t-rx: Add rx100 under multi library matches option for nofpu
18640 option.
18641
18642 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
18643
18644 PR tree-optimization/57441
18645 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
18646 Don't limit size of incr_vec to number of candidates.
18647
18648 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
18649
18650 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
18651 and mips16 directories.
18652 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
18653 (MULTILIB_DIRNAMES): Ditto.
18654 (MULTILIB_EXCEPTIONS): Add new exceptions.
18655 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
18656 (MULTILIB_DIRNAMES): Ditto.
18657 (MULTILIB_EXCEPTIONS): Add new exceptions.
18658
18659 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18660 Marcus Shawcroft <marcus.shawcroft@arm.com>
18661
18662 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
18663 SYMBOL_TINY_ABSOLUTE.
18664 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
18665 SYMBOL_TINY_ABSOLUTE.
18666 (aarch64_expand_mov_immediate): Likewise.
18667 (aarch64_classify_symbol): Likewise.
18668 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
18669 Permit SYMBOL_TINY_ABSOLUTE.
18670 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
18671
18672 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18673 Marcus Shawcroft <marcus.shawcroft@arm.com>
18674
18675 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
18676 Refactor if/switch. Replace gcc_assert with if.
18677
18678 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
18679
18680 * config/i386/i386.c (initial_ix86_tune_features): Enable
18681 FP Reassociation for AMD bdver1 and bdver2.
18682
18683 2013-05-29 Martin Jambor <mjambor@suse.cz>
18684
18685 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
18686 and IMAGPART_EXPR do not occur within other handled_components.
18687
18688 2013-05-29 Richard Biener <rguenther@suse.de>
18689
18690 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
18691 access on whether the use is in the BB we currently try to
18692 vectorize.
18693 (vect_bb_vectorization_profitable_p): Pass the BB we currently
18694 vectorize to vect_bb_slp_scalar_cost.
18695
18696 2013-05-29 Richard Biener <rguenther@suse.de>
18697
18698 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
18699 computing scalar cost offsetted by stmts that are kept live
18700 by scalar uses.
18701 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
18702 for computation of scalar cost.
18703
18704 2013-05-28 Steve Ellcey <sellcey@mips.com>
18705
18706 * config/mips/mips-cpus.def (mips32r2): Change processor type.
18707
18708 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
18709
18710 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
18711 array notation built-in reduction functions.
18712 * doc/passes.texi (Passes): Added documentation about changes done
18713 for Cilk Plus.
18714 * doc/invoke.texi (C Dialect Options): Added documentation about
18715 the -fcilkplus flag.
18716 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
18717 (BUILTINS_DEF): Depend on cilkplus.def.
18718 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
18719 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
18720 * cilkplus.def: New file.
18721
18722 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
18723
18724 PR rtl-optimization/57439
18725 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
18726
18727 2013-05-28 Easwaran Raman <eraman@google.com>
18728
18729 PR tree-optimization/57337
18730 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
18731 (find_insert_point): Correctly identify the insertion point
18732 when two statements with the same UID is compared.
18733
18734 2013-05-28 Richard Biener <rguenther@suse.de>
18735
18736 PR tree-optimization/56787
18737 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
18738 from the list of data references.
18739 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
18740 clobbers.
18741 (vect_analyze_loop_operations): Likewise.
18742 (vect_transform_loop): Remove clobbers.
18743
18744 2013-05-28 Martin Jambor <mjambor@suse.cz>
18745
18746 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
18747 and REALPART_EXPRs have scalar type.
18748
18749 2013-05-28 Richard Biener <rguenther@suse.de>
18750
18751 PR tree-optimization/57411
18752 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
18753 virtual operands.
18754 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
18755 virtual operand propagation.
18756
18757 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18758
18759 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
18760 destination register for bmasksi_vis.
18761 (vector_init_bshuffle): Likewise.
18762 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
18763
18764 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
18765
18766 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
18767 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
18768 mode if the instruction isn't available in the original mode.
18769 * config/sparc/sparc.opt (mfix-ut699): New option.
18770 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
18771 (divdf3): Turn into expander.
18772 (divdf3_nofix): New insn.
18773 (divdf3_fix): Likewise.
18774 (divsf3): Disable if -mfix-ut699.
18775 (sqrtdf2): Turn into expander.
18776 (sqrtdf2_nofix): New insn.
18777 (sqrtdf2_fix): Likewise.
18778 (sqrtsf2): Disable if -mfix-ut699.
18779
18780 2013-05-27 Richard Biener <rguenther@suse.de>
18781
18782 PR middle-end/57412
18783 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
18784 block for the new loop.
18785
18786 2013-05-27 Richard Biener <rguenther@suse.de>
18787
18788 PR tree-optimization/57343
18789 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
18790 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
18791 (number_of_iterations_cond): Do not build the folded tree.
18792
18793 2013-05-27 Richard Biener <rguenther@suse.de>
18794
18795 Revert
18796 PR middle-end/57381
18797 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
18798 OEP_CONSTANT_ADDRESS_OF retained.
18799
18800 PR tree-optimization/57417
18801 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
18802 for unchanged base.
18803 (set_ssa_val_to): Compare addresses using
18804 get_addr_base_and_unit_offset.
18805
18806 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
18807
18808 PR rtl-optimization/56833
18809 * postreload.c (move2add_record_mode): New function.
18810 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
18811 (move2add_use_add2_insn): Use move2add_record_sym_value.
18812 (move2add_use_add3_insn): Likewise.
18813 (reload_cse_move2add): Use move2add_valid_value_p and
18814 move2add_record_mode. Invalidate call-clobbered and REG_INC
18815 affected regs by setting reg_mode to VOIDmode.
18816 (move2add_note_store): Don't pretend the inside of a SUBREG is
18817 the actual destination. Invalidate single/leading registers by
18818 setting reg_mode to VOIDmode.
18819 Use move2add_record_sym_value, move2add_valid_value_p and
18820 move2add_record_mode.
18821
18822 2013-05-27 Richard Biener <rguenther@suse.de>
18823
18824 PR tree-optimization/57396
18825 * tree-affine.c (double_int_constant_multiple_p): Properly
18826 return false for val == 0 and div != 0.
18827
18828 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18829
18830 * config/mips/mips.h: Use #elif in preprocessor conditions.
18831
18832 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18833
18834 PR target/53916
18835 * config/mips/constraints.md (kl): New constraint.
18836 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
18837 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
18838 constraint for operand 0. Split after CSE for MIPS16. Emit a move
18839 from LO for MIPS16.
18840 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
18841
18842 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
18843
18844 PR target/55777
18845 * config/mips/mips.c (mips_can_inline_p): New function.
18846 (TARGET_CAN_INLINE_P): Define.
18847
18848 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
18849
18850 * sched-int.h (ds_t, dw_t): Make unsigned int.
18851 Fix documentation that describes how all the ds_t bits are used.
18852 Reserve the last bit for delayed-branch scheduling.
18853 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
18854 (BITS_PER_DEP_WEAK): Fix definition and documentation.
18855 (gen_dep_weak_1): Remove prototype.
18856 * sched-deps.c (get_dep_weak_1): Make static.
18857 * target.def (speculate_insn, needs_block_p, gen_spec_check,
18858 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
18859 * doc/tm.texi: Regenerate.
18860 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
18861
18862 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
18863
18864 PR debug/56950
18865 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
18866
18867 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
18868 Sandra Loosemore <sandra@codesourcery.com>
18869
18870 * config.gcc (powerpc-*): Allow native for with-cpu.
18871
18872 2013-05-24 Jeff Law <law@redhat.com>
18873
18874 PR tree-optimization/57124
18875 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
18876 conversion feeding a condition if the range has an overflow
18877 if -fstrict-overflow. Add warnings for when we do make the
18878 transformation.
18879
18880 2013-05-24 Dehao Chen <dehao@google.com>
18881
18882 * tree-cfg.c (locus_discrim_map): Fix the typo.
18883 (locus_discrim_hasher): Likewise.
18884 (locus_discrim_hasher::hash): Likewise.
18885 (locus_discrim_hasher::equal): Likewise.
18886
18887 2013-05-24 Martin Jambor <mjambor@suse.cz>
18888
18889 PR tree-optimization/57294
18890 * cgraph.h (ipa_record_stmt_references): Declare.
18891 * cgraphbuild.c (ipa_record_stmt_references): New function.
18892 (build_cgraph_edges): Use ipa_record_stmt_references.
18893 (rebuild_cgraph_edges): Likewise.
18894 (cgraph_rebuild_references): Likewise.
18895 * ipa-prop.c (ipa_modify_call_arguments): Discard references
18896 associated with the old statement and build references from the
18897 newly built statements.
18898 * ipa-ref.c (ipa_remove_stmt_references): New function.
18899 * ipa-ref.h (ipa_remove_stmt_references): Declare.
18900
18901 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
18902
18903 * lra-constraints.c (emit_spill_move): Use smaller mode for
18904 mem-mem moves.
18905 (check_and_process_move): Consider mem-reg moves for secondary
18906 too.
18907 (curr_insn_transform): Don't lose insns emitted before for
18908 secondary memory moves.
18909 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
18910 reg set up in the current insn.
18911
18912 2013-05-24 Dehao Chen <dehao@google.com>
18913
18914 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
18915 hash function.
18916 (locus_descrim_hasher::equal): Likewise.
18917 (build_gimple_cfg): New discriminator assignment algorithm.
18918 (make_edges): Likewise.
18919 (next_discriminator_for_locus): Likewise.
18920 (same_line_p): Likewise.
18921 (assign_discriminators): Likewise.
18922 (make_cond_expr_edges): Likewise.
18923 (make_gimple_switch_edges): Likewise.
18924 (make_goto_expr_edges): Likewise.
18925 (make_gimple_asm_edges): Likewise.
18926
18927 2013-05-24 Ian Bolton <ian.bolton@arm.com>
18928
18929 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
18930 X format specifier to only display bottom 16 bits.
18931 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
18932 immediate to match for operand 2, since it will be masked.
18933
18934 2013-05-24 Richard Biener <rguenther@suse.de>
18935
18936 PR tree-optimization/57287
18937 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
18938 all SSA names that occur in abnormal PHIs.
18939
18940 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
18941
18942 PR tree-ssa/57385
18943 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
18944 that index is not negative.
18945
18946 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
18947
18948 PR rtl-optimization/55177
18949 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
18950 (simplify_byte_swapping_operation): New.
18951 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
18952 (simplify_relational_operation_1): Deal with BSWAP.
18953
18954 2013-05-23 Richard Henderson <rth@redhat.com>
18955
18956 PR target/56742
18957 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
18958 (ix86_reorg): Call it.
18959
18960 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
18961
18962 PR target/57379
18963 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
18964 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
18965 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
18966
18967 2013-05-23 Christian Bruel <christian.bruel@st.com>
18968
18969 PR debug/57351
18970 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
18971
18972 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18973 Marcus Shawcroft <marcus.shawcroft@arm.com>
18974
18975 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
18976 * config/aarch64/constraints.md (Usa): Remove.
18977 * doc/md.texi (AArch64 Usa): Remove.
18978
18979 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
18980 Marcus Shawcroft <marcus.shawcroft@arm.com>
18981
18982 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
18983 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
18984 * config/aarch64/predicates.md (aarch64_const_address): Remove.
18985 (aarch64_mov_operand): Use aarch64_mov_operand_p.
18986
18987 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
18988
18989 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
18990 instruction (AdvSIMD).
18991 * config/aarch64/aarch64-builtins.c
18992 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
18993 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
18994
18995 2013-05-23 Martin Jambor <mjambor@suse.cz>
18996
18997 PR middle-end/57347
18998 * tree.h (contains_bitfld_component_ref_p): Declare.
18999 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
19000 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
19001 caller.
19002 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
19003 not access a bit-field. Assert all final offsets are byte-aligned.
19004
19005 2013-05-23 Richard Biener <rguenther@suse.de>
19006
19007 PR tree-optimization/57380
19008 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
19009 least one invariant or re-used load.
19010 * passes.c (init_optimization_passes): Move pass_phiprop before
19011 pass_forwprop.
19012
19013 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
19014
19015 * config/aarch64/aarch64-simd.md
19016 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
19017
19018 2013-05-23 Richard Biener <rguenther@suse.de>
19019
19020 PR middle-end/57381
19021 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
19022 OEP_CONSTANT_ADDRESS_OF retained.
19023
19024 2013-05-23 Jakub Jelinek <jakub@redhat.com>
19025
19026 PR middle-end/57344
19027 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
19028 don't lower unit. Handle unit not being always BITS_PER_WORD.
19029
19030 2013-05-23 Richard Biener <rguenther@suse.de>
19031
19032 PR rtl-optimization/57341
19033 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
19034 instead of true_dependence.
19035
19036 2013-05-22 David Malcolm <dmalcolm@redhat.com>
19037
19038 * bb-reorder.c (branch_threshold): Make const.
19039 (exec_threshold): Ditto.
19040
19041 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19042 Pat Haugen <pthaugen@us.ibm.com>
19043 Peter Bergner <bergner@vnet.ibm.com>
19044
19045 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
19046 documentation for the power8 crypto builtins.
19047
19048 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
19049
19050 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
19051 macros for defining power8 builtin functions.
19052 (BU_P8V_AV_2): Likewise.
19053 (BU_P8V_AV_P): Likewise.
19054 (BU_P8V_VSX_1): Likewise.
19055 (BU_P8V_OVERLOAD_1): Likewise.
19056 (BU_P8V_OVERLOAD_2): Likewise.
19057 (BU_CRYPTO_1): Likewise.
19058 (BU_CRYPTO_2): Likewise.
19059 (BU_CRYPTO_3): Likewise.
19060 (BU_CRYPTO_OVERLOAD_1): Likewise.
19061 (BU_CRYPTO_OVERLOAD_2): Likewise.
19062 (XSCVSPDP): Fix typo, point to the correct instruction.
19063 (VCIPHER): Add power8 crypto builtins.
19064 (VCIPHERLAST): Likewise.
19065 (VNCIPHER): Likewise.
19066 (VNCIPHERLAST): Likewise.
19067 (VPMSUMB): Likewise.
19068 (VPMSUMH): Likewise.
19069 (VPMSUMW): Likewise.
19070 (VPERMXOR_V2DI): Likewise.
19071 (VPERMXOR_V4SI: Likewise.
19072 (VPERMXOR_V8HI: Likewise.
19073 (VPERMXOR_V16QI: Likewise.
19074 (VSHASIGMAW): Likewise.
19075 (VSHASIGMAD): Likewise.
19076 (VPMSUM): Likewise.
19077 (VPERMXOR): Likewise.
19078 (VSHASIGMA): Likewise.
19079
19080 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
19081 __CRYPTO__ if the crypto instructions are available.
19082 (altivec_overloaded_builtins): Add support for overloaded power8
19083 builtins.
19084
19085 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
19086 support for power8 crypto builtins.
19087 (builtin_function_type): Likewise.
19088 (altivec_init_builtins): Add support for builtins that take vector
19089 long long (V2DI) arguments.
19090
19091 * config/rs6000/crypto.md: New file, define power8 crypto
19092 instructions.
19093
19094 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
19095 Pat Haugen <pthaugen@us.ibm.com>
19096 Peter Bergner <bergner@vnet.ibm.com>
19097
19098 * doc/invoke.texi (Option Summary): Add power8 options.
19099 (RS/6000 and PowerPC Options): Likewise.
19100
19101 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
19102 constraints.md instead of rs6000.h. Reorder w* constraints. Add
19103 wm, wn, wr documentation.
19104
19105 * config/rs6000/constraints.md (wm): New constraint for VSX
19106 registers if direct move instructions are enabled.
19107 (wn): New constraint for no registers.
19108 (wq): New constraint for quad word even GPR registers.
19109 (wr): New constraint if 64-bit instructions are enabled.
19110 (wv): New constraint if power8 vector instructions are enabled.
19111 (wQ): New constraint for quad word memory locations.
19112
19113 * config/rs6000/predicates.md (const_0_to_15_operand): New
19114 constraint for 0..15 for crypto instructions.
19115 (gpc_reg_operand): If VSX allow registers in VSX registers as well
19116 as GPR and floating point registers.
19117 (int_reg_operand): New predicate to match only GPR registers.
19118 (base_reg_operand): New predicate to match base registers.
19119 (quad_int_reg_operand): New predicate to match even GPR registers
19120 for quad memory operations.
19121 (vsx_reg_or_cint_operand): New predicate to allow vector logical
19122 operations in both GPR and VSX registers.
19123 (quad_memory_operand): New predicate for quad memory operations.
19124 (reg_or_indexed_operand): New predicate for direct move support.
19125
19126 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
19127 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
19128 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
19129 (POWERPC_MASKS): Add power8 options.
19130 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
19131 various options.
19132
19133 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
19134 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
19135
19136 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
19137 (-mpower8-fusion): New power8 options.
19138 (-mpower8-fusion-sign): Likewise.
19139 (-mpower8-vector): Likewise.
19140 (-mcrypto): Likewise.
19141 (-mdirect-move): Likewise.
19142 (-mquad-memory): Likewise.
19143
19144 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
19145 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
19146 registers.
19147 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
19148 (rs6000_debug_vector_unit): Add p8_vector.
19149 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
19150 definitions. Also print fusion state.
19151 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
19152 (rs6000_builtin_mask_calculate): Add power8 builtin support.
19153 (rs6000_option_override_internal): Add support for power8.
19154 (rs6000_common_init_builtins): Add debugging for skipped builtins
19155 if -mdebug=builtin.
19156 (rs6000_adjust_cost): Add power8 support.
19157 (rs6000_issue_rate): Likewise.
19158 (insn_must_be_first_in_group): Likewise.
19159 (insn_must_be_last_in_group): Likewise.
19160 (force_new_group): Likewise.
19161 (rs6000_register_move_cost): Likewise.
19162 (rs6000_opt_masks): Likewise.
19163
19164 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
19165 power8 capable assembler, default to power7 options.
19166 (TARGET_DIRECT_MOVE): Likewise.
19167 (TARGET_CRYPTO): Likewise.
19168 (TARGET_P8_VECTOR): Likewise.
19169 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
19170 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
19171 (VECTOR_MEM_P8_VECTOR_P): Likewise.
19172 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
19173 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
19174 (TARGET_XSCVDPSPN): Likewise.
19175 (TARGET_XSCVSPDPN): Likewsie.
19176 (TARGET_SYNC_HI_QI): Likewise.
19177 (TARGET_SYNC_TI): Likewise.
19178 (MASK_CRYPTO): Likewise.
19179 (MASK_DIRECT_MOVE): Likewise.
19180 (MASK_P8_FUSION): Likewise.
19181 (MASK_P8_VECTOR): Likewise.
19182 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
19183 temporary used by some of the direct move instructions to get two FP
19184 temporary registers does not force creation of a stack frame.
19185 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
19186 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
19187 that any VSX registers are tieable, even if they are also an
19188 Altivec vector mode.
19189 (r6000_reg_class_enum): Add wm, wr, wv constraints.
19190 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
19191 (RS6000_BTM_CRYPTO): Likewise.
19192 (RS6000_BTM_COMMON): Likewise.
19193
19194 * config/rs6000/rs6000.md (cpu attribute): Add power8.
19195 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
19196 (enum rs6000_vector): Add power8 vector support.
19197
19198 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19199
19200 PR target/19599
19201 PR target/57340
19202 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
19203 (any_sibcall_could_use_r3): this and handle indirect calls.
19204 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
19205
19206 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
19207
19208 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
19209
19210 2013-05-22 Richard Biener <rguenther@suse.de>
19211
19212 PR middle-end/57349
19213 * profile.c (branch_prob): Do not split blocks that are
19214 abnormally receiving from ECF_RETURNS_TWICE functions.
19215
19216 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
19217
19218 * recog.c (offsettable_address_addr_space_p): Fix calculation of
19219 address mode. Move pointer mode initialization to the same place.
19220
19221 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
19222
19223 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
19224 while it has any effect.
19225
19226 2013-05-21 Easwaran Raman <eraman@google.com>
19227
19228 PR tree-optimization/57322
19229 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
19230 UID of the statement added to the BB to be 1.
19231
19232 2013-05-21 Jakub Jelinek <jakub@redhat.com>
19233
19234 PR tree-optimization/57331
19235 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
19236 of conversion from pointer type to integral type with integer.
19237
19238 2013-05-21 Martin Jambor <mjambor@suse.cz>
19239
19240 PR lto/57289
19241 * ipa-prop.c (ipa_read_node_info): Process param_used and
19242 controlled_uses in the same order as when writing.
19243
19244 2013-05-21 Magnus Granberg <baldrick@free.fr>
19245
19246 PR plugins/56754
19247 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
19248
19249 2013-05-21 Richard Biener <rguenther@suse.de>
19250
19251 PR tree-optimization/57318
19252 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
19253 estimate stmts with side-effects as likely eliminated.
19254
19255 2013-05-21 Richard Biener <rguenther@suse.de>
19256
19257 PR tree-optimization/57330
19258 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
19259 preserve the call stmts fntype.
19260
19261 2013-05-21 Richard Biener <rguenther@suse.de>
19262
19263 PR tree-optimization/57303
19264 * tree-ssa-sink.c (statement_sink_location): Improve killing
19265 stmt detection and properly handle self-assignments.
19266
19267 2013-05-21 Christian Bruel <christian.bruel@st.com>
19268
19269 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
19270 spanning registers. LEAF_REG_REMAP is supported only for contiguous
19271 registers. Set register size out of the PARALLEL loop.
19272
19273 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
19274
19275 PR target/56547
19276 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
19277 (*fmasf4, *fmasf4_media): New insns.
19278
19279 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19280
19281 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
19282 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
19283 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
19284 (mips_idiv_insns): Update the comments to say that the returned
19285 instruction counts are in units of BASE_INSN_LENGTH.
19286 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
19287 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
19288 using 2 rather than 4 as the length of indirect MIPS16 and
19289 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
19290 length of a NOP. Don't divide MIPS16 lengths by 2.
19291 (mips16_split_long_branches): Assume a branch is long if the
19292 length is greater than 4 rather than 8.
19293 * config/mips/mips.md (length): Give MIPS16 lengths directly,
19294 rather than multiplying them by 2. Multiply instruction counts
19295 by BASE_INSN_LENGTH rather than 4.
19296 (*jump_mips16, tls_get_tp_mips16_<mode>)
19297 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
19298
19299 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19300
19301 * config/mips/mips.md (extended_mips16): Remove branch case.
19302 (length): Remove duplicated extended_mips16 test.
19303
19304 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
19305
19306 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
19307
19308 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
19309
19310 * recog.h (Recog_data): Rename to...
19311 (recog_data_d): ...this.
19312 (recog_data): Update accordingly.
19313 * recog.c (recog_data): Likewise.
19314 * reload.c (save_recog_data): Likewise.
19315 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
19316 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
19317
19318 2013-05-17 Julian Brown <julian@codesourcery.com>
19319
19320 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
19321 found in a REG_EQUAL note, invalidate it.
19322
19323 2013-05-17 Easwaran Raman <eraman@google.com>
19324
19325 * tree-ssa-reassoc.c (find_insert_point): New function.
19326 (insert_stmt_after): Likewise.
19327 (get_def_stmt): Likewise.
19328 (ensure_ops_are_available): Likewise.
19329 (not_dominated_by): Likewise.
19330 (rewrite_expr_tree): Do not move statements beyond what is
19331 necessary. Remove call to swap_ops_for_binary_stmt...
19332 (reassociate_bb): ... and move it here.
19333 (build_and_add_sum): Assign UIDs for new statements.
19334 (linearize_expr): Likewise.
19335 (do_reassoc): Renumber gimple statement UIDs.
19336
19337 2013-05-17 Jan Hubicka <jh@suse.cz>
19338
19339 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
19340 weakrefs.
19341 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
19342 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
19343 weakrefs.
19344 (output_weakrefs): Update.
19345
19346 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
19347 Martin Jambor <mjambor@suse.cz>
19348
19349 PR middle-end/57276
19350 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
19351 value that corresponds to the given aggval is found in values vector.
19352
19353 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
19354
19355 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
19356 sse, sse2, sse3, ssse3 and sse4a flags to options.
19357
19358 2013-05-17 David Malcolm <dmalcolm@redhat.com>
19359
19360 * gengtype-state.c: (s_expr_writer): New class, to handle
19361 prettifying of output layout of s-expressions.
19362 (state_writer): New class, to write out gtype.state.
19363 (state_written_type_count): Move this variable into member data of
19364 state_writer.
19365 (s_expr_writer::s_expr_writer): New code: constructor for new class
19366 (state_writer::state_writer(): ditto
19367 (s_expr_writer::write_new_line): New function
19368 (s_expr_writer::write_any_indent): ditto
19369 (s_expr_writer::begin_s_expr): ditto
19370 (s_expr_writer::end_s_expr): ditto
19371 (write_state_fileloc): convert to method of state_writer...
19372 (state_writer:: write_state_fileloc): ...and use methods of
19373 s_expr_writer to write indentation into the gtype.state output file
19374 to visually represent the hierarchical structure of the list
19375 structures
19376 (write_state_fields): ditto, renaming to...
19377 (state_writer::write_state_fields)
19378 (write_state_a_string): ditto, renaming to...
19379 (state_writer::write_state_a_string)
19380 (write_state_string_option): ditto, renaming to...
19381 (state_writer::write_state_string_option)
19382 (write_state_type_option): ditto, renaming to...
19383 (state_writer::write_state_type_option)
19384 (write_state_nested_option): ditto, renaming to...
19385 (state_writer::write_state_nested_option)
19386 (write_state_option): ditto, renaming to...
19387 (state_writer::write_state_option)
19388 (write_state_options): ditto, renaming to...
19389 (state_writer::write_state_options)
19390 (write_state_lang_bitmap): ditto, renaming to...
19391 (state_writer::write_state_lang_bitmap)
19392 (write_state_version): ditto, renaming to...
19393 (state_writer::write_state_version)
19394 (write_state_scalar_type): ditto, renaming to...
19395 (state_writer::write_state_scalar_type)
19396 (write_state_string_type): ditto, renaming to...
19397 (state_writer::write_state_string_type)
19398 (write_state_undefined_type): ditto, renaming to...
19399 (state_writer::write_state_undefined_type)
19400 (write_state_struct_union_type): ditto, renaming to...
19401 (state_writer::write_state_struct_union_type)
19402 (write_state_struct_type): ditto, renaming to...
19403 (state_writer::write_state_struct_type)
19404 (write_state_user_struct_type): ditto, renaming to...
19405 (state_writer::write_state_user_struct_type)
19406 (write_state_lang_struct_type): ditto, renaming to...
19407 (state_writer::write_state_lang_struct_type)
19408 (write_state_param_struct_type): ditto, renaming to...
19409 (state_writer::write_state_param_struct_type)
19410 (write_state_pointer_type): ditto, renaming to...
19411 (state_writer::write_state_pointer_type)
19412 (write_state_array_type): ditto, renaming to...
19413 (state_writer::write_state_array_type)
19414 (write_state_gc_used): ditto, renaming to...
19415 (state_writer::write_state_gc_used)
19416 (write_state_common_type_content): ditto, renaming to...
19417 (state_writer::write_state_common_type_content)
19418 (write_state_type): ditto, renaming to...
19419 (state_writer::write_state_type)
19420 (write_state_pair_list): ditto, renaming to...
19421 (state_writer::write_state_pair_list)
19422 (write_state_pair): ditto, renaming to...
19423 (state_writer::write_state_pair)
19424 (write_state_typedefs): ditto, renaming to...
19425 (state_writer::write_state_typedefs)
19426 (write_state_structures): ditto, renaming to...
19427 (state_writer::write_state_structures)
19428 (write_state_param_structs): ditto, renaming to...
19429 (state_writer::write_state_param_structs)
19430 (write_state_variables): ditto, renaming to...
19431 (state_writer::write_state_variables)
19432 (write_state_srcdir): ditto, renaming to...
19433 (state_writer::write_state_srcdir)
19434 (write_state_files_list): ditto, renaming to...
19435 (state_writer::write_state_files_list)
19436 (write_state_languages): ditto, renaming to...
19437 (state_writer::write_state_languages)
19438 (write_state): create a state_writer instance and use it when
19439 writing out the state file
19440
19441 2013-05-17 Mike Stump <mikestump@comcast.net>
19442
19443 PR rtl-optimization/57304
19444 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
19445 accessing DF_REF_REAL_LOC.
19446
19447 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19448
19449 PR rtl-optimization/57281
19450 PR rtl-optimization/57300
19451 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
19452 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
19453 what the other splitter did if the registers are dead.
19454
19455 2013-05-17 Richard Biener <rguenther@suse.de>
19456
19457 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
19458 MEM_REF offsets.
19459
19460 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19461
19462 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
19463 linking.
19464
19465 2013-05-17 Marek Polacek <polacek@redhat.com>
19466
19467 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
19468 length when doing non-zero store of storing '\0' to '\0'.
19469
19470 2013-05-17 Jakub Jelinek <jakub@redhat.com>
19471
19472 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
19473 vect_external_def oprnd1 with loop_vinfo, try to emit
19474 optional cast, negation and and stmts on the loop preheader
19475 edge instead of into the pattern def seq.
19476
19477 PR tree-optimization/57051
19478 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
19479 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
19480
19481 2013-05-16 Nick Clifton <nickc@redhat.com>
19482
19483 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
19484 (rl78_is_naked_func): New function.
19485 (rl78_expand_prologue): Skip prologue generation for naked functions.
19486 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
19487 * doc/extend.texi (naked): Add RL78 to the list of processors
19488 that supports this attribute.
19489
19490 2013-05-16 Jeff Law <law@redhat.com>
19491
19492 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
19493
19494 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19495
19496 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
19497 cache parameters using detect_caches_amd also for CYRIX,
19498 NSC and TM2 signatures.
19499
19500 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
19501 Dzianis Kahanovich <mahatma@eu.by>
19502
19503 PR target/45359
19504 PR target/46396
19505 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
19506 VIA/Centaur processors and determine their cache parameters
19507 using detect_caches_amd.
19508
19509 2013-05-16 Teresa Johnson <tejohnson@google.com>
19510
19511 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
19512 (rtl_verify_edges): New function.
19513 (rtl_verify_bb_insns): Ditto.
19514 (rtl_verify_bb_pointers): Ditto.
19515 (rtl_verify_bb_insn_chain): Ditto.
19516 (rtl_verify_fallthru): Ditto.
19517 (rtl_verify_bb_layout): Ditto.
19518 (rtl_verify_flow_info_1): Outline checks into new functions.
19519 (rtl_verify_flow_info): Ditto.
19520
19521 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
19522
19523 * cfghooks.c (copy_bbs): Add update_dominance argument.
19524 * cfghooks.h (copy_bbs): Update prototype.
19525 * tree-cfg.c (gimple_duplicate_sese_region):
19526 Add update_dominance argument.
19527 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
19528 * tree-ssa-loop-ch.c (copy_loop_headers): Update
19529 gimple_duplicate_sese_region call.
19530 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
19531 Update copy_bbs call.
19532 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
19533 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
19534
19535 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19536
19537 * tree-vectorizer.h (NUM_PATTERNS): Increment.
19538 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
19539 vect_recog_rotate_pattern.
19540 (vect_recog_rotate_pattern): New function.
19541
19542 2013-05-16 Jason Merrill <jason@redhat.com>
19543
19544 * Makefile.in (LLINKER): New variable.
19545 (mostlyclean): Remove link mutex.
19546 * configure.ac: Handle --enable-link-mutex.
19547 * lock-and-run.sh: New script.
19548
19549 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19550
19551 PR target/19599
19552 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
19553 for NULL decl.
19554
19555 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19556
19557 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
19558
19559 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
19560
19561 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
19562 * config/arm/arm.c (next_consecutive_mem): New function.
19563 (gen_movmem_ldrd_strd): Likewise.
19564 * config/arm/arm.md (movmemqi): Update condition and code.
19565 (unaligned_loaddi, unaligned_storedi): New patterns.
19566
19567 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19568
19569 * config.gcc: Obsolete *-*-solaris2.9*.
19570 * doc/install.texi (Specific, *-*-solaris2*): Document it.
19571
19572 2013-05-16 Richard Biener <rguenther@suse.de>
19573
19574 * passes.c (init_optimization_passes): Move pass_parallelize_loops
19575 earlier, after GRAPHITE transforms and IV canonicalization.
19576
19577 2013-05-16 Jakub Jelinek <jakub@redhat.com>
19578
19579 * omp-low.c (extract_omp_for_data): For collapsed loops,
19580 if at least one of the loops is known at compile time to
19581 iterate zero times, set count to 0.
19582 (expand_omp_regimplify_p): New function.
19583 (expand_omp_for_generic): For collapsed loops, if at least
19584 one of the loops isn't known to iterate at least once,
19585 add runtime check with setting count to 0.
19586 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
19587 For unsigned types if it isn't known at compile time that
19588 the loop will iterate at least once, add runtime check to bypass
19589 the whole loop if initial condition isn't true.
19590
19591 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
19592
19593 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
19594
19595 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
19596
19597 PR middle-end/57286
19598 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
19599 transformations to avoid an infinite loop.
19600
19601 2013-05-16 Marek Polacek <polacek@redhat.com>
19602
19603 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
19604
19605 2013-05-15 Leif Ekblad <leif@rdos.net>
19606
19607 * config/i386/i386.c (ix86_decompose_address): Use
19608 DEFAULT_TLS_SEG_REG to access TLS segment register.
19609 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
19610 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
19611 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
19612
19613 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
19614
19615 PR target/57260
19616 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
19617 sibling calls to functions that would normally be lazily bound,
19618 unless $gp is call-clobbered.
19619
19620 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
19621
19622 * config/i386/i386.c (ix86_option_override_internal): Update
19623 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
19624 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
19625 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
19626 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
19627 of TARGET_3DNOW.
19628 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
19629
19630 2013-05-15 Andreas Schwab <schwab@suse.de>
19631
19632 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
19633 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
19634 third operand.
19635
19636 2013-05-15 Teresa Johnson <tejohnson@google.com>
19637
19638 * loop-unroll.c (report_unroll_peel): Check decision before
19639 emitting unroll/peel message.
19640
19641 2013-05-15 Teresa Johnson <tejohnson@google.com>
19642
19643 * function.h (has_bb_partition): New rtl_data flag.
19644 (bb_reorder_complete): Ditto.
19645 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
19646 instead of flag_reorder_blocks_and_partition.
19647 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
19648 with some enhancements.
19649 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
19650 * bb-reorder.c (connect_traces): Check for has_bb_partition
19651 instead of flag_reorder_blocks_and_partition.
19652 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
19653 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
19654 verify_hot_cold_block_grouping.
19655 (partition_hot_cold_basic_blocks): Set has_bb_partition.
19656
19657 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
19658
19659 PR target/19599
19660 * config/arm/predicates.md (call_insn_operand): New predicate.
19661 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
19662 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
19663 if insn is not a tail call.
19664 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
19665 registers.
19666 * config/arm/arm.h (enum reg_class): New caller save register class.
19667 (REG_CLASS_NAMES): Likewise.
19668 (REG_CLASS_CONTENTS): Likewise.
19669 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
19670 without decls.
19671
19672 2013-05-15 Richard Biener <rguenther@suse.de>
19673
19674 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
19675 of MSG_OPTIMIZED_LOCATIONS.
19676 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
19677 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
19678 message.
19679 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
19680 of MSG_OPTIMIZED_LOCATIONS.
19681 (execute_vect_slp): Likewise.
19682 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
19683 (vect_create_cond_for_alias_checks): Likewise.
19684 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
19685 (vect_recog_widen_mult_pattern): Likewise.
19686 (vect_recog_widen_sum_pattern): Likewise.
19687 (vect_recog_over_widening_pattern): Likewise.
19688 (vect_recog_widen_shift_pattern): Likewise.
19689 (vect_recog_vector_vector_shift_pattern): Likewise.
19690 (vect_recog_divmod_pattern): Likewise.
19691 (vect_recog_mixed_size_cond_pattern): Likewise.
19692 (vect_recog_bool_pattern): Likewise.
19693 (vect_pattern_recog_1): Likewise.
19694
19695 2013-05-15 Martin Jambor <mjambor@suse.cz>
19696
19697 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
19698 non-functions to builtin_unreachable.
19699 * ipa-inline-transform.c (inline_call): Do not assert estimates were
19700 correct when new direct edges were discovered.
19701
19702 2013-05-15 Martin Jambor <mjambor@suse.cz>
19703
19704 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
19705 header, print symbol order instead of node uid, print more information
19706 about indirect edge targets.
19707 (ipa_make_edge_direct_to_target): Print symbol order instead of node
19708 uids.
19709 (ipa_make_edge_direct_to_target): Likewise.
19710 (remove_described_reference): Likewise.
19711 (propagate_controlled_uses): Likewise.
19712 (ipa_print_node_params): Also print symbol order.
19713 (ipcp_transform_function): Print symbol order instead of node uids.
19714 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
19715 (cgraph_get_create_real_symbol_node): Likewise.
19716 * ipa-cp.c (print_lattice): Likewise.
19717 (print_all_lattices): Likewise.
19718 (determine_versionability): Likewise.
19719 (initialize_node_lattices): Likewise.
19720 (estimate_local_effects): Likewise.
19721 (update_profiling_info): Likewise.
19722 (create_specialized_node): Likewise.
19723 (perhaps_add_new_callers): Likewise.
19724 (decide_about_value): Likewise.
19725 (decide_whether_version_node): Likewise.
19726 (identify_dead_nodes): Likewise.
19727 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
19728 (dump_inline_summary): Likewise.
19729 (estimate_node_size_and_time): Likewise.
19730 (inline_analyze_function): Likewise.
19731 * ipa-inline.c (report_inline_failed_reason): Likewise.
19732 (want_early_inline_function_p): Likewise.
19733 (edge_badness): Likewise.
19734 (update_edge_key): Likewise.
19735 (inline_small_functions): Likewise. Add dumping of order to two other
19736 dumps.
19737 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
19738 instead of node uids.
19739 (propagate_pure_const): Likewise.
19740 (propagate_pure_const): Likewise.
19741 * ipa-utils.c (dump_cgraph_node_set): Likewise.
19742 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
19743 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
19744 of node uids.
19745 * tree-pretty-print.c (dump_function_header): Likewise.
19746 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
19747 Print symbol order instead of node uids.
19748
19749 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19750
19751 * config/s390/s390.c (s390_register_move_cost): Don't impose the
19752 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
19753
19754 2013-05-15 Richard Biener <rguenther@suse.de>
19755
19756 PR tree-optimization/57275
19757 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
19758 return value for fail to do runtime alias checks for gather loads.
19759
19760 2013-05-15 Jan Hubicka <jh@suse.cz>
19761
19762 PR lto/57038
19763 PR lto/47375
19764 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
19765 weakrefs are not external.
19766 (lto_symtab_merge_decls): Fix thinko when dealing with
19767 non-lto_symtab decls.
19768 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
19769 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
19770 * varpool.c (dump_varpool_node): Dump more flags.
19771
19772 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19773
19774 * config/i386/i386.c (processor_alias_table): Add instruction
19775 FSGSBASE for AMD bdver3 architecture.
19776
19777 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19778
19779 * tree.c (warn_deprecated_use): Print file:line using locus color.
19780 * diagnostic.c (diagnostic_report_current_module): Print file:line
19781 and file:line:column using locus color.
19782
19783 2013-05-14 Mike Stump <mikestump@comcast.net>
19784
19785 * gdbinit.in: Add __null.
19786
19787 2013-05-14 Mike Stump <mikestump@comcast.net>
19788
19789 * recog.h: Rename struct recog_data to Recog_data.
19790 * recog.c: Likewise.
19791 * reload.c (can_reload_into): Likewise.
19792 * config/picochip/picochip.c: Likewise.
19793
19794 2013-05-14 Mike Stump <mikestump@comcast.net>
19795
19796 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
19797
19798 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
19799
19800 * resource.h (struct resources): Remove unch_memory member.
19801 (CLEAR_RESOURCE): Don't clear unch_memory.
19802 * resource.c (mark_referenced_resources): Don't set it.
19803 (mark_set_resources): Likewise.
19804 (mark_target_live_regs): Don't clear it.
19805 (init_resource_info): Likewise.
19806 * reorg.c (resource_conflicts_p): Don't compare it.
19807 (redundant_insn): Don't set it.
19808
19809 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
19810 Remove prototypes.
19811 * emit-rtl.c (next_label): Remove unused function.
19812 (skip_consecutive_labels, link_cc0_insns): Move to ...
19813 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
19814 only place where these functions are used, and make them static.
19815
19816 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19817
19818 * fold-const.c (fold_negate_expr): Handle vectors.
19819 (fold_truth_not_expr): Make it static.
19820 (fold_invert_truthvalue): New static function.
19821 (invert_truthvalue_loc): Handle vectors. Do not call
19822 fold_truth_not_expr directly.
19823 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
19824 <TRUTH_NOT_EXPR>: Do not cast to boolean.
19825 (fold_comparison): Handle vector constants.
19826 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
19827 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
19828 * tree.h (fold_truth_not_expr): Remove declaration.
19829
19830 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
19831
19832 * config/aarch64/aarch64-simd.md
19833 (aarch64_vcond_internal<mode>): Rename to...
19834 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
19835 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
19836 float modes. Clarify all iterator modes.
19837 (vcond<mode><mode>): Use new name for vcond expanders.
19838 (vcond<v_cmp_result><mode>): Likewise.
19839 (vcondu<mode><mode>: Likewise.
19840 * config/aarch64/iterators.md (VDQF_COND): New.
19841
19842 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
19843
19844 PR bootstrap/57266
19845 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
19846 variable for the shift amount. Check that we shift by non-negative
19847 amounts.
19848
19849 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
19850
19851 PR target/42017
19852 * config/arm/arm.h (EPILOGUE_USES): Only return true
19853 for LR_REGNUM after epilogue_completed.
19854
19855 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
19856
19857 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
19858 is error_mark_node.
19859
19860 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19861
19862 PR target/57261
19863 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
19864 and Solaris 11+/x86 with gld.
19865 * configure: Regenerate.
19866
19867 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19868
19869 * expmed.c (expand_shift_1): Canonicalize rotates by
19870 constant bitsize / 2 to bitsize - 1.
19871 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
19872 case ROTATERT>: Likewise.
19873
19874 Revert:
19875 2013-05-10 Jakub Jelinek <jakub@redhat.com>
19876
19877 * config/i386/i386.md (rotateinv): New code attr.
19878 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
19879 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
19880 roll $31, %eax, etc.
19881
19882 2013-05-14 Richard Biener <rguenther@suse.de>
19883
19884 PR middle-end/57235
19885 * tree-eh.c (sink_clobbers): Give up for successors with
19886 multiple predecessors and no virtual uses.
19887
19888 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
19889
19890 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
19891 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
19892
19893 2013-05-14 Jakub Jelinek <jakub@redhat.com>
19894
19895 PR middle-end/57251
19896 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
19897 the case when both op0 and op1 have VOIDmode.
19898
19899 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
19900
19901 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
19902 in multiply-accumulate mode.
19903
19904 2013-05-13 Guozhi Wei <carrot@google.com>
19905
19906 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
19907
19908 2013-05-13 Kai Tietz <ktietz@redhat.com>
19909
19910 PR target/56975
19911 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
19912 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
19913 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
19914 * config/i386/i386.c (ix86_option_override_internal): Likewise.
19915 (ix86_expand_prologue): Likewise.
19916 (ix86_expand_split_stack_prologue): Likewise.
19917 (legitimate_pic_address_disp_p): Likewise.
19918 (legitimize_pic_address): Likewise.
19919 (legitimize_tls_address): Likewise.
19920 (legitimize_pe_coff_symbol): Likewise.
19921 (output_pic_addr_const): Likewise.
19922 (construct_plt_address): Likewise.
19923 (ix86_expand_call): Likewise.
19924 (x86_output_mi_thunk): Likewise.
19925 (x86_function_profiler): Likewise.
19926
19927 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
19928
19929 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
19930 similar switch cases.
19931 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
19932 (aarch64_simd_mov_to_<mode>low): Delete.
19933 (aarch64_simd_mov_to_<mode>high): Delete.
19934 (move_lo_quad_<mode>): Add w<-r alternative.
19935 (aarch64_simd_move_hi_quad_<mode>): Likewise.
19936 (aarch64_simd_mov_from_*): Update type attribute.
19937 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
19938 statement.
19939
19940 2013-05-13 Jan Hubicka <jh@suse.cz>
19941
19942 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
19943 * config/i386/i386.c (ix86_compute_frame_layout,
19944 ix86_expand_epilogue, emit_i387_cw_initialization,
19945 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
19946 ix86_local_alignment): Fix use of size/speed predicates.
19947
19948 2013-05-13 Jakub Jelinek <jakub@redhat.com>
19949
19950 PR tree-optimization/45216
19951 PR tree-optimization/57157
19952 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
19953 the (-Y) & (B - 1) variant if OP is |.
19954 * expmed.c (expand_shift_1): For rotations by const0_rtx just
19955 return shifted. Use (-op1) & (prec - 1) as other_amount
19956 instead of prec - op1.
19957
19958 2013-05-13 Martin Jambor <mjambor@suse.cz>
19959
19960 PR middle-end/42371
19961 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
19962 (ipa_constant_data): New type.
19963 (ipa_jump_func): Use ipa_constant_data to hold information about
19964 constant jump functions.
19965 (ipa_get_jf_constant): Adjust to jump function type changes.
19966 (ipa_get_jf_constant_rdesc): New function.
19967 (ipa_param_descriptor): New field controlled_uses.
19968 (ipa_get_controlled_uses): New function.
19969 (ipa_set_controlled_uses): Likewise.
19970 * ipa-ref.h (ipa_find_reference): Declare.
19971 * ipa-prop.c (ipa_cst_ref_desc): New type.
19972 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
19973 changes.
19974 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
19975 New parameter cs. Adjust all callers.
19976 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
19977 (remove_described_reference): New function.
19978 (jfunc_rdesc_usable): Likewise.
19979 (try_make_edge_direct_simple_call): Decrement controlled use count,
19980 attempt to remove reference if it hits zero.
19981 (combine_controlled_uses_counters): New function.
19982 (propagate_controlled_uses): Likewise.
19983 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
19984 (ipa_edge_duplication_hook): Duplicate reference descriptions.
19985 (ipa_print_node_params): Print described use counter.
19986 (ipa_write_jump_function): Adjust to jump function type changes.
19987 (ipa_read_jump_function): New parameter CS, pass it to
19988 ipa_set_jf_constant. Adjust caller.
19989 (ipa_write_node_info): Stream controlled use count
19990 (ipa_read_node_info): Likewise.
19991 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
19992 asserting.
19993 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
19994 count. Remove cloning-added reference if it reaches zero.
19995 * ipa-ref.c (ipa_find_reference): New function.
19996
19997 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
19998
19999 * config/i386/i386.c (processor_target_table): Modified default
20000 alignment values for AMD BD and BT architectures.
20001
20002 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
20003
20004 * tree-vect-generic.c (uniform_vector_p): Move ...
20005 * tree.c (uniform_vector_p): ... here.
20006 * tree.h (uniform_vector_p): Declare it.
20007 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
20008 into a scalar.
20009
20010 2013-05-13 Jakub Jelinek <jakub@redhat.com>
20011
20012 PR tree-optimization/57230
20013 * tree-ssa-strlen.c (handle_char_store): Record length for
20014 array store from STRING_CST.
20015
20016 PR tree-optimization/57230
20017 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
20018 check.
20019
20020 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
20021
20022 * config/epiphany/epiphany.c (epiphany_init): Check size of
20023 NUM_MODES_FOR_MODE_SWITCHING.
20024 (epiphany_expand_prologue):
20025 Remove CONFIG_REGNUM initial value handling code.
20026 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
20027 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
20028 (emit_set_fp_mode, epiphany_mode_after): Likewise.
20029 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
20030 Don't return 1 for FP_MODE_NONE.
20031 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
20032 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
20033 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
20034 * config/epiphany/epiphany.md (save_config): New pattern.
20035
20036 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
20037
20038 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
20039
20040 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
20041
20042 * config/i386/i386.md (memory): Handle sseishft1.
20043 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
20044 (*vec_extractv2di_1): Ditto.
20045
20046 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
20047
20048 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
20049 saved registers.
20050
20051 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
20052
20053 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
20054 Add mthumb/march=armv7-a multilib.
20055 Add mthumb/march=armv7-r multilib.
20056 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
20057
20058 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
20059
20060 * config/v850/t-rtems: Add more multilibs.
20061
20062 2013-05-10 Richard Biener <rguenther@suse.de>
20063
20064 PR tree-optimization/57214
20065 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
20066 not propagate from SSA names that occur in abnormal PHI nodes.
20067
20068 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
20069
20070 * stor-layout.c (element_precision): New function.
20071 * machmode.h (element_precision): Declare it.
20072 * tree.c (build_minus_one_cst): New function.
20073 (element_precision): Likewise.
20074 * tree.h (build_minus_one_cst): Declare new function.
20075 (element_precision): Likewise.
20076 * fold-const.c (operand_equal_p): Use element_precision.
20077 (fold_binary_loc): Handle vector types.
20078 * convert.c (convert_to_integer): Use element_precision.
20079 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
20080 separately.
20081
20082 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
20083
20084 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
20085 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
20086 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
20087 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20088 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
20089 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
20090 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
20091 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
20092 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
20093 (Uuw8): New constraints.
20094 (Usb4): Move into alphabetical order.
20095 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
20096 (sd8_operand, ub8_operand, uw8_operand): New predicates.
20097 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
20098 previously unnamed patterns.
20099 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
20100 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
20101 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
20102 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
20103 of set_attr_alternative/if_then_else. Use extended_mips16 instead
20104 of specific lengths.
20105
20106 2013-05-10 Jakub Jelinek <jakub@redhat.com>
20107
20108 * config/i386/i386.md (rotateinv): New code attr.
20109 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
20110 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
20111 roll $31, %eax, etc.
20112
20113 PR tree-optimization/45216
20114 PR tree-optimization/57157
20115 * tree-ssa-forwprop.c (simplify_rotate): New function.
20116 (ssa_forward_propagate_and_combine): Call it.
20117
20118 2013-05-10 Richard Biener <rguenther@suse.de>
20119
20120 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
20121 disable peeling when we version for aliasing.
20122 (vector_alignment_reachable_p): Honor explicit user alignment.
20123 (vect_supportable_dr_alignment): Likewise.
20124 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
20125 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
20126 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
20127 then peeling to arrange for the cost-model check to come first.
20128
20129 2013-05-10 Alan Modra <amodra@gmail.com>
20130
20131 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
20132 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
20133 * configure: Regenerate.
20134
20135 2013-05-10 Alan Modra <amodra@gmail.com>
20136
20137 PR target/55033
20138 * varasm.c (default_elf_select_section): Move !DECL_P check..
20139 (get_named_section): ..to here before calling get_section_name.
20140 Adjust assertion.
20141 (default_section_type_flags): Add DECL_P check.
20142 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
20143 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
20144
20145 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
20146
20147 * config/epiphany/epiphany.c (epiphany_expand_prologue):
20148 When using gen_stack_adjust_str with a register offset, add a
20149 REG_FRAME_RELATED_EXPR note.
20150
20151 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
20152
20153 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
20154 (*vec_extractv4si_zext_mem): Ditto.
20155 (*vec_extractv2di): Add 0->x and x->x alternatives.
20156 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
20157 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
20158
20159 2013-05-09 Jason Merrill <jason@redhat.com>
20160
20161 N3639 C++1y VLA support
20162 * gimplify.c (gimplify_vla_decl): Don't touch an existing
20163 DECL_VALUE_EXPR.
20164
20165 * tree.c (build_constructor_va): New.
20166 * tree.h: Declare it.
20167
20168 2013-05-09 Martin Jambor <mjambor@suse.cz>
20169
20170 PR lto/57084
20171 * gimple-fold.c (canonicalize_constructor_val): Call
20172 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
20173
20174 2013-05-09 Jan Hubicka <jh@suse.cz>
20175 Richard Biener <rguenther@suse.de>
20176
20177 PR lto/54095
20178 * symtab.c (symtab_make_decl_local): Do not add private names.
20179
20180 2013-05-09 Jan Hubicka <jh@suse.cz>
20181
20182 PR lto/54095
20183 * symtab.c (insert_to_assembler_name_hash): Handle clones.
20184 (unlink_from_assembler_name_hash): Likewise.
20185 (symtab_prevail_in_asm_name_hash, symtab_register_node,
20186 symtab_unregister_node, symtab_initialize_asm_name_hash,
20187 change_decl_assembler_name): Update.
20188
20189 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
20190
20191 * config/aarch64/aarch64.md: New movtf split.
20192 (*movtf_aarch64): Update.
20193 (aarch64_movdi_tilow): Handle TF modes and rename to
20194 aarch64_movdi_<mode>low.
20195 (aarch64_movdi_tihigh): Handle TF modes and rename to
20196 aarch64_movdi_<mode>high
20197 (aarch64_movtihigh_di): Handle TF modes and rename to
20198 aarch64_mov<mode>high_di
20199 (aarch64_movtilow_di): Handle TF modes and rename to
20200 aarch64_mov<mode>low_di
20201 (aarch64_movtilow_tilow): Remove spurious whitespace.
20202 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
20203 splits.
20204 (aarch64_print_operand): Update.
20205
20206 2013-05-09 Alan Modra <amodra@gmail.com>
20207
20208 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
20209 powerpc64le.
20210 * configure: Regenerate.
20211
20212 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20213
20214 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
20215 splitter preparation statements.
20216 * config/i386/sse.md (*vec_extract* splitters): Ditto.
20217 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
20218 adjust_address_nv.
20219
20220 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20221
20222 * gimple-ssa-strength-reduction.c (count_candidates): Change
20223 return value to int.
20224 (analyze_candidates_and_replace): Change type of length to int.
20225
20226 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
20227
20228 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
20229 (*vec_extract<mode>): Use VI12_128 mode iterator.
20230 (*vec_extract<mode>_mem): Ditto.
20231 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
20232 attribute.
20233
20234 2013-05-08 Diego Novillo <dnovillo@google.com>
20235
20236 PR bootstrap/54659
20237
20238 Revert:
20239 2012-08-17 Diego Novillo <dnovillo@google.com>
20240
20241 PR bootstrap/54281
20242 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
20243 * config.in: Regenerate.
20244 * configure: Regenerate.
20245 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
20246
20247 2013-05-08 Jan Hubicka <jh@suse.cz>
20248
20249 PR lto/54095
20250 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
20251 * cgraph.h (symtab_node_base): Add unique_name.
20252 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
20253 input_overwrite_node, input_varpool_node): Stream unique_name.
20254 * cgraphclones.c (cgraph_create_virtual_clone,
20255 cgraph_function_versioning): Set unique_name.
20256 * ipa.c (function_and_variable_visibility): Set unique_name.
20257
20258 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20259
20260 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
20261 (alloc_cand_and_find_basis): Restrict conditional candidate
20262 processing to CAND_MULTs.
20263
20264 2013-05-08 Jan Hubicka <jh@suse.cz>
20265
20266 PR lto/54095
20267 lto-symtab.c (lto_symtab_symbol_p): New function.
20268 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
20269 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
20270 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
20271 Skip static symbols.
20272
20273 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
20274
20275 PR tree-optimization/57200
20276 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
20277 Only call inform if the preceding warning_at returns true.
20278
20279 2013-05-07 Han Shen <shenhan@google.com>
20280
20281 * cfgexpand.c (record_or_union_type_has_array_p): New function.
20282 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
20283 * common.opt (fstack-protector-strong): New option.
20284 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
20285 * doc/invoke.texi (Optimization Options): Document
20286 "-fstack-protector-strong".
20287 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
20288
20289 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20290
20291 * config/mips/mips.c (mips_machine_reorg2): Return 0.
20292
20293 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
20294
20295 * ira.c (update_equiv_regs): Add insn having equiv memory even if
20296 it is not lhs of the insn.
20297 (setup_reg_equiv): Remove insn having equiv memory which it is not
20298 lhs of the insn.
20299 * lra-constraints.c (process_address): Try to improve generation
20300 code for address base + disp.
20301 (lra_constraints): Make correct the code for checking insn setting
20302 up backward equivalence. Remove insn only if it is in the init
20303 insn list.
20304 * lra-eliminations.c (update_reg_eliminate): Change return value.
20305 (lra_eliminate): Use the result.
20306
20307 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
20308
20309 * config/i386/sse.md (ssescalarnummask): New mode attribute.
20310 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
20311 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
20312 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
20313 register target operands.
20314 (*vec_extractv8hi_sse2): New pattern.
20315 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
20316 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
20317 (*vec_extract<mode>_mem): New insn and split pattern.
20318
20319 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
20320
20321 * config/arm/arm.c (arm_asan_shadow_offset): New function.
20322 (TARGET_ASAN_SHADOW_OFFSET): Define.
20323 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
20324 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
20325
20326 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20327
20328 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
20329 (incr_vec_index): Return -1 if increment not found.
20330 (create_add_on_incoming_edge): Assert if increment not found.
20331 (record_increment): Limit number of increments recorded.
20332 (all_phi_incrs_profitable): Return false if an increment not found.
20333 (replace_profitable_candidates): Don't process increments that were
20334 not recorded.
20335 (analyze_candidates_and_replace): Limit size of incr_vec.
20336
20337 2013-05-07 Richard Biener <rguenther@suse.de>
20338
20339 * calls.c (special_function_p): setjmp-like functions are leaf.
20340 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
20341 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
20342
20343 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
20344
20345 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
20346 (aarch64_simd_mov<mode>): New expander.
20347 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
20348 (aarch64_simd_mov_to_<mode>high): Likewise.
20349 (aarch64_simd_mov_from_<mode>low): Likewise.
20350 (aarch64_simd_mov_from_<mode>high): Likewise.
20351 (aarch64_dup_lane<mode>): Update.
20352 (aarch64_dup_lanedi): New instruction pattern.
20353 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
20354 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
20355
20356 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20357
20358 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
20359 (replace_mult_candidate): Remove unnecessary argument; remove
20360 unnecessary parameter from call to introduce_cast_before_cand.
20361 (replace_unconditional_candidate): Remove unnecessary parameter
20362 from call to replace_mult_candidate.
20363 (replace_conditional_candidate): Likewise.
20364 (insert_initializers): Use make_temp_ssa_name.
20365 (introduce_cast_before_cand): Remove unnecessary argument; use
20366 make_temp_ssa_name.
20367 (replace_one_candidate): Remove unnecessary argument; remove
20368 unnecessary parameter from calls to introduce_cast_before_cand.
20369 (replace_profitable_candidates): Remove unnecessary parameters
20370 from calls to replace_one_candidate.
20371
20372 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20373
20374 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
20375 phi def as possibly hiding a basis for a CAND_ADD whose operands
20376 have been commuted in the analysis.
20377 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
20378
20379 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20380
20381 * config/aarch64/aarch64.md
20382 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
20383 shift value between 0-4.
20384
20385 2013-05-07 Richard Biener <rguenther@suse.de>
20386
20387 * double-int.h (rshift): New overload.
20388 * double-int.c (rshift): New function.
20389 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
20390 (create_reference_ops_from_ref): Remove.
20391 (vn_reference_insert): Use shared ops for constructing the
20392 reference and copy it.
20393
20394 2013-05-07 Richard Biener <rguenther@suse.de>
20395
20396 PR middle-end/57190
20397 * tree-eh.c (sink_clobbers): Properly propagate
20398 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
20399
20400 2013-05-07 Jakub Jelinek <jakub@redhat.com>
20401
20402 PR tree-optimization/57149
20403 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
20404 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
20405 collect_phi_def_edges, execute_late_warn_uninitialized): Use
20406 uninit_undefined_value_p instead of ssa_undefined_value_p.
20407
20408 PR debug/57184
20409 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
20410 for modifier == EXPAND_INITIALIZER.
20411
20412 2013-05-07 Anton Blanchard <anton@samba.org>
20413
20414 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
20415 for powerpc64 little endian.
20416 * configure: Regenerate.
20417
20418 2013-05-06 Graham Stott <grahams@btinternet.com>
20419
20420 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
20421 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
20422 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
20423 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
20424
20425 2013-05-06 Graham Stott <grahams@btinternet.com>
20426
20427 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
20428 codes which allow non-lvalues.
20429
20430 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
20431
20432 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
20433 components are all 1s.
20434 (integer_minus_onep): New function.
20435 * tree.h (integer_minus_onep): Declare it.
20436 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
20437 integer_minus_onep instead of integer_all_onesp.
20438
20439 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20440
20441 PR target/52933
20442 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
20443 variations of these patterns.
20444
20445 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20446
20447 * config/i386/i386.md (isa): Add x64_sse4 member.
20448 (enabled): Handle x64_sse4.
20449 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
20450 instruction for 64bit SSE4_1 targets. Update insn attributes.
20451 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
20452 instruction for SSE4_1 targets. Update insn attributes.
20453 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20454 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
20455 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
20456 const_1 selector.
20457 (*vec_extractv4si): Rename from *sse4_1_pextrd.
20458 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
20459 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
20460
20461 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20462
20463 PR target/57108
20464 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
20465
20466 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
20467
20468 * final.c (do_assembler_dialects): Don't handle curly braces and
20469 vertical bar escaped by % as dialect delimiters.
20470 (output_asm_insn): Print curly braces and vertical bar if escaped
20471 by % and ASSEMBLER_DIALECT defined.
20472 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
20473 * doc/tm.texi: Regenerated.
20474
20475 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
20476
20477 * config/mips/mips.c: Include tree-pass.h.
20478 (mips_reorg): Split in pre- and post-dbr_schedule parts.
20479 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
20480 (pass_mips_machine_reorg2): New machine specific pass.
20481 (insert_pass_mips_machine_reorg2): New pass plugin definition.
20482 (mips_option_override): Register the new pass.
20483 * rtl.h (cleanup_barriers): Remove prototype.
20484 (dbr_schedule): Likewise.
20485 * jump.c (cleanup_barriers): Make static.
20486 * reorg.c (dbr_schedule): Likewise.
20487
20488 2013-05-06 Richard Biener <rguenther@suse.de>
20489
20490 PR tree-optimization/57185
20491 * tree-parloops.c (add_field_for_reduction): Handle anonymous
20492 SSA names properly.
20493
20494 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
20495
20496 PR target/57106
20497 * config/i386/i386.c (add_parameter_dependencies): Add dependence
20498 between "first_arg" and "insn", not "last" and "insn".
20499
20500 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20501
20502 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
20503 (find_candidates_in_block): Re-enable slsr_process_phi.
20504 (create_phi_basis): Fix double counting of candidate adjustment.
20505
20506 2013-05-06 Richard Biener <rguenther@suse.de>
20507
20508 PR middle-end/57147
20509 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
20510 the edge is also fallthru, preserve it and just clear the
20511 abnormal flag.
20512 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
20513 also complex, preserve that and just clear the fallthru flag.
20514 * tree-inline.c (update_ssa_across_abnormal_edges): Also
20515 update virtual operands.
20516
20517 2013-05-06 Alan Modra <amodra@gmail.com>
20518
20519 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
20520 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
20521 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
20522 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
20523 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20524 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
20525
20526 2013-05-06 Alan Modra <amodra@gmail.com>
20527
20528 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
20529 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
20530 (DEFAULT_ASM_ENDIAN): Define.
20531 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
20532 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
20533 Update -K PIC clause from sysv4.h.
20534 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
20535 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
20536
20537 2013-05-06 Alan Modra <amodra@gmail.com>
20538
20539 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
20540 twice for little-endian.
20541 (ashrdi3_no_power, ashrdi3): Support little-endian.
20542
20543 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
20544
20545 PR target/55303
20546 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
20547 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
20548 related expanders.
20549 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
20550 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
20551 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
20552 New predicates.
20553
20554 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
20555 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20556
20557 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
20558 * config/pa/pa.opt: Make mbig-switch a no-op.
20559 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
20560 (CASE_VECTOR_MODE): Always return SImode.
20561 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
20562 for the !TARGET_BIG_SWITCH case.
20563 * config/pa/pa-linux.h: Likewise.
20564 * config/pa/pa-openbsd.h: Likewise.
20565 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
20566 * config/pa/pa.md (short_jump): Remove define_insn.
20567 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
20568 (casesi0): Remove define_insn.
20569 (type): Remove btable_branch.
20570 (pa_combine_type): Likewise.
20571 (in_nullified_branch_delay): Likewise.
20572 (in_call_delay): Likewise.
20573 (define_delay): Likewise.
20574 (define_insn_reservation "Z3"): Likewise.
20575 (define_insn_reservation "Z4"): Likewise.
20576 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
20577 (pa_adjust_insn_length): Remove adjustment for btable branches.
20578 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
20579 and mno-big-switch
20580
20581 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
20582
20583 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
20584 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
20585 Add m->r,x alternatives.
20586 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
20587 splitters using SWI48x mode iterator.
20588 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
20589 TARGET_64BIT. Add m->x alternative.
20590 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
20591 Add o->x alternative. Enable for TARGET_SSE.
20592 (sse_storeq): Remove expander.
20593 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
20594 with memory input operand.
20595 (*vec_extractv2di_1 splitter): New.
20596 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
20597 * config/i386/i386.md (ssevecmodelower): New mode attribute.
20598
20599 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20600
20601 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
20602 (INT_LOWPART): Delete.
20603 (extract_MB): Adjust.
20604 (extract_ME): Adjust.
20605 (print_operand): Adjust.
20606
20607 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
20608
20609 * config/rs6000/predicates.md (reg_or_add_cint_operand,
20610 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20611 (reg_or_logical_cint_operand, easy_fp_constant,
20612 logical_const_operand): Delete "CONST_DOUBLE" case.
20613 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
20614 "HOST_BITS_PER_WIDE_INT == 64" test.
20615 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
20616 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
20617 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
20618 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
20619 test.
20620 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
20621 CONST_DOUBLE DImode/VOIDmode case.
20622 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
20623 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20624 CONST_DOUBLE VOIDmode case.
20625 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
20626 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
20627 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
20628 Delete CONST_DOUBLE case.
20629 (splitters for mov FMOVE64 const_double): Delete
20630 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
20631 "HOST_BITS_PER_WIDE_INT >= 64" test.
20632 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
20633 case.
20634 (mov DI const_double): Delete.
20635
20636 2013-05-04 Jakub Jelinek <jakub@redhat.com>
20637
20638 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
20639 on op shows all bits zero in mode of a lowpart subreg, return zero.
20640
20641 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
20642
20643 PR target/57150
20644 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
20645 to save TFmode registers and DImode to save TImode registers for
20646 caller save operations.
20647 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
20648 mark being partially clobbered since they only use the first
20649 double word.
20650
20651 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
20652 and TDmode only use the upper 64-bits of each VSX register.
20653
20654 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20655
20656 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
20657 (find_candidates_in_block): Disable slsr_process_phi.
20658
20659 2013-05-03 Guozhi Wei <carrot@google.com>
20660
20661 * coverage.c (coverage_obj_init): Move the construction of gcov
20662 constructor to ...
20663 (build_init_ctor): ... here.
20664
20665 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20666
20667 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
20668 (slsr_cand_d): Redefine def_phi.
20669 (stride_status, phi_adjust_status, count_phis_status): New enums.
20670 (find_phi_def): New.
20671 (find_basis_for_base_expr): New.
20672 (find_basis_for_candidate): Handle hidden bases.
20673 (alloc_cand_and_find_basis): Handle phi candidates.
20674 (slsr_process_phi): New.
20675 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
20676 (create_mul_imm_cand): Likewise.
20677 (create_add_ssa_cand): Exclude phi base candidates.
20678 (create_add_imm_cand): Likewise.
20679 (slsr_process_cast): Likewise.
20680 (slsr_process_copy): Likewise.
20681 (find_candidates_in_block): Handle phi candidates.
20682 (dump_candidate): Likewise.
20683 (unconditional_cands): Delete.
20684 (unconditional_cands_with_known_stride_p): Delete.
20685 (phi_dependent_cand_p): New.
20686 (cand_increment): Handle phi-dependent candidates.
20687 (replace_dependent): Delete.
20688 (replace_mult_candidate): New.
20689 (replace_unconditional_candidate): New.
20690 (incr_vec_index): Move to avoid forward reference.
20691 (create_add_on_incoming_edge): New.
20692 (create_phi_basis): New.
20693 (replace_dependents): Delete.
20694 (replace_conditional_candidate): New.
20695 (phi_add_costs): New.
20696 (replace_uncond_cands_and_profitable_phis): New.
20697 (record_increment): Handle phi adjustments.
20698 (record_phi_increments): New.
20699 (record_increments): Handle phi adjustments.
20700 (phi_incr_cost): New.
20701 (lowest_cost_path): Handle phis.
20702 (total_savings): Likewise.
20703 (analyze_increments): Likewise.
20704 (ncd_with_phi): New.
20705 (ncd_of_cand_and_phis): New.
20706 (nearest_common_dominator_for_cands): Handle phi increments.
20707 (all_phi_incrs_profitable): New.
20708 (replace_profitable_candidates): Handle phi-dependent candidates.
20709 (analyze_candidates_and_replace): Likewise.
20710
20711 2013-05-03 Teresa Johnson <tejohnson@google.com>
20712
20713 PR bootstrap/57154
20714 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
20715 do not exceed REG_BR_PROB_BASE.
20716
20717 2013-05-03 Jeff Law <law@redhat.com>
20718
20719 PR tree-optimization/57144
20720 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
20721 operand of the condition will bit into the new type when eliminating
20722 a cast feeding a condition.
20723
20724 2013-05-03 Jakub Jelinek <jakub@redhat.com>
20725
20726 PR rtl-optimization/57130
20727 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
20728 of COMPARE as in_code to the recursive call if needed.
20729
20730 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
20731
20732 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
20733 (enabled): Handle new members.
20734 * config/i386/sse.md (*vec_concatv2si): Merge from
20735 *vec_concatv2si_sse2 and vec_concatv2si_sse.
20736 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
20737
20738 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
20739
20740 PR tree-optimization/57027
20741 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
20742 for fnms opportunity, check we got the prerequisite kind
20743 of tree / gimple before using accessor functions.
20744
20745 2013-05-03 Richard Biener <rguenther@suse.de>
20746
20747 * double-int.h (lshift): New overload without precision
20748 and arith argument.
20749 (operator *=, operator +=, operator -=): Move ...
20750 * double-int.c (operator *=, operator +=, operator -=): ... here
20751 and implement more efficiently.
20752 (mul_double_with_sign): Remove.
20753 (lshift_double): Adjust to take unsinged shift argument, push
20754 dispatching code to callers.
20755 (mul_double_wide_with_sign): Add early out for callers that
20756 are not interested in high parts or overflow.
20757 (lshift): New function.
20758 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
20759 dispatch code here.
20760 (lrotate, rrotate): Use logical shifts.
20761 * expr.c (get_inner_reference): Use lshift.
20762 * fixed-value.c (do_fixed_divide): Likewise.
20763 * tree-dfa.c (get_ref_base_and_extent): Likewise.
20764 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
20765 (indirect_refs_may_alias_p): Likewise.
20766 (stmt_kills_ref_p_1): Likewise.
20767
20768 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20769
20770 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
20771
20772 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
20773
20774 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
20775 scalar form of FABD instruction.
20776
20777 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
20778
20779 * lra-constraints.c (process_alt_operands): Add checking alt
20780 number to choose the best alternative.
20781
20782 2013-05-02 Richard Biener <rguenther@suse.de>
20783
20784 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
20785 bitmap and its handling.
20786 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
20787
20788 2013-05-02 Richard Biener <rguenther@suse.de>
20789
20790 PR middle-end/57140
20791 * tree-inline.c (copy_loops): Properly handle removed loops.
20792 (copy_cfg_body): Mark destination loops for fixup if source
20793 loops needed fixup.
20794
20795 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
20796
20797 PR target/56732
20798 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
20799 generating simple_return for naked functions.
20800
20801 2013-05-02 Martin Jambor <mjambor@suse.cz>
20802
20803 PR middle-end/56988
20804 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
20805 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
20806 flags match.
20807 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
20808 ipa_agg_replacement_value structures.
20809 (known_aggs_to_agg_replacement_list): Likewise.
20810 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
20811 (read_agg_replacement_chain): Likewise.
20812 (ipcp_transform_function): Also check that by_ref flags match.
20813
20814 2013-05-02 Richard Biener <rguenther@suse.de>
20815
20816 * graphds.h (struct graph): Add obstack member.
20817 * graphds.c (new_graph): Initialize obstack and allocate
20818 vertices from it.
20819 (add_edge): Allocate edge from the obstack.
20820 (free_graph): Free the obstack instead of all edges and vertices.
20821
20822 2013-05-02 Teresa Johnson <tejohnson@google.com>
20823
20824 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
20825 divides.
20826 * cfg.c (update_bb_profile_for_threading): Ditto.
20827 * tree-inline.c (copy_bb): Ditto.
20828 (copy_edges_for_bb): Ditto.
20829 (initialize_cfun): Ditto.
20830 (copy_cfg_body): Ditto.
20831 (expand_call_inline): Ditto.
20832 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
20833 (estimate_node_size_and_time): Ditto.
20834 (inline_merge_summary): Ditto.
20835 * cgraphclones.c (cgraph_clone_edge): Ditto.
20836 (cgraph_clone_node): Ditto.
20837 * sched-rgn.c (compute_dom_prob_ps): Ditto.
20838 (compute_trg_info): Ditto.
20839
20840 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20841
20842 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
20843 S reg when fp attribute set.
20844 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
20845
20846 2013-05-02 Ian Bolton <ian.bolton@arm.com>
20847
20848 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
20849 New pattern.
20850 (*and_one_cmplsi3_compare0_uxtw): Likewise.
20851 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
20852 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
20853
20854 2013-05-02 Richard Biener <rguenther@suse.de>
20855
20856 * tree-scalar-evolution.c (scev_info_hasher): Remove.
20857 (struct instantiate_cache_entry): New type.
20858 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
20859 (struct instantiate_cache_type): New type.
20860 (set_instantiated_value, get_instantiated_value): Remove.
20861 (get_instantiated_value_entry): New function.
20862 (instantiate_scev_name): Use the new cache and adjust.
20863 (instantiate_scev_poly): Adjust.
20864 (instantiate_scev_binary): Likewise.
20865 (instantiate_array_ref): Likewise.
20866 (instantiate_scev_convert): Likewise.
20867 (instantiate_scev_not): Likewise.
20868 (instantiate_scev_3): Likewise.
20869 (instantiate_scev_2): Likewise.
20870 (instantiate_scev_r): Likewise.
20871 (instantiate_scev): Likewise.
20872 (resolve_mixers): Likewise.
20873
20874 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
20875
20876 PR target/57091
20877 * lra-constraints.c (best_small_class_operands_num): Remove.
20878 (process_alt_operands): Remove small_class_operands_num. Take
20879 small classes operands into losers and only if the operand is not
20880 matched. Modify debugging output.
20881 (curr_insn_transform): Remove best_small_class_operands_num.
20882 Print insn name.
20883
20884 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20885
20886 * config/aarch64/aarch64-builtins.c
20887 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
20888 * config/aarch64/aarch64-simd-builtins.def
20889 (reduc_splus_): Add new modes.
20890 (reduc_uplus_): New.
20891 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
20892 (reduc_uplus_v4sf): Likewise.
20893 (reduc_splus_v4sf): Likewise.
20894 (aarch64_addv<mode>): Likewise.
20895 (reduc_uplus_<mode>): Likewise.
20896 (reduc_splus_<mode>): Likewise.
20897 (aarch64_addvv2di): Likewise.
20898 (reduc_uplus_v2di): Likewise.
20899 (reduc_splus_v2di): Likewise.
20900 (aarch64_addvv2si): Likewise.
20901 (reduc_uplus_v2si): Likewise.
20902 (reduc_splus_v2si): Likewise.
20903 (reduc_<sur>plus_<mode>): New.
20904 (reduc_<sur>plus_v2di): Likewise.
20905 (reduc_<sur>plus_v2si): Likewise.
20906 (reduc_<sur>plus_v4sf): Likewise.
20907 (aarch64_addpv4sf): Likewise.
20908 * config/aarch64/arm_neon.h
20909 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
20910 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
20911 add UNSPEC_SADDV, UNSPEC_UADDV.
20912 (SUADDV): New.
20913 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
20914
20915 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20916
20917 * config/aarch64/arm_neon.h
20918 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
20919
20920 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20921
20922 * config/aarch64/aarch64-builtins
20923 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
20924
20925 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20926
20927 * config/aarch64/aarch64-simd-builtins.def
20928 (reduc_smax_): New.
20929 (reduc_smin_): Likewise.
20930 (reduc_umax_): Likewise.
20931 (reduc_umin_): Likewise.
20932 (reduc_smax_nan_): Likewise.
20933 (reduc_smin_nan_): Likewise.
20934 (fmax): Remove.
20935 (fmin): Likewise.
20936 (smax): Update for V2SF, V4SF and V2DF modes.
20937 (smin): Likewise.
20938 (smax_nan): New.
20939 (smin_nan): Likewise.
20940 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
20941 (<su><maxmin><mode>3): ...This, refactor.
20942 (s<maxmin><mode>3): New.
20943 (<maxmin_uns><mode>3): Likewise.
20944 (reduc_<maxmin_uns>_<mode>): Refactor.
20945 (reduc_<maxmin_uns>_v4sf): Likewise.
20946 (reduc_<maxmin_uns>_v2si): Likewise.
20947 (aarch64_<fmaxmin><mode>: Remove.
20948 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
20949 new builtin names.
20950 (vmin<q>_f<32,64>): Likewise.
20951 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20952 (FMAXMIN): New.
20953 (su): Add mappings for smax, smin, umax, umin.
20954 (maxmin): New.
20955 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
20956 (FMAXMIN): Rename as...
20957 (FMAXMIN_UNS): ...This.
20958 (maxminv): Remove.
20959 (fmaxminv): Likewise.
20960 (fmaxmin): Likewise.
20961 (maxmin_uns): New.
20962 (maxmin_uns_op): Likewise.
20963
20964 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20965
20966 * config/aarch64/arm_neon.h
20967 (vac<ge, gt><sd>_f<32, 64>): Rename to...
20968 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
20969 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
20970
20971 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20972
20973 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
20974 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
20975
20976 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20977
20978 * config/aarch64/aarch64-simd.md
20979 (vcond<mode>_internal): Handle special cases for constant masks.
20980 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
20981 (vcondu<mode><mode>): Likewise.
20982 (vcond<v_cmp_result><mode>): New.
20983
20984 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20985
20986 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
20987 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
20988 * config/aarch64/aarch64-simd-builtins.def
20989 (cmeq): Update to BUILTIN_VALLDI.
20990 (cmgt): Likewise.
20991 (cmge): Likewise.
20992 (cmle): Likewise.
20993 (cmlt): Likewise.
20994 * config/aarch64/arm_neon.h
20995 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
20996 to builtins or C as appropriate.
20997
20998 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
20999
21000 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
21001 (cmgeu): ...This.
21002 (cmhi): Rename to...
21003 (cmgtu): ...This.
21004 * config/aarch64/aarch64-simd.md
21005 (simd_mode): Add SF.
21006 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
21007 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
21008 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
21009 (cstore<mode>_neg): ...This.
21010 * config/aarch64/iterators.md
21011 (VALLF): new.
21012 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
21013 (COMPARISONS): New.
21014 (UCOMPARISONS): Likewise.
21015 (optab): Add missing comparisons.
21016 (n_optab): New.
21017 (cmp_1): Likewise.
21018 (cmp_2): Likewise.
21019 (CMP): Likewise.
21020 (cmp): Remove.
21021 (VCMP_S): Likewise.
21022 (VCMP_U): Likewise.
21023 (V_cmp_result): Add DF, SF modes.
21024 (v_cmp_result): Likewise.
21025 (v): Likewise.
21026 (vmtype): Likewise.
21027 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
21028
21029 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
21030
21031 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
21032 define_insn to define_insn_and_split.
21033 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
21034 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
21035 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
21036 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
21037 (thumb2_negscc): Likewise.
21038
21039 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21040
21041 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
21042
21043 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
21044
21045 * config/arm/thumb2.md: Remove trailing whitespaces.
21046
21047 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
21048
21049 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
21050 Use gen_int_mode rather than GEN_INT.
21051
21052 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
21053
21054 * value-prof.c (stream_in_histogram_value): Remove the strayed
21055 debug_gimple_stmt.
21056
21057 2013-04-30 Richard Biener <rguenther@suse.de>
21058
21059 PR middle-end/57122
21060 * cfghooks.c (split_edge): Properly check for the loop latch edge.
21061
21062 2013-04-30 Richard Biener <rguenther@suse.de>
21063
21064 PR middle-end/57107
21065 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
21066
21067 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
21068
21069 PR rtl-optimization/56957
21070 PR rtl-optimization/57105
21071 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
21072 variable. Use just INSN_UID for determining whether an insn
21073 should be only disconnected from the insn stream.
21074 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
21075
21076 2013-04-30 Jakub Jelinek <jakub@redhat.com>
21077
21078 PR tree-optimization/57104
21079 * tsan.c (instrument_expr): Don't instrument accesses to
21080 DECL_HARD_REGISTER VAR_DECLs.
21081
21082 2013-04-30 Richard Biener <rguenther@suse.de>
21083
21084 * function.h (loops_for_fn): New inline function.
21085 (set_loops_for_fn): Likewise.
21086 * cfgloop.h (place_new_loop): Add struct function parameter.
21087 (get_loop): Likewise.
21088 (get_loops): Likewise.
21089 (number_of_loops): Likewise.
21090 (fel_next): Adjust.
21091 (fel_init): Likewise.
21092 * cfg.c (get_loop_copy): Adjust.
21093 * cfgloop.c (flow_loops_dump): Likewise.
21094 (record_loop_exits): Likewise.
21095 (verify_loop_structure): Likewise.
21096 * cfgloopanal.c (mark_irreducible_loops): Likewise.
21097 (estimate_reg_pressure_cost): Likewise.
21098 (mark_loop_exit_edges): Likewise.
21099 * cfgloopmanip.c (place_new_loop): Likewise.
21100 (add_loop): Likewise.
21101 (duplicate_loop): Likewise.
21102 * graph.c (draw_cfg_nodes): Likewise.
21103 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
21104 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
21105 (extract_affine_chrec): Likewise.
21106 (build_scop_iteration_domain): Likewise.
21107 * graphite.c (graphite_initialize): Likewise.
21108 * ira-build.c (create_loop_tree_nodes): Likewise.
21109 (more_one_region_p): Likewise.
21110 (rebuild_regno_allocno_maps): Likewise.
21111 (mark_loops_for_removal): Likewise.
21112 (mark_all_loops_for_removal): Likewise.
21113 (remove_unnecessary_regions): Likewise.
21114 (ira_build): Likewise.
21115 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
21116 * loop-init.c (fix_loop_structure): Likewise.
21117 (gate_rtl_move_loop_invariants): Likewise.
21118 (gate_rtl_unswitch): Likewise.
21119 (gate_rtl_unroll_and_peel_loops): Likewise.
21120 (rtl_doloop): Likewise.
21121 * lto-streamer-in.c (input_cfg): Likewise.
21122 * lto-streamer-out.c (output_cfg): Likewise.
21123 * modulo-sched.c (sms_schedule): Likewise.
21124 * predict.c (tree_estimate_probability): Likewise.
21125 (tree_estimate_probability_driver): Likewise.
21126 (estimate_loops): Likewise.
21127 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
21128 (move_sese_region_to_fn): Likewise.
21129 (debug_loop_num): Likewise.
21130 * tree-chrec.c (chrec_evaluate): Likewise.
21131 (hide_evolution_in_other_loops_than_loop): Likewise.
21132 (chrec_component_in_loop_num): Likewise.
21133 (reset_evolution_in_loop): Likewise.
21134 (evolution_function_is_invariant_rec_p): Likewise.
21135 * tree-if-conv.c (main_tree_if_conversion): Likewise.
21136 * tree-inline.c (copy_loops): Likewise.
21137 (copy_cfg_body): Likewise.
21138 (tree_function_versioning): Likewise.
21139 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
21140 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
21141 Likewise.
21142 (add_to_evolution_1): Likewise.
21143 (scev_const_prop): Likewise.
21144 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
21145 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
21146 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
21147 (tree_ssa_lim_initialize): Likewise.
21148 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
21149 (verify_loop_closed_ssa): Likewise.
21150 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
21151 (tree_ssa_loop_im): Likewise.
21152 (tree_ssa_loop_unswitch): Likewise.
21153 (tree_vectorize): Likewise.
21154 (check_data_deps): Likewise.
21155 (tree_ssa_loop_ivcanon): Likewise.
21156 (tree_ssa_loop_bounds): Likewise.
21157 (tree_complete_unroll): Likewise.
21158 (tree_complete_unroll_inner): Likewise.
21159 (tree_parallelize_loops): Likewise.
21160 (tree_ssa_loop_prefetch): Likewise.
21161 (tree_ssa_loop_ivopts): Likewise.
21162 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
21163 * tree-vectorizer.c (vectorize_loops): Likewise.
21164
21165 2013-04-29 Mike Frysinger <vapier@gentoo.org>
21166
21167 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
21168 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
21169 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
21170 with EABI_LINK_SPEC.
21171
21172 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21173
21174 PR target/44578
21175 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
21176 alternative.
21177
21178 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
21179
21180 PR target/57097
21181 * lra-constraints.c (process_alt_operands): Discourage a bit more
21182 using memory for pseudos. Print cost dump for alternatives.
21183 Modify cost values for conflicts with early clobbers.
21184 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
21185
21186 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21187
21188 PR target/57098
21189 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
21190
21191 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21192
21193 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
21194 from/to S register.
21195 (movdi_aarch64): Support LDR/STR from/to D register.
21196
21197 2013-04-29 Ian Bolton <ian.bolton@arm.com>
21198
21199 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
21200 or higher by default.
21201
21202 2013-04-29 Richard Biener <rguenther@suse.de>
21203
21204 PR middle-end/57075
21205 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
21206 even if not adding abnormal edges for calls that can make
21207 abnormal gotos.
21208
21209 2013-04-29 Richard Biener <rguenther@suse.de>
21210
21211 PR middle-end/57103
21212 * tree-cfg.c (move_stmt_op): Fix condition under which to update
21213 TREE_BLOCK.
21214 (move_stmt_r): Remove redundant checking.
21215
21216 2013-04-29 Teresa Johnson <tejohnson@google.com>
21217
21218 PR bootstrap/57077
21219 * basic-block.h (apply_scale): New function.
21220 (apply_probability): Use apply_scale.
21221 * gimple-streamer-in.c (input_bb): Ditto.
21222 * lto-streamer-in.c (input_cfg): Ditto.
21223 * lto-cgraph.c (merge_profile_summaries): Ditto.
21224 * tree-optimize.c (execute_fixup_cfg): Ditto.
21225 * tree-inline.c (copy_bb): Update comment to use apply_scale.
21226 (copy_edges_for_bb): Ditto.
21227 (copy_cfg_body): Ditto.
21228
21229 2013-04-29 Tom de Vries <tom@codesourcery.com>
21230
21231 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
21232 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
21233 (tail_merge_optimize): Handle current_loops == NULL.
21234
21235 2013-04-26 Jeff Law <law@redhat.com>
21236
21237 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
21238 (simplify_cond_using_ranges): Generalize code to simplify
21239 COND_EXPRs where one argument is a constant and the other
21240 is an SSA_NAME created by an integral type conversion.
21241
21242 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21243
21244 * config/arm/arm.md (store_minmaxsi): Use only when
21245 optimize_insn_for_size_p.
21246
21247 2013-04-29 Christian Bruel <christian.bruel@st.com>
21248
21249 PR target/57108
21250 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
21251
21252 2013-04-29 Richard Biener <rguenther@suse.de>
21253
21254 PR middle-end/57089
21255 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
21256 loop tree make sure to schedule a fixup for the child as well.
21257 (expand_omp_for_generic): Properly add loops.
21258 (expand_omp_for_static_nochunk): Likewise.
21259 (expand_omp_for_static_chunk): Likewise.
21260 (expand_omp_for): For the degenerate case fixup loops.
21261 (expand_omp_sections): Fix default bb placement in loops.
21262 (expand_omp_atomic_pipeline): Properly add loops.
21263
21264 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21265
21266 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
21267
21268 2013-04-29 Tom de Vries <tom@codesourcery.com>
21269
21270 * tree-ssa-tail-merge.c: Update header comment.
21271
21272 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21273
21274 * config/aarch64/arm_neon.h
21275 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
21276 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
21277 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
21278 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
21279 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
21280 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
21281 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
21282 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
21283
21284 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21285
21286 * config/aarch64/aarch64-simd.md
21287 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
21288 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
21289 fix_trunc, fixuns_trunc.
21290 (ftrunc<VDQF:mode>2): New.
21291 * config/aarch64/iterators.md (optab): Add fix, fixuns.
21292 (fix_trunc_optab): New.
21293
21294 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21295
21296 * config/aarch64/aarch64-builtins.c
21297 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
21298 iceilf, lround, iroundf.
21299
21300 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
21301
21302 PR target/54349
21303 * config/i386/i386.h (enum ix86_tune_indices)
21304 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
21305 New, split from X86_TUNE_INTER_UNIT_MOVES.
21306 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
21307 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
21308 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
21309 (TARGET_INTER_UNIT_MOVES): Remove.
21310 * config/i386/i386.c (initial_ix86_tune_features): Update.
21311 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
21312 (ix86_expand_convert_uns_didf_sse): Use
21313 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21314 (ix86_expand_vector_init_one_nonzero): Ditto.
21315 (ix86_expand_vector_init_interleave): Ditto.
21316 (inline_secondary_memory_needed): Return true for moves from SSE class
21317 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
21318 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
21319 * config/i386/constraints.md (Yi, Ym): Depend on
21320 TARGET_INTER_UNIT_MOVES_TO_VEC.
21321 (Yj, Yn): New constraints.
21322 * config/i386/i386.md (*movdi_internal): Change constraints of
21323 operand 1 from Yi to Yj and from Ym to Yn.
21324 (*movsi_internal): Ditto.
21325 (*movdf_internal): Ditto.
21326 (*movsf_internal): Ditto.
21327 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
21328 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
21329 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
21330 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
21331 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
21332 * config/i386/sse.md (movdi_to_sse): Ditto.
21333 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
21334 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
21335 TARGET_INTER_UNIT_MOVES.
21336 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
21337 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
21338 instead of TARGET_INTER_UNIT_MOVES.
21339 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
21340 operand 1 from Yi to Yj and from Ym to Yn.
21341
21342 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21343
21344 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
21345 (float_truncate_hi_): Likewise.
21346 (float_extend_lo_): Likewise.
21347 (float_truncate_lo_): Likewise.
21348 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
21349 (aarch64_float_extend_lo_v2df): Likewise.
21350 (vec_unpacks_hi_v4sf): Likewise.
21351 (aarch64_float_truncate_lo_v2sf): Likewise.
21352 (aarch64_float_truncate_hi_v4sf): Likewise.
21353 (vec_pack_trunc_v2df): Likewise.
21354 (vec_pack_trunc_df): Likewise.
21355
21356 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21357
21358 * config/aarch64/aarch64-builtins.c
21359 (aarch64_fold_builtin): Fold float conversions.
21360 * config/aarch64/aarch64-simd-builtins.def
21361 (floatv2si, floatv4si, floatv2di): New.
21362 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
21363 * config/aarch64/aarch64-simd.md
21364 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
21365 * config/aarch64/iterators.md (FLOATUORS): New.
21366 (optab): Add float, floatuns.
21367 (su_optab): Likewise.
21368
21369 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21370
21371 * config/aarch64/aarch64-builtins.c
21372 (aarch64_builtin_vectorized_function): Use new names for
21373 fcvt builtins.
21374 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
21375 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
21376 (fcvtzu): Split as...
21377 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
21378 (fcvtas): Split as...
21379 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
21380 (fcvtau): Split as...
21381 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
21382 (fcvtps): Split as...
21383 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
21384 (fcvtpu): Split as...
21385 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
21386 (fcvtms): Split as...
21387 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
21388 (fcvtmu): Split as...
21389 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
21390 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
21391 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
21392 (lfrintnusf, lfrintnudf): Likewise.
21393 * config/aarch64/aarch64-simd.md
21394 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
21395 define_insn.
21396 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
21397 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
21398 (fcvt_pattern): Likewise.
21399
21400 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21401
21402 * config/aarch64/aarch64-simd.md
21403 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
21404 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
21405
21406 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21407
21408 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
21409 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
21410 (vrnd<a,m,n,p>_f32): Implement using builtins.
21411 (vrnd<i,x><q>_f<32, 64>): New.
21412
21413 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
21414
21415 * config/aarch64/aarch64-builtins.c
21416 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
21417 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
21418 (frintz): Rename to...
21419 (btrunc): ...this.
21420 (frintp): Rename to...
21421 (ceil): ...this.
21422 (frintm): Rename to...
21423 (floor): ...this.
21424 (frinti): Rename to...
21425 (nearbyint): ...this.
21426 (frintx): Rename to...
21427 (rint): ...this.
21428 (frinta): Rename to...
21429 (round): ...this.
21430 * config/aarch64/aarch64-simd.md
21431 (aarch64_frint<frint_suffix><mode>): Delete.
21432 (<frint_pattern><mode>2): Convert to insn.
21433 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
21434 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
21435 (frint_pattern): Likewise.
21436 (frint_suffix): Likewise.
21437
21438 2013-04-29 Richard Biener <rguenther@suse.de>
21439
21440 PR tree-optimization/57081
21441 * loop-init.c: Include tree-flow.h.
21442 (loop_optimizer_finalize): Free number of iteration estimates.
21443 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
21444
21445 2013-04-29 Jakub Jelinek <jakub@redhat.com>
21446
21447 PR tree-optimization/57083
21448 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
21449 non-singleton shift count range, zero extend low_bound for uns case.
21450
21451 * config/i386/predicates.md (general_vector_operand): New predicate.
21452 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
21453 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
21454 if they aren't nonimmediate operands. If their original values
21455 satisfy const_vector_equal_evenodd_p, don't shift them.
21456 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
21457 predicates. For the SSE4.1 case force operands[{1,2}] into registers
21458 if not nonimmediate_operand.
21459 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
21460 instead of register_operand.
21461 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
21462
21463 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
21464
21465 * stor-layout.c (finalize_size_functions): Allocate a structure and
21466 reset cfun before dumping the functions.
21467
21468 2013-04-27 Jakub Jelinek <jakub@redhat.com>
21469
21470 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
21471
21472 PR target/56866
21473 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
21474 use xop_pmacsdqh if uns_p.
21475 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
21476 the immediate rotate count.
21477
21478 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
21479
21480 * rtl.h (struct rtx_def): Add comment for field jump.
21481 (LRA_SUBREG_P): New macro.
21482 * recog.c (register_operand): Check LRA_SUBREG_P.
21483 * lra.c (lra): Add note at the end of RTL code. Align non-empty
21484 stack frame.
21485 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
21486 (lra_final_code_change): Skip subreg change for operators.
21487 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
21488 if there are no operand changes.
21489 * lra-constraints.c (curr_insn_set): New.
21490 (match_reload): Set LRA_SUBREG_P.
21491 (emit_spill_move): Ditto.
21492 (check_and_process_move): Use curr_insn_set. Process only single
21493 set insns. Don't initialize sec_mem_p and change_p.
21494 (simplify_operand_subreg): Use LRA_SUBREG_P.
21495 (reg_in_class_p): New function.
21496 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
21497 of #ifdef. Add code to remove cycling.
21498 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
21499 non-null disp. Reload inner instead of disp when base and index
21500 are null. Try to put lo_sum into register.
21501 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
21502 (check_and_process_move): Move code for move cost check to
21503 simple_move_p. Remove equiv_substitution.
21504 (simple_move_p): New function.
21505 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
21506 curr_insn_set. Call check_and_process_move only for single set
21507 insns. Use the new function. Move call of check_and_process_move
21508 after operand equiv substitution and address process.
21509
21510 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21511
21512 PR go/57045
21513 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
21514 with nonlocal goto receivers or returns twice calls, ignore
21515 unininitialized values from abnormal edges to nl goto receiver
21516 or returns twice call.
21517
21518 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21519
21520 PR tree-optimization/57051
21521 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
21522 and VEC_RSHIFT_EXPR if shift count is a multiple of element
21523 bitsize.
21524
21525 2013-04-26 Richard Biener <rguenther@suse.de>
21526
21527 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
21528 (expand_omp_taskreg): Likewise. Mark loops for fixup.
21529 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
21530 (fixup_loop_arrays_after_move): New function.
21531 (move_sese_region_to_fn): Properly outline the loop tree parts
21532 of the SESE region.
21533
21534 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
21535
21536 * config/i386/i386.md (type, unit): Fix long lines.
21537
21538 2013-04-26 Richard Biener <rguenther@suse.de>
21539
21540 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
21541 (lto-streamer-out.o): Likewise.
21542 * cfgloop.c (init_loops_structure): Export, add struct function
21543 argument and adjust.
21544 (flow_loops_find): Adjust.
21545 * cfgloop.h (enum loop_estimation): Add EST_LAST.
21546 (init_loops_structure): Declare.
21547 * lto-streamer-in.c: Include cfgloop.h.
21548 (input_cfg): Input the loop tree.
21549 * lto-streamer-out.c: Include cfgloop.h.
21550 (output_cfg): Output the loop tree.
21551 (output_struct_function_base): Do not drop PROP_loops.
21552
21553 2013-03-26 Richard Biener <rguenther@suse.de>
21554
21555 * tree-cfg.c (execute_build_cfg): Build the loop tree.
21556 (pass_build_cfg): Provide PROP_loops.
21557 (move_sese_region_to_fn): Remove loops that are outlined into fn
21558 for now.
21559 * tree-inline.c: Include cfgloop.h.
21560 (initialize_cfun): Do not drop PROP_loops.
21561 (copy_loops): New function.
21562 (copy_cfg_body): Copy loop structure.
21563 (tree_function_versioning): Initialize destination loop tree.
21564 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
21565 (pass_parallelize_loops): Do IL verification.
21566 * loop-init.c (loop_optimizer_init): Fixup loops if required.
21567 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
21568 the CFG make sure we fixup loops as well.
21569 * tree-ssa-tail-merge.c: Include cfgloop.h.
21570 (replace_block_by): When merging loop latches mark loops for fixup.
21571 * lto-streamer-out.c (output_struct_function_base): Drop
21572 PROP_loops for now.
21573 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
21574 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
21575 * ipa-split.c: Include cfgloop.h.
21576 (split_function): Add the new return block to the loop tree root.
21577 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
21578 whether we have removed the forwarder block.
21579 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
21580 * cfgloop.h (place_new_loop): Declare.
21581 * cfgloopmanip.c (place_new_loop): Export.
21582 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
21583 (tree-switch-conversion.o): Likewise.
21584 (tree-complex.o): Likewise.
21585 (tree-inline.o): Likewise.
21586 (tree-ssa-tailmerge.o): Likewise.
21587 (ipa-split.o): Likewise.
21588 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
21589 (tree-ssa-copy.o): Likewise.
21590 * tree-switch-conversion.c: Include cfgloop.h
21591 (process_switch): If we emit a bit-test cascade, schedule loops
21592 for fixup.
21593 * tree-complex.c: Include cfgloop.h.
21594 (expand_complex_div_wide): Properly add new basic-blocks to loops.
21595 * asan.c: Include cfgloop.h.
21596 (create_cond_insert_point): Properly add new basic-blocks to
21597 loops, schedule loop fixup.
21598 * cfgloop.c (verify_loop_structure): Check that looks are not
21599 marked for fixup.
21600 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
21601 to loops.
21602 (expand_omp_for_generic): Likewise.
21603 (expand_omp_sections): Likewise.
21604 (expand_omp_atomic_pipeline): Schedule loops for fixup.
21605 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
21606 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
21607 is initialized, not when loops are present.
21608 * tree-parloops.c (parallelize_loops): Remove checking here.
21609 * passes.c (init_optimization_passes): Schedule a copy-propagation
21610 pass before complete unrolling of inner loops.
21611
21612 2013-04-26 Jakub Jelinek <jakub@redhat.com>
21613
21614 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
21615 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
21616 (colorize_init): Add argument to _WIN32 version.
21617 * toplev.c: Include diagnostic-color.h.
21618 (process_options): Default to -fdiagnostics-color=auto if
21619 GCC_COLORS env var is in the environment.
21620 * common.opt (fdiagnostics-color=): Add Var and Init.
21621 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
21622 env var is in the environment, the default is auto rather than never.
21623
21624 * diagnostic.h (file_name_as_prefix): Add context argument.
21625 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
21626 the string as locus.
21627 * langhooks.c (lhd_print_error_function): Adjust caller.
21628
21629 2013-04-25 Lawrence Crowl <crowl@google.com>
21630
21631 * var-tracking.c (shared_hash_def::htab):
21632 Change type to hash_table. Update dependent calls and types.
21633
21634 2013-04-25 Lawrence Crowl <crowl@google.com>
21635
21636 * Makefile.in: Update as needed below.
21637
21638 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
21639 Move declaration to after the type's method definitons.
21640
21641 * attribs.c (htab_t scoped_attributes::attribute_hash):
21642 Change type to hash_table. Update dependent calls and types.
21643
21644 * bitmap.c (htab_t bitmap_desc_hash):
21645 Change type to hash_table. Update dependent calls and types.
21646
21647 * cselib.c (htab_t cselib_hash_table):
21648 Change type to hash_table. Update dependent calls and types.
21649
21650 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
21651 (hash_string_slot_node): Move implementation into lto-streamer.h
21652 struct string_slot_hasher.
21653 (eq_string_slot_node): Likewise.
21654
21655 * data-streamer-out.c: Update output_block::string_hash_table
21656 dependent calls and types.
21657
21658 * dwarf2cfi.c (htab_t trace_index):
21659 Change type to hash_table. Update dependent calls and types.
21660
21661 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
21662 Change type to hash_table. Update dependent calls and types.
21663 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
21664 (htab_t optimize_external_refs::map): Likewise.
21665 (htab_t output_comp_unit::extern_map): Likewise.
21666 (htab_t output_comdat_type_unit::extern_map): Likewise.
21667 (htab_t output_macinfo::macinfo_htab): Likewise.
21668 (htab_t optimize_location_lists::htab): Likewise.
21669 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
21670
21671 * except.c (htab_t ehspec_hash_type):
21672 Change type to hash_table. Update dependent calls and types.
21673 (assign_filter_values::ttypes): Likewise.
21674 (assign_filter_values::ehspec): Likewise.
21675 (sjlj_assign_call_site_values::ar_hash): Likewise.
21676 (convert_to_eh_region_ranges::ar_hash): Likewise.
21677
21678 * gcse.c (htab_t pre_ldst_table):
21679 Change type to hash_table. Update dependent calls and types.
21680
21681 * ggc-common.c (htab_t saving_htab):
21682 Change type to hash_table. Update dependent calls and types.
21683 (htab_t loc_hash): Likewise.
21684 (htab_t ptr_hash): Likewise.
21685 (call_count): Rename ggc_call_count.
21686 (call_alloc): Rename ggc_call_alloc.
21687 (loc_descriptor): Rename make_loc_descriptor.
21688 (add_statistics): Rename ggc_add_statistics.
21689
21690 * ggc-common.c (saving_htab):
21691 Change type to hash_table. Update dependent calls and types.
21692
21693 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
21694 (push_gimplify_context): Likewise.
21695 (pop_gimplify_context): Likewise.
21696 (struct gimple_temp_hash_elt): Added.
21697 (struct gimplify_hasher): Likewise.
21698 (struct gimplify_ctx.temp_htab):
21699 Change type to hash_table. Update dependent calls and types.
21700
21701 * gimple-fold.c: Include gimplify-ctx.h.
21702
21703 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
21704 Change type to hash_table. Update dependent calls and types.
21705 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
21706 avoid potential global name collision.
21707
21708 * gimplify.c: Include gimplify-ctx.h.
21709 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
21710 (htab_t gimplify_ctx::temp_htab):
21711 Update dependent calls and types for new type hash_table.
21712 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
21713 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
21714
21715 * gimplify-ctx.h: New.
21716 (struct gimple_temp_hash_elt): Move from gimplify.c.
21717 (class gimplify_hasher): New.
21718 (struct gimplify_ctx): Move from gimple.h.
21719 (htab_t gimplify_ctx::temp_htab):
21720 Change type to hash_table. Update dependent calls and types.
21721
21722 * graphite-clast-to-gimple.c: Include graphite-htab.h.
21723 (htab_t ivs_params::newivs_index):
21724 Change type to hash_table. Update dependent calls and types.
21725 (htab_t ivs_params::params_index): Likewise.
21726 (htab_t print_generated_program::params_index): Likewise.
21727 (htab_t gloog::newivs_index): Likewise.
21728 (htab_t gloog::params_index): Likewise.
21729
21730 * graphite.c: Include graphite-htab.h.
21731 4htab_t graphite_transform_loops::bb_pbb_mapping):
21732 Change type to hash_table. Update dependent calls and types.
21733
21734 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
21735 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
21736 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
21737
21738 * graphite-dependences.c: Include graphite-htab.h.
21739 (loop_is_parallel_p): Change hash table type of parameter.
21740
21741 * graphite-htab.h: New.
21742 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
21743 (extern find_pbb_via_hash): Move from graphite-poly.h.
21744 (extern loop_is_parallel_p): Move from graphite-poly.h.
21745 (extern get_loop_body_pbbs): Move from graphite-poly.h.
21746
21747 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
21748 (extern loop_is_parallel_p): Move to graphite-htab.h.
21749 (extern get_loop_body_pbbs): Move to graphite-htab.h.
21750
21751 * haifa-sched.c (htab_t delay_htab):
21752 Change type to hash_table. Update dependent calls and types.
21753 (htab_t delay_htab_i2): Likewise.
21754
21755 * ira-color.c (htab_t allocno_hard_regs_htab):
21756 Change type to hash_table. Update dependent calls and types.
21757
21758 * ira-costs.c (htab_t cost_classes_htab):
21759 Change type to hash_table. Update dependent calls and types.
21760
21761 * loop-invariant.c (htab_t merge_identical_invariants::eq):
21762 Change type to hash_table. Update dependent calls and types.
21763
21764 * loop-iv.c (htab_t bivs):
21765 Change type to hash_table. Update dependent calls and types.
21766
21767 * loop-unroll.c (htab_t opt_info::insns_to_split):
21768 Change type to hash_table. Update dependent calls and types.
21769 (htab_t opt_info::insns_with_var_to_expand): Likewise.
21770
21771 * lto-streamer.h (struct string_slot): Move from data-streamer.h
21772 (struct string_slot_hasher): New.
21773 (htab_t output_block::string_hash_table):
21774 Change type to hash_table. Update dependent calls and types.
21775
21776 * lto-streamer-in.c (freeing_string_slot_hasher): New.
21777 (htab_t file_name_hash_table):
21778 Change type to hash_table. Update dependent calls and types.
21779
21780 * lto-streamer-out.c: Update output_block::string_hash_table dependent
21781 calls and types.
21782
21783 * lto-streamer.c (htab_t tree_htab):
21784 Change type to hash_table. Update dependent calls and types.
21785
21786 * omp-low.c: Include gimplify-ctx.h.
21787
21788 * passes.c (htab_t name_to_pass_map):
21789 Change type to hash_table. Update dependent calls and types.
21790 (pass_traverse): Rename to passes_pass_traverse.
21791
21792 * plugin.c (htab_t event_tab):
21793 Change type to hash_table. Update dependent calls and types.
21794
21795 * postreload-gcse.c (htab_t expr_table):
21796 Change type to hash_table. Update dependent calls and types.
21797 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
21798
21799 * sese.c (debug_rename_map_1): Make extern.
21800 (htab_t copy_bb_and_scalar_dependences::rename_map):
21801 Change type to hash_table. Update dependent calls and types.
21802
21803 * sese.h (extern debug_rename_map): Move to .c file.
21804
21805 * store-motion.c (htab_t store_motion_mems_table):
21806 Change type to hash_table. Update dependent calls and types.
21807
21808 * trans-mem.c (htab_t tm_new_mem_hash):
21809 Change type to hash_table. Update dependent calls and types.
21810
21811 * tree-browser.c (htab_t TB_up_ht):
21812 Change type to hash_table. Update dependent calls and types.
21813
21814 * tree-cfg.c (htab_t discriminator_per_locus):
21815 Change type to hash_table. Update dependent calls and types.
21816
21817 * tree-complex.c: Include tree-hasher.h
21818 (htab_t complex_variable_components):
21819 Change type to hash_table. Update dependent calls and types.
21820
21821 * tree-eh.c (htab_t finally_tree):
21822 Change type to hash_table. Update dependent calls and types.
21823
21824 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
21825 struct int_tree_hasher.
21826 (extern int_tree_map_eq): Likewise.
21827 (uid_decl_map_hash): Removed.
21828 (extern decl_tree_map_eq): Likewise.
21829
21830 * tree-hasher.h: New.
21831 (struct int_tree_hasher): New.
21832 (typedef int_tree_htab_type): New.
21833
21834 * tree-inline.c: Include gimplify-ctx.h.
21835
21836 * tree-mudflap.c: Include gimplify-ctx.h.
21837
21838 * tree-parloops.c: Include tree-hasher.h.
21839 (htab_t eliminate_local_variables_stmt::decl_address):
21840 Change type to hash_table. Update dependent calls and types.
21841 (htab_t separate_decls_in_region::decl_copies): Likewise.
21842
21843 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
21844 Change type to hash_table. Update dependent calls and types.
21845
21846 * tree-sra.c (candidates):
21847 Change type to hash_table. Update dependent calls and types.
21848
21849 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
21850 in tree-flow.h.
21851 (int_tree_map_hash): Likewise.
21852
21853 * tree-ssa-dom.c (htab_t avail_exprs):
21854 Change type to hash_table. Update dependent calls and types.
21855
21856 * tree-ssa-live.c (var_map_base_init::tree_to_index):
21857 Change type to hash_table. Update dependent calls and types.
21858
21859 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
21860 Change type to hash_table. Update dependent calls and types.
21861
21862 * tree-ssa-phiopt.c (seen_ssa_names):
21863 Change type to hash_table. Update dependent calls and types.
21864
21865 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
21866 Change type to hash_table. Update dependent calls and types.
21867
21868 * tree-ssa-uncprop.c (equiv):
21869 Change type to hash_table. Update dependent calls and types.
21870
21871 2013-04-25 Jakub Jelinek <jakub@redhat.com>
21872
21873 PR rtl-optimization/57003
21874 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
21875 call note_stores with kill_clobbered_value callback again after
21876 killing regs_invalidated_by_call.
21877
21878 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21879
21880 * config/aarch64/aarch64-simd.md
21881 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
21882 (aarch64_simd_bsl<mode>): Likewise.
21883 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
21884
21885 2013-04-25 Marek Polacek <polacek@redhat.com>
21886
21887 PR tree-optimization/57066
21888 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
21889
21890 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21891
21892 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
21893
21894 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21895
21896 * config/aarch64/aarch64-builtins.c
21897 (aarch64_fold_builtin): New.
21898 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
21899 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
21900 * config/aarch64/aarch64-simd-builtins.def (abs): New.
21901 * config/aarch64/arm_neon.h
21902 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
21903
21904 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21905 Tejas Belagod <tejas.belagod@arm.com>
21906
21907 * config/aarch64/aarch64-builtins.c
21908 (aarch64_gimple_fold_builtin): New.
21909 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
21910 * config/aarch64/aarch64-simd-builtins.def (addv): New.
21911 * config/aarch64/aarch64-simd.md (addpv4sf): New.
21912 (addvv4sf): Update.
21913 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
21914
21915 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21916
21917 * config/aarch64/aarch64.md
21918 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
21919
21920 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21921
21922 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
21923 (*ngcsi_uxtw): New pattern.
21924
21925 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21926 Julian Brown <julian@codesourcery.com>
21927
21928 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
21929 (TB_DREG): Add T_V4HF.
21930 (v4hf_UP): New macro.
21931 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21932 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21933 Handle initialisation of V4HF. Adjust initialisation of reinterpret
21934 built-ins.
21935 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
21936 (arm_vector_mode_supported_p): Handle V4HF.
21937 (arm_mangle_map): Handle V4HFmode.
21938 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
21939 * config/arm/arm_neon_builtins.def: Add entries for
21940 vcvtv4hfv4sf, vcvtv4sfv4hf.
21941 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
21942 (neon_vcvtv4hfv4sf): Likewise.
21943 * config/arm/neon-gen.ml: Handle half-precision floating point
21944 features.
21945 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
21946 * config/arm/arm_neon.h: Regenerate.
21947 * config/arm/neon.ml (type elts): Add F16.
21948 (type vectype): Add T_float16x4, T_floatHF.
21949 (type vecmode): Add V4HF.
21950 (type features): Add Requires_FP_bit feature.
21951 (elt_width): Handle F16.
21952 (elt_class): Likewise.
21953 (elt_of_class_width): Likewise.
21954 (mode_of_elt): Refactor.
21955 (type_for_elt): Handle F16, fix error messages.
21956 (vectype_size): Handle T_float16x4.
21957 (vcvt_sh): New function.
21958 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
21959 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
21960 (string_of_mode): Handle V4HF.
21961 * doc/arm-neon-intrinsics.texi: Regenerate.
21962
21963 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
21964
21965 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
21966 format specifier in 'X' case.
21967
21968 2013-04-25 Alan Modra <amodra@gmail.com>
21969
21970 PR target/57052
21971 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
21972 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
21973 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
21974 Repeat for many other rotate/shift and mask patterns using subregs.
21975 Name lshiftrt insns.
21976 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
21977 on WORDS_BIG_ENDIAN.
21978
21979 2013-04-25 Alan Modra <amodra@gmail.com>
21980
21981 * config.gcc: Support little-endian powerpc-linux targets.
21982 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
21983 (LINK_OS_LINUX_SPEC): Define.
21984 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
21985 Preserve MASK_LITTLE_ENDIAN.
21986 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
21987 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
21988 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
21989 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
21990 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
21991 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
21992 Correct fp word order for little-endian. Don't shift toc entries
21993 smaller than a word for little-endian.
21994 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
21995 (bswapdi2 splits): Correct low-part subreg for little-endian.
21996 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
21997 low/high where such is correct only for be.
21998 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
21999 little-endian for -mcall-aixdesc.
22000
22001 2013-04-25 Alan Modra <amodra@gmail.com>
22002
22003 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
22004 replace_equiv_address_nv.
22005
22006 2013-04-25 Alan Modra <amodra@gmail.com>
22007
22008 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
22009
22010 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22011
22012 Revert:
22013 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22014 * rtl.h (struct rtx_def): ...
22015
22016 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22017
22018 PR rtl-optimizations/57046
22019 * lra-constraints (split_reg): Set up lra_risky_transformations_p
22020 for multi-reg splits.
22021
22022 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
22023
22024 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
22025
22026 2013-04-24 Sterling Augustine <saugustine@google.com>
22027
22028 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
22029 (comp_dir_string, debug_str_dwo_section): New.
22030 (DEBUG_STR_DWO_SECTION): Rename to ...
22031 (DEBUG_DWO_STR_SECTION): ... this.
22032 (DEBUG_NORM_STR_SECTION): Delete.
22033 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
22034 (DEBUG_STR_DWO_SECTION_FLAGS): New.
22035 (find_AT_string): Move most logic to ...
22036 (find_AT_string_in_table): ... here. New.
22037 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
22038 add_skeleton_AT_string. Delete logic.
22039 (output_skeleton_debug_sections): Remove call to
22040 add_top_level_skeleton_die_attrs.
22041 (add_comp_dir_attribute): Move logic to comp_dir_string.
22042 (dwarf2out_init): Initialize debug_str_dwo_section.
22043 (output_indirect_string): Call find_string_form.
22044 (output_indirect_strings): Rewrite.
22045 (prune_unused_types): Empty skeleton_debug_str_hash.
22046 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
22047 (dwarf2out_finish): Call output_indirect_strings.
22048
22049 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22050
22051 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
22052
22053 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
22054
22055 * rtl.h (struct rtx_def): Add comment for field jump.
22056 (LRA_SUBREG_P): New macro.
22057 * recog.c (register_operand): Check LRA_SUBREG_P.
22058 * lra.c (lra): Add note at the end of RTL code. Align non-empty
22059 stack frame.
22060 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
22061 (lra_final_code_change): Skip subreg change for operators.
22062 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
22063 if there are no operand changes.
22064 * lra-constraints.c (curr_insn_set): New.
22065 (match_reload): Set LRA_SUBREG_P.
22066 (emit_spill_move): Ditto.
22067 (check_and_process_move): Use curr_insn_set. Process only single
22068 set insns. Don't initialize sec_mem_p and change_p.
22069 (simplify_operand_subreg): Use LRA_SUBREG_P.
22070 (reg_in_class_p): New function.
22071 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
22072 of #ifdef. Add code to remove cycling.
22073 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
22074 non-null disp. Reload inner instead of disp when base and index
22075 are null. Try to put lo_sum into register.
22076 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
22077 (check_and_process_move): Move code for move cost check to
22078 simple_move_p. Remove equiv_substitution.
22079 (simple_move_p): New function.
22080 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
22081 curr_insn_set. Call check_and_process_move only for single set
22082 insns. Use the new function. Move call of check_and_process_move
22083 after operand equiv substitution and address process.
22084
22085 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
22086
22087 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
22088 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
22089 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
22090
22091 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
22092
22093 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
22094
22095 2013-04-24 Marek Polacek <polacek@redhat.com>
22096
22097 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
22098 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
22099 (select_loops_exit_conditions): Likewise.
22100 (number_of_iterations_for_all_loops): Likewise.
22101 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
22102 (scev_analysis): Likewise.
22103
22104 2013-04-02 Catherine Moore <clm@codesourcery.com>
22105 Chao-ying Fu <fu@mips.com>
22106
22107 * config/mips/micromips.md (jraddiusp): New pattern.
22108 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
22109 instruction if possible.
22110
22111 2013-04-24 Alan Modra <amodra@gmail.com>
22112
22113 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
22114
22115 2013-04-24 Julian Brown <julian@codesourcery.com>
22116 Chung-Lin Tang <cltang@codesourcery.com>
22117
22118 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
22119 dependency behavior in enumeration type DIE generation. Add TODO note
22120 to comments about future DW_FORM_sdata/udata re-work of related code.
22121
22122 2013-04-23 Lawrence Crowl <crowl@google.com>
22123
22124 * Makefile.in: Update as needed below.
22125
22126 * hash-table.h (class hash_table):
22127 Correct many methods with parameter types compare_type to the correct
22128 value_type. (Correct code was unlikely to notice the change.)
22129 (hash_table::elements_with_deleted) New.
22130 (class hashtable::iterator): New.
22131 (hashtable::begin()): New.
22132 (hashtable::end()): New.
22133 (FOR_EACH_HASH_TABLE_ELEMENT): New.
22134
22135 * statistics.c (statistics_hashes):
22136 Change type to hash_table. Update dependent calls and types.
22137
22138 * tree-into-ssa.c (var_infos):
22139 Change type to hash_table. Update dependent calls and types.
22140
22141 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
22142 Change type to hash_table. Update dependent calls and types.
22143
22144 * tree-ssa-loop-im.c (struct mem_ref.refs):
22145 Change type to hash_table. Update dependent calls and types.
22146
22147 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
22148 Change type to hash_table. Update dependent calls and types.
22149
22150 * tree-ssa-sccvn.c (vn_tables_s::nary):
22151 Change type to hash_table. Update dependent calls and types.
22152 (vn_tables_s::phis): Likewise.
22153 (vn_tables_s::references): Likewise.
22154
22155 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
22156 (vn_reference_eq): Update parameter and return types.
22157
22158 * tree-ssa-structalias.c (pointer_equiv_class_table):
22159 Change type to hash_table. Update dependent calls and types.
22160 (location_equiv_class_table): Likewise.
22161
22162 * tree-vect-data-refs.c: Consequential changes for making
22163 peeling a hash_table.
22164
22165 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
22166 (destroy_loop_vec_info): Dependent hash_table update.
22167
22168 * tree-vectorizer.h (peeling_htab):
22169 Change type to hash_table. Update dependent calls and types.
22170
22171 2013-04-23 Shiva Chen <shiva0217@gmail.com>
22172
22173 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
22174 to check the register content is equal or not.
22175 * lra-constraints.c (match_reload): Use lra_assign_reg_val
22176 to assign register content record.
22177 * lra-eliminations.c (update_reg_eliminate): Use
22178 lra_update_reg_val_offset to update register content offset.
22179 * lra-int.h (struct lra_reg): Add offset member.
22180 (lra_reg_val_equal_p): New static inline function.
22181 (lra_update_reg_val_offset): New static inline function.
22182 (lra_assign_reg_val): New static inline function.
22183 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
22184 to assign register content record.
22185 (initialize_lra_reg_info_element): Initial offset to zero.
22186
22187 2013-04-23 Catherine Moore <clm@codesourcery.com>
22188
22189 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
22190 operands. Record compression.
22191
22192 2013-04-23 Xinliang David Li <davidxl@google.com>
22193
22194 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
22195
22196 2013-04-23 Richard Biener <rguenther@suse.de>
22197
22198 PR middle-end/57036
22199 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
22200 parameter, only add abnormal goto edges from the copied body
22201 if the call could perform abnormal gotos.
22202 (copy_cfg_body): Adjust.
22203
22204 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
22205
22206 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
22207
22208 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
22209
22210 * coretypes.h (gimple_stmt_iterator): Add struct to make
22211 compatible with C.
22212
22213 2013-04-23 Richard Biener <rguenther@suse.de>
22214
22215 PR tree-optimization/57026
22216 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
22217 from SSA names occuring in abnormal PHI nodes.
22218
22219 2013-04-22 Andi Kleen <ak@linux.intel.com>
22220
22221 * lto/lto.c (print_lto_report_1): Fix LTO report names.
22222
22223 2013-04-22 Andi Kleen <ak@linux.intel.com>
22224
22225 * lto/lto.c (print_lto_report_1): Declare early.
22226 (read_cgraph_and_symbols): Call print_lto_report_1 early.
22227
22228 2013-04-22 Andi Kleen <ak@linux.intel.com>
22229
22230 * common.opt (-flto-report-wpa): Add.
22231 * doc/invoke.texi (-flto-report-wpa): Add.
22232 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
22233 (lto_main): dito.
22234
22235 2013-04-22 Xinliang David Li <davidxl@google.com>
22236
22237 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
22238 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
22239 * Makefile.in: New dependency
22240
22241 David Daney <ddaney.cavm@gmail.com>
22242
22243 * configure.ac (gcc_cv_as_micromips_support): Use the
22244 --fatal-warnings option.
22245 * configure: Regenerate.
22246
22247 2013-04-22 Marek Polacek <polacek@redhat.com>
22248
22249 PR sanitizer/56990
22250 * tsan.c (instrument_expr): Don't instrument expression
22251 in case its size is zero.
22252
22253 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
22254
22255 PR target/57032
22256 Revert:
22257 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22258
22259 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22260
22261 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22262
22263 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
22264 (gimple_stmt_iterator): New typedef.
22265 * gimple.h (gimple_stmt_iterator): Rename to...
22266 (gimple_stmt_iterator_d): ... This.
22267 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
22268 trees be valid for GIMPLE and GENERIC.
22269 (TARGET_GIMPLE_FOLD_BUILTIN): New.
22270 * gimple-fold.c (gimple_fold_call): Call target hook
22271 gimple_fold_builtin.
22272 * hooks.c (hook_bool_gsiptr_false): New.
22273 * hooks.h (hook_bool_gsiptr_false): New.
22274 * target.def (fold_stmt): New.
22275 * doc/tm.texi: Regenerate.
22276
22277 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
22278
22279 PR target/57018
22280 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
22281 a set sp if no stack realignment.
22282
22283 2013-04-22 Nick Clifton <nickc@redhat.com>
22284
22285 * config.gcc (tilegx-linux): Extend extra_objs rather than
22286 overwriting it.
22287 (tilepro-linux): Likewise.
22288
22289 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22290
22291 * config/aarch64/aarch64-builtins.c
22292 (CF): Remove.
22293 (CF0, CF1, CF2, CF3, CF4, CF10): New.
22294 (VAR<1-12>): Add MAP parameter.
22295 (BUILTIN_*): Likewise.
22296 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
22297 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
22298 (aarch64_ushl_n<mode>): Likewise.
22299 (aarch64_sshr_n<mode>): Likewise.
22300 (aarch64_ushr_n<mode>): Likewise.
22301 (aarch64_<maxmin><mode>): Likewise.
22302 (aarch64_sqrt<mode>): Likewise.
22303 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
22304 (vshr<q>_n_*): Likewise.
22305
22306 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
22307
22308 * config/aarch64/aarch64-builtins.c
22309 (aarch64_simd_builtin_type_mode): Handle SF types.
22310 (sf_UP): Define.
22311 (BUILTIN_GPF): Define.
22312 (aarch64_init_simd_builtins): Handle SF types.
22313 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
22314 (frecps): Likewise.
22315 (frecpx): Likewise.
22316 * config/aarch64/aarch64-simd.md
22317 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
22318 (aarch64_frecpe<mode>): New.
22319 (aarch64_frecps<mode>): Likewise.
22320 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
22321 (v8type): Add frecp<esx>.
22322 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
22323 (aarch64_frecps<mode>): Likewise.
22324 * config/aarch64/iterators.md (FRECP): New.
22325 (frecp_suffix): Likewise.
22326 * config/aarch64/arm_neon.h
22327 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
22328
22329 2013-04-22 Christian Bruel <christian.bruel@st.com>
22330
22331 PR target/56995
22332 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
22333 (REG_CLASS_NAMES): Idem.
22334 (REG_CLASS_CONTENTS): Idem.
22335 (REGCLASS_HAS_FP_REG): Idem.
22336 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
22337 (sh_conditional_register_usage): Idem.
22338
22339 2013-04-21 Jeff Law <law@redhat.com>
22340
22341 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
22342 (ssa_forward_propagate_and_combine): Use it.
22343
22344 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22345
22346 * lra.c: Update the flow chart diagram.
22347
22348 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
22349
22350 PR rtl-optimization/56847
22351 * lra-constraints.c (process_alt_operands): Discourage alternative
22352 with non-matche doffsettable memory constraint fro memory with
22353 known offset.
22354
22355 2013-04-19 Richard Biener <rguenther@suse.de>
22356
22357 PR tree-optimization/56982
22358 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
22359 function.
22360 * gimplify.c (gimplify_call_expr): Notice special calls.
22361 (gimplify_modify_expr): Likewise.
22362 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
22363 abnormal control flow receivers.
22364 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
22365 in the same way as cfun->has_nonlocal_labels.
22366 (gimple_purge_dead_abnormal_call_edges): Likewise.
22367 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
22368 receivers start a basic-block.
22369
22370 2013-04-19 Richard Biener <rguenther@suse.de>
22371
22372 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
22373 member ...
22374 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
22375 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
22376 (SLP_TREE_LOAD_PERMUTATION): Add.
22377 (vect_transform_slp_perm_load): Adjust prototype.
22378 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
22379 (vect_free_slp_instance): Likewise.
22380 (vect_create_new_slp_node): Likewise.
22381 (vect_supported_slp_permutation_p): Remove.
22382 (vect_slp_rearrange_stmts): Adjust.
22383 (vect_supported_load_permutation_p): Likewise. Inline
22384 vect_supported_slp_permutation_p here.
22385 (vect_analyze_slp_instance): Compute load permutations per
22386 slp node instead of per instance.
22387 (vect_get_slp_defs): Adjust.
22388 (vect_transform_slp_perm_load): Likewise.
22389 (vect_schedule_slp_instance): Remove redundant code.
22390 (vect_schedule_slp): Remove hack for PR56270, add it ...
22391 * tree-vect-stmts.c (vectorizable_load): ... here, do not
22392 CSE loads for SLP. Adjust.
22393
22394 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22395
22396 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
22397 spelling in two comments.
22398
22399 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
22400
22401 PR target/56797
22402 * config/arm/arm.c (load_multiple_sequence): Require SP
22403 as base register for loads if SP is in the register list.
22404
22405 2013-04-19 Martin Jambor <mjambor@suse.cz>
22406
22407 PR tree-optimization/56718
22408 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
22409 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
22410 and made public. Adjusted all callers.
22411 (ipa_intraprocedural_devirtualization): New function.
22412 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
22413 (ipa_intraprocedural_devirtualization): Likewise.
22414 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
22415
22416 2013-04-19 Richard Biener <rguenther@suse.de>
22417
22418 PR tree-optimization/57000
22419 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
22420
22421 2013-04-19 Terry Guo <terry.guo@arm.com>
22422
22423 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
22424 Replace with ...
22425 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
22426 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
22427 (cortex_m4_fmacs): Use new reservations.
22428 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
22429
22430 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
22431
22432 PR rtl-optimization/56999
22433 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
22434 related code.
22435 (lra_coalesce): Remove split_origin_bitmap and related code.
22436 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
22437 ranges if necessary.
22438
22439 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
22440
22441 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
22442 New array.
22443 (ix86_expand_call): Remove clobbered_registers array and use
22444 x86_64_ms_sysv_extra_clobbered_registers instead.
22445 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
22446 Declare here.
22447 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
22448 predicate.
22449 * config/i386/i386.md (*call_rex64_ms_sysv): Use
22450 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
22451 (*call_value_rex64_ms_sysv): Ditto.
22452
22453 2013-04-18 Cary Coutant <ccoutant@google.com>
22454
22455 * dwarf2out.c (output_pubnames): Check die_perennial_p of
22456 parent instead of die_mark.
22457
22458 2013-04-18 Diego Novillo <dnovillo@google.com>
22459
22460 * gimple.c (create_gimple_tmp): New.
22461 (get_expr_type): New.
22462 (build_assign): New.
22463 (build_type_cast): New.
22464 * gimple.h (enum ssa_mode): Define.
22465 (gimple_seq_set_location): New.
22466 * asan.c (build_check_stmt): Change some gimple_build_* calls
22467 to use build_assign and build_type_cast.
22468
22469 2013-04-18 Richard Biener <rguenther@suse.de>
22470
22471 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
22472 handle negative step. Remove redundant checks.
22473 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
22474 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
22475 for negative step and grouped loads fail to vectorize.
22476
22477 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22478
22479 * emit-rtl.c (reset_insn_used_flags): New function.
22480 (reset_all_used_flags): Use it.
22481 (verify_insn_sharing): New function.
22482 (verify_rtl_sharing): Fix verification for SEQUENCEs.
22483
22484 2013-04-18 Jakub Jelinek <jakub@redhat.com>
22485
22486 PR tree-optimization/56984
22487 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
22488 and (x >> M) >= N don't register any assertion if N << M is the
22489 minimum value.
22490
22491 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
22492
22493 * lower-subreg.c (resolve_simple_move): If called self-recursive,
22494 do not delete_insn insns that have not yet been emitted, only
22495 unlink them with remove_insn.
22496 * df-scan.c (df_insn_delete): Revert r197492.
22497
22498 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
22499
22500 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
22501 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
22502
22503 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22504
22505 * config/arm/arm.md (movsicc_insn): Convert define_insn into
22506 define_insn_and_split.
22507 (and_scc,ior_scc,negscc): Likewise.
22508 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
22509
22510 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
22511
22512 * config/arm/arm.c (use_return_insn): Return 0 for targets that
22513 can benefit from using a sequence of LDRD instructions in epilogue
22514 instead of a single LDM instruction.
22515
22516 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
22517
22518 PR 45688
22519 * doc/extend.texi: Fix typo.
22520
22521 2013-04-17 Richard Biener <rguenther@suse.de>
22522
22523 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
22524 (vect_build_slp_tree): ... here.
22525 (vect_build_slp_tree_1): Compute which stmts of the SLP group
22526 match. Remove special-casing of mismatched complex loads.
22527 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
22528 re-try the match with swapped commutative operands.
22529 (vect_supported_load_permutation_p): Remove special-casing of
22530 mismatched complex loads.
22531 (vect_analyze_slp_instance): Adjust.
22532
22533 2013-04-17 Richard Biener <rguenther@suse.de>
22534
22535 PR rtl-optimization/56921
22536 * cfgloop.h (struct loop): Add simple_loop_desc member.
22537 (struct niter_desc): Mark with GTY(()).
22538 (simple_loop_desc): Do not use aux field but simple_loop_desc.
22539 * loop-iv.c (get_simple_loop_desc): Likewise.
22540 (free_simple_loop_desc): Likewise.
22541
22542 Revert
22543 2013-04-16 Richard Biener <rguenther@suse.de>
22544
22545 PR rtl-optimization/56921
22546 * loop-init.c (pass_rtl_move_loop_invariants): Add
22547 TODO_do_not_ggc_collect to todo_flags_finish.
22548 (pass_rtl_unswitch): Same.
22549 (pass_rtl_unroll_and_peel_loops): Same.
22550 (pass_rtl_doloop): Same.
22551
22552 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
22553
22554 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
22555 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
22556 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
22557 references.
22558 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
22559 * tree-streamer.c (record_common_node): Adjust reference in comment.
22560
22561 2013-04-17 Terry Guo <terry.guo@arm.com>
22562
22563 * config/arm/cortex-m4.md: Add a new bypass.
22564
22565 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22566
22567 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
22568 New pattern.
22569 (*subs_<optab><mode>_multp2): New pattern.
22570 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22571 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
22572
22573 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22574
22575 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
22576 (*subs_mul_imm_<mode>): New pattern.
22577
22578 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
22579
22580 PR target/56948
22581 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
22582 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
22583 (vsx_movti_32bit): Change j->wa to O->wa.
22584
22585 2013-04-16 Richard Biener <rguenther@suse.de>
22586
22587 PR rtl-optimization/56921
22588 * loop-init.c (pass_rtl_move_loop_invariants): Add
22589 TODO_do_not_ggc_collect to todo_flags_finish.
22590 (pass_rtl_unswitch): Same.
22591 (pass_rtl_unroll_and_peel_loops): Same.
22592 (pass_rtl_doloop): Same.
22593
22594 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
22595
22596 * config/arm/arm.c (emit_multi_reg_push): New declaration
22597 for an existing function.
22598 (arm_emit_strd_push): New function.
22599 (arm_expand_prologue): Used here.
22600 (arm_emit_ldrd_pop): New function.
22601 (arm_expand_epilogue): Used here.
22602 (arm_get_frame_offsets): Update condition.
22603 (arm_emit_multi_reg_pop): Add a special case for load of a single
22604 register with writeback.
22605
22606 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
22607
22608 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
22609 description.
22610
22611 2013-04-16 Richard Biener <rguenther@suse.de>
22612
22613 PR tree-optimization/56756
22614 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
22615 (first_mem_ref_loc): New.
22616 (execute_sm): Place the load temporarily before a previous
22617 access instead of in the latch edge to ensure its SSA dependencies
22618 are defined at points dominating the load.
22619
22620 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
22621
22622 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
22623 correct fix by moving header and footer insn to the footer of
22624 the merged basic block. Clear BB_END of the merged-away block.
22625
22626 PR middle-end/43631
22627 * emit-rtl.c (make_note_raw): New function.
22628 (link_insn_into_chain): New static inline function.
22629 (add_insn): Use it.
22630 (add_insn_before, add_insn_after): Factor insn chain linking code...
22631 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
22632 using link_insn_into_chain.
22633 (note_outside_basic_block_p): New helper function for emit_note_after
22634 and emit_note_before.
22635 (emit_note_after): Use nobb variant of add_insn_after if the note
22636 should not be contained in a basic block.
22637 (emit_note_before): Use nobb variant of add_insn_before if the note
22638 should not be contained in a basic block.
22639 (emit_note_copy): Use make_note_raw.
22640 (emit_note): Likewise.
22641 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
22642 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
22643 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
22644 the moved barrier the tail of the basic block it follows.
22645 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
22646
22647 2013-04-15 Jakub Jelinek <jakub@redhat.com>
22648
22649 PR tree-optimization/56962
22650 * gimple-ssa-strength-reduction.c (record_increment): Only set
22651 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
22652 either rhs1 or rhs2 is equal to c->base_expr.
22653
22654 2013-04-15 Richard Biener <rguenther@suse.de>
22655
22656 PR tree-optimization/56933
22657 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
22658 member.
22659 (GROUP_READ_WRITE_DEPENDENCE): Remove.
22660 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
22661 * tree-vect-data-refs.c (vect_analyze_group_access): Move
22662 dependence check ...
22663 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
22664 ... here.
22665 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
22666 GROUP_READ_WRITE_DEPENDENCE.
22667
22668 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22669
22670 * emit-rtl.c (reset_all_used_flags): New function.
22671 (verify_rtl_sharing): Call reset_all_used_flags before and after
22672 performing the checks.
22673
22674 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
22675
22676 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
22677 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
22678 * config/arm/constraints.md (De): New constraint.
22679 * config/arm/neon.md (anddi3_neon): Delete.
22680 (neon_vand<mode>): Expand to standard anddi3 pattern.
22681 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
22682 Move earlier in the file.
22683 (neon_inv_logic_op2): Likewise.
22684 (arm_anddi_operand_neon): New predicate.
22685
22686 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
22687
22688 * configure.ac (gcc_cv_ld_as_needed): Set
22689 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
22690 Use -z ignore, -z record on *-*-solaris2*.
22691 (HAVE_LD_AS_NEEDED): Update comment.
22692 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
22693 * configure: Regenerate.
22694 * config.in: Regenerate.
22695 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
22696 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
22697 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
22698 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
22699 equivalents. Fix markup.
22700 * doc/tm.texi: Regenerate.
22701
22702 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
22703
22704 * config/i386/i386.opt: New option mstack-protector-guard=.
22705 * config/i386/i386-opts.h: Add enum stack_protector_guard.
22706 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
22707 TARGET_SSP_TLS_GUARD.
22708 * config/i386/i386.c (ix86_option_override_internal): Set
22709 ix86_stack_protector_guard.
22710 * config/i386/i386.md (stack_protect_set): Enable for
22711 TARGET_SSP_TLS_GUARD only.
22712 (stack_protect_set_<mode>): Ditto.
22713 (stack_protect_test): Ditto.
22714 (stack_protect_test_<mode>): Ditto.
22715 * doc/invoke.texi (i386 Option): Document.
22716
22717 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
22718
22719 PR target/56890
22720 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
22721 (S_MODES): Set H_MODE bit.
22722 (SF_MODES): Set only S_MODE and SF_MODE bits.
22723 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
22724 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
22725 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
22726 <MODE_FLOAT>: Likewise.
22727
22728 2013-04-15 Joey Ye <joey.ye@arm.com>
22729
22730 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
22731
22732 2013-04-15 Joey Ye <joey.ye@arm.com>
22733
22734 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
22735 for real far jump.
22736 (thumb_far_jump_used_p): Count instruction size and set
22737 far_jump_used.
22738
22739 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
22740
22741 * reorg.c (fill_simple_delay_slots): Reindent block of code.
22742 * resource.c (mark_target_live_regs): Reformat conditional block.
22743
22744 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22745
22746 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
22747 notes, they are emitted only just before final.
22748 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
22749
22750 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
22751
22752 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
22753 * cfgrtl.c (delete_insn): Call it here instead.
22754 * lra-spills.c (lra_final_code_change): Use delete_insn.
22755 * haifa-sched.c (sched_remove_insn): Likewise.
22756 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
22757 returning to the nop pool.
22758 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
22759 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
22760
22761 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22762
22763 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
22764 * doc/tm.texi: Regenerated.
22765
22766 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
22767
22768 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
22769 QImode checks.
22770
22771 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
22772
22773 * df-core.c (df_find_def): Compare register numbers.
22774 (df_find_use): Likewise.
22775
22776 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
22777
22778 PR target/56903
22779 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
22780 lra_in_progress for return.
22781
22782 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22783
22784 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
22785 define_insn into define_insn_and_split and emit movsicc patterns.
22786
22787 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
22788
22789 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
22790
22791 2013-04-12 Richard Biener <rguenther@suse.de>
22792
22793 * tree-pass.h (TODO_do_not_ggc_collect): New.
22794 * passes.c (execute_one_ipa_transform_pass): Honor
22795 TODO_do_not_ggc_collect.
22796 (execute_one_pass): Likewise.
22797
22798 Revert
22799 2013-04-10 Richard Biener <rguenther@suse.de>
22800
22801 * passes.c (init_optimization_passes): Remove reload pass.
22802 * ira.c (do_reload): Merge into ...
22803 (ira): ... this.
22804 (rest_of_handle_reload): Remove.
22805 (pass_reload): Likewise.
22806 * config/i386/i386.c (ix86_option_override): Refer to ira instead
22807 of reload for vzeroupper pass placement.
22808
22809 2013-04-12 Jakub Jelinek <jakub@redhat.com>
22810
22811 PR tree-optimization/56918
22812 PR tree-optimization/56920
22813 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
22814 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
22815 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
22816 use wide_mul_with_sign method.
22817
22818 2013-04-12 Richard Biener <rguenther@suse.de>
22819
22820 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
22821 not be considered a gimple constant.
22822
22823 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
22824
22825 * fold-const.c (const_binop): Handle vector shifts by a scalar.
22826 (fold_binary_loc): Call const_binop also for mixed vector-scalar
22827 operations.
22828
22829 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
22830 Jakub Jelinek <jakub@redhat.com>
22831
22832 * opts.c: Include diagnostic-color.h.
22833 (common_handle_option): Handle OPT_fdiagnostics_color_.
22834 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
22835 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
22836 (diagnostic-color.o): New.
22837 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
22838 (diagnostic_color_rule): New enum.
22839 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
22840 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
22841 the location string.
22842 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
22843 either NULL, or color kind.
22844 * diagnostic-color.c: New file.
22845 * diagnostic-color.h: New file.
22846 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
22847 arguments.
22848 * doc/invoke.texi (-fdiagnostics-color): Document.
22849 * pretty-print.h (pp_show_color): Define.
22850 (struct pretty_print_info): Add show_color field.
22851 * diagnostic.c: Include diagnostic-color.h.
22852 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
22853 macros. Colorize error:, warning: etc. strings and also the location
22854 string.
22855 (diagnostic_show_locus): Colorize the caret line.
22856 * pretty-print.c: Include diagnostic-color.h.
22857 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
22858 inside of %< %> quotes or quoted through q format modifier.
22859
22860 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22861
22862 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
22863
22864 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22865
22866 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
22867 code in CC_NZ mode.
22868 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
22869 pattern.
22870
22871 2013-04-11 Marek Polacek <polacek@redhat.com>
22872
22873 PR tree-optimization/48184
22874 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
22875
22876 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
22877
22878 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
22879 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
22880 (skip_simple_arithmetic): Tidy up.
22881 * tree.h (skip_simple_constant_arithmetic): Declare.
22882
22883 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
22884
22885 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
22886
22887 2013-04-11 Richard Biener <rguenther@suse.de>
22888
22889 * tree-vect-loop.c (get_initial_def_for_induction): Properly
22890 generate vector constants.
22891
22892 2013-04-11 Richard Biener <rguenther@suse.de>
22893
22894 PR tree-optimization/56878
22895 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
22896 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
22897 New function.
22898 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
22899 Prefer to align the DR with the most invariant base address.
22900
22901 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22902
22903 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
22904 comment.
22905
22906 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
22907
22908 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
22909 floating-point vector comparisons against 0.
22910
22911 2013-04-11 Jakub Jelinek <jakub@redhat.com>
22912
22913 PR tree-optimization/56899
22914 * fold-const.c (extract_muldiv_1): Apply distributive law
22915 only if TYPE_OVERFLOW_WRAPS (ctype).
22916
22917 2013-04-11 Bin Cheng <bin.cheng@arm.com>
22918
22919 PR target/56124
22920 * ira-costs.c (scan_one_insn): Check whether the source rtx of
22921 loading has side effect.
22922
22923 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
22924
22925 * config/sparc/sparc.c: Include tree-pass.h.
22926 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
22927 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
22928 head of file. Change return type. Split off gate function.
22929 (sparc_gate_work_around_errata): New function.
22930 (pass_work_around_errata): New pass definition.
22931 (insert_pass_work_around_errata) New pass insert definition to
22932 insert pass_work_around_errata just after delayed-branch scheduling.
22933 (sparc_option_override): Insert the pass.
22934 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
22935
22936 2013-04-10 David S. Miller <davem@davemloft.net>
22937
22938 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
22939 or -mcpu=hypersparc.
22940
22941 * target.def (cstore_mode): New hook.
22942 * target.h: Include insn-codes.h
22943 * targhooks.c: Likewise.
22944 (default_cstore_mode): New function.
22945 * targhooks.h: Declare it.
22946 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
22947 * doc/tm.texi: Rebuild.
22948 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
22949 target hook, rather than inspecting the insn_data.
22950 * config/sparc/sparc.c (sparc_cstore_mode): New function.
22951 (TARGET_CSTORE_MODE): Redefine.
22952 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
22953 result patterns.
22954 * config/sparc/predicates.md (cstore_result_operand): New special
22955 predicate.
22956 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
22957 Use it for operand 0.
22958 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
22959 (*snesi_special): Likewise.
22960 (*snesi_zero): Likewise.
22961 (*seqsi_zero): Likewise.
22962 (*sltu_insn): Likewise.
22963 (*sgeu_insn): Likewise.
22964 (*seqdi_special): Make operand 0 and comparison operation be of
22965 DImode.
22966 (*snedi_special): Likewise.
22967 (*snedi_special_vis3): Likewise.
22968 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
22969 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
22970 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
22971 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
22972 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
22973 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
22974 (*sltu_extend_sp64): Likewise.
22975 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
22976 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
22977 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
22978 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
22979 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
22980
22981 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
22982
22983 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
22984 (aarch64_start_file): Use the new function.
22985
22986 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
22987 Jason Merrill <jason@redhat.com>
22988
22989 * common.opt: Add -gdwarf.
22990 * opts.c (common_handle_option): Handle it.
22991 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
22992
22993 2013-04-10 Richard Biener <rguenther@suse.de>
22994
22995 * passes.c (execute_todo): Do not call ggc_collect conditional here.
22996 (execute_one_ipa_transform_pass): But unconditionally here.
22997 (execute_one_pass): And here.
22998 (init_optimization_passes): Remove reload pass.
22999 * tree-pass.h (TODO_ggc_collect): Remove.
23000 (pass_reload): Likewise.
23001 * ira.c (do_reload): Merge into ...
23002 (ira): ... this.
23003 (rest_of_handle_reload): Remove.
23004 (pass_reload): Likewise.
23005 * config/i386/i386.c (ix86_option_override): Refer to ira instead
23006 of reload for vzeroupper pass placement.
23007 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
23008 and todo_flags_finish of all passes.
23009
23010 2013-04-10 Richard Biener <rguenther@suse.de>
23011
23012 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
23013 first_const_oprnd field, rename first_def_type to first_op_type.
23014 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
23015 (vect_get_and_check_slp_defs): Always use the type of the
23016 operand. Allow mixed vect_external_def, vect_constant_def types.
23017 (vect_get_constant_vectors): Handle mixed vect_external_def,
23018 vect_constant_def types.
23019
23020 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
23021
23022 PR tree-optimization/55524
23023 * tree-ssa-math-opts.c
23024 (convert_mult_to_fma): Don't use an fms construct
23025 when we don't have an fms operation, but fnma, and it looks
23026 likely that we'll be able to use the latter.
23027
23028 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
23029
23030 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
23031 function.
23032 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
23033 inline fail caused by overwritable functions.
23034
23035 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
23036
23037 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
23038 unnecessary bits in the constant power of two case.
23039
23040 2013-04-10 Richard Biener <rguenther@suse.de>
23041
23042 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
23043 broken code swapping operands.
23044 (vect_build_slp_tree): Do not compute load permutations here.
23045 (vect_analyze_slp_instance): Compute load permutations here,
23046 after building the SLP tree.
23047
23048 2013-04-09 Christian Bruel <christian.bruel@st.com>
23049
23050 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
23051 of next/prev_real_insn.
23052
23053 2013-04-09 Jan Hubicka <jh@suse.cz>
23054
23055 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
23056 Drop aliased parameter.
23057 (function_and_variable_visibility): Do not handle alias pairs.
23058 * cgraph.c (varpool_externally_visible_p): Update prototype.
23059 * varpool.c (varpool_add_new_variable): Update.
23060
23061 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23062
23063 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
23064
23065 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
23066
23067 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
23068
23069 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
23070
23071 2013-04-09 Marek Polacek <polacek@redhat.com>
23072
23073 PR tree-optimization/48762
23074 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
23075
23076 2013-04-09 Richard Biener <rguenther@suse.de>
23077
23078 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
23079 dealing with cost.
23080 (vect_build_slp_tree): Likewise.
23081 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
23082 calculating the cost of a SLP instance.
23083 (vect_analyze_slp_instance): Use it from here, after building
23084 the SLP tree.
23085
23086 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23087
23088 PR middle-end/56883
23089 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
23090 expand_omp_for_static_chunk): Use simple_p = true in
23091 force_gimple_operand_gsi calls when assigning to addressable decls.
23092
23093 2013-04-09 Jeff Law <law@redhat.com>
23094
23095 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
23096 when the boolean was created by converting a wider object which
23097 had a boolean range.
23098
23099 2013-04-09 Richard Biener <rguenther@suse.de>
23100
23101 * tree-vectorizer.h (slp_void_p): Remove.
23102 (slp_tree): Typedef before _slp_tree declaration.
23103 (struct _slp_tree): Use a vector of slp_tree as children.
23104 (vect_get_place_in_interleaving_chain): Remove.
23105 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
23106 Move ...
23107 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
23108 and make static.
23109 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
23110 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
23111 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
23112 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
23113 Use slp_node instead of slp_void_p and adjust.
23114
23115 2013-04-09 Richard Biener <rguenther@suse.de>
23116
23117 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
23118 work that is not necessary.
23119
23120 2013-04-09 Jakub Jelinek <jakub@redhat.com>
23121
23122 PR tree-optimization/56854
23123 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
23124 forward into clobber stmts if it would change MEM_REF lhs into
23125 non-MEM_REF.
23126
23127 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
23128
23129 * tree.c (type_hash_lookup, type_hash_add): Make static.
23130 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
23131
23132 2013-04-09 Richard Biener <rguenther@suse.de>
23133
23134 * tree.h (unsave_expr_now): Remove.
23135 * tree-inline.c (mark_local_for_remap_r): Remove.
23136 (unsave_expr_1): Likewise.
23137 (unsave_r): Likewise.
23138 (unsave_expr_now): Likewise.
23139 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
23140 (propagate_tree_value): Likewise.
23141
23142 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
23143
23144 * doc/rtl.texi (sequence): Rewrite documentation to match the
23145 current use of SEQUENCE rtl objects.
23146 * rtl.def (SEQUENCE): Likewise.
23147
23148 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
23149 Update documentation.
23150 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
23151 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
23152
23153 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
23154
23155 2013-04-08 Teresa Johnson <tejohnson@google.com>
23156
23157 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
23158 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
23159 methods.
23160 (estimate_edge_size_and_time): Add comment to suggest using rounding
23161 methods.
23162 (estimate_node_size_and_time): Ditto.
23163 (remap_edge_change_prob): Use helper rounding divide methods.
23164 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
23165 (gimple_mod_pow2_value_transform): Ditto.
23166 (gimple_mod_subtract_transform): Ditto.
23167 (gimple_ic_transform): Ditto.
23168 (gimple_stringops_transform): Ditto.
23169 * stmt.c (conditional_probability): Ditto.
23170 (emit_case_dispatch_table): Ditto.
23171 * lto-cgraph.c (merge_profile_summaries): Ditto.
23172 * tree-optimize.c (execute_fixup_cfg): Ditto.
23173 * cfgcleanup.c (try_forward_edges): Ditto.
23174 * cfgloopmanip.c (scale_loop_profile): Ditto.
23175 (loopify): Ditto.
23176 (duplicate_loop_to_header_edge): Ditto.
23177 (lv_adjust_loop_entry_edge): Ditto.
23178 * tree-vect-loop.c (vect_transform_loop): Ditto.
23179 * profile.c (compute_branch_probabilities): Ditto.
23180 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
23181 * lto-streamer-in.c (input_cfg): Ditto.
23182 * gimple-streamer-in.c (input_bb): Ditto.
23183 * ipa-cp.c (update_profiling_info): Ditto.
23184 (update_specialized_profile): Ditto.
23185 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
23186 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
23187 rounding methods.
23188 * sched-rgn.c (compute_dom_prob_ps): Ditto.
23189 (compute_trg_info): Ditto.
23190 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
23191 (purge_dead_edges): Ditto.
23192 * loop-unswitch.c (unswitch_loop): Ditto.
23193 * cgraphclones.c (cgraph_clone_edge): Ditto.
23194 (cgraph_clone_node): Ditto.
23195 * tree-inline.c (copy_bb): Ditto.
23196 (copy_edges_for_bb): Ditto.
23197 (initialize_cfun): Ditto.
23198 (copy_cfg_body): Ditto.
23199 (expand_call_inline): Ditto.
23200
23201 2013-04-08 Kai Tietz <ktietz@redhat.com>
23202
23203 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
23204 TARGET_CYGWIN64 by TARGET_64BIT.
23205
23206 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23207
23208 * config/epiphany/epiphany.md (GPR_1): New constant.
23209 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
23210 * config/epiphany/epiphany.c (gen_compare_reg):
23211 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
23212 is already in place.
23213 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
23214 Don't require being called during rtl expansion; If y operlaps r0,
23215 return 0.
23216 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
23217 (epiphany_expand_epilogue): Likewise.
23218
23219 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
23220 Don't use CC_FPmode for ORDERED / UNORDERED.
23221 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
23222
23223 * config/epiphany/constraints.md (CnL): New constraint.
23224 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
23225 * config/epiphany/predicates.md (add_operand): Allow 1024.
23226
23227 * config/epiphany/epiphany.md (logical_op): New code iterator.
23228 (op_mnc): New code attribute.
23229 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
23230 (mov_f+1, mov_f+2): New peephole2 patterns.
23231
23232 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
23233 (cstoresi4): Also allow re-use of zero result when doing a NE
23234 comparison to a non-zero operand.
23235 Use (clobber (scratch)) for first insn if the gpr output is not needed.
23236
23237 * config/epiphany/epiphany.md (<insn_opname>v2si3):
23238 Use gen_addsi3_i / gen_subsi3_i.
23239
23240 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23241
23242 PR c++/34949
23243 PR c++/50243
23244 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
23245 contain anything but clobbers, at most one __builtin_stack_restore,
23246 optionally debug stmts and final resx, and if it has at least one
23247 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
23248 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
23249 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
23250 which isn't defaut definition, remove them.
23251 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
23252 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
23253 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
23254 with MEM_REF LHS with SSA_NAME address.
23255
23256 2013-04-08 Jeff Law <law@redhat.com>
23257
23258 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
23259
23260 2013-04-08 Richard Biener <rguenther@suse.de>
23261
23262 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
23263 extra newline.
23264 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
23265 determined vector type.
23266 (vect_analyze_data_refs): Likewise.
23267 (vect_get_new_vect_var): Adjust.
23268 (vect_create_destination_var): Preserve SSA name versions.
23269 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
23270 not dump anything here.
23271
23272 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
23273
23274 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
23275 Add member lr_slot_known.
23276 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
23277 if necessary.
23278 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
23279 Remove code that sets lr_slot_offset according to what a previous
23280 version of epiphany_emit_save_restore used to do.
23281 (epiphany_emit_save_restore): When doing an lr save or restore,
23282 set/verify lr_slot_known and lr_slot_offset.
23283
23284 2013-04-08 Xinyu Qi <xyqi@marvell.com>
23285
23286 PR target/54338
23287 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
23288 in ALL_REGS.
23289
23290 2013-04-08 Richard Biener <rguenther@suse.de>
23291
23292 * alias.c (find_base_term): Fix thinko in previous change.
23293
23294 2013-04-08 Jakub Jelinek <jakub@redhat.com>
23295
23296 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
23297 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
23298 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
23299 if possible to compute val.
23300 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
23301 For QImode integers don't require anything about precision. Use
23302 const_with_all_bytes_same to find out if the constant doesn't have
23303 repeated bytes in it.
23304
23305 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
23306
23307 * config/s390/s390.c (s390_expand_insv): Only accept insertions
23308 within mode size.
23309
23310 2013-04-08 Marek Polacek <polacek@redhat.com>
23311
23312 PR rtl-optimization/48182
23313 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
23314 value to 1.
23315
23316 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23317
23318 PR target/55487
23319 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
23320 nuses, make sure we have a label.
23321
23322 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
23323
23324 PR target/56843
23325 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
23326 (rs6000_emit_swdiv_low_precision): Remove.
23327 (rs6000_emit_swdiv): Rewrite to handle between one and four
23328 iterations of Newton-Raphson generally; modify required number of
23329 iterations for some cases.
23330 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
23331
23332 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
23333
23334 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
23335 set-but-unused variable.
23336
23337 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
23338 basic blocks of released function bodies garbage-collectable.
23339
23340 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
23341 (struct rtl_opt_pass): Add TODO_df_finish.
23342
23343 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
23344
23345 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23346
23347 * config/arm/constraints.md (q): New constraint.
23348 * config/arm/ldrdstrd.md: New file.
23349 * config/arm/arm.md (ldrdstrd.md) New include.
23350 (arm_movdi): Use "q" instead of "r" constraint
23351 for double-word memory access.
23352 (movdf_soft_insn): Likewise.
23353 * config/arm/vfp.md (movdi_vfp): Likewise.
23354 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
23355 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
23356 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
23357 (mem_ok_for_ldrd_strd): Likewise.
23358 (output_move_double): Update assertion.
23359
23360 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23361
23362 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
23363
23364 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23365
23366 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
23367 define_insn_and_split.
23368 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
23369
23370 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23371
23372 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
23373 define_insn_and_split.
23374 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
23375 (shiftsi3_compare): New pattern.
23376 (rrx): New pattern.
23377 * config/arm/unspecs.md (UNSPEC_RRX): New.
23378
23379 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23380
23381 * config/arm/arm.md (negdi_extendsidi): New pattern.
23382 (negdi_zero_extendsidi): Likewise.
23383
23384 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23385
23386 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
23387 define_insn_and_split.
23388 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
23389 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
23390
23391 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23392
23393 * config/arm/arm.md (arm_subdi3): Convert define_insn into
23394 define_insn_and_split.
23395 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
23396 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
23397
23398 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23399
23400 * config/arm/arm.md (subsi3_carryin): New pattern.
23401 (subsi3_carryin_const): Likewise.
23402 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
23403 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
23404
23405 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23406
23407 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
23408
23409 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
23410
23411 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
23412 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
23413
23414 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23415
23416 * config/arm/arm.c (arm_expand_builtin): Change fcode
23417 type to unsigned int.
23418
23419 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23420
23421 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
23422
23423 2013-04-04 Ian Lance Taylor <iant@google.com>
23424
23425 * doc/standards.texi (Standards): The Go frontend supports the Go 1
23426 language standard.
23427
23428 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
23429
23430 PR middle-end/56729
23431 * df-scan.c (df_insn_delete): Disable failing assert.
23432
23433 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23434
23435 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
23436 New function prototype.
23437 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
23438 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
23439 (arm_builtin_vectorized_function): New function.
23440
23441 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23442
23443 * config/arm/arm_neon_builtins.def: New file.
23444 * config/arm/arm.c (neon_builtin_data): Move contents to
23445 arm_neon_builtins.def.
23446 (enum arm_builtins): Include neon builtin definitions.
23447 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
23448 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
23449
23450 2013-04-04 Marek Polacek <polacek@redhat.com>
23451
23452 PR tree-optimization/48186
23453 * predict.c (maybe_hot_frequency_p): Return false if
23454 HOT_BB_FREQUENCY_FRACTION is 0.
23455 (cgraph_maybe_hot_edge_p): Likewise.
23456
23457 2013-04-04 Richard Biener <rguenther@suse.de>
23458
23459 PR tree-optimization/56826
23460 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
23461 more accurately.
23462
23463 2013-04-04 Richard Biener <rguenther@suse.de>
23464
23465 PR tree-optimization/56213
23466 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
23467 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
23468
23469 2013-04-04 Richard Biener <rguenther@suse.de>
23470
23471 PR tree-optimization/56837
23472 * tree-loop-distribution.c (classify_partition): For non-zero
23473 values require that the value has the same precision as its
23474 mode to be useful as memset value.
23475
23476 2013-04-03 Nick Clifton <nickc@redhat.com>
23477
23478 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
23479 (fmssf4): Use fmsf.s on E3V5 architectures.
23480 (fnmasf4): Use fnmaf.s on E3V5 architectures.
23481 (fnmssf4): Use fnmsf.s on E3V5 architectures.
23482
23483 2013-04-03 Jeff Law <law@redhat.com>
23484
23485 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
23486 (lra-eliminations.o): Likewise.
23487
23488 2013-04-03 Teresa Johnson <tejohnson@google.com>
23489
23490 * gcov-io.c (compute_working_sets): Moved most of body of old
23491 compute_working_sets here from profile.c.
23492 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
23493 (gcov_working_set_t): Moved typedef here from basic-block.h
23494 (compute_working_set): Declare.
23495 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
23496 (get_working_sets): Renamed from compute_working_set,
23497 replace most of body with call to new compute_working_sets.
23498 (get_exec_counts): Replace call to compute_working_sets
23499 to get_working_sets.
23500 * profile.h (get_working_sets): Renamed from compute_working_set.
23501 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
23502 to get_working_sets.
23503 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
23504 * gcov-dump.c (dump_working_sets): New function.
23505
23506 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
23507
23508 * hwint.c (sext_hwi, zext_hwi): New functions.
23509 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
23510 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
23511 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
23512 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
23513 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
23514 (sext_hwi, zext_hwi): New functions.
23515
23516 2013-04-03 Jeff Law <law@redhat.com>
23517
23518 PR tree-optimization/56799
23519 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
23520 back test for widening conversion erroneously dropped in prior change.
23521
23522 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23523
23524 PR target/56809
23525 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
23526 instead of next_real_insn.
23527
23528 2013-04-03 Marek Polacek <polacek@redhat.com>
23529
23530 PR sanitizer/55702
23531 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
23532
23533 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
23534
23535 PR target/56809
23536 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
23537 next_real_insn.
23538 (thumb1_output_casesi): Likewise.
23539 (thumb2_output_casesi): Likewise.
23540
23541 2013-04-03 Richard Biener <rguenther@suse.de>
23542
23543 PR tree-optimization/56817
23544 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
23545 Split out ...
23546 (tree_unroll_loops_completely_1): ... new function to manually
23547 walk the loop tree, properly defering outer loops of unrolled
23548 loops to later iterations.
23549
23550 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23551
23552 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
23553 (vectorizable_load): Likewise.
23554 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
23555 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
23556
23557 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23558
23559 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
23560 BIT_FIELD_REF.
23561
23562 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
23563
23564 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
23565
23566 2013-04-03 Bin Cheng <bin.cheng@arm.com>
23567
23568 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
23569
23570 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23571
23572 PR tree-optimization/56790
23573 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
23574 folding.
23575
23576 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
23577
23578 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
23579 Handle VEC_MERGE.
23580 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
23581 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
23582 equal arguments.
23583
23584 2013-04-03 Jakub Jelinek <jakub@redhat.com>
23585
23586 PR c/19449
23587 * tree.h (force_folding_builtin_constant_p): New decl.
23588 * builtins.c (force_folding_builtin_constant_p): New variable.
23589 (fold_builtin_constant_p): Fold immediately also if
23590 force_folding_builtin_constant_p.
23591
23592 2013-04-03 Richard Biener <rguenther@suse.de>
23593
23594 PR tree-optimization/56812
23595 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
23596 DRs of the same interleaving chain are independent.
23597
23598 2013-04-02 Jason Merrill <jason@redhat.com>
23599
23600 * gdbinit.in (pbb): Use debug fn.
23601
23602 2013-04-02 Lawrence Crowl <crowl@google.com>
23603
23604 * sese.h (struct ivtype_map_elt_s): Remove unused.
23605 (extern debug_ivtype_map): Remove unused.
23606 (extern eq_ivtype_map_elts): Remove unused.
23607 * sese.c (debug_ivtype_map): Removed unused.
23608 (debug_ivtype_map_1): Removed unused.
23609 (debug_ivtype_elt): Remove unused.
23610 (eq_ivtype_map_elts): Remove unused.
23611
23612
23613 2013-04-02 Kai Tietz <ktietz@redhat.com>
23614
23615 PR target/52790
23616 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
23617 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
23618 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
23619 function.
23620 (legitimize_pe_coff_symbol): Likewise.
23621 (is_imported_p): New helper-function.
23622 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
23623 for Windows x64 targets.
23624 (ix86_expand_prologue): Optimize for pe-coff targets.
23625 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
23626 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
23627 medium/large code-model.
23628 (legitimize_pic_address): Likewise.
23629 (legitimize_tls_address): Likewise.
23630 (ix86_expand_call): Likewise.
23631 (x86_output_mi_thunk): Likewise.
23632 (get_dllimport_decl): Add new beimport argument.
23633 (construct_plt_address): Don't assert for x64 pe-coff targets.
23634 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
23635 targets.
23636 (SYMBOL_FLAG_STUBVAR): New macro.
23637 (SYMBOL_REF_STUBVAR_P): Likewise.
23638 * config/i386/winnt.c (stub_list): New structure.
23639 (stub_head): New local variable.
23640 (i386_pe_record_stub): New function.
23641 (i386_pe_file_end): Emit refptr-stubs.
23642
23643 2013-04-02 Jakub Jelinek <jakub@redhat.com>
23644
23645 PR rtl-optimization/56745
23646 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
23647 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
23648
23649 PR c++/34949
23650 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
23651 and both of them are MEM_REFs, just compare first argument for
23652 equality and attempt to deal even with differing offsets.
23653
23654 PR c++/34949
23655 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
23656 of gimple_clobber_p to be MEM_REF.
23657 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
23658 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
23659 after gimplification.
23660 * asan.c (get_mem_ref_of_assignment): Don't instrument
23661 gimple_clobber_p stmts.
23662 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
23663 gimple_clobber_p stmt if they have MEM_REF lhs and
23664 are dead because of another gimple_clobber_p stmt.
23665 * tree-ssa-live.c (clear_unused_block_pointer): Treat
23666 gimple_clobber_p stmts like debug stmts.
23667 (remove_unused_locals): Remove clobbers with MEM_REF lhs
23668 that refer to unused VAR_DECLs or uninitialized values.
23669 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
23670 gimple_clobber_p stmts if they refer to removed parameters.
23671 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
23672 formatting.
23673
23674 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
23675
23676 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
23677 using SWI48 mode attribute.
23678
23679 2013-04-02 Wei Mi <wmi@google.com>
23680
23681 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
23682 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
23683 *<rotate_insn><mode>3_mask in i386.md.
23684
23685 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
23686
23687 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
23688
23689 2013-04-02 Richard Biener <rguenther@suse.de>
23690
23691 PR tree-optimization/56778
23692 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
23693 Runtime alias tests are not supported for gather loads.
23694 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
23695 stmts referenced from SSA operands before updating SSA form.
23696
23697 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
23698 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23699
23700 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
23701 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
23702 * config/arm/cortex-a53.md: New file.
23703 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
23704 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
23705 * config/arm/arm.c (arm_issue_rate): Likewise.
23706 * config/arm/arm-tune.md: Regenerate
23707 * config/arm/arm-tables.opt: Regenerate.
23708 * config/arm/arm-cores.def: Add cortex-a53.
23709
23710 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
23711
23712 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
23713 non-static link.
23714
23715 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23716
23717 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
23718 scalar load/store operations using B/H registers.
23719 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
23720
23721 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
23722
23723 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
23724 scalar move.
23725 * config/aarch64/aarch64.c
23726 (aarch64_simd_scalar_immediate_valid_for_move): New.
23727 * config/aarch64/aarch64-protos.h
23728 (aarch64_simd_scalar_immediate_valid_for_move): New.
23729 * config/aarch64/constraints.md (Dh, Dq): New.
23730 * config/aarch64/iterators.md (hq): New.
23731
23732 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
23733
23734 * reorg.c (get_branch_condition): Deal with conditional returns.
23735 (fill_simple_delay_slots): Remove dead code dealing with jumps.
23736
23737 2013-04-01 Wei Mi <wmi@google.com>
23738
23739 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
23740 Truncate operand 2 using %b asm operand modifier.
23741 (*<shift_insn><mode>3_mask): Ditto.
23742 (*<rotate_insn><mode>3_mask): Ditto.
23743
23744 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
23745
23746 PR middle-end/56798
23747 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
23748
23749 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
23750
23751 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
23752 of next_real_insn.
23753 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
23754
23755 2013-03-30 Lawrence Crowl <crowl@google.com>
23756
23757 * dse.c (clear_alias_sets): Remove never set.
23758 (disqualified_clear_alias_sets): Remove never set.
23759 (clear_alias_mode_pool): Remove never set.
23760 (dse_step0): Remove condition that is never true.
23761 (canon_address): Remove condition that is never true.
23762 (dse_step7): Remove condition that is never true.
23763 (rest_of_handle_dse): Remove condition that is never true.
23764 (rest_of_handle_dse::did_global): Remove never read from above.
23765 (dse_step2_spill): Remove never called from above.
23766 (dse_step5_spill): Remove never called from above.
23767
23768 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
23769
23770 * doc/md.texi (Standard Names) <casesi>: Update documentation for
23771 JUMP_TABLE_DATA changes.
23772 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
23773 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
23774 (Insns) <jump_table_data>: New entry.
23775 * doc/tm.texi: Regenerate.
23776
23777 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
23778
23779 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
23780 for table jump at the end of a basic block using tablejump_p.
23781 * targhooks.c (default_invalid_within_doloop): Likewise.
23782 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
23783 target hook implementation that is identical to the default hook.
23784 (rs6000_invalid_within_doloop): Remove.
23785
23786 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
23787 unused variable from tablejump_p call.
23788
23789 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
23790 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
23791 (INSN_DELETED_P): Likewise.
23792 (emit_jump_table_data): New prototype.
23793 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
23794 after 4th as unused.
23795 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
23796 * sched-vis.c (print_insn): Likewise.
23797 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
23798 insn for compatibility with back ends that use next_active_insn to
23799 identify jump table data.
23800 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
23801 (remove_insn): Likewise.
23802 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
23803 to be emitted.
23804 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
23805 (emit_jump_table_data): New function.
23806
23807 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
23808 basic block, a JUMP_TABLE_DATA never is.
23809 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
23810 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
23811 off from code handling real insns.
23812 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
23813 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
23814 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
23815 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
23816 is not a NONDEBUG_INSN_P.
23817 * ira-costs.c (scan_one_insn): Likewise.
23818 * jump.c (mark_all_labels): Likewise.
23819 (mark_jump_label_1): Likewise.
23820 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23821 * lra.c (get_insn_freq): Expect all insns reaching here to be in
23822 a basic block.
23823 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
23824 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
23825 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
23826 JUMP_TABLE_DATA_P insns.
23827 (calculate_elim_costs_all_insns): Likewise.
23828 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
23829 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
23830 (delete_output_reload): Code style fixups.
23831 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
23832 insn flags on this non-insn.
23833 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
23834 as scheduling barriers, for pre-change compatibility.
23835 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
23836 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
23837
23838 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
23839 redundant JUMP_TABLE_DATA_P test.
23840 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
23841 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23842 (frv_for_each_packet): Likewise.
23843 * config/i386/i386.c (min_insn_size): Likewise.
23844 (ix86_avoid_jump_mispredicts): Likewise.
23845 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23846 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
23847 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23848 (mips16_insn_length): Robustify.
23849 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
23850 (mips16_split_long_branches): Likewise.
23851 * config/pa/pa.c (pa_combine_instructions): Likewise.
23852 * config/rs6000/rs6000.c (get_next_active_insn): Treat
23853 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
23854 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
23855 as contributing to pool range lengths.
23856 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
23857 Remove redundant JUMP_TABLE_DATA_P test.
23858 (sh_loop_align): Likewise.
23859 (split_branches): Likewise.
23860 (sh_insn_length_adjustment): Likewise.
23861 * config/spu/spu.c (get_branch_target): Likewise.
23862
23863 2013-03-29 Jan Hubicka <jh@suse.cz>
23864
23865 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
23866 gcov streaming; stream hot bb threshold to ltrans.
23867 * predict.c (get_hot_bb_threshold): Break out from ....
23868 (maybe_hot_count_p): ... here.
23869 (set_hot_bb_threshold): New function.
23870 * lto-section-in.c (lto_section_name): Add profile.
23871 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
23872 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
23873 and data-streamer.h
23874 (histogram_entry): New structure.
23875 (histogram, histogram_pool): New global vars.
23876 (histogram_hash): New structure.
23877 (histogram_hash::hash): New method.
23878 (histogram_hash::equal): Likewise.
23879 (account_time_size): New function.
23880 (cmp_counts): New function.
23881 (dump_histogram): New function.
23882 (ipa_profile_generate_summary): New function.
23883 (ipa_profile_write_summary): New function.
23884 (ipa_profile_read_summary): New function.
23885 (ipa_profile): Decide on threshold.
23886 (pass_ipa_profile): Add ipa_profile_write_summary and
23887 ipa_profile_read_summary.
23888 * Makefile.in (ipa.o): Update dependencies.
23889 * lto-streamer.h (LTO_section_ipa_profile): New section.
23890
23891 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
23892
23893 * tree.h (VAR_P): New.
23894
23895 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
23896
23897 PR lto/56777
23898 * doc/invoke.texi ([-fwhole-program]): Fix typo.
23899
23900 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
23901
23902 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
23903 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23904 (control_flow_insn_p): Likewise.
23905 * cfgrtl.c (duplicate_insn_chain): Likewise.
23906 * final.c (get_attr_length_1): Likewise.
23907 (shorten_branches): Likewise.
23908 (final_scan_insn): Likewise.
23909 * function.c (instantiate_virtual_regs): Likewise.
23910 * gcse.c (insert_insn_end_basic_block): Likewise.
23911 * ira-costs.c (scan_one_insn): Likewise.
23912 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
23913 * lra.c (check_rtl): Likewise.
23914 * reload1.c (elimination_costs_in_insn): Likewise.
23915 * reorg.c (follow_jumps): Likewise.
23916
23917 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
23918 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
23919 (thumb_far_jump_used_p): Likewise.
23920 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
23921 (workaround_speculation): Likewise.
23922 (add_sched_insns_for_speculation): Likewise.
23923 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
23924 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
23925 (frv_for_each_packet): Likewise.
23926 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
23927 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
23928 (final_emit_insn_group_barriers): Likewise.
23929 * config/m32r/m32r.c (m32r_is_insn): Likewise.
23930 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
23931 (mips16_insn_length): Likewise.
23932 * config/pa/pa.c (pa_reorg): Likewise.
23933 (pa_combine_instructions): Likewise.
23934 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
23935 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
23936 (sh_reorg): Likewise.
23937 (split_branches): Likewise.
23938 * config/spu/spu.c (get_branch_target): Likewise.
23939
23940 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
23941 JUMP_TABLE_DATA_P.
23942
23943 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
23944
23945 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
23946 Fix declaration name.
23947
23948 2013-03-28 Lawrence Crowl <crowl@google.com>
23949
23950 * graphds.h (struct graph.indicies): Remove unused.
23951 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
23952 (SCOP_ORIGINAL_PDDRS): Remove unused.
23953 * sese.h (extern insert_loop_close_phis): Removed unused.
23954 (extern insert_guard_phis): Removed unused.
23955 (extern ivtype_map_elt_info): Removed unused.
23956 (new_ivtype_map_elt): Removed unused.
23957 * sese.c (ivtype_map_elt_info): Removed unused.
23958
23959 2013-03-28 Lawrence Crowl <crowl@google.com>
23960
23961 * Makefile.in: Add several missing include dependences.
23962 (DUMPFILE_H): New.
23963 (test-dump.o): New. This object is not added to any executable,
23964 but is present for ad-hoc testing.
23965 * bitmap.c
23966 (debug (const bitmap_head_def &)): New.
23967 (debug (const bitmap_head_def *)): New.
23968 * bitmap.h
23969 (extern debug (const bitmap_head_def &)): New.
23970 (extern debug (const bitmap_head_def *)): New.
23971 * cfg.c
23972 (debug (edge_def &)): New.
23973 (debug (edge_def *)): New.
23974 * cfghooks.c
23975 (debug (basic_block_def &)): New.
23976 (debug (basic_block_def *)): New.
23977 * dumpfile.h
23978 (dump_node (const_tree, int, FILE *)): Correct source file.
23979 * dwarf2out.c
23980 (debug (die_struct &)): New.
23981 (debug (die_struct *)): New.
23982 * dwarf2out.h
23983 (extern debug (die_struct &)): New.
23984 (extern debug (die_struct *)): New.
23985 * gimple-pretty-print.c
23986 (debug (gimple_statement_d &)): New.
23987 (debug (gimple_statement_d *)): New.
23988 * gimple-pretty-print.h
23989 (extern debug (gimple_statement_d &)): New.
23990 (extern debug (gimple_statement_d *)): New.
23991 * ira-build.c
23992 (debug (ira_allocno_copy &)): New.
23993 (debug (ira_allocno_copy *)): New.
23994 (debug (ira_allocno &)): New.
23995 (debug (ira_allocno *)): New.
23996 * ira-int.h
23997 (extern debug (ira_allocno_copy &)): New.
23998 (extern debug (ira_allocno_copy *)): New.
23999 (extern debug (ira_allocno &)): New.
24000 (extern debug (ira_allocno *)): New.
24001 * ira-lives.c
24002 (debug (live_range &)): New.
24003 (debug (live_range *)): New.
24004 * lra-int.h
24005 (debug (lra_live_range &)): New.
24006 (debug (lra_live_range *)): New.
24007 * lra-lives.c
24008 (debug (lra_live_range &)): New.
24009 (debug (lra_live_range *)): New.
24010 * omega.c
24011 (debug (omega_pb_d &)): New.
24012 (debug (omega_pb_d *)): New.
24013 * omega.h
24014 (extern debug (omega_pb_d &)): New.
24015 (extern debug (omega_pb_d *)): New.
24016 * print-rtl.c
24017 (debug (const rtx_def &)): New.
24018 (debug (const rtx_def *)): New.
24019 * print-tree.c
24020 (debug_tree (tree): Move within file.
24021 (debug_raw (const tree_node &)): New.
24022 (debug_raw (const tree_node *)): New.
24023 (dump_tree_via_hooks (const tree_node *, int)): New.
24024 (debug (const tree_node &)): New.
24025 (debug (const tree_node *)): New.
24026 (debug_verbose (const tree_node &)): New.
24027 (debug_verbose (const tree_node *)): New.
24028 (debug_head (const tree_node &)): New.
24029 (debug_head (const tree_node *)): New.
24030 (debug_body (const tree_node &)): New.
24031 (debug_body (const tree_node *)): New.
24032 (debug_vec_tree (tree): Move and reimplement in terms of dump.
24033 (debug (vec<tree, va_gc> &)): New.
24034 (debug (vec<tree, va_gc> *)): New.
24035 * rtl.h
24036 (extern debug (const rtx_def &)): New.
24037 (extern debug (const rtx_def *)): New.
24038 * sbitmap.c
24039 (debug_raw (simple_bitmap_def &)): New.
24040 (debug_raw (simple_bitmap_def *)): New.
24041 (debug (simple_bitmap_def &)): New.
24042 (debug (simple_bitmap_def *)): New.
24043 * sbitmap.h
24044 (extern debug (simple_bitmap_def &)): New.
24045 (extern debug (simple_bitmap_def *)): New.
24046 (extern debug_raw (simple_bitmap_def &)): New.
24047 (extern debug_raw (simple_bitmap_def *)): New.
24048 * sel-sched-dump.c
24049 (debug (vinsn_def &)): New.
24050 (debug (vinsn_def *)): New.
24051 (debug_verbose (vinsn_def &)): New.
24052 (debug_verbose (vinsn_def *)): New.
24053 (debug (expr_def &)): New.
24054 (debug (expr_def *)): New.
24055 (debug_verbose (expr_def &)): New.
24056 (debug_verbose (expr_def *)): New.
24057 (debug (vec<rtx> &)): New.
24058 (debug (vec<rtx> *)): New.
24059 * sel-sched-dump.h
24060 (extern debug (vinsn_def &)): New.
24061 (extern debug (vinsn_def *)): New.
24062 (extern debug_verbose (vinsn_def &)): New.
24063 (extern debug_verbose (vinsn_def *)): New.
24064 (extern debug (expr_def &)): New.
24065 (extern debug (expr_def *)): New.
24066 (extern debug_verbose (expr_def &)): New.
24067 (extern debug_verbose (expr_def *)): New.
24068 (extern debug (vec<rtx> &)): New.
24069 (extern debug (vec<rtx> *)): New.
24070 * sel-sched-ir.h
24071 (_list_iter_cond_expr): Make inline instead of static.
24072 * sreal.c
24073 (debug (sreal &)): New.
24074 (debug (sreal *)): New.
24075 * sreal.h
24076 (extern debug (sreal &)): New.
24077 (extern debug (sreal *)): New.
24078 * tree.h
24079 (extern debug_raw (const tree_node &)): New.
24080 (extern debug_raw (const tree_node *)): New.
24081 (extern debug (const tree_node &)): New.
24082 (extern debug (const tree_node *)): New.
24083 (extern debug_verbose (const tree_node &)): New.
24084 (extern debug_verbose (const tree_node *)): New.
24085 (extern debug_head (const tree_node &)): New.
24086 (extern debug_head (const tree_node *)): New.
24087 (extern debug_body (const tree_node &)): New.
24088 (extern debug_body (const tree_node *)): New.
24089 (extern debug (vec<tree, va_gc> &)): New.
24090 (extern debug (vec<tree, va_gc> *)): New.
24091 * tree-cfg.c
24092 (debug (struct loop &)): New.
24093 (debug (struct loop *)): New.
24094 (debug_verbose (struct loop &)): New.
24095 (debug_verbose (struct loop *)): New.
24096 * tree-dump.c: Add header dependence.
24097 * tree-flow.h
24098 (extern debug (struct loop &)): New.
24099 (extern debug (struct loop *)): New.
24100 (extern debug_verbose (struct loop &)): New.
24101 (extern debug_verbose (struct loop *)): New.
24102 * tree-data-ref.c
24103 (debug (data_reference &)): New.
24104 (debug (data_reference *)): New.
24105 (debug (vec<data_reference_p> &)): New.
24106 (debug (vec<data_reference_p> *)): New.
24107 (debug (vec<ddr_p> &)): New.
24108 (debug (vec<ddr_p> *)): New.
24109 * tree-data-ref.h
24110 (extern debug (data_reference &)): New.
24111 (extern debug (data_reference *)): New.
24112 (extern debug (vec<data_reference_p> &)): New.
24113 (extern debug (vec<data_reference_p> *)): New.
24114 (extern debug (vec<ddr_p> &)): New.
24115 (extern debug (vec<ddr_p> *)): New.
24116 * tree-ssa-alias.c
24117 (debug (pt_solution &)): New.
24118 (debug (pt_solution *)): New.
24119 * tree-ssa-alias.h
24120 (extern debug (pt_solution &)): New.
24121 (extern debug (pt_solution *)): New.
24122 * tree-ssa-alias.c
24123 (debug (_var_map &)): New.
24124 (debug (_var_map *)): New.
24125 (debug (tree_live_info_d &)): New.
24126 (debug (tree_live_info_d *)): New.
24127 * tree-ssa-alias.h
24128 (extern debug (_var_map &)): New.
24129 (extern debug (_var_map *)): New.
24130 (extern debug (tree_live_info_d &)): New.
24131 (extern debug (tree_live_info_d *)): New.
24132
24133 2013-03-28 Jan Hubicka <jh@suse.cz>
24134
24135 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
24136
24137 2013-03-28 Ian Bolton <ian.bolton@arm.com>
24138
24139 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
24140 record only when desired or required.
24141
24142 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
24143
24144 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
24145 *vec_extractv2di_1_rex64. Use x64 isa attribute.
24146
24147 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24148
24149 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
24150 (*andsi3_compare0_uxtw): New pattern.
24151 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
24152 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
24153
24154 2013-03-28 Jan Hubicka <jh@suse.cz>
24155
24156 * data-streamer-in.c (streamer_read_gcov_count): New function.
24157 * gimple-streamer-out.c: Include value-prof.h.
24158 (output_gimple_stmt): Output histogram.
24159 (output_bb): Use streamer_write_gcov_count.
24160 * value-prof.c: Include data-streamer.h
24161 (dump_histogram_value): Add HIST_TYPE_MAX.
24162 (stream_out_histogram_value): New function.
24163 (stream_in_histogram_value): New function.
24164 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
24165 (stream_out_histogram_value, stream_in_histogram_value): Declare.
24166 * data-streamer-out.c (streamer_write_gcov_count): New function.
24167 (streamer_write_gcov_count_stream): New function.
24168 * lto-cgraph.c (lto_output_edge): Update counter streaming.
24169 (lto_output_node): Likewise.
24170 (input_node, input_edge): Likewise.
24171 * lto-streamer-out.c (output_cfg): Update streaming.
24172 * lto-streamer-in.c (input_cfg): Likewise.
24173 * data-streamer.h (streamer_write_gcov_count,
24174 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
24175 * gimple-streamer-in.c: Include value-prof.h
24176 (input_gimple_stmt): Input histograms.
24177 (input_bb): Update profile streaming.
24178
24179 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24180
24181 * genmodes.c (emit_max_int): New function.
24182 (emit_insn_modes_h): Added call to emit_max_function.
24183 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
24184 Added doc.
24185 * machmode.def: Fixed comment.
24186
24187 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
24188
24189 * combine.c (try_combine): Removed useless assert.
24190 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
24191
24192 2013-03-28 Marek Polacek <polacek@redhat.com>
24193 Richard Biener <rguenther@suse.de>
24194
24195 PR tree-optimization/56695
24196 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
24197 build signed result of a vector comparison.
24198 * tree-cfg.c (verify_gimple_comparison): Check that a result
24199 of a vector comparison has signed type.
24200
24201 2013-03-28 Richard Biener <rguenther@suse.de>
24202
24203 PR tree-optimization/37021
24204 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
24205 do not restrict gaps between groups.
24206 * tree-vect-stmts.c (vectorizable_load): Properly account for
24207 a gap between groups.
24208
24209 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
24210
24211 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
24212 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
24213 is not enabled.
24214
24215 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
24216
24217 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
24218 * doc/extend.texi (Named Address Spaces): Ditto.
24219 (Variable Attributes): Ditto.
24220
24221 2013-03-27 Kai Tietz <ktietz@redhat.com>
24222
24223 * config.build: Add support for cygwin x64 target.
24224 * config.gcc: Likewise.
24225 * config.host: Likewise.
24226 * configure.ac: Likewise
24227 * configure: Regenerated.
24228
24229 2013-03-27 Kai Tietz <ktietz@redhat.com>
24230
24231 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
24232 * config/i386/t-cygwin-w64: New file.
24233 * config/i386/cygwin-w64.h: New file.
24234 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
24235 and add support for x64-cygwin target.
24236 (CPP_SPEC): Likewise.
24237 (CXX_WRAP_SPEC_LIST): Undefine before define.
24238 (LIBGCJ_SONAME): Use 15 as version.
24239
24240 2013-03-27 Richard Biener <rguenther@suse.de>
24241
24242 PR tree-optimization/56716
24243 * tree-ssa-structalias.c (perform_var_substitution): Adjust
24244 dumping for ref nodes.
24245
24246 2013-03-27 Martin Jambor <mjambor@suse.cz>
24247
24248 PR tree-optimization/55334
24249 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
24250 restricted pointers to arrays.
24251
24252 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
24253
24254 * Makefile.in (.SUFFIXES): Add .cc.
24255 (.c.o): Apply same recipe for implicit rule .cc.o.
24256
24257 2013-03-27 Richard Biener <rguenther@suse.de>
24258
24259 PR tree-optimization/37021
24260 * tree-vect-data-refs.c (vect_check_strided_load): Allow
24261 REALPART/IMAGPART_EXPRs around the supported refs.
24262 * tree-ssa-structalias.c (find_func_aliases): Assume that
24263 floating-point values are not used to transfer pointers.
24264
24265 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
24266
24267 * target.def (TARGET_HAS_IFUNC_P): New target hook.
24268 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
24269 * doc/tm.texi: Regenerate.
24270 * targhooks.h (default_has_ifunc_p): New.
24271 * targhooks.c (default_has_ifunc_p): Ditto.
24272 * config/linux-protos.h: New file.
24273 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
24274 hook for linux which disables support of indirect functions in android.
24275 * config/linux-android.c: New file.
24276 * config/t-linux-android.c: Ditto.
24277 * config.gcc: Added new object file linux-android.o.
24278 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
24279 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
24280 * varasm.c (do_assemble_alias): Likewise.
24281 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
24282 doesn't support indirect functions.
24283 * configure: Regenerate.
24284
24285 2013-03-27 Bin Cheng <bin.cheng@arm.com>
24286
24287 PR target/56102
24288 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
24289 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
24290 mult-word mode.
24291
24292 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
24293
24294 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
24295
24296 2013-03-27 Terry Guo <terry.guo@arm.com>
24297
24298 * config/arm/arm-cores.def: Added core cortex-r7.
24299 * config/arm/arm-tune.md: Regenerated.
24300 * config/arm/arm-tables.opt: Regenerated.
24301 * doc/invoke.texi: Added entry for core cortex-r7.
24302
24303 2013-03-27 Walter Lee <walt@tilera.com>
24304
24305 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
24306 double-decrement of next_scratch_regno.
24307
24308 2013-03-27 Walter Lee <walt@tilera.com>
24309
24310 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
24311 input operands.
24312 (insn_v1mulus): Ditto.
24313 (insn_v2muls): Ditto.
24314
24315 2013-03-27 Walter Lee <walt@tilera.com>
24316
24317 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
24318 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
24319
24320 2013-03-27 Walter Lee <walt@tilera.com>
24321
24322 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
24323 (*sibcall_value): Ditto.
24324
24325 2013-03-27 Walter Lee <walt@tilera.com>
24326
24327 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
24328 (insn_mnz_v8qi): ... this ...
24329 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
24330 vector equivalent.
24331 (insn_v<n>mnz): Replaced by ...
24332 (insn_v1mnz): ... this ...
24333 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
24334 equivalent.
24335 (insn_mz_<mode>): Replaced by ...
24336 (insn_mz_v8qi): ... this ...
24337 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
24338 vector equivalent.
24339 (insn_v<n>mz): Replaced by ...
24340 (insn_v1mz): ... this ...
24341 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
24342 equivalent.
24343
24344 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
24345
24346 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
24347
24348 2013-03-26 Roland McGrath <mcgrathr@google.com>
24349
24350 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
24351 than fprintf with a non-constant, non-format string.
24352
24353 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24354
24355 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
24356 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
24357 operand 0 predicate.
24358 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
24359 attribute. Use general_x64nomem_operand as operand 1 predicate.
24360 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
24361 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24362 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
24363 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
24364 (mov<mode>_insv_1): Remove expander. Merge insn with
24365 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
24366 Use general_x64nomem_operand as operand 1 predicate.
24367 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
24368 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
24369 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24370 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
24371 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
24372 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
24373 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
24374 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
24375 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
24376 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
24377 (general_x64nomem_operand): Ditto.
24378
24379 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
24380
24381 * config/rtems.opt: Add -pthread option.
24382
24383 2013-03-26 Richard Biener <rguenther@suse.de>
24384
24385 * alias.c (find_base_term): Avoid redundant and not used recursion.
24386 (base_alias_check): Get the initial base term from the caller.
24387 (true_dependence_1): Compute and pass base terms to base_alias_check.
24388 (write_dependence_p): Likewise.
24389 (may_alias_p): Likewise.
24390
24391 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
24392
24393 * config/aarch64/aarch64.c (aarch64_classify_address): Support
24394 PC-relative load in SI modes and above only.
24395
24396 2013-03-26 Xinyu Qi <xyqi@marvell.com>
24397
24398 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
24399 * config/arm/iwmmxt.md (WCGR0): Update.
24400 (WCGR1, WCGR2, WCGR3): Likewise.
24401
24402 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
24403
24404 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
24405 Use x64 and nox64 isa attributes.
24406
24407 2013-03-26 Richard Biener <rguenther@suse.de>
24408
24409 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
24410 alignment computations and rely on get_object_alignment_1
24411 for the !TYPE_P case.
24412 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
24413
24414 2013-03-26 Walter Lee <walt@tilera.com>
24415
24416 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
24417 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
24418
24419 2013-03-25 Jeff Law <law@redhat.com>
24420
24421 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
24422 check for INTEGRAL_TYPE_P that was missing due to checking in
24423 wrong version of prior patch.
24424
24425 2013-03-25 Walter Lee <walt@tilera.com>
24426
24427 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
24428 TILEGX_INSN_SHUFFLEBYTES1.
24429 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
24430 shufflebytes1.
24431 (tilegx_builtins): Ditto.
24432 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
24433
24434 2013-03-25 Walter Lee <walt@tilera.com>
24435
24436 * config/tilegx/tilegx.md (floatsisf2): New pattern.
24437 (floatunssisf2): New pattern.
24438 (floatsidf2): New pattern.
24439 (floatunssidf2): New pattern.
24440
24441 2013-03-25 Walter Lee <walt@tilera.com>
24442
24443 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
24444 tests for constraint J, K, N, P.
24445
24446 2013-03-25 Walter Lee <walt@tilera.com>
24447
24448 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
24449 Use indirect/pcrel encoding.
24450 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
24451 Ditto.
24452
24453 2013-03-25 Steve Ellcey <sellcey@mips.com>
24454
24455 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
24456 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
24457 * config/mips/mips.c (mips_option_override): Set IMADD default.
24458 * config/mips/mips.h (PTF_AVOID_IMADD): New.
24459 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
24460 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
24461 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
24462 * doc/invoke.texi (-mimadd/-mno-imadd): New.
24463
24464 2013-03-25 Jeff Law <law@redhat.com>
24465
24466 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
24467 slightly to avoid creating and folding useless trees. Simplify
24468 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
24469
24470 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
24471
24472 * config/i386/i386.md (*zero_extendsidi2): Merge with
24473 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
24474 * config/i386/predicates.md (x86_64_zext_operand): Rename from
24475 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
24476 targets. Clarify comment.
24477
24478 2013-03-25 Martin Jambor <mjambor@suse.cz>
24479
24480 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
24481 pass-through jump functions differently.
24482 (ipa_read_jump_function): Likewise. Also use setter functions to set
24483 up jump functions.
24484
24485 2013-03-25 Martin Jambor <mjambor@suse.cz>
24486
24487 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
24488 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
24489 process it.
24490 (ipa_get_indirect_edge_target): New function.
24491 (devirtualization_time_bonus): New parameter known_aggs, pass it to
24492 ipa_get_indirect_edge_target. Update all callers.
24493 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
24494 ipa_get_indirect_edge_target_1 instead of calling
24495 ipa_get_indirect_edge_target.
24496 (create_specialized_node): Pass aggvlas to
24497 ipcp_discover_new_direct_edges.
24498
24499 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24500
24501 * config/arm/arm.md (f_sels, f_seld): New types.
24502 (*cmov<mode>): New pattern.
24503 * config/arm/predicates.md (arm_vsel_comparison_operator): New
24504 predicate.
24505
24506 2013-03-25 Kai Tietz <ktietz@redhat.com>
24507
24508 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
24509 POSIX-printf for mingw-hosted builds.
24510
24511 2013-03-25 Richard Biener <rguenther@suse.de>
24512
24513 PR middle-end/56694
24514 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
24515 must-not-throw stmt location.
24516
24517 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24518
24519 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
24520 Emit load-acquire versions when acq is true.
24521 (arm_emit_store_exclusive): Add rel parameter.
24522 Emit store-release versions when rel is true.
24523 (arm_split_compare_and_swap): Use acquire-release instructions
24524 instead.
24525 of barriers when appropriate.
24526 (arm_split_atomic_op): Likewise.
24527 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
24528 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
24529 (VUNSPEC_SLX): Likewise.
24530 (VUNSPEC_LDA): Likewise.
24531 (VUNSPEC_STL): Likewise.
24532 * config/arm/sync.md (atomic_load<mode>): New pattern.
24533 (atomic_store<mode>): Likewise.
24534 (arm_load_acquire_exclusive<mode>): Likewise.
24535 (arm_load_acquire_exclusivesi): Likewise.
24536 (arm_load_acquire_exclusivedi): Likewise.
24537 (arm_store_release_exclusive<mode>): Likewise.
24538
24539 2013-03-25 Catherine Moore <clm@codesourcery.com>
24540
24541 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
24542 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
24543 * config/mip/predicates.md (lwsp_swsp_operand,
24544 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
24545 sb16_operand, db4_operand, db7_operand, ib3_operand,
24546 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
24547 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
24548 andi16_operand): New predicates.
24549 * config/mips/mips.md (compression): New attribute.
24550 (enabled): New attribute.
24551 (length): Consider compression in computing length.
24552 (shift_compression): New code attribute.
24553 (*add<mode>3): New operands. Record compression.
24554 (sub<mode>3): Likewise.
24555 (one_cmpl<mode>2): Likewise.
24556 (*and<mode>3): Likewise.
24557 (*ior<mode>3): Likewise.
24558 (unnamed pattern for xor): Likewise.
24559 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
24560 (*<optab><mode>3): Likewise.
24561 (*mov<mode>_internal: Likewise.
24562 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
24563 (mips_unsigned_immediate_p): New.
24564 (umips_lwsp_swsp_address_p): New.
24565 (m16_based_address_p): New.
24566 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
24567 (mips_unsigned_immediate_p): New prototype.
24568 (lwsp_swsp_address_p): New prototype.
24569 (m16_based_address_p): New prototype.
24570 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
24571 (mips_signed_immediate_p): New function.
24572 (m16_based_address_p): New function.
24573 (lwsp_swsp_address_p): New function.
24574 (mips_print_operand_punctuation): Recognize short delay slot insns
24575 for microMIPS.add<mode>3"
24576
24577 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
24578
24579 PR target/56720
24580 * config/arm/iterators.md (v_cmp_result): New mode attribute.
24581 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
24582
24583 2013-03-25 Richard Biener <rguenther@suse.de>
24584
24585 PR tree-optimization/56689
24586 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
24587 any edge.
24588
24589 2013-03-25 Richard Biener <rguenther@suse.de>
24590
24591 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
24592 of bitmap.
24593 (memory_references): Likewise.
24594 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
24595 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
24596 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
24597 (gather_mem_refs_in_loops): Fold into ...
24598 (analyze_memory_references): ... this. Move initialization
24599 to tree_ssa_lim_initialize.
24600 (fill_always_executed_in): Rename to ...
24601 (fill_always_executed_in_1): ... this.
24602 (fill_always_executed_in): Move contains_call computation to
24603 this new function from ...
24604 (tree_ssa_lim_initialize): ... here.
24605 (tree_ssa_lim): Call fill_always_executed_in.
24606
24607 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24608
24609 * postreload.c (reload_combine): Fix code detecting returns.
24610
24611 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
24612
24613 * function.c (emit_use_return_register_into_block): On cc0 targets,
24614 do not emit the sequence between cc0 setter and user.
24615
24616 2013-03-25 Kai Tietz <ktietz@redhat.com>
24617
24618 * config/i386/predicates.md (local_symbolic_operand): Interpret
24619 dll-imported symbols as none-local.
24620
24621 2013-03-25 Richard Biener <rguenther@suse.de>
24622
24623 * tree-ssa-loop-im.c (struct depend): Remove.
24624 (struct lim_aux_data): Make depends a vec of gimples.
24625 (free_lim_aux_data): Adjust.
24626 (add_dependency): Likewise.
24627 (set_level): Likewise.
24628
24629 2013-03-25 Richard Biener <rguenther@suse.de>
24630
24631 PR middle-end/56434
24632 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
24633 the pointer returned by calls with ECF_MALLOC set.
24634
24635 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24636
24637 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
24638
24639 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
24640
24641 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
24642 using MMXMODE mode iterator.
24643 (*move<mode>_internal): Merge with *movv2sf_internal and
24644 *movv2sf_internal_rex64 using MMXMODE mode iterator.
24645
24646 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24647
24648 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
24649 (record_last_mem_set_info): Likewise.
24650
24651 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
24652 of XNEWVEC followed by memset.
24653 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
24654
24655 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
24656
24657 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
24658 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
24659 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
24660 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
24661 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
24662 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
24663 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
24664 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
24665 BARRIER_P instead of GET_CODE.
24666
24667 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24668
24669 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
24670 inaccuracy in the probing code.
24671
24672 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
24673 (ctrapdi4): Likewise.
24674
24675 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24676
24677 * calls.c (expand_call): Add missing guard to code handling return
24678 of non-BLKmode structures in MSB.
24679 * function.c (expand_function_end): Likewise.
24680
24681 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
24682
24683 * combine.c (try_combine): Adjust comment. Do not add the set of
24684 insn #0 if the destination indirectly is set or dies in insn #2.
24685 Tidy up code to distribute a new note.
24686
24687 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24688
24689 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
24690 also for alternatives 16 and 17.
24691
24692 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
24693
24694 * config/i386/sse.md (*mov<mode>_internal): Merge with
24695 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
24696 Emit insn template depending on type attribute. Use
24697 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24698 movd instead of movq mnemonic for interunit moves. Rewrite mode
24699 attribute calculation. Remove unit attribute calculation.
24700 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
24701 Set prefix_data16 attribute for DImode ssemov types.
24702 Use Ym instead of y for SSE-MMX conversion alternatives.
24703 Reorder operand constraints.
24704
24705 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
24706
24707 * df.h (df_insn_delete): Adjust prototype.
24708 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
24709 and let it decide whether mark the basic block dirty.
24710 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
24711 * df-scan.c (df_insn_info_delete): New helper function, split
24712 off from df_insn_delete.
24713 (df_scan_free_bb_info): Use it.
24714 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
24715 Likewise.
24716 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
24717 that the insn is actually an insn and it has a non-NULL basic block.
24718 Do not mark basic block dirty if only deleting a DEBUG_INSN.
24719
24720 2013-03-22 Richard Biener <rguenther@suse.de>
24721
24722 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
24723 dep_ref members.
24724 (mem_ref_alloc): Do not allocate them.
24725 (refs_independent_p): Do not query or maintain a cache.
24726
24727 2013-03-22 Richard Biener <rguenther@suse.de>
24728
24729 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
24730 (gather_mem_refs_in_loops): Do not compute it.
24731 (analyze_memory_references): Do not allocate it.
24732 (tree_ssa_lim_finalize): Do not free it.
24733 (for_all_locs_in_loop): Do not query all_refs_in_loop.
24734
24735 2013-03-22 Richard Biener <rguenther@suse.de>
24736
24737 * is-a.h (as_a): Use gcc_checking_assert.
24738
24739 2013-03-22 Ian Bolton <ian.bolton@arm.com>
24740
24741 * config/aarch64/aarch64.c (aarch64_print_operand): New
24742 format specifier for printing a constant in hex.
24743 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
24744 format specifier for printing second operand.
24745
24746 2013-03-22 Richard Biener <rguenther@suse.de>
24747
24748 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
24749 bitmaps.
24750 (gather_mem_refs_in_loops): Perform store accumulation here.
24751 (create_vop_ref_mapping_loop): Remove.
24752 (create_vop_ref_mapping): Likewise.
24753 (analyze_memory_references): Initialize refs_stored_in_loop.
24754 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
24755 (record_indep_loop): Remove.
24756 (record_dep_loop): New function.
24757 (ref_indep_loop_p_1): Adjust to only walk over references
24758 in the loop, not its subloops.
24759 (ref_indep_loop_p): Rename to ...
24760 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
24761 maintaining a more fine-grained cache.
24762 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
24763 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
24764
24765 2013-03-22 Richard Biener <rguenther@suse.de>
24766
24767 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
24768 (struct mem_ref): Make accesses_in_loop a vec of a vec of
24769 aggregate mem_ref_loc.
24770 (free_mem_ref_locs): Inline into ...
24771 (memref_free): ... this and adjust.
24772 (mem_ref_alloc): Adjust.
24773 (mem_ref_locs_alloc): Remove.
24774 (record_mem_ref_loc): Adjust.
24775 (get_all_locs_in_loop): Rewrite into ...
24776 (for_all_locs_in_loop): ... this iterator.
24777 (rewrite_mem_ref_loc): New functor.
24778 (rewrite_mem_refs): Use for_all_locs_in_loop.
24779 (sm_set_flag_if_changed): New functor.
24780 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
24781 (ref_always_accessed): New functor.
24782 (ref_always_accessed_p): Use for_all_locs_in_loop.
24783
24784 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24785
24786 * tree-pass.h (PROP_gimple_lvec): New.
24787 * passes.c (dump_properties): Handle PROP_gimple_lvec.
24788 (init_optimization_passes): Move pass_lower_vector.
24789 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
24790 PROP_gimple_lvec.
24791 (pass_lower_vector): Provide PROP_gimple_lvec.
24792 (pass_lower_vector_ssa): Likewise.
24793 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
24794
24795 2013-03-21 Mark Wielaard <mjw@redhat.com>
24796
24797 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
24798
24799 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24800
24801 * config/i386/i386.md (*movdi_internal): Disparage slightly
24802 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
24803 conversion alternatives.
24804
24805 2013-03-21 Jakub Jelinek <jakub@redhat.com>
24806
24807 PR middle-end/48087
24808 * diagnostic.def (DK_WERROR): New kind.
24809 * diagnostic.h (werrorcount): Define.
24810 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
24811 promoted to DK_ERROR, increment DK_WERROR counter instead of
24812 DK_ERROR counter.
24813 * toplev.c (toplev_main): Call print_ignored_options even if
24814 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
24815 even if just werrorcount is non-zero.
24816
24817 PR debug/55608
24818 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
24819 on failure.
24820 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
24821 (string_cst_pool_decl): New function.
24822 (optimize_one_addr_into_implicit_ptr): New function.
24823 (resolve_addr_in_expr): Optimize DWARF location expression
24824 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
24825 which doesn't live in memory, but has DW_AT_location or
24826 DW_AT_const_value, or refers to a string literal, into
24827 DW_OP_GNU_implicit_pointer.
24828 (optimize_location_into_implicit_ptr): New function.
24829 (resolve_addr): If removing DW_AT_location of a variable because
24830 it was DW_OP_addr of address of the variable, but the variable doesn't
24831 live in memory, try to emit const value attribute for the initializer.
24832
24833 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24834
24835 * tree.h (VECTOR_TYPE_P): New macro.
24836 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
24837 TYPE_MODE): Use it.
24838 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
24839 VEC_COND_EXPR cannot be lvalues.
24840 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
24841
24842 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
24843
24844 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
24845 Restrict the transformation to equal modes.
24846
24847 2013-03-21 Richard Biener <rguenther@suse.de>
24848
24849 PR tree-optimization/39326
24850 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
24851 (MEM_ANALYZABLE): Adjust.
24852 (record_mem_ref_loc): Move bitmap ops ...
24853 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
24854 unanalyzable refs, do not record locations for it.
24855 (analyze_memory_references): Allocate ref zero as shared
24856 unanalyzable ref.
24857 (refs_independent_p): Do not test for unanalyzed mems here.
24858 (ref_indep_loop_p_1): Special-case disambiguation against
24859 the unanalyzed ref.
24860 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
24861
24862 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
24863
24864 * config/arm/arm-protos.h (tune_params): Add
24865 prefer_neon_for_64bits field.
24866 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
24867 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
24868 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
24869 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
24870 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
24871 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
24872 (arm_option_override): Handle -mneon-for-64bits new option.
24873 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
24874 (prefer_neon_for_64bits): Declare new variable.
24875 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
24876 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
24877 (arch_enabled): Handle new arch types. Remove support for onlya8
24878 and nota8.
24879 (one_cmpldi2): Use new arch names.
24880 (zero_extend<mode>di2, extend<mode>di2): Ditto.
24881 * config/arm/arm.opt (mneon-for-64bits): Add option.
24882 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
24883 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
24884 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
24885 of onlya8.
24886 * doc/invoke.texi (-mneon-for-64bits): Document.
24887
24888 2013-03-21 Richard Biener <rguenther@suse.de>
24889
24890 PR tree-optimization/39326
24891 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
24892 (sort_bbs_in_loop_postorder_cmp): New function.
24893 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
24894
24895 2013-03-21 Richard Biener <rguenther@suse.de>
24896
24897 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
24898 (vect_insert_into_interleaving_chain): Likewise.
24899 (vect_drs_dependent_in_basic_block): Inline ...
24900 (vect_slp_analyze_data_ref_dependence): ... here. New function,
24901 split out from ...
24902 (vect_analyze_data_ref_dependence): ... here. Simplify.
24903 (vect_check_interleaving): Simplify.
24904 (vect_analyze_data_ref_dependences): Likewise. Split out ...
24905 (vect_slp_analyze_data_ref_dependences): ... this new function.
24906 (dr_group_sort_cmp): New function.
24907 (vect_analyze_data_ref_accesses): Compute data-reference groups
24908 here instead of in vect_analyze_data_ref_dependence. Use
24909 a more efficient algorithm.
24910 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
24911 vect_slp_analyze_data_ref_dependences. Call
24912 vect_analyze_data_ref_accesses earlier.
24913 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
24914 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
24915 (vect_slp_analyze_data_ref_dependences): New prototype.
24916
24917 2013-03-21 Richard Biener <rguenther@suse.de>
24918
24919 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
24920 ref is stored in the loop.
24921 (find_refs_for_sm): Walk only over all stores.
24922 (store_motion_loop): Allocate from lim_bitmap_obstack.
24923 (store_motion): Likewise.
24924
24925 2013-03-21 Richard Biener <rguenther@suse.de>
24926
24927 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
24928 Update virtual SSA form.
24929
24930 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24931
24932 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
24933 * configure: Regenerate.
24934 * config.in: Regenerate.
24935 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
24936 if !HAVE_LD_EH_FRAME_CIEV3.
24937
24938 2013-03-21 Richard Biener <rguenther@suse.de>
24939
24940 * tree-cfg.c (verify_expr_no_block): New function.
24941 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
24942 nor DECL_VALUE_EXPR have locations with associated blocks.
24943 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
24944 (clear_unused_block_pointer): Remove code dealing with
24945 blocks in DECL_DEBUG_EXPR locations.
24946
24947 2013-03-21 Richard Biener <rguenther@suse.de>
24948
24949 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
24950 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
24951 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
24952 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
24953 instead of DECL_DEBUG_EXPR_IS_FROM.
24954 * gimplify.c (gimplify_modify_expr): Likewise.
24955 * tree-cfg.c (verify_expr_location_1): Likewise.
24956 * tree-complex.c (create_one_component_var): Likewise.
24957 * tree-sra.c (create_access_replacement): Likewise.
24958 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
24959 (clear_unused_block_pointer): Likewise.
24960 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
24961 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
24962 * var-tracking.c (var_debug_decl): Likewise.
24963 (track_expr_p): Likewise.
24964 * tree-inline.c (add_local_variables): Likewise. Set
24965 DECL_HAS_DEBUG_EXPR_P after copying it.
24966 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
24967 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
24968
24969 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
24970
24971 PR bootstrap/56656
24972 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
24973 * configure: Regenerate.
24974 * config.in: Regenerate.
24975 * config/i386/i386.md (*movdf_internal): Use
24976 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
24977 movd instead of movq mnemonic for interunit moves.
24978 (*movdi_internal): Ditto.
24979
24980 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24981
24982 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
24983 (abd<mode>_3): New pattern.
24984 (aba<mode>_3): New pattern.
24985 (fabd<mode>_3): New pattern.
24986
24987 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
24988
24989 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
24990 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
24991 occurrence of REGISTER_PREFIX as its empty string.
24992
24993 2013-03-20 Jeff Law <law@redhat.com>
24994
24995 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
24996 addititional equivalences for equality comparisons between an SSA_NAME
24997 and a constant where the SSA_NAME was set from a widening conversion.
24998
24999 2013-03-20 Walter Lee <walt@tilera.com>
25000
25001 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
25002
25003 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25004
25005 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
25006 depending on type attribute.
25007 (*movti_internal): Ditto.
25008 (*movtf_internal): Ditto.
25009 (*movxf_internal): Ditto.
25010 (*movdf_internal): Ditto.
25011 (*movsf_internal): Ditto.
25012
25013 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25014
25015 * config/i386/i386.md (*movti_internal): Set prefix attribute to
25016 maybe_vex for sselog1 and ssemov types.
25017 (*movdi_internal): Reorder operand constraints.
25018 (*movsi_internal): Ditto. Set prefix attribute to
25019 maybe_vex for sselog1 and ssemov types.
25020 (*movtf_internal): Set prefix attribute to maybe_vex
25021 for sselog1 and ssemov types.
25022 (*movdf_internal): Ditto. Set prefix_data16 attribute for
25023 DImode ssemov types. Reorder operand constraints.
25024 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
25025 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
25026 attribute for SImode ssemov types. Reorder operand constraints.
25027
25028 2013-03-20 Martin Jambor <mjambor@suse.cz>
25029
25030 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
25031 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
25032
25033 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
25034
25035 * config/rs6000/predicates.md (indexed_address, update_address_mem
25036 update_indexed_address_mem): New predicates.
25037 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
25038 attribute for load/store instructions.
25039 * config/rs6000/dfp.md (movsd_store): Likewise.
25040 (movsd_load): Likewise.
25041 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
25042 (unnamed HI->DI extend define_insn): Likewise.
25043 (unnamed SI->DI extend define_insn): Likewise.
25044 (unnamed QI->SI extend define_insn): Likewise.
25045 (unnamed QI->HI extend define_insn): Likewise.
25046 (unnamed HI->SI extend define_insn): Likewise.
25047 (unnamed HI->SI extend define_insn): Likewise.
25048 (extendsfdf2_fpr): Likewise.
25049 (movsi_internal1): Likewise.
25050 (movsi_internal1_single): Likewise.
25051 (movhi_internal): Likewise.
25052 (movqi_internal): Likewise.
25053 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
25054 attribute for load/store instructions.
25055 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
25056 instructions.
25057 (mov<mode>_softfloat): Likewise.
25058 (mov<mode>_hardfloat32): Likewise.
25059 (mov<mode>_hardfloat64): Likewise.
25060 (mov<mode>_softfloat64): Likewise.
25061 (movdi_internal32): Likewise.
25062 (movdi_internal64): Likewise.
25063 (probe_stack_<mode>): Likewise.
25064
25065 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
25066
25067 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
25068 floating point, and decimal floating point to reload iterator.
25069
25070 * config/rs6000/constraints.md (wl constraint): New constraints to
25071 return FLOAT_REGS if certain options are used to reduce the number
25072 of separate patterns that exist in the file.
25073 (wx constraint): Likewise.
25074 (wz constraint): Likewise.
25075
25076 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25077 -mdebug=reg, print wg, wl, wx, and wz constraints.
25078 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25079 Initialize the reload functions for 64-bit binary/decimal floating
25080 point types.
25081 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25082 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25083 create the buffer on the stack to overcome not having a 32-bit
25084 load and store.
25085 (rs6000_emit_move): Likewise.
25086 (rs6000_secondary_memory_needed_rtx): Likewise.
25087 (rs6000_alloc_sdmode_stack_slot): Likewise.
25088 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25089 via xxlxor, just like DFmode 0.0.
25090
25091 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
25092 define as 1 if we are running on a power7 or newer.
25093 (enum r6000_reg_class_enum): Add new constraints.
25094
25095 * config/rs6000/dfp.md (movsd): Delete, combine with binary
25096 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25097 with other moves by using conditional constraits (wg). Use LFIWZX
25098 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
25099 (movsd splitter): Likewise.
25100 (movsd_hardfloat): Likewise.
25101 (movsd_softfloat): Likewise.
25102
25103 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
25104 binary and decimal floating point moves.
25105 (fmove_ok): New attributes to combine binary and decimal floating
25106 point moves, and to combine power6x (mfpgpr) moves along normal
25107 floating moves.
25108 (real_value_to_target): Likewise.
25109 (f32_lr): Likewise.
25110 (f32_lm): Likewise.
25111 (f32_li): Likewise.
25112 (f32_sr): Likewise.
25113 (f32_sm): Likewise.
25114 (f32_si): Likewise.
25115 (movsf): Combine binary and decimal floating point moves. Combine
25116 power6x (mfpgpr) moves with other moves by using conditional
25117 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
25118 (mov<mode> for SFmode/SDmode); Likewise.
25119 (SFmode/SDmode splitters): Likewise.
25120 (movsf_hardfloat): Likewise.
25121 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
25122 (movsf_softfloat): Likewise.
25123 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
25124
25125 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
25126 wx and wz constraints.
25127
25128 * config/rs6000/constraints.md (wg constraint): New constraint to
25129 return FLOAT_REGS if -mmfpgpr (power6x) was used.
25130
25131 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
25132 constraint.
25133
25134 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
25135 -mdebug=reg, print wg, wl, wx, and wz constraints.
25136 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
25137 Initialize the reload functions for 64-bit binary/decimal floating
25138 point types.
25139 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
25140 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
25141 create the buffer on the stack to overcome not having a 32-bit
25142 load and store.
25143 (rs6000_emit_move): Likewise.
25144 (rs6000_secondary_memory_needed_rtx): Likewise.
25145 (rs6000_alloc_sdmode_stack_slot): Likewise.
25146 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
25147 via xxlxor, just like DFmode 0.0.
25148
25149 * config/rs6000/dfp.md (movdd): Delete, combine with binary
25150 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
25151 with other moves by using conditional constraits (wg). Use LFIWZX
25152 and STFIWX for loading SDmode on power7.
25153 (movdd splitters): Likewise.
25154 (movdd_hardfloat32): Likewise.
25155 (movdd_softfloat32): Likewise.
25156 (movdd_hardfloat64_mfpgpr): Likewise.
25157 (movdd_hardfloat64): Likewise.
25158 (movdd_softfloat64): Likewise.
25159
25160 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
25161 64-bit binary and decimal floating point moves.
25162 (FMOVE64X): Likewise.
25163 (movdf): Combine 64-bit binary and decimal floating point moves.
25164 Combine power6x (mfpgpr) moves with other moves by using
25165 conditional constraits (wg).
25166 (mov<mode> for DFmode/DDmode): Likewise.
25167 (DFmode/DDmode splitters): Likewise.
25168 (movdf_hardfloat32): Likewise.
25169 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
25170 (movdf_softfloat32): Likewise.
25171 (movdf_hardfloat64_mfpgpr): Likewise.
25172 (movdf_hardfloat64): Likewise.
25173 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
25174 (movdf_softfloat64): Likewise.
25175 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
25176 (reload_<mode>_load): Move to later in the file so they aren't in
25177 the middle of the floating point move insns.
25178 (reload_<mode>_store): Likewise.
25179
25180 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
25181 constraint.
25182
25183 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
25184 constraint if -mdebug=reg.
25185 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
25186 Enable using dd reload support if needed.
25187
25188 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
25189 binary and decimal floating point moves in rs6000.md.
25190 (movtd_internal): Likewise.
25191
25192 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
25193 decimal floating point moves.
25194 (movtf): Likewise.
25195 (movtf_internal): Likewise.
25196 (mov<mode>_internal, TDmode/TFmode): Likewise.
25197 (movtf_softfloat): Likewise.
25198 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
25199
25200 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
25201 movdi_internal64, using wg constraint for move direct operations.
25202 (movdi_internal64): Likewise.
25203
25204 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
25205 MODES_TIEABLE_P for selected modes. Print the numerical value of
25206 the various virtual registers. Use GPR/FPR first/last values,
25207 instead of hard coding the register numbers. Print which modes
25208 have reload functions registered.
25209 (rs6000_option_override_internal): If -mdebug=reg, trace the options
25210 settings before/after setting cpu, target and subtarget settings.
25211 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
25212 and for secondary reload failures in rs6000_secondary_reload_inner.
25213 (rs6000_secondary_reload_fail): Likewise.
25214 (rs6000_secondary_reload_inner): Likewise.
25215
25216 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
25217 macros for first/last GPR and FPR registers.
25218 (LAST_GPR_REGNO): Likewise.
25219 (FIRST_FPR_REGNO): Likewise.
25220 (LAST_FPR_REGNO): Likewise.
25221
25222 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
25223 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
25224 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
25225 (vcond<mode><mode>): Likewise.
25226 (vcondu<mode><mode>): Likewise.
25227 (vector_gtu<mode>): Likewise.
25228 (vector_gte<mode>): Likewise.
25229 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
25230 to prevent the compiler from converting DImode operations to TImode.
25231 (ior<mode>3): Likewise.
25232 (and<mode>3): Likewise.
25233 (one_cmpl<mode>2): Likewise.
25234 (nor<mode>3): Likewise.
25235 (andc<mode>3): Likewise.
25236
25237 * config/rs6000/constraints.md (wt constraint): New constraint
25238 that returns VSX_REGS if TImode is allowed in VSX registers.
25239
25240 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
25241 constant under VSX.
25242
25243 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
25244 similar to TImode, but it is restricted to being in the GPRs.
25245
25246 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
25247 TImode to occupy a single VSX register.
25248
25249 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
25250 -mvsx-timode for power7/power8.
25251 (power7 cpu): Likewise.
25252 (power8 cpu): Likewise.
25253
25254 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
25255 sure that TFmode/TDmode take up two registers if they are ever
25256 allowed in the upper VSX registers.
25257 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
25258 registers.
25259 (rs6000_init_hard_regno_mode_ok): Likewise.
25260 (rs6000_debug_reg_global): Add debugging for PTImode and wt
25261 constraint. Print if LRA is turned on.
25262 (rs6000_option_override_internal): Give an error if -mvsx-timode
25263 and VSX is not enabled.
25264 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
25265 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
25266 to reg+offset addressing. Use PTImode when checking offset
25267 addresses for validity.
25268 (reg_offset_addressing_ok_p): Likewise.
25269 (rs6000_legitimate_offset_address_p): Likewise.
25270 (rs6000_legitimize_address): Likewise.
25271 (rs6000_legitimize_reload_address): Likewise.
25272 (rs6000_legitimate_address_p): Likewise.
25273 (rs6000_eliminate_indexed_memrefs): Likewise.
25274 (rs6000_emit_move): Likewise.
25275 (rs6000_secondary_reload): Likewise.
25276 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
25277 reloads to fpr registers to continue to use reg+offset addressing,
25278 but 64-bit reloads to altivec registers need reg+reg addressing.
25279 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
25280 it. Treat LO_SUM like a PLUS operation.
25281 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
25282 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
25283 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
25284 registers to share a register with a smaller sized type, since VSX
25285 puts scalars in the upper 64-bits.
25286 (print_operand): Add support for PTImode.
25287 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
25288 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
25289 registers, but don't have arithmetic support.
25290 (rs6000_memory_move_cost): Add test for VSX.
25291 (rs6000_opt_masks): Add -mvsx-timode.
25292
25293 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
25294 for TImode.
25295 (VSs): Likewise.
25296 (VSr): Use wt constraint for TImode.
25297 (VSv): Drop TImode support.
25298 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
25299 (vsx_movti_64bit): Likewise.
25300 (vsx_movti_32bit): Likewise.
25301 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
25302 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
25303 one '?' on the appropriate output constraint. Do not allow TImode
25304 logical operations on 32-bit systems.
25305 (vsx_ior<mode>3): Likewise.
25306 (vsx_xor<mode>3): Likewise.
25307 (vsx_one_cmpl<mode>2): Likewise.
25308 (vsx_nor<mode>3): Likewise.
25309 (vsx_andc<mode>3): Likewise.
25310 (vsx_concat_<mode>): Likewise.
25311 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
25312
25313 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
25314 OPTION_MASK_VSX_TIMODE.
25315 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
25316 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
25317
25318 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
25319 (TI2 iterator): New iterator for TImode, PTImode.
25320 (wd mode attribute): Add values for vector types.
25321 (movti_string): Replace TI move operations with operations for TImode
25322 and PTImode. Add support for TImode being allowed in VSX registers.
25323 (mov<mode>_string, TImode/PTImode): Likewise.
25324 (movti_ppc64): Likewise.
25325 (mov<mode>_ppc64, TImode/PTImode): Likewise.
25326 (TI mode splitters): Likewise.
25327
25328 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
25329 constraint.
25330
25331 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
25332
25333 PR tree-optimization/56355
25334 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
25335 Also handle integers with undefined overflow.
25336
25337 2013-03-20 Catherine Moore <clm@codesourcery.com>
25338 Maciej W. Rozycki <macro@codesourcery.com>
25339 Tom de Vries <tom@codesourcery.com>
25340 Nathan Sidwell <nathan@codesourcery.com>
25341 Iain Sandoe <iain@codesourcery.com>
25342 Nathan Froyd <froydnj@codesourcery.com>
25343 Chao-ying Fu <fu@mips.com>
25344
25345 * doc/extend.texi: (micromips, nomicromips, nocompression):
25346 Document new function attributes.
25347 * doc/invoke.texi (minterlink-compressed, mmicromips,
25348 m14k, m14ke, m14kec): Document new options.
25349 (minterlink-mips16): Update documentation.
25350 * doc/md.texi (ZC, ZD): Document new constraints.
25351 * configure.ac (gcc_cv_as_micromips): Check if linker
25352 supports the .set micromips directive.
25353 * configure: Regenerate.
25354 * config.in: Regenerate.
25355 * config/mips/mips-tables.opt: Regenerate.
25356 * config/mips/micromips.md: New file.
25357 * constraints.md (ZC, ZD): New constraints.
25358 * config/mips/predicates.md (movep_src_register): New predicate.
25359 (movep_src_operand): New predicate.
25360 (non_volatile_mem_operand): New predicate.
25361 * config/mips/mips.md (multimem): New type.
25362 (length): Differentiate between 17-bit and 18-bit branch offsets.
25363 (MOVEP1, MOVEP2): New mode iterator.
25364 (mov_<load>l): Use ZC constraint.
25365 (mov_<load>r): Likewise.
25366 (mov_<store>l): Likewise.
25367 (mov_<store>r): Likewise.
25368 (*branch_equality<mode>_inverted): Add microMIPS support.
25369 (*branch_equality<mode>): Likewise.
25370 (*jump_absolute): Likewise.
25371 (indirect_jump_<mode>): Likewise.
25372 (tablejump_<mode>): Likewise.
25373 (<optab>_internal): Likewise.
25374 (sibcall_internal): Likewise.
25375 (sibcall_value_internal): Likewise.
25376 (prefetch): Use constraint ZD.
25377 * config/mips/mips.opt (minterlink-compressed): New option.
25378 (minterlink-mips16): Now an alias for minterlink-compressed.
25379 (mmicromips): New option.
25380 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
25381 (compare_and_swap_12): Likewise.
25382 (sync_add<mode>): Likewise.
25383 (sync_<optab>_12): Likewise.
25384 (sync_old_<optab>_12): Likewise.
25385 (sync_new_<optab>_12): Likewise.
25386 (sync_nand_12): Likewise.
25387 (sync_old_nand_12): Likewise.
25388 (sync_new_nand_12): Likewise.
25389 (sync_sub<mode>): Likewise.
25390 (sync_old_add<mode>): Likewise.
25391 (sync_old_sub<mode>): Likewise.
25392 (sync_new_add<mode>): Likewise.
25393 (sync_new_sub<mode>): Likewise.
25394 (sync_<optab><mode>): Likewise.
25395 (sync_old_<optab><mode>): Likewise.
25396 (sync_new_<optab><mode>): Likewise.
25397 (sync_nand<mode>): Likewise.
25398 (sync_old_nand<mode>): Likewise.
25399 (sync_new_nand<mode>): Likewise.
25400 (sync_lock_test_and_set<mode>): Likewise.
25401 (test_and_set_12): Likewise.
25402 (atomic_compare_and_swap<mode>): Likewise.
25403 (atomic_exchange<mode>_llsc): Likewise.
25404 (atomic_fetch_add<mode>_llsc): Likewise.
25405 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
25406 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
25407 (umips_save_restore_pattern_p): Likewise.
25408 (umips_load_store_pair_p): Likewise.
25409 (umips_output_load_store_pair): Likewise.
25410 (umips_movep_target_p): Likewise.
25411 (umips_12bit_offset_address_p): Likewise.
25412 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
25413 (mips_base_mips16): Rename this...
25414 (mips_base_compression_flags): ...to this. Update all uses.
25415 (mips_attribute_table): Add micromips, nomicromips and nocompression.
25416 (mips_mips16_decl_p): Delete.
25417 (mips_nomips16_decl_p): Delete.
25418 (mips_get_compress_on_flags): New function.
25419 (mips_get_compress_off_flags): New function.
25420 (mips_get_compress_mode): New function.
25421 (mips_get_compress_on_name): New function.
25422 (mips_get_compress_off_name): New function.
25423 (mips_insert_attributes): Support multiple compression types.
25424 (mips_merge_decl_attributes): Likewise.
25425 (umips_12bit_offset_address_p): New function.
25426 (mips_start_function_definition): Emit .set micromips directive.
25427 (mips_call_may_need_jalx_p): New function.
25428 (mips_function_ok_for_sibcall): Add microMIPS support.
25429 (mips_print_operand_punctuation): Support short delay slots and
25430 compact jumps.
25431 (umips_swm_mask, umips_swm_encoding): New.
25432 (umips_build_save_restore): New function.
25433 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
25434 (was_mips16_p): Remove.
25435 (old_compression_mode): New.
25436 (mips_set_compression_mode): New function.
25437 (mips_set_current_function): Add microMIPS support.
25438 (mips_option_override): Likewise.
25439 (umips_save_restore_pattern_p): New function.
25440 (umips_output_save_restore): New function.
25441 (umips_load_store_pair_p_1): New function.
25442 (umips_load_store_pair_p): New function.
25443 (umips_output_load_store_pair_1): New function.
25444 (umips_output_load_store_pair): New function.
25445 (umips_movep_target_p) New function.
25446 (mips_prepare_pch_save): Add microMIPS support.
25447 * config/mips/mips.h (TARGET_COMPRESSION): New.
25448 (TARGET_CPU_CPP_BUILTINS): Update macro
25449 to use new compression flags and to support microMIPS.
25450 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
25451 (MIPS_ARCH_FLOAT_SPEC): Likewise.
25452 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
25453 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
25454 (ASM_SPEC): Support mmicromips and mno-micromips.
25455 (M16STORE_REG_P): New macro.
25456 (MIPS_CALL): Support TARGET_MICROMIPS.
25457 (MICROMIPS_J): New macro.
25458 (mips_base_mips16): Rename this...
25459 (mips_base_compression_flags): ...to this.
25460 (UMIPS_12BIT_OFFSET_P): New macro.
25461 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
25462 (MULTILIB_DIRNAMES): Likewise.
25463 2013-03-20 Richard Biener <rguenther@suse.de>
25464
25465 PR tree-optimization/56661
25466 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
25467 the result does not have to be distinct.
25468
25469 2013-03-20 Richard Biener <rguenther@suse.de>
25470
25471 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
25472 remap_gimple_op_r.
25473
25474 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
25475 Steven Bosscher <steven@gcc.gnu.org>
25476
25477 PR rtl-optimization/56605
25478 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
25479
25480 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
25481
25482 PR bootstrap/56656
25483 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
25484 that require movd instead of movq.
25485
25486 2013-03-20 Richard Biener <rguenther@suse.de>
25487
25488 * tree-ssa-structalias.c (struct variable_info): Add pointer
25489 to the first field of an aggregate with sub-vars. Make
25490 this and the pointer to the next subfield its ID.
25491 (vi_next): New function.
25492 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
25493 storedanything_id, integer_id): Increment by one.
25494 (new_var_info, get_call_vi, lookup_call_clobber_vi,
25495 get_call_clobber_vi): Adjust.
25496 (solution_set_expand): Simplify and speedup.
25497 (solution_set_add): Inline into ...
25498 (set_union_with_increment): ... this. Adjust accordingly.
25499 (do_sd_constraint): Likewise.
25500 (do_ds_constraint): Likewise.
25501 (do_complex_constraint): Simplify.
25502 (build_pred_graph): Adjust.
25503 (solve_graph): Likewise. Simplify and speedup.
25504 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
25505 get_constraint_for_component_ref, get_constraint_for_1,
25506 first_vi_for_offset, first_or_preceding_vi_for_offset,
25507 create_function_info_for, create_variable_info_for_1,
25508 create_variable_info_for, intra_create_variable_infos): Adjust.
25509 (init_base_vars): Push NULL for ID zero.
25510 (compute_points_to_sets): Adjust.
25511
25512 2013-03-20 Richard Biener <rguenther@suse.de>
25513
25514 * cfgloop.c (verify_loop_structure): Streamline and avoid
25515 ICEing on corrupt loop tree.
25516 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
25517 loop tree.
25518
25519 2013-03-20 Richard Biener <rguenther@suse.de>
25520
25521 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
25522 check whether an SSA update is needed.
25523
25524 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
25525
25526 * config/mips/constraints.md (T): Rename to...
25527 (Yf): ...this.
25528 (U): Rename to...
25529 (Yd): ...this.
25530 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
25531 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
25532
25533 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25534
25535 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
25536 (*subsi3_carryin_uxtw): Likewise.
25537
25538 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25539
25540 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
25541 (*rorsi3_insn_uxtw): Likewise.
25542
25543 2013-03-19 Ian Bolton <ian.bolton@arm.com>
25544
25545 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
25546 (*extrsi5_insn_uxtw): Likewise.
25547
25548 2013-03-19 Richard Biener <rguenther@suse.de>
25549
25550 PR tree-optimization/56273
25551 * passes.c (init_optimization_passes): Move second VRP after DOM.
25552
25553 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
25554
25555 * config/i386/i386.md (*movti_internal): Merge from
25556 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
25557 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
25558 nox64 isa attributes.
25559
25560 2013-03-18 Richard Biener <rguenther@suse.de>
25561
25562 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
25563 (unite): Likewise.
25564 (merge_node_constraints): Likewise.
25565 (build_succ_graph): Likewise.
25566 (valid_graph_edge): Inline into single caller.
25567 (unify_nodes): Likewise. Use bitmap_set_bit return value
25568 and cache varinfo.
25569 (scc_visit): Fix formatting and variable use.
25570 (do_sd_constraint): Use gcc_checking_assert.
25571 (do_ds_constraint): Likewise.
25572 (do_complex_constraint): Likewise.
25573 (condense_visit): Likewise. Cleanup.
25574 (dump_pred_graph): New function.
25575 (perform_var_substitution): Dump the pred-graph before
25576 variable substitution.
25577 (find_equivalent_node): Use gcc_checking_assert.
25578 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
25579
25580 2013-03-18 Richard Biener <rguenther@suse.de>
25581
25582 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
25583 Remove cond_expr_stmt_list argument and do not gimplify the
25584 built expression.
25585 (vect_loop_versioning): Adjust.
25586 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
25587 Cleanup to use less temporaries.
25588 (vect_create_data_ref_ptr): Cleanup.
25589
25590 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25591
25592 PR tree-optimization/56635
25593 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
25594 require types_compatible_p types.
25595
25596 2013-03-18 Nick Clifton <nickc@redhat.com>
25597
25598 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
25599 spurious backslash.
25600
25601 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
25602 Add missing line to comment describing function.
25603
25604 2013-03-18 Richard Biener <rguenther@suse.de>
25605
25606 PR tree-optimization/56210
25607 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
25608 Handle string / character search functions.
25609 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
25610
25611 2013-03-18 Richard Biener <rguenther@suse.de>
25612
25613 PR middle-end/56483
25614 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
25615 and implement properly.
25616 * gimple.h (gimple_cond_single_var_p): Remove.
25617
25618 2013-03-18 Richard Biener <rguenther@suse.de>
25619
25620 * tree-data-ref.h (find_data_references_in_loop): Declare.
25621 * tree-data-ref.c (get_references_in_stmt): Use a stack
25622 vector pre-allocated in the callers.
25623 (find_data_references_in_stmt): Adjust.
25624 (graphite_find_data_references_in_stmt): Likewise.
25625 (create_rdg_vertices): Likewise.
25626 (find_data_references_in_loop): Export.
25627 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
25628 Compute dependences here...
25629 (vect_analyze_data_refs): ...not here. When we encounter
25630 a non-vectorizable data reference in basic-block vectorization
25631 truncate the data reference vector. Do not bother to
25632 fixup data-dependence information for gather loads.
25633 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
25634 of data references, as reported.
25635
25636 2013-03-18 Richard Biener <rguenther@suse.de>
25637
25638 PR tree-optimization/3713
25639 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
25640 has_constants and expr.
25641 (stmt_has_constants): Properly valueize SSA names when deciding
25642 whether the stmt has constants.
25643
25644 2013-03-18 Richard Biener <rguenther@suse.de>
25645
25646 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
25647 whole function when there is nothing to do.
25648 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
25649 * tree-vectorizer.c (vectorize_loops): Update virtual and
25650 loop-closed SSA once.
25651 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
25652
25653 2013-03-18 Richard Biener <rguenther@suse.de>
25654
25655 PR middle-end/56113
25656 * domwalk.c (bb_postorder): New global static.
25657 (cmp_bb_postorder): New function.
25658 (walk_dominator_tree): Replace scheme imposing an order for
25659 visiting dominator sons by one sorting them at the time they
25660 are pushed on the stack.
25661
25662 2013-03-18 Richard Biener <rguenther@suse.de>
25663
25664 PR tree-optimization/39326
25665 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
25666 (struct mem_ref): Replace mem member with ao_ref typed member.
25667 (MEM_ANALYZABLE): Adjust.
25668 (memref_eq): Likewise.
25669 (mem_ref_alloc): Likewise.
25670 (gather_mem_refs_stmt): Likewise.
25671 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
25672 (execute_sm_if_changed_flag_set): Adjust.
25673 (execute_sm): Likewise.
25674 (ref_always_accessed_p): Likewise.
25675 (refs_independent_p): Likewise.
25676 (can_sm_ref_p): Likewise.
25677
25678 2013-03-18 Jakub Jelinek <jakub@redhat.com>
25679
25680 PR c/56566
25681 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
25682 return 1 even for !unsignedp.
25683
25684 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25685
25686 * config/i386/i386.md (isa): Add x64 and nox64.
25687 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
25688 (*pushtf): Enable *roF alternative for x64 isa only.
25689 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
25690 mode attribute of integer alternatives to DImode for TARGET_64BIT.
25691 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
25692 (*movtf_internal): Merge from *movtf_internal_rex64 and
25693 *movtf_internal_sse. Use x64 and nox64 isa attributes.
25694 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
25695 nox64 isa attributes.
25696 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
25697 nox64 isa attributes.
25698 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
25699
25700 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
25701
25702 * config/alpha/alpha.c (TARGET_LRA_P): New define.
25703
25704 2013-03-17 Jakub Jelinek <jakub@redhat.com>
25705
25706 PR target/56640
25707 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
25708 class names. Remove trailing comma after "ALL_REGS".
25709
25710 2013-03-16 Jan Hubicka <jh@suse.cz>
25711
25712 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
25713 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
25714 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
25715 of cgraph_get_create_node.
25716 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
25717
25718 2013-03-16 Jason Merrill <jason@redhat.com>
25719
25720 PR debug/49090
25721 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
25722 with DW_AT_default_value.
25723
25724 2013-03-16 Jakub Jelinek <jakub@redhat.com>
25725
25726 * BASE-VER: Set to 4.9.0.
25727
25728 2013-03-14 Andi Kleen <ak@linux.intel.com>
25729
25730 PR target/56619
25731 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
25732 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
25733 Document _x* TSX intrinsics.
25734
25735 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
25736 David Holsgrove <david.holsgrove@xilinx.com>
25737
25738 * configure.ac: Add MicroBlaze TLS support detection.
25739 * configure: Regenerate.
25740 * config/microblaze/microblaze-protos.h
25741 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
25742 symbol_mentioned_p, label_mentioned_p): Add prototypes.
25743 * config/microblaze/microblaze.c (microblaze_address_type): Add
25744 ADDRESS_TLS and tls_reloc address types.
25745 (microblaze_address_info): Add tls_reloc.
25746 (TARGET_HAVE_TLS): Define.
25747 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
25748 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
25749 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
25750 load_tls_operand, microblaze_call_tls_get_addr,
25751 microblaze_legitimize_tls_address): New functions.
25752 (microblaze_classify_unspec): Handle UNSPEC_TLS.
25753 (get_base_reg): Use microblaze_tls_symbol_p.
25754 (microblaze_classify_address): Handle TLS.
25755 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
25756 label_mentioned_p and microblaze_tls_referenced_p.
25757 (microblaze_legitimize_address): Handle TLS.
25758 (microblaze_address_insns): Handle ADDRESS_TLS.
25759 (pic_address_needs_scratch): Handle TLS.
25760 (print_operand_address): Handle TLS.
25761 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
25762 (microblaze_expand_move): Handle TLS.
25763 (microblaze_legitimate_constant_p): Check
25764 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
25765 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
25766 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
25767 (PIC_OFFSET_TABLE_REGNUM): Set.
25768 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
25769 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
25770 (addsi3, movsi_internal2, movdf_internal): Update constraints
25771 * config/microblaze/predicates.md (arith_plus_operand): Define
25772 (move_operand): Redefine as move_src_operand,
25773 check microblaze_tls_referenced_p.
25774
25775 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25776
25777 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
25778 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
25779
25780 2013-03-14 Ian Bolton <ian.bolton@arm.com>
25781
25782 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
25783 CC mode for AND.
25784
25785 2013-03-14 Jakub Jelinek <jakub@redhat.com>
25786
25787 PR tree-optimization/53265
25788 * common.opt (Waggressive-loop-optimizations): New option.
25789 * tree-ssa-loop-niter.c: Include tree-pass.h.
25790 (do_warn_aggressive_loop_optimizations): New function.
25791 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
25792 if number_of_latch_executions returned constant.
25793 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
25794 early. If number_of_latch_executions returned constant, set
25795 nb_iterations_upper_bound back to it.
25796 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
25797 field.
25798 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
25799 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
25800
25801 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
25802 (MULTILIB_OSDIRNAMES): Set.
25803 * genmultilib: If defaultosdirname doesn't start with :: , set
25804 defaultosdirname2 instead, clear it and emit two . multilib_raw
25805 entries instead of just one.
25806
25807 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
25808
25809 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25810 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
25811 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
25812 (SUBTARGET_OVERRIDE_OPTIONS): New.
25813
25814 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
25815
25816 PR target/49880
25817 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
25818 (musermode): Convert to Var(TARGET_USERMODE).
25819 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
25820 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
25821 * config/sh/sh.c (sh_option_override): Use
25822 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
25823 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
25824 condition.
25825 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
25826 TARGET_SH4.
25827 (udivsi3_i4_single, divsi3_i4_single): Use
25828 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
25829
25830 2013-03-13 Dave Korn <dave.korn.cygwin@gmail.com>
25831
25832 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
25833 default setting.
25834
25835 2013-03-13 Richard Biener <rguenther@suse.de>
25836
25837 PR tree-optimization/56608
25838 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
25839 calls when vectorizing basic-blocks.
25840
25841 2013-03-13 Jakub Jelinek <jakub@redhat.com>
25842
25843 PR plugins/45078
25844 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
25845 tm_file.
25846
25847 2013-03-12 Jakub Jelinek <jakub@redhat.com>
25848
25849 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
25850
25851 2013-03-11 Jan Hubicka <jh@suse.cz>
25852
25853 PR lto/56557
25854 * lto-streamer-out.c (output_symbol_p): Skip references from
25855 constructors of external variables.
25856
25857 2013-03-11 Jan Hubicka <jh@suse.cz>
25858
25859 PR middle-end/56571
25860 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
25861 from pseudos.
25862 * emit-rtl.c (verify_rtx_sharing): Likewise.
25863 (copy_insn_1): Likewise.
25864 * rtl.c (copy_rtx): Likewise.
25865
25866 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
25867
25868 PR target/56591
25869 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
25870 output_operand_lossage message.
25871
25872 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
25873
25874 PR target/56470
25875 * arm.c (shift_op): Validate RTL pattern on the fly.
25876 (arm_print_operand, case 'S'): Don't use shift_operator to validate
25877 the RTL.
25878
25879 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25880
25881 PR target/56347
25882 * config/pa/pa.md (call_value): Check for calls to powf and direct to
25883 new call patterns that clobber %fr12.
25884 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
25885 split and postreload patterns.
25886 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
25887 registers %fr12 and %fr12R as call used.
25888
25889 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
25890
25891 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
25892 (canon_address, record_store, replace_read, check_mem_read_rtx,
25893 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
25894 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
25895 rest_of_handle_dse): Likewise.
25896
25897 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
25898
25899 PR middle-end/56524
25900 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
25901 Add base_optabs.
25902 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
25903 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
25904 (save_optabs_if_changed): Replace with...
25905 (init_tree_optimization_optabs): ...this.
25906 * optabs.c (save_optabs_if_changed): Rename to...
25907 (init_tree_optimization_optabs): ...this. Take the optimization node
25908 as argument. Do nothing if the base optabs are already correct.
25909 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
25910 to recompute optabs.
25911 * function.h (function): Remove optabs field.
25912 * function.c (invoke_set_current_function_hook): Call
25913 init_tree_optimization_optabs. Use the result to initialize
25914 this_fn_optabs.
25915
25916 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
25917
25918 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
25919 if GTMA_HAS_NO_INSTRUMENTATION.
25920 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
25921 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
25922 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
25923 * gimple-pretty-print.c (dump_gimple_transaction): Handle
25924 GTMA_HAS_NO_INSTRUMENTATION.
25925
25926 2013-03-08 Jakub Jelinek <jakub@redhat.com>
25927
25928 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
25929 libasan_preinit.o.
25930
25931 2013-03-08 Marek Polacek <polacek@redhat.com>
25932 Jakub Jelinek <jakub@redhat.com>
25933
25934 PR tree-optimization/56478
25935 * predict.c (is_comparison_with_loop_invariant_p): Change the
25936 type of loop_step to tree.
25937 (predict_loops): Adjust.
25938 (predict_iv_comparison): Perform the computations on double_ints.
25939
25940 2013-03-08 Richard Biener <rguenther@suse.de>
25941
25942 PR tree-optimization/56570
25943 * tree-cfg.c (verify_expr_location_1): Verify locations for
25944 DECL_DEBUG_EXPR.
25945 * tree-sra.c (create_access_replacement): Strip locations
25946 from DECL_DEBUG_EXPRs.
25947
25948 2013-03-08 Richard Biener <rguenther@suse.de>
25949
25950 * tree-inline.c (expand_call_inline): Do not associate
25951 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
25952 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
25953
25954 2013-03-08 Richard Biener <rguenther@suse.de>
25955
25956 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
25957 or block changes with -Og. Fix for location / block encoding
25958 changes and PHI arguments with locations.
25959
25960 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
25961
25962 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
25963 for all counters.
25964 (struct output_info): Likewise.
25965 (register_overhead): Remove bad gcc_assert.
25966 (bitmap_find_bit): If there is only a single bitmap element, do not
25967 count a miss as a search.
25968 (print_statistics): Update for counter type changes.
25969 (dump_bitmap_statistics): Likewise. Print headers such that they
25970 are properly lined up with the printed counters.
25971
25972 2013-03-07 Jakub Jelinek <jakub@redhat.com>
25973
25974 PR tree-optimization/56559
25975 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
25976 check that it has only a single use.
25977
25978 2013-03-07 Richard Biener <rguenther@suse.de>
25979
25980 * doc/invoke.texi (fwhole-program): Discourage use in combination
25981 with -flto.
25982
25983 2013-03-06 Jakub Jelinek <jakub@redhat.com>
25984
25985 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
25986
25987 PR tree-optimization/56539
25988 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
25989 instead of GSI_CONTINUE_LINKING as last argument to
25990 force_gimple_operand_gsi. Adjust function comment.
25991
25992 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
25993 aarch64-cores.def.
25994
25995 PR middle-end/56548
25996 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
25997 promoted mode, convert the result back to the original mode.
25998
25999 2013-03-06 Richard Biener <rguenther@suse.de>
26000
26001 PR middle-end/56294
26002 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
26003 (insert_updated_phi_nodes_compare_uids): New function.
26004 (update_ssa): Sort symbols_to_rename after UID before
26005 traversing it to insert PHI nodes.
26006
26007 2013-03-06 Richard Biener <rguenther@suse.de>
26008
26009 PR middle-end/50494
26010 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
26011 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
26012
26013 Revert
26014 2013-02-13 Richard Biener <rguenther@suse.de>
26015
26016 PR lto/50494
26017 * varasm.c (output_constant_def_1): Get the decl representing
26018 the constant as argument.
26019 (output_constant_def): Wrap output_constant_def_1.
26020 (make_decl_rtl): Use output_constant_def_1 with the decl
26021 representing the constant.
26022 (build_constant_desc): Optionally re-use a decl already
26023 representing the constant.
26024 (tree_output_constant_def): Adjust.
26025
26026 2013-03-06 Joey Ye <joey.ye@arm.com>
26027
26028 PR lto/50293
26029 * gcc.c (convert_white_space): New function.
26030 (main): Handles white space in function name.
26031
26032 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
26033
26034 PR target/56529
26035 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
26036 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
26037 to SH_DIV_CALL_TABLE for TARGET_SH2.
26038 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
26039 list.
26040 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
26041 call-table options.
26042
26043 2013-03-05 Sterling Augustine <saugustine@google.com>
26044 Cary Coutant <ccoutant@google.com>
26045
26046 PR debug/55364
26047 * dwarf2out.c (resolve_addr): Don't call
26048 remove_loc_list_addr_table_entries a second time for the same
26049 expression.
26050
26051 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26052
26053 PR debug/56510
26054 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
26055 (avoid_complex_debug_insns): New function.
26056 (expand_debug_locations): Call it.
26057
26058 PR rtl-optimization/56484
26059 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
26060 lifetimes of hard registers on small register class machines.
26061
26062 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26063
26064 * config/microblaze/microblaze-protos.h: Rename
26065 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
26066 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
26067 fast_interrupt.
26068 (microblaze_fast_interrupt_function_p): New function.
26069 (microblaze_is_interrupt_handler): Rename to
26070 microblaze_is_interrupt_variant and add fast_interrupt check.
26071 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
26072 (save_restore_insns): Likewise.
26073 (compute_frame_size): Likewise.
26074 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
26075 (microblaze_globalize_label): Likewise.
26076 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
26077 * config/microblaze/microblaze.md: Use wrapper
26078 microblaze_is_interrupt_variant.
26079
26080 2013-03-05 Kai Tietz <ktietz@redhat.com>
26081
26082 * sdbout.c (sdbout_one_type): Switch to current function's section
26083 supporting cold/hot.
26084
26085 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
26086
26087 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
26088 -mxl-reorder.
26089
26090 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26091
26092 PR middle-end/56461
26093 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
26094 if VALGRIND_GET_VBITS is defined, temporarily make object
26095 memory all defined, and restore previous valgrind addressability
26096 and definability afterwards. Free this_object at the end.
26097
26098 PR middle-end/56461
26099 * lra.c (lra): Call lra_clear_live_ranges if live_p,
26100 right before calling lra_create_live_ranges, also call it
26101 when clearing live_p. Only call lra_clear_live_ranges
26102 at the end if live_p.
26103
26104 PR middle-end/56461
26105 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
26106
26107 2013-03-05 Richard Biener <rguenther@suse.de>
26108
26109 PR tree-optimization/56521
26110 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
26111 value-id.
26112
26113 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
26114
26115 PR c++/55135
26116 * except.h (remove_unreachable_eh_regions): New prototype.
26117 * except.c (remove_eh_handler_splicer): New function, split out
26118 of remove_eh_handler.
26119 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
26120 warning about running it on many EH regions one at a time.
26121 (remove_unreachable_eh_regions_worker): New function, walk the
26122 EH tree in depth-first order and remove non-marked regions.
26123 (remove_unreachable_eh_regions): New function.
26124 * tree-eh.c (mark_reachable_handlers): New function, split out
26125 from remove_unreachable_handlers.
26126 (remove_unreachable_handlers): Use mark_reachable_handlers and
26127 remove_unreachable_eh_regions.
26128 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
26129 and remove_unreachable_eh_regions.
26130
26131 2013-03-05 Richard Biener <rguenther@suse.de>
26132
26133 PR middle-end/56525
26134 * loop-init.c (fix_loop_structure): Remove loops in two stages,
26135 not freeing them until the end.
26136
26137 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
26138
26139 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
26140
26141 2013-03-05 Richard Biener <rguenther@suse.de>
26142
26143 PR tree-optimization/56270
26144 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
26145 of loads after scheduling an SLP instance.
26146
26147 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26148
26149 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
26150 tic6x.exp.
26151 (check_gcc_parallelize): Run guality.exp as a separate job from
26152 vect.exp with unsorted.exp and $(dg_target_exps) separately from
26153 struct-layout-1.exp with stackalign.exp.
26154
26155 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
26156
26157 PR middle-end/56461
26158 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
26159 load_index sbitmap even if some bit in it isn't set.
26160
26161 PR middle-end/56461
26162 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
26163 (discover_iteration_bound_by_body_walk): Change queues to
26164 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
26165 spelling in comment. Call safe_push on queues[bound_index] directly.
26166 Release queues[queue_index] in every iteration unconditionally.
26167 Release bounds vector.
26168
26169 PR middle-end/56461
26170 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
26171 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
26172 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
26173 inner_phis vector.
26174
26175 2013-03-05 Richard Biener <rguenther@suse.de>
26176
26177 PR lto/56515
26178 * tree-inline.c (remap_blocks_to_null): New function.
26179 (expand_call_inline): When expanding a call stmt without
26180 an associated block inline remap all callee blocks to NULL.
26181
26182 2013-03-05 Jakub Jelinek <jakub@redhat.com>
26183
26184 PR rtl-optimization/56494
26185 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
26186 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
26187 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
26188
26189 PR middle-end/56461
26190 * sel-sched-ir.c (free_sched_pools): Release
26191 succs_info_pool.stack[succs_info_pool.max_top] vectors too
26192 if succs_info_pool.max_top isn't -1.
26193
26194 PR bootstrap/56509
26195 * opts.c (opts_obstack, opts_concat): Moved to...
26196 * opts-common.c (opts_obstack, opts_concat): ... here.
26197
26198 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26199
26200 PR middle-end/56461
26201 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
26202
26203 2013-03-04 Martin Jambor <mjambor@suse.cz>
26204
26205 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
26206 all appropriate places.
26207
26208 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
26209
26210 PR tree-optimization/56424
26211 * ipa-split.c (split_function): Do not set the RSO flag if result is
26212 not by reference and its type is a register type.
26213
26214 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
26215
26216 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
26217 (microblaze_legitimate_pic_operand): Likewise
26218 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
26219 new function microblaze_legitimate_pic_operand
26220 * config/microblaze/microblaze-protos.h
26221 (microblaze_legitimate_pic_operand): Declare.
26222
26223 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26224
26225 * config/microblaze/predicates.md (call_insn_simple_operand):
26226 New predicate for supported rtx code types.
26227 * config/microblaze/microblaze.md (call_internal1): Use
26228 call_insn_simple_operand predicate.
26229
26230 2013-03-04 Jakub Jelinek <jakub@redhat.com>
26231
26232 PR middle-end/56461
26233 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
26234 partitions.ordered_remove.
26235
26236 PR middle-end/56461
26237 * tree-vect-stmts.c (vectorizable_conversion): Don't call
26238 vec_oprnds0.create (1) for modifier == NONE.
26239
26240 PR middle-end/56461
26241 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
26242 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
26243 vec_oprnds1 right before pushing anything to it for
26244 scalar_shift_arg.
26245
26246 PR middle-end/56461
26247 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
26248 set nbbs to 0 instead of having separate code path.
26249 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
26250 instead of false as last argument if returning NULL.
26251
26252 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
26253
26254 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
26255 the attribute is now called "target" instead of "option".
26256 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
26257 * doc/tm.texi.in (Target Attributes): Likewise document the correct
26258 attribute/pragma name for TARGET_OPTION_VALID_P and
26259 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
26260 * doc/tm.texi: Regenerated.
26261
26262 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
26263
26264 * config/microblaze/microblaze.c:
26265 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
26266 * config/microblaze/microblaze.h: Add -mxl-reorder to
26267 DRIVER_SELF_SPECS.
26268 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
26269 instructions emitted if TARGET_REORDER.
26270 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
26271 or 0 for -m/-mno case, but initialises as 2 to detect default use case
26272 separately.
26273
26274 2013-03-01 Xinliang David Li <davidxl@google.com>
26275
26276 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
26277 walk length.
26278
26279 2013-03-01 Jakub Jelinek <jakub@redhat.com>
26280
26281 PR middle-end/56461
26282 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
26283 vector even when returning true. Fix up function comment formatting.
26284
26285 PR middle-end/56461
26286 * ira-build.c (ira_loop_nodes_count): New variable.
26287 (create_loop_tree_nodes): Initialize it.
26288 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
26289
26290 PR middle-end/56461
26291 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
26292 method on dr_chain and result_chain.
26293 * tree-vect-stmts.c (vectorizable_store): Only call
26294 result_chain.create if j == 0.
26295
26296 PR middle-end/56461
26297 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
26298 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
26299 before overwriting it.
26300
26301 2013-03-01 Tobias Burnus <burnus@net-b.de>
26302
26303 * doc/extended.texi (C Extensions): Change order in @menu
26304 to match @node.
26305 (Other MIPS Built-in Functions): Move last MIPS entry before
26306 "picoChip Built-in Functions".
26307 (SH Built-in Functions): Move after RX Built-in Functions.
26308 * doc/gcc.texi (Introduction): Change order in @menu
26309 to match @node.
26310 * doc/md.texi (Constraints): Ditto.
26311 * gty.texi (Type Information): Ditto.
26312 (User-provided marking routines for template types): Make
26313 subsection.
26314 * doc/invoke.texi (AArch64 Options): Move before
26315 "Adapteva Epiphany Options".
26316
26317 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
26318 Jakub Jelinek <jakub@redhat.com>
26319
26320 PR sanitizer/56454
26321 * asan.c (gate_asan): Lookup no_sanitize_address instead of
26322 no_address_safety_analysis attribute.
26323 * doc/extend.texi (no_address_safety_attribute): Rename to
26324 no_sanitize_address attribute, mention no_address_safety_analysis
26325 attribute as deprecated alias.
26326
26327 2013-02-28 Jakub Jelinek <jakub@redhat.com>
26328
26329 PR middle-end/56461
26330 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
26331 type to vec<vec<tree> > *.
26332 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
26333 to be vec<tree> instead of vec<tree> *, set vec_defs
26334 to vNULL and call vec_defs.create (number_of_vects), adjust other
26335 uses of vec_defs.
26336 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
26337 vectorizable_condition): Adjust vect_get_slp_defs callers.
26338
26339 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26340
26341 * config/aarch64/aarch64.c
26342 (aarch64_float_const_representable): Remove unused variable.
26343
26344 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26345
26346 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
26347
26348 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26349
26350 * config/aarch64/aarch64-builtins.c
26351 (aarch64_init_simd_builtins): Make static.
26352
26353 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
26354
26355 * config/aarch64/aarch64.c
26356 (aarch64_simd_make_constant): Make static.
26357
26358 2013-02-28 Martin Jambor <mjambor@suse.cz>
26359
26360 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
26361 with no initialization to the RHS of debug statements.
26362
26363 2013-02-28 Martin Jambor <mjambor@suse.cz>
26364
26365 PR tree-optimization/56294
26366 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
26367 Adjust dumping.
26368 (get_access_replacement): Do not call create_access_replacement.
26369 Assert a replacement exists.
26370 (get_repl_default_def_ssa_name): Create the replacement declaration
26371 itself.
26372
26373 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
26374
26375 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
26376 final_end_function.
26377
26378 2013-02-28 Marek Polacek <polacek@redhat.com>
26379
26380 PR rtl-optimization/56466
26381 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
26382 if we're changing a loop.
26383 (peel_loops_completely): Likewise.
26384
26385 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
26386
26387 PR c++/55813
26388 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
26389
26390 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26391
26392 PR target/56445
26393 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
26394 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
26395 INTX_FTYPE_FX, FX_FTYPE_INTX.
26396 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
26397
26398 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26399
26400 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
26401 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
26402 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
26403 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
26404 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
26405 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
26406 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
26407 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
26408 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
26409 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
26410 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
26411 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
26412 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
26413 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
26414 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
26415 (avrxmega6): Increase max flash segments from 5 to 6.
26416 * config/avr/t-multilib: Regenerate.
26417 * config/avr/avr-tables.opt: Regenerate.
26418 * doc/avr-mmcu.texi: Regenerate.
26419
26420 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
26421
26422 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
26423 (avr_device_to_arch): Rename to avr_device_to_ld.
26424 (avr_device_to_as): New prototype.
26425 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
26426 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
26427 * config/avr/driver-avr.c (avr_device_to_as): New.
26428 (avr_device_to_arch): Rename to avr_device_to_ld.
26429
26430 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26431
26432 PR middle-end/56461
26433 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
26434 method on dr_chain and result_chain.
26435
26436 PR middle-end/56461
26437 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
26438 pointer_set_destroy on not_executed_last_iteration.
26439
26440 PR middle-end/56461
26441 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
26442
26443 PR middle-end/56461
26444 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
26445 FOR_EACH_DEFINED_FUNCTION when freeing state.
26446
26447 PR middle-end/56461
26448 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
26449 pool_free.
26450 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
26451 overwriting it.
26452
26453 PR middle-end/56461
26454 * ipa-cp.c (decide_whether_version_node): Call vec_free on
26455 known_aggs[i].items and release known_aggs vector.
26456
26457 PR middle-end/56461
26458 * ipa-reference.c (propagate): Free node_info even for alias nodes.
26459
26460 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
26461
26462 * config/microblaze/microblaze.c (microblaze_emit_compare):
26463 Use xor for EQ/NE comparisions.
26464 * config/microblaze/microblaze.md (cstoresf4): Add constraints
26465 (cbranchsf4): Adjust operator to comparison_operator.
26466
26467 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26468
26469 PR middle-end/56461
26470 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
26471 vector.
26472 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
26473 vec_safe_push, always update *slot.
26474 (redirect_edge_var_map_clear): Use vec_free.
26475 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
26476 (free_var_map_entry): Use vec_free.
26477 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
26478 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
26479
26480 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
26481
26482 PR middle-end/45472
26483 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
26484 when the may_trap_p bit of the exprs being merged differs.
26485 Reorder tests for speculativeness in the logical and operator.
26486
26487 2013-02-27 Jakub Jelinek <jakub@redhat.com>
26488
26489 * incpath.c (add_standard_paths): Use reconcat instead of concat
26490 where appropriate and avoid leaking memory.
26491
26492 * opts.h: Include obstack.h.
26493 (opts_concat): New prototype.
26494 (opts_obstack): New declaration.
26495 * opts.c (opts_concat): New function.
26496 (opts_obstack): New variable.
26497 (init_options_struct): Call gcc_init_obstack on opts_obstack.
26498 (finish_options): Use opts_concat instead of concat
26499 and XOBNEWVEC instead of XNEWVEC.
26500 * opts-common.c (generate_canonical_option, decode_cmdline_option,
26501 generate_option): Likewise.
26502 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
26503 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
26504
26505 PR target/56455
26506 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
26507 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
26508
26509 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26510
26511 PR middle-end/56461
26512 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
26513
26514 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26515
26516 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
26517 (arm_block_move_unaligned_straight): Likewise.
26518 (arm_adjust_block_mem): Likewise.
26519
26520 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
26521
26522 PR target/48901
26523 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
26524 temp, cond and label.
26525 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
26526
26527 PR target/52500
26528 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
26529 * config/c6x/c6x.h (dbx_register_map): Update declaration.
26530
26531 PR target/52501
26532 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
26533 of prologue/epilogue functions.
26534
26535 PR target/52550
26536 * config/tilegx/tilegx.c (tilegx_expand_prologue):
26537 Remove unused variable cfa_offset.
26538 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
26539
26540 PR target/54639
26541 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
26542 type promotion to unsigned.
26543
26544 PR target/54640
26545 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
26546 for HOST_WIDE_INT of 32 bit / same size as int.
26547 (arm_block_move_unaligned_straight): Likewise.
26548 (arm_adjust_block_mem): Likewise.
26549
26550 PR target/54662
26551 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
26552 ALL_CFLAGS.
26553
26554 2013-02-26 Marek Polacek <polacek@redhat.com>
26555
26556 PR tree-optimization/56426
26557 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
26558
26559 2013-02-26 Richard Biener <rguenther@suse.de>
26560
26561 PR target/56444
26562 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
26563 unused variable loops.
26564
26565 2013-02-26 Jakub Jelinek <jakub@redhat.com>
26566
26567 PR tree-optimization/56448
26568 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
26569 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
26570 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
26571 later operands of the references, or even first operand for
26572 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
26573
26574 PR tree-optimization/56443
26575 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
26576 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
26577 to type_for_mode langhook.
26578
26579 2013-02-25 Matt Turner <mattst88@gmail.com>
26580
26581 * doc/invoke.texi: Document r4700.
26582
26583 2013-02-25 Richard Biener <rguenther@suse.de>
26584
26585 PR tree-optimization/56175
26586 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
26587 split out from ...
26588 (simplify_bitwise_binary): ... here. Also guard the conversion
26589 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
26590
26591 2013-02-25 Catherine Moore <clm@codesourcery.com>
26592
26593 Revert:
26594 2013-02-24 Catherine Moore <clm@codesourcery.com>
26595 Maciej W. Rozycki <macro@codesourcery.com>
26596 Tom de Vries <tom@codesourcery.com>
26597 Nathan Sidwell <nathan@codesourcery.com>
26598 Iain Sandoe <iain@codesourcery.com>
26599 Nathan Froyd <froydnj@codesourcery.com>
26600 Chao-ying Fu <fu@mips.com>
26601
26602 * doc/extend.texi: (micromips, nomicromips, nocompression):
26603 Document new function attributes.
26604 * doc/invoke.texi (minterlink-compressed, mmicromips,
26605 m14k, m14ke, m14kec): Document new options.
26606 (minterlink-mips16): Update documentation.
26607 * doc/md.texi (ZC, ZD): Document new constraints.
26608 * configure.ac (gcc_cv_as_micromips): Check if linker
26609 supports the .set micromips directive.
26610 * configure: Regenerate.
26611 * config.in: Regenerate.
26612 * config/mips/mips-tables.opt: Regenerate.
26613 * config/mips/micromips.md: New file.
26614 * constraints.md (ZC, AD): New constraints.
26615 * config/mips/predicates.md (movep_src_register): New predicate.
26616 (movep_src_operand): New predicate.
26617 (non_volatile_mem_operand): New predicate.
26618 * config/mips/mips.md (multimem): New type.
26619 (length): Differentiate between 17-bit and 18-bit branch offsets.
26620 (MOVEP1, MOVEP2): New mode iterator.
26621 (mov_<load>l): Use ZC constraint.
26622 (mov_<load>r): Likewise.
26623 (mov_<store>l): Likewise.
26624 (mov_<store>r): Likewise.
26625 (*branch_equality<mode>_inverted): Add microMIPS support.
26626 (*branch_equality<mode>): Likewise.
26627 (*jump_absolute): Likewise.
26628 (indirect_jump_<mode>): Likewise.
26629 (tablejump_<mode>): Likewise.
26630 (<optab>_internal): Likewise.
26631 (sibcall_internal): Likewise.
26632 (sibcall_value_internal): Likewise.
26633 (prefetch): Use constraint ZD.
26634 * config/mips/mips.opt (minterlink-compressed): New option.
26635 (minterlink-mips16): Now an alias for minterlink-compressed.
26636 (mmicromips): New option.
26637 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26638 (compare_and_swap_12): Likewise.
26639 (sync_add<mode>): Likewise.
26640 (sync_<optab>_12): Likewise.
26641 (sync_old_<optab>_12): Likewise.
26642 (sync_new_<optab>_12): Likewise.
26643 (sync_nand_12): Likewise.
26644 (sync_old_nand_12): Likewise.
26645 (sync_new_nand_12): Likewise.
26646 (sync_sub<mode>): Likewise.
26647 (sync_old_add<mode>): Likewise.
26648 (sync_old_sub<mode>): Likewise.
26649 (sync_new_add<mode>): Likewise.
26650 (sync_new_sub<mode>): Likewise.
26651 (sync_<optab><mode>): Likewise.
26652 (sync_old_<optab><mode>): Likewise.
26653 (sync_new_<optab><mode>): Likewise.
26654 (sync_nand<mode>): Likewise.
26655 (sync_old_nand<mode>): Likewise.
26656 (sync_new_nand<mode>): Likewise.
26657 (sync_lock_test_and_set<mode>): Likewise.
26658 (test_and_set_12): Likewise.
26659 (atomic_compare_and_swap<mode>): Likewise.
26660 (atomic_exchange<mode>_llsc): Likewise.
26661 (atomic_fetch_add<mode>_llsc): Likewise.
26662 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26663 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26664 (umips_save_restore_pattern_p): Likewise.
26665 (umips_load_store_pair_p): Likewise.
26666 (umips_output_load_store_pair): Likewise.
26667 (umips_movep_target_p): Likewise.
26668 (umips_12bit_offset_address_p): Likewise.
26669 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26670 (mips_base_mips16): Rename this...
26671 (mips_base_compression_flags): ...to this. Update all uses.
26672 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26673 (mips_mips16_decl_p): Delete.
26674 (mips_nomips16_decl_p): Delete.
26675 (mips_get_compress_on_flags): New function.
26676 (mips_get_compress_off_flags): New function.
26677 (mips_get_compress_mode): New function.
26678 (mips_get_compress_on_name): New function.
26679 (mips_get_compress_off_name): New function.
26680 (mips_insert_attributes): Support multiple compression types.
26681 (mips_merge_decl_attributes): Likewise.
26682 (umips_12bit_offset_address_p): New function.
26683 (mips_start_function_definition): Emit .set micromips directive.
26684 (mips_call_may_need_jalx_p): New function.
26685 (mips_function_ok_for_sibcall): Add microMIPS support.
26686 (mips_print_operand_punctuation): Support short delay slots and
26687 compact jumps.
26688 (umips_swm_mask, umips_swm_encoding): New.
26689 (umips_build_save_restore): New function.
26690 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26691 (was_mips16_p): Remove.
26692 (old_compression_mode): New.
26693 (mips_set_compression_mode): New function.
26694 (mips_set_current_function): Add microMIPS support.
26695 (mips_option_override): Likewise.
26696 (umips_save_restore_pattern_p): New function.
26697 (umips_output_save_restore): New function.
26698 (umips_load_store_pair_p_1): New function.
26699 (umips_load_store_pair_p): New function.
26700 (umips_output_load_store_pair_1): New function.
26701 (umips_output_load_store_pair): New function.
26702 (umips_movep_target_p) New function.
26703 (mips_prepare_pch_save): Add microMIPS support.
26704 * config/mips/mips.h (TARGET_COMPRESSION): New.
26705 (TARGET_CPU_CPP_BUILTINS): Update macro
26706 to use new compression flags and to support microMIPS.
26707 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26708 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26709 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26710 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26711 (ASM_SPEC): Support mmicromips and mno-micromips.
26712 (M16STORE_REG_P): New macro.
26713 (MIPS_CALL): Support TARGET_MICROMIPS.
26714 (MICROMIPS_J): New macro.
26715 (mips_base_mips16): Rename this...
26716 (mips_base_compression_flags): ...to this.
26717 (UMIPS_12BIT_OFFSET_P): New macro.
26718 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26719 (MULTILIB_DIRNAMES): Likewise.
26720
26721 2013-02-25 Tom de Vries <tom@codesourcery.com>
26722
26723 PR rtl-optimization/56131
26724 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
26725 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
26726 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
26727
26728 2013-02-25 Tobias Burnus <burnus@net-b.de>
26729
26730 * doc/invoke.texi (-fsanitize=): Move from optimization
26731 to debugging options.
26732
26733 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26734
26735 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
26736
26737 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
26738 Alexander Monakov <amonakov@ispras.ru>
26739
26740 PR middle-end/56077
26741 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
26742 flush pending lists also on non-jumps. Adjust comment.
26743
26744 2013-02-24 Catherine Moore <clm@codesourcery.com>
26745 Maciej W. Rozycki <macro@codesourcery.com>
26746 Tom de Vries <tom@codesourcery.com>
26747 Nathan Sidwell <nathan@codesourcery.com>
26748 Iain Sandoe <iain@codesourcery.com>
26749 Nathan Froyd <froydnj@codesourcery.com>
26750 Chao-ying Fu <fu@mips.com>
26751
26752 * doc/extend.texi: (micromips, nomicromips, nocompression):
26753 Document new function attributes.
26754 * doc/invoke.texi (minterlink-compressed, mmicromips,
26755 m14k, m14ke, m14kec): Document new options.
26756 (minterlink-mips16): Update documentation.
26757 * doc/md.texi (ZC, ZD): Document new constraints.
26758 * configure.ac (gcc_cv_as_micromips): Check if linker
26759 supports the .set micromips directive.
26760 * configure: Regenerate.
26761 * config.in: Regenerate.
26762 * config/mips/mips-tables.opt: Regenerate.
26763 * config/mips/micromips.md: New file.
26764 * constraints.md (ZC, AD): New constraints.
26765 * config/mips/predicates.md (movep_src_register): New predicate.
26766 (movep_src_operand): New predicate.
26767 (non_volatile_mem_operand): New predicate.
26768 * config/mips/mips.md (multimem): New type.
26769 (length): Differentiate between 17-bit and 18-bit branch offsets.
26770 (MOVEP1, MOVEP2): New mode iterator.
26771 (mov_<load>l): Use ZC constraint.
26772 (mov_<load>r): Likewise.
26773 (mov_<store>l): Likewise.
26774 (mov_<store>r): Likewise.
26775 (*branch_equality<mode>_inverted): Add microMIPS support.
26776 (*branch_equality<mode>): Likewise.
26777 (*jump_absolute): Likewise.
26778 (indirect_jump_<mode>): Likewise.
26779 (tablejump_<mode>): Likewise.
26780 (<optab>_internal): Likewise.
26781 (sibcall_internal): Likewise.
26782 (sibcall_value_internal): Likewise.
26783 (prefetch): Use constraint ZD.
26784 * config/mips/mips.opt (minterlink-compressed): New option.
26785 (minterlink-mips16): Now an alias for minterlink-compressed.
26786 (mmicromips): New option.
26787 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
26788 (compare_and_swap_12): Likewise.
26789 (sync_add<mode>): Likewise.
26790 (sync_<optab>_12): Likewise.
26791 (sync_old_<optab>_12): Likewise.
26792 (sync_new_<optab>_12): Likewise.
26793 (sync_nand_12): Likewise.
26794 (sync_old_nand_12): Likewise.
26795 (sync_new_nand_12): Likewise.
26796 (sync_sub<mode>): Likewise.
26797 (sync_old_add<mode>): Likewise.
26798 (sync_old_sub<mode>): Likewise.
26799 (sync_new_add<mode>): Likewise.
26800 (sync_new_sub<mode>): Likewise.
26801 (sync_<optab><mode>): Likewise.
26802 (sync_old_<optab><mode>): Likewise.
26803 (sync_new_<optab><mode>): Likewise.
26804 (sync_nand<mode>): Likewise.
26805 (sync_old_nand<mode>): Likewise.
26806 (sync_new_nand<mode>): Likewise.
26807 (sync_lock_test_and_set<mode>): Likewise.
26808 (test_and_set_12): Likewise.
26809 (atomic_compare_and_swap<mode>): Likewise.
26810 (atomic_exchange<mode>_llsc): Likewise.
26811 (atomic_fetch_add<mode>_llsc): Likewise.
26812 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
26813 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
26814 (umips_save_restore_pattern_p): Likewise.
26815 (umips_load_store_pair_p): Likewise.
26816 (umips_output_load_store_pair): Likewise.
26817 (umips_movep_target_p): Likewise.
26818 (umips_12bit_offset_address_p): Likewise.
26819 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
26820 (mips_base_mips16): Rename this...
26821 (mips_base_compression_flags): ...to this. Update all uses.
26822 (mips_attribute_table): Add micromips, nomicromips and nocompression.
26823 (mips_mips16_decl_p): Delete.
26824 (mips_nomips16_decl_p): Delete.
26825 (mips_get_compress_on_flags): New function.
26826 (mips_get_compress_off_flags): New function.
26827 (mips_get_compress_mode): New function.
26828 (mips_get_compress_on_name): New function.
26829 (mips_get_compress_off_name): New function.
26830 (mips_insert_attributes): Support multiple compression types.
26831 (mips_merge_decl_attributes): Likewise.
26832 (umips_12bit_offset_address_p): New function.
26833 (mips_start_function_definition): Emit .set micromips directive.
26834 (mips_call_may_need_jalx_p): New function.
26835 (mips_function_ok_for_sibcall): Add microMIPS support.
26836 (mips_print_operand_punctuation): Support short delay slots and
26837 compact jumps.
26838 (umips_swm_mask, umips_swm_encoding): New.
26839 (umips_build_save_restore): New function.
26840 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
26841 (was_mips16_p): Remove.
26842 (old_compression_mode): New.
26843 (mips_set_compression_mode): New function.
26844 (mips_set_current_function): Add microMIPS support.
26845 (mips_option_override): Likewise.
26846 (umips_save_restore_pattern_p): New function.
26847 (umips_output_save_restore): New function.
26848 (umips_load_store_pair_p_1): New function.
26849 (umips_load_store_pair_p): New function.
26850 (umips_output_load_store_pair_1): New function.
26851 (umips_output_load_store_pair): New function.
26852 (umips_movep_target_p) New function.
26853 (mips_prepare_pch_save): Add microMIPS support.
26854 * config/mips/mips.h (TARGET_COMPRESSION): New.
26855 (TARGET_CPU_CPP_BUILTINS): Update macro
26856 to use new compression flags and to support microMIPS.
26857 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
26858 (MIPS_ARCH_FLOAT_SPEC): Likewise.
26859 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
26860 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
26861 (ASM_SPEC): Support mmicromips and mno-micromips.
26862 (M16STORE_REG_P): New macro.
26863 (MIPS_CALL): Support TARGET_MICROMIPS.
26864 (MICROMIPS_J): New macro.
26865 (mips_base_mips16): Rename this...
26866 (mips_base_compression_flags): ...to this.
26867 (UMIPS_12BIT_OFFSET_P): New macro.
26868 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
26869 (MULTILIB_DIRNAMES): Likewise.
26870
26871 2013-02-24 Jakub Jelinek <jakub@redhat.com>
26872
26873 PR target/52555
26874 * target-globals.c (save_target_globals): For init_reg_sets and
26875 target_reinit remporarily set this_fn_optabs to this_target_optabs.
26876
26877 2013-02-22 James Greenhalgh <james.greenhalgh@arm.com>
26878
26879 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
26880 * config/aarch64/t-aarch64
26881 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
26882
26883 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
26884
26885 PR inline-asm/56148
26886 * lra-constraints.c (process_alt_operands): Reload operand
26887 conflicting with earlier clobber only if no more other conflicting
26888 operands.
26889
26890 2013-02-22 Jakub Jelinek <jakub@redhat.com>
26891
26892 PR sanitizer/56393
26893 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
26894 if not linking a shared library.
26895
26896 2013-02-22 Seth LaForge <sethml@google.com>
26897
26898 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
26899
26900 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
26901
26902 * config/arm/arm.md (split for extendsidi): Update condition.
26903 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
26904 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
26905 (qhs_zextenddi_cstr): Likewise.
26906
26907 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26908
26909 PR middle-end/56420
26910 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
26911 avoid signed wrapping.
26912 (expand_mult): Handle properly multiplication by
26913 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
26914 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
26915 in the compiler if coeff is HOST_WIDE_INT_MIN.
26916 (expand_divmod): Don't make ext_op1 static, change it's type to
26917 uhwi. Avoid undefined behavior in -INTVAL (op1).
26918
26919 PR rtl-optimization/50339
26920 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
26921 field.
26922 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
26923 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
26924 into splitting_ashiftrt field.
26925 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
26926 ASHIFTRT.
26927 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
26928 choices.
26929
26930 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
26931
26932 PR middle-end/56108
26933 * trans-mem.c (execute_tm_mark): Do not expand transactions that
26934 are sure to go irrevocable.
26935
26936 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
26937
26938 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
26939 scalars are valid operands.
26940
26941 2013-02-21 Martin Jambor <mjambor@suse.cz>
26942
26943 PR tree-optimization/56310
26944 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
26945 only matching indices and non-negative final offsets.
26946 (intersect_aggregates_with_edge): Pass src_idx to
26947 agg_replacements_to_vector. Pass src_idx insstead of index to
26948 intersect_with_agg_replacements.
26949
26950 2013-02-21 Martin Jambor <mjambor@suse.cz>
26951
26952 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
26953 instead of hard-wired defaults.
26954
26955 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
26956
26957 * doc/invoke.texi (MIPS Options): Update documentation of the
26958 floating-point multiply-accumulate instruction restrictions.
26959
26960 2013-02-21 Kostya Serebryany <kcc@google.com>
26961
26962 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
26963 asan_shadow_offset on x86_64 linux.
26964
26965 2013-02-21 Richard Biener <rguenther@suse.de>
26966
26967 PR tree-optimization/56415
26968 Revert
26969 2013-02-11 Richard Biener <rguenther@suse.de>
26970
26971 PR tree-optimization/56273
26972 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
26973 first VRP run.
26974
26975 2013-02-21 Jakub Jelinek <jakub@redhat.com>
26976
26977 PR bootstrap/56258
26978 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
26979 instead of @itemx.
26980
26981 PR inline-asm/56405
26982 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
26983 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
26984
26985 2013-02-20 Jan Hubicka <jh@suse.cz>
26986
26987 PR tree-optimization/56265
26988 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
26989 when target is referenced for first time.
26990
26991 2013-02-20 Richard Biener <rguenther@suse.de>
26992
26993 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
26994 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
26995 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
26996 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
26997 not return anything.
26998 (rename_ssa_copies): Do not remove unused locals.
26999 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
27000 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
27001 * passes.c (execute_function_todo): Do not schedule unused locals
27002 removal if cleanup_tree_cfg did something.
27003 * tree-ssa-live.c (remove_unused_locals): Dump statistics
27004 about the number of removed locals.
27005
27006 2013-02-20 Richard Biener <rguenther@suse.de>
27007
27008 PR tree-optimization/56398
27009 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
27010
27011 2013-02-20 Martin Jambor <mjambor@suse.cz>
27012
27013 PR tree-optimization/55334
27014 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
27015 restricted pointers to arrays.
27016
27017 2013-02-20 Richard Biener <rguenther@suse.de>
27018 Jakub Jelinek <jakub@redhat.com>
27019
27020 PR tree-optimization/56396
27021 * tree-ssa-ccp.c (n_const_val): New static variable.
27022 (get_value): Return NULL for SSA names we don't have a lattice
27023 entry for.
27024 (ccp_initialize): Initialize n_const_val.
27025 * tree-ssa-copy.c (n_copy_of): New static variable.
27026 (init_copy_prop): Initialize n_copy_of.
27027 (get_value): Return NULL_TREE for SSA names we don't have a
27028 lattice entry for.
27029
27030 2013-02-20 Martin Jambor <mjambor@suse.cz>
27031
27032 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
27033
27034 2013-02-20 Richard Biener <rguenther@suse.de>
27035
27036 * genpreds.c (write_lookup_constraint): Do not compare first
27037 letter of the constraint again.
27038
27039 2013-02-20 Richard Biener <rguenther@suse.de>
27040
27041 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
27042 and ceil_log2.
27043 (get_use_iv_cost): Terminate hashtable walk when coming across
27044 an empty entry.
27045
27046 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
27047
27048 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
27049 reassociation for avx2 targets.
27050
27051 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27052
27053 * config/microblaze/microblaze.c: microblaze_has_clz = 0
27054 Add version check for v8.10.a to enable microblaze_has_clz
27055 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
27056 version and TARGET_PATTERN_COMPARE check
27057 * config/microblaze/microblaze.md: New clzsi2 instruction
27058
27059 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27060
27061 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
27062 function before branching.
27063
27064 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27065
27066 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
27067 DUMP_INSN_RTX_UID.
27068 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
27069
27070 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
27071
27072 PR middle-end/55889
27073 * sel-sched.c: Include ira.h.
27074 (implicit_clobber_conflict_p): New function.
27075 (moveup_expr): Use it.
27076 * Makefile.in (sel-sched.o): Depend on ira.h.
27077
27078 2013-02-19 Richard Biener <rguenther@suse.de>
27079
27080 PR tree-optimization/56384
27081 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
27082 (vn_hash_type): Split out from ...
27083 (vn_hash_constant_with_type): ... here.
27084 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
27085 (vn_phi_eq): Compare types from vn_phi_s structure.
27086 (vn_phi_lookup): Populate vn_phi_s type.
27087 (vn_phi_insert): Likewise.
27088
27089 2013-02-19 Jakub Jelinek <jakub@redhat.com>
27090
27091 PR tree-optimization/56350
27092 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
27093 if haven't found reduction or nested cycle operand, rather than
27094 asserting we must find it.
27095
27096 PR tree-optimization/56381
27097 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
27098 to fold_build3.
27099
27100 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
27101 Jakub Jelinek <jakub@redhat.com>
27102
27103 PR target/52555
27104 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
27105 (swap_optab_enable): Same.
27106 (init_all_optabs): Use argument instead of global.
27107 * tree.h (struct tree_optimization_option): New field target_optabs.
27108 * expr.h (init_all_optabs): Add argument to prototype.
27109 (TREE_OPTIMIZATION_OPTABS): New.
27110 (save_optabs_if_changed): Protoize.
27111 * optabs.h: Declare this_fn_optabs.
27112 * optabs.c (save_optabs_if_changed): New.
27113 Declare this_fn_optabs.
27114 (init_optabs): Add argument to init_all_optabs() call.
27115 * function.c (invoke_set_current_function_hook): Handle per
27116 function optabs.
27117 * function.h (struct function): New field optabs.
27118 * config/mips/mips.c (mips_set_mips16_mode): Handle when
27119 optimization_current_node has changed.
27120 * target-globals.h (save_target_globals_default_opts): Protoize.
27121 * target-globals.c (save_target_globals_default_opts): New.
27122
27123 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27124
27125 PR target/56347
27126 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
27127 registers %fr12 and %fr12R as call used.
27128
27129 PR target/56214
27130 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
27131 and HImode, require all displacements to be an integer multiple of
27132 their mode size.
27133 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
27134 only allow QImode and HImode when reload is in progress and strict is
27135 true. Likewise for symbolic addresses. Use base14_operand to check
27136 displacements in REG+BASE addresses.
27137
27138 2013-02-18 Richard Biener <rguenther@suse.de>
27139
27140 PR tree-optimization/56366
27141 * tree-vect-loop.c (get_initial_def_for_induction): Properly
27142 handle sign-conversion of outer-loop initial induction value.
27143
27144 2013-02-18 Richard Biener <rguenther@suse.de>
27145
27146 PR middle-end/56349
27147 * cfghooks.c (merge_blocks): If we merge a latch into another
27148 block adjust references to it.
27149 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
27150 (verify_loop_structure): Verify that a recorded latch is in fact
27151 a latch.
27152
27153 2013-02-18 Richard Biener <rguenther@suse.de>
27154
27155 PR tree-optimization/56321
27156 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
27157 order SSA name release and virtual operand unlinking.
27158
27159 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27160
27161 * config/microblaze/microblaze.md (save_stack_block): Define.
27162 (restore_stack_block): Likewise.
27163
27164 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27165
27166 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
27167 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
27168 * config/microblaze/microblaze.c (microblaze_option_override):
27169 Bail out early for PIC modes when target does not support PIC.
27170
27171 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27172
27173 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
27174 Replace with a microblaze version.
27175 (microblaze_trampoline_init): Adapt for microblaze.
27176 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
27177 microblaze.
27178
27179 2013-02-16 Jakub Jelinek <jakub@redhat.com>
27180 Dodji Seketeli <dodji@redhat.com>
27181
27182 PR asan/56330
27183 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
27184 (instrument_mem_region_access): Do not forget to always put
27185 instrumentation of the of 'base' and 'base + len' in a "if (len !=
27186 0) statement, even for cases where either 'base' or 'base + len'
27187 are not instrumented -- because they have been previously
27188 instrumented. Simplify the logic by putting all the statements
27189 instrument 'base + len' inside a sequence, and then insert that
27190 sequence right before the current insertion point. Then, to
27191 instrument 'base + len', just get an iterator on that statement.
27192 And do not forget to update the pointer to iterator the function
27193 received as argument.
27194
27195 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
27196
27197 PR rtl-optimization/56348
27198 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
27199
27200 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
27201
27202 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
27203 (clean_graph_dump_file): Pass base to start_graph_dump.
27204
27205 2013-02-14 Richard Henderson <rth@redhat.com>
27206
27207 PR target/55941
27208 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
27209
27210 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
27211
27212 * collect2-aix.h: Define F_LOADONLY.
27213
27214 2013-02-14 Richard Biener <rguenther@suse.de>
27215
27216 PR lto/50494
27217 * varasm.c (output_constant_def_1): Get the decl representing
27218 the constant as argument.
27219 (output_constant_def): Wrap output_constant_def_1.
27220 (make_decl_rtl): Use output_constant_def_1 with the decl
27221 representing the constant.
27222 (build_constant_desc): Optionally re-use a decl already
27223 representing the constant.
27224 (tree_output_constant_def): Adjust.
27225
27226 2013-02-14 Dodji Seketeli <dodji@redhat.com>
27227
27228 Fix an asan crash
27229 * asan.c (instrument_builtin_call): Really put the length of the
27230 second source argument into src1_len.
27231
27232 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27233
27234 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
27235 argument. If it is false, don't create edge from then_bb to
27236 fallthru_bb.
27237 (insert_if_then_before_iter): Pass true to it.
27238 (build_check_stmt): Pass false to it.
27239 (transform_statements): Flush hash table only on extended basic
27240 block boundaries, rather than at the beginning of every bb.
27241 Don't flush hash table on nonfreeing_call_p calls.
27242 * tree-flow.h (nonfreeing_call_p): New prototype.
27243 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
27244
27245 2013-02-13 David S. Miller <davem@davemloft.net>
27246
27247 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
27248
27249 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
27250
27251 PR target/56184
27252 * ira.c (max_regno_before_ira): Move from ...
27253 (ira): ... here.
27254 (fix_reg_equiv_init): Use max_regno_before_ira instead of
27255 vec_safe_length.
27256
27257 2013-02-13 Jakub Jelinek <jakub@redhat.com>
27258
27259 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
27260
27261 2013-02-13 Richard Biener <rguenther@suse.de>
27262
27263 PR lto/56295
27264 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
27265 globals in MEM_REFs.
27266
27267 2013-02-13 Richard Biener <rguenther@suse.de>
27268
27269 * loop-init.c (loop_optimizer_init): Clear loop state when
27270 re-initializing preserved loops.
27271 * loop-unswitch.c (unswitch_single_loop): Return whether
27272 we unswitched the loop. Do not verify loop state here.
27273 (unswitch_loops): When we unswitched a loop discover new loops.
27274
27275 2013-02-13 Kostya Serebryany <kcc@google.com>
27276
27277 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
27278 on x86_64 linux.
27279 * sanitizer.def: Rename __asan_init to __asan_init_v1.
27280
27281 2013-02-12 Dodji Seketeli <dodji@redhat.com>
27282
27283 Avoid instrumenting duplicated memory access in the same basic block
27284 * Makefile.in (asan.o): Add new dependency on hash-table.h
27285 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
27286 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
27287 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
27288 (free_mem_ref_resources, has_mem_ref_been_instrumented)
27289 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
27290 (get_mem_ref_of_assignment): New functions.
27291 (get_mem_refs_of_builtin_call): Extract from
27292 instrument_builtin_call and tweak a little bit to make it fit with
27293 the new signature.
27294 (instrument_builtin_call): Use the new
27295 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
27296 of is_gimple_builtin_call.
27297 (instrument_derefs, instrument_mem_region_access): Insert the
27298 instrumented memory reference into the hash table.
27299 (maybe_instrument_assignment): Renamed instrument_assignment into
27300 this, and change it to advance the iterator when instrumentation
27301 actually happened and return true in that case. This makes it
27302 homogeneous with maybe_instrument_assignment, and thus give a
27303 chance to callers to be more 'regular'.
27304 (transform_statements): Clear the memory reference hash table
27305 whenever we enter a new BB, when we cross a function call, or when
27306 we are done transforming statements. Use
27307 maybe_instrument_assignment instead of instrumentation. No more
27308 need to special case maybe_instrument_assignment and advance the
27309 iterator after calling it; it's now handled just like
27310 maybe_instrument_call. Update comment.
27311
27312 2013-02-13 Richard Biener <rguenther@suse.de>
27313
27314 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
27315 Fix loop discovery code.
27316
27317 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
27318
27319 PR inline-asm/56148
27320 * lra-constraints.c (process_alt_operands): Match early clobber
27321 operand with itself. Check conflicts with earlyclobber only if
27322 the operand is not reloaded. Prefer to reload conflicting operand
27323 if earlyclobber and matching operands are the same.
27324
27325 2013-02-12 Richard Biener <rguenther@suse.de>
27326
27327 PR lto/56297
27328 * lto-streamer-out.c (write_symbol): Do not output symbols
27329 for hard register variables.
27330
27331 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
27332
27333 PR target/54222
27334 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
27335 (umulsidi3_insn, mulsidi3_insn): New insns.
27336
27337 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
27338
27339 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
27340 (struct tune_params): Add vec_costs field.
27341 * config/arm/arm.c (arm_builtin_vectorization_cost)
27342 (arm_add_stmt_cost): New functions.
27343 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
27344 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
27345 (arm_default_vec_cost): New struct of type cpu_vec_costs.
27346 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
27347 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
27348 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
27349 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
27350
27351 2013-02-12 Richard Biener <rguenther@suse.de>
27352
27353 PR lto/56295
27354 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
27355 decls again if possible.
27356
27357 2013-02-12 Richard Biener <rguenther@suse.de>
27358
27359 PR middle-end/56288
27360 * tree-ssa.c (verify_ssa_name): Fix check, move
27361 SSA_NAME_IN_FREE_LIST check up.
27362
27363 2013-02-12 Jakub Jelinek <jakub@redhat.com>
27364 Steven Bosscher <steven@gcc.gnu.org>
27365
27366 PR rtl-optimization/56151
27367 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
27368 equal to op0 or op1, and last_insn pattern is CODE operation
27369 with MEM dest and one of the operands matches that MEM.
27370
27371 2013-02-11 Sriraman Tallam <tmsriram@google.com>
27372
27373 * doc/extend.texi: Document Function Multiversioning and "default"
27374 parameter string to target attribute.
27375 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
27376 target attribute parameter is "default".
27377 (ix86_compare_version_priority): Remove checks for target attribute.
27378 (ix86_mangle_function_version_assembler_name): Change error to sorry.
27379 Remove check for target attribute equal to NULL. Add assert.
27380 (ix86_generate_version_dispatcher_body): Change error to sorry.
27381
27382 2013-02-11 Iain Sandoe <iain@codesourcery.com>
27383 Jack Howarth <howarth@bromo.med.uc.edu>
27384 Patrick Marlier <patrick.marlier@gmail.com>
27385
27386 PR libitm/55693
27387 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
27388 define ENDFILE_SPEC as TM_DESTRUCTOR.
27389 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
27390
27391 2013-02-11 Alexander Potapenko <glider@google.com>
27392 Jack Howarth <howarth@bromo.med.uc.edu>
27393 Jakub Jelinek <jakub@redhat.com>
27394
27395 PR sanitizer/55617
27396 * config/darwin.c (cdtor_record): Rename ctor_record.
27397 (sort_cdtor_records): Rename sort_ctor_records.
27398 (finalize_dtors): New routine to sort destructors by
27399 priority before use in assemble_integer.
27400 (machopic_asm_out_destructor): Use finalize_dtors if needed.
27401
27402 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
27403
27404 PR rtl-optimization/56275
27405 * simplify-rtx.c (avoid_constant_pool_reference): Check that
27406 offset is non-negative and less than cmode size before
27407 calling simplify_subreg.
27408
27409 2013-02-11 Richard Biener <rguenther@suse.de>
27410
27411 PR tree-optimization/56264
27412 * cfgloop.h (fix_loop_structure): Adjust prototype.
27413 * loop-init.c (fix_loop_structure): Return the number of
27414 newly discovered loops.
27415 * tree-cfgcleanup.c (repair_loop_structures): When new loops
27416 are discovered, do a full loop-closed SSA rewrite.
27417
27418 2013-02-11 Richard Biener <rguenther@suse.de>
27419
27420 PR tree-optimization/56273
27421 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
27422 first VRP run.
27423 (check_array_ref): Fix missing newline in dumps.
27424 (search_for_addr_array): Likewise.
27425
27426 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
27427
27428 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
27429
27430 2013-02-09 Jakub Jelinek <jakub@redhat.com>
27431
27432 PR target/56256
27433 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
27434
27435 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
27436
27437 PR rtl-optimization/56246
27438 * lra-constraints.c (simplify_operand_subreg): Try to reuse
27439 reload pseudo.
27440 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
27441 constraints are satisfied.
27442
27443 2013-02-08 Jeff Law <law@redhat.com>
27444
27445 PR debug/53948
27446 * emit-rtl.c (reg_is_parm_p): New function.
27447 * regs.h (reg_is_parm_p): New prototype.
27448 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
27449 callee-clobbered registers.
27450
27451 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
27452
27453 PR target/56043
27454 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
27455 If there is no implicit builtin declaration, just return NULL.
27456
27457 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27458
27459 * config/i386/sse.md (FMAMODEM): New mode iterator.
27460 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
27461 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
27462
27463 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27464
27465 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
27466 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
27467 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
27468
27469 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
27470
27471 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
27472 (microblaze*-*-elf): Likewise.
27473 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
27474 LINK_SPEC.
27475 * config/microblaze/microblaze-c.c: Add builtin defines for
27476 _LITTLE_ENDIAN and _BIG_ENDIAN.
27477 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
27478 add to TARGET_DEFAULT flags.
27479 Expand ASM_SPEC and LINK_SPEC.
27480 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
27481 * config/microblaze/microblaze.md: Update extendsidi2 and
27482 movdi_internal instructions to use low-order / high-order reg
27483 print_operands.
27484 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
27485 options and inversemask / mask of LITTLE_ENDIAN.
27486 * config/microblaze/t-microblaze: Expand multilib options to
27487 include mlittle-endian (le) and update exceptions patterns.
27488
27489 2013-02-08 Jakub Jelinek <jakub@redhat.com>
27490
27491 PR rtl-optimization/56195
27492 * lra-constraints.c (get_reload_reg): Don't reuse regs
27493 if they have smaller mode than requested, if they have
27494 wider mode than requested, try to return a SUBREG.
27495
27496 PR tree-optimization/56250
27497 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
27498 if type is unsigned and code isn't MULT_EXPR.
27499
27500 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27501
27502 PR tree-optimization/56064
27503 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
27504 bits according to mode.
27505 * fixed-value.h (fixed_from_double_int)
27506 (const_fixed_from_double_int): Adjust comments.
27507
27508 2013-02-08 Richard Biener <rguenther@suse.de>
27509
27510 PR lto/56231
27511 * lto-streamer.h (struct data_in): Remove current_file, current_line
27512 and current_col members.
27513 * lto-streamer-out.c (lto_output_location): Stream changed bits
27514 en-block for efficiency.
27515 * lto-streamer-in.c (clear_line_info): Remove.
27516 (lto_input_location): Cache current file, line and column
27517 globally via local statics. Read changed bits en-block.
27518 (input_function): Do not call clear_line_info.
27519 (lto_read_body): Likewise.
27520 (lto_input_toplevel_asms): Likewise.
27521
27522 2013-02-08 Michael Matz <matz@suse.de>
27523
27524 PR tree-optimization/52448
27525 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
27526 (nt_call_phase): New static.
27527 (add_or_mark_expr): Only mark accesses with newer phase than any
27528 call seen.
27529 (nonfreeing_call_p): New.
27530 (nt_init_block): Update nt_call_phase, mark blocks as visited.
27531 (nt_fini_block): Keep blocks marked as visited.
27532 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
27533
27534 2013-02-08 Richard Biener <rguenther@suse.de>
27535
27536 * ira.c (ira): Free broken dominator information.
27537
27538 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
27539
27540 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
27541
27542 2013-02-08 Marek Polacek <polacek@redhat.com>
27543
27544 * cfgloop.c (verify_loop_structure): Add more checking of headers.
27545
27546 2013-02-08 Richard Biener <rguenther@suse.de>
27547
27548 PR middle-end/56181
27549 * cfgloop.h (flow_loops_find): Adjust.
27550 (bb_loop_header_p): Declare.
27551 * cfgloop.c (bb_loop_header_p): New function split out from ...
27552 (flow_loops_find): ... here. Adjust function signature,
27553 support incremental loop structure update.
27554 (verify_loop_structure): Cleanup. Verify a loop is a loop.
27555 * cfgloopmanip.c (fix_loop_structure): Move ...
27556 * loop-init.c (fix_loop_structure): ... here.
27557 (apply_loop_flags): Split out from ...
27558 (loop_optimizer_init): ... here.
27559 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
27560 in incremental mode, only remove dead loops here.
27561
27562 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
27563
27564 PR target/54222
27565 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
27566 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
27567 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
27568 (*round<mode>3.libgcc): New insns for fixed-modes.
27569 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
27570 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
27571 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
27572 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
27573 implementations. Define to __builtin_avr_absFX,
27574 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
27575 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
27576 __builtin_avr_countlsFX, respectively.
27577 * config/avr/avr-c.c (target.h): Include it.
27578 (enum avr_builtin_id): New enum.
27579 (avr_resolve_overloaded_builtin): New static function.
27580 (avr_register_target_pragmas): Use it to set
27581 targetm.resolve_overloaded_builtin.
27582 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
27583 tree nodes used by DEF_BUILTIN.
27584 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
27585 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
27586 <AVR_BUILTIN_xxBITS>: Same.
27587
27588 2013-02-08 Richard Biener <rguenther@suse.de>
27589
27590 * cfgloop.c (verify_loop_structure): Properly handle
27591 a loop exiting to another loop header.
27592 * ira-int.h (ira_loops): Remove.
27593 * ira.c (ira_loops): Remove.
27594 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
27595 (do_reload): Use loop_optimizer_finalize.
27596 * ira-build.c (create_loop_tree_nodes): Use get_loops and
27597 number_of_loops to access the loop tree.
27598 (more_one_region_p): Likewise.
27599 (finish_loop_tree_nodes): Likewise.
27600 (rebuild_regno_allocno_maps): Likewise.
27601 (mark_loops_for_removal): Likewise.
27602 (mark_all_loops_for_removal): Likewise.
27603 (remove_unnecessary_regions): Likewise.
27604 (ira_build): Likewise.
27605 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
27606
27607 2013-02-08 Richard Biener <rguenther@suse.de>
27608
27609 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
27610 * ipa-pure-const.c (analyze_function): Avoid calling
27611 mark_irreducible_loops twice.
27612 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
27613
27614 2013-02-07 David S. Miller <davem@davemloft.net>
27615
27616 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
27617 on 'reg'.
27618 * var-tracking.c (vt_add_function_parameter): Test the presence of
27619 HAVE_window_save properly and do not remap argument registers when
27620 we have a leaf function.
27621
27622 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
27623
27624 PR bootstrap/56227
27625 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
27626 instead of "ll".
27627 * config/i386/i386.c (ix86_print_operand): Ditto.
27628
27629 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27630
27631 * lra-constraints.c (process_alt_operands): Fix recently added comment.
27632
27633 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
27634
27635 PR rtl-optimization/56225
27636 * lra-constraints.c (process_alt_operands): Check that reload hard
27637 reg can hold value for strict_low_part.
27638
27639 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27640
27641 PR debug/56154
27642 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
27643 dwarf2out_end_function.
27644 (in_first_function_p, maybe_at_text_label_p,
27645 first_loclabel_num_not_at_text_label): New variables.
27646 (dwarf2out_var_location): In the first function find out
27647 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
27648 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
27649 functions.
27650
27651 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
27652
27653 PR rtl-optimization/56178
27654 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
27655 SUBREG of a register. Tidy up related block of code.
27656 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
27657 note if the source is a register or a SUBREG of a register.
27658
27659 2013-02-07 Jakub Jelinek <jakub@redhat.com>
27660
27661 PR target/56228
27662 * config/rs6000/rs6000.md (ptrm): New mode attr.
27663 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
27664 call_value_indirect_aix<pttrsize>,
27665 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
27666 m in constraints.
27667
27668 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
27669
27670 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
27671 if -bnortl. Convert to strcmp and strncmp.
27672
27673 2013-02-07 Alan Modra <amodra@gmail.com>
27674
27675 PR target/54009
27676 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
27677 addresses won't wrap when offsetting.
27678 (rs6000_secondary_reload): Provide secondary reloads needed for
27679 wrapping LO_SUM addresses.
27680
27681 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
27682
27683 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
27684 MACH, just __MACH__.
27685
27686 2013-02-06 Richard Biener <rguenther@suse.de>
27687
27688 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
27689 instead of calling fix_loop_structure.
27690
27691 2013-02-06 Jakub Jelinek <jakub@redhat.com>
27692
27693 PR middle-end/56217
27694 * omp-low.c (use_pointer_for_field): Return false if
27695 lower_send_shared_vars doesn't generate any copy-out code.
27696
27697 2013-02-06 Tom de Vries <tom@codesourcery.com>
27698
27699 PR rtl-optimization/56131
27700 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
27701 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
27702 of the label is NULL. Add comment.
27703
27704 2013-02-05 Jakub Jelinek <jakub@redhat.com>
27705
27706 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
27707
27708 PR sanitizer/55374
27709 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
27710 (STATIC_LIBTSAN_LIBS): Likewise.
27711 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
27712 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
27713 is defined, don't add anything else beyond that.
27714 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
27715 (LINK_COMMAND_SPEC): Use them.
27716
27717 PR tree-optimization/56205
27718 * tree-stdarg.c (check_all_va_list_escapes): Return true if
27719 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
27720 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
27721
27722 2013-02-05 Richard Biener <rguenther@suse.de>
27723
27724 PR tree-optimization/53342
27725 PR tree-optimization/53185
27726 * tree-vectorizer.h (vect_check_strided_load): Remove.
27727 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
27728 not disallow peeling for vectorized strided loads.
27729 (vect_check_strided_load): Make static and simplify.
27730 (vect_analyze_data_refs): Adjust.
27731 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
27732 correctly when vectorizing strided loads.
27733
27734 2013-02-05 Richard Biener <rguenther@suse.de>
27735
27736 * doc/install.texi: Refer to ISL, not PPL.
27737
27738 2013-02-05 Jan Hubicka <jh@suse.cz>
27739
27740 PR tree-optimization/55789
27741 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
27742
27743 2013-02-05 Jan Hubicka <jh@suse.cz>
27744
27745 PR tree-optimization/55789
27746 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
27747 the dead call anyway.
27748
27749 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
27750
27751 PR sanitizer/55374
27752 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
27753
27754 2013-02-04 Alexander Potapenko <glider@google.com>
27755 Jack Howarth <howarth@bromo.med.uc.edu>
27756 Jakub Jelinek <jakub@redhat.com>
27757
27758 PR sanitizer/55617
27759 * config/darwin.c (sort_ctor_records): Stabilized qsort
27760 on constructor priority by using original position.
27761 (finalize_ctors): New routine to sort constructors by
27762 priority before use in assemble_integer.
27763 (machopic_asm_out_constructor): Use finalize_ctors if needed.
27764
27765 2013-02-04 Jakub Jelinek <jakub@redhat.com>
27766
27767 PR libstdc++/54314
27768 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
27769 about visibility on artificial decls.
27770 * config/sol2.c (solaris_assemble_visibility): Likewise.
27771
27772 2013-02-04 Kai Tietz <ktietz@redhat.com>
27773
27774 PR target/56186
27775 * config/i386/i386.c (function_value_ms_64): Add additional valtype
27776 argument and improve checking of return-argument types for 16-byte
27777 modes.
27778 (ix86_function_value_1): Add additional valtype argument on call
27779 of function_value_64.
27780 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
27781 handling infunction_value_64 function.
27782
27783 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
27784
27785 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
27786
27787 2013-02-04 Richard Biener <rguenther@suse.de>
27788
27789 PR tree-optimization/56188
27790 * tree-ssa-structalias.c (label_visit): Consider case with
27791 initially non-empty points-to set.
27792 (perform_var_substitution): Dump node mapping and clean up.
27793
27794 2013-02-04 Richard Guenther <rguenther@suse.de>
27795
27796 PR lto/56168
27797 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
27798 node prevail as last resort.
27799 (lto_symtab_merge_decls): Remove guard on LTRANS here.
27800 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
27801
27802 2013-02-04 Richard Biener <rguenther@suse.de>
27803
27804 PR tree-optimization/56113
27805 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
27806 Merge into ...
27807 (equiv_class_lookup_or_add): ... this.
27808 (label_visit): Adjust and fix error in previous patch.
27809 (perform_var_substitution): Adjust.
27810
27811 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
27812
27813 * config/sh/divtab.c: Fix formatting and comments throughout the file.
27814 * config/sh/sh4-300.md: Likewise.
27815 * config/sh/sh4a.md: Likewise.
27816 * config/sh/constraints.md: Likewise.
27817 * config/sh/sh.md: Likewise.
27818 * config/sh/netbsd-elf.h: Likewise.
27819 * config/sh/predicates.md: Likewise.
27820 * config/sh/sh-protos.h: Likewise.
27821 * config/sh/ushmedia.h: Likewise.
27822 * config/sh/linux.h: Likewise.
27823 * config/sh/sh.c: Likewise.
27824 * config/sh/superh.h: Likewise.
27825 * config/sh/elf.h: Likewise.
27826 * config/sh/sh4.md: Likewise.
27827 * config/sh/sh.h: Likewise.
27828
27829 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
27830
27831 * config/pa/constraints.md: Adjust unused letters. Change "T"
27832 constraint to match_test floating_point_store_memory_operand().
27833 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
27834 (base14_operand): New.
27835 (floating_point_store_memory_operand): New.
27836 (integer_store_memory_operand): Revise to use base14_operand and
27837 reg_plus_base_memory_operand.
27838 (move_dest_operand): Allow symbolic_memory_operands.
27839 (symbolic_memory_operand): Check for LO_SOM.
27840 (symbolic_operand): Change default case to break.
27841 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
27842 CONST_DOUBLE values to be reloaded by putting them into memory when
27843 the destination is a floating point register.
27844 (movdf): Remove code to handle CONST_DOUBLE.
27845 (movsf): Likewise.
27846 (reload_indf_r1): New.
27847 (reload_insf_r1): New.
27848 Consistently use "Q" and "T" constraints with integer and floating
27849 point move instructions, respectively.
27850 (movdi): Remove FAIL.
27851 Change predicate for source operand unamed DImode move from
27852 general_operand to move_src_operand.
27853 (umulsidi3): Change predicate for destination operand to
27854 register_operand.
27855 Likewise for similar unamed patterns.
27856 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
27857 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
27858 (hppa_legitimize_address): Simplify mask calculation.
27859 (pa_emit_move_sequence): Revised handling of secondary reloads from
27860 REG+D addresses for floating point loads and stores. Directly handle
27861 loading CONST0_RTX (mode) to a floating point register.
27862 (pa_secondary_reload): Handle reloading DF and SFmode constant values
27863 to floating point registers. Don't restrict secondary reloads to
27864 floating point registers to integer modes. Revise some comments and
27865 cleanup some code.
27866 (TARGET_LEGITIMATE_ADDRESS_P): Define.
27867 (pa_legitimate_address_p): New.
27868 (pa_legitimize_reload_address): New.
27869 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
27870 (STRICT_REG_OK_FOR_BASE_P): New.
27871 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
27872 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
27873
27874 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
27875 Andrew Dixie <andrewd@gentrack.com>
27876
27877 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
27878 flag set.
27879
27880 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
27881
27882 * expmed.c (extract_bit_field_1): Pass the full width of the
27883 structure to get_best_reg_extraction_insn.
27884
27885 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27886
27887 PR target/54601
27888 * configure.ac (use_cxa_atexit): Add AIX.
27889 * configure: Regenerate.
27890
27891 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
27892
27893 2013-02-01 Jakub Jelinek <jakub@redhat.com>
27894
27895 PR debug/54793
27896 * final.c (need_profile_function): New variable.
27897 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
27898 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
27899 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
27900 notes, targetm.asm_out.function_prologue doesn't emit anything,
27901 HAVE_prologue and profiler should be emitted before prologue,
27902 set need_profile_function instead of emitting it.
27903 (final_scan_insn): If need_profile_function, emit
27904 profile_function on the first NOTE_INSN_BASIC_BLOCK or
27905 NOTE_INSN_FUNCTION_BEG note.
27906
27907 2013-02-01 Richard Henderson <rth@redhat.com>
27908
27909 * config/rs6000/rs6000.md (smulditi3): New.
27910 (umulditi3): New.
27911
27912 * config/alpha/alpha.md (umulditi3): New.
27913
27914 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
27915
27916 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
27917 (ASM_OUTPUT_ALIGNED_LOCAL): New.
27918
27919 2013-02-01 Richard Biener <rguenther@suse.de>
27920
27921 PR tree-optimization/56113
27922 * tree-ssa-structalias.c (label_visit): Reduce work for
27923 single-predecessor nodes.
27924
27925 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
27926
27927 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
27928 range isn't testing for zero.
27929
27930 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
27931
27932 PR middle-end/56113
27933 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
27934
27935 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
27936 Nick Clifton <nickc@redhat.com>
27937
27938 * config/v850/constraints.md (Q): Define as a memory constraint.
27939 * config/v850/predicates.md (label_ref_operand): New predicate.
27940 (e3v5_shift_operand): New predicate.
27941 (ior_operator): New predicate.
27942 * config/v850/t-v850: Add e3v5 multilib.
27943 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
27944 (v850_gen_movdi): Prototype.
27945 * config/v850/v850.c: Add support for e3v5 architecture.
27946 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
27947 TARGET_V850E_UP.
27948 (construct_save_jarl): Add e3v5 long JARL support.
27949 (v850_adjust_insn_length): New function. Adjust length of call
27950 insns when using e3v5 instructions.
27951 (v850_gen_movdi): New function: Generate instructions to move a
27952 DImode value.
27953 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
27954 (CPP_SPEC): Define __v850e3v5__ as appropriate.
27955 (TARGET_USE_FPU): Enable for e3v5.
27956 (CONST_OK_FOR_W): New macro.
27957 (ADJUST_INSN_LENGTH): Define.
27958 * config/v850/v850.md (UNSPEC_LOOP): Define.
27959 (attr cpu): Add v850e3v5.
27960 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
27961 (movdi): New pattern.
27962 (movdi_internal): New pattern.
27963 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
27964 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
27965 (cstoresf4): Likewise.
27966 (cstoredf4): Likewise.
27967 (insv): New pattern.
27968 (rotlso3_a): New pattern.
27969 (rotlsi3_b): New pattern
27970 (rotlsi3_v850e3v5): New pattern.
27971 (doloop_begin): New pattern.
27972 (fix_loop_counter): New pattern.
27973 (doloop_end): New pattern.
27974 (branch_normal): Add e3v5 long branch support.
27975 (branch_invert): Likewise.
27976 (branch_z_normal): Likewise.
27977 (branch_z_invert): Likewise.
27978 (branch_nz_normal): Likewise.
27979 (branch_nz_invert): Likewise.
27980 (call_internal_short): Add e3v5 register-indirect JARL support.
27981 (call_internal_long): Likewise.
27982 (call_value_internal_short): Likewise.
27983 (call_value_internal_long): Likewise.
27984 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
27985 (mloop): New option.
27986 * config.gcc: Add support for configuring v840e3v5 target.
27987 * doc/invoke.texi: Document new v850 specific command line options.
27988
27989 2013-01-31 Paul Koning <ni1d@arrl.net>
27990
27991 PR debug/55059
27992 PR debug/54508
27993 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
27994 children if parent is a class.
27995 (prune_unused_types_prune): Don't add DW_AT_declaration.
27996
27997 2013-01-31 Richard Biener <rguenther@suse.de>
27998
27999 PR tree-optimization/56157
28000 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
28001 match up operand with SLP child.
28002
28003 2013-01-31 Jason Merrill <jason@redhat.com>
28004
28005 PR debug/54410
28006 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
28007 parameters the first time.
28008 (gen_scheduled_generic_parms_dies): Check completeness here.
28009
28010 2013-01-31 Richard Biener <rguenther@suse.de>
28011
28012 PR middle-end/53073
28013 * common.opt (faggressive-loop-optimizations): New flag,
28014 enabled by default.
28015 * doc/invoke.texi (faggressive-loop-optimizations): Document.
28016 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
28017 infer_loop_bounds_from_undefined by it.
28018
28019 2013-01-31 Richard Biener <rguenther@suse.de>
28020
28021 PR tree-optimization/56150
28022 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
28023 visit virtual operands.
28024 (find_uses_to_rename_bb): Likewise.
28025
28026 2013-01-31 Richard Biener <rguenther@suse.de>
28027
28028 PR tree-optimization/56150
28029 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
28030 mixed store non-store stmts.
28031
28032 2013-01-30 Jakub Jelinek <jakub@redhat.com>
28033
28034 PR sanitizer/55374
28035 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
28036 LIBASAN_EARLY_SPEC is defined.
28037 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
28038 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
28039 before %o.
28040 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
28041
28042 PR c++/55742
28043 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
28044 invalid args instead of ICEing on it.
28045 (ix86_valid_target_attribute_tree): Return error_mark_node if
28046 ix86_valid_target_attribute_inner_p failed.
28047 (ix86_valid_target_attribute_p): Return false only if
28048 ix86_valid_target_attribute_tree returned error_mark_node. Allow
28049 target("default") attribute.
28050 (sorted_attr_string): Change argument from const char * to tree,
28051 merge in all target attribute arguments rather than just one.
28052 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
28053 instead of free. Avoid using strcat.
28054 (ix86_mangle_function_version_assembler_name): Mangle
28055 target("default") as if no target attribute is present. Adjust
28056 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
28057 instead of xmalloc and XDELETEVEC instead of free.
28058 (ix86_function_versions): Don't return true if one of the decls
28059 doesn't have target attribute. If they don't and one of the decls
28060 is DECL_FUNCTION_VERSIONED, report an error. Adjust
28061 sorted_attr_string caller. Use XDELETEVEC instead of free.
28062 (ix86_supports_function_versions): Remove.
28063 (make_name): Fix up formatting.
28064 (make_dispatcher_decl): Remove resolver_name and its initialization.
28065 Avoid leaking memory.
28066 (is_function_default_version): Return true if there is
28067 target("default") attribute rather than no target attribute at all.
28068 (make_resolver_func): Avoid leaking memory.
28069 (ix86_generate_version_dispatcher_body): Likewise.
28070 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
28071 * target.def (supports_function_versions): Remove.
28072 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
28073 * doc/tm.texi: Regenerated.
28074
28075 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
28076
28077 PR rtl-optimization/56144
28078 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
28079 for values with side effects.
28080
28081 2013-01-30 Richard Biener <rguenther@suse.de>
28082
28083 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
28084 (sparseset_pop): Likewise.
28085 * cfganal.c (compute_idf): Likewise. Increase work-stack size
28086 to be able to use quick_push in the worker loop.
28087
28088 2013-01-30 Marek Polacek <polacek@redhat.com>
28089
28090 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
28091
28092 2013-01-30 Richard Biener <rguenther@suse.de>
28093
28094 PR lto/56147
28095 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
28096
28097 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
28098
28099 PR tree-optimization/56064
28100 * fixed-value.c (fixed_from_double_int): New function.
28101 * fixed-value.h (fixed_from_double_int): New prototype.
28102 (const_fixed_from_double_int): New static inline function.
28103 * fold-const.c (native_interpret_fixed): New static function.
28104 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
28105 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
28106 (native_encode_fixed): New static function.
28107 (native_encode_expr) <FIXED_CST>: Use it.
28108 (native_interpret_int): Move double_int worker code to...
28109 * double-int.c (double_int::from_buffer): ...this new static method.
28110 * double-int.h (double_int::from_buffer): Prototype it.
28111
28112 2013-01-30 Richard Biener <rguenther@suse.de>
28113
28114 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
28115 New pointer-map and obstack.
28116 (init_alias_vars): Allocate pointer-map and obstack.
28117 (delete_points_to_sets): Free them.
28118 (find_what_var_points_to): Cache result.
28119 (find_what_p_points_to): Adjust for changed interface of
28120 find_what_var_points_to.
28121 (compute_points_to_sets): Likewise.
28122 (ipa_pta_execute): Likewise.
28123
28124 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
28125
28126 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
28127 * configure: Regenerate.
28128 * config.in: Regenerate.
28129 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
28130 #nobits/#progbits if supported.
28131
28132 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
28133
28134 PR target/56121
28135 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
28136 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
28137 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
28138
28139 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28140
28141 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
28142 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
28143
28144 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28145
28146 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
28147 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
28148
28149 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28150
28151 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
28152 declaration.
28153 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
28154 * config/arm/cortex-a7.md: New bypasses using
28155 arm_mac_accumulator_is_result.
28156
28157 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28158
28159 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
28160 (cortex_a7_neon_mla): Likewise.
28161 (cortex_a7_fpfmad): New reservation.
28162 (cortex_a7_fpmacs): Use ffmas and update required units.
28163 (cortex_a7_fpmuld): Update required units and latency.
28164 (cortex_a7_fpmacd): Likewise.
28165 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
28166 (cortex_a7_neon). Likewise.
28167 (bypass) Update participating units.
28168
28169 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
28170
28171 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
28172 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
28173 from fmac to ffma.
28174 * config/arm/vfp11.md (vfp_farith): Use ffmas.
28175 (vfp_fmul): Use ffmad.
28176 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
28177 (cortex_r4_fmacd): Use ffmad.
28178 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
28179 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
28180 (cortex_a9_fmacd): Use ffmad.
28181 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
28182 (cortex_a8_vfp_macd): Use ffmad.
28183 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
28184 (cortex_a5_fpmacd): Use ffmad.
28185 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
28186 (cortex_a15_vfp_macd): Use ffmad.
28187 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
28188
28189 2013-01-29 Jason Merrill <jason@redhat.com>
28190
28191 PR libstdc++/54314
28192 * varasm.c (default_assemble_visibility): Don't warn about
28193 visibility on artificial decls.
28194
28195 2013-01-29 Richard Biener <rguenther@suse.de>
28196
28197 PR tree-optimization/56113
28198 * tree-ssa-structalias.c (equiv_class_lookup): Also return
28199 the bitmap leader.
28200 (label_visit): Free duplicate bitmaps and record the leader instead.
28201 (perform_var_substitution): Adjust.
28202
28203 2013-01-29 Richard Biener <rguenther@suse.de>
28204
28205 PR tree-optimization/55270
28206 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
28207 the CFG, schedule loops for fixup.
28208
28209 2013-01-29 Nick Clifton <nickc@redhat.com>
28210
28211 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
28212 SP_REG.
28213
28214 2013-01-28 Leif Ekblad <leif@rdos.net>
28215
28216 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
28217 * config/i386/i386.h (TARGET_RDOS): New macro.
28218 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
28219 * config/i386/i386.c (ix86_option_override_internal): For 64bit
28220 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
28221 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
28222 DEFAULT_LARGE_SECTION_THRESHOLD.
28223 * config/i386/i386.md (R14_REG, R15_REG): New constants.
28224 * config/i386/rdos.h: New file.
28225 * config/i386/rdos64.h: New file.
28226
28227 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
28228
28229 PR other/54814
28230 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
28231 TEST_HARD_REG_BIT.
28232
28233 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28234
28235 PR rtl-optimization/56117
28236 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
28237 call cselib_lookup_from_insn on the MEM before calling
28238 add_insn_mem_dependence.
28239
28240 2013-01-28 Richard Biener <rguenther@suse.de>
28241
28242 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
28243 to a stmt that didn't have one.
28244 (copy_phis_for_bb): Likewise for PHI arguments.
28245 (copy_debug_stmt): Likewise for debug stmts.
28246
28247 2013-01-28 Richard Biener <rguenther@suse.de>
28248
28249 PR tree-optimization/56034
28250 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
28251 (partition_builtin_p): Adjust.
28252 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
28253 it is the last partition.
28254 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
28255 up the vertex for the definition.
28256 (classify_partition): Classify whether a partition is a
28257 PKIND_REDUCTION, thus has uses outside of the loop.
28258 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
28259 Merge all PKIND_REDUCTION partitions into the last partition.
28260 (tree_loop_distribution): Seed partitions from reductions as well.
28261
28262 2013-01-28 Jakub Jelinek <jakub@redhat.com>
28263
28264 PR tree-optimization/56125
28265 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
28266 pow(x,c) into sqrt(x) * powi(x, n/2) or
28267 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
28268 optimizing for size.
28269 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
28270 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
28271 integer.
28272
28273 PR tree-optimization/56094
28274 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
28275 to UNKNOWN_LOCATION while gimplifying expr.
28276
28277 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
28278
28279 PR target/56114
28280 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
28281 operand 0 in movabs insn template for -masm=intel asm alternative.
28282 (*movabs<mode>_2): Ditto for operand 1.
28283
28284 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
28285
28286 PR target/54663
28287 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
28288 of microblaze-c.o
28289
28290 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28291
28292 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
28293 tm_file.
28294
28295 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
28296
28297 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
28298 Undef to avoid warning.
28299
28300 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
28301
28302 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
28303 * configure: Regenerate.
28304
28305 2013-01-25 Jakub Jelinek <jakub@redhat.com>
28306
28307 PR tree-optimization/56098
28308 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
28309 for stmts with volatile ops.
28310 (cond_store_replacement): Don't optimize if assign has volatile ops.
28311 (cond_if_else_store_replacement_1): Don't optimize if either
28312 then_assign or else_assign have volatile ops.
28313 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
28314 volatile ops.
28315
28316 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28317
28318 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
28319
28320 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28321
28322 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
28323 missing ':' in asm example.
28324
28325 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
28326
28327 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
28328 entries into lane and laneq entries.
28329 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
28330 Remove AdvSIMD scalar modes.
28331 (aarch64_sq<r>dmulh_laneq<mode>): New.
28332 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
28333 modes.
28334 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
28335 builtin implementations to relfect changes in RTL in aarch64-simd.md.
28336 * config/aarch64/iterators.md (VCOND): New.
28337 (VCONQ): New.
28338
28339 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
28340
28341 PR target/54222
28342 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
28343 Add NULL LIBNAME argument to existing definitions.
28344 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
28345 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
28346 * config/avr/avr.c (DEF_BUILTIN): Same.
28347 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
28348 (avr_expand_builtin): Expand to a vanilla call if a libgcc
28349 implementation is available (DECL_ASSEMBLER_NAME is set).
28350 (avr_fold_absfx): New static function.
28351 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
28352 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
28353 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
28354 AVR_BUILTIN_ABSLLK.
28355 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
28356 (abshk, absk, abslk, absllk): Provide as static inline functions.
28357
28358 2013-01-25 Marek Polacek <polacek@redhat.com>
28359
28360 PR tree-optimization/56035
28361 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
28362
28363 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
28364
28365 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
28366 (*movtf_internal_rex64): Add (!o,C) alternative
28367 (*movxf_internal_rex64): Ditto.
28368 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
28369
28370 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
28371
28372 * doc/invoke.texi: fix typo.
28373 * doc/objc.texi: fix typo.
28374
28375 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
28376
28377 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
28378 for the first two alternatives.
28379
28380 2013-01-24 Diego Novillo <dnovillo@google.com>
28381
28382 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
28383 (ggc-zone.o): Remove.
28384 * configure.ac: Remove option --with-gc.
28385 * configure: Re-generate.
28386 * doc/install.texi: Remove documentation for --with-gc.
28387 * gengtype.c (write_enum_defn): Remove. Update all users.
28388 (write_Types_process_field): Remove generation of gt_e_* argument.
28389 (output_type_enum): Remove. Update all users.
28390 (write_enum_defn): Remove. Update all users.
28391 (enum alloc_zone): Remove. Update all users.
28392 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
28393 * ggc-common.c (ggc_splay_alloc): Remove first argument.
28394 Update all callers.
28395 (struct ptr_data): Remove field TYPE. Update all users.
28396 (gt_pch_note_object): Remove argument TYPE. Update all users.
28397 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
28398 Update all users.
28399 * ggc-none.c (ggc_alloc_typed_stat): Remove.
28400 (struct alloc_zone): Remove.
28401 (ggc_internal_alloc_zone_stat): Remove.
28402 (ggc_internal_cleared_alloc_zone_stat): Remove.
28403 * ggc-page.c (ggc_alloc_typed_stat): Remove.
28404 (ggc_pch_count_object): Remove last argument. Update all users.
28405 (ggc_pch_alloc_object): Remove last argument. Update all users.
28406 (struct alloc_zone): Remove.
28407 * ggc-zone.c: Remove.
28408 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
28409 (struct alloc_zone): Remove.
28410 (ggc_alloc_typed_stat): Remove.
28411 (ggc_alloc_typed): Remove.
28412 (ggc_splay_alloc): Remove first argument.
28413 (rtl_zone): Remove. Update all users.
28414 (tree_zone): Remove. Update all users.
28415 (tree_id_zone): Remove. Update all users.
28416 (ggc_internal_zone_alloc_stat): Remove. Update all users.
28417 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
28418 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
28419 * tree-ssanames.c: Remove references to zone allocator in comments.
28420
28421 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
28422
28423 * config/avr/avr.c (avr_out_fract): Make register numbers that
28424 might be outside of source operand signed.
28425
28426 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
28427
28428 * config/i386/constraints.md (Yf): New constraint.
28429 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
28430 of f constraint to conditionaly disable x87 register preferences.
28431 (*movdf_internal): Ditto.
28432 (*movsf_internal): Ditto.
28433
28434 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
28435
28436 PR inline-asm/55934
28437 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
28438 that have operands with impossible constraints.
28439 Add a FIXME for a speed-up opportunity.
28440 * lra-constraints.c (process_alt_operands): Verify that a class
28441 selected from constraints on asms is valid for the operand mode.
28442 (curr_insn_transform): Remove incorrect comment.
28443
28444 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
28445
28446 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
28447 TOC operand is a valid symbol ref in the constant pool.
28448
28449 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
28450
28451 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
28452
28453 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
28454
28455 PR target/54222
28456 * config/avr/stdfix.h: New file.
28457 * t-avr (stdfix-gcc.h): New rule to build it.
28458 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
28459
28460 2013-01-23 Kostya Serebryany <kcc@google.com>
28461
28462 * config/darwin.h: remove dependency on
28463 CoreFoundation (asan on Mac OS).
28464
28465 2013-01-23 Jakub Jelinek <jakub@redhat.com>
28466
28467 PR target/49069
28468 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
28469 instead of cmpdi_operand for first comparison operand.
28470 Don't assert that comparison operands aren't both constants.
28471
28472 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28473
28474 * doc/install.texi (Downloading the Source): Update references to
28475 downloading separate components.
28476
28477 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
28478
28479 * doc/extend.texi (__int128): Improve grammar.
28480
28481 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
28482
28483 PR target/56028
28484 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
28485 alternative to (o,r).
28486 (*movdi_internal_rex64): Remove (!o,n) alternative.
28487 (DImode immediate->memory splitter): Remove.
28488 (DImode immediate->memory peephole2): Remove.
28489 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
28490 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
28491 alternative to (!o,*r).
28492 (*movtf_internal_sse): New pattern.
28493 (*movxf_internal_rex64): New pattern.
28494 (*movxf_internal): Disable for TARGET_64BIT.
28495 (*movdf_internal_rex64): Remove (!o,F) alternative.
28496
28497 2013-01-22 Jakub Jelinek <jakub@redhat.com>
28498
28499 PR middle-end/56074
28500 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
28501 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
28502 * tree-vect-loop-manip.c (find_loop_location): Also ignore
28503 stmt locations where LOCATION_LOCUS of the stmt location is
28504 UNKNOWN_LOCATION or BUILTINS_LOCATION.
28505
28506 PR target/55686
28507 * config/i386/i386.md (UNSPEC_STOS): New.
28508 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
28509 *strsetqi_1): Add UNSPEC_STOS.
28510
28511 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
28512
28513 PR c++/56067
28514 * doc/invoke.texi: Remove left over -Wsynth example.
28515
28516 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28517
28518 PR tree-optimization/56051
28519 * fold-const.c (fold_binary_loc): Don't fold
28520 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
28521 a narrowing conversion, or widening conversion from signed
28522 to unsigned.
28523
28524 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28525
28526 PR rtl-optimization/56023
28527 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
28528 dependent on debug instruction.
28529
28530 2013-01-21 Martin Jambor <mjambor@suse.cz>
28531
28532 PR middle-end/56022
28533 * function.c (allocate_struct_function): Call
28534 invoke_set_current_function_hook earlier.
28535
28536 2013-01-21 Jakub Jelinek <jakub@redhat.com>
28537
28538 * reload1.c (init_reload): Only initialize reload_obstack
28539 during the first call.
28540
28541 2013-01-21 Marek Polacek <polacek@redhat.com>
28542
28543 * cfgloop.c (verify_loop_structure): Fix up grammar.
28544
28545 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
28546
28547 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
28548 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
28549
28550 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28551
28552 PR target/56058
28553 * config/arm/marvell-pj4.md: Update copyright year.
28554 Fix up use of alu to alu_reg and simple_alu_imm.
28555
28556 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
28557
28558 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
28559
28560 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
28561
28562 PR target/55433
28563 * lra-constraints.c (curr_insn_transform): Don't reuse original
28564 insn for secondary memory move when memory mode should be different.
28565
28566 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
28567
28568 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
28569 atomic_storedi_1): New patterns.
28570
28571 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
28572
28573 btver2 pipeline descriptions.
28574 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
28575 descriptions.
28576 * config/i386/i386.md (btver2_decode): New type attributes.
28577 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
28578 type attributes.
28579 * config/i386/btver2.md: New file describing btver2 pipelines.
28580
28581 2013-01-19 Andrew Pinski <apinski@cavium.com>
28582
28583 PR tree-optimization/52631
28584 * tree-ssa-sccvn (visit_use): Before looking up the original
28585 statement, try looking up the simplified expression.
28586
28587 2013-01-19 Anthony Green <green@moxielogic.com>
28588
28589 * config/moxie/moxie.c (moxie_expand_prologue): Set
28590 current_function_static_stack_size.
28591
28592 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28593
28594 PR tree-optimization/56029
28595 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
28596 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
28597
28598 2013-01-18 Sharad Singhai <singhai@google.com>
28599
28600 PR tree-optimization/55995
28601 * dumpfile.c (dump_loc): Print location only if available.
28602 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
28603
28604 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
28605
28606 PR target/55433
28607 * lra-constraints.c (curr_insn_transform): Reuse original insn for
28608 secondary memory move.
28609 (inherit_reload_reg): Use rclass instead of cl for
28610 check_secondary_memory_needed_p.
28611
28612 2013-01-18 Jakub Jelinek <jakub@redhat.com>
28613
28614 PR middle-end/56015
28615 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
28616 the case where writing real complex part of target modifies op1.
28617
28618 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
28619
28620 * config/aarch64/aarch64-simd.md
28621 (aarch64_vcond_internal<mode>): Handle unordered cases.
28622 * config/aarch64/iterators.md (v_cmp_result): New.
28623
28624 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
28625 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
28626
28627 * config/arm/marvell-pj4.md: New file.
28628 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
28629 * config/arm/arm.md (generic_sched): Add marvell_pj4.
28630 (generic_vfp): Likewise.
28631 * config/arm/arm-cores.def: Add marvell-pj4.
28632 * config/arm/arm-tune.md: Regenerate.
28633 * config/arm/arm-tables.opt: Regenerate.
28634 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
28635 * doc/invoke.texi: Document marvell-pj4.
28636
28637 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
28638
28639 * config/aarch64/arm_neon.h: Map scalar types to standard types.
28640
28641 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28642
28643 PR debug/54114
28644 PR debug/54402
28645 PR debug/49888
28646 * var-tracking.c (negative_power_of_two_p): New.
28647 (global_get_addr_cache, local_get_addr_cache): New.
28648 (get_addr_from_global_cache, get_addr_from_local_cache): New.
28649 (vt_canonicalize_addr): Rewrite using the above. Adjust the
28650 heading comment.
28651 (vt_stack_offset_p): Remove.
28652 (vt_canon_true_dep): Always canonicalize loc's address.
28653 (clobber_overlapping_mems): Make sure we have a MEM.
28654 (local_get_addr_clear_given_value): New.
28655 (val_reset): Clear local cached entries.
28656 (compute_bb_dataflow): Create and release the local cache.
28657 Disable duplicate MEMs clobbering.
28658 (emit_notes_in_bb): Clobber MEMs likewise.
28659 (vt_emit_notes): Create and release the local cache.
28660 (vt_initialize, vt_finalize): Create and release the global
28661 cache, respectively.
28662 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
28663
28664 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28665
28666 PR libmudflap/53359
28667 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
28668 not found in the symtab.
28669
28670 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
28671
28672 PR debug/56006
28673 PR rtl-optimization/55547
28674 PR rtl-optimization/53827
28675 PR debug/53671
28676 PR debug/49888
28677 * alias.c (offset_overlap_p): New, factored out of...
28678 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
28679 the conservative special case for symbolic constants. Don't
28680 adjust zero sizes on alignment.
28681
28682 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
28683
28684 PR rtl-optimization/52573
28685 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
28686 REG_UNUSED for the same register.
28687
28688 2013-01-17 Richard Biener <rguenther@suse.de>
28689 Marek Polacek <polacek@redhat.com>
28690
28691 PR rtl-optimization/55833
28692 * loop-unswitch.c (unswitch_loops): Move loop verification...
28693 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
28694 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
28695 Set it to true when we're removing a loop from hierarchy tree in
28696 an irreducible region.
28697 (fix_bb_placements): Adjust caller.
28698 (fix_loop_placements): Likewise.
28699
28700 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
28701
28702 * config/avr/builtins.def (DEF_BUILTIN): Factor out
28703 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
28704 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
28705 Remove ID. Adjust comments.
28706 * config/avr/avr-c.c (avr_builtin_name): Remove.
28707 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
28708 * config/avr/avr.c (avr_tolower): New static function.
28709 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
28710 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
28711 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
28712 default expansion.
28713
28714 2013-01-17 Jan Hubicka <jh@suse.cz>
28715
28716 PR tree-optimization/55273
28717 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
28718
28719 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
28720
28721 PR target/55981
28722 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
28723 store through atomic_store<mode>_1.
28724 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
28725
28726 2013-01-17 Martin Jambor <mjambor@suse.cz>
28727
28728 PR tree-optimizations/55264
28729 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
28730 for virtual methods.
28731 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
28732 virtual methods before inlining is over.
28733 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
28734 virtual functions.
28735 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
28736 non-virtual.
28737
28738 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
28739
28740 PR rtl-optimization/56005
28741 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
28742 pending reads for prefetch.
28743
28744 2013-01-16 Ian Bolton <ian.bolton@arm.com>
28745
28746 * config/aarch64/aarch64.md
28747 (*cstoresi_neg_uxtw): New pattern.
28748 (*cmovsi_insn_uxtw): New pattern.
28749 (*<optab>si3_uxtw): New pattern.
28750 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
28751 (*<optab>si3_insn_uxtw): New pattern.
28752 (*bswapsi2_uxtw): New pattern.
28753
28754 2013-01-16 Richard Biener <rguenther@suse.de>
28755
28756 * tree-inline.c (tree_function_versioning): Remove set but
28757 never used variable.
28758
28759 2013-01-16 Richard Biener <rguenther@suse.de>
28760
28761 PR tree-optimization/55964
28762 * tree-flow.h (rename_variables_in_loop): Remove.
28763 (rename_variables_in_bb): Likewise.
28764 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
28765 (copy_loop_before): Adjust and delete update-ssa status.
28766 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
28767 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
28768 (rename_variables_in_loop): Remove.
28769 (slpeel_update_phis_for_duplicate_loop): Likewise.
28770 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
28771 use available cfg machinery instead of duplicating it.
28772 Update PHI nodes and perform poor-mans SSA update here.
28773 (slpeel_tree_peel_loop_to_edge): Adjust.
28774
28775 2013-01-16 Richard Biener <rguenther@suse.de>
28776
28777 PR tree-optimization/54767
28778 PR tree-optimization/53465
28779 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
28780 (vrp_visit_phi_node): For PHI arguments coming via backedges
28781 drop all symbolical range information.
28782 (execute_vrp): Compute backedges.
28783
28784 2013-01-16 Richard Biener <rguenther@suse.de>
28785
28786 * doc/install.texi: Update CLooG and ISL requirements to
28787 0.18.0 and 0.11.1.
28788
28789 2013-01-16 Christian Bruel <christian.bruel@st.com>
28790
28791 PR target/55301
28792 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
28793 (broken_move): Handle UNSPECV_SP_SWITCH_B.
28794 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
28795
28796 2013-01-16 DJ Delorie <dj@redhat.com>
28797
28798 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
28799 (UNSPECV_SP_SWITCH_E): New.
28800 (sp_switch_1): Change to an unspec.
28801 (sp_switch_2): Change to an unspec. Don't use post-inc when we
28802 replace $r15.
28803
28804 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
28805
28806 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
28807 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
28808 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
28809 (expand_mem_thread_fence): Ditto.
28810 (expand_mem_signal_fence): Ditto.
28811 (expand_atomic_load): Ditto.
28812 (expand_atomic_store): Ditto.
28813
28814 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
28815
28816 PR rtl-optimization/55547
28817 PR rtl-optimization/53827
28818 PR debug/53671
28819 PR debug/49888
28820 * alias.c (memrefs_conflict_p): Set sizes to negative after
28821 AND adjustments.
28822
28823 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28824
28825 PR target/55940
28826 * function.c (thread_prologue_and_epilogue_insns): Always
28827 add crtl->drap_reg to set_up_by_prologue.set, even if
28828 stack_realign_drap is false.
28829
28830 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
28831
28832 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
28833 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
28834 *call): Fix indention.
28835
28836 2013-01-15 Tom de Vries <tom@codesourcery.com>
28837
28838 PR target/55876
28839 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
28840 Update comment.
28841
28842 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
28843
28844 PR rtl-optimization/55153
28845 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
28846
28847 2013-01-15 Martin Jambor <mjambor@suse.cz>
28848
28849 PR tree-optimization/55920
28850 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
28851 accesses as grp_to_be_debug_replaced.
28852
28853 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28854
28855 PR tree-optimization/55920
28856 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
28857 there is non-useless type conversion needed from debug rhs to lhs,
28858 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
28859
28860 2013-01-15 Joseph Myers <joseph@codesourcery.com>
28861 Mikael Pettersson <mikpe@it.uu.se>
28862
28863 PR target/43961
28864 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
28865 Thumb.
28866 (ASM_OUTPUT_CASE_LABEL): Remove.
28867 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
28868 * final.c (shorten_branches): Update alignment of labels before
28869 jump tables if CASE_VECTOR_SHORTEN_MODE.
28870
28871 2013-01-15 Richard Biener <rguenther@suse.de>
28872
28873 PR bootstrap/55961
28874 * system.h: Do not include gmp.h for building host tools.
28875
28876 2013-01-15 Richard Biener <rguenther@suse.de>
28877
28878 PR middle-end/55882
28879 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
28880 account for bitpos when computing alignment.
28881
28882 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
28883
28884 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
28885 (ix86_target_macros_internal): Likewise.
28886
28887 * config/i386/i386.c (m_CORE2I7): Removed.
28888 (m_CORE_HASWELL): New macro.
28889 (m_CORE_ALL): Likewise.
28890 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
28891 (initial_ix86_arch_features): Likewise.
28892 (processor_target_table): Initializations for Core avx2.
28893 (cpu_names): New names "core-avx2".
28894 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
28895 PROCESSOR_CORE_HASWELL.
28896 (ix86_issue_rate): New case.
28897 (ia32_multipass_dfa_lookahead): Likewise.
28898 (ix86_sched_init_global): Likewise.
28899
28900 * config/i386/i386.h (TARGET_HASWELL): New macro.
28901 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
28902 (processor_type): New PROCESSOR_HASWELL.
28903
28904 2013-01-15 Jakub Jelinek <jakub@redhat.com>
28905
28906 PR tree-optimization/55955
28907 * tree-vect-loop.c (vectorizable_reduction): Give up early on
28908 *SHIFT_EXPR and *ROTATE_EXPR codes.
28909
28910 PR tree-optimization/48766
28911 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
28912 -ftrapv disable -fwrapv.
28913
28914 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28915
28916 PR target/55974
28917 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
28918 etc. to 1 and not to __flash.
28919 Use LL suffix for __INT24_MAX__ with -mint8.
28920 Use ULL suffix for __UINT24_MAX__ with -mint8.
28921
28922 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28923
28924 * config/avr/avr-arch.h
28925 (struct base_arch_s): Use typedef avr_arch_t instead.
28926 (struct arch_info_s): Use typedef avr_arch_info_t instead.
28927 (struct mcu_type_s): Use typedef avr_mcu_t instead.
28928 * config/avr/avr.c: Same.
28929 * config/avr/avr-devices.c: Same.
28930 * config/avr/driver-avr.c: Same.
28931 * config/avr/gen-avr-mmcu-texi.c: Same.
28932 * config/avr/avr-mcus.def: Adjust comment.
28933
28934 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
28935
28936 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
28937 * config/aarch64/iterators.md (VALLDI): New.
28938
28939 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
28940 Andi Kleen <ak@linux.intel.com>
28941
28942 PR target/55948
28943 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
28944 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
28945 memmodel flag.
28946
28947 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28948
28949 * config/avr/avr-stdint.h: Remove trailing blanks.
28950 * config/avr/avr-log.h: Same.
28951 * config/avr/avr-arch.h: Same.
28952 * config/avr/avr-devices.c: Same.
28953 * config/avr/avr-dimode.md: Same.
28954 * config/avr/predicates.md: Same.
28955 * config/avr/avr-c.c: Same. And fix typo.
28956
28957 * config/avr/avr-protos.h: Same. And:
28958 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
28959 (init_cumulative_args): Rename to avr_init_cumulative_args.
28960 (expand_prologue): Rename to avr_expand_prologue.
28961 (expand_epilogue): Rename to avr_expand_epilogue.
28962 (adjust_insn_length): Rename to avr_adjust_insn_length.
28963 (notice_update_cc): Rename to avr_notice_update_cc.
28964 (final_prescan_insn): Rename to avr_final_prescan_insn.
28965 * config/avr/avr.c: Same.
28966 * config/avr/avr.h: Same.
28967 * config/avr/avr.md: Remove trailing blanks.
28968 (prologue): Use avr_expand_prologue.
28969 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
28970
28971 2013-01-14 Richard Biener <rguenther@suse.de>
28972
28973 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
28974 verify_location, collect_subblocks): New functions.
28975 (verify_gimple_in_cfg): Verify that locations only reference
28976 BLOCKs in the functions BLOCK tree.
28977
28978 2013-01-14 Richard Biener <rguenther@suse.de>
28979
28980 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
28981 PHI argument.
28982 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
28983 unshare reference.
28984 (insert_out_of_ssa_copy_on_edge): Likewise.
28985 (rewrite_close_phi_out_of_ssa): Likewise.
28986 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
28987 debug expressions.
28988 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
28989 propagated constants.
28990 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
28991 can not be shared.
28992
28993 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
28994
28995 * config/avr/avr-modes.def: Add GPL copyright notice.
28996
28997 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
28998
28999 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
29000 MEMMODEL_MASK to determine memory model.
29001 (atomic_store<mode>): Ditto from operands[2].
29002 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
29003
29004 2013-01-13 Jakub Jelinek <jakub@redhat.com>
29005
29006 PR fortran/55935
29007 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
29008 (fold_gimple_assign): Don't call unshare_expr here.
29009 (fold_ctor_reference): Call unshare_expr.
29010
29011 2013-01-13 Terry Guo <terry.guo@arm.com>
29012
29013 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
29014 * doc/fragments.texi: Document MULTILIB_REUSE.
29015 * gcc.c (multilib_reuse): New internal spec.
29016 (set_multilib_dir): Also search multilib from multilib_reuse.
29017 * genmultilib (tmpmultilib3): Refactor code.
29018 (tmpmultilib4): Ditto.
29019 (multilib_reuse): New multilib argument.
29020
29021 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
29022
29023 * Makefile.in: Update copyright.
29024
29025 2013-01-12 Tom de Vries <tom@codesourcery.com>
29026
29027 PR middle-end/55890
29028 * calls.c (expand_call): Check if arg_nr is valid.
29029
29030 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
29031
29032 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
29033 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
29034 documentation. Add missing '__' in front of
29035 __builtin_ia32_packssdw256.
29036
29037 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29038
29039 PR target/55719
29040 * config/s390/s390.c (s390_preferred_reload_class): Do not return
29041 NO_REGS for larl operands.
29042 (s390_reload_larl_operand): Use s390_load_address instead of
29043 emit_move_insn.
29044
29045 2013-01-11 Richard Biener <rguenther@suse.de>
29046
29047 * tree-cfg.c (verify_node_sharing_1): Split out from ...
29048 (verify_node_sharing): ... here.
29049 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
29050
29051 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
29052
29053 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
29054 Substitute TREECHECKING.
29055 * configure: Regenerate.
29056 * Makefile.in (TREECHECKING): New.
29057
29058 2013-01-11 Richard Guenther <rguenther@suse.de>
29059
29060 PR tree-optimization/44061
29061 * tree-vrp.c (extract_range_basic): Compute zero as
29062 value-range for __builtin_constant_p of function parameters.
29063
29064 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
29065
29066 Update copyright years.
29067
29068 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
29069
29070 PR rtl-optimization/55672
29071 * lra-eliminations.c (mark_not_eliminable): Permit addition with
29072 const to be eliminable.
29073
29074 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
29075
29076 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
29077 * configure: Regenerate.
29078
29079 2013-01-10 Richard Biener <rguenther@suse.de>
29080
29081 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
29082
29083 2013-01-10 Richard Biener <rguenther@suse.de>
29084
29085 PR bootstrap/55792
29086 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
29087 locations for virtual PHI arguments.
29088 (rewrite_update_phi_arguments): Likewise.
29089
29090 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
29091
29092 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
29093 on to assembler.
29094
29095 2013-01-10 Jakub Jelinek <jakub@redhat.com>
29096
29097 PR tree-optimization/55921
29098 * tree-complex.c (expand_complex_asm): New function.
29099 (expand_complex_operations_1): Call it for GIMPLE_ASM.
29100
29101 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
29102
29103 PR target/55718
29104 * config/s390/s390.c (s390_symref_operand_p)
29105 (s390_loadrelative_operand_p): Merge the two functions.
29106 (s390_check_qrst_address, print_operand_address): Add parameters
29107 to s390_loadrelative_operand_p invokation.
29108 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
29109 (s390_reload_larl_operand, s390_secondary_reload): Use
29110 s390_loadrelative_operand_p instead of s390_symref_operand_p.
29111 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
29112
29113 2013-01-09 Mike Stump <mikestump@comcast.net>
29114
29115 * dse.c (record_store): Remove unnecessary assert.
29116
29117 2013-01-09 Jan Hubicka <jh@suse.cz>
29118
29119 PR tree-optimization/55569
29120 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
29121 * cfgloop.h (scale_loop_profile): Likewise.
29122
29123 2013-01-09 Jan Hubicka <jh@suse.cz>
29124
29125 PR lto/45375
29126 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
29127 functions.
29128 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
29129
29130 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
29131
29132 PR middle-end/55114
29133 * expr.h (maybe_emit_group_store): Declare.
29134 * expr.c (maybe_emit_group_store): New function.
29135 * builtins.c (expand_builtin_int_roundingfn): Call it.
29136 (expand_builtin_int_roundingfn_2): Likewise.
29137
29138 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
29139
29140 PR rtl-optimization/55829
29141 * lra-constraints.c (match_reload): Add code for absent output.
29142 (curr_insn_transform): Add code for reloads of matched inputs
29143 without output.
29144
29145 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
29146
29147 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
29148 attribute of movddup insn to DF.
29149 (*vec_interleave_lowv2df): Ditto.
29150 (vec_dupv2df): Ditto.
29151
29152 2013-01-09 Jan Hubicka <jh@suse.cz>
29153
29154 PR tree-optimiation/55875
29155 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
29156 EVERY_ITERATION parameter.
29157 (number_of_iterations_exit): Check if exit is executed every iteration.
29158 (idx_infer_loop_bounds): Similarly here.
29159 (n_of_executions_at_most): Simplify
29160 to only test for cases where statement is dominated by the
29161 particular bound; handle correctly the "postdominance" test.
29162 (scev_probably_wraps_p): Use max loop iterations info
29163 as a global bound first.
29164
29165 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
29166 Nick Clifton <nickc@redhat.com>
29167
29168 * config/v850/v850.md (cbranchsf4): New pattern.
29169 (cstoresf4): New pattern.
29170 (cbranchdf4): New pattern.
29171 (cstoredf4): New pattern.
29172 (movsicc): Disallow floating point comparisons.
29173 (cmpsf_le_insn): Fix order of operators.
29174 (cmpsf_lt_insn): Likewise.
29175 (cmpsf_eq_insn): Likewise.
29176 (cmpdf_le_insn): Likewise.
29177 (cmpdf_lt_insn): Likewise.
29178 (cmpdf_eq_insn): Likewise.
29179 (cmpsf_ge_insn): Use LE comparison.
29180 (cmpdf_ge_insn): Likewise.
29181 (cmpsf_gt_insn): Use LT comparison.
29182 (cmpdf_gt_insn): Likewise.
29183 (cmpsf_ne_insn): Delete pattern.
29184 (cmpdf_ne_insn): Delete pattern.
29185 * config/v850/v850.c (v850_gen_float_compare): Use
29186 gen_cmpdf_eq_insn for NE comparison.
29187 (v850_float_z_comparison_operator)
29188 (v850_float_nz_comparison_operator): Move from here ...
29189 * config/v850/predicates.md: ... to here. Move GT and GE
29190 comparisons into v850_float_z_comparison_operator.
29191 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
29192 Delete prototype.
29193 (v850_float_nz_comparison_operator): Likewise.
29194
29195 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29196
29197 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
29198 with calls to gen_insvsi/gen_insvdi.
29199
29200 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
29201
29202 * config/i386/i386.c (initial_ix86_tune_features): Set up
29203 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
29204
29205 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
29206 Jakub Jelinek <jakub@redhat.com>
29207
29208 PR tree-optimization/48189
29209 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
29210 If nitercst is 0, don't predict the exit edge.
29211
29212 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
29213
29214 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
29215 in asm_fprintf with reg_names.
29216 (aarch64_print_operand_address): Likewise.
29217 (aarch64_return_addr): Likewise.
29218 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
29219
29220 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29221
29222 * config/pa/pa.h (VAL_U6_BITS_P): Define.
29223 (INT_U6_BITS): Likewise.
29224 * config/pa/predicates.md (uint6_operand): New predicate.
29225 (shift5_operand, shift6_operand): Likewise.
29226 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
29227 arith32_operand.
29228 (lshrdi3): Use shift6_operand.
29229 (shrpsi4, shrpdi4): New insn patterns.
29230 (extzv): Delete expander.
29231 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
29232 predicates in unamed zero extract patterns. Tighten common constraint.
29233 (extv): Delete expander.
29234 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
29235 predicates in unamed sign extract patterns. Tighten common constraint.
29236 (insv): Delete expander.
29237 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
29238 predicates in unamed insert patterns. Tighten common constraint.
29239 Change uint32_operand predicate to uint6_operand predicate in unamed
29240 DImode pattern to insert constant values of type 1...1xxxx.
29241
29242 2013-01-04 Jan Hubicka <jh@suse.cz>
29243
29244 PR tree-optimization/55823
29245 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
29246 issue.
29247
29248 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29249 Uros Bizjak <ubizjak@gmail.com>
29250
29251 PR rtl-optimization/55845
29252 * df-problems.c (can_move_insns_across): Stop scanning at
29253 volatile_insn_p source instruction or give up if
29254 across_from .. across_to range contains any volatile_insn_p
29255 instructions.
29256
29257 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29258
29259 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
29260 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
29261 Declare.
29262 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
29263 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
29264
29265 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29266
29267 PR fortran/55341
29268 * asan.c (asan_clear_shadow): New function.
29269 (asan_emit_stack_protection): Use it.
29270
29271 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
29272
29273 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
29274 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
29275 with tab instead of space.
29276
29277 2013-01-08 Nick Clifton <nickc@redhat.com>
29278
29279 * config/rl78/rl78.c (rl78_expand_prologue): Always select
29280 register bank 0 at the start of an interrupt handler.
29281 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
29282 MDBH registers.
29283
29284 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29285
29286 * config/aarch64/aarch64-simd.md
29287 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
29288 (aarch64_simd_bsl): Likewise.
29289 (aarch64_vcond_internal<mode>): Likewise.
29290 (vcond<mode><mode>): Likewise.
29291 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
29292 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
29293
29294 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
29295
29296 * config/aarch64/aarch64-builtins.c
29297 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
29298
29299 2013-01-08 Martin Jambor <mjambor@suse.cz>
29300
29301 PR debug/55579
29302 * tree-sra.c (analyze_access_subtree): Return true also after
29303 potentially creating a debug-only replacement.
29304
29305 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29306
29307 PR middle-end/55890
29308 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
29309
29310 PR tree-optimization/54120
29311 * tree-vrp.c (range_fits_type_p): Don't allow
29312 src_precision < precision from signed vr to unsigned_p
29313 if vr->min or vr->max is negative.
29314 (simplify_float_conversion_using_ranges): Test can_float_p
29315 against CODE_FOR_nothing.
29316
29317 2013-01-08 Jakub Jelinek <jakub@redhat.com>
29318 Richard Biener <rguenther@suse.de>
29319
29320 PR middle-end/55851
29321 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
29322 types instead of just INTEGER_TYPE types.
29323
29324 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
29325
29326 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
29327 TF_SIZE): Define.
29328
29329 2013-01-07 Steve Ellcey <sellcey@mips.com>
29330
29331 PR target/42661
29332 * config/mips/mips.opt: Change mad to mmad to match documentation.
29333
29334 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29335
29336 PR target/55897
29337 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
29338 .progmemx.data now.
29339
29340 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29341
29342 PR target/55897
29343 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
29344 (avr_addrspace_t): Add .section_name field.
29345 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
29346 array size.
29347 (avr_addrspace): Same. Initialize .section_name. Remove last
29348 NULL entry. Put __memx into .progmemx.data.
29349 (progmem_section_prefix): Remove.
29350 (avr_asm_init_sections): No need to initialize progmem_section.
29351 (avr_asm_named_section): Use avr_addrspace[].section_name to get
29352 section name prefix.
29353 (avr_asm_select_section): Ditto. And use get_unnamed_section to
29354 retrieve the progmem section.
29355 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
29356 boundary to run over avr_addrspace[].
29357 (avr_register_target_pragmas): Ditto.
29358
29359 2013-01-06 Jakub Jelinek <jakub@redhat.com>
29360
29361 * varasm.c (output_constant_def_contents): For asan_protect_global
29362 protected strings, adjust DECL_ALIGN if needed, before testing for
29363 anchored symbols.
29364 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
29365 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
29366 normal decls.
29367 (output_object_block): For asan protected decls, emit asan padding
29368 after their contents.
29369 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
29370 (asan_finish_file): Test it here instead.
29371
29372 2013-01-07 Nick Clifton <nickc@redhat.com>
29373 Matthias Klose <doko@debian.org>
29374 Doug Kwan <dougkwan@google.com>
29375 H.J. Lu <hongjiu.lu@intel.com>
29376
29377 PR driver/55470
29378 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
29379
29380 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
29381
29382 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
29383
29384 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
29385
29386 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
29387
29388 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29389
29390 PR target/54461
29391 * doc/install.texi (Cross-Compiler-Specific Options): Document
29392 --with-avrlibc.
29393
29394 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
29395
29396 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
29397 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
29398 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
29399 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
29400 vqmovun_high_s64): Fix source operand number and update copyright.
29401
29402 2013-01-07 Richard Biener <rguenther@suse.de>
29403
29404 PR middle-end/55890
29405 * gimple.h (gimple_call_builtin_p): New overload.
29406 * gimple.c (validate_call): New function.
29407 (gimple_call_builtin_p): Likewise.
29408 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
29409 Use gimple_call_builtin_p.
29410 (find_func_clobbers): Likewise.
29411 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
29412 (strlen_optimize_stmt): Likewise.
29413
29414 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29415
29416 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
29417 (vld1q_dup_*): Likewise.
29418 (vld1_*): Likewise.
29419 (vld1q_*): Likewise.
29420 (vld1_lane_*): Likewise.
29421 (vld1q_lane_*): Likewise.
29422
29423 2013-01-07 Richard Biener <rguenther@suse.de>
29424
29425 * lto-streamer.h (LTO_minor_version): Bump to 2.
29426
29427 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
29428
29429 * config/aarch64/aarch64-protos.h
29430 (aarch64_const_double_zero_rtx_p): Rename to...
29431 (aarch64_float_const_zero_rtx_p): ...this.
29432 (aarch64_float_const_representable_p): New.
29433 (aarch64_output_simd_mov_immediate): Likewise.
29434 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
29435 move immediate case.
29436 * config/aarch64/aarch64.c
29437 (aarch64_const_double_zero_rtx_p): Rename to...
29438 (aarch64_float_const_zero_rtx_p): ...this.
29439 (aarch64_print_operand): Allow printing of new constants.
29440 (aarch64_valid_floating_const): New.
29441 (aarch64_legitimate_constant_p): Check for valid floating-point
29442 constants.
29443 (aarch64_simd_valid_immediate): Likewise.
29444 (aarch64_vect_float_const_representable_p): New.
29445 (aarch64_float_const_representable_p): Likewise.
29446 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
29447 (aarch64_output_simd_mov_immediate): New.
29448 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
29449 (*movdf_aarch64): Likewise.
29450 * config/aarch64/constraints.md (Ufc): New.
29451 (Y): call aarch64_float_const_zero_rtx.
29452 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
29453
29454 2013-01-07 Richard Biener <rguenther@suse.de>
29455
29456 PR tree-optimization/55888
29457 PR tree-optimization/55862
29458 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
29459 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
29460 not if it is contained therein.
29461
29462 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29463
29464 * config/avr/t-avr: Typo.
29465
29466 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
29467
29468 PR55243
29469 * config/avr/t-avr: Don't automatically rebuild
29470 $(srcdir)/config/avr/t-multilib
29471 $(srcdir)/config/avr/avr-tables.opt
29472 $(srcdir)/doc/avr-mmcu.texi
29473 (avr-mcus): New phony target to build them on request.
29474 (s-avr-mlib, s-avr-mmcu-texi): Remove.
29475 * avr/avr-mcus.def: Adjust comments.
29476
29477 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
29478
29479 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
29480
29481 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29482
29483 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
29484
29485 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
29486
29487 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
29488
29489 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
29490
29491 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
29492 to generate profiling.
29493 * config/rs6000/aix64.h (LIB_SPEC): Same.
29494
29495 2013-01-04 Andrew Pinski <apinski@cavium.com>
29496
29497 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
29498 New function.
29499 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
29500
29501 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
29502
29503 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
29504 unconditionally.
29505 (ix86_expand_move): Ditto.
29506 (ix86_zero_extend_to_Pmode): Ditto.
29507 (ix86_expand_call): Ditto.
29508 (ix86_expand_special_args_builtin): Ditto.
29509 (ix86_expand_builtin): Ditto.
29510
29511 2013-01-04 Richard Biener <rguenther@suse.de>
29512
29513 PR tree-optimization/55862
29514 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
29515 translating them through PHI nodes.
29516
29517 2013-01-04 Martin Jambor <mjambor@suse.cz>
29518
29519 PR tree-optimization/55755
29520 * tree-sra.c (sra_modify_assign): Do not check that an access has no
29521 children when trying to avoid producing a VIEW_CONVERT_EXPR.
29522
29523 2013-01-04 Marek Polacek <polacek@redhat.com>
29524
29525 PR middle-end/55859
29526 * opts.c (default_options_optimization): Clarify error message.
29527
29528 2013-01-04 Richard Biener <rguenther@suse.de>
29529
29530 PR middle-end/55863
29531 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
29532 reassociation.
29533
29534 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29535
29536 PR target/53789
29537 * config/pa/pa.md (movsi): Revert previous change.
29538 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
29539 references.
29540
29541 2013-01-03 Richard Henderson <rth@redhat.com>
29542
29543 * config/i386/i386.c (ix86_expand_move): Always assign to op1
29544 after eliminating TLS symbols.
29545
29546 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
29547
29548 PR bootstrap/50167
29549 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
29550 * graphite-poly.c (debug_gmp_value): Likewise.
29551
29552 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
29553
29554 PR target/55712
29555 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
29556 selected code model, define __code_mode_small__, __code_model_medium__,
29557 __code_model_large__, __code_model_32__ or __code_model_kernel__.
29558 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
29559 xchg temporary register with %k. Declare temporary register as
29560 early clobbered.
29561 [__x86_64__]: For medium and large code models, preserve %rbx register.
29562
29563 2013-01-03 Richard Biener <rguenther@suse.de>
29564
29565 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
29566 (dump_subscript): Adjust.
29567 (finalize_ddr_dependent): Do not dump redundant info.
29568 (analyze_siv_subscript): Adjust.
29569 (subscript_dependence_tester): Likewise.
29570 (compute_affine_dependence): Likewise.
29571
29572 2013-01-03 Richard Biener <rguenther@suse.de>
29573
29574 Revert
29575 2013-01-03 Richard Biener <rguenther@suse.de>
29576
29577 PR tree-optimization/55857
29578 * tree-vect-stmts.c (vectorizable_load): Do not setup
29579 re-alignment for invariant loads.
29580
29581 2013-01-02 Richard Biener <rguenther@suse.de>
29582
29583 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29584 invariant load do not generate a vector load from the scalar location.
29585
29586 2013-01-03 Richard Biener <rguenther@suse.de>
29587
29588 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
29589 for not vectorizing.
29590 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
29591 not build INDIRECT_REFs, call get_name once only.
29592 (vect_create_data_ref_ptr): Likewise. Dump base object kind
29593 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
29594
29595 2013-01-03 Richard Biener <rguenther@suse.de>
29596
29597 PR tree-optimization/55857
29598 * tree-vect-stmts.c (vectorizable_load): Do not setup
29599 re-alignment for invariant loads.
29600
29601 2013-01-03 Richard Biener <rguenther@suse.de>
29602
29603 PR lto/55848
29604 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
29605 prefer a built-in decl.
29606
29607 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29608
29609 * gcc.c (process_command): Update copyright notice dates.
29610 * gcov.c (print_version): Likewise.
29611 * gcov-dump.c (print_version): Likewise.
29612
29613 PR rtl-optimization/55838
29614 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
29615 iv0.step, iv1.step and step.
29616
29617 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29618 Marc Glisse <marc.glisse@inria.fr>
29619
29620 PR tree-optimization/55832
29621 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
29622 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
29623 integer_{one,zero}_node.
29624
29625 2013-01-03 Jakub Jelinek <jakub@redhat.com>
29626
29627 PR debug/54402
29628 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
29629 * var-tracking.c (reverse_op): Don't add reverse ops to
29630 VALUEs that have already
29631 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
29632
29633 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
29634
29635 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
29636
29637 2013-01-02 Teresa Johnson <tejohnson@google.com>
29638
29639 * dumpfile.c (dump_loc): Print filename with location.
29640 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
29641 new location_t parameter to emit complete unroll message with
29642 new dump framework.
29643 (canonicalize_loop_induction_variables): Compute loops location
29644 and pass to try_unroll_loop_completely.
29645 * loop-unroll.c (report_unroll_peel): New function.
29646 (peel_loops_completely): Use new dump format with location
29647 for main dumpfile message, and invoke report_unroll_peel on success.
29648 (decide_unrolling_and_peeling): Ditto.
29649 (decide_peel_once_rolling): Remove old dumpfile message subsumed
29650 by report_unroll_peel.
29651 (decide_peel_completely): Ditto.
29652 (decide_unroll_constant_iterations): Ditto.
29653 (decide_unroll_runtime_iterations): Ditto.
29654 (decide_peel_simple): Ditto.
29655 (decide_unroll_stupid): Ditto.
29656 * cfgloop.c (get_loop_location): New function.
29657 * cfgloop.h (get_loop_location): Declare.
29658
29659 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29660
29661 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
29662 NULL.
29663
29664 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
29665
29666 PR middle-end/55198
29667 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
29668 BLKmode objects when EXPAND_MEMORY is specified.
29669
29670 2013-01-02 Sriraman Tallam <tmsriram@google.com>
29671
29672 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
29673 in loop predicate.
29674 (fold_builtin_cpu): Do not share cpu model decls across statements.
29675
29676 2013-01-02 Jason Merrill <jason@redhat.com>
29677
29678 PR c++/55804
29679 * tree.c (build_array_type_1): Revert earlier change.
29680
29681 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
29682
29683 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
29684 "cortex-a57".
29685 * config/aarch64/aarch64-tune.md: Re-generate.
29686
29687 2013-01-02 Richard Biener <rguenther@suse.de>
29688
29689 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
29690 invariant load do not generate a vector load from the scalar location.
29691
29692 2013-01-02 Richard Biener <rguenther@suse.de>
29693
29694 PR bootstrap/55784
29695 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
29696 * configure: Regenerate.
29697
29698 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29699
29700 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
29701 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
29702 (expand_builtin_int_roundingfn_2): Keep the original target around
29703 for the fallback case.
29704
29705 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
29706
29707 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
29708 to be clear for sign changes.
29709
29710 2013-01-01 Jan Hubicka <jh@suse.cz>
29711
29712 * ipa-inline-analysis.c: Fix formatting.
29713
29714 2013-01-01 Jakub Jelinek <jakub@redhat.com>
29715
29716 PR tree-optimization/55831
29717 * tree-vect-loop.c (get_initial_def_for_induction): Use
29718 gsi_after_labels instead of gsi_start_bb.
29719 \f
29720 Copyright (C) 2013 Free Software Foundation, Inc.
29721
29722 Copying and distribution of this file, with or without modification,
29723 are permitted in any medium without royalty provided the copyright
29724 notice and this notice are preserved.